UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pre-commercial thinning and repeated fertilization of young lodgepole pine stands : long-term impacts… Lindgren, Pontus Maurits Fredrik 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_spring_lindgren_pontus.pdf [ 3.24MB ]
Metadata
JSON: 24-1.0165656.json
JSON-LD: 24-1.0165656-ld.json
RDF/XML (Pretty): 24-1.0165656-rdf.xml
RDF/JSON: 24-1.0165656-rdf.json
Turtle: 24-1.0165656-turtle.txt
N-Triples: 24-1.0165656-rdf-ntriples.txt
Original Record: 24-1.0165656-source.json
Full Text
24-1.0165656-fulltext.txt
Citation
24-1.0165656.ris

Full Text

PRE-COMMERCIAL THINNING AND REPEATED FERTILIZATION OF YOUNG LODGEPOLE PINE STANDS: LONG-TERM IMPACTS ON TREE GROWTH, PLANT DIVERSITY, AND RANGE  by Pontus Maurits Fredrik Lindgren  B.Sc., University of British Columbia, 1995 M.Sc., University of British Columbia, 2000  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF  Doctor of Philosophy in THE FACULTY OF GRADUATE AND POSTGRADUATE STUDIES (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  October 2013  ? Pontus Maurits Fredrik Lindgren, 2013     ii  Abstract The 16-year research program studied the impacts of pre-commercial thinning (PCT) and fertilization on tree growth, plant diversity, forage production, and ungulate habitat use.  Ecological effects of cattle grazing were also studied.   Three study areas were located in south-central British Columbia, Canada.  Each study area was comprised of densely stocked stands of young lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) thinned to 250, 500, 1000, and 2000 stems/ha.  Half of each thinned unit was fertilized five times over 10 years.  An unthinned stand completed the experimental design.  Fifteen-year increments of crop-tree diameter at breast height (DBH) as well as tree and stand volume were enhanced by fertilization, but not by density level.  Non crop trees became abundant, particularly within the heavily thinned stands.  Fertilization enhanced herb abundance but decreased herb and shrub diversity, at least temporarily.  Plant species richness was unaffected by density levels or fertilization.  Structural diversity of the tree layer was greater in the heavily than lightly thinned stands, and overall structural diversity was enhanced by fertilization.  Forage yield was enhanced by fertilization, but only within heavily thinned stands.  Fertilization also resulted in increased crude protein levels of pinegrass (Calamagrostis rubescens Buckley), an important forage for cattle (Bos taurus L.) for up to six years after the final fertilization.  Habitat use by cattle was enhanced by both fertilization and PCT, and this increased use did not appear to have any negative impacts on the native ungulates using these stands.  Relative to ungrazed plots, cattle grazing increased both richness and diversity of herbs and shrubs within fertilized stands, whereas grazing decreased shrub richness within unfertilized stands. Results indicated that increased forest and range productivity is possible with treatments of repeated fertilization and, to a lesser extent, PCT.  Ecological effects and management implications of these treatments are discussed.      iii  Preface  Study design, statistical analysis, interpretation, and write-up for all chapters in this dissertation were performed by Pontus Lindgren, under the supervision of Dr. T.P. Sullivan.  Versions of all five research chapters are in various states of publication (all with Dr. Sullivan as co-author).  Those that have been published or accepted for publication are: ? Chapter 2 ? In press o Lindgren, P.M.F. and Sullivan, T.P.  2013.  Long-term responses of tree and stand growth of young lodgepole pine to pre-commercial thinning and repeated fertilization.  Forest Ecol. Manage.   ? Chapter 3 ? In press o Lindgren, P.M.F. and Sullivan, T.P.  2013.  Influence of stand thinning and repeated fertilization on plant community abundance and diversity in young lodgepole pine stands: 15-year results. Forest Ecol. Manage. ? Chapter 6 ? Published o Lindgren, P.M.F. and Sullivan, T.P.  2012.  Response of plant community abundance and diversity during 10 years of cattle exclusion within silvopasture systems. Can. J. For. Res. 42: 451?462.  In addition, significant contributions to this thesis were provided by Rob Brockley1 and Dr. Don Thompson2 who assisted with laboratory analyses of crop tree foliar samples (reported in Chapter 2) and forage samples (reported in Chapter 4), respectively.                                                      1 Research Scientist: Kalamalka Research Station, Ministry of Forests and Range, 3401 Reservoir Road, Vernon, British Columbia, Canada V1B 2C7 2 Range Ecologist: Agriculture and Agri-Food Canada, Range Research Unit, 3015 Ord Road, Kamloops, BC, Canada V2B 8A9    iv  Table of contents Abstract ........................................................................................................................................... ii Preface ........................................................................................................................................... iii Table of contents ........................................................................................................................... iv List of tables ................................................................................................................................ viii List of figures ............................................................................................................................... xii Acknowledgements ..................................................................................................................... xvii Dedication.................................................................................................................................. xviii 1 General introduction .............................................................................................................. 1 1.1 Thesis outline ............................................................................................................................. 1 1.2 Background ................................................................................................................................ 2 1.3 Study hypotheses ....................................................................................................................... 5 1.4 Methods ...................................................................................................................................... 7 1.4.1 Study areas ............................................................................................................................................. 7 1.4.2 Experimental design ............................................................................................................................ 17 1.4.3 Treatments ........................................................................................................................................... 17 1.4.4 Data collection and calculations .......................................................................................................... 21 1.4.5 Statistical analysis ................................................................................................................................ 22 2 Long-term responses of tree and stand growth of young lodgepole pine to pre-commercial thinning and repeated fertilization .............................................................................................. 24 2.1 Introduction ............................................................................................................................. 24 2.2 Methods .................................................................................................................................... 26 2.2.1 Study areas ........................................................................................................................................... 26 2.2.2 Experimental design ............................................................................................................................ 26 2.3 Results ....................................................................................................................................... 30 2.3.1 Tree density ......................................................................................................................................... 30 2.3.2 Fertilizer ............................................................................................................................................... 36 2.3.3 Growth of crop trees ............................................................................................................................ 37    v  2.4 Discussion ................................................................................................................................. 54 2.4.1 Height class distribution ...................................................................................................................... 54 2.4.2 Crop trees ............................................................................................................................................. 57 2.4.3 Carbon sequestration and resilience to climate change ........................................................................ 61 2.5 Conclusions .............................................................................................................................. 63 3 Plant community diversity and structural response to pre-commercial thinning and repeated fertilization of young lodgepole pine stands: 15 years of monitoring ......................... 65 3.1 Introduction ............................................................................................................................. 65 3.2 Methods .................................................................................................................................... 67 3.2.1 Study areas ........................................................................................................................................... 67 3.2.2 Experimental design ............................................................................................................................ 69 3.3 Results ....................................................................................................................................... 74 3.3.1 Abundance ........................................................................................................................................... 74 3.3.2 Species diversity .................................................................................................................................. 79 3.3.3 Structural diversity ............................................................................................................................... 84 3.3.4 Productivity-diversity relationship ....................................................................................................... 94 3.4 Discussion ............................................................................................................................... 103 3.4.1 Abundance ......................................................................................................................................... 103 3.4.2 Species diversity ................................................................................................................................ 106 3.4.3 Structural diversity ............................................................................................................................. 108 3.4.4 Stand productivity and diversity relationships ................................................................................... 109 3.5 Conclusions ............................................................................................................................ 112 4 Forage yield and quality response to pre-commercial thinning and repeated fertilization of young lodgepole pine stands: implications for silvopasture management .......................... 115 4.1 Introduction ........................................................................................................................... 115 4.2 Methods .................................................................................................................................. 117 4.2.1 Study areas ......................................................................................................................................... 117 4.2.2 Experimental design .......................................................................................................................... 118 4.3 Results ..................................................................................................................................... 121 4.3.1 Canopy closure .................................................................................................................................. 121 4.3.2 Yield .................................................................................................................................................. 122 4.3.3 Crude protein (CP) ............................................................................................................................. 126    vi  4.3.4 Total crude protein (TCP) .................................................................................................................. 126 4.3.5 Acid detergent fiber (ADF) ................................................................................................................ 127 4.3.6 Relationships between canopy closure and forage attributes ............................................................. 127 4.4 Discussion ............................................................................................................................... 130 4.4.1 Canopy closure .................................................................................................................................. 130 4.4.2 Yield .................................................................................................................................................. 130 4.4.3 Quality (CP and ADF) ....................................................................................................................... 133 4.4.4 Implications for cattle grazing ........................................................................................................... 135 4.5 Conclusions ............................................................................................................................ 136 5 Relative habitat use by cattle and mule deer in response to forest thinning and fertilization ................................................................................................................................. 138 5.1 Introduction ........................................................................................................................... 138 5.2 Methods .................................................................................................................................. 142 5.2.1 Study areas ......................................................................................................................................... 142 5.2.2 Experimental design .......................................................................................................................... 143 5.3 Results ..................................................................................................................................... 146 5.3.1 Relative habitat use ............................................................................................................................ 146 5.3.2 Cattle impact on mule deer ................................................................................................................ 149 5.3.3 Overstory trees and winter habitat use by mule deer ......................................................................... 153 5.4 Discussion ............................................................................................................................... 154 5.4.1 Cattle and forest management ............................................................................................................ 154 5.4.2 Cattle impact on mule deer ................................................................................................................ 156 5.5 Conclusions ............................................................................................................................ 158 6 Response of plant community abundance and diversity during 10 years of cattle exclusion within managed stands of young lodgepole pine ...................................................................... 160 6.1 Introduction ........................................................................................................................... 160 6.2 Methods .................................................................................................................................. 162 6.2.1 Study areas ......................................................................................................................................... 162 6.2.2 Experimental design .......................................................................................................................... 162 6.3 Results ..................................................................................................................................... 166 6.3.1 Lack of pre-treatment exclosure data ................................................................................................. 166    vii  6.3.2 Herb layer .......................................................................................................................................... 166 6.3.3 Shrub layer ......................................................................................................................................... 171 6.3.4 Tree layer ........................................................................................................................................... 172 6.4 Discussion ............................................................................................................................... 173 6.4.1 Abundance of vegetation ................................................................................................................... 173 6.4.2 Species richness and diversity ........................................................................................................... 175 6.5 Conclusions ............................................................................................................................ 176 7 General conclusions ........................................................................................................... 178 7.1 Study strengths and limitations ............................................................................................ 178 7.2 Future opportunities.............................................................................................................. 180 7.3 Contributions to forest and range management ................................................................. 181 7.4 General conclusions and management implications ........................................................... 182 References .................................................................................................................................. 186       viii  List of tables Table 1.1.  Split-plot allocation of density and fertilizer treatments and stand descriptions by replicate study area (blocks). ...................................................................................................................................................... 11 Table 2.1.  Summary of split-split plot ANOVA investigating the effects of PCT density (main plot; four levels; target of 250, 500, 1000, and 2000 crop trees/ha), fertilizer (split plot; two levels; unfertilized and repeatedly fertilized), and time (split-split plot; three levels; 1998, 2003, and 2008) on density of all trees among four height classes. ..................................................................................................................................... 31 Table 2.2.  Adjusted mean (n = 2 replicate study areas) (SE) 15-year increments (1993 to 2008) of crop tree attributes for young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts...................................................................................................................................................... 39 Table 2.3.  Adjusted mean (n = 2 replicate study areas) (SE) change in growth rate observed during the final 10 years for crop tree attributes of young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts........................................................................................................................................ 40 Table 2.4.  Adjusted mean (n = 2 replicate study areas) (SE) height-to-diameter ratio (cm:cm) of crop tree attributes for young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. ..... 43 Table 2.5.  Mean (n = 2 replicate study areas) (SE) crown area and volume of young lodgepole pine crop trees among four density and two fertilization treatments measured in 2008; 15 years after the onset of treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. ............................................................................................................... 51 Table 3.1.  Mean (n = 3 replicate study areas) (SE) crown volume index (m3/0.01 ha) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The overall sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction was observed for this overall analysis, the 15-year post-PCT period was split into three five-year periods, and assessed separately.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts........................................................................................................................................ 76 Table 3.2.  Mean (n = 3 replicate study areas) (SE) species richness (number of species) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.     ix  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods. ..................................................... 80 Table 3.3.  Mean (n = 3 replicate study areas) (SE) species diversity (Simpson index) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. ..... 81 Table 3.4.  Mean (n = 3 replicate study areas) (SE) structural richness (number of vegetation layers) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilizer effects, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts........................................................................................................................................ 87 Table 3.5.  Mean (n = 3 replicate study areas) (SE) structural diversity (Simpson index) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. .................................................................................................................................................................... 90 Table 3.6.  Summary of linear and quadratic models for mean total species richness regressed on mean total crown volume index for two different time periods and study areas.  Pattern of relationships are described as N (negative linear), H (hump-shaped quadratic), or NS (not significant). ..................................................... 95 Table 3.7. Summary of linear and quadratic models for mean total species diversity (Simpson index) regressed on mean total crown volume index for two different time periods and study areas.  Pattern of relationships are described as N (negative linear), H (hump-shaped quadratic), U (U-shaped), or NS (not significant). .................................................................................................................................................................... 98    x  Table 3.8.  Summary of linear models for mean total structural diversity (Simpson index) regressed on mean total crown volume index by Summerland, Kelowna, and Cariboo study areas and sample years.  Patterns of relationships (significant models indicated in bold text) are described as P (positive linear), N (negative linear), or NS (not significant). ................................................................................................................. 101 Table 3.9. Summary of quadratic models for mean total structural diversity (Simpson index) regressed on mean total crown volume index by Summerland, Kelowna, and Cariboo study areas and sample years.  Patterns of relationships (significant models indicated in bold text) are described as H (hump-shaped), U (U-shaped), or NS (not significant). ............................................................................................................... 101 Table 4.1.  Mean (n = 2 replicate study areas) (SE) percent canopy closure (CC) by density and fertilizer treatment for lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Associated split-plot ANOVA results are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted for multiple contrasts, if necessary (lower and uppercase letters for density and fertilizer effects, respectively). ...... 122 Table 4.2.  Mean (n = 2 replicate study areas) (SE) yield (oven-dried kg/ha), percent crude protein (CP), total crude protein (TCP; kg/ha), and percent acid detergent fiber (ADF) for various forage components by density and fertilizer treatment in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Associated split-plot ANOVA results are also provided for each attribute.  Within a row, mean values with different letters were significantly different by Duncan?s multiple range test, adjusted for multiple contrasts, if necessary (lower and uppercase letters for density and fertilizer effects, respectively). ................................................................................................................................. 125 Table 4.3.  Relationship between percent canopy closure (CC) and several forage attributes observed among 26- to 28-year-old lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer treatment.  Only significant (ANOVA; P=0.05) relationships were presented.  Data were pooled across the two study areas (Summerland and Kelowna) only if models were statistically similar among the two areas.  If models were different (i.e., different intercepts and/or slopes), relationships were only presented if both models had the same form. ............................................................................................................ 127 Table 5.1.  Mean (n = 2 replicate study areas) (SE) indices of relative habitat use c by cattle among five levels of density and two levels of fertilization treatments sampled over six years.  Split-plot ANOVA results are also provided.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. ................................................................................... 148 Table 5.2.  Relationship between relative habitat use by mule deer (y; inferred by number of pellet-groups/ha) and cattle (x; inferred by proportion of sample plots with cowpies present) among nine young, lodgepole pine treatment stands.  Cattle were present only during summers; therefore, the summer models represent concurrent use by these two herbivores, whereas the winter models represent the relationship between habitat use by cattle during the summer and the use of that same habitat by mule deer the following    xi  winter.  Linear (y = Const. + B1x) and quadratic (y = Const. + B1x + B2x2) models were explored; however, only the best fit (most significant) were presented. ................................................................... 151 Table 5.3.  Relationship between density of trees taller than 3 m (x; trees/ha) and relative habitat use by mule deer (y; inferred by number of pellet-groups/ha) among nine young, lodgepole pine treatment stands over five winters.  Linear (y = Const. + B1x) and quadratic (y = Const. + B1x + B2x2) models were explored; however, only the best fit (most significant) were presented. ................................................................... 153 Table 6.1.  Summary of split-plot ANOVA results for plant community response (crown volume index, species richness, and Shannon-Wiener species diversity) to repeated fertilization and cattle grazing treatments (grazed vs. 10 years of cattle exclusion).  Unless otherwise specified, effects correspond to the entire 10-year period of this study (1994-2003).  Significant effects are indicated by bold text. ............................. 168       xii  List of figures Figure 1.1.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Summerland study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 10700 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE. ....................................................... 9 Figure 1.2.  Summerland study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand. ................................................................. 10 Figure 1.3.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Kelowna study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 4700 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE. ..................................................... 12 Figure 1.4.  Kelowna study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand. .................................................................................. 13 Figure 1.5.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Cariboo study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 2915 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE. ..................................................... 15 Figure 1.6.  Cariboo study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand. .................................................................................. 16 Figure 1.7.  Gantt chart depicting the timing of significant events (treatments, samples, and disturbances) that have occurred during the initial 16 years of this research project (1993 to 2008). ............................................ 20 Figure 2.1.  Mean total density (stems/ha) of trees within the dominant tree layer (taller than 3 m) among the four density treatments measured at five-year intervals; 1998, 2003, and 2008.  Densities include crop trees, as well as other trees that grew into this upper height class with time.  Numbers along the x-axis represent target PCT densities (stems/ha).  Error bars indicate SE and are based on two replicate study areas (n = 2). .............................................................................................................................................. 33 Figure 2.2.  Mean total tree density (stems/ha) within three understory tree layers (i.e., excluding crop trees) among the eight treatment stands sampled in 1998, 2003, and 2008.  Numbers along the x-axis represent target PCT densities (crop trees/ha) and the ?F? denotes stands that were repeatedly fertilized.  Error bars indicate SE and are based on two replicate study areas (n = 2). ............................................................... 35 Figure 2.3.  Mean foliar N concentration for young lodgepole pine crop trees within unfertilized and repeatedly fertilized stands measured annually from 1995 to 2003.  Values were averaged across all densities and two replicate study areas (n = 2).  Error bars indicate SE......................................................................... 36    xiii  Figure 2.4.  Adjusted mean 15-year increments (1993 to 2008) of a) height and b) DBH of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). ............................................. 38 Figure 2.5.  Adjusted mean height-to-diameter ratio (HDR; cm:cm) of young lodgepole pine crop trees among nine treatment stands in 2008; 15 years after onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). .... 42 Figure 2.6.  Adjusted mean 15-year increments (1993 to 2008) of a) tree and b) stand BA of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). ............................................. 45 Figure 2.7.  Adjusted mean 15-year increments (1993 to 2008) of a) tree and b) stand volume of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). ............................ 47 Figure 2.8.  Mean (n = 2 replicate sites) crop tree crown dimensions as measured at five-year intervals (1998, 2003, and 2008) among the nine treatment stands. .............................................................................................. 49 Figure 2.9.  Mean crown area of young lodgepole pine crop trees at the a) tree and b) stand levels among nine treatment stands measured in 2008; 15 years following the onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). ........................................................................................................................................................ 50 Figure 2.10.  Mean crown volume of young lodgepole pine crop trees at the a) tree and b) stand levels among nine treatment stands measured in 2008; 15 years following the onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2). ........................................................................................................................................................ 53 Figure 3.1.  Mean (n = 3) crown volume index (m3/0.01 ha) of a) herbs and b) shrubs by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. ............................................................................................................................................. 77    xiv  Figure 3.2.  Mean (n = 3) crown volume index (m3/0.01 ha) of the tree layer by a) density and b) fertilizer treatments from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. ..................................................................................................................... 78 Figure 3.3.  Mean (n = 3) species diversity (Simpson index) of the a) herb and b) shrub layers from 1993 to 2008 by fertilizer treatment.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. .................................................................................................................................... 82 Figure 3.4.  Mean (n = 3) species diversity (Simpson index) of the a) tree and b) total layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. .................................................................................................................................... 83 Figure 3.5.  Mean (n = 3) structural richness (number of vegetation layers) of the tree layer by density treatment, from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. .................................................................................................................................... 85 Figure 3.6.  Mean (n = 3) structural richness (number of vegetation layers) of the a) herb, b) shrub, and c) tree layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. ....................................................................................... 88 Figure 3.7.  Mean (n = 3) structural diversity (Simpson index) of the a) tree and b) total (herbs, shrubs, and trees combined) layers by density treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization. ........................................................... 91 Figure 3.8.  Mean (n = 3) structural diversity (Simpson index) of the a) herb, b) shrub, and c) tree, and total (herbs, shrubs, and trees combined) layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization................................ 93 Figure 3.9.  Relationship (or lack thereof) between species richness and an index of stand productivity (total crown volume index; m3/0.01 ha) among managed stands of young lodgepole pine.  Relationships were assessed for two separate datasets (a) Summerland and Kelowna study areas combined and b) Cariboo study area) and two sample periods (initial six, and final six sample years).  Data points represent mean values for a given year from each of nine treatment stands.  Trendlines and R2 values are presented for significant regression models. ...................................................................................................................................... 96 Figure 3.10.  Relationship (or lack thereof) between species diversity (Simpson index) and an index of stand productivity (total crown volume index; m3/0.01 ha) among managed stands of young lodgepole pine.  Relationships were assessed for two separate datasets (a) Summerland and Kelowna study areas combined and b) Cariboo study area) and two sample periods (initial six, and final six sample years).  Data points represent mean values for a given year from each of nine treatment stands.  Trendlines and R2-values are presented for significant regression models. ........................................................................ 99 Figure 3.11.  Example of relationships (or lack thereof) between total structural diversity (Simpson index) and an index of total stand productivity (crown volume index; m3/0.01 ha) for six of the 12 sample years at the Summerland study area.  Data points represent mean values for each of nine treatment stands.  PCT target densities (250, 500, 1000, 2000 stems/ha and unthinned) are indicated for reference only as    xv  regression analyses applied to all treatment stands.  Trendlines and R2 values are only presented for significant regression models. .................................................................................................................. 102 Figure 4.1.  Mean percent canopy closure (CC) in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Error bars indicate SE and are based on two replicate study areas (n = 2). ...................................................................................................................................................... 122 Figure 4.2.  Mean yield (oven-dried kg/ha) of a) pinegrass, b) total grasses, and c) herbs (grasses and forbs combined) in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Error bars indicate SE and are based on two replicate study areas (n = 2).  Columns with different letters were significantly (P=0.05) different (lowercase and uppercase letters indicate density and fertilizer effects, respectively) ............................................................................................................ 124 Figure 4.3.  Relationship between pinegrass yield (oven-dried kg/ha) and percent canopy closure in 26- to 28-year-old lodgepole pine stands 14 years post PCT and five years following the most recent fertilizer treatment at Summerland and Kelowna. ................................................................................................................... 128 Figure 4.4.  Relationship between percent crude protein (CP) and percent canopy closure for pinegrass and herbs in 26- to 28-year-old lodgepole pine stands 14 years post PCT and five years following the most recent fertilizer treatment.  Data were pooled across Summerland and Kelowna. ............................................. 129 Figure 4.5.  Relationship between percent acid detergent fiber (ADF) for grasses and percent canopy closure in 26- to 28-year-old lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer treatment.  Data were pooled across Summerland and Kelowna. ............................................................ 129 Figure 5.1.  Mean index of relative habitat use for cattle (inferred from cowpie samples) among nine young, lodgepole pine stands.  Error bars represent SE and are based on two replicate study areas (n = 2).  Data were collected annually over six years. .................................................................................................... 147 Figure 5.2.  Mean index of relative habitat use for mule deer (inferred from pellet-group counts) among nine, young, lodgepole pine treatment stands during a) summer and b) winter periods.  Error bars represent SE and are based on three replicate study areas (n = 3).  Data were collected annually over five years. ........... 150 Figure 5.3.  Relationship between habitat use by cattle in summer 1999 (inferred from proportion of sample plots with cowpies) with habitat use by mule deer (inferred from density of pellet-groups/ha) during the following winter (1999-2000) at a) Summerland and b) Cariboo study areas. ........................................ 152 Figure 5.4.  Relationship between density of tall (> 3 m) trees and winter habitat use by deer during the winter of 2002/2003 (inferred from density of pellet-groups) at Summerland. ........................................................ 154 Figure 5.5.  Relationship between density of tall (> 3 m) trees and winter habitat use by deer during the winters of 1999/2000, 2000/2001, and 2001/2002 (inferred from density of pellet-groups) at Cariboo................... 154 Figure 6.1.  Mean (n = 2) plant community attributes for herbs, shrubs, and trees by fertilization and cattle grazing treatments (averaged across stand densities and years).  Attributes are a) crown volume index, b) species richness, and c) Shannon-Wiener species diversity index.  Differences among cattle grazing treatments (i.e., open range vs. cattle excluded) are indicated by asterisks; * or ** for marginally significant (0.05 < P < 0.10) and significant (P ? 0.05) effects, respectively.  Error bars represent SE. .............................. 169    xvi  Figure 6.2.  Mean (n = 2) attributes of the herb layer in repeatedly fertilized stands by sample year and cattle grazing treatments (averaged across stand densities). Attributes are a) crown volume index, b) species richness, and c) Shannon-Wiener species diversity index. Differences among cattle grazing treatments (i.e., open range vs. cattle excluded) for a given period are indicated by asterisks; * or ** for marginally significant (0.05 < P < 0.10) and significant (P ? 0.05) effects, respectively.  Error bars represent SE. 170       xvii  Acknowledgements I would like to thank the Silviculture Branch, BC Ministry of Forests (MoF), Victoria, BC, the Canada-British Columbia Partnership Agreement on Forest Resource Development (FRDA II) for financial support during the first four years of the project, and Forest Renewal BC through the BC Science Council, Forest Innovation Investment, Gorman Bros. Lumber Ltd., Tolko Industries Ltd. (formerly Riverside Forest Products Ltd.) and the Alex Fraser Research Forest, University of British Columbia for support throughout.  Operational treatments were conducted by the Silviculture sections of Penticton and Horsefly Forest Districts (MoF).   I would like to thank my committee members M. Feller and P. Marshall for their guidance and constructive criticism over the duration of this dissertation.  A very special thanks to my supervisor, T. Sullivan, who has been, and will continue to be, a mentor that leads by example, and the driving force behind all of my professional accomplishments.  I would also like thank A. Kozak and L. Zabek for their statistical advice. I owe a huge debt of gratitude to the army of personnel that have contributed to the 16 years of fieldwork on which this thesis is based.  For their assistance with tree measurements, pellet sampling, and data recording, I would like to thank N. Handford, J. Hickson, C. Houwers, C. Kohler, S. Lang, T. Lindgren, F. Moreau, C. Nowotny,  D. Ransome, B. Runciman, H. Sullivan,  B. Tilling, and L. Zabek.  I apologize if I have missed anyone.  For their assistance with the forage study (setup, sampling, and laboratory analysis), I would like to thank L. Liggins, D. Ransome, D. Thompson, D. Trotter, B. Wheatley, and L. Zabek. For his tireless work ethic that I can only hope has partially rubbed off on me, I would like to thank my dad.  For keeping me grounded and encouraging me to always think critically, I would like to thank my mom.  But my biggest thanks goes to my beautiful wife, Paula.  She has been supportive and patient for all these years and, as the smarter half of our duo, has also contributed to the project in many ways.  Countless proofreads of earlier drafts, many hours of data entry, and even a few days of counting pellets (which she has indicated that she will not be volunteering for again).       xviii  Dedication  This thesis is dedicated to my two perfect boys, Tristan and Paxton.  You have been the primary motivation for getting through the toughest parts of this journey.  Thank you for keeping my spirits high and for always being a welcomed reminder of what is truly important.        1  1 General introduction 1.1 Thesis outline This thesis is based on 16 years of research that has studied the effects of intensive management treatments of pre-commercial thinning (PCT) and repeated fertilization of young lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) stands on a wide range of ecosystem attributes, including tree growth, plant diversity, forage production, and cattle (Bos taurus L.).  The impact of cattle on the plant community and relative habitat use by mule deer (Odocoileus hemionus Raf.) is also addressed. This general introductory chapter provides an overview of the thesis? subject matter, identifies knowledge gaps, and presents the hypotheses used to address the objectives of the thesis.  A detailed description of study areas and treatments is provided that will be referenced throughout the thesis.  Finally, a brief description of the experimental design, sampling methods, and statistical analyses is provided, which will be described in detail within the relevant research chapters that follow. The principal questions addressed by the research chapters are as follows: ? Can we develop a deeper understanding of the effects of PCT and repeated fertilization on: o tree growth (Chapter 2),  o abundance, species diversity, and structural diversity of plants (Chapter 3),  o forage (Chapter 4), and  o habitat use by cattle (Chapter 5)? ? Similarly, about the effects of cattle grazing on: o mule deer (Chapter 5) and  o abundance, species diversity, and structural diversity of plants (Chapter 6)? The general concluding chapter (Chapter 7) provides a summary of the results and the implications of these findings for both forest and range management.       2  1.2 Background  With very few exceptions, Canada?s forest industry has been based on extensive silviculture practices that have taken advantage of large tracts of unmanaged, old forests, often with little or no investment in silviculture treatments beyond those required for stand establishment.  Intensive silviculture, including such treatments as thinning and fertilization, has been demonstrated to be an economically beneficial strategy within plantations throughout the southern United States (US; Albaugh et al. 2004; Jokela et al. 2004) and northern Europe (Nabuurs et al. 2007; Bergh et al. 2008), with growth rates more than doubled and rotation lengths cut by more than 50% (Fox et al. 2007).  However, the benefits of intensive silviculture remain largely unrealized throughout northern latitudes, and within the boreal and sub-boreal forests of North America in particular (Lautenschlager 2000; Park and Wilson 2007).  Given the prediction of increased global demand for wood products (Raunikar et al. 2010), concurrent with conservation strategies that endeavor to significantly increase protected areas (Hunter and Schmiegelow 2011) and unpredictable large-scale reductions to existing timber supplies (e.g., losses to insect epidemics; Walton et al. 2009), enhanced wood production may become increasingly necessary to mitigate current and future wood supply shortfalls (Brooks 1997; Sutton 1999).  Of all tree species in the inland areas of the Pacific Northwest (PNW) of North America, lodgepole pine likely has the greatest potential to respond favourably to silvicultural treatments such as thinning and fertilization.  Lodgepole pine often regenerates over-abundantly, resulting in excessive stand densities and reduced stand and tree growth.  By concentrating growth on a smaller number of stems, PCT offers the forest manager some control over the rotation, yield, and value of the future crop (Johnstone 1985; Cole and Koch 1996).  Extensive lodgepole pine forests are usually perpetuated through repeated fire disturbance, and therefore often occupy sites of low-N status (Brockley et al. 1992).  Not surprisingly, this species responds well to conventional, single-application fertilization with nitrogen alone, and in combination with other elements (Weetman 1988; Brockley 1996).  As a result, large gains in yield at harvest are expected by thinning and fertilizing dense stands of lodgepole pine (Farnden and Herring 2002; Lindgren et al. 2007).  Although a single fertilizer application typically produces only a temporary increase in growth, sustained growth responses to repeated fertilization have been    3  demonstrated in field experiments with lodgepole pine (Kishchuk et al. 2002; Brockley 2005) and other Pinus species (Malkonen and Kukkola 1991; Tamm et al. 1999; Albaugh et al. 2004; Sword Sayer et al. 2004).   An improved understanding of the effects that PCT and repeated fertilization have on the growth of lodgepole pine would have direct implications for several important management issues, including timber supply (Brooks 1997; Sutton 1999), biofuel production, carbon sequestration (Sedjo 1999; Raunikar et al. 2010), and resiliency to climate change (Miller et al. 2007).  While there are important benefits associated with enhanced tree growth, there are also concerns regarding potential negative impacts of intensive management on biodiversity (McDonald and Lane 2004) and, therefore, the sustainability of such practices (Nakashizuka 2007).   Intensive management is often associated with decreased diversity, particularly of plants.  However, responses have ranged from decreased (Hansen et al. 1991; Gilliam and Roberts 1995; Elliott et al. 1997) to increased plant diversity (Thomas et al. 1999; Thysell and Carey 2001; Battles et al. 2001; Lindgren et al. 2006) following intensive management.  Mechanisms such as the intermediate disturbance hypothesis (Battles et al. 2001) and the hump-shaped productivity-diversity relationship (Waide et al. 1999; Mittelbach et al. 2001) have been suggested to explain the relationship between management and diversity; however, the predictive power of these mechanisms is low.  As such, the effect of intensive management on biodiversity should be determined by site- and context-specific investigation rather than generalizations or extrapolations from other studies.   Plants are a critical component of biodiversity unto themselves; however, they also impact other components of biodiversity by modifying the environment, both above (air temperature, wind speed, humidity, and shading) and below ground (soil temperature, moisture, and nutrient content) (Berch et al. 2006) and provide much of the physical attributes of habitat for all species of wildlife (e.g., hunting grounds, cover, nesting sites, forage) (Carey et al. 1999; Sullivan et al. 2001).  I have located no studies that have investigated the long-term plant community response to treatments of PCT and repeated fertilization.  Such information would enhance our understanding of the impacts that these treatments have on both the plant community itself and other ecological attributes that are directly influenced by plant    4  composition, abundance, and structure (e.g., wildlife habitat) and provide valuable insight into the sustainability of these practices.   Cattle grazing is common throughout most inland forests of North America (McLean 1983; Bock et al. 1993; Fleischner 1994).  Following a disturbance, regenerating stands provide suitable range conditions for a limited number of years before the tree crowns expand and shade out the understory forage (Basile and Jensen 1971; Wikeem et al. 1993b).  By manipulating canopy closure and nutrient regimes, treatments of thinning and fertilization can enhance the production of both forage (Freyman and van Ryswyk 1969; Scotter 1980; Bedunah et al. 1988; Wikeem et al. 1993b) and trees (Farnum et al. 1983; Weetman et al. 1995; Sedjo 1999; Farnden and Herring 2002; Jokela et al. 2004; Lindgren et al. 2007).  Silvopastoralism (the deliberate integration of trees with forage and livestock production) has been practiced sustainably throughout the world (Knowles et al. 1973; Msika and Etienne 1989; Sibbald et al. 1989; Papanastasis et al. 1995) and with enhanced economic benefits compared to single resource management strategies (McDonald and Fiddler 1993; Clason and Sharrow 2000; Husak and Grado 2002).  However, the potential for silvopastoralism is vastly underutilized throughout the PNW (Bedunah et al. 1988) and has not been studied in lodgepole pine forests.  An improved understanding of how PCT and fertilization impact both the forage resource (quantity, quality, and longevity) and cattle habitat use would help determine the suitability of these treatments for promoting silvopastoralism within lodgepole pine stands. There are serious concerns regarding the potential negative impacts of livestock grazing on the environment, including loss of biodiversity, change in physical characteristics of habitats, and lowering of population densities (Bock et al. 1993; Fleischner 1994).  Given that rangeland management is expected to intensify to meet the growing global demands for agricultural productivity (Watkinson and Ormerod 2001; Steinfeld et al. 2006), range managers will require a better understanding of the impacts that livestock are having within grazed ecosystems if they are to be managed in a sustainable manner (Hunter 1999).  Cervids are excellent indicators of environmental impacts as they require a variety of habitats, are sensitive to changes in their environment (Hanley 1996), and have the potential to be negatively impacted by grazing cattle, particularly by decreasing the value of critical winter range habitats (Darambazar et al. 2003).  By studying diet overlap between cattle and deer, several studies have investigated the potential    5  for competition among these herbivores (Stuth and Winward 1977; Willms et al. 1980b; Bedunah et al. 1988; Gibbs et al. 2004; Findholt et al. 2005).  However, I am not aware of any studies that have investigated impacts of cattle on deer within lodgepole pine stands, as inferred by patterns of relative habitat use.  An improved understanding of the interactions between cattle and the mule deer that coexist in lodgepole pine stands would provide valuable information for determining the suitability and sustainability of treatments that increase habitat use by cattle.  Studies investigating the impact of livestock grazing on plant community composition have found a range of responses, from an increase in exotic plant species richness (McIntyre et al. 2003), to no change (Stohlgren et al. 1999), to an increase in native species richness (Germano et al. 2001).  In addition, plant community responses to grazing disturbance can be very long-lived and require even longer periods to recover (Valone et al. 2001; Courtois et al. 2004).  Therefore, it is unreasonable to expect to be able to accurately predict the consequences of grazing disturbance on plant community diversity without long-term monitoring of a specific management regime.  I am not aware of any investigation that has compared the long-term effects of both cattle grazing and cattle exclusion on plant communities within lodgepole stands.  Therefore, an improved understanding of the impacts that cattle grazing has on plant communities (abundance, species diversity, structural diversity) would provide valuable information for determining the sustainability of range management within lodgepole pine forests.   1.3 Study hypotheses No large-scale or long-term studies have investigated the environmental effects of PCT and repeated fertilization of lodgepole pine stands.  Therefore, the primary objective of this thesis was to enhance our understanding of the effects that these silviculture treatments have on several important ecosystem components.  Specifically, this thesis investigated the effects of PCT and fertilization on growth of lodgepole pine trees and stands, plant abundance, plant diversity, forage quality, and cattle habitat use.  A secondary objective of this thesis was to study the effects that cattle grazing had on habitat use by native ungulates and the plant community.  The chapters and specific hypotheses (phrased as predictions) used to address these objectives are as follows:    6  ? Chapter 2: lodgepole pine growth o H1: Application of a range of PCT intensities and repeated fertilization with optimum nutrition formulations will enhance the 15-year growth increments of lodgepole pine crop trees at both tree and stand levels. ? Chapter 3:  response of the plant community o H1: PCT will increase abundance of the herb and shrub layers. o H2: Fertilization treatments will increase abundance of all plant layers (herbs, shrubs, trees, and combined total). o H3: PCT and fertilization treatments will decrease species and structural diversity of all plant layers. o H4: Enhanced abundance of vegetation will result in a decline in total species diversity. ? Chapter 4: response of forage o H1: PCT thinning will enhance both the yield and quality of forage. o H2: Repeated fertilization treatments of thinned stands will further enhance both the yield and quality of forage. ? Chapter 5: habitat use by cattle and their impact on native ungulates  o H1: PCT and repeated fertilization will increase relative habitat use by cattle. o H2: Increased use by cattle will decrease use by native ungulates. ? Chapter 6: response of plant community to cattle exclusion o H1: Abundance of understory plants (herbs and shrubs) will increase in response to cattle exclusion. o H2: Cattle grazing will increase plant species richness and diversity. In summary, the aim of my thesis was to improve understanding of the potential gains and consequences of intensive management prescriptions of PCT and repeated fertilization of lodgepole pine stands.  The results and conclusions of the following chapters should enhance our ability to sustainably manage lodgepole pine forests, as well as integrate silvicultural objectives with those of range management and biodiversity conservation. This study is part of a larger research program that has been investigating the effects of PCT and repeated fertilization on a number of ecosystem attributes.  The response of non-timber    7  forest products (Clason et al. 2008), habitat use by mule deer and moose (Alces alces L.) (Sullivan et al. 2006a), snowshoe hare (Lepus americanus Erxleben) (Sullivan et al. 2006b), and small mammals (Sullivan et al. 2012) are reported elsewhere. 1.4 Methods 1.4.1 Study areas Three study areas were chosen on the basis of having candidate stands of young (12 to 14 year old) lodgepole pine that had relatively uniform tree cover, comparable diameter, height, and density of trees prior to stand treatments.  Location, proximity (boundaries), and size of candidate stands were determined by a balance between adequate interspersion of experimental units (Hurlbert 1984) and the logistics and access for conducting the operational-scale treatments of PCT and fertilization.  The study areas are located in south-central British Columbia (BC), Canada and are named after nearby towns (Summerland and Kelowna) or forest region in which they are located (Cariboo).  Both the Summerland and Kelowna study areas are located within the Montane Spruce biogeoclimatic zone (dry mild subzone; MSdm) (Meidinger and Pojar 1991).  The MS zone has a cool, continental climate with cold winters and moderately short, warm summers.  Mean annual temperature is 0.5-4.7?C and precipitation ranges from 380 to 900 mm.  The MS landscape has extensive, young and maturing seral stages of lodgepole pine, which have regenerated after wildfire.  Hybrid interior spruce (Picea engelmannii ? P. glauca (Moench) Voss) and subalpine fir (Abies lasiocarpa (Hook.) Nutt.) are the dominant shade-tolerant climax trees.  Douglas-fir (Pseudotsuga menziesii (Mirbel) Franco. var glauca (Beissn.) Franco) is an important seral species in zonal ecosystems and is a climax species on warm south-facing slopes in the driest ecosystems.  Trembling aspen (Populus tremuloides Michx.) is a common seral species and black cottonwood (Populus trichocarpa T. & G.) occurs on some moist sites (Meidinger and Pojar 1991).   The Summerland study area is located in the Bald Range 25 km west of Summerland (49?40?N; 119?53?W).  Topography of this area is gently rolling to flat with sandy loam soil at 1450 to 1530 m elevation (a.s.l.).  Clearcut harvesting of lodgepole pine with some uniform and group seed-tree reserves of Douglas-fir began in this area in 1978 in response to an outbreak of mountain pine beetle (Dendroctonus ponderosae Hopk.).  Depending on the original    8  composition of the harvested stands and the degree of windthrow after harvesting, the number of residual Douglas-fir ranged from none to one or two trees per ha.  Lodgepole pine regenerated naturally after harvesting and was the dominant tree species in these young stands.  Three harvested units with pre-thinning stand densities ranging from 9,980 to 11,150 stems/ha were divided into eight treatment stands and one control stand, as per the experimental design.  Minor components of the stands included Douglas-fir, interior spruce, subalpine fir, ponderosa pine (Pinus ponderosa Laws.), willow (Salix spp.), Sitka alder (Alnus viridis ssp. sinuata (Chaix) DC. (Regel) A. L?ve & D. L?ve), and trembling aspen.  Predictive ecosystem mapping (PEM) provided by the forest licensee operating in this area (Gorman Bros. Lumber Ltd., Summerland, BC) indicated that my treatment stands are predominantly zonal with site series 01 (i.e., zonal nutrient and moisture regimes; Steen and Coupe 1997).  A wetter area (site series 07) does exist within the 1000 stems/ha unit (both unfertilized and fertilized); however, sample plots avoided this area. Seasonal cattle (Bos taurus L.) grazing is prominent throughout the Summerland study area.  Range management of the 6050 ha Bald Range Summer Pasture, which includes the Summerland study area, was described as 125 cow/calf pairs, plus 6 bulls, grazed from June 8 to August 31, yielding 366 animal unit months (AUM).   At the start of the study (1993), DBH (diameter at breast height, 1.3 m above the forest floor) ranged from 2.2 ? 0.1 cm (mean ? S.E.) to 4.1 ? 0.1 cm with a mean age of 12 to 14 years.  Stand height ranged from 2.3 ? 0.1 m to 3.4 ? 0.1 m (Figure 1.1).  Stand areas ranged from 4.4 to 11.3 ha (Figure 1.2; Table 1.1).      9  Figure 1.1.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Summerland study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 10700 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE.   012345012345250250F500500F10001000F20002000FUnthinned250250F500500F10001000F20002000FUnthinnedDBH HeightHeight (m) DBH (cm) Summerland    10  Figure 1.2.  Summerland study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand.  250 500F 500 250F UT - 500 m -    11  Table 1.1.  Split-plot allocation of density and fertilizer treatments and stand descriptions by replicate study area (blocks).   Treatments Area 1993 stand density (stems/ha) Block Density? (main plot) Fertilizer (split plot) (ha) pre-thinning post-thinning Summerland 250 unfertilized 11.3 9980 to 11150 268  fertilized 11.3  500 unfertilized 7.6 9980 to 11150 511  fertilized 7.6  1000 unfertilized 4.5 9980 to 11150 936  fertilized 4.5  2000 unfertilized 4.4 9980 to 11150 1774  fertilized 4.4  unthinned unfertilized 5.0 10700 10700 Kelowna 250 unfertilized 10.0 8686 286  fertilized 10.0  500 unfertilized 11.0 8686 619  fertilized 11.0  1000 unfertilized 9.5 8686 1004  fertilized 9.5  2000 unfertilized 11.9 8686 1739  fertilized 11.9  unthinned unfertilized 12.6 4029 4029 Caribou 250 unfertilized 1.5 3333 292  fertilized 1.5  500 unfertilized 4.5 3333 470  fertilized 4.5  1000 unfertilized 3.2 3333 980  fertilized 3.2  2000 unfertilized 4.3 3333 1240  fertilized 4.3  unthinned unfertilized 3.3 2915 2915 ?  Target PCT density (stems/ha)  The Kelowna study area is located 37 km northwest of Kelowna, BC (50?04?N; 119?34?W).  Topography of this area is gently rolling to flat with sandy loam soil at 1230 to 1260 m elevation.  This area was clearcut harvested in 1979-1980 and regenerated naturally to lodgepole pine with the other coniferous species, including western larch (Larix occidentalis Nutt.), as minor components.  PEM data provided by the forest licensee operating in this area (Tolko Industries Ltd., Armstrong, BC) indicated that my treatment stands were representative of the MSdm biogeoclimatic subzone as site series were predominantly 01 (i.e., zonal nutrient and moisture regimes; Steen and Coupe 1997).  Cattle grazing within the Kelowna study area was minimal. In 1993, the mean stand DBH and height ranged from 3.1 ? 0.1 cm to 4.7 ? 0.1 cm and 3.0 ? 0.1 m to 4.1 ? 0.1 m, respectively, with a mean stand age of 12-13 years (Figure 1.3).  Area    12  of stands ranged from 9.5 to 12.6 ha (Table 1.1).  One large unit (84.8 ha) with a pre-thinning density of 8,686 stems/ha was shaped in a horseshoe with an unharvested riparian management zone separating the two arms (Figure 1.4).  This riparian zone had a steep ravine and varied in width from 75 to 300 m.  The overall unit was separated into eight treatment stands.  An additional unthinned stand completed the experimental design.  Figure 1.3.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Kelowna study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 4700 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE.  012345012345250250F500500F10001000F20002000FUnthinned250250F500500F10001000F20002000FUnthinnedDBH HeightHeight (m) DBH (cm) Kelowna    13  Figure 1.4.  Kelowna study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand.    250 250F 1000 2000 1000F 500F 500 UT 2000F - 500 m -    14  The Cariboo study area is within the Sub-Boreal Spruce biogeoclimatic zone (dry warm subzone; SBSdw) (Meidinger and Pojar 1991) and is located within the Gavin Lake block of University of British Columba?s Alex Fraser Research Forest, 75 km northeast of Williams Lake, BC (52?29?N; 121?45?W).  The SBS zone is characterized by mean annual temperature from 1.7 to 5?C and mean annual precipitation from 440-900 mm, of which ca. 25-50% is snow.  In mature stands, hybrid interior spruce, subalpine fir, and some Douglas-fir are mixed with extensive stands of lodgepole pine, which regenerated after wildfires.  The general topography of the study area is gently rolling to flat at 830 to 910 m elevation.   Range management of the 3000 ha range unit, which includes the Cariboo study area, was described as 240 cow/calf pairs, plus 10 bulls, grazed from May 16 to June 15, and 50 cow/calf pairs, plus 5 bulls, grazed from September 1 to October 15, yielding 332 AUMs. All treatment stands were located within an 80 ha unit that was clear-cut harvested in 1976 followed by some natural regeneration and some planting of lodgepole pine in 1983.  In 1993, pre-thinning stand density was 3333 stems/ha with a mean stand age of 13 years.  Mapping indicated that treatment stands were representative of the SBSdw subzone, as site series were predominantly 01 (i.e., zonal nutrient and moisture regimes; Lloyd et al. 1990; Klinka et al 2004).  Post-thinning stand DBH and height ranged from 4.4 ? 0.2 cm to 7.2 ? 0.3 cm and 3.4 ? 0.1 m to 5.4 ? 0.2 m, respectively (Figure 1.5).  Area of stands ranged from 1.5 to 4.5 ha (Table 1.1; Figure 1.6).        15  Figure 1.5.  Mean DBH (cm) and top height (m) of crop trees immediately post-thinning (fall of 1993) at the Cariboo study area.  Numbers along the x-axis represent target thinning densities (stems/ha) for each of the treatment stands.  The unthinned stand had a density of ca. 2915 stems/ha. The F identifies stands selected for repeated fertilization treatments.  Error bars indicate SE.   012345678012345678250250F500500F10001000F20002000FUnthinned250250F500500F10001000F20002000FUnthinnedDBH HeightHeight (m) DBH (cm) Cariboo    16  Figure 1.6.  Cariboo study area (2004 Google Earth image) with treatment stand boundaries sketched in yellow.  The numbers indicate the target PCT density (stems/ha), ?F? indicates stands that were repeatedly fertilized, and ?UT? indicated the unthinned stand.   UT - 500 m - 500 500F 1000 1000F 250 250F    17  In 2005, an outbreak of mountain pine beetle began to impact all of the stands throughout the Cariboo study area.  As a result, data collected from this replicate post 2004 were omitted from analyses. 1.4.2 Experimental design The three study areas acted as regional replicates (blocks).  Within each replicate, there were five experimental plots which had lodgepole pine stands pre-commercially thinned (PCT) in the following randomized block design: very low density (target 250 stems/ha), low density (target 500 stems/ha), medium density (target 1000 stems/ha), high density (target 2000 stems/ha), and unthinned (at least 3000 stems/ha).  Fertilization treatments were applied to one half of each of the thinned stands, resulting in a total of nine experimental units per replicate as follows: 1) 250 stems/ha, 2) 250 stems/ha with fertilization, 3) 500 stems/ha, 4) 500 stems/ha with fertilization, 5) 1000 stems/ha, 6) 1000 stems/ha with fertilization, 7) 2000 stems/ha, 8) 2000 stems/ha with fertilization, and 9) unthinned (Table 1.1).  A fertilized unthinned experimental unit was not included in this design as this treatment combination would not be part of any management prescription.  The restriction on randomization for the allocation of fertilizer treatment (i.e., applied to half of each density treatment) resulted in an overall split-plot design, with density as the main-plot and fertilization as the split-plot.  Time was treated as a split-split plot factor for datasets including samples over multiple years. Treatments of PCT and fertilization were applied at operational scales (i.e., several hectares), which, because of the restrictive amount of resources required, limited the number of replicates to three.  Accordingly, statistical power will be lower compared to studies with higher numbers of replicates.  Therefore, conclusions are worded conservatively and should be interpreted as indications rather than hard-core inferences.    Additional details regarding experimental design components relevant to specific studies are provided within the research chapters that follow. 1.4.3 Treatments 1.4.3.1 Density The initial treatment applied to this study was PCT of young stands of pine at an appropriate time to maximize growth response potential before trees experience severe growth    18  repression.  PCT, using chainsaws, was carried out in the late summer-early fall of 1993 (Figure 1.7), at which time stands were 12 to 14 years old.  Thinned trees were left on site.  To provide rigorous measurements of the various response variables, treatments were applied at a spatial scale large enough to be comparable to operational applications.  The usual prescription (>90% of stands) for PCT of lodgepole pine in BC is 1600-2000 stems/ha, with variations from this range tending to be towards higher rather than lower densities.  The broad range of densities applied during this study (250 to 2000 stems/ha) was clearly testing extremes beyond standard operational prescriptions and was deemed necessary to cause measurable changes in attributes of crop tree productivity as well as other non-timber values of interest to this research project such as measures of plant community diversity, structural diversity, and ungulate habitat use.   Within heavily thinned stands (? 1000 stems/ha), pruning was carried out five years following thinning (from September to December, 1998) to enhance wood quality (Ballard and Long 1988; Kellomaki et al. 1989) and stem form of crop trees (Muhairwe 1994).  Pruning to a height of approximately 3.0 m was carried out using manual pruning saws and all pruning debris (slash) was left on site. 1.4.3.2 Fertilization  The fertilization treatments were designed as large-scale applications of previously established ?optimum nutrition? fertilization field experiments developed in Sweden (Tamm et al. 1999) and BC (Brockley 1992; Kishchuk et al. 2002).  The aim was to maintain elevated foliar N levels (~ 1.3%), with levels of all other nutrients in proportional balance with N (Linder 1995).      Fertilization treatments were initiated in November 1994 using a blended fertilizer formulated to provide 100 kg/ha Nitrogen (100 N), 100 kg/ha Phosphorus (100 P), 100 kg/ha Potassium (100 K), 50 kg/ha Sulfur (50 S), 25 kg/ha Magnesium (25 Mg), and 1.5 kg/ha Boron (1.5 B).  The blended product (11-25-13-5.5S-2.7Mg-0.17B) was applied at a rate of 906 kg/ha to each of the four treatment stands (fertilized half of the 250, 500, 1000, and 2000 stems/ha units) at all three study areas.  Fertilizer was applied by helicopter at the Summerland and Kelowna study areas.  Fertilizer was spread by hand at the Cariboo study area because of the relatively small size of the treatment stands throughout this replicate.  Fertilization of the Cariboo site was completed prior to initiation of tree growth in April 1995.     19  During the fall of 1995 (one year following initial fertilization), replicated samples of current year?s foliage were collected from all stands.  Foliar sampling was carried out to monitor the nutrient status of the crop trees and to develop appropriate multi-nutrient formulations for subsequent fertilizer applications.   Fertilizers were reapplied every two years for a 10-year period, resulting in a total of five applications (Figure 1.7).  During May of 1997, two growing seasons after initial application, stands were re-fertilized with a N+S blended fertilizer (36-0-0-9S) at an application rate of 547 kg/ha (200 N and 50 S).  In October of 1998, two growing seasons after the second application, stands were re-fertilized with a blended product (37-0-0-6.1S-0.7B) at an application rate of 404 kg/ha (150 N, 25 S, and 3 B).  During the fall of 2000, two growing seasons after the third application, stands were re-fertilized with a blended product (31.1-0-0-11.3S) at an application rate of 439.4 kg/ha (150 N and 50 S).  And finally, during the spring of 2003, two growing seasons after the fourth application, stands were re-fertilized with a blended product (44.6-0-0-0.45B) at an application rate of 336.1 kg/ha (150 N and 1.5 B). Application methods described for the initial fertilization treatment (i.e., aerial vs. manual) remained constant for the duration of this study.       20  Figure 1.7.  Gantt chart depicting the timing of significant events (treatments, samples, and disturbances) that have occurred during the initial 16 years of this research project (1993 to 2008). Event Year ?93 ?94 ?95 ?96 ?97 ?98 ?99 ?00 ?01 ?02 ?03 ?04 ?05 ?06 ?07 ?08 Treatments                                    Pre-commercial thinning                                    Fertilization                                 Samples                                    Crop tree foliage (nutrient content)                                    Crop tree stems (DBH, height, BA, volume)                                    Crop tree crowns (area, volume)                                    Stand structure (density of all tree species by height class)                                    Plant community, grazed (crown volume index)                                    Plant community, cattle excluded (crown volume index)                                    Habitat use (spring and fall fecal pellets counts)                                    Forage quality and quantity                                 Disturbances                                    Mountain pine beetle infestation (Cariboo replicate only)                                      21  1.4.4 Data collection and calculations Data collection took place over a period of 16 years (1993 to 2008).  Below is an overview of the methods used to collect and calculate the data used to measure various treatment effects reported throughout this thesis.  Further details of these methods are provided within the relevant chapters that follow. 1.4.4.1 Crop trees Sampling of lodgepole pine crop trees (trees left standing following PCT treatments) was designed to measure the effects of PCT and repeated fertilization on growth and yield of crop tree stems, as well as crown characteristics.  Sampling was carried out within systematically located, variable-radius, permanent plots; each containing 10 crop trees.  A target of 20 plots (i.e., 200 crop trees) was established within each treatment stand.  Sample trees were permanently tagged with individually numbered metal tags.  Initial sampling was carried out immediately after thinning in the fall of 1993 and DBH and height of all sample trees measured.  Subsequent samples were carried out at five-year intervals during the fall of 1998, 2003, and 2008, and included additional measurements of crown attributes (crown width and heights to the base and widest parts of the crown) (Figure 1.7).   1.4.4.2 Stand structure  Density of trees (crop trees and any other trees that were not culled during thinning treatments or established post-thinning) was estimated among four height classes during the final 10 years (1998 to 2008) using 20, fixed-radius (100 m2), non-permanent plots per treatment stand.  Sample plots were systematically located every 50 m along compass lines (50 m apart).  Within each plot, all trees were tallied by species into four height classes; 0-1.0, 1.1-2.0, 2.1-3.0, and >3.0 m.  Three separate samples were carried out at five-year intervals; 1998, 2003, and 2008 (Figure 1.7). 1.4.4.3 Plant community Vegetation sampling (herbs, shrubs, and trees) was designed to measure the effects of PCT and repeated fertilization on individual plant species, as well as community-level attributes such as species and structural diversity.  This was accomplished by transect sampling that provided estimates of abundance by species.  A visual estimate of percent cover was made for each species-height class combination.  These data were then used to calculate a crown volume    22  index (m3/0.01 ha) for each plant species (Stickney 1985).  Three permanent vegetation transects were randomly located in each treated stand during 1993 and sampled prior to PCT.  To isolate any effects that cattle disturbance may have had on plant response, fenced exclosures were constructed; each containing a single vegetation transect.  One such exclosure was randomly located within each treatment stand and was established during 1994.  Sampling was carried out during July-August, and continued annually until 2003.  A final sample took place in 2008 (Figure 1.7).  1.4.4.4 Forage quantity and quality In 2007, forage quantity and quality was estimated by collecting and analyzing herbs growing within grazing-proof cages (10 plots per stand).  All herbaceous vegetation was clipped at ground level and sorted into various classes of forage.  Samples were then analyzed in a laboratory to determine accurate estimates of quantity and quality (crude protein and acid detergent fiber). 1.4.4.5 Ungulate habitat use Relative habitat use by ungulates was inferred by scat counts carried out semi-annually over a period of 6 years (1998 to 2003) within permanent, 5-m2 sample plots (target 100 plots per stand). 1.4.5 Statistical analysis  Regional replicates functioned as blocks and were assigned as a random factor.  The experimental design restricted the randomness of fertilizer treatment allocation (i.e., applied to one-half of each of the thinned stands), and hence a split-plot analysis of variance (ANOVA) (SPSS Institute Inc. 2007) was used to compare treatment means.  The density and fertilizer treatments were assigned as the main- and split-plots, respectively.  Time (year of sample) was assigned as a split-split plot factor when analyzing data over multiple years.  Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare mean values based on ANOVA results (Saville 1990).  In all analyses, the level of significance was ? = 0.05.  P-values ranging from 0.06 to 0.10 were reported as marginally significant.  Regression analyses were used to investigate relationships.  Where appropriate, data were pooled across replicates and sample years.  Before pooling, regression relationships were    23  assessed for differences in intercept and slope among study areas and sample years.  Data were analyzed separately where significant differences were detected.  The strength of a relationship (i.e., degree of correlation) was described with the coefficient of determination and referred to as very weak (R2=0.00-0.03), weak (R2=0.04-0.14), modest (R2=0.15-0.47), strong (R2=0.48-0.78), and very strong (R2=0.79-1.00) as per Fowler et al. (1998).       24  2 Long-term responses of tree and stand growth of young lodgepole pine to pre-commercial thinning and repeated fertilization 2.1 Introduction With very few exceptions, Canada?s forest industry has been based on extensive silviculture practices that have been able to take advantage of large tracts of unmanaged, old forests, often with little or no investment in silviculture treatments beyond stand establishment.  Intensive silviculture has been clearly demonstrated as an economically beneficial strategy within pine (Pinus spp.) plantations throughout the southern United States (US) with growth rates more than doubled and rotation lengths cut by more than 50% (Fox et al. 2007).  However, the benefits of intensive silviculture remain largely unrealized throughout northern latitudes, and within boreal and sub-boreal forests in particular (Lautenschlager 2000; Park and Wilson 2007).  There is a predicted increase in global demand for timber production and forest cover to produce conventional wood products, biofuels production, and sequester carbon in response to climate change (Sedjo 1999; Raunikar et al. 2010).  This demand is concurrent with conservation strategies that endeavor to increase the size of protected areas and conserve biodiversity (Hunter and Schmiegelow 2011), while balancing the unpredictable large-scale loss of existing timber to natural disturbances such as losses to wildfire and insect epidemics (Agee 1993; Walton et al. 2009).  As a result, enhanced wood production is predicted to become increasingly necessary to mitigate current and future wood supply shortfalls (Brooks 1997; Sutton 1999).  Intensive silvicultural practices such as pre-commercial thinning (PCT), commercial thinning, and fertilization have the potential to sustain wood and biomass production while creating a diversity of forest habitat conditions to meet the goals of biodiversity conservation (Moore and Allen 1999; Hartley 2002; Sullivan et al. 2009).  These silvicultural treatments have been used successfully around the world to increase biomass production in existing even-aged forests (Allen et al. 1990; Oliver and Larson 1996), across northern Europe (Nabuurs et al. 2007; Bergh et al. 2008), the southeastern US (Albaugh et al. 2004; Jokela et al. 2004), and inland lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) forests of the Pacific Northwest (PNW) of North America (NA) (Sullivan et al. 2006c; Lindgren et al. 2007).    25  Early- to mid-seral (1-40 years old) lodgepole pine is the dominant coniferous tree species across a vast area of the inland PNW of NA (Koch 1996; Sullivan et al. 2001).  This species likely has the greatest potential to respond favorably to silvicultural treatments designed to enhance the growth of crop trees within stands (Johnstone 1985; Brockley 2005).  Lodgepole pine often regenerates over-abundantly, after wildfire or clearcut harvesting, with excessive stand densities that reduce tree growth and stand productivity.  PCT and CT concentrate growth on a smaller number of stems and provide some control over the rotation, yield, and value of the future crop (Johnstone 1985; Cole and Koch 1996).  Because they originated from repeated fire disturbance, lodgepole pine forests usually occupy sites of low-N status (Brockley et al. 1992), and hence respond well to conventional, single applications of nitrogen, as well as in combination with other elements (Weetman 1988; Brockley 1996).  Sustained growth responses to fertilization with optimum nutrition formulations have been demonstrated in field experiments with lodgepole pine (Brockley 2005; Lindgren et al. 2007) and other Pinus species (Malkonen and Kukkola 1991; Tamm et al. 1999; Fox et al. 2007). Operational scale nutrition experiments apply nutrients infrequently, usually in a larger amount.  Thinning and repeated fertilization treatments are applied over an entire ecosystem, and hence they have the potential to significantly increase stand-level wood production and structural diversity.  Small-scale studies have demonstrated the concept of ?steady-state? nutrition, whereby small, balanced supplies of nutrients are provided at optimum rates consistent with estimated demand (Linder 1987; Raison and Myers 1992; Brockley 2005).  Stand production and structural diversity may be enhanced by maintaining steady state nutrition with repeated optimum nutrient applications.  With respect to stand structure and biodiversity, we have barely begun to explore the possibilities (Sullivan et al. 2013).   This study was designed to test the hypothesis that, among managed stands, application of a range of PCT intensities and repeated fertilization with optimum nutrition formulations will enhance the 15-year growth increments of lodgepole pine crop trees at both tree and stand levels.       26  2.2 Methods 2.2.1 Study areas Three study areas each containing several lodgepole pine stands were originally established in 1993.  The Summerland study area is located in the Bald Range, 25 km west of Summerland in south-central British Columbia (BC), Canada (49?40?N; 119?53?W).  The Kelowna study area is located 37 km northwest of Kelowna, BC (50?04?N; 119?34?W).  Both areas are in the Montane Spruce biogeoclimatic zone (dry mild subzone; MSdm) (Meidinger and Pojar 1991).  A third study area near Williams Lake, BC, reported in Lindgren et al. (2007), was decimated by mountain pine beetle (Dendroctonus pondersosae Hopk.) in 2005, and therefore excluded from this analysis.   Additional study area details are provided in Lindgren et al. (2007) and Chapter 1. 2.2.2 Experimental design The two study areas were replicates (blocks).  Within each replicate, there were five experimental lodgepole pine stands PCT in the following randomized block design: very low density (target 250 stems/ha), low density (target 500 stems/ha), medium density (target 1000 stems/ha), high density (target 2000 stems/ha), and unthinned (at least 4000 stems/ha).  Fertilization treatments were applied to one half of each of the thinned units, resulting in a total of nine stands per study area: 1) 250 stems/ha, 2) 250 stems/ha with fertilization, 3) 500 stems/ha, 4) 500 stems/ha with fertilization, 5) 1000 stems/ha, 6) 1000 stems/ha with fertilization, 7) 2000 stems/ha, 8) 2000 stems/ha with fertilization, and 9) unthinned.  The restriction on randomization for the allocation of fertilizer treatment (i.e., applied to one-half of each density treatment) resulted in an overall split-plot design, with density as the main-plot effect and fertilization as the split-plot effect.  A fertilized unthinned experimental unit was not included as this treatment combination would not be part of any management prescription.   Additional details of the experimental design are provided in Chapter 1. 2.2.2.1 Density and fertilization treatments  The initial treatment was PCT of pine during fall of 1993.  Pruning (3-m lift) was carried out within all stands with densities < 2000 stems/ha in 1998, five years after PCT.      27  Aerial (helicopter) applications of fertilizer were initiated in November 1994.  Fertilizers were reapplied every two years for a 10-year period, resulting in a total of five applications.  The objective was to maintain a foliar N level of 1.3%, with foliar levels of all other nutrients in proportional balance with foliar N concentration.  Results from foliar analyses were used to develop appropriate blends and application rates for treatments to ensure that this objective was being met.  Foliar sampling was conducted annually from 1995 to 2003.  Foliage was collected from 10 representative healthy dominant or co-dominant crop trees evenly distributed within each stand.  Samples of current year?s growth were collected from the lower portion of the top third of the live crown, consistent with standardized foliar sampling guidelines (Brockley 2001).  Individual foliage samples were frozen following field collection, and then dried in a forced-air oven at 70o C for 20 hours before analysis.  Two composite samples per block, each sample consisting of equal amounts of foliage from five of the trees, were prepared for chemical analysis.  Dried composite samples were ground in an electric coffee grinder and sent to a commercial laboratory for nutrient analysis.    Composite samples for all years were digested using a variation of the sulphuric acid ? hydrogen peroxide procedure described by Parkinson and Allen (1975).  The digests were analyzed colorimetrically for N and P on a Technicon Autoanalyzer, N using the Kjeldahl procedure and P using a procedure based on the reduction of the ammonium molybdophosphate complex by ascorbic acid (Watanabe and Olson 1965).  Total K, Ca., Mg, and Mn were determined by atomic absorption spectrophotometry.  Total S was determined by combustion using a LECO sulphur analyzer.  Separate sub-samples were dry-ashed, and Cu, Zn, and Fe concentrations were determined by atomic absorption spectrophotometry.  After dry-ashing, B was determined colorimetrically using the azomethine-H method described by Gaines and Mitchell (1979).  Addition details of PCT and fertilization treatments are provided in Lindgren et al. (2007) and Chapter 1. 2.2.2.2 Tree measurements  In each stand, sampling of lodgepole pine crop trees was done within permanent tree plots with variable-radius to accommodate variations in stand density.  The initial sampling measured DBH (cm) and tree height (m) and was conducted after PCT during fall of 1993.     28  Subsequent samples were carried out at five-year intervals during fall 1998, 2003, and 2008, and included additional measurements of crown width and heights to the base and widest parts of the crown.  In 1998, tree measurements were completed immediately prior to pruning.  Crown measurements were made within one-half of the tree plots.  Initial tree height measurements (1993) were made using a telescopic height pole, which measured top height to the nearest 0.01 m.  Re-measurements of tree heights in 1998, 2003, and 2008 were to the nearest 0.1 m using a digital hypsometer (Forestor Vertex).  At the tree-level, growth response of the stem to thinning and fertilizer treatments was described by 15-year increments (1993 to 2008) of top height, DBH, basal area (BA), and volume.  Five- and 10-year increments are reported in Lindgren et al. (2007).  Tree volume was calculated using a variable-exponent taper function developed for lodgepole pine in the central interior of BC (Kozak 1988):  [1]  V = a1 + a2D2H  where V is gross total inside bark volume (m3), D is DBH (m), and H is top height of tree (m).  The two constants, a1 and a2, varied depending on the study area and range of DBH and tree heights observed within a given sample period.  Pre-treatment stem volumes were estimated using a1 and a2 values of 1254 and 0.349, respectively.  Stem volumes for trees sampled in 2008 were estimated using a1 and a2 values of 3633 and 0.3474, respectively. At the stand-level, BA/ha and volume/ha were estimated by multiplying the mean tree values by an estimate of stand density (stems/ha) for each stand.  Initial stand density was determined immediately following PCT in 1993.  Stand densities in 2008 were estimated based on the observed percentage of dead or severely snowpressed trees among the permanently tagged crop trees.  This estimate of mortality was used to adjust the initial stand density by a percentage assumed to be representative of the mortality experienced throughout a given stand.  As a result, stand-level estimates of BA/ha and volume/ha within thinned stands apply only to the residual    29  cohort of dominant and codominant lodgepole pine trees selected during thinning in 1993 and do not account for any ingress that occurred during the 15 years following thinning.   Crown measurements were initially made immediately prior to pruning in 1998.  Crown area represented the widest horizontal cross-section of a crown, and was calculated using the widest crown diameter measurement.  Crown characteristics following pruning (1998) were estimated based on the assumption that the new base of the live crown created by pruning was equal to the base of the live crown in 2003 (Figure 2.8).  These estimates of post-pruning crown dimensions should be reasonably accurate because lodgepole pines do not regrow lower branches by epicormic branching (Ballard and Long 1988) and pruned stands had yet to experience any self-pruning as of 2003.  To account for the effects that pruning had on crown characteristics, crown development was calculated relative to the 1998 post-pruning estimates rather than the 1998 pre-pruning measurements.   Addition crop tree measurement details are provided in Lindgren et al. (2007).  Density of coniferous trees (crop trees and any other trees that were not culled during thinning treatments or established post-thinning) was estimated among four height classes during the final 10 years (1998 to 2008) using 20 fixed-radius (100 m2) plots within the permanent crop tree plots.  Within each plot, all trees were tallied by species and height class (0-1.0, 1.1-2.0, 2.1-3.0, and >3.0 m).  Three separate samples were carried out at five-year intervals; 1998, 2003, and 2008.  2.2.2.3 Statistical analysis  The experimental design restricted the randomness of fertilizer treatment allocation (i.e., applied to one-half of each of the thinned stands), and hence a split-plot analysis of variance (ANOVA) was used to compare treatment means.  The density and fertilizer treatments were assigned as the main- and split-plots, respectively.  When several years of data were analyzed together, the time factor was assigned as a split-split plot.  The two replicates functioned as blocks.  Because tree growth is correlated with initial tree size (Johnstone 1985), a split-plot analysis of covariance (ANCOVA) was used to evaluate stem growth rate responses to the treatments using initial tree size as the covariate.      30  Before the adjusted means could be used, the assumption of homogeneity of regression coefficients had to be tested.  Testing this assumption was carried out using the same blocked, split-plot ANCOVA design described above.  A significant covariate ? density ? fertilizer interaction indicated that the assumption was violated and precluded any further analyses using ANCOVA.  When this occurred, separate regression equations were generated for each stand (independent factor regressed on the covariate) and used to adjust tree growth to the same rate.  These adjusted values were then analyzed using a blocked, split-plot ANOVA.  Effects of treatments on stand density, foliar N concentration, and crown characteristics were analyzed without a covariate, using a blocked, split-plot ANOVA. Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare treatment means (Saville 1990).  In all analyses, the level of significance was ? = 0.05.  P-values ranging from 0.06 to 0.10 were also considered of interest and are reported as marginally significant. The experimental design was unbalanced because there were no unthinned, fertilized stands.  To maintain the power and sensitivity of a balanced design, the unthinned level of the density treatment was omitted from all statistical analyses.  Data from unthinned stands are presented in graphs to allow for a visual comparison with managed stands. 2.3 Results 2.3.1 Tree density  Estimates of post-PCT tree densities among thinned stands made immediately following thinning (Table 1.1) indicate that thinning treatments had a significant (F3,3=430.19; P<0.01) effect on the residual tree (i.e., crop tree) density.  While actual post-thinning densities were slightly different than the PCT target densities, each of the four thinning levels resulted in significantly different (DMRT; P=0.05) crop tree densities.  The post-PCT crop tree density estimated in 1993 changed little during the following 15 years as only 5.75 ? 2.06% of crop trees (averaged over all densities) were lost during this period; primarily due to snowpress.    31  Table 2.1.  Summary of split-split plot ANOVA investigating the effects of PCT density (main plot; four levels; target of 250, 500, 1000, and 2000 crop trees/ha), fertilizer (split plot; two levels; unfertilized and repeatedly fertilized), and time (split-split plot; three levels; 1998, 2003, and 2008) on density of all trees among four height classes. Height Class density fertilizer density ? fertilizer time time ? density time ? fertilizer F3,3 P F1,4 P F3,4 P F2,16 P F6,16 P F2,16 P dominant layer                  > 3.0 m 2982.89 <0.01 0.66 0.46 1.50 0.34 31.76 <0.01 2.56 0.06 0.84 0.45 understory layers                  2.1 ? 3.0 m 0.95 0.51 2.19 0.21 0.24 0.87 84.55 <0.01 8.79 <0.01 5.17 0.02      1.1 ? 3.0 m 2.48 0.24 9.89 0.03 0.03 0.99 12.74 <0.01 0.89 0.53 19.59 <0.01      0.0 to 1.0 m 0.87 0.55 42.60 <0.01 1.14 0.43 39.00 <0.01 0.26 0.95 4.45 0.03       32  Because very few understory saplings and/or seedlings existed throughout the stands prior to thinning and because the objective of thinning was to remove all trees that could compete with future crop trees, post-thinning tree composition was initially made up of a single stratum of crop trees, which ranged in height from approximately 2.5 to 4.0 m.  Therefore, immediately following thinning, the tallest height class (taller than 3.0 m) included only crop trees.  However, from 1998 to 2008 (five to 10 years post-thinning), the density of the tallest height class was significantly affected (F3,3=2982.89; P<0.01) by PCT.  As was reported for 1993, each of the four density levels had a significantly different mean density of this dominant trees strata (DMRT; P = 0.05) according to PCT intensity (i.e., 2000 > 1000 > 500 > 250 stems/ha; Figure 2.1).  During this period, the mean density of the dominant tree layer was also significantly affected by time (F2,16=31.76; P<0.01); increasing as shorter trees grew into this upper height class.  A marginally significant time ? density interaction (F6,16=2.56; P=0.06) resulted from dominant tree layer densities increasing far more rapidly in the heavily thinned stands compared to those lightly thinned (Figure 2.1).  From 1998 to 2008, the dominant tree layer density increased by factor of 2.82, 2.34, 1.33, and 1.09 among the 250, 500, 1000, and 2000 stems/ha stands, respectively (Figure 2.1).  Fertilizer treatments did not significantly affect the dominant tree layer density (Table 2.1).      33  Figure 2.1.  Mean total density (stems/ha) of trees within the dominant tree layer (taller than 3 m) among the four density treatments measured at five-year intervals; 1998, 2003, and 2008.  Densities include crop trees, as well as other trees that grew into this upper height class with time.  Numbers along the x-axis represent target PCT densities (stems/ha).  Error bars indicate SE and are based on two replicate study areas (n = 2).   During the final 10 years, mean density of the understory tree layers (? 3.0 m in height) increased with increasing PCT intensity and decrease with fertilization; a trend that became more pronounced with time (Figure 2.2).  PCT density treatments did not have an overall significant effect on the densities of any of the understory tree layers (Table 2.1).  However, a significant time ? density effect (F6,16=8.79; P<0.01) for the tallest understory strata (2.1 to 3.0 m) revealed that that the density effect, while only marginally significant five years post-PCT (F3,3=5.68; P=0.09), became significant 10 (F3,3=11.26; P=0.04) and 15 years post-PCT (F3,3=9.29; P=0.05).  Ten years following PCT, mean density of the 2.1 to 3.0-m tall tree strata (Figure 2.2a) was significantly greater within the three most heavily thinned stands (250, 500, and 1000 stems/ha) compared to the least thinned stands (2000 stems/ha) (DMRT; P=0.05).  By 15 years post-PCT, the significant difference was only evident between the two most heavily thinned stands (250 and 500 stems/ha) compared to both of the more lightly thinned stands (DMRT; P=0.05).  A 025050075010001250150017502000250 500 1000 2000stems/ha Target stand density (stems/ha) 1998 2003 2008   34  significant time ? fertilizer effect for all three understory tree strata (Table 2.1) resulted from a lack of significant fertilizer effect observed in 1998 and 2003 compared to significantly less density observed within fertilized compared to unfertilized stands in 2008 for the 0 to 1.0 m (F1,4=59.97; P<0.01), 1.1 to 2.0 (F1,4=35.05; P<0.01) and 2.1 to 3.0 m height classes (F1,4=12.61; P=0.02).      35  Figure 2.2.  Mean total tree density (stems/ha) within three understory tree layers (i.e., excluding crop trees) among the eight treatment stands sampled in 1998, 2003, and 2008.  Numbers along the x-axis represent target PCT densities (crop trees/ha) and the ?F? denotes stands that were repeatedly fertilized.  Error bars indicate SE and are based on two replicate study areas (n = 2).      0100200300400500600250 250F 500 500F 1000 1000F 2000 2000Fstems/ha a) 2.1 - 3.0 m height class 1998 2003 2008020040060080010001200250 250F 500 500F 1000 1000F 2000 2000Fstems/ha b) 1.1 - 2.0 m height class 04008001200160020002400250 250F 500 500F 1000 1000F 2000 2000Fstems/ha c) 0.0 - 1.0 m height class    36  2.3.2 Fertilizer  Foliar N levels within fertilized stands remained higher than in unfertilized stands throughout the nine-year period that foliar analyses were conducted (1995 to 2003) and peaked in the year following each re-fertilization (Figure 2.3).  Density and the density ? fertilizer interaction did not have a significant effect on foliar N levels.  Averaged across all years, repeated fertilizer treatments resulted in significantly (F1,1=286.17; P=0.04) greater foliar N levels (1.25 ? 0.01%) compared to no fertilizer treatments (1.02 ? 0.01%).  Additional details regarding foliar concentrations of S, B, K, P, and Mg are provided in Lindgren et al. (2007).  No nutrient imbalances or deficiencies were detected as a result of the fertilization treatments.  Figure 2.3.  Mean foliar N concentration for young lodgepole pine crop trees within unfertilized and repeatedly fertilized stands measured annually from 1995 to 2003.  Values were averaged across all densities and two replicate study areas (n = 2).  Error bars indicate SE.      0.911.11.21.31.41.5199519961997199819992000200120022003[Foliar N] Year unfertilized fertilized=  timing of fertilization    37  2.3.3 Growth of crop trees 2.3.3.1 Stem attributes Adjusted mean 15-year height increments varied less than 0.75 m among all of the treated stands (from 5.32 ? 0.54 to 6.04 ? 0.52 m; Figure 2.4a).  Restated, a mean annual increment of less than five cm height growth separated the fastest (fertilized 1000 stems/ha) from slowest (fertilized 250 stems/ha) growing stands.  Neither density nor fertilizer treatments had a significant effect on 15-year height increment (Table 2.2).  Rate of height growth generally decreased during the final five years relative to the previous five-year increment.  Density did not change the rate of height growth; however, the decrease in rate of height growth was significantly less within fertilized than unfertilized stands (F1,3=26.79; P<0.01; Table 2.3).  Relative to the 1998 to 2003 five-year increment, fertilized stands grew a mean of four ? 28 cm in height more than the unfertilized stands during the final five years of this study (2003 to 2008).     38  Figure 2.4.  Adjusted mean 15-year increments (1993 to 2008) of a) height and b) DBH of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).         012345678250 500 1000 2000 unthinnedm a) Height, 15-yr increment unfertilized fertilized03691215250 500 1000 2000 unthinnedcm Target stand density (stems/ha) b) DBH, 15-yr increment    39  Table 2.2.  Adjusted mean (n = 2 replicate study areas) (SE) 15-year increments (1993 to 2008) of crop tree attributes for young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  Estimated marginal means (adjusted for covariate)  Split-plot ANCOVA  Density Fertilizer Density ? Fertilizer  F3,2 P F1,3 P F3,3 P HT (m) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  0.12 0.94 0.20 0.69 2.76 0.21 5.52 (0.42) 5.68 (0.92) 5.71 (0.45) 5.77 (0.79)        unfertilized fertilized        5.64 (0.56) 5.70 (0.74)        DBH  (cm) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  3.77 0.22 134.08 <0.01 4.10 0.14 13.25 (0.12) 12.20 (0.82) 10.83 (0.64) 9.83 (0.30)        unfertilized fertilized        10.64b (0.08) 12.42a (0.08)        BA/tree (m2x10-2)  250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  3.44 0.23 121.51 <0.01 4.33 0.13 2.24 (0.06) 2.00 (0.24) 1.64 (0.13) 1.42 (0.05)        unfertilized fertilized        1.60b (0.03) 2.05a (0.03)        BA/ha (m2) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  4.94 0.17 8.35 0.06 0.87 0.54 8.30 (0.24) 12.09 (1.70) 14.39 (1.58) 19.38 (1.16)        unfertilized fertilized        12.33 (0.25) 14.75 (0.64)        Vol./tree (m3) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  1.70 0.39 23.96 0.02 1.75 0.33 0.10 (0.01) 0.09 (0.02) 0.07 (0.00) 0.07 (0.00)        unfertilized fertilized        0.07b  (0.01) 0.09a (0.01)        Vol./ha (m3) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  2.24 0.32 10.82 0.05 1.83 0.32 48.02 (2.29) 63.06 (12.03) 67.38 (3.27) 81.47 (5.25)        unfertilized fertilized        59.72b (4.81) 70.24a (3.33)              40  Table 2.3.  Adjusted mean (n = 2 replicate study areas) (SE) change in growth rate observed during the final 10 years for crop tree attributes of young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  Adjusted mean change in growth rate?  Split-plot ANCOVA  Density Fertilizer Density ? Fertilizer  F3,2 P F1,3 P F3,3 P HT (m) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  0.60 0.67 26.79 0.01 34.67 <0.01 -0.07 (0.24) 0.05 (0.16) -0.29 (0.37) -0.09 (0.01)        unfertilized fertilized        -0.12b (0.19) -0.08a (0.20)        DBH (cm) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  7.76 0.12 3.30 0.21 0.85 0.58 -0.15 (0.45) 0.11 (0.29) -0.54 (0.34) -0.88 (0.14)        unfertilized fertilized        -0.22 (0.21) -0.50 (0.40)        BA/tree (m2x10-2)  250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  13.59 0.07 0.06 0.82 0.49 0.71 0.22 (0.08) 0.22 (0.03) 0.04 (0.07) -0.03 (0.03)        unfertilized fertilized        0.12 (0.03) 0.11 (0.08)        BA/ha (m2) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  0.53 0.71 0.96 0.40 0.26 0.85 0.33 (0.17) 0.96 (0.14) -0.13 (1.41) -0.24 (0.88)        unfertilized fertilized        0.58 (0.11) -0.12 (1.19)        Vol./tree (m3) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  3.83 0.21 8.27 0.06 1.10 0.47 0.02 (0.00) 0.02 (0.00) 0.01 (0.00) 0.01 (0.00)        unfertilized fertilized        0.01 (0.00) 0.02 (0.00)        Vol./ha (m3) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  0.34 0.80 0.10 0.78 0.08 0.97 5.70 (0.39) 10.81 (1.92) 6.23 (7.07) 10.81 (1.81)        unfertilized fertilized        8.86 (0.29) 7.92 (3.97)        ?  Change in growth rate was calculated by comparing the final five-year increment (2003 to 2008) with that of the previous five years (1998 to 2003).  A mean of zero would indicate a constant growth rate among the two five-year increments, whereas a positive result would indicate the amount by which growth had accelerated during the final five years, and vice versa.  Growth rates were adjusted for covariate effect.     41  Adjusted mean 15-year DBH increments among treated stands were positively correlated with PCT intensity and further increased by fertilization (Figure 2.4b).  Mean DBH increment was 1.51 times greater within stands with the fastest (fertilized 250 stems/ha) compared to the slowest (unfertilized 2000 stems/ha) diameter growth; 14.01 ? 0.14 vs. 9.31 ? 0.07 cm, respectively.  While DBH increment increased with decreasing stand density, the density effect was not significant (Table 2.2).  The effect of repeated fertilization on 15-year DBH increment was significant (F1,3=134.08; P<0.01), with 1.17 times greater diameter growth resulting from fertilizer treatments; 12.42 ? 0.08 vs. 10.64 ? 0.08 cm, respectively (Table 2.2).  The rate of diameter growth generally decreased during the final five years relative to the previous five-year increment; however, neither density nor fertilizer treatments significantly affected this change in diameter growth rate (Table 2.3). In 2008, the adjusted mean height-to-diameter ratios (HDR) of crop trees among treated stands were positively correlated with crop tree density and decreased by fertilization (Figure 2.5).  Mean HDR was 1.48 times greater within the unfertilized 2000 stems/ha compared to the fertilized 250 stems/ha stands; 71.66 ? 7.45 vs. 48.49 ? 4.25 cm:cm, respectively.  The density effect on HDR in 2008 was not significant (Table 2.4).  However, with 2.24 times greater HDR within unfertilized compared to fertilized stands (64.00 ? 4.08 vs. 57.38 ? 6.39 cm:cm, respectively), the fertilizer effect was significant (F1,3=2.16; P=0.02).  The decrease in HDR caused by density and fertilization treatments occurred within the initial five-year period and has remained relatively constant thereafter.  During the final 10 years, there were no significant changes to HDR resulting from either density (F3,2=3.04; P=0.26) or fertilization treatments (F1,3=4.90; P=0.11).       42   Figure 2.5.  Adjusted mean height-to-diameter ratio (HDR; cm:cm) of young lodgepole pine crop trees among nine treatment stands in 2008; 15 years after onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).  020406080100250 500 1000 2000 unthinnedcm:cm Target stand density (stems/ha) unfertilized fertilized   43  Table 2.4.  Adjusted mean (n = 2 replicate study areas) (SE) height-to-diameter ratio (cm:cm) of crop tree attributes for young lodgepole pine among four density and two fertilization treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  HDR in 2008  Split-plot ANCOVA  Density Fertilizer Density ? Fertilizer  F3,2 P F1,3 P F3,3 P 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  4.07 0.20 20.16 0.02 0.49 0.71 52.54 (2.87) 57.67 (2.76) 63.26 (6.74) 69.31 (8.57)        unfertilized fertilized        64.00a (4.08) 57.38b (6.39)              44  Adjusted mean 15-year tree BA increments among treated stands were positively correlated with PCT intensity and further increased by fertilization (Figure 2.6a).  Tree BA was 1.88 times greater within fertilized 250 stems/ha compared to unfertilized 2000 stems/ha stands; 2.43 ? 0.11 vs. 1.29 ? 0.01 m2 ? 10-2, respectively.  While the density effect was not significant, the effect of fertilizer was (F1,3=121.51; P<0.01), with a mean of 1.28 times greater tree BA growth resulting with fertilizer treatments; 2.05 ? 0.03 vs. 1.60 ? 0.03 m2 ? 10-2, respectively (Table 2.2).  Relative to the 1998 to 2003 five-year increment, the rate of tree BA growth increased during the final five-year increment (2003 to 2008) among the heavily thinned stands.  Both the 250 and 500 stems/ha stands experienced a mean of 0.22 m2 ? 10-2 (? 0.08 and 0.03 m2 ? 10-2, respectively) more BA growth per tree during the final five years compared to the previous five years (Table 2.3).  The rate of tree BA growth continued to increase within the 1000 stems/ha stands, but much more slowly (0.04 ? 0.07 m2 ? 10-2) than within the heavily thinned stands, and decreased slightly within the 2000 stems/ha stands (-0.03 ? 0.03 m2 ? 10-2).  Despite these differences, the effect of density on the changing rate of tree BA growth during the final 10 years was only marginally significant (F3,2=13.59; P=0.07; Table 2.3).  Fertilizer treatments did not significantly affect any changes in tree BA growth rates during the final 10 years (Table 2.3).     45  Figure 2.6.  Adjusted mean 15-year increments (1993 to 2008) of a) tree and b) stand BA of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).     00.511.522.53250 500 1000 2000 unthinnedm2 ? 10-2 a) BA/tree, 15-yr increment unfertilized fertilized05101520253035250 500 1000 2000 unthinnedm2/ha Target stand density (stems/ha) b) BA/ha, 15-yr increment    46  Adjusted mean 15-year stand BA increments among treated stands were positively correlated with crop tree density and increased by fertilization (Figure 2.6b).  Stand BA increment was 2.51 times greater within the fertilized 2000 stems/ha compared to the unfertilized 250 stems/ha stands; 20.64 ? 2.26 vs. 8.22 ? 0.17 m2/ha, respectively.  Despite the notable effect of PCT treatments on stand BA growth, the density effect was not significant (Table 2.2).  The effect of fertilizer on 15-year stand BA increment was marginally significant (F1,3=8.35; P=0.06), with 1.20 times greater BA growth resulting from fertilizer treatments; 14.75 ? 0.64 vs. 12.33 ? 0.25 m2/ha, respectively (Table 2.2).  Relative to the 1998 to 2003 five-year increment, the rate of stand BA growth increased during the final five-year increment (2003 to 2008) among the heavily thinned (250 and 500 stems/ha) as well as fertilized stands, while growth rates slowed within the more heavily stocked (1000 and 2000 stems/ha) as well as unfertilized stands (Table 2.3).  However, density and fertilizer effects did not significantly affect any changes in stand BA growth rates during the final 10 years (Table 2.3). Adjusted mean 15-year tree volume increments among treated stands were positively correlated with PCT intensity and further increased by fertilization (Figure 2.7a) and was 1.75 times greater within fertilized 250 stems/ha compared to unfertilized 2000 stems/ha stands; 10.62 ? 1.36 vs. 6.08 ? 0.46 m3 ? 10-2, respectively.  While the density effect was not significant, the effect of fertilizer was (F1,3=23.96; P=0.02), with 1.27 times greater tree volume growth resulting with fertilizer treatments; 9.18 ? 1.00 vs. 7.24 ? 0.54 m3 ? 10-2, respectively (Table 2.2).  All treated stands experienced greater volume of crop tree growth during the final five-year increment (2003 to 2008) than during the previous five-year increment (1998 to 2003).  This increased growth rate was particularly evident among the heavily thinned stands (250 and 500 stems/ha) relative to the lightly thinned stands (1000 and 2000 stems/ha); however, the density effect was not significant (Table 2.3).  The fertilizer effect was marginally significant (F1,3=8.27; P=0.06), with a mean of 1.19 times more tree volume growth experienced during the final five-year increment than in the unfertilized stands; 1.60 ? 0.11 vs. 1.35 ? 0.04 m3 ? 10-2 (Table 2.3).        47  Figure 2.7.  Adjusted mean 15-year increments (1993 to 2008) of a) tree and b) stand volume of young lodgepole pine crop trees among nine treatment stands.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).      00.020.040.060.080.10.120.14250 500 1000 2000 unthinnedm3 a) Volume/tree, 15-yr increment unfertilized fertilized020406080100250 500 1000 2000 unthinnedm3 Target stand density (stems/ha) b) Volume/ha, 15-yr increment    48  Adjusted mean 15-year stand volume increments among treated stands were positively correlated with crop tree density and increased by fertilization (Figure 2.7b).  Stand volume was 1.89 times greater within the fertilized 2000 stems/ha compared to the fertilized 250 stems/ha stands; 86.95 ? 1.77 vs. 46.07 ? 3.42 m3/ha, respectively.  Despite the notable effect of PCT treatments on stand volume increment, the density effect was not significant (Table 2.2).  The effect of fertilizer on 15-year stand volume increment was significant (F1,3=10.82; P=0.05), with 1.18 times greater stand volume growth resulting from fertilizer treatments; 70.24 ? 3.33 vs. 59.72 ? 4.81 m3/ha, respectively (Table 2.2).  The rate of stand volume growth increased during the final five-year increment (2003 to 2008) relative to the 1998 to 2003 increment among all stands.  Neither thinning nor fertilizer treatments significantly affected the rate of this increased growth (Table 2.3).   No significant density effect was noted among PCT stands for mean stand volume increment.  Although this statistical comparison did not include the unthinned stands, a visual inspection of the treatment means (Figure 2.7b; unthinned increment being within the range reported for the PCT stands) suggests that the mean 15-year stand volume increment of the unthinned stands was similar to that observed among the PCT stands. 2.3.3.2 Crown attributes  Crop tree crown dimensions (width, length, and live crown ratio) were noticeably affected by thinning and fertilization treatments and, of course, pruning (see Lindgren et al. (2007) for details regarding the initial 10 years) (Figure 2.8).  In 2008 (15 years after PCT and five years after the most recent fertilization application), mean crown area per tree among treated stands was positively correlated with PCT intensity and further increased by fertilization (Figure 2.9a).  Tree crown area was 1.83 times greater within fertilized 250 stems/ha compared to unfertilized 2000 stems/ha stands; 11.71 ? 0.86 vs. 6.39 ? 1.13 m2, respectively.  While the density effect was not significant, the effect of fertilizer was marginally significant (F1,4=6.79; P=0.06), with a mean of 1.20 times greater tree crown area resulting with fertilizer treatments; 9.92 ? 0.68 vs. 8.30 ? 1.09 m2, respectively (Table 2.5).      49  Figure 2.8.  Mean (n = 2 replicate sites) crop tree crown dimensions as measured at five-year intervals (1998, 2003, and 2008) among the nine treatment stands.               1998                              1998                           2003                        2008   (pre-pruning)              (post-pruning) height (m) pruned to a 3-m lift in 1998                                       not pruned uf                      f                               uf                     f                              uf                    f                            uf                   f                             uf     250                                                        500                                                1000                                              2000                             unthinned 10        8        6       4       2         0 PCT target density (stems/ha) uf = unfertilized                             f = fertilized 1 m    50   Figure 2.9.  Mean crown area of young lodgepole pine crop trees at the a) tree and b) stand levels among nine treatment stands measured in 2008; 15 years following the onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).         036912250 500 1000 2000 unthinnedm2 a) Crown area/tree unfertilized fertilized051015202530250 500 1000 2000 unthinnedm2 x 103 Target stand density (stems/ha) b) Crown area/ha    51  Table 2.5.  Mean (n = 2 replicate study areas) (SE) crown area and volume of young lodgepole pine crop trees among four density and two fertilization treatments measured in 2008; 15 years after the onset of treatments.  Results of split-plot analysis of covariance (ANCOVA) are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  Mean crown values  Split-plot ANOVA  Density Fertilizer Density ? Fertilizer  F3,3 P F1,4 P F3,4 P Area/tree (m2) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  3.49 0.17 6.79 0.06 0.12 0.94 10.72 (0.99) 10.37 (2.16) 8.44 (0.07) 6.90 (0.32)        unfertilized fertilized        8.30 (1.09) 9.92 (0.68)        Area/ha (m2 ? 103) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  28.08 0.01 0.63 0.47 0.05 0.98 2.72c (0.33) 5.64bc (1.53) 7.75b (0.01) 11.49a (0.34)        unfertilized fertilized        6.53 (1.07) 7.27 (0.02)        Volume/tree (m3) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  1.29 0.42 1.66 0.27 0.51 0.70 22.48 (3.86) 24.03 (9.02) 18.53 (2.28) 16.05 (2.38)        unfertilized fertilized        18.85 (4.60) 21.69 (4.17)        Volume/ha (m3 ? 103) 250 stems . ha-1 500 stems . ha-1 1000 stems . ha-1 2000 stems . ha-1  18.24 0.02 0.02 0.91 0.31 0.82 5.73c (1.13) 13.21bc (5.69) 16.98b (1.92) 26.79a (3.64)        unfertilized fertilized        15.49 (4.54) 15.87 (1.66)             52  At the stand level, mean crown area per ha among treated stands was positively correlated with crop tree density and increased by fertilization (Figure 2.9b).  Stand crown area was 4.49 times greater within fertilized 2000 stems/ha compared to unfertilized 250 stems/ha stands; 11.94 ? 0.84 vs. 2.66 ? 0.26 m2, respectively.  While the fertilizer effect was not significant, the density effect was (F3,3=28.08; P=0.01).  The 2000 stems/ha stands had significantly greater crop tree crown area than any of the lower density stands, while the 500 and 1000 stems/ha stands were statistically similar (DMRT; P=0.05) (Table 2.5).  Although heavy thinning and fertilization appeared to increase mean crown volume per tree (Figure 2.10a), neither the density nor fertilizer effect was significant (Table 2.5).  At the stand level, mean crown volume per ha among treated stands was significantly (F3,3=18.24; P=0.02) affected by PCT treatments (Figure 2.10b) with 4.67 times greater stand crown volume found within the 2000 compared to the 250 stems/ha stands; 26.79 ? 3.64 vs. 5.73 ? 1.13 m3 ? 103, respectively.  The 2000 stems/ha stands had significantly greater crop tree crown volume than any of the lower density stands, while the 500 and 1000 stems/ha stands were statistically similar (DMRT; P=0.05) (Table 2.5).      53  Figure 2.10.  Mean crown volume of young lodgepole pine crop trees at the a) tree and b) stand levels among nine treatment stands measured in 2008; 15 years following the onset of treatments.  Treatment stands include four PCT densities, both unfertilized and repeatedly fertilized, and an unthinned stand for comparison.  Numbers along the x-axis represent target thinning densities (stems/ha).  Unthinned stand densities ranged from 4000 to 10700 stems/ha.  Error bars indicate SE and are based on two replicate study areas (n = 2).      051015202530250 500 1000 2000 unthinnedm3 a) Crown volume/tree unfertilized fertilized01020304050250 500 1000 2000 unthinnedm3 x 103 Target stand density (stems/ha) b) Crown volume/ha    54  2.4 Discussion Intensive management is the norm within pine plantations of the southern US, where the economic benefits of incremental silviculture practices such as PCT and fertilization have been clearly demonstrated for loblolly pine plantations (Albaugh et al. 2004; Sword Sayer et al. 2004; Fox et al. 2007).  In contrast, management of plantations throughout Canada continues to be extensive rather than intensive, with little or no silviculture intervention beyond stand establishment, despite the clear potential for many benefits possible with increased tree culture (Lautenschlager 2000; Park and Wilson 2007).  Intensive silviculture would require a significant increase in long-term investments over extensive silviculture, which, is legitimately viewed as risky because of constantly changing economic, ecologic, and social constraints (Brown et al 1999).  Therefore, an improved understanding of treatment effects and potential benefits associated with intensive silviculture is needed, particularly for northern ecosystems (Lautenschlager 2000), before managers are likely to make intensive silviculture a part of standard practices. Prior to this study, the potential for intensive management within Canadian pine plantations remained largely theoretical, or demonstrated only at a small-scale (< 0.1 ha sample plots), as no other studies have reported on long-term tree responses (> 10 years) to operational-scale (several hectares) applications of intensive silviculture treatments.  This study represents a large effort that provides 15 years of tree growth response data from the most intensively managed plantations in Canada.  The total area of the treated stands (combined over two replicate study areas) was 158 ha, of which 140 ha were thinned, 108 ha were pruned, and 70 ha were repeatedly fertilized.  The five applications of fertilizers totaled 750 kg N/ha (as well as smaller amounts of P, K, S, B, and Mg) resulting in more than 52 tonnes of N being applied over an eight-year period. Treatment effects were monitored using a total of 360 variable-radius plots including 3600 permanently tagged crop trees (for quantifying growth of crop trees; sampled four times at five-year intervals) as well as 360 fixed radius plots (for quantifying total tree densities among height classes; sampled three times at five-year intervals). 2.4.1 Height class distribution When this study was initiated in 1993, stands were predominantly comprised of a single cohort of densely stocked (more than 4000 and 10000 stems/ha at the Summerland and Kelowna    55  study areas, respectively), 12- to 14-year old lodgepole pine trees, that were approximately 2.5 to 4.0 m in height.  As intended, the PCT treatments had a dramatic and significant effect on stand density (Table 1.1) and fertilizer treatments significantly increased the foliar N levels (Figure 2.3) without creating any nutrient imbalances or deficiencies (Lindgren et al. 2007).  The reduced competition among trees following thinning treatments appears to have also resulted in very low rates of mortality as very few crop trees were lost during the 15-year period post-PCT.  However, as tree canopies continue to close with fast growing crop trees and ingress of a new cohort of non-crop trees, increased competition is expected to initiate self-thinning among these managed stands.  While density of tall trees among the heavily thinned stands continued to increase 15 years post-PCT, densities have leveled off, or even decreased within stands that received the lightest thinning treatment (2000 stems/ha), indicating that self-thinning is just beginning within these dense stands.   The change in height class distribution can largely be explained by the increased growing space initially provided by the thinning treatments and, later, by the different rates at which this growing space was re-occupied, which was accelerated by fertilization.  Decreased competition for light, nutrient, and moisture resources provided by PCT would have enhanced both the growth of crop trees (discussed in detail below) and establishment of a new cohort of trees (i.e., ingress of non-crop trees).  Initially, the influx of non-crop trees was most pronounced within the shortest height class.  This was expected as this cohort was initially dominated by suppressed trees that were small enough to have avoided thinning, or newly germinated seedlings that established post-PCT.  The survival rate of this young cohort of non-crop trees differed markedly among the treatments.  Increased shading by taller trees and completion for soil resources were the most likely causes of the increased mortality of understory trees observed with time.  While PCT did have an expected significant impact on the number of trees in the dominant tree layer, it did not significantly affect the number of understory trees.  However, by 2008, fertilizer treatments resulted in significantly fewer trees in all understory height classes than observed within unfertilized stands.  This decrease in numbers of understory trees may at least partially be explained by the larger crowns observed among the overstory crop trees of fertilized compared to unfertilized stands (discussed below), which presumably increased the understory shading and mortality of shade intolerant understory trees.  In addition to increased crown size, fertilization    56  has also been shown to increase the leaf area index, which further decreased the amount of light available to the understory plants (Brockley 2007). It is the dominant, not understory, tree layers that are of particular interest to silviculturalists practicing even?aged management (the predominant management system applied to these stands).  Similarly, it is the stem volume and quality of these trees that determine the value of the stands for future harvest.  Fifteen years post-PCT, the number of trees in the dominant tree layer remained significantly different among all four levels of PCT.  However, the trend suggested that the number of trees in the dominant tree layer was converging among all of the density treatments.  This decrease in density effect occurred as shorter trees grew into this upper height class, taking advantage of space created by heavy thinning, while to a lesser extent, self-thinning began to decrease density within the densely stocked stands.  With time, the once obvious distinctions among the different thinning treatments (i.e., density and size of dominant tree layer) diminished.  The ability to accurately distinguish a crop- from non-crop tree within heavily thinned stands was already becoming difficult 15 years post-PCT, and any remaining differences will likely become further obscured with time.  Unfortunately, the contribution of ingress to total stand volume was not inferred from the height class distribution data; particularly when the height class which contributed the majority of volume was not finite (i.e., > 3.0 m).  Intuitively, it would seem that there will be considerable fiber value provided by the non-crop trees; however, without specific data, one is left to speculate as to the significance of this ingress.  A pilot study carried out within treatment stands at the Kelowna study area during the fall of 2009 (16 years post-PCT) estimated the stem volume of all trees taller than 1.3 m using 20 fixed-radius plots (3.99 m radius; 50 m2) per treatment stand and distinguished crop- from non-crop trees (unpublished data).  This study found that thinning increased the proportional contribution of ingress to total stand volume by as much as 1.47 times within the 250 stems/ha unfertilized stands, and that this proportion decreased with increasing crop tree density.  This same study found the fertilizer effect on total stand volume appeared to be negligible; however, a density effect was evident.  When averaged across unfertilized and fertilized stands, the impact of PCT was apparent, with a mean of 39, 31, 12, and 2% of the total stand volume attributed to ingress within the 250, 500, 1000, and 2000 stems/ha stands, respectively.  A portion of the stand volume contributed by ingress will undoubtedly be    57  non-merchantable (i.e., too small).  However, the stand-level productivity of crop trees (discussed below) should not be interpreted as total productivity, as ingress will clearly mitigate some of the stand volume losses resulting from intensive PCT.  2.4.2 Crop trees Height growth of non-repressed lodgepole pine has been recognized by other studies to not benefit from increased resources (Johnstone 1985; Brockley 2007), as trees allocate increased photosynthate production into crown and stem radial growth more than height (Tamm et al. 1999).  Amponsah et al. (2004) also hypothesized that the higher flow capacity (increased hydraulic conductivity, sapwood permeability, and leaf specific conductivity) observed within lower branches following repeated fertilizations, may cause water stress and reduced growth for the upper portions of the tree, including the terminal leader.  Brockley (2007) reported that annual fertilization applications to lodgepole pine resulted in less height growth than periodic applications.  Increased resources (light from thinning and nutrients from fertilization) may also stimulate the growth of non-crop plants, which then increase competition, possibly eliminating much of the desired growth response intended for the crop trees (Powers and Reynolds 1999).    For these reasons, excessive thinning and fertilization may, with time, result in decreased height growth. Although mean annual height increments decreased among all stands at a similar rate throughout the first 10 years (Lindgren et al. 2007), the rate of decrease during the final five years became significantly less within the fertilized compared to unfertilized stands.  It is speculated that the increased competition and shading resulting from accelerated crown development within fertilized stands is beginning to stimulate a vertical growth response or, conversely, slow its retardation.  The significantly greater DBH and BA per tree increments observed among heavily thinned stands reported for the initial 10-years post-PCT (Lindgren et al. 2007) was not evident when considering the entire 15-year growth increment.  The loss of the density effect for DBH growth for this longer period likely resulted from a greater decrease in rates of diameter and BA growth during the final five years within the heavily compared to lightly thinned stands.  A likely cause was increased competition from non-crop trees as well as well-developed herb and shrub layers.  Mean BA per tree growth continued to accelerate during years 11-15 among all    58  treatments, except for the slowest growing stand (2000 stems/ha).  The significantly enhanced DBH and BA growth rate resulting from fertilization reported for the initial 10-year increment (Lindgren et al. 2007) continued for the duration of this 15-year study.  The positive response in diameter growth without a similar response in height growth resulted in decreased HDR; particularly within fertilized stands.  This decrease in HDR (i.e., increase in stem taper) was established within the first five-year period and remained statistically similar thereafter.  Because HDR remained relatively constant during the final 10 years, even though this period included three of the five fertilization treatments, fertilization effects on HDR appear to be fully expressed following one or two applications and do not increase with additional fertilization.  With time, HDR should increase among fertilized stands as accelerated crown development and, to a lesser extent, dominance of non-crop trees, increases competition and stimulates height growth among crop trees. At the tree level, mean 15-year crop tree volume increment followed the same trend as reported for the initial 10-year increment (Lindgren et al. 2007); the increment was enhanced by thinning (although not significantly) and significantly increased by fertilization.  Most studies evaluate the effectiveness of thinning treatments relative to no thinning and nearly all indicate significantly enhanced tree volume growth rates with thinning (e.g., Weetman et al. 1995; Valinger et al. 2000; Blevins et al. 2005; Zhang et al. 2006; Jim?nez et al. 2011; Soucy et al. 2012).  While my experimental design did not allow a comparison of growth rates between thinned and unthinned stands, it is reasonable to expect that tree volume increment was enhanced by thinning as has been consistently reported for other studies.  This expectation was also supported by a visual comparison of tree volume increment provided by Figure 2.7a. The hypothesis that 15-year, tree-level growth increments of lodgepole pine crop trees would be enhanced by PCT and repeated fertilization was partially supported by the results.  While visual comparison of the density effect on DBH, BA, and tree volume were in support of this hypothesis (FiguresFigure 2.4b, Figure 2.6a, and Figure 2.7a), the statistical comparison, which did not include the unthinned control, did not detect a significant density effect among the thinned stands.  The fertilizer effect did support the hypothesis in that DBH, BA, and tree volume, were all significantly enhanced by repeated fertilization.       59  At the stand level, despite the significant impact that PCT had on crop tree numbers, 15-year BA and volume increments were statistically similar across density treatments.  This contrasted with the initial 10-year increment, where higher density stands had greater BA/ha (marginally significant; P=0.08) and volume/ha increments (P<0.01) than lower density stands (Lindgren et al. 2007).  This pattern indicated that the enhanced growth rates observed at the tree level among heavily thinned stands, while not significant, was enough to compensate for some of the losses of stand level crop tree productivity caused by heavy thinning treatments.  If the contributions that non-crop trees make to net total stand volume were included, it became apparent that initial losses associated with thinning can decrease substantially over time.  Although net total stand volume can be significantly greater than merchantable volume, the significance of non-merchantable trees should not be discounted.  For several reasons, non-merchantable trees may have considerable value as, for example, future crop trees, advanced regeneration (should anything damage the overstory crop, such as a mountain pine beetle infestation), carbon sinks, habitat structure, and for visual quality objectives.  Alternatively, the development of an understory of non-merchantable trees may be counter to the objectives of a stand.  A manager may prescribe that such ingress be controlled (i.e., thinned), for example, to decrease competition with overstory trees, enhance and prolong suitable range conditions for grazing ungulates, and decrease fire risk by reducing ?ladder fuels?.   Fertilizer treatments can further mitigate losses associated with heavy thinning.  Averaged over 15 years, the fertilizer treatments resulted in an increase to mean annual crop tree volume increment of 0.70 ? 0.39 m3/ha/year.  The longevity of this beneficial fertilizer effect is of key interest as the repeated applications had the objective of achieving a sustained rather than temporary increase in tree and stand growth (Tamm et al. 1999; Brockley 2007).  It is encouraging to note that the significantly enhanced rate of growth within fertilized stands was maintained during the final five years (2003 to 2008); a period that did not include any fertilizer applications.  However, only future sampling will be able to determine how long this enhanced growth will be sustained.  Weetman et al. (1995) reported a significant fertilizer effect lasting for seven to nine years following a fertilization regime similar to this study (total of 672 kg N/ha applied over 10 years) applied to jack pine (Pinus banksiana Lamb.).  While my study assessed only a single level of fertilizer application, Brockley (2007) investigated several levels of fertilizers (both different formulations as well as frequencies of application) over a period of 12    60  years.  The response of lodgepole pine crop trees in his study showed clear evidence of diminishing returns with increasing intensity of fertilization.  Annual applications of fertilizer provided only slightly increased, or even decreased mean annual stand volume increments relative to various fertilizer formulations applied at six-year intervals (Brockley 2007).  Weetman et al. (1995) also reported a negative dose response as jack pine treated with a total of 1344 kg N/ha (repeated applications over 10 years) was significantly less productive than controls, primarily because of the combined effects of competition-induced mortality.  Interestingly, such negative dose responses have not been reported for southern pines such as loblolly, which respond favorably to large, and frequent, nutrient additions (Albaugh et al. 2004).  With only a single level of fertilization treatment, the current study is unable to comment on whether or not a negative dose response has occurred.  However, the conclusions of Brockley (2007) and Weetman et al. (1995) suggest that even larger gains in wood production may be possible with a less intensive fertilization prescription. The hypothesis that 15-year, stand-level growth increments of lodgepole pine crop trees would be enhanced by PCT and repeated fertilization was partially supported by the results.  The density effect on stand BA and volume growth was not significant, while fertilization significantly enhanced both.  However, the lack of significant density effect can, be interpreted as being in support of the hypothesis that thinning initially caused a significant reduction is stand-level growth rates (Lindgren et al. 2007).  The loss of density effect with time indicates that stand-level growth rates among heavily thinned stands must have increased relative to the less intensively thinned stands. In 2008, no density effect remained for mean crop tree crown area or volume.  Given that pruning 10 years earlier removed more than 40% of crown area and 60% of crown volume from the 250, 500, and 1000 stems/ha stands, and none from the 2000 stems/ha stands (Lindgren et al. 2007), the lack of density effect can only be explained by rapid regrowth of crowns within the pruned stands, as well as the self-pruning that occurred within the 2000 stems/ha stands after 1998.  The increase in crown area of crop trees following fertilization of thinned stands was expected and can be explained by the increased photosynthetic productivity (more photosynthates available for growth), as well as the relatively greater allocation of photosynthates to root systems for trees growing within nutrient poor vs. rich sites (Power et al.    61  2012).  While fertilizer treatments resulted in only a marginally significant increase in tree crown area, the effect on leaf area index (LAI) was likely substantial and has been reported in other studies (Vose and Allen 1988; Brockley 2007).  Although no measurements of LAI, crown density, or understory light conditions were made during my study, other correlated observations did support the suggestion that fertilization decreased understory light conditions.  For example, the accelerated rate of self-pruning observed within the fertilized vs. unfertilized 2000 stems/ha stands during the final five years (Figure 2.8) resulted in a 19% shorter crown length in 2008 (crown length of 5.3 vs. 7.3 m) and could be explained by increased understory shading (Lotan and Critchfield 1990; Power et al. 2012).  In addition, the significantly fewer understory trees within the fertilized stands observed in 2008 (Figure 2.2) could also be explained by increased understory shading.  The 1.20 times increase in tree crown area associated with fertilizer treatments, coupled with an increase in LAI, may largely explain the positive growth response observed for the crop trees as photosynthetic productivity is primarily a function of these crown attributes (Oliver and Larson 1996; Power et al. 2012). At the stand level, the density treatment that received the least thinning and no pruning (2000 stems/ha stands) had significantly greater crop tree crown area and volume than the lower density stands, which had also been pruned.  However, the lack of significant difference in total crop tree crown area and volume among the 500 compared to both the 250 and 1000 stems/ha stands was interesting as it suggested that the enhanced crown development observed within heavily thinned stands, while not significant, had mitigated some of the losses to total crop tree crown quantity associated with PCT.  The contributions that the non-crop tree crowns make would also have further minimized any effects that PCT had on total tree crown area and volume.  The negative influences that treatments of PCT and fertilization can have on wood form (i.e., increased taper) and wood quality (e.g., decreased density, increased branching), and the mitigating effects of pruning were discussed in detail by Lindgren et al. (2007). 2.4.3 Carbon sequestration and resilience to climate change Carbon sequestration is one way to moderate the continuing increase in atmospheric carbon dioxide (CO2), and therefore is an important process for mitigating climate change.  Trees provide 74% of all aboveground terrestrial carbon and are an important sink for atmospheric CO2    62  (Birdsey 1992).  The enhanced carbon sequestration associated with increased stand volume of crop trees is another non-timber value affected by PCT and fertilization.  Jim?nez et al. (2011) noted that intense PCT (more than 90% reduction in density) of densely stocked, post-fire maritime pine (Pinus pinaster) in northwestern Spain resulted in increased foliar efficiency and growth of residual saplings that compensated for much of the carbon storage lost by thinning, at least in the short-term (five years post-thinning).  Blevins et al. (2005) noted that stand-level volume growth rates returned to control levels just four years following thinning within repressed stands of lodgepole pine in BC.  Stand-level carbon storage losses resulting from PCT are further mitigated if accounting for the contributions of residues and understory vegetation (Campbell et al. 2009; Jim?nez et al. 2011).  While stand-level growth rates may be restored soon after thinning, total stand volume may take longer to catch up to unthinned levels (Soucy et al. 2012).   Huettl and Zoettl (1992) noted that the potential to increase carbon storage capacity in forests is often limited by poor nutrient availability.  In addition, Hoen and Solberg (1994) indicated fertilization as one of the best methods to increase the net CO2 fixation.  Enhanced forest productivity resulting from fertilization can increase carbon storage capacity not only by increasing volume and biomass of trees (as was reported in my study), but also by increasing the carbon input to forest soils (Jandl et al. 2007; Nave et al. 2009).  Forest soils are by far the largest carbon sink within the forested ecosystem, contributing 1.4 times the carbon storage compared to all aboveground components put together (Birdsey 1992). In addition to mitigating the effects of climate change through enhanced carbon sequestration, incremental silviculture treatments may enhance the resiliency of managed stands to the difficult growing conditions predicted as a result of climate change (Miller et al. 2007).  For example, the increased intensity of summer droughts associated with climate change would negatively impact the health and survival of young and older forests alike; particularly those growing in areas that historically experience water stress, such as much of the lodgepole pine forests of BC?s interior.  Thinning of densely stocked forest stands has been suggested as an adaptive management strategy to enhance soil water content and reduce water stress for forests growing within dry environments (Papadopol 2000).  Jim?nez et al. (2008) reported that young maritime pine trees retained during intensive PCT treatments, while experiencing short-term    63  decrease in transpiration (shock associated with sudden increase in crown illumination, exposure to wind, and decreased stem water conductivity), were able to take advantage of soil water availability provided by decreased competition and rain interception relative to unthinned stands.   Fertilization may similarly enhance the resiliency of trees growing in nutrient poor conditions; however, negative impacts to ectomycorrhizal (EM) fungi and the stress tolerance that they impart to their tree symbiont (e.g., enhanced nutrient and water absorption, tolerance to low pH soil conditions, protection from root pathogens), could offset anticipated benefits of fertilization, and should be carefully considered.  Studies have noted strong associations of EM fungal species with N availability, suggesting that fertilization will cause shifts in EM species composition and potential loss of beneficial species (Leckie et al. 2004; Berch et al. 2006; Kranabetter et al. 2009).  Of particular interest for lodgepole pine forests, is the impact that fertilization will have on the Suillus tomentosus/Pinus contorta tuberculate EM relationship, which has recently been discovered to have biologically significant nitrogen fixing potential and nitrogenase activity equal to 10 to 15% of that reported for the symbiotic relationship of Alnus/Frankia (Paul et al. 2007). 2.5 Conclusions  My results indicated a trend of increasing tree-level growth (15-year increment of DBH, BA, and volume) with increasing levels of PCT; however, the density effect among thinned stands was not significant.  Over the 15-year period following PCT, the increased tree-level growth, while not statistically significant, resulted in similar stand-level increments of BA and volume.  This lack of stand-level density effect 15-years post PCT was important considering the striking difference in crop tree density following thinning treatments (i.e., each increasing level of thinning intensity corresponded with a decrease in crop tree density of 50%).  While stand-level productivity of crop trees was still clearly correlated with crop tree density 15 years post-PCT, the lack of significant difference indicated that more crop tree wood was being produced within the heavily than lightly thinned stands, which mitigated the initial loss of productivity associated with intensive thinning.    The five fertilization treatments, which were applied at two-year intervals during the initial nine years, had a consistent and significant positive effect on 15-year growth increments    64  of DBH, BA/tree, volume/tree, BA/ha, and volume/ha (increased by factors of 1.17, 1.28, 1.27, 1.20, and 1.18, respectively).  Because tree stem growth was primarily governed by the production of photosynthates, the enhanced growth of crop trees can be largely explained by the enhanced crown characteristics observed following thinning (increased crown area) and fertilization (increased crown area and presumably LAI).  It is interesting to note that the increased productivity of trees within the heavily thinned stands occurred despite pruning five years post-thinning (i.e., significant decrease in crown size).  This pattern indicated that the negative effects on stem form and wood quality properties associated with intensive thinning and fertilization (e.g., increased taper, decreased clear wood and wood density) can by mitigated with pruning without any detectable negative impact on the growth potential of crop trees only 10 years after pruning.  Stand-level productivity was only quantified accurately for crop trees selected during the initial PCT treatment in 1993.  However, the contributions of non-crop trees to total stand productivity appeared to be substantial, particularly within the heavily thinned stands.  Considering the contributions of non-crop trees (assuming that these trees are of a merchantable species) and the trend of increased tree growth rates, the initial loss of total stand productivity following thinning to below full stocking levels may be short-lived.    Intensive silviculture treatments of PCT and repeated fertilization, as has been demonstrated in the southern US, have the potential to significantly enhance wood production of lodgepole pine stands.  A range of PCT intensities (even below full stocking) and levels of fertilization are available to managers, the various combinations of which could provide for stand conditions and growth trajectories suited for a variety of stand objectives.  A regional-level management strategy that utilizes a diverse range of silvicultural treatments (ranging from no management to intensive management) should be able to integrate several objectives including, for example, enhanced fiber and round wood production, reduced rotation lengths, and increased carbon sequestration.        65  3 Plant community diversity and structural response to pre-commercial thinning and repeated fertilization of young lodgepole pine stands: 15 years of monitoring 3.1 Introduction  Given the prediction of increasing global demand for wood products (Raunikar et al. 2010) concurrent with the increasing constraints on the harvestable land base imposed by measures to conserve biodiversity (Hunter and Schmiegelow 2011; Leroux and Kerr 2013), enhanced forest productivity may be required if supply of wood fiber is to keep pace with demand.  This is particularly true for the forest industry throughout boreal and sub-boreal forests of North America, where management is predominantly extensive rather than intensive (Lautenschlager 2000; Park and Wilson 2007).  The enhanced productivity possible with intensive management has been clearly demonstrated for pine (Pinus spp.) plantations throughout the southern United States (US), where tree improvement programs and treatments of thinning (pre-commercial and commercial) and fertilization have resulted in tree growth rates that have more than doubled and rotation lengths cut by more than 50% (Fox et al. 2007).  Similar management strategies have also significantly increased yields of pine and spruce (Picea spp.) stands throughout the boreal forests of Scandinavia (Nohrstedt 2001; Saarsalmi and M?lk?nen 2001).  Intensive management of Canada?s second-growth forests may provide similar benefits as observed in the southern US and Scandinavia and, therefore, should likely be promoted to ensure a competitive forest industry (Brooks 1997; Sutton 1999).  While the need for enhanced forest productivity is clear, there are concerns regarding potential negative impacts of intensive management on biodiversity (McDonald and Lane 2004).   Silviculture, because of its historic narrow focus on enhanced productivity of one or two species of crop trees, is often associated with a decrease in biodiversity, particularly plant diversity (Gilliam and Roberts 1995).  However, there is little support for this generalization as studies have reported a range of results, from increased (Thomas et al. 1999; Thysell and Carey 2001; Battles et al. 2001; Lindgren et al. 2006) to decreased levels of plant diversity following management compared to unmanaged controls (Hansen et al. 1991; Elliott et al. 1997).  A potential unifying mechanism that could accommodate this range of observations is the    66  intermediate disturbance hypothesis, which suggests that moderate disturbance prevents a few species from dominating resources, but that severe disturbance creates an environment that few species can tolerate (Battles et al. 2001).  Because silvicultural practices (e.g., thinning) represent a type of disturbance, it follows that intensive silviculture may decrease diversity (Wang and Chen 2010), raising concerns about the suitability of such treatments and their impact on biodiversity. Because intensive silviculture treatments strive to increase crop tree productivity, the suitability of such treatments is questioned on the grounds that increased levels of productivity are often associated with decreased levels of diversity.  Although the form of the productivity-diversity relationship (PDR) has been debated for decades (e.g., see exchange between Adler at al. (2011), Fridley et al. (2012), Pan et al. (2012) and Grace et al. (2012)), a predominant view is that of a hump-shaped model, where diversity is increased at low levels of productivity, but decreased at high levels of productivity (Waide et al. 1999; Mittelbach et al. 2001).  The suggested mechanisms are that, increased productivity leads to increased diversity at low productivity, as enhanced resource availability facilitates larger populations with low rates of extirpation.  At high productivity, further increases to productivity leads to decreased diversity as the heterogeneity of limiting resources decreases and competitive exclusion increases.  Recently, studies are beginning to challenge the hump-shaped PDR, suggesting instead a positive linear relationship (Gillman and Wright 2006; Bai et al. 2007) or no general relationship at all (Adler et al. 2011; Grace et al. 2012).  The continued debate surrounding the PDR form clearly indicates that accurate predictions regarding diversity should not be expected from existing models and that decisions regarding management should be based on real data gained from context-specific monitoring.  This study was designed to address the concern over potential negative impacts of intensive management on biodiversity by monitoring long-term plant community response to incremental silviculture treatments applied to young stands of lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.).  Lodgepole pine is the major coniferous tree species in inland areas of the Pacific Northwest where it occupies ~20 million ha in Canada (mostly in British Columbia and Alberta) and about six million ha in the western US (Cole and Koch 1996).  Lodgepole pine is a pioneer species that is perpetuated through repeated fire disturbance, and    67  stands often occupies sites of low-N status (Brockley et al. 1992).  In addition, this species often regenerates at very high densities, resulting in conditions with repressed growth rates (Blevins et al., 2005).  Given the overcrowded and low nutrient conditions of young lodgepole pine stands, silvicultural treatments of pre-commercial thinning (PCT; Johnstone 1985; Cole and Koch 1996) and fertilization (Weetman 1988; Brockley 1996) have significant potential to increase crop tree productivity (Lindgren et al. 2007; Chapter 2) and, therefore, provide a useful stand type for studying the effects of intensive management. In this chapter, I examine the impacts of PCT and repeated fertilization on plant community (herbs, shrubs, trees, and combined total) abundance, species diversity, and structural diversity as these attributes provide both direct and indirect measures of stand-level biodiversity.  Not only is the plant community itself a major component of biodiversity, but it also provides much of the physical attributes of habitat for all species of wildlife (e.g., forage, hunting grounds, cover, nesting sites) (Carey et al. 1999; Sullivan et al. 2001).  It also modifies the environmental conditions of these habitats, both above (air temperature, wind speed, humidity, and shading) and below ground (soil temperature, moisture, and nutrient content) (Berch et al. 2006).  An improved understanding of how plant community attributes respond to PCT and fertilization would enhance our ability to determine the suitability of these silvicultural treatments for sustainably managing our forests.    This study was designed to test four hypotheses (H), phrased as predictions: (H1) the application of a range of PCT intensities will increase abundance of the herb and shrub layers; (H2) fertilization treatments will increase abundance of all plant layers (herbs, shrubs, trees, and combined total); (H3) PCT and fertilization treatments will decrease species and structural diversity of all plant layers; and (H4) enhanced abundance of vegetation will result in a decline in total species diversity. 3.2 Methods 3.2.1 Study areas Three study areas were chosen on the basis of having candidate stands of young (12 to 14 year old) lodgepole pine that had relatively uniform tree cover, comparable diameter, height, and density of trees prior to stand treatments.  Location, proximity (boundaries), and size of    68  candidate stands were determined by a balance between adequate interspersion of experimental units (Hurlbert 1984) and the logistics and access for conducting the operational-scale treatments of PCT and fertilization.  The study areas are named after nearby towns (Summerland and Kelowna) or forest region in which they are located (Cariboo).  Both the Summerland and Kelowna study areas are located within the Montane Spruce biogeoclimatic zone (dry mild subzone; MSdm), whereas the Cariboo study area is within the Sub-Boreal Spruce biogeoclimatic zone (dry warm subzone; SBSdw) (Meidinger and Pojar 1991). Prominent herb species throughout all three study areas included yarrow (Achillea millefolium L.), rosy pussytoes (Antennaria microphylla Rydb.), field pussytoes (A. neglecta Greene), racemose pussytoes (A. racemose Hook.), heart-leaved arnica (Arnica cordifolia Hook.), fireweed (Epilobium angustifolium L.), bunchberry (Cornus canadensis L.), wild strawberry (Fragaria virginiana Duchesne), whiteflowered hawkweed (Hieracium albiflorum Hook.), Arctic lupine (Lupinus arcticus S. Wats.), pinegrass (Calamagrostis rubescens Buckley), and common dandelion (Taraxacum officinale Weber).  Additional important herb species at the Cariboo study area only included fringed aster (Aster ciliolatus Lindl.), northern bedstraw (Galium boreale L.), creamy peavine (Lathyrus ochroleucus Hook.), red-clover (Trifolium pratense L.), white clover (T. repens L.), American vetch (Vicia americana Muhl.) and smooth brome (Bromus inermis Leyss).   Prominent shrub species included Sitka alder (Alnus viridis ssp. sinuata (Chaix) DC. (Regel) A. L?ve & D. L?ve), twinflower (Linnaea borealis L.), black twinberry (Lonicera involucrata (Rich.) Banks), red twinberry (L. utahensis Wats.), falsebox (Pachistima myrsinites (Pursh) Raf.), black gooseberry (Ribes lacustre (Pers.) Poir.), red raspberry (Rubus idaeus L.), kinnikinnick (Arctostaphylos uva-ursi (L.) Spreng.), willow (Salix spp.), birch-leaved spirea (Spiraea betulifolia Pall.), and several species of Vaccinium (L.). Other common shrub species in the Cariboo study area included Saskatoon berry (Amelanchier alnifolia Nutt.), tall Oregon-grape (Mahonia aquifolium Pursh), prickly rose (Rosa acicularis Lindl.), and soopolallie (Shepherdia canadensis (L.) Nutt.).   Prominent trees, in addition to lodgepole pine, included Douglas-fir (Pseudotsuga menziesii (Mirbel) Franco. var glauca (Beissn.) Franco), subalpine fir (Abies lasiocarpa (Hook.)    69  Nutt.), hybrid interior spruce (Picea engelmannii ?  P. glauca (Moench) Voss), and trembling aspen (Populus tremuloides Michx.). A complete description of these three study areas is provided in chapter 1. 3.2.2 Experimental design The three study areas acted as regional replicates (blocks).  Within each replicate, there were five experimental plots where lodgepole pine stands were treated (PCT) in the following randomized block design: very low density (target 250 stems/ha), low density (target 500 stems/ha), medium density (target 1000 stems/ha), high density (target 2000 stems/ha), and unthinned (at least 4000 stems/ha).  Fertilization treatments were applied to one half of each of the thinned units, resulting in a total of nine stands per study area, as follows: 1) 250 stems/ha, 2) 250 stems/ha with fertilization, 3) 500 stems/ha, 4) 500 stems/ha with fertilization, 5) 1000 stems/ha, 6) 1000 stems/ha with fertilization, 7) 2000 stems/ha, 8) 2000 stems/ha with fertilization, and 9) unthinned.  A fertilized, unthinned experimental unit was not included in this design as this treatment combination would not be part of any management prescription.  Additional details regarding the nine experimental units are provided in chapter 1. 3.2.2.1 Treatments PCT to target densities was done at all study areas in the late summer-early fall of 1993, at which time stands were 12 to 14 years old.  A more complete description of the density treatment is provided in Chapter 1 and Lindgren et al. (2007).   Five years following thinning, heavily thinned stands (? 1000 stems/ha) were pruned to a height of approximately 3.0 m using manual pruning saws.  All thinning and pruning debris (slash) was left on site.  Fertilization treatments were initiated during fall 1994 and were repeated at two-year intervals for a total of five applications; spring 1997, fall 1998, fall 2000, and spring 2003.  These five treatments applied 100, 200, 150, 150, and 150 kg/ha nitrogen, respectively (total of 750 kg N/ha), in conjunction with other nutrients.  Complete descriptions of density and fertilization treatments are provided in Chapter 1. 3.2.2.2 Vegetation sampling Three permanent vegetation transects were randomly established throughout each treatment stand during 1993 and sampled prior to PCT.  Sampling was carried out during the    70  period of peak productivity (July-August), and continued annually until 2003.  A final sample took place in 2008.  Plant species were identified in accordance with Hitchcock and Cronquist (1973).  All grass species were grouped together and collectively referred to as ?grasses?.  To ensure consistency of any sampling bias, I conducted all of the vegetation sampling.  Vegetation sampling (herbs, shrubs, and trees) was designed to measure the effects of PCT and repeated fertilization on individual plants species as well community-level attributes such as species diversity and structural diversity.  This was accomplished by transect sampling that provided estimates of abundance by species.  A vegetation transect measured 5 m wide by 25 m long and was made up of five contiguous 5 ? 5-m plots. Each plot contained three sizes of nested sub-plots: the entire 5 ? 5-m plot for sampling trees; a 3 ? 3-m sub-plot for sampling shrubs; and a 1 ? 1-m sub-plot for sampling herbs.  All plants were subdivided into six height classes: 0-0.25, 0.25-0.5, 0.5-1.0, 1.0-2.0, 2.0-3.0, and > 3.0 m (Walmsley et al. 1980).  A given species (an individual plant or group of plants of that species), was described by a visual estimate of percent cover within the appropriate height class.  Estimates were made to the nearest 1 percent, and cover values ? 0.5% were recorded as ?trace? and were given a value of 0.5% within subsequent calculations.  Height class was determined by the tallest point of the plant.  Therefore, depending on the range of heights for plants of a given species, a species could be described by separate estimates of percent cover among different height classes.  Percent cover was quantified, not by the outline (or drip line) of the plant, but rather by an estimate of the percent of ground covered by a vertical projection of the plant?s silhouette on the ground.  Because 0.01 ha equals 100 m2, the estimated percent cover can also interpreted as m2/0.01 ha.  Crown volume index was then calculated for each plant species as the product of percent cover and its corresponding height (Stickney 1985).  In this way, crown volume index provided the volume of a cylinder and represents the space occupied by the plant in the community (m3/0.01 ha).   3.2.2.3 Diversity measures Plant communities were described in terms of their composition (species richness), relative abundance (diversity), and vertical complexity (structural richness and diversity).  All diversity measures were calculated separately for herb, shrub, and tree layers, as well as for a combined total layer, which included all species.  Species richness was the total number of species sampled within a stand.  Estimates of total species richness were calculated as the sum of the species richness from the herb, shrub, and tree layers.      71  Species diversity was estimated using both the Shannon-Wiener (Pielou 1966) and Simpson indices.  The Shannon-Wiener and Simpson indices are more sensitive to changes in rare and common species, respectively (Magurran 2004).  However, for the plant community data, interpretation of either of these diversity measures resulted in the same conclusions regarding treatment effect.  Therefore, only a single diversity index (Simpson) is presented.  This diversity index has good discriminant ability and is well represented in the ecological literature (Burton et al. 1992).  For all diversity calculations (including structural diversity), crown volume index was used for the relative abundance component of the diversity index.      Structural diversity (foliage height diversity; MacArthur and MacArthur 1961) was also described using the Simpson index.  The structural diversity index was calculated with height classes acting as ?species?.  Thus, structural richness was the total number of height classes occupied by the various vegetative layers.  To account for the unequal ranges of height classes, abundance was standardized on a per meter basis. All diversity estimates were calculated at the subplot level and averaged across the five subplots on a given transect.  Means from the three independent transects were then averaged to provide treatment means for a given study area. 3.2.2.4 Crown volume index as a proxy for productivity  A direct measure of net primary productivity such as actual evapotranspiration (a measure of the biologically available energy; Rosenzweig 1995) would have been ideal for investigating the relationship between productivity and diversity; however, such data were not available.  Instead, a proxy for productivity was used.  Maximum standing biomass is commonly used as an index of productivity (Bai et al. 2007; Guo 2007), although Gillman and Wright (2006) have pointed out that this is most appropriate for communities dominated by annual vegetation as the relationship between biomass and productivity can become decoupled in perennial vegetation.   I carried out three separate pilot studies (unpublished data) that found strong and significant correlations between crown volume index and biomass.  The first study was carried out in 2007 and involved clipping 180 herb plots and correlating the estimated crown volume index with the corresponding oven-dried biomass.  This herb study revealed a very strong and    72  significant positive linear relationship for both graminoids (R2 = 0.95; F1,17 = 154.50; P < 0.01) and forbs (R2 = 0.79; F1,17 = 26.12; P < 0.01).  The second study was carried out in 2010 and investigated the correlation between crown volume index and biomass for shrubs using similar methods as described for herbs.  This shrub study revealed a strong and significant positive linear relationship for dwarf (R2 = 0.46; F1,95 = 84.16; P < 0.01), low (R2 = 0.77; F1,98 = 331.92; P < 0.01), and tall shrubs (R2 = 0.90; F1,136 = 312.77; P < 0.01).  The third study involved detailed measurements of diameters and heights of all trees as a means to estimate mean total biomass of the tree layer (stem, branches, and foliage) among my nine treatment stands.  These biomass data were collected in 2009 and were correlated with the corresponding mean crown volume index data of the tree layer collected in 2008.  I assumed a minimal difference in tree layer abundance following one year of growth and found a strong and significant positive linear relationship (R2 = 0.78; F1,7 = 24.79; P < 0.01).  While a replicated study correlating total crown volume index to total biomass is lacking, these three pilot studies clearly indicate that a consistent and strong linear relationship exists between crown volume index and direct measures of biomass.  Consequently, total crown volume index data were deemed appropriate as a surrogate for productivity for investigating relationships between species diversity and site productivity. 3.2.2.5 Statistical analysis  The three regional replicates functioned as blocks and were assigned as a random factor.  The experimental design restricted the randomness of fertilizer treatment allocation (i.e., applied to one-half of each of the thinned stands), and hence a split-plot analysis of variance (ANOVA) was used to compare treatment means.  The density and fertilizer treatments were assigned as the main- and split-plots, respectively.  Time (year of sample) was assigned as a split-split plot factor when analyzing data over multiple years.  Before performing any analyses, data not conforming to properties of normality and equal variance were subjected to transformations to better approximate the assumptions required by ANOVA (Zar 1999).  For count data (e.g., species richness), logarithmic, square root, or an area hyperbolic sine function was used, depending on the presence of zero counts and the magnitude of variance relative to the mean.  For variables based on percent cover data (e.g., crown volume index), an arcsine transformation was used (Fowler et al. 1998).    73  Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare mean values based on ANOVA results (Saville 1990).  In all analyses, the level of significance was at least ? = 0.05.  P-values ranging from 0.06 to 0.10 were reported as marginally significant.  The large size of experimental units, high cost of treatment installations, and lack of operational feasibility of certain treatment combinations, resulted in an experimental design that lacked specific experimental units.  Most notably, the experimental design was unbalanced because of the lack of a fertilized unit for the unthinned density.  In order to preserve the ability to compare treatment means among managed (PCT and fertilized) and unmanaged stands (unthinned with no fertilizer), data were analyzed together.  The resulting degrees of freedom were decreased accordingly to account for the unbalanced design. In 2005, a mountain pine beetle (Dendroctonus ponderosae Hopk.) epidemic caused a significant disturbance throughout the Cariboo study area, affecting the 2008 vegetation sample from this replicate.  As a result, the 2008 Cariboo data was omitted from the analyses, resulting in reduced degrees of freedom to account for this missing data.  Regression analyses were used to investigate relationships between mean total crown volume index (volume of herbs, shrubs, and trees combined used as an index of stand productivity) and several different measures of diversity, including species richness, species diversity, and structural diversity.  As recommended by Gillman and Wright (2006), ordinary least squares regression analyses were used to investigate these relationships.  Regression models were limited to linear and second order polynomials.  Where appropriate, data were pooled across replicates and sample years.  Before pooling, regression relationships were assessed for differences in intercept and slope among study areas and sample years.  Data were analyzed separately where significant differences were detected.  The strength of a relationship (i.e., degree of correlation) was described with the coefficient of determination and referred to as very weak (R2=0.00-0.03), weak (R2=0.04-0.14), modest (R2=0.15-0.47), strong (R2=0.48-0.78), and very strong (R2=0.79-1.00) as per Fowler et al. (1998).      74  3.3 Results 3.3.1 Abundance The mean crown volume index of the herb layer, averaged over the entire 15-year post-treatment period, appeared to be positively correlated with thinning intensity, with the 250 stems/ha stands having 1.65 times greater abundance than the unthinned stands (Table 3.1).  However, differences among densities were not statistically significant (Table 3.1).  Greater herb volume among all of the PCT stands compared with unthinned stands was a trend that remained consistent for the first 10 years; however, the lightly thinned stands (1000 and 2000 stems/ha) had the least herb volume of all densities 15 years post-thinning, which resulted in a marginally significant density effect (F4,4=1.33; P=0.08; Table 3.1).  A significant (F1,8=13.84; P<0.01) fertilizer effect was evident when averaged across all 15 years, with 2.67 times greater herb volume observed within fertilized than unfertilized stands (Table 3.1).  A significant time ? fertilizer interaction (F10,171=4.95; P<0.01) resulted from a lack of significant fertilizer effect during 1994 and 1995 (pre- and post-fertilization samples) followed by significantly enhanced herb volume resulting from fertilization from 1996 to 2003, and a loss of fertilizer effect in 2008 (Figure 3.1a; Table 3.1).  While there were yearly fluctuations, mean herb volume tended to increase within all treated stands for the first seven years post-PCT.  Herb volume peaked in 2000 among all stands and decreased thereafter (Figure 3.1a; annual data for density treatments not shown).   Mean crown volume index of shrubs was not significantly affected by density treatments (Table 3.1).  Shrub volume initially decreased within all thinned stands (data not shown); however, a lack of significant time ? density interaction suggests that this decrease was not significant (Table 3.1).  Following the initial decrease, shrub volume generally increased for the next nine years (until 2003), particularly within the fertilized stands.  While there was a lag of at least 5 years, fertilization treatments eventually resulted in a noticeably increased shrub volume compared to unfertilized stands (Figure 3.1b).  Shrub volume within fertilized stands was 1.64 times greater than in unfertilized  for the 1999 to 2003 period (Table 3.1); however, this increase was only marginally significant (F1,8=3.74; P=0.09).  During the last five years, shrub volume decreased in all stands and most dramatically within fertilized stands.  From 2003 to 2008, shrub    75  volume decreased by factors of 0.79 and 0.36 within unfertilized and fertilized stands, respectively (Figure 3.1b). Averaged over the 15-year post-PCT period (1994 to 2008), mean crown volume index of the tree layer was significantly (F4,8=13.92; P<0.01) affected by density treatments.  The unthinned stands had 3.24 times more tree crown volume index than the most heavily thinned stands (250 stems/ha) (Table 3.1).   Tree volume was inversely related to thinning intensity (Figure 3.2a).  Mean tree volume was statistically similar among the heavily to moderately thinned stands (250, 500, and 1000 stems/ha), but was significantly less (DMRT; P?0.05) than the most lightly thinned (2000 stems/ha) and unthinned stands (Table 3.1).  Fertilizer effect was significant (F1,8=7.76; P=0.02) with unfertilized stands having 1.03 times greater tree crown volume index than fertilized stands.         76  Table 3.1.  Mean (n = 3 replicate study areas) (SE) crown volume index (m3/0.01 ha) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The overall sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction was observed for this overall analysis, the 15-year post-PCT period was split into three five-year periods, and assessed separately.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  Plant layer  Treatment Means (SE)  Split-Split-Plot ANOVA Results  Densitya Fertilizerb Density Fertilizer Density ? Fertilizer Time ? Density Time ? Fertilizer Period 250 500 1000 2000 Unthinned UF F Herbs 94-08 30.68 (4.84) 26.86 (8.26) 22.64 (4.81) 23.99 (5.21) 18.60 (2.24) 14.49B (1.41) 38.62A (11.99) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 1.48 0.29 13.84 <0.01 0.26 0.85 0.32 1.00 4.95 <0.01  94-98 26.69 (4.70) 22.07 (4.07) 20.92 (3.51) 22.49 (2.71) 17.81 (1.74) 15.14B (2.28) 31.61A (7.29) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.60 0.67 11.10 0.01 0.36 0.78 0.25 1.00 3.93 <0.01  99-03 36.99 (6.94) 33.52 (14.61) 25.75 (8.56) 27.71 (10.18) 20.24 (4.30) 14.33B (1.52) 49.13A (19.64) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 1.33 0.34 16.13 <0.01 0.26 0.85 0.94 0.53 1.78 0.14  2008 14.07 (1.01) 11.69 (2.10) 9.51 (0.33) 6.39 (0.42) 13.34 (2.95) 9.47 (0.71) 12.32 (3.23) F4,4 P F1,4 P F3,4 P / / 4.97 0.08 1.10 0.35 0.68 0.61 Shrubs 94-08 21.69 (2.32) 18.27 (8.70) 14.29 (3.41) 17.49 (8.88) 18.73 (7.58) 15.22 (3.03) 21.53 (8.71) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.31 0.86 2.90 0.13 1.29 0.34 0.35 1.00 2.10 0.03  94-98 15.63 (2.21) 12.14 (5.49) 10.48 (2.22) 10.70 (4.14) 15.39 (5.55) 12.03 (2.40) 13.28 (3.69) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.45 0.77 1.15 0.32 2.31 0.15 0.31 0.99 0.60 0.67  99-03 27.48 (3.35) 23.45 (12.38) 17.79 (5.12) 24.58 (13.55) 22.53 (9.74) 18.07 (3.96) 29.69 (13.86) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.27 0.89 3.74 0.09 0.88 0.49 0.44 0.97 3.39 0.01  2008 21.48 (1.23) 16.97 (8.66) 13.54 (2.95) 6.80 (3.52) 11.65 (8.14) 14.83 (0.76) 13.77 (5.08) F4,4 P F1,4 P F3,4 F4,4 / / 0.59 0.69 0.14 0.73 0.48 0.71 Trees 94-08 52.61b (2.13) 69.20b (15.52) 88.97b (4.72) 138.95a (4.75) 170.31a (24.10) 97.94A (3.05) 95.02B (2.35) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 13.92 <0.01 7.76 0.02 1.39 0.31 1.39 0.08 1.44 0.17 Total 94-08 104.98b (5.06) 114.33b (26.93) 125.90b (5.57) 180.43a (8.28) 207.64a (15.65) 127.65B (6.38) 155.17A (14.34) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 10.28 <0.01 31.91 <0.01 0.55 0.66 0.85 0.72 4.10 <0.01  94-98 78.81c (4.97) 88.76c (25.37) 100.17bc (3.64) 132.05ab (8.45) 176.90a (18.43) 105.09B (9.77) 112.76A (7.43) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 7.94 <0.01 21.96 <0.01 1.57 0.27 0.20 1.00 1.37 0.25  99-03 123.68b (9.80) 132.13b (35.85) 142.87b (18.16) 223.00a (17.93) 232.17a (13.83) 141.60B (9.32) 191.88A (27.64) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 10.76 <0.01 37.19 <0.01 0.32 0.81 1.93 0.03 2.37 0.06  2008 157.48c (9.89) 147.54c (7.39) 189.64b (3.48) 219.30b (3.74) 264.10a (2.04) 186.11 (3.99) 190.37 (4.03) F4,4 P F1,4 P F3,4 F4,4 / / 36.99 <0.01 2.36 0.20 0.13 0.94 a  Density treatment: target PCT density = 250, 500, 1000, and 2000 stems/ha and unthinned  b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized       77  Figure 3.1.  Mean (n = 3) crown volume index (m3/0.01 ha) of a) herbs and b) shrubs by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.      0102030405060701993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008m3/0.01 ha unfertilizedfertilizedCrown volume index a) herbs 0510152025303540451993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008m3/0.01 ha Year (PCT)                         (Fert.) b) shrubs    78  Figure 3.2.  Mean (n = 3) crown volume index (m3/0.01 ha) of the tree layer by a) density and b) fertilizer treatments from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.     0501001502002503001993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008m3/0.01 ha 250 5001000 2000unthinnedTree layer crown volume index a) density effect 0204060801001201401601801993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008m3/0.01 ha Year unfertilized fertilized(PCT)                         (Fert.) b) fertilizer effect    79  Averaged over the 15-year post-PCT period, mean crown volume index of the total plant community was significantly (F4,8=10.28; P<0.01) affected by density treatments and showed the same inverse relationship with thinning intensity as observed for the tree layer; significantly (DMRT; P?0.05) greater volume among the unthinned and lightly thinned (2000 stems/ha) stands compared to the more heavily thinned stands (Table 3.1).  Total volume was significantly enhanced by fertilizer during the 1994 to 1998 (F1,8=21.96; P<0.01) and 1999 to 2003 periods (F1,8=37.19; P<0.01); however, this fertilizer effect was not significant during the final sample (Table 3.1).  3.3.2 Species diversity  Averaged over the 15-year post-PCT period, mean species richness of the herb, shrub, tree, and combined total layers was unaffected by density treatments (Table 3.2).  Although there was a trend for fertilization to decrease the species richness of all plant layers, no significant differences were observed.  The most significant fertilization effect was observed for shrub species richness during the 1999 to 2003 period (13.03 ? 1.48 and 11.33 ? 0.67 species within unfertilized and fertilized stands, respectively); however this difference was only marginally significant (F1,8=4.04; P=0.08) (Table 3.2).  The overall mean species diversity (Simpson index) measurements of the herb, shrub, tree, and combined total layers were not significantly affected by density treatments (Table 3.3).  The density effect approached significance for total species diversity and was 67 ? 20% greater within the 250 than 2000 stems/ha stands (Simpson index = 0.69 ? 0.05 vs. 0.41 ? 0.08).  However, this difference was only marginally significant (F4,8=3.56; P=0.06). Fertilization treatments significantly decreased species diversity for the herb and shrub layers (Figure 3.3).  In contrast, fertilization appeared to increase species diversity of the tree layer, and even more so for the combined total layer (Figure 3.4).  However, this enhanced species diversity was not significant (Table 3.3).       80  Table 3.2.  Mean (n = 3 replicate study areas) (SE) species richness (number of species) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Plant layer  Treatment Means (SE)  Split-Split-Plot ANOVA Results  Density Fertilizer Density Fertilizer Density ? Fertilizer Time ?  Density Time ?  Fertilizer Period 250 500 1000 2000 Unthinned UF F Herbs 94-08 16.13 (3.74) 15.17 (5.1) 16.16 (4.23) 13.78 (3.92) 17.48 (3.8) 16.06 (4.70) 14.92 (3.22) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.72 0.60 0.00 0.98 0.15 0.93 0.49 1.00 1.44 0.17 Shrubs 94-08 13.28 (0.69) 12.16 (1.69) 12.5 (0.52) 10.87 (1.41) 12.21 (1.60) 12.59 (1.31) 11.71 (0.77) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 1.60 0.26 1.31 0.29 0.73 0.56 1.05 0.40 3.45 <0.01  94-98 12.93 (0.85) 12.17 (1.64) 12.63 (0.44) 10.97 (1.34) 11.87 (1.52) 12.21 (1.11) 12.05 (0.89) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 1.23 0.37 0.05 0.82 0.78 0.54 1.01 0.46 0.61 0.65  99-03 13.57 (0.64) 12.17 (1.77) 12.47 (0.57) 10.7 (1.53) 12.67 (1.68) 13.03 (1.48) 11.33 (0.67) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 1.85 0.21 4.04 0.08 0.66 0.60 1.03 0.44 1.19 0.32  2008 13.00 (1.22) 10.75 (1.02) 11.25 (0.20) 10.50 (1.22) 10.00 (0.82) 11.00 (0.98) 11.50 (0.82) F4,4 P F1,4 P F3,4 F4,4 / / 3.56 0.12 0.13 0.74 0.48 0.71 Trees 94-08 4.89 (0.46) 4.26 (0.57) 4.83 (1.3) 4.62 (0.58) 4.51 (0.46) 4.82 (0.65) 4.41 (0.67) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.37 0.83 2.04 0.19 0.85 0.51 1.05 0.40 1.43 0.17 Total 94-08 34.30 (4.63) 31.60 (6.60) 33.50 (3.89) 29.26 (4.87) 34.19 (5.63) 33.47 (5.89) 31.04 (3.63) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 1.31 0.35 0.51 0.49 0.21 0.89 0.52 0.99 1.35 0.21 a  Density treatment: target PCT density = 250, 500, 1000, and 2000 stems/ha and unthinned  b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized      81  Table 3.3.  Mean (n = 3 replicate study areas) (SE) species diversity (Simpson index) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. Plant layer  Treatment Means (SE)  Split-Split-Plot ANOVA Results  Densitya Fertilizerb Density Fertilizer Density ? Fertilizer Time ?  Density Time ?  Fertilizer Period 250 500 1000 2000 Unthinned UF F Herbs 94-08 0.50 (0.05) 0.53 (0.04) 0.56 (0.08) 0.54 (0.02) 0.51 (0.09) 0.59A (0.03) 0.45B (0.08) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.93 0.49 12.29 <0.01 0.30 0.82 0.38 1.00 4.66 <0.01  94-98 0.51 (0.04) 0.54 (0.04) 0.58 (0.07) 0.54 (0.04) 0.51 (0.08) 0.57A (0.04) 0.49B (0.06) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.97 0.48 9.01 0.02 0.69 0.58 0.59 0.89 2.91 0.03  99-03 0.49 (0.06) 0.53 (0.05) 0.54 (0.09) 0.53 (0.04) 0.51 (0.11) 0.61A (0.03) 0.41B (0.11) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.70 0.61 12.49 <0.01 0.11 0.95 0.51 0.93 1.13 0.35  2008 0.52 (0.02) 0.50 (0.02) 0.57 (0.03) 0.52 (0.07) 0.41 (0.03) 0.51 (0.00) 0.52 (0.05) F4,4 P F1,4 P F3,4 F4,4 / / 1.38 0.38 0.07 0.81 0.08 0.97 Shrubs 94-08 0.62 (0.02) 0.60 (0.03) 0.64 (0.04) 0.57 (0.03) 0.62 (0.08) 0.65A (0.01) 0.56B (0.04) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.64 0.65 8.86 0.02 2.42 0.14 0.56 0.98 2.95 <0.01  94-98 0.68 (0.01) 0.65 (0.05) 0.67 (0.04) 0.62 (0.03) 0.63 (0.09) 0.67A (0.03) 0.63B (0.03) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.61 0.67 7.90 0.02 4.81 0.03 0.38 0.98 1.96 0.11  99-03 0.57 (0.02) 0.57 (0.03) 0.61 (0.04) 0.52 (0.04) 0.63 (0.06) 0.63A (0.02) 0.50B (0.05) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.76 0.58 7.12 0.03 1.17 0.38 0.97 0.50 0.90 0.47  2008 0.51 (0.01) 0.49 (0.00) 0.55 (0.03) 0.51 (0.05) 0.60 (0.09) 0.56 (0.01) 0.48 (0.03) F4,4 P F1,4 P F3,4 F4,4 / / 0.29 0.87 2.95 0.16 4.40 0.09 Trees 94-08 0.31 (0.05) 0.16 (0.02) 0.23 (0.10) 0.09 (0.03) 0.22 (0.08) 0.19 (0.04) 0.21 (0.03) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 2.04 0.18 0.29 0.61 0.44 0.73 1.14 0.28 0.83 0.60 Total 94-08 0.69 (0.05) 0.60 (0.04) 0.57 (0.08) 0.41 (0.08) 0.46 (0.12) 0.53 (0.04) 0.59 (0.07) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 3.56 0.06 2.36 0.16 0.26 0.85 0.59 0.98 0.79 0.64 a  Density treatment: target PCT density = 250, 500, 1000, and 2000 stems/ha and unthinned  b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized      82  Figure 3.3.  Mean (n = 3) species diversity (Simpson index) of the a) herb and b) shrub layers from 1993 to 2008 by fertilizer treatment.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.      00.10.20.30.40.50.60.70.81993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D unfertilizedfertilizedSpecies diversity (Simpson index) a) herbs 00.10.20.30.40.50.60.70.81993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D Year (PCT)                         (Fert.) b) shrubs    83   Figure 3.4.  Mean (n = 3) species diversity (Simpson index) of the a) tree and b) total layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.     00.050.10.150.20.250.31993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D unfertilizedfertilizedSpecies diversity (Simpson index) a) trees 00.10.20.30.40.50.60.71993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D Year (PCT)                         (Fert.) b) total (herbs, shrubs, and trees combined)    84  For herb species diversity, significant time ? fertilizer interactions (Table 3.3) revealed that a significant fertilizer effect was not observed until 1996, whereas from 1996 to 2003, fertilizer treatments significantly (F1,8=13.42; P=0.01) decreased herb species diversity.  A fertilizer effect was not evident in 2008 (Figure 3.3a; Table 3.3).   For shrub species diversity, a significant time ? fertilizer interaction (Table 3.3) revealed that differences between unfertilized and fertilized stands were not consistently significant until 1999 although fertilizer treatments tended to decrease diversity.  From 1999 to 2003, shrub diversity was significantly (F1,8=7.12; P=0.03) decreased by fertilizer treatments.  A fertilizer effect was not evident in 2008 (Figure 3.3b).  The significant density ? fertilization interaction observed for the 1994 to 1998 period (Table 3.3) resulted from the fertilized stands having greater shrub diversity than unfertilized stands within the 250 and 2000 stems/ha stands, but less within the 500 and 1000 stems/ha stands.  Because differences were not significant and this interaction occurred prior to fertilization (i.e., cannot be an interaction between density and fertilization treatments), it is not considered important to the interpretation of the results. 3.3.3 Structural diversity  Averaged over the 15-year post-PCT period, mean structural richness of the herb and shrub layers were unaffected by density treatments (Table 3.4).  A significant time ? density interaction for structural richness of the tree layer revealed that unthinned and lightly thinned stands (2000 stems/ha) tended to decrease in structural richness with time, whereas the more heavily thinned stands (250 and 1000 stems/ha stands, in particular), after an initial decrease, were relatively high in structural richness throughout the study (Figure 3.5).  These differences resulted in significant density effects during 2000 (F4,8=4.46; P=0.04) and 2003 (F4,8=5.57; P=0.02) when structural richness was significantly greater (DMRT; P?0.05) within the 250 stems/ha stands than the 2000 stems/ha and unthinned stands.      85  Figure 3.5.  Mean (n = 3) structural richness (number of vegetation layers) of the tree layer by density treatment, from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.    A significant time ? density interaction for total structural richness revealed a lack of density effect during the 1994 to 1998 period and an overall significant density effect (F4,8=16.47; P<0.01) for the 1999 to 2003 period.  While a mean of less than one-half of a structural class separated the richest (unthinned) from least rich (500 stems/ha) stands (6.00 ? 0.00 vs. 5.69 ? 0.04 vegetation layers, respectively), this difference was statistically different (DMRT; P?0.05) (Table 3.4).  There was no significant density effect for total structural richness in 2008.  Averaged over the 15-year post-PCT period, mean structural richness of the herb and shrub layers were enhanced by fertilization treatments (Figure 3.6a and b); however, only the herb layer significantly so (F1,8=22.85; P<0.01) (Table 3.4).  Fertilizer had the opposite effect on tree layer structural richness, as unfertilized stands tended to have greater structural richness than fertilized stands.  Moreover, tree layer structural richness appeared to decline over time within fertilized stands but remained relatively constant within unfertilized stands (Figure 3.6c).  As a 34561993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008No. of vegetation layers Year 250 5001000 2000unthinnedStructural richness of tree layer (PCT)                         (Fert.)    86  result, the significance of the fertilizer effect increased with time (Table 3.4).  Total richness was unaffected by fertilizer treatments.    87  Table 3.4.  Mean (n = 3 replicate study areas) (SE) structural richness (number of vegetation layers) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilizer effects, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. Plant layer  Treatment Means (SE)  Split-Split-Plot ANOVA Results  Densitya Fertilizerb Density Fertilizer Density ? Fertilizer Time ?  Density Time ?  Fertilizer Period 250 500 1000 2000 Unthinned UF F Herbs 94-08 3.43 (0.11) 3.39 (0.24) 3.44 (0.22) 3.38 (0.24) 3.35 (0.18) 3.24B (0.16) 3.61A (0.24) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.41 0.80 22.85 <0.01 1.24 0.36 0.89 0.66 1.27 0.25 Shrubs 94-08 3.85 (0.11) 3.45 (0.72) 3.90 (0.30) 3.40 (0.77) 3.83 (0.53) 3.59 (0.42) 3.77 (0.52) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.63 0.66 2.76 0.14 0.80 0.53 1.01 0.46 0.46 0.91 Trees 94-08 5.09 (0.06) 4.70 (0.35) 5.09 (0.17) 4.50 (0.44) 4.74 (0.45) 5.21A (0.16) 4.36B (0.33) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 1.85 0.21 11.03 0.01 1.23 0.36 3.72 <0.01 11.28 <0.01  94-98 4.96 (0.09) 4.73 (0.28) 5.24 (0.21) 4.91 (0.51) 5.31 (0.41) 5.19A (0.17) 4.77B (0.35) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.69 0.62 5.12 0.05 0.74 0.56 0.79 0.70 0.31 0.87  99-03 5.22 (0.08) 4.71 (0.43) 5.03 (0.15) 4.23 (0.45) 4.29 (0.50) 5.26A (0.15) 4.09B (0.38) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 3.47 0.06 17.29 <0.01 1.50 0.29 0.91 0.57 2.10 0.09  2008 5.08 (0.20) 4.75 (0.34) 4.42 (0.20) 3.83 (0.00) 4.17 (0.95) 5.13A (0.16) 3.67B (0.07) F4,4 P F1,4 P F3,4 F4,4 / / 1.78 0.30 184.45 <0.01 7.33 0.04 Total 94-08 5.60cd (0.04) 5.49d (0.04) 5.77b (0.07) 5.66bc (0.03) 5.97a (0.03) 5.66 (0.04) 5.67 (0.05) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 11.42 <0.01 1.51 0.25 0.73 0.56 2.93 <0.01 2.85 <0.01  94-98 5.27 (0.05) 5.23 (0.05) 5.62 (0.14) 5.61 (0.07) 5.93 (0.07) 5.46 (0.05) 5.52 (0.09) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.93 0.49 0.70 0.43 0.97 0.45 0.71 0.77 0.33 0.86  99-03 5.88b (0.04) 5.69c (0.04) 5.91ab (0.06) 5.71c (0.06) 6.00a (0.00) 5.82 (0.06) 5.82 (0.02) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 16.47 <0.01 0.63 0.45 0.57 0.65 0.60 0.87 1.07 0.38  2008 5.92 (0.07) 5.92 (0.07) 5.92 (0.07) 5.67 (0.14) 6.00 (0.00) 5.97 (0.03) 5.75 (0.07) F4,4 P F1,4 P F3,4 F4,4 / / 1.37 0.38 3.57 0.13 2.43 0.21 a  Density treatment: target PCT density = 250, 500, 1000, and 2000 stems/ha and unthinned  b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized    88  Figure 3.6.  Mean (n = 3) structural richness (number of vegetation layers) of the a) herb, b) shrub, and c) tree layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.       23451993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008No. of vegetation layers unfertilizedfertilizedStructural richness a) herbs 23451993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008No. of vegetation layers b) shrubs 34561993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008No. of vegetation layers Year (PCT)                         (Fert.) c) trees    89   Averaged over the 15-year post-PCT period, mean structural diversity of the herb and shrub layers were unaffected by density treatments (Table 3.5).  A significant density effect for tree layer structural diversity (F4,8=6.61; P=0.01) indicated that the heavily thinned stands (250 stems/ha) had significantly greater (DMRT; P?0.05) structural diversity than the lightly thinned (2000 stems/ha) and unthinned stands (Table 3.5).  A significant time ? density interaction (F40,171=2.64; P<0.01) revealed that tree layer structural diversity among all densities generally decreased with time, and that this rate of decrease was more rapid among the unthinned and lightly thinned (2000 stems/ha) stands (Figure 3.7a).  For the 1994 to 2003 period, the mean structural diversity of the tree layer remained significantly (F4,8=6.32; P=0.01) greater within the most heavily thinned stands (250 stems/ha) compared to the unthinned and lightly thinned (2000 stems/ha) stands, and statistically similar to that of the 500 and 1000 stems/ha stands (DMRT; P?0.05) (Figure 3.7a).  During 2008, the difference between the greatest and least tree structural diversity (observed within the 500 and 2000 stems/ha stands, respectively) was only marginally significant (Table 3.5).  A significant overall time ? density interaction (F40,171=1.54; P=0.03) observed for total structural diversity revealed that diversity remained relatively constant over time within the unthinned stands and that PCT appeared to initially decrease structural diversity, particularly within the heavily thinned stands (Figure 3.7b).  This resulted in a significant density effect (F4,8=4.33; P=0.04) for the 1994 to 1998 period, when mean total structural diversity was significantly less (DMRT; P?0.05) within the 500 stems/ha stands compared to the unthinned stands (Table 3.5).  From 1999 to 2003, heavily thinned stands (250 and 500 stems/ha) increased in total structural diversity (Figure 3.7b).  This trend continued during 2008; however; differences were not significant (Table 3.5).      90   Table 3.5.  Mean (n = 3 replicate study areas) (SE) structural diversity (Simpson index) of herbs, shrubs, trees, and all layers combined (total) among five levels of density and two levels of fertilization treatments.  Separate sets of means, and associated ANOVA results, are provided for different sampling periods.  The initial sampling period (94-08, entire 15-year post-treatment period) includes 11 sample years (1994-2003 and 2008).  If a significant time ? treatment interaction is observed for this overall analysis, the 15-year post-treatment period is analyzed among three five-year periods.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts. Plant layer  Treatment Means (SE)  Split-Split-Plot ANOVA Results  Densitya Fertilizerb Density Fertilizer Density ? Fertilizer Time ?  Density Time ? Fertilizer Period 250 500 1000 2000 Unthinned UF F Herbs 94-08 0.43 (0.06) 0.39 (0.05) 0.39 (0.08) 0.40 (0.07) 0.38 (0.07) 0.34B (0.06) 0.47A (0.07) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 1.64 0.26 15.64 <0.01 0.38 0.77 0.38 1.00 1.08 0.38 Shrubs 94-08 0.48 (0.06) 0.42 (0.16) 0.48 (0.10) 0.39 (0.15) 0.47 (0.12) 0.42B (0.10) 0.48A (0.13) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 0.70 0.62 13.63 <0.01 1.52 0.28 0.70 0.91 1.68 0.09 Trees 94-08 0.50a (0.04) 0.38ab (0.12) 0.33abc (0.01) 0.19c (0.04) 0.22bc (0.05) 0.38A (0.05) 0.28B (0.04) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 6.61 0.01 22.78 <0.01 0.54 0.67 2.64 <0.01 3.92 <0.01  94-98 0.54a (0.03) 0.42ab (0.12) 0.37ab (0.02) 0.27b (0.06) 0.32b (0.08) 0.42A (0.06) 0.36B (0.05) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 3.86 0.05 10.28 0.01 0.80 0.53 1.69 0.07 1.42 0.24  99-03 0.48a (0.05) 0.37a (0.13) 0.32ab (0.02) 0.13c (0.02) 0.13bc (0.04) 0.36A (0.05) 0.24B (0.05) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 8.31 <0.01 31.37 <0.01 0.73 0.56 1.33 0.21 0.29 0.88  2008 0.25 (0.05) 0.32 (0.09) 0.12 (0.01) 0.08 (0.02) 0.10 (0.04) 0.26A (0.04) 0.09B (0.01) F4,4 P F1,4 P F3,4 F4,4 / / 5.22 0.07 8.89 0.04 0.47 0.72 Total 94-08 0.62 (0.02) 0.59 (0.03) 0.61 (0.04) 0.63 (0.03) 0.64 (0.05) 0.59B (0.03) 0.65A (0.03) F4,8 P F1,8 P F3,8 P F40,171 P F10,171 P 2.10 0.17 19.83 <0.01 1.84 0.22 1.54 0.03 1.36 0.20  94-98 0.56ab (0.04) 0.52b (0.04) 0.57ab (0.06) 0.61ab (0.04) 0.64a (0.06) 0.55B (0.04) 0.60A (0.05) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 4.33 0.04 15.88 <0.01 1.24 0.36 0.35 0.99 0.81 0.52  99-03 0.66 (0.01) 0.66 (0.03) 0.66 (0.04) 0.67 (0.03) 0.65 (0.05) 0.63B (0.03) 0.70A (0.03) F4,8 P F1,8 P F3,8 P F16,72 P F4,72 P 0.32 0.86 25.64 <0.01 1.73 0.24 2.34 <0.01 1.26 0.29  2008 0.70 (0.00) 0.67 (0.02) 0.61 (0.01) 0.54 (0.04) 0.60 (0.09) 0.64 (0.02) 0.62 (0.03) F4,4 P F1,4 P F3,4 F4,4 / / 3.91 0.11 1.74 0.26 1.94 0.27 a  Density treatment: target PCT density = 250, 500, 1000, and 2000 stems/ha and unthinned  b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized      91  Figure 3.7.  Mean (n = 3) structural diversity (Simpson index) of the a) tree and b) total (herbs, shrubs, and trees combined) layers by density treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.     00.10.20.30.40.50.60.71993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D 250 5001000 2000unthinnedStructural diversity (Simpson index) a) tree layer 0.40.50.60.70.81993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D Year (PCT)                         (Fert.) b) total (herb, shrub, and tree layers combined)    92  Averaged over the 15-year post-PCT period, mean structural diversity of the herb, shrub, tree, and total layers were significantly affected by fertilization treatments (Figure 3.8; Table 3.5).  Fertilization enhanced the overall structural diversity of the herb (F1,8=15.64; P<0.01) and shrub layers (F1,8=13.63; P<0.01); however, it had the opposite effect on the tree layer (F1,8=22.78; P<0.01) (Figure 3.8a, b, and c).  A significant time ? fertilization interaction (F10,171=3.92; P<0.01) was observed for the overall tree layer structural diversity and revealed that diversity appeared to decline over time among all stands; however, the rate of decrease was greater among fertilized stands (Figure 3.8c).  The overall mean structural diversity of the total layer was significantly (F1,8=19.83; P<0.01) enhanced by fertilizer treatments (Figure 3.8d).       93  Figure 3.8.  Mean (n = 3) structural diversity (Simpson index) of the a) herb, b) shrub, and c) tree, and total (herbs, shrubs, and trees combined) layers by fertilizer treatment from 1993 to 2008.  Arrows along x-axis indicate timing of treatments; PCT = pre-commercial thinning, Fert. = fertilization.                 0.20.30.40.50.61993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D unfertilizedfertilizedStructural diversity (Simpson index) a) herb layer 0.30.40.50.61993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D b) shrub layer 00.20.40.61993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D c) tree layer 0.50.60.70.81993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2008D Year (PCT)                         (Fert.) d) total (herb, shrub, and tree layers combined)    94  3.3.4 Productivity-diversity relationship 3.3.4.1 Species richness Using pooled data from all nine treated stands, three study areas, and 12 sample years (n=324), mean total species richness regressed on an index of productivity (mean total crown volume index) resulted in a marginally significant positive linear (F1,322=3.06; P=0.08) and significant humped-shaped quadratic (F2,321=17.67; P<0.01) relationship.  The strength of the linear relationship was very weak (R2=0.01) and that of the quadratic relationship was weak (R2=0.10).  Conclusions based on these pooled data may be questioned due to the significantly different regression models observed among study areas (i.e., relationships with different intercepts and/or slopes; data not shown).  Among years, linear and quadratic models were similar among the two southern study areas (Summerland and Kelowna) and different from those for the Cariboo study area.   As a result, separate regression analyses were conducted using pooled data from the Summerland and Kelowna study areas and the Cariboo data on its own.  Linear and quadratic models were statistically similar among the initial six (1993-1998) and the final six sample years (1999-2003 and 2008) for both of these datasets.  As a result, data were pooled within these two time periods.   Data from the initial six sample years indicated regression models for species richness vs. crown volume had weak and non-significant relationships (Table 3.6; Figure 3.9).  Data from the final six sample years also indicated very weak and non-significant relationship for the pooled Summerland - Kelowna dataset (Table 3.6); however, significant negative linear (R2=0.22; F1,52=14.76; P<0.01) and humped-shaped quadratic (R2=0.28; F2,51=10.13; P<0.01) relationships were observed for the Cariboo data (Table 3.6; Figure 3.9b).     95  Table 3.6.  Summary of linear and quadratic models for mean total species richness regressed on mean total crown volume index for two different time periods and study areas.  Pattern of relationships are described as N (negative linear), H (hump-shaped quadratic), or NS (not significant).    Linear  Quadratic Time period Data source (study areas) n R2 F P Pattern  R2 F P Pattern Initial 6 sample years; 1993-1998 Summerland and Kelowna 108 0.01 F1, 106=0.56 0.46 NS  0.01 F2, 105=0.67 0.52 NS  Cariboo 54 0.01 F1, 52=0.32 0.57 NS  0.02 F2, 51=0.40 0.67 NS Final 6 sample years; 1999-2003 & 2008 Summerland and Kelowna 108 0.01 F1, 106=0.64 0.43 NS  0.01 F2, 105=0.75 0.48 NS  Cariboo 54 0.22 F1, 52=14.76 <0.01 N  0.28 F2, 51=10.13 <0.01 H       96  152025303540450 100 200 300 400152025303540450 50 100 150 200 250 300species richness R? = 0.22 R? = 0.28 253035404550556050 150 250 350species richness crown volume index 25303540455055600 50 100 150 200 250species richness crown volume index Figure 3.9.  Relationship (or lack thereof) between species richness and an index of stand productivity (total crown volume index; m3/0.01 ha) among managed stands of young lodgepole pine.  Relationships were assessed for two separate datasets (a) Summerland and Kelowna study areas combined and b) Cariboo study area) and two sample periods (initial six, and final six sample years).  Data points represent mean values for a given year from each of nine treatment stands.  Trendlines and R2 values are presented for significant regression models.  a) Summerland and Kelowna study areas combined (n = 108) 1993-1998      1999-2003 and 2008          b) Cariboo study area (n = 54) 1993-1998      1999-2003 and 2008        97  3.3.4.2 Species diversity  Regressing mean total species diversity on mean total crown volume index resulted in highly significant negative linear (F1,322=18.47; P<0.01) and U-shaped quadratic (F2,321=9.41; P<0.01) models with weak relationships (R2=0.05 for both models) when using the total pooled dataset.  Although the form of linear and quadratic relationships were consistent among all three study areas (negative and U-shaped), there were some significant differences among regression models from one study area to the next that suggested that interpretations of the regression models resulting from the total pooled data should be made with caution.  Within sample years, regression models resulting from the Summerland and Kelowna data were found to be statistically similar and, therefore, were pooled.  Regression models from the Cariboo data were significantly different from both Summerland and Kelowna and, therefore, were analyzed separately.  Pooling data among sample years was deemed appropriate after confirming statistically similar regression models among the initial six and final six sample years. Data from the initial six sample years indicated that regression models for species diversity vs. crown volume had a significant negative linear relationship with modest strength for both the pooled Summerland - Kelowna (R2=0.16; F1,106=20.49; P<0.01) and Cariboo data (R2=0.35; F1,52=28.36; P<0.01) (Figure 3.10).  The form of the quadratic relationships differed among study areas as the pooled Summerland - Kelowna data indicated a significant U-shaped relationship of modest strength (R2=0.18; F2,105=11.36; P<0.01), whereas the Cariboo data indicated a significant hump-shaped relationship of modest strength (R2=0.35; F2,51=13.98; P<0.01).  For the final six sample years, a significant negative linear relationship of weak strength (R2=0.08; F1,106=8.83; P<0.01) and a significant U-shaped relationship of weak strength (R2=0.09; F2,105=5.01; P=0.01) was observed for the pooled Summerland - Kelowna dataset (Figure 3.10a).  For the Cariboo study area, data from the last six sample years indicated no significant relationships (Table 3.7; Figure 3.10b).     98    Table 3.7. Summary of linear and quadratic models for mean total species diversity (Simpson index) regressed on mean total crown volume index for two different time periods and study areas.  Pattern of relationships are described as N (negative linear), H (hump-shaped quadratic), U (U-shaped), or NS (not significant).    Linear  Quadratic Time period Data source (study areas) N R2 F P Pattern  R2 F P Pattern Initial 6 sample years; 1993-1998 Summerland and Kelowna 108 0.16 F1, 106=20.49 <0.01 N  0.18 F2, 105=11.36 <0.01 U  Cariboo 54 0.35 F1, 52=28.36 <0.01 N  0.35 F2, 51=13.98 <0.01 H Final 6 sample years; 1999-2003 & 2008 Summerland and Kelowna 108 0.08 F1, 106=8.83 <0.01 N  0.09 F2, 105=5.01 0.01 U  Cariboo 54 0.02 F1, 52=1.08 0.30 NS  0.02 F2, 51=0.56 0.58 NS      99  R? = 0.08 R? = 0.09 0.00.10.20.30.40.50.60.70.80.90 100 200 300 400crown volume index R? = 0.16 R? = 0.18 0.00.10.20.30.40.50.60.70.80.90 50 100 150 200 250 300species diversity crown volume index 0.30.40.50.60.70.80.950 100 150 200 250 300 350crown volume index R? = 0.35 R? = 0.35 0.30.40.50.60.70.80.925 75 125 175species diversity crown volume index Figure 3.10.  Relationship (or lack thereof) between species diversity (Simpson index) and an index of stand productivity (total crown volume index; m3/0.01 ha) among managed stands of young lodgepole pine.  Relationships were assessed for two separate datasets (a) Summerland and Kelowna study areas combined and b) Cariboo study area) and two sample periods (initial six, and final six sample years).  Data points represent mean values for a given year from each of nine treatment stands.  Trendlines and R2-values are presented for significant regression models.  a) Summerland and Kelowna study areas (n = 108) 1993-1998      1999-2003 and 2008           b) Cariboo study area (n = 54) 1993-1998      1999-2003 and 2008              100  3.3.4.3 Structural diversity   Regressing mean total structural diversity on mean total crown volume index resulted in a highly significant positive linear relationship of modest strength (R2=0.15; F1,322=56.42; P<0.01) and a hump-shaped quadratic relationship of modest strength (R2=0.17; F2,321=33.68; P<0.01) when using the total pooled dataset (i.e., data from all three replicates and all 12 sample years combined).  The form of linear and quadratic relationships (positive and hump-shaped, respectively) were consistent among all three study areas.  However, the regression models differed significantly among the study areas and among most of the 12 sample years.  Because the appropriateness of pooling data across areas and years may be questioned, structural diversity-crown volume relationships were also assessed separately by study area and sample years.  The linear relationship between structural diversity and crown volume index was similar among all study areas in that all indicated a positive relationship, at least initially.  The positive relationship weakened after seven and eight years at Summerland and Kelowna, respectively, and eventually turned into a negative relationship (Figure 3.11).  This negative relationship was significant (R2=0.50; F1,7=7.34; P=0.03) only for Summerland during the final sample year (Table 3.8; Figure 3.11).  At the Cariboo study area, the positive relationship between structural diversity and productivity was maintained throughout all 12 sample years, although significance varied from year to year (Table 3.8).  The R2-values indicate that strength of quadratic relationships between total structural diversity and total crown volume index were consistently greater than, or equal to, that for linear relationships (Table 3.9).  The form of these relationships was nearly always hump-shaped.  The hump-shaped relationships suggest that structural diversity decreases at the highest levels of crown volume index.  However, the peak of structural diversity changed with time within both the Summerland (Figure 3.11) and Kelowna study areas.  For these areas, structural diversity was initially greatest at the upper range of crown volume index; however, with time, this peak appeared to gradually move to the lower range of volume (Figure 3.11).  The peak of structural diversity remained within the upper range of the volume range for the Cariboo study area throughout this study.      101   Table 3.8.  Summary of linear models for mean total structural diversity (Simpson index) regressed on mean total crown volume index by Summerland, Kelowna, and Cariboo study areas and sample years.  Patterns of relationships (significant models indicated in bold text) are described as P (positive linear), N (negative linear), or NS (not significant).  Summerland  Kelowna  Cariboo Year R2 F1,7 P Pattern  R2 F1,7 P Pattern  R2 F1,7 P Pattern 1993 0.35 3.79 0.10 P  0.11 0.89 0.38 NS  0.28 2.69 0.15 NS 1994 0.76 23.11 <0.01 P  0.19 1.66 0.24 NS  0.46 5.88 0.05 P 1995 0.93 99.81 <0.01 P  0.46 5.93 0.05 P  0.34 3.63 0.10 P 1996 0.78 21.60 <0.01 P  0.71 17.03 <0.01 P  0.57 9.08 0.02 P 1997 0.37 4.08 0.08 P  0.70 16.28 0.01 P  0.09 0.70 0.43 NS 1998 0.37 7.53 0.03 P  0.61 10.75 0.01 P  <0.01 0.02 0.89 NS 1999 0.52 7.50 0.03 P  0.46 6.05 0.04 P  0.30 3.04 0.13 NS 2000 0.12 0.98 0.36 NS  0.44 5.47 0.05 P  0.63 11.89 0.01 P 2001 0.03 0.23 0.65 NS  0.27 2.58 0.15 NS  0.79 26.98 <0.01 P 2002 0.42 5.01 0.06 N  <0.01 0.03 0.88 NS  0.51 7.23 0.03 P 2003 0.19 1.67 0.24 NS  0.12 0.95 0.36 NS  0.73 19.20 <0.01 P 2008 0.50 7.34 0.03 N  0.13 1.04 0.34 NS  0.29 2.81 0.14 NS  Table 3.9. Summary of quadratic models for mean total structural diversity (Simpson index) regressed on mean total crown volume index by Summerland, Kelowna, and Cariboo study areas and sample years.  Patterns of relationships (significant models indicated in bold text) are described as H (hump-shaped), U (U-shaped), or NS (not significant).  Summerland  Kelowna  Cariboo Year R2 F2,6 P Pattern  R2 F2,6 P Pattern  R2 F2,6 P Pattern 1993 0.54 3.49 0.10 H  0.16 0.58 0.59 NS  0.28 1.19 0.37 NS 1994 0.84 14.99 0.01 H  0.20 0.77 0.50 NS  0.61 4.76 0.06 U 1995 0.95 51.78 <0.01 H  0.67 6.13 0.04 H  0.66 5.92 0.04 H 1996 0.79 9.71 0.01 H  0.84 16.25 <0.01 H  0.76 9.68 0.01 H 1997 0.41 2.08 0.21 NS  0.73 8.17 0.02 H  0.50 2.97 0.13 NS 1998 0.68 6.79 0.03 H  0.71 7.41 0.02 H  0.01 0.02 0.98 NS 1999 0.93 40.96 <0.01 H  0.71 7.28 0.03 H  0.51 3.11 0.12 NS 2000 0.69 6.81 0.03 H  0.80 12.31 0.01 H  0.70 7.04 0.03 H 2001 0.28 1.15 0.38 NS  0.68 6.40 0.03 H  0.81 12.65 0.01 U 2002 0.59 4.24 0.07 H  0.87 19.33 <0.01 H  0.52 3.24 0.11 NS 2003 0.42 2.19 0.19 NS  0.39 1.88 0.23 NS  0.74 8.37 0.02 U 2008 0.57 4.00 0.08 H  0.50 2.97 0.13 NS  0.29 1.23 0.36 NS        102  R? = 0.37 R? = 0.68 0.400.500.600.700.800 100 200 3000.400.500.600.700.800 100 200 300structural diversity crown volume index R? = 0.78 R? = 0.79 0.400.500.600.700.800 100 200 300R? = 0.76 R? = 0.84 0.40.50.60.70.80 50 100 150structural diversity 250 5001000 2000unthinnedR? = 0.50 R? = 0.57 0.400.500.600.700.800 100 200 300R? = 0.69 0.40.50.60.70.80 100 200 300structural diversity Figure 3.11.  Example of relationships (or lack thereof) between total structural diversity (Simpson index) and an index of total stand productivity (crown volume index; m3/0.01 ha) for six of the 12 sample years at the Summerland study area.  Data points represent mean values for each of nine treatment stands.  PCT target densities (250, 500, 1000, 2000 stems/ha and unthinned) are indicated for reference only as regression analyses applied to all treatment stands.  Trendlines and R2 values are only presented for significant regression models.  1994          1996        1998        2000           2002        2008         103  3.4 Discussion 3.4.1 Abundance  While herb volume appeared to be enhanced by PCT, the lack of a significant density effect was surprising given that other studies have observed significant increases in herb abundance following thinning (Thomas et al. 1999; Thysell and Carey 2001; Lindgren et al. 2006; Cole et al. 2010).  Because the overall herb volume within the most heavily thinned stands (250 stems/ha) was nearly twice that of the unthinned stands, the lack of overall density effect was likely obscured by the large variance associated with the treatment means.  For example, the overall mean ? 95% confidence limits for herb volumes within the 250 stems/ha and unthinned stands were 30.68 ? 20.82 and 18.60 ? 9.64 m3/0.01 ha, respectively.  Herb volume within unfertilized stands remained remarkably constant throughout the 16-year sample period; however, volume changed dramatically with fertilization.  The increase in herb volume immediately following fertilization strongly suggested that there were species present at all study areas that were able to quickly take advantage of the added nutrient resource.  Increased competition for nutrients among herbs, shrubs, and trees may have contributed to the decrease in herb volume observed after 2000, six years after the initial fertilization treatment.  However, the post-2000 decrease in herb volume occurred despite the application of an additional two fertilizer treatments, suggesting limitations imposed by resources other than nutrients contributed to the decline.  Both soil moisture and understory light were likely the limiting factors.    Mean abundance of total herbs, grasses, and fireweed all increased dramatically with fertilization in the lightly grazed stands and similar results have been recorded elsewhere, at least where overstory canopy cover was sufficiently open (Kellner and Redbo-Torstensson 1995; VanderSchaaf et al. 2000).  Nams et al. (1993) and Turkington, et al. (2001) also reported overall increases in grasses and some herbs in response to annual fertilizer applications over nine years in Yukon boreal forest The trend of increasing shrub abundance observed among fertilized stands beginning in 2000 (discussed below) coincided with the period of decreasing herb volume among these same stands and suggested that shrub competition, in particular, may have played a role in the herb layer decline.     104   A slight decrease in shrub volume was observed within all thinned stands immediately following PCT treatments and was likely caused by thinning debris, which was left on site and may have either physically damaged shrubs and/or impeded the growth of the plants that were buried and shaded by the debris.  While this effect was not significant, over the longer term, this slight decrease in shrub abundance may have decreased the shrub potential among these stands and contributed to the lack of density effect observed throughout this study.  Because shrubs are much slower than herbs to respond to changing environmental conditions (Halpern 1989; Thomas et al. 1999), the initial decrease in shrub abundance may have indirectly promoted the herb layer and contributed to the herb-dominated community that was observed for at least the first 10 post-treatment years.  Restated, the increased resources provided by thinning were exploited by the herb layer first due to their fast growth and the thinning debris, which may have further slowed the response of shrubs.  Generalizations regarding shrub volume response to thinning are difficult to make as this response has been reported as decreased (Wilson et al. 2009), unaffected (Lindgren et al. 2006), or enhanced relative to unthinned controls (Thomas et al. 1999).  This suggests that site-specific conditions (e.g., plant community composition, level of competition from other plants), treatment intensity (heavy vs. light thinning), and timing of sampling (shrubs respond slowly to treatments) may have considerable influence on the results (Wilson et al. 2009).  Nevertheless, for young stands of lodgepole pine, shrub response to thinning may have been enhanced if excessive thinning debris had been managed. Conversely, if enhanced herb production is a management objective, leaving debris on site may be beneficial.    Shrub volume in unfertilized stands gradually increased throughout the initial 10 years, followed by a gradual decline.  Among fertilized stands, the significantly increased productivity of the herb layer may have contributed to the lack of shrub response during the initial five-year period.  However, the fertilizer effect became marginally significant during the subsequent five-year period as relative growth rates steadily increased, becoming markedly greater than within unfertilized stands.  I suggest that the initial five years represented a period in which slow-growing, shade-tolerant shrubs developed beneath the shade of the more productive, shade-intolerant herbs.  Following this period, some shrubs within the fertilized stands appeared to have escaped the herb layer suppression and increased in abundance, apparently at the expense of shade-intolerant herbs.  The dramatic and synchronous decrease in shrub and herb volume observed from 2003 to 2008, particularly within the fertilized stands, suggested that both of these    105  understory layers were being limited by a third component of the plant community; trees.  Fertilization over a 9-year period in boreal forest in the Yukon increased the growth rates of all shrubs by about 25-30% over control values (Krebs et al. 2001).  Although positive responses were noted at 1-2 years after fertilization, there tended to be a lag time of 5 to 6 years before stable patterns of response appeared (Turkington et al. 1998).   PCT treatments significantly affected tree layer volume in a manner to be expected, with tree volume being greatest within unthinned stands and decreased with increasing thinning intensity.  However, it was not expected that that fertilization would significantly decrease tree volume.  This is counter to the results presented in Chapter 2, where I reported that fertilization enhanced area and volume of the crowns of lodgepole pine crop trees (although not significantly); however, these estimates of tree abundance did not include non-crop trees (ingress).  It is therefore likely that non-crop trees were responsible for the opposite treatment effect observed in this study.  I suggest that the enhanced productivity of the herb and shrub layers may have impeded the establishment of non-crop trees within fertilized stands, whereas such non-crop trees became comparatively abundant within the unfertilized stands, particularly within the thinned stands.  The significantly greater density of ingress for the unfertilized compared to fertilized stands in 2008 supports this suggestion (Chapter 2).  The negative fertilizer response observed for the tree layer volume was not observed for total volume, indicating that the enhanced herb and shrub productivity within fertilized stands was more than enough to compensate for the decrease in tree layer volume.   The enhanced stand-level productivity suggested by an increased crown volume index for the total plant community was to be expected given the large amount of nutrients that were added to these stands (total of 750 kg/ha).  Interestingly, the significant fertilizer effect on total volume disappeared by the final sample, six years after the most recent fertilizer application.  While studies have reported on short-term fertilizer effects, particularly following single applications (e.g., McIntosh 1982), the repeated fertilizer treatments applied during this study were designed to result in long-term increased productivity (Weetman et al. 1995; Kishchuk et al. 2002, Brockley 2005, Lindgren et al. 2007).  It is possible that the sampling methods used to describe the plant community provided a less accurate estimate of plant abundance as succession continued and tree layer dominance increased.  For example, the tallest height class of > 3 m    106  may have limited the ability to distinguish plant communities that differ primarily at heights taller than 3 m. H1, that the application of a range of PCT intensities would increase the crown volume index of herbs and shrubs, was only partially supported as thinning effects were short-term for herbs and not significant for shrubs.  H2, that fertilization treatments would increase crown volume index of all plant layers, was also partially supported, as the abundance of herbs, shrubs, and total plants were all significantly enhanced by fertilization.  Herbs responded faster than shrubs and dominated the understory; however, following a lag-period of approximately five years, shrubs within fertilized stands made a much greater contribution to the understory than within unfertilized stands.  Crown volume index data from the tree layer did not support this hypothesis as it was decreased by fertilization, likely due to the suppression of ingress caused by the abundance of the herb and shrub layers.  3.4.2 Species diversity  Despite the intense nature of some of the density treatments (e.g., PCT to 250 stems/ha), thinning did not result in any significant changes to species richness or diversity.  While not significant, heavy thinning appeared to enhance herb and shrub richness and supported the results of other studies that have found significant increases in species richness following thinning (Thomas et al. 1999; Thysell and Carey 2001; Lindgren et al. 2006).  However, the intermediate disturbance hypothesis (Battles et al. 2001) would have predicted a decrease in species richness following an extreme disturbance in the environment caused by such heavy thinning.  Interestingly, the most significant density effect (P = 0.06) was observed for total species diversity, which was highest within the three most heavily thinned stands and least within the unthinned and most lightly thinned stands.  This suggested that the dominance of the species that thrived under thinned conditions did not negatively influence the evenness of the plant community to the point that diversity suffered.  These data suggest an example of how some forms of disturbance, even at extreme levels, may not necessarily fit the intermediate disturbance hypothesis.   Widenfalk and Weslien (2009) concluded that early PCT of pine and spruce forests throughout central and northern Sweden promoted and sustained high levels of plant species diversity associated with young seral stages.  I suggest that the relative difference between    107  heavily thinned and lightly or unthinned stands will continue to gain in significance with time as the more heavily stocked stands progress more quickly into the stem-exclusion phase of succession.  This suggestion is supported by many other studies that recommend PCT as a means of enhancing habitat and biodiversity conservation because of long-term increased species diversity (as well as enhanced structural attributes) (Sullivan et al. 2001; Lindh and Muir 2004; Homyack et al. 2005; Lindgren et al. 2006; Sullivan et al. 2009; Taki et al. 2010). The high level of fertilization (total of 750 kg/ha added over 10 years) resulted in decreasing species richness; however, decreases were only marginally significant for shrubs during the 1999 to 2003 period.  Species diversity of the understory (herb and shrub layers) was significantly decreased by fertilization and was consistent with the results of several studies (Thomas et al. 1999; VanderSchaaf et al. 2000; Strengbom and Nordin 2008; Oswald et al. 2009).  The suggested mechanism for this decrease in diversity is competitive exclusion, where a few, likely nitrophilic, species were successful at taking advantage of the added nutrients from fertilization and thrived at the expense of other species.  Although competition for light may be a driving force for competitive exclusion (particularly in ecosystems with potential for tall vegetation), other factors such as root competition can also be important (Rajaniemi 2002).  Another process that may have contributed to the decrease in species diversity is one in which fertilization has direct negative effects on some species or groups of species.  Ericaceous plants, for example, have been shown to decrease following fertilization, not simply due to increased competition, but because of decreased fitness and competitive ability brought about by fertilizer-induced changes to soil properties (e.g., pH) and interruption of ericoid mycorrhizae relationships (Prescott et al. 1993; He and Barclay 2000).   Competitive exclusion may have decreased the evenness of the fertilized plant community; however, lack of a corresponding decrease in species richness suggested that species were not lost from the community.  Also noteworthy was the observation that the negative effect of fertilizer on species diversity was not evident during the final year of sampling for either the herb or shrub layers.  This was likely due to a decreasing effect of fertilizer five years following the most recent application.  Also contributing to the lack of fertilizer effect in 2008 was the decrease in diversity within the unfertilized stands, presumably a successional change as shade-intolerant seral species were shaded out by the closing canopy of trees.    108  The loss of a significant treatment effect so soon after the cessation of fertilization differs from the findings of Stengbom and Nordin (2008) who reported significant effects extending more than one forest generation and > 20 years after the most recent application.  Fertilizer effects included increased abundance, decreased richness and diversity, fewer dwarf shrubs, and more grasses and nitrophilous herbs.  Long-lasting and negative impacts of fertilization have serious implications for the sustainability of such forestry treatments and highlight the importance of continued monitoring of these stands to observe whether or not the fertilizer effects have indeed subsided, or perhaps simply been temporarily obscured by successional changes to stand composition.  Neither tree layer nor total species diversity were affected by fertilization, suggesting that the dominance of a few herb and shrub species, which caused the decreased diversity of their corresponding layers, was not as prominent when considering the plant community as a whole. 3.4.3 Structural diversity PCT did not appear to significantly affect the structural richness or diversity of either herbs or shrubs.  This result was somewhat surprising given that several studies found significantly increased structural complexity of understory plants following PCT (Homyack et al. 2005; Sullivan et al. 2001; Lindh and Muir 2004; Lindgren et al. 2006; Sullivan et al. 2009; Taki et al. 2010).  However, significantly greater structural richness and diversity of trees was observed within the 250 stems/ha stands compared with lightly thinned or unthinned stands and was consistent with results of other studies (Lindgren et al. 2006; Moya et al. 2009; Sullivan et al. 2009; Taki et al. 2010).  The enhanced structure of the tree layer following thinning was presumed to have resulted from a larger component of ingress that established and grew within the relatively open growing conditions of the heavily thinned stands compared to the lightly thinned stands.  This assumption was supported by observations of significantly greater densities of trees below the crop tree layer among heavily thinned compared to lightly thinned stands (Chapter 2).  Density effects on total structural richness and diversity were observed during the 1999-2003 and 1994-1998 periods, respectively; however, lack of trend among densities (i.e., increasing or decreasing with thinning intensity) made it difficult to speculate as to the mechanism for these apparent differences.    109  Fertilizer treatments increased the structural complexity of the plant community.  The increased volume of the herb and shrub layers following fertilization suggested that total structural complexity was being enhanced by plants growing into taller height classes with plant abundance being distributed more evenly throughout the plant community profile.  The increased volume and stature of the understory herb and shrub layers and accelerated canopy closure of the tree layer with fertilization (Lindgren et al. 2007; Chapter 2) likely contributed to the suppression of ingress of trees, which in turn contributed to the decrease in structural richness and diversity of the tree layer.  The significantly fewer understory trees in fertilized compared to unfertilized stands reported in Chapter 2 supports this observation.  It was not surprising that that total structural richness was unaffected by fertilizer treatments considering that each of the six height classes would, on average, be represented by at least some plants, resulting in the statistically similar and near maximal structural richness observed for both fertilized and unfertilized stands.  Because of the enhanced structural complexity of the herb and shrub layers, total structural diversity was significantly enhanced by fertilization despite the dominance of the tallest tree layer.  However, as the dominance of the tallest tree layer increased with time, contributions of the understory components were expected to decline as shade-intolerant herbs, shrubs, and trees were shaded below the closing canopy.  Moreover, this successional change was expected to occur more rapidly within the fertilized stands, suggesting that the positive effects of fertilization may be limited to the first 15 years of post-treatment stand development without further density management. H3, that PCT and fertilization treatments would decrease species and structural diversity of all plant layers, was generally not supported.  In terms of species diversity, only herbs and shrubs appeared to be negatively impacted by fertilization, and this effect was short-lived.  On the contrary, heavy thinning and fertilization tended to increase both total species richness and diversity, although not significantly.  In terms of structural diversity, only the tree layer responded negatively to fertilizer treatments, whereas both PCT and fertilizer either enhanced or had no effect on structural richness and diversity for the other plant layers.      3.4.4 Stand productivity and diversity relationships The range of stand productivity (as indicated by total crown volume index) resulting from PCT and fertilization treatments allowed the relationships between productivity and various    110  measures of species and structural diversity to be explored.  The lack of a significant relationship between stand productivity and total species richness for three of four datasets was surprising.  Therefore, as suggested by Bai et al. (2007), the range of species richness among stands may have been more a function of the inherent species composition of the stand than its productivity.  In addition, because the abundance of plants generally increased with fertilization while species richness remained unaffected, species present within these young pine stands appeared to be resilient to forces of competitive exclusion (i.e., species may have decreased with competition; however, they were generally not eliminated from the community).   The possibility that a productivity-species richness relationship would have been evident if a greater range of productivity had been sampled is unlikely due to the wide range of productivity that was provided by the PCT and fertilization treatments.  Because fertilization treatments accelerated tree layer crown development and canopy closure (Lindgren et al. 2007; Chapter 2), the successional shift from young to mature forest conditions, and the associated decrease in species richness as early seral, shade-intolerant species are outcompeted by fewer, taller, and shade-tolerant species (Oliver 1981), is expected to occur sooner with than without fertilizers.  As a result, a negative relationship between productivity and species richness may become evident with time.   The negative linear and hump-shaped quadratic relationships between productivity and species richness for the latter six sample years at Cariboo were consistent with the findings of several studies (Thomas et al. 1999; VanderSchaaf et al. 2000; Strengbom and Nordin 2008; Oswald et al. 2009; Fridley et al. 2012).  A few species likely thrived with the increased resources and outcompeted those species that were not as well adapted to competing within these new growing conditions.  The similar R2-values for the linear and quadratic models (0.22 and 0.28, respectively) indicated that the relationship was best described as linear for these treatment stands.  That a productivity-richness relationship was not evident until several years post-treatment can likely be explained by the time lag inherent for eliminating resident species, sometimes referred to as ?biological inertia? (Von Holle et al. 2003).  The lack of any productivity-richness relationship from the pooled Summerland - Kelowna dataset, even 15 years post-treatment, indicated either greater biological inertia or, more likely, a plant community comprised of species with greater resistance to competitive exclusion (e.g., more shade-tolerant    111  plants) compared with the Cariboo study area.  Alternatively, species lost to competition at higher levels of productivity may have been replaced by new species that were well-adapted for competing within these productive growing conditions, resulting in no net change to species richness.  A species composition analysis through time was required to address such speculation and was not done within the current study. Three of the four datasets assessed for relationships between productivity and total species diversity indicated significant negative linear models of weak to modest strength.  Because these three datasets indicated no productivity-richness relationship, decrease in diversity may have resulted from decreased evenness of the plant community and not loss of species.  Interestingly, the only dataset to reveal a productivity-richness relationship (negative linear relationship observed for Cariboo during the final six sample years) was also the only dataset to not indicate a productivity-diversity relationship.  This suggested that dominance of a few species within species rich stands, and/or a high degree of evenness among species within the species poor stands, obscured any relationship between productivity and diversity among these stands although increased productivity may be associated with fewer species.  The model that consistently best described the relationship between productivity and structural diversity was a hump-shaped quadratic.  The peak in diversity, or vertex of the hump-shaped parabola, changed from the upper to lower range of productivity with time, at least among stands at the Summerland and Kelowna study areas.  This change in vertex from high to low productivity over time also resulted in a change in linear models from positive to negative, which also fit the data well.  Because the change in the productivity-structural diversity relationship was not due to any obvious change in the range of data, succession appeared to be at least partially responsible for the change.  Early in the study, the stands that had received little or no thinning (i.e., stands with the largest tree component and, therefore, greatest estimates of productivity) also had the greatest structural diversity as they would have included a range of tree heights (particularly within the unthinned stands).  These stands also had a diverse understory of herbs and shrubs that were able to grow under the still open canopies.  At this same time, stands that received heavy thinnings would have had a narrower range of tree heights as only the largest trees were retained during PCT.  While heavy thinning would have stimulated the understory herb layer, the contribution to total structural complexity made by this layer was    112  not enough to compensate for the simplification of the tree layer.  In addition, the abundant herb layer and thinning slash impeded shrub growth, all of which resulted in total structural diversity initially being less than within the lightly and unthinned stands.  As trees grew, the lightly thinned and unthinned stands, which remained high in terms of productivity, experienced a gradual decrease in structural complexity as plant abundance became increasingly concentrated within the upper height classes of the taller trees.  This shift in plant abundance was associated with increased canopy closure, decreased understory light, and a simplified understory plant community (reduction of shade-intolerant plants).  Meanwhile, the prolonged open growing conditions provided by heavy thinnings allowed understory herbs and eventually shrubs and tree ingress to develop a complex understory structure despite the continued growth and canopy expansion of the upper tree canopy.  By the end of the study, the lightly thinned and unthinned stands had changed from the most to least structurally diverse.  Restated, the natural decrease in structural diversity observed for forest stands as they progress through the stand initiation and stem exclusion physiognomic stages of forest succession (Oliver 1981) was delayed by heavy thinning. H4, that the relationships between productivity and both species and structural diversity would be negative, was only partially supported.  In terms of species richness, only one of the four available datasets indicated a negative linear relationship, whereas the other three showed no relationship.  A trend of decreasing species diversity with increasing productivity was indicated for three of the four datasets and supported this hypothesis.  The changing relationship between productivity and structural diversity indicated that structure was more related to succession than productivity and, therefore, did not support the hypothesis. 3.5 Conclusions  Intensive management of North America?s boreal and sub-boreal forests, although currently rare, may become an increasingly common strategy for addressing timber supply shortfalls and enhancing economic returns from an increasingly constrained harvestable land base.  This study investigated the long-term effects of PCT and repeated fertilization of young lodgepole pine stands on plant community attributes to increase our knowledge regarding the suitability and, ultimately, sustainability of these intensive treatments.  By manipulating resources of light, moisture, and nutrients, treatments of PCT and fertilization should influence    113  plant community abundance, composition, diversity, and structure.  However, the magnitude and longevity of these changes, prior to my investigation, had not been assessed by a replicated, long-term, and large-scale study.  PCT, even to very low densities (e.g., 250 stems/ha), had very little effect on long-term plant community attributes and significant effects were limited to enhanced structural diversity of the tree layer and a short-term (five-year post-PCT) decrease in total structural diversity.  The enhanced tree layer structure following heavy thinning became more pronounced with time as ingress of trees and understory herbs and shrubs were promoted by open canopy conditions.  The short-term decrease in total structural diversity likely resulted from the initial simplification of the tree layer by thinning of smaller trees.  Fertilization treatments had many more significant effects on plant community attributes; however, most had waned by the end of the 15-year study.  Significant short-term effects of fertilization treatments included enhanced herb volume, which likely contributed to the short-term decrease in herb and shrub species diversity via increased competitive exclusion.  Shrub volume also increased with fertilization; however, this response took five years to be expressed (likely impeded by herb competition and thinning slash).  Increased shrub volume was also short-lived and only marginally significant.  Only tree volume remained significantly affected by fertilizer treatment during the final sample in 2008, five years following the most recent fertilization application.  A decrease in crown volume index of the total tree layer with fertilization resulted from suppression of ingress and should not be interpreted as having decreased the stem volume of crop trees, which was enhanced by fertilization (Chapter 2).   No significant fertilizer effects on species diversity remained at the end of this study.  However, fertilizer effects on plant community structure appeared to be longer lived, with decreased structural diversity observed within the tree layer and enhanced structural diversity for the herb, shrub, and total layers in the heavily thinned stands.  Negative linear relationships between productivity and species diversity were more common than with productivity and species richness, suggesting that the evenness of the plant communities decreased with increased productivity.  However, such changes were often not associated with a decrease in the number of species.  Total structural diversity, while generally enhanced by fertilization, appeared to be affected more by successional change than total plant community abundance.  In addition, by    114  delaying canopy closure, heavy thinning appeared to prolong the enhanced structural diversity associated with younger seral stages compared with the closed canopy conditions of the more densely stocked stands.    Negative impacts to biodiversity and habitat structure by intensive management was not supported by this study.  In particular, no treatment effects were detected for species richness, and negative impacts to other measures of species diversity were short-lived.  In addition, longer term impacts to habitat structure could be interpreted as beneficial for habitat, such as increased total structural diversity.  Therefore, my results suggested that PCT and repeated fertilization of young lodgepole pine stands appear to be management strategies compatible with objectives of biodiversity conservation. It is important to point out that that my assessment of PCT and repeated fertilization treatments were made among young, managed stands, and therefore are relevant to intensive vs. conventional management strategies.  My suggestion that intensive management could be part of a sustainable forest management strategy should not be interpreted as having anything to do with rate of harvest.  Instead, results indicate that intensive management of young lodgepole pine stands could be applied without negatively impacting biodiversity conservation relative to current extensive management.  A conservative forest management strategy should be comprised of a range of management practices, including areas set aside from harvest, as well as areas that receive extensive and intensive management.      115  4 Forage yield and quality response to pre-commercial thinning and repeated fertilization of young lodgepole pine stands: implications for silvopasture management 4.1 Introduction  Understanding how to manage forage yield and quality is essential for effective management of the herbivores that rely on this resource.  Even though seasonal grazing of upland forested range is common throughout the Pacific Northwest (PNW) (McLean 1983), forage production in these forests is poorly understood due to the complex interactions between the overstory trees and understory forage (Krueger 1981; Buergler et al. 2006).  In British Columbia (BC), the cattle (Bos taurus L.) ranching industry could benefit greatly from an improved understanding of these interactions as nearly 80% of this province?s 11 million ha of rangeland is forested (Wikeem et al. 1993a).  Despite the long history of forestry and cattle ranching in BC (Galbraith and Anderson 1971), management of these two resources has remained separate for the most part.  As a result, the potential for enhanced productivity, financial gains, and economic stability considered possible with integrated management (McDonald and Fiddler 1993; Husak and Grado 2002) has been largely overlooked in BC and throughout much of the PNW.   In general, forage productivity peaks soon after any disturbance that removes tree cover (e.g., wildfire, harvesting) and predictably decreases as the tree layer regains dominance and shades out the understory forage.  In BC, natural succession limits the post-disturbance period when the forage resource is capable of sustaining grazing cattle to approximately 10 to 20 years (Basile and Jensen 1971; Wikeem et al. 1993b).  The transient nature of forested range is often further abbreviated by silvicultural objectives that accelerate the development of crop trees.  The fact that both trees and forage compete for the same and often limited resources of sunlight, nutrients, and moisture presents difficulties for successful integration of tree and cattle production on the same landbase.  Nonetheless, silvopastoralism (the deliberate integration of trees with forage and livestock production) has long been successfully practiced throughout the world (Knowles et al. 1973; Msika and Etienne 1989; Sibbald et al. 1989; Papanastasis et al. 1995).  In addition, this form of agroforestry has the potential to be more profitable than single    116  resource management scenarios (McDonald and Fiddler 1993; Clason and Sharrow 2000; Husak and Grado 2002).  Bedunah et al. (1988) noted that silvopastoralism is vastly underutilized and that integrated management could greatly increase productivity within forested range throughout western North America.  The primary means to modify grazing conditions within forested range is to manipulate the density of trees by thinning (Bedunah et al. 1988).  The impact of thinning on understory forage productivity can be dramatic and can be largely attributed to increased levels of understory light.  Removing trees also delays canopy closure, thus prolonging the period of forage production relative to unthinned stands (Pase 1958; Bedunah et al. 1988; Moore and Deiter 1992; Peitz et al. 2001).  Lindgren et al. (2006) reported delayed canopy closure and increased herb production 12 to 14 years following pre-commercial thinning (PCT) of lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) in the central and southern interior of BC.   Although trees compete with forage plants for resources, they also modify understory microclimate and may enhance forage production by providing reduced wind speeds, a more moderate temperature regime, higher humidity, lower rates of evapotranspiration, and higher surface soil moisture levels than in treeless pastures (Lin et al. 2001).  Depending on the conditions of a given stand, these factors may be as important as light levels in determining forage productivity.  Therefore, predicting a forage response to a thinning treatment is not simply related to light conditions.  In addition, improving the light conditions may not necessarily enhance forage production as other resources may still be, or become, limiting.  For example, forage productivity within a very dry or very nutrient deficient site may benefit only slightly, or not at all, from tree removal.  Restated, forage response to thinning is not a simple function of understory light conditions, rather is site specific and will be influenced by a myriad of interactions between light, moisture, nutrient, and temperature gradients, all of which are impacted by succession.  Another silvicultural treatment that can have direct impacts on the forage resource is fertilization.  Fertilization treatments are well-documented to enhance the productivity of either trees (Malkonen and Kukkola 1991; Tamm et al. 1999; Kishchuk et al. 2002; Albaugh et al. 2004; Sword Sayer et al. 2004; Brockley 2005; Lindgren et al. 2007) or forage (Freyman and van    117  Ryswyk 1969; Wikeem et al. 1993b).  However, the benefits of fertilization may become less predictable if managing for both trees and forage.  For example, enhanced growth rates following fertilization may initially enhance the productivity of both trees and understory forage.  However, the accelerated crown closure resulting from fertilization (Lindgren et al. 2007; Chapter 2) will also accelerate understory shading and eventually result in decreased forage productivity relative to unfertilized stands.  The response of forage to fertilization is therefore not linear or easily predicted.   The potential benefits of PCT and fertilization for forage and cattle management are many, including enhanced forage quantity and quality that may be capable of sustaining increased cattle stocking rates over longer periods of time.   This study was designed to test two hypotheses, phrased as predictions that: (H1) PCT  thinning will enhance both the yield and quality of the forage resource within lodgepole pine stands 14 years post-PCT and (H2) repeated fertilization treatments will further enhance both the yield and quality of this forage resource over the same period. 4.2 Methods  4.2.1 Study areas  Two study areas (Summerland and Kelowna) were chosen on the basis of having candidate stands of young lodgepole pine with relatively uniform tree cover, comparable diameter, height, and density of trees prior to stand treatments.  Both areas are located within the Montane Spruce biogeoclimatic zone (dry mild subzone; MSdm) (Meidinger and Pojar 1991).  The Summerland study area was located in the Bald Range 25 km west of Summerland in south-central BC, Canada (49? 40? N; 119? 53? W).  The Kelowna study area was located 37 km northwest of Kelowna, BC (50?04?N; 119?34?W).  Cattle grazing within this study area was very minimal.    To enhance forage production for cattle, both study areas were likely aerially seeded with an agronomic grass-legume mix including species such as timothy (Phleum pretense L.), orchardgrass (Dactylis glomerata L.), smooth brome (Bromus inermis Leyss), intermediate wheatgrass (Thinopyrum intermedium (Host) Barkworth & D.R. Dewey), and alsike clover (Trifolium hybridum L.) (Quinton 1984; Wikeem et al. 1993b).  Until recently, such seeding has    118  been a common range improvement practice throughout the interior of BC since the 1960s (Wikeem et al. 1993a).  Despite this seeding, none of these agronomic species were very prevalent within either study area.  The most abundant graminoid species was pinegrass (Calamagrostis rubescens Buckley) and the most abundant forb species were fireweed (Epilobium angustifolium L.) and arctic lupine (Lupinus arcticus S. Watson). A complete description of the study areas is provided in chapter 1. 4.2.2 Experimental design The two study areas acted as regional replicates (blocks).  Within each replicate, there were three experimental units where lodgepole pine stands were treated (PCT) in the following randomized block design: low density (target 500 stems/ha), medium density (target 1000 stems/ha), and high density (target 2000 stems/ha).  Fertilization treatments were applied to one-half of each of the thinned units, resulting in a total of six stands per study area: 1) 500 stems/ha, 2) 500 stems/ha with fertilization, 3) 1000 stems/ha, 4) 1000 stems/ha with fertilization, 5) 2000 stems/ha, and 6) 2000 stems/ha with fertilization.  These six stands represented a subset of a larger experimental design which included lower and higher density stands.  Additional details for all experimental units are provided in chapter 1 and Lindgren et al. (2007). 4.2.2.1 Treatments PCT was applied in the late summer-early fall of 1993, at which time stands were 12 to 14 years old.   Five years following thinning, heavily thinned stands (? 1000 stems/ha) were pruned to a height of approximately 3.0 m using manual pruning saws.  All thinning and pruning debris (slash) was left on site.  Fertilization treatments were initiated during the fall of 1994 and were repeated at two-year intervals for a total of five applications; spring 1997, fall 1998, fall 2000, and spring 2003.  These five treatments applied 100, 200, 150, 150, and 150 kg/ha nitrogen, respectively (total of 750 kg N/ha), in conjunction with other nutrients.  Complete descriptions of density and fertilization treatments are provided in Chapter 1. 4.2.2.2 Forage clipping  Forage clipping was conducted during early to mid-July 2007 (14 years post-thinning and five years following the most recent fertilizer application), and was designed to coincide with estimated peak forage productivity.  Sampling was completed within two to three days at both    119  study areas.  At each study area, 10 1-m ? 1-m forage clipping plots were located within each stand using systematic placement based on previously established permanent crop-tree plots.  Twenty crop-tree plots (each plot consisting of 10 neighboring crop trees and each tree permanently tagged with individually numbered metal tags), were previously established every 50 m along straight line transects that traverse each stand.  Alternate tree plots were used for general location of each forage clipping plot.  Specific location of a forage clipping plot was determined by placing the plot midway between tree #1 and #10.  To prevent cattle from impacting the clipping plots, a metal, cattle-proof cage was constructed around each clipping plot prior to cattle turnout.   During the clipping sample, cages were removed and all herbaceous vegetation was clipped at ground level using sheep shears and a square 1-m ? 1-m metal frame to accurately delineate sample plot boundaries.  The herbaceous vegetation was sorted into one of four classes: agronomic grasses; pinegrass; other native grasses; and forbs (non-grass herbs).  Due to a lack of dead litter within the sites, only live plant material was removed for analyses.  Samples were placed into appropriately labeled paper bags, which were then loaded into large burlap sacks that were placed in cold-storage at the end of each day.  4.2.2.3 Laboratory analyses  All forage samples were processed at the Agriculture and Agri-Food Canada Research Station in Kamloops, BC.  Samples were oven-dried at 60?C for 48 hours, to achieve a moisture content of 0%, and then ground to pass through a 1 mm sieve using a Wiley Mill.  Samples were individually analyzed to provide measures of dry mass (to nearest 0.001 g), N concentration (nearest 0.001%), and acid detergent fibre (ADF) fraction (nearest 0.001%).    The concentration of N was determined by the Dumas dry combustion technique using a LECO TruMac N Combustion Nitrogen Determinator (LECO; St Joseph, Michigan, USA).  This method exposes the forage sample to O2 and very high temperature (ca. 900 ?C) and combusts the sample, releasing CO2, H2O, and N2.  The N is then quantified using a thermal conductivity detector.  Because protein is approximately 16% N, a conversion factor of 6.25, or 100/16 (also called the N-factor), was multiplied by the sample N concentration to estimate the percentage of protein.  Nitrogen content determined by the Dumas process cannot distinguish protein from non-protein N, and therefore the product of N concentration and N-factor is referred to as crude    120  protein (CP).  As CP is nearly 100% digestible, the higher the CP concentration, the higher the digestibility of the forage.  CP is an important value for describing forage quality; however, it does not provide information as to the amount of CP provided by the forage resource.  Total crude protein (TCP) is the product of crude protein (% CP by oven-dried forage mass) and forage yield (kg/ha oven-dried forage) and is presented as kg CP/ha.   The percentage of a forage sample comprised of fiber (cellulose and hemicellulose) was determined using the micro-digestion procedure that isolates the amount of fiber in a forage sample by boiling it with an acid-detergent solution (Waldern 1971).  The concentration of ADF describes the fraction of the forage comprised of nearly indigestible cellulose and hemicellulose, and hence is inversely related to digestibility (i.e., forages with low ADF usually have higher energy).   4.2.2.4 Canopy closure  At each forage clipping plot, percent canopy closure (CC) was estimated by three spherical densiometer measurements (Lemmon 1956; Englund et al. 2000).  One measurement was taken directly above the sample plot and the other two were taken 3 m either side of sample plot, measured perpendicular to the direction of travel.  This resulted in 30 densiometer measurements per stand. 4.2.2.5 Statistical analyses The two regional replicates functioned as blocks and were assigned as a random factor.  The experimental design restricted the randomness of fertilizer treatment allocation (i.e., applied to one-half of each of the thinned stands), and hence a split-plot analysis of variance (ANOVA) was used to compare treatment means.  The density and fertilizer treatments were assigned as the main- and split-plots, respectively.  Before performing any analyses, data not conforming to properties of normality and equal variance were subjected to transformations to better approximate the assumptions required by ANOVA (Zar 1999).  For variables based on percent cover data (e.g., crown volume index), an arcsine transformation was used (Fowler et al. 1998).  Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare mean values based on ANOVA results (Saville 1990).  In all analyses, the level of significance    121  was at least ? = 0.05.  P-values ranging from 0.06 to 0.10 were reported as marginally significant. Ordinary least squares regression analyses were used to investigate relationships between mean percent crown closure and several different forage attributes, including yield, CP, TCP, and ADF.  Regression models were limited to linear and second order polynomials.  Before pooling data across study areas, regression relationships were assessed for differences in intercept and slope among study areas.  Data were analyzed separately where significant differences were detected.  The strength of a relationship (i.e., degree of correlation) was described with the coefficient of determination and referred to as very weak (R2=0.00-0.03), weak (R2=0.04-0.14), modest (R2=0.15-0.47), strong (R2=0.48-0.78), and very strong (R2=0.79-1.00) as per Fowler et al. (1998). 4.3 Results 4.3.1 Canopy closure   Both PCT and fertilization had significant effects on mean CC 14 years post-PCT and five years following the most recent fertilizer application (Figure 4.1).  Mean CC was significantly (F2,2 =27.20; P=0.04) affected by PCT and increased with stand density.  CC was significantly (DMRT; P=0.05) less within the 500 than both the 1000 and 2000 stems/ha stands (Table 4.1).  Fertilization significantly (F1,3=16.78; P=0.03) increased the mean CC by a factor of 1.20; 64.18 ? 11.23 vs. 76.71 ? 12.49% (mean ? SE).      122  Figure 4.1.  Mean percent canopy closure (CC) in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Error bars indicate SE and are based on two replicate study areas (n = 2).    Table 4.1.  Mean (n = 2 replicate study areas) (SE) percent canopy closure (CC) by density and fertilizer treatment for lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Associated split-plot ANOVA results are also provided.  Within a row, mean values with different letters are significantly different by Duncan?s multiple range test, adjusted for multiple contrasts, if necessary (lower and uppercase letters for density and fertilizer effects, respectively). Treatment Means (SE) Split-plot ANOVA results Density (stems/ha) Fertilizer Density Fertilizer Density ? Fertilizer 500 1000 2000 unfertilized fertilized F2,2 P F1,3 P F2,3 P 53.63b (10.71) 76.62a (8.87) 81.09a (8.13) 64.18B (11.23) 76.71A (12.49) 27.20 0.04 16.78 0.03 0.36 0.72   4.3.2 Yield  Pinegrass, total grasses, and herbs (forbs and grasses combined) all showed a similar response to PCT and fertilization in that yield among heavily thinned stands (500 stems/ha) was enhanced by fertilization.  This response was reversed within the more lightly thinned stands (1000 and 2000 stems/ha) (Figure 4.2).  This resulted in significant density ? fertilization interactions for pinegrass (F2,3=17.27; P=0.02), total grasses (F2,3=11.37; P=0.04), and herbs (F2,3=11.03; P=0.04) (Table 4.2).  Among fertilized stands, the 500 and 1000 stems/ha stands provided 2.31 and 1.80 times greater pinegrass yield than the 2000 stems/ha stands, respectively 0102030405060708090100500 1000 2000CC (%) Target stand density (stems/ha) unfertilizedfertilized   123  (188.22 ? 12.58 and 14.202 ? 4.28 vs. 81.57 ? 11.69 kg/ha), which represented a significant difference (F2,2=32.15; P=0.03) (Figure 4.2a).  Within the 500 stems/ha stands, pinegrass yield was 1.18 times greater with fertilization than without (188.22 ? 12.58 vs. 158.89 ? 7.70 kg/ha); however, this effect was not significant.  Conversely, within the 2000 stems/ha stands, pinegrass yield was significantly (F1,1=3117.99; P=0.01) decreased by fertilization, with unfertilized stands providing 2.69 times greater yield than fertilized stands (219.61 ? 14.16 vs. 81.57 ? 11.69 kg/ha) (Figure 4.2a).  Forb yield was not significantly affected by PCT or fertilizer (Table 4.2).      124  050100150200250500 1000 2000Dry weight (kg/ha) PCT density (stems/ha) a) pinegrass unfertilizedfertilized050100150200250300350500 1000 2000Yield (dried kg/ha) PCT density (stems/ha) b) total grasses 0100200300400500600500 1000 2000Dry weight (kg/ha) Target stand density (stems/ha) c) herbs Figure 4.2.  Mean yield (oven-dried kg/ha) of a) pinegrass, b) total grasses, and c) herbs (grasses and forbs combined) in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Error bars indicate SE and are based on two replicate study areas (n = 2).  Columns with different letters were significantly (P=0.05) different (lowercase and uppercase letters indicate density and fertilizer effects, respectively)    a a b A B B A    125  Table 4.2.  Mean (n = 2 replicate study areas) (SE) yield (oven-dried kg/ha), percent crude protein (CP), total crude protein (TCP; kg/ha), and percent acid detergent fiber (ADF) for various forage components by density and fertilizer treatment in lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer application.  Associated split-plot ANOVA results are also provided for each attribute.  Within a row, mean values with different letters were significantly different by Duncan?s multiple range test, adjusted for multiple contrasts, if necessary (lower and uppercase letters for density and fertilizer effects, respectively). Forage attribute Treatment Means (SE) Split-plot ANOVA results  Density (stems/ha)  Fertilizer Density Fertilizer Density ? Fertilizer 500 1000 2000 unfertilized fertilized F2,2 P F1,3 P F2,3 P Yield (oven-dried kg/ha)             Pinegrass 173.55a (14.69) 163.52a (22.25) 150.59b (57.34) 186.11A (24.88) 139.00B (34.93) 47.36 0.02 16.06 0.03 17.27 0.02  Total grasses 208.96 (38.56) 174.37 (21.55) 153.03 (56.38) 193.03 (22.99) 164.54 (54.25) 7.33 0.12 2.64 0.20 11.37 0.04  Forbs 203.17 (78.58) 150.45 (34.38) 94.51 (44.87) 144.70 (28.73) 154.06 (84.45) 2.11 0.32 0.08 0.80 2.91 0.20  Herbs 412.13 (85.19) 324.83 (37.48) 247.54 (93.39) 337.72 (17.17) 318.61 (124.30) 6.37 0.14 0.30 0.62 11.03 0.04 CP (%)             Pinegrass 10.42 (0.84) 10.98 (0.58) 12.49 (1.11) 10.37B (0.63) 12.22A (0.93) 15.12 0.06 86.44 <0.01 1.83 0.30  Total grasses 9.96 (0.82) 9.98 (0.93) 11.91 (0.66) 9.89B (0.88) 11.34A (0.89) 2.18 0.31 26.97 0.01 0.20 0.83  Forbs 13.77 (0.36) 15.07 (1.05) 15.34 (1.02) 15.24 (0.95) 14.21 (0.84) 1.16 0.46 2.77 0.20 0.26 0.79  Herbs 11.63b (0.59) 12.90ab (0.97) 14.14a (1.02) 12.45B (0.95) 13.33A (1.24) 22.81 0.04 9.92 0.05 1.76 0.31 TCP (oven-dried kg CP/ha)             Pinegrass 18.34 (3.14) 19.60 (1.54) 19.05 (5.50) 20.28 (3.57) 17.71 (3.34) 2.10 0.32 4.85 0.12 27.11 0.01  Total grasses 20.34 (4.10) 18.21 (2.34) 18.05 (5.34) 19.33 (3.44) 18.41 (4.44) 17.10 0.06 0.28 0.64 13.89 0.03  Forbs 28.53 (9.79) 22.20 (6.43) 15.96 (7.60) 22.81 (5.13) 21.64 (11.04) 1.57 0.39 0.04 0.85 1.84 0.30  Herbs 48.87 (11.64) 40.41 (5.58) 34.01 (11.84) 42.14 (4.57) 40.05 (14.36) 2.48 0.29 0.14 0.74 6.03 0.09 ADF (%)             Pinegrass 42.55 (0.90) 43.80 (0.99) 43.34 (1.18) 43.53 (0.91) 42.93 (1.14) 0.49 0.67 0.76 0.45 0.20 0.83  Total grasses 41.60 (1.05) 36.45 (3.05) 36.25 (4.11) 39.99A (3.18) 36.21B (3.05) 2.23 0.31 10.09 0.05 7.34 0.07  Forbs 26.06 (2.43) 25.11 (2.63) 20.12 (6.44) 23.69 (1.81) 23.84 (6.12) 1.50 0.40 0.00 0.97 1.63 0.33  Herbs 35.84 (0.27) 35.13 (0.45) 34.10 (2.66) 36.30A (0.59) 33.75B (1.65) 2.67 0.27 166.61 <0.01 92.85 <0.01      126  4.3.3 Crude protein (CP)  Mean concentration of CP for pinegrass, total grasses, and forbs responded similarly to PCT and fertilization with no interaction among these treatments.  CP concentration decreased with increasing intensity of PCT; however, the density effect was significant (F2,2=22.81; P=0.04) for herbs only.  Herb CP was 1.22 times greater within the 2000 than 500 stems/ha stands (14.14 ? 1.02 vs. 11.63 ? 0.59 %), which represented a significant difference (DMRT; P = 0.05).  The 1000 stems/ha stands had herb CP levels intermediate between the 500 and 2000 stems/ha stands (12.90 ? 0.97%).  Density had a marginally significant (F2,2=15.12; P=0.06) effect on concentration of pinegrass CP (Table 4.2).  CP of pinegrass was significantly (F1,3=86.44; P<0.01) enhanced by fertilization and was 1.18 times greater within fertilized than unfertilized stands (12.22 ? 0.93 vs. 10.37 ? 0.63%).  A similar fertilizer effect was also evident for total grasses and herbs (Table 4.2).  Fertilization reduced the mean CP levels of forbs; however, this effect was not significant (Table 4.2).     4.3.4 Total crude protein (TCP)  Mean TCP of pinegrass and total grasses showed a similar response to PCT and fertilization in that TCP among heavily thinned stands (500 stems/ha) was enhanced by fertilization.  This response was reversed within the more lightly thinned stands (1000 and 2000 stems/ha), resulting in significant density ? fertilization interactions for both pinegrass (F2,3=27.11; P=0.01) and total grasses (F2,3=13.89; P=0.03).  Among unfertilized stands, the density effect on pinegrass TCP was significant (F2,2=76.95; P=0.01), inversely related to PCT intensity, and significantly (DMRT; P=0.05) different among the 2000, 1000, and 500 stems/ha stands (25.67 ? 0.51, 20.49 ? 0.64, and 14.58 ? 0.54 kg TCP/ha, respectively).  Among fertilized stands, PCT enhanced pinegrass TCP as the 500 stems/ha stands provided 1.79 times greater TCP than the 2000 stems/ha stands (22.10 ? 1.05 vs. 12.33 ? 0.62 kg TCP/ha).  However, this density effect was only marginally significantly (F2,2=11.77; P=0.08).  Within the 500 stems/ha stands, fertilization provided 1.51 times greater pinegrass TCP than unfertilized stands (22.10 ? 1.05 vs. 14.58 ? 0.54 kg TCP/ha); however, this effect was not significant.  Conversely, within the 2000 stems/ha stands, pinegrass TCP was significantly (F1,1=139.97; P=0.05) decreased by fertilization as unfertilized stands provided 2.09 times greater TCP than fertilized stands (25.76 ? 0.51 vs. 12.33 ? 0.62 kg TCP/ha).  Mean TCP of herbs and forbs was not significantly affected by PCT or fertilizer (Table 4.2).    127  4.3.5 Acid detergent fiber (ADF)  PCT thinning did not have any significant overall effect on ADF levels for any of the forage components (Table 4.2).  Mean ADF of total grasses was significantly (F1,3=10.09; P=0.05) affected by fertilizer as ADF was 1.10 times greater within unfertilized than fertilized stands (39.99 ? 3.18 vs. 36.21 ? 3.05%).  A significant density ? fertilizer interaction for herb ADF (F2,3=92.85; P<0.01) revealed that the fertilizer effect was significant (F1,1=168.79; P=0.05) among the 2000 stems/ha stands, but not the other densities.  The mean ADF level of forbs was unaffected by PCT or fertilization treatments (Table 4.2). 4.3.6 Relationships between canopy closure and forage attributes  Several forage attributes had strong and significant relationships with CC (Table 4.3).  The form of relationship between pinegrass yield and CC was best described as hump-shaped, with yield peaking at an intermediate CC and tapering off at higher and lower levels of CC (Figure 4.3).  The Summerland model peaked with a slightly greater yield and at a lower CC than did the Kelowna model (ca. 23 kg/ha at 60% CC vs. 20 kg/ha at 75% CC); therefore, relationships were assessed separately for the two study areas (Figure 4.3).      Table 4.3.  Relationship between percent canopy closure (CC) and several forage attributes observed among 26- to 28-year-old lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer treatment.  Only significant (ANOVA; P=0.05) relationships were presented.  Data were pooled across the two study areas (Summerland and Kelowna) only if models were statistically similar among the two areas.  If models were different (i.e., different intercepts and/or slopes), relationships were only presented if both models had the same form.  Forage attribute being related to CC Sample area (n) Relationship Model (y = Const. + B1x + B2x2) R2 F P Form Const. B1 B2 Yield (dry kg/ha)             Pinegrass Summerland (6) 0.91 F2,3=14.81 0.03 quadratic, hump-shaped -44.03 2.34 -0.02 Kelowna (6) 0.88 F2,3=10.69 0.04 quadratic, hump-shaped -86.71 2.99 -0.02 CP (%)              Pinegrass Summerl. & Kelowna (12) 0.70 F2,9=10.42 0.01 quadratic, u-shaped 12.90 -0.13 0.00      Herbs Summerl. & Kelowna (12) 0.89 F2,9=37.40 <0.01 quadratic, u-shaped 13.22 -0.11 0.00 ADF (%)              Grasses Summerl. & Kelowna (12) 0.74 F2,9=12.94 <0.01 quadratic, hump-shaped 20.84 0.81 -0.01       128  Figure 4.3.  Relationship between pinegrass yield (oven-dried kg/ha) and percent canopy closure in 26- to 28-year-old lodgepole pine stands 14 years post PCT and five years following the most recent fertilizer treatment at Summerland and Kelowna.     The relationships between percent CP and CC were similar across study areas for both pinegrass and herbs.  This relationship was U-shaped, with CP increasing with increasing CC.  While the herb and pinegrass models were similar in form, CP was consistently greater for herbs than pinegrass by ca. 1.5 to 2% (Figure 4.4).  The relationship between ADF of grasses and CC was similar across study areas (hump-shaped), with ADF peaking at ca. 55% CC and decreasing as CC increased (Figure 4.5).   R? = 0.91 R? = 0.88 051015202530 40 50 60 70 80 90 100Pinegrass yield (dried kg/ha) canopy closure (%) SummerlandKelowna   129  Figure 4.4.  Relationship between percent crude protein (CP) and percent canopy closure for pinegrass and herbs in 26- to 28-year-old lodgepole pine stands 14 years post PCT and five years following the most recent fertilizer treatment.  Data were pooled across Summerland and Kelowna.     Figure 4.5.  Relationship between percent acid detergent fiber (ADF) for grasses and percent canopy closure in 26- to 28-year-old lodgepole pine stands 14 years post-PCT and five years following the most recent fertilizer treatment.  Data were pooled across Summerland and Kelowna.    R? = 0.89 R? = 0.70 68101214161830 40 50 60 70 80 90 100CP (%) canopy closure (%) herbspinegrassR? = 0.74 25303540455030 40 50 60 70 80 90 100grass ADF (%) canopy closure (%)    130  4.4 Discussion 4.4.1 Canopy closure  By modifying the understory microclimate (e.g., light levels, temperature, wind speeds, soil moisture, nutrient cycling), tree CC is, in large part, responsible for determining understory forage potential.  While thinning trees undoubtedly decreases canopy closure, this effect may be short-lived, as the remaining trees take advantage of increased resources and begin to occupy empty space (expand crown width), eventually returning to pre-thinning levels of canopy closure (Chapter 2; Lindgren et al. 2007).  The more intense thinning is, the longer it will take for thinned stands to attain levels of CC similar to that of unthinned stands (Lindgren et al. 2006).  Indeed, very heavy thinning may result in stands that never close canopy, assuming ingress is restricted by additional thinnings.  Results indicated that the stand density effect on canopy closure was still significant (P=0.04) 14 years post-PCT, with mean canopy closure increasing with stand density.  Fertilization, by increasing the growth rates of trees, including lateral growth of branches, had a direct and positive impact on canopy closure.  With more nutrients available for growth, fertilized trees were able to expand into unoccupied growing space created by thinning and close canopy more rapidly than unfertilized trees (Chapter 2; Lindgren et al. 2007).  A significant fertilizer effect (P=0.03) was evident as fertilizer applications caused canopy closure to increase by 20% compared to unfertilized stands.    4.4.2 Yield  Enhanced forage yield associated with decreased tree density (Chapter 3; Burner and Brauer 2003; Lindgren et al. 2006) and fertilization (Chapter 3; Freyman and van Ryswyk 1969; Wikeem et al. 1993b) has been well-documented.  Forage yield that is limited by light (i.e., overstory shading) can be dramatically enhanced by thinning, and this treatment effect will persist until such time as CC returns to that of unthinned conditions.  Therefore, the short-term effectiveness of thinning for enhancing forage yield is a function of both the environmental conditions that limit forage productivity and the intensity of thinning.  The long-term effectiveness of thinning for enhancing forage yield will be largely determined by the rate at which CC returns to levels similar to that of unthinned stands.  As a result, intensive thinning should result in a greater forage yield (greater increase in understory light conditions) and a more    131  prolonged treatment effect (delayed CC) than light thinning.  This statement was partially supported by Lindgren et al (2006) who reported a significant increase in herb abundance following PCT of young lodgepole pine stands relative to unthinned stands.  However, differences among the three levels of PCT (500, 1000, and 2000 stems/ha) were not statistically significant 12 to 14 years post-PCT (Lindgren et al. 2006).  Similarly, the current study did not detect a significant difference in herb yield among these same levels of PCT, 14 years post-PCT.  However, Chapter 3 reported that herb abundance was 1.83 times greater within the 500 than the 2000 stems/ha stands 15 years post-PCT (11.69 ? 2.10 vs. 6.39 ? 0.42 m3/0.01 ha).  While not statistically significant, this density effect may be biologically and economically important for forage management.   Fertilization can dramatically enhance the yield of forage plants, as well as that of other plants (e.g., shrubs and trees) if productivity is limited by nutrient availability.  A fertilizer effect will last until such time as nutrient availability returns to that of unfertilized stands, which following a single application, is often only one or two years.  Following an application of 200 kg N/ha, Freyman and van Ryswyk (1969) reported forage yield increases of 225 and 229% during the first and second year, respectively; however, yield returned to control levels by year three.  Similarly, Wikeem et al. (1993b) noted that an application of 400 kg N/ha quadrupled forage yields on clearcut-lodgepole pine sites; however, response peaked one year following treatment and was short-lived.  Chapter 3 noted that mean herb abundance (averaged across the 15-year sample period) was significantly enhanced by repeated fertilization.  However, a closer inspection revealed that the fertilizer effect on herb yield was waning by the end of the 15-year sample period. Forage yield data indicated a significant density ? fertilizer interaction, which resulted from different fertilizer effects among the three PCT densities.  For example, within the 500 stems/ha stands, fertilized stands provided 1.18 times greater pinegrass yield than unfertilized stands.  However, within the 2000 stems/ha stands, the unfertilized stands that provided 2.69 times greater pinegrass yield than the fertilized stands.  This interaction can be explained by the direct effect of that fertilization had on CC and the associated effect on understory conditions.  Fertilization has been shown to significantly accelerate the rate of tree crown expansion (i.e., directly related to CC) among these same stands 10 (Lindgren et al. 2007) and 15 years post-PCT    132  (Chapter 2) and was confirmed by specific measurements of CC taken during the current study.  Therefore, results suggest that the initial positive effects of fertilization on understory yield will eventually become offset, and even reversed, by the negative effects of increased shading that result from accelerated growth of overstory trees.  Several other studies have noted the interaction between density and fertilization treatments on understory production (Chapter 3; Hart et al. 1970; Thomas et al. 1999; VanderSchaaf et al. 2002; Lindgren et al. 2007). While the effects of density and fertilization on herb yield were not significant, the grass component of the herb layer (and pinegrass in particular) was significantly affected.  Pinegrass yield was affected by the same density ? fertilizer interaction discussed for herbs.  Among fertilized stands, PCT to 500 and 1000 stems/ha significantly enhanced yield of pinegrass compared to the 2000 stems/ha stands, but that the opposite trend (decreasing yield with increasing PCT intensity) occurred within unfertilized stands.   Fertilization increased pinegrass yield within the 500 stems/ha stands, but this effect was not significant.  Conversely, fertilization resulted in a significant decrease in pinegrass yield within the 2000 stems/ha stands.  Hart et al. (1970) reported a similar negative fertilization response of grass yield following fertilization of a five-year-old slash pine (Pinus elliottii Engelm.) plantation.  Grass yield was initially enhanced, but by the third and fourth year, light interception by trees resulted in less yield than within unfertilized stands.  These negative fertilizer responses contradict the dramatic increases in pinegrass yield often reported to follow fertilization (Freyman and van Ryswyk 1969; Wikeem et al.1993b).  However, these studies measured yield over relatively short periods of time (i.e., less than five years) and test plots were not impacted by overstory trees.   The very strong and significant relationship between pinegrass yield and CC further illustrates the direct influence that the tree layer has on the productivity of this forage species.  Interestingly, the hump-shaped form of this relationship indicated that decreasing CC increased pinegrass yield but, at a given level of CC, further decreases were not associated with increased yields.  Buergler et al. (2005) also indicated a hump-shaped relationship between trees and forage productivity as they found that forage yield was highest at a medium densities.  The point of maximal yield may indicate the condition where resources other than light (e.g., soil moisture) become limiting and positive interactions between trees and the forage crop are greatest.    133  These results suggest that pinegrass yield can be enhanced by PCT and that the longevity of this effect is directly related to the intensity of thinning.  However, yield may be decreased as beneficial contributions of the tree layer (i.e., reduced wind speeds, a more moderate temperature regime, higher humidity, lower rates of evapotranspiration, and higher surface soil moisture levels) may be lost at very high levels of PCT.  In addition, fertilization can enhance yield; however, the effect will be short-lived and even negative if not associated with intensive PCT.   4.4.3 Quality (CP and ADF) The mean concentration of CP among herbs was significantly (P=0.04) decreased by PCT.  The positive correlation between CP and tree density has been reported by many studies (Kephart and Buxton 1993; Lin et al. 2001; Burner and Bauer 2003; Buergler et al. 2006) and several hypotheses have been suggested to explain this relationship.  Temperature and light are both important factors governing plant development, and phenological development of plants is known to be inversely related to the nutritional quality of forage (Oelberg 1956).  Higher temperatures experienced by forage growing in heavily thinned (open) stands would promote a more rapid synthesis of new cells and accelerate maturity and stem development relative to forage growing within shaded environments of lightly thinned or unthinned stands.  This delayed development of shaded forage could explain the increased CP concentrations of forages growing within lightly vs. heavily thinned stands (Kephart and Buxton 1993; Lin et al. 2001).  Higher surface soil moisture levels associated with the more moderate soil temperatures of shaded environments may result in a faster rate of N mineralization, litter breakdown, and turnover of N (Lin et al. 2001).  As a result, more densely stocked stands may have greater N available for protein synthesis compared to heavily thinned stands.  In addition, hot and dry conditions of heavily thinned stands may stimulate a plant survival strategy, whereby drought-stressed understory plants reduce evapotranspiration, which leads to reduced plant nutrient demand, nutrient uptake, and possibly remobilization of nutrients to belowground structures, all of which would reduce concentration of CP within clipped foliage samples (Kering et al. 2011).  A final potential mechanism may be that low light conditions cause cell size reduction with a near-constant quantity of N per cell, which would result in a concentrating effect for N (Kephart and Buxton 1993).  The reported decrease in CP concentration associated with PCT may be a function of several, or all of these mechanisms.  The CP concentration of pinegrass showed a similar response to PCT as herbs, but differences were only marginally significant (P=0.06).      134  Freyman and van Ryswyk (1969) noted that, compared to unfertilized plots, fertilizer applications of 55 and 165 kg N/ha caused significant increases in CP of pinegrass (13.4 and 15.8%, respectively, vs. 9.7%) and a non-significant decrease in ADF (35.8 and 33.1%, respectively, vs. 37.9%) during the summer following spring fertilization.  The higher rate of fertilization also resulted in significantly improved utilization by cattle, indicating enhanced palatability of fertilized pinegrass.  This same study noted further increases in CP when N was applied in conjunction with S; again, these fertilization treatments did not significantly decrease ADF.  The largest fertilization effect was noted at 200 kg N/ha plus 200 kg S/ha, which resulted in increased CP levels of 17.0 vs. 8.5 and 10.0 vs. 8.8% during the first and second growing seasons, respectively.  Similarly, Wikeem et al. (1993b) noted a 28% increase in CP following application of 400 kg N/ha (11.4% vs. 8.9%), although this increase was only significant during the first year following fertilization.  Again, fertilizer effects on ADF were not significant.  My results agree with these studies, as pinegrass CP concentration was significantly (P<0.01) increased by fertilization from 10.37 ? 0.63 to 12.22 ? 0.93% (an increase of 18%).  However, this is the first study to report a significant fertilizer effect on CP content persisting for five years following the most recent fertilizer application.  The relatively high CP concentration of pinegrass growing within fertilized stands and the longevity of this treatment effect was likely a direct result of increased protein synthesis made possible by enhanced plant nutrition sustained by repeated fertilizer applications. While ADF of pinegrass decreased with fertilization (indicating increased digestibility), differences were not significant (P=0.45), similar to the findings of other studies (Freyman and van Ryswyk 1969; Wikeem et al. 1993b).  Silica, which is found in high concentrations in pinegrass, is a component of plant cells that likely decreases both palatability and digestibility of forage.  Freyman and van Ryswyk (1969) noted that fertilization decreased the concentration of silica within pinegrass.  Therefore, fertilization may enhance the digestibility of pinegrass, even though this effect would not be revealed by measurements of ADF.  Although the fertilizer effect on pinegrass ADF was not significant, fertilizer treatments did significantly decrease ADF for the combined grasses (P=0.05) and for herbs within the 2000 stems/ha stands (P=0.05).  Significantly decreased ADF of grasses has been reported following fertilization of silvopastures (Dasci and Comakli 2011; Kering et al. 2011) and is speculated to result from delayed    135  maturation of grasses growing at lower temperatures and with less solar radiation than grasses growing beneath relatively sparse canopies of unfertilized trees.  While several mechanisms have been suggested to explain the response of CP and ADF to density and fertilizer treatments, nearly all are based on changes to temperature and/or solar radiation regimes resulting from tree canopy response to these treatments.  Therefore, it was not surprising to find strong and significant relationships between canopy closure and both CP and ADF.  The U-shaped and hump-shaped relationships for CP and ADF concentrations, respectively, were also reported by Burner and Brauer (2003) for grasses growing within loblolly pine (Pinus taeda L.) stands.  The impact of thinning and fertilization on TCP of pinegrass and total grasses paralleled that of yield.  TCP was enhanced by fertilization at low densities, but was decreased by fertilization at high densities suggests that enhanced yield of crude protein is best achieved by fertilization in conjunction with aggressive thinning. H1, that PCT would enhance both the yield and quality of the forage resource within lodgepole pine stands 14 years post-PCT was generally not supported.  Yield was enhanced by PCT, but only within fertilized stands.  Forage quality was not enhanced by PCT as concentrations of CP decreased within thinning intensity and concentrations of ADF were unaffected.  H2, that repeated fertilization treatments would enhance both the yield and quality of the forage resource within lodgepole pine stands five years following the most recent application was partially supported.  Fertilization did enhance forage yield within the most heavily thinned stands; however, this effect was not significant.  Conversely, fertilization decreased forage yield within the lightly thinned stands, significantly for pinegrass and total grasses.  Forage quality, on the other hand, was enhanced by fertilization as CP of grasses was significantly increased and ADF of grasses and total herbs significantly decreased.  4.4.4 Implications for cattle grazing While pinegrass is cited as being readily consumed by cattle through June, by mid-August it apparently is not (McLean 1967; 1983).  As the grass matures, the CP drops to levels below minimums required for active growth of calves (12%) and eventually below minimum levels required for lactating cows (8.3%) (McLean et al. 1969).  My data indicated that fertilizer applications enhanced CP content of pinegrass to 12.22% compared to 10.37% within unfertilized stands in mid-July.  During this time of year, Freyman and van Ryswyk (1969) noted    136  CP levels of only 8.5 and 8.8%.  Therefore, CP levels provided by fertilized pinegrass should adequately provide the minimum nutritional requirements of a lactating cow beyond the estimated mid-August timeframe indicated for unfertilized pinegrass.  It is important to note that elevated CP levels were sampled five years after the most recent fertilization treatment, whereas a single fertilizer application did not maintain such elevated levels of CP beyond the initial growing season (Freyman and van Ryswyk 1969).  Cattle may not utilize forage under dense forest canopies, regardless of the forage quality, preferring instead to spend the majority of their time within openings or sparse canopy cover (McLean 1983).  The reason for the observed low utilization of dense stands has not been determined, although it is speculated that this behavior may be related to difficulties with locomotion (McLean 1983).  Therefore, in addition to managing for enhanced forage yield and quality, forest range managers may also have to consider the range selection criteria of cattle if the forage resource is to be effectively utilized. 4.5 Conclusions Response of forage yield and quality to treatments of PCT and fertilization are affected by highly complex interactions between the treatments and their effects on microclimate, and, therefore, are difficult to generalize (Buergler et al. 2006).  Forage yield will likely be enhanced by PCT that increases understory light conditions.  However, as target thinning densities decrease a threshold level of light may be reached, where resource limitations other than light (e.g., nutrients and moisture) may negate the short-term benefits of additional thinning.  Moreover, yield may be decreased as beneficial contributions of the tree layer (i.e., buffering effects) are lost at very high levels of PCT.  However, the benefits of delayed canopy closure, and therefore prolonged forage productivity, would still be realized following extreme PCT intensities.  Fertilization can address nutrient deficits and enhance forage yield; however, canopy closure will occur more rapidly, eventually shading out the understory forage.  Therefore, fertilization, for the objective of enhanced forage yield, should be accompanied by aggressive thinning, as higher densities will soon close canopy and negate any benefits of either the fertilization or PCT treatments.  Significantly improved quality of pinegrass following repeated fertilization, even five years after the most recent application, indicates that intensive fertilization prescriptions, coupled    137  with aggressive thinning may 1) extend the period of suitable range condition, 2) allow for increased stocking densities, and 3) allow for extending the grazing season further into the fall.  138 5 Relative habitat use by cattle and mule deer in response to forest thinning and fertilization 5.1 Introduction  Livestock grazing is the most widespread land management practice in western North America and concerns over its potential negative environmental impacts are many (Bock et al. 1993; Fleischner 1994).  Studies have indicated that the ecological consequences of livestock grazing can range from detrimental (Ogawa and Mitamura 1984; Bock et al. 1993; Fleischner 1994) to beneficial (McNaughton 1993; West 1993; Humphrey and Patterson 2000; Pyk?l? 2000), making it difficult to generalize regarding the suitability and sustainability of this practice (e.g., Bock et al. 1993; West 1993).  The impacts of cattle (Bos taurus L.) grazing on native ungulates, such as mule deer (Odocoileus hemionus Raf.), elk (Cervus canadensis L.), moose (Alces alces L.), and bighorn sheep (Ovis canadensis Shaw) have been vigorously debated for centuries (Stover 1985).  Again, disagreements as to the nature of the interactions between cattle and native ungulates have been fueled by wide-ranging and often contradictory conclusions reported by a large number of studies.  At one extreme, cattle grazing has been reported as beneficial for native ungulates; primarily as the result of forage conditioning (Willms et al. 1980b; Willms et al. 1981; Yeo et al. 1993; Vavra 2005).  At the other extreme, grazing by cattle competes with native ungulates and reduces their fitness.  Competition for a limited forage resource is the most frequently cited form of competition (Austin and Urness 1986; Kie et al. 1991; Findholt et al. 2005).  Willms et al. (1980a) noted that diet overlap between cattle and mule deer was most pronounced when forage was abundant.  When forage was limited, diet overlap was only 1 to 10% (Austin and Urness 1986) as these herbivores were able to exploit different forage niches; cattle utilizing graminoids and mule deer utilizing forbs (Willms et al. 1980a; Bedunah et al. 1988; Gibbs et al. 2004).  However, even a small portion of diet overlap can represent an important source of competition when forage opportunities are limited.  Austin and Urness (1986) reported a shift in deer feeding behavior from forb-dominated in areas without cattle, to browse-dominated in areas with cattle; partly due to decreased    139 availability of forbs, but also because of a lack of grass.  During the winter, when forb abundance was severely limited, browse availability for deer becomes critical (Darambazar et al. 2003).   Because of their anatomical and digestive attributes, mule deer are expected to have a more selective diet and choose higher quality forages than cattle (Hofmann 1988).  Willms et al. (1980a) reported that compositional changes of mule deer diet were less than that of cattle when preferred forages became limited, indicating a greater ability for deer to be selective.  Austin and Urness (1986) reported that the high mobility and forage selectivity of unconfined mule deer resulted in minimal impacts of cattle grazing on the diet of deer, thereby strengthening the suggestion that mule deer are able to adapt to changes in their forage resource caused by grazing cattle.  As a result, the potential for competition between cattle and mule deer was often reported as minimal (Austin and Urness 1986; Wallace and Krausman 1987; Coe et al. 2001).   However, Findholt et al. (2005) reported that mule deer switched their diet the least of three herbivores (cattle, elk, and mule deer) in response to previous grazing by cattle.  This lack of flexibility in their diet could result in increased competition with cattle in areas that have high cattle densities or low forage production.  Kie et al. (1991) demonstrated that cattle can compete with deer if cattle are present at high stocking rates and particularly during years with low precipitation.  Forage selection by mule deer is relatively unrestricted during spring and early summer when variety and abundance of forage was at its peak; however, selection becomes more restrictive during late summer (Austin and Urness 1986).  Therefore, livestock grazing is expected to be most critical during late summer when forage and cover are at a minimum (Austin and Urness 1986; Kie et al. 1991; Findholt et al. 2005).   Commensalism among cattle and deer is another potential interaction that can occur if the actions of one species benefit another.  Forage conditioning is an example of commensalism that results if the disturbance caused by cattle grazing enhances the forage resource for native ungulates.  This process may involve removing old growth to stimulate new growth that is more palatable to deer, reducing shrub heights to levels that are available to deer, removing old growth that blocks access to palatable forage for deer,    140 grazing on forage less palatable to deer, or a combination of these effects.  Forage conditioning has been most widely documented for sympatric cattle and elk (Anderson and Scherzinger 1975; Clark et al. 2000; Damiran et al. 2003; Vavra 2005), but has also been demonstrated for mule deer (Willms et al. 1980b).  Willms et al. (1981) suggested that light grazing by cattle could be used to enhance forage for wintering mule deer and, therefore, may increase the use of grazed habitats during the winter (Yeo et al. 1993).  Similarly, Vavra (2005) noted that the properly timed entry with cattle can facilitate a change in the competitive relationships between grasses and forbs or shrubs, suggesting that cattle can be used to enhance winter forage production for mule deer. Cattle grazing may also negatively impact native ungulates by reducing critical vegetative cover, for example, within mule deer fawning habitat (Loft et al. 1987).  The reported impacts of livestock grazing on native ungulates are understandably varied as the nature of the interaction is undoubtedly a function of many factors, including species of livestock, species of native ungulates, livestock grazing intensity, season of livestock grazing, ecosystem, site productivity, and year-to-year variability in weather (e.g., precipitation). Within the Pacific Northwest (PNW), interactions between livestock and native ungulates are primarily a concern within the upland forested range where cattle often graze during spring, summer, and sometimes fall months.  In British Columbia (BC), grazing of forested range is particularly important for the ranching industry as nearly 80% of the 11 million ha provincial-crown range resource is forested (Wikeem et al. 1993a).  Given that rangeland management is expected to intensify to meet the growing global demands for agricultural productivity and sustainability, managers will require a better understanding of the impacts of livestock on grazed ecosystems (Hunter 1999; Watkinson and Ormerod 2001; Steinfeld et al. 2006).  Because cervids are good indicators of ecosystem health (Hanley 1996), monitoring the response of native ungulates to livestock should provide valuable information to improve range management.   Bedunah et al. (1988) noted that the grazing potential of forested range is vastly underutilized throughout the PNW, primarily because the focus has historically been on    141 silvicultural, not range productivity.  Although integrated management of trees and forage (i.e., silvopastoralism) is ecologically and economically feasible (Clason and Sharrow 2000; Husak and Grado 2002), these two resources are often viewed as competitive.  Therefore, management practices that benefit both timber and forage production are of particular interest for promoting silvopastoralism in the PNW.  Pre-commercial (PCT) thinning and fertilization are two proven strategies for enhancing tree growth (Farnum et al. 1983; Weetman et al. 1995; Sedjo 1999; Farnden and Herring 2002; Jokela et al. 2004; Lindgren et al. 2007; Chapter 2) as well as forage production (Chapters 3 and 4) for livestock (Bedunah et al. 1988) and native ungulates (Scotter 1980). Many studies have clearly demonstrated that thinning of overstory trees can significantly increase forage production, primarily by increasing understory light levels (Pase 1958; McConnel and Smith 1965; 1970; Bedunah et al. 1988; Hall 1988; Moore and Deiter 1992; Peitz et al. 2001; Lindgren et al. 2006; Chapter 4).  In addition to enhancing forage production, thinning to low densities (i.e., <1000 stems/ha) can delay canopy closure, and consequently extend the period of suitable range conditions (Lindgren et al. 2006; Chapter 4).   Fertilization treatments can also have dramatic effects on forage quantity and quality.  Freyman and van Ryswyk (1969) found that applications of 100 and 200 kg N/ha enhanced yield, nutritive value, and palatability of pinegrass (Calamagrostis rubescens Buckley) and that this response could be enhanced by also including S with N.  Similarly, Wikeem et al. (1993b) noted that applications of 400 kg N/ha quadrupled forage yields and increased pinegrass crude protein content by nearly 30% on clearcut-lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) sites; however, response peaked one year following treatment and was short-lived.  As was found by Freyman and van Ryswyk (1969), applications of S further enhanced the positive effect of N (Wikeem et al. 1993b).  Five applications of fertilizer over an eight year period were found to provide significantly enhanced forage quality (crude protein levels of pinegrass) in BC lodgepole pine forests five years after the most recent fertilizer application (Chapter 4).     142 Within BC, considerable range management information is available for grasslands and low-elevation forested zones of the southern interior (Wikeem et al. 1993a).  However, very little is known about the upland forested zones.  This study is the first to investigate the impacts of PCT and repeated fertilization on relative habitat use by cattle in these zones.  The wide range of forage opportunities provided by these treatments also made it possible to investigate potential interactions between cattle and relative habitat use by mule deer.    This study was designed to test two hypotheses (H), phrased as predictions that: (H1) PCT and repeated fertilization of young lodgepole pine stands will increase relative habitat use by cattle; and (H2) increased use of forested range by cattle will result in decreased use by mule deer.  5.2 Methods 5.2.1 Study areas Three study areas were chosen on the basis of having candidate stands of young (12 to 14 year old) lodgepole pine that had relatively uniform tree cover, comparable diameter, height, and density of trees prior to stand treatments.  Location, proximity (boundaries), and size of candidate stands were determined by a balance between adequate interspersion of experimental units (Hurlbert 1984) and the logistics and access for conducting the operational-scale treatments of PCT and fertilization.  The study areas are named after nearby towns (Summerland and Kelowna) or forest region in which they are located (Cariboo).  Both the Summerland and Kelowna study areas are located within the Montane Spruce biogeoclimatic zone (dry mild subzone; MSdm), whereas the Cariboo study area is within the Sub-Boreal Spruce biogeoclimatic zone (dry warm subzone; SBSdw) (Meidinger and Pojar 1991). Cattle grazing within the Kelowna study area was minimal.  Range management of the 6050 ha Bald Range Summer Pasture, which includes the Summerland study area, was described as 125 cow/calf pairs, plus 6 bulls, grazed from June 8 to August 31, yielding 366 animal unit months (AUM).  Range management of the 3000 ha range unit, which includes the Cariboo study area, was described as 240 cow/calf pairs, plus 10    143 bulls, grazed from May 16 to June 15, and 50 cow/calf pairs, plus 5 bulls, grazed from September 1 to October 15, yielding 332 AUMs.   A complete description of these three study areas is provided in Chapter 1. 5.2.2 Experimental design The study areas acted as regional replicates (blocks).  Within each replicate, there were five experimental plots where lodgepole pine stands were treated (PCT) in the following randomized block design: very low density (target 250 stems/ha), low density (target 500 stems/ha), medium density (target 1000 stems/ha), high density (target 2000 stems/ha), and unthinned (at least 4000 stems/ha).  Fertilization treatments were applied to one half of each of the thinned units, resulting in a total of nine stands per study area, as follows: 1) 250 stems/ha, 2) 250 stems/ha with fertilization, 3) 500 stems/ha, 4) 500 stems/ha with fertilization, 5) 1000 stems/ha, 6) 1000 stems/ha with fertilization, 7) 2000 stems/ha, 8) 2000 stems/ha with fertilization, and 9) unthinned.  A fertilized, unthinned experimental unit was not included in this design as this treatment combination would not be part of any management prescription.  Additional details regarding the nine experimental units are provided in Chapter 1. 5.2.2.1 Treatments PCT to target densities was done at all study areas in the late summer-early fall of 1993, at which time stands were 12 to 14 years old.  Five years following thinning, heavily thinned stands (? 1000 stems/ha) were pruned to a height of approximately 3.0 m using manual pruning saws.  All thinning and pruning debris (slash) was left on site.  Fertilization treatments were initiated during the fall of 1994 and were repeated at two-year intervals for a total of five applications: spring 1997, fall 1998, fall 2000, and spring 2003.  These five treatments applied 100, 200, 150, 150, and 150 kg/ha N, respectively (total of 750 kg N/ha), in conjunction with other nutrients.  Complete descriptions of density and fertilization treatments are provided in Chapter 1 and Lindgren et al. (2007). 5.2.2.2 Relative habitat use  Sampling of fecal cowpies and pellet-groups was used to measure relative habitat use by cattle and mule deer, respectively.  Cowpies were counted for cattle in summer    144 (May to September) periods and pellet-groups for deer in summer and winter (October to April) periods 1998 to 2003. All new cowpies and pellet-groups (minimum of 20 pellets per group) were counted on cleared permanent 5-m2 (r = 1.26 m) circular plots (Loft and Kie 1988; Edge and Marcum 1989).  Plots were located systematically, in 5-plot arrays installed at stations every 50 m, throughout each stand at the three study areas.  Numbers of sample plots per stand ranged from 55 to 145 at Summerland, 60 to 140 at Kelowna, and 35 to 100 at Cariboo.  Plots were permanently marked with a flagged aluminum ?pig-tail? stake and a small painted rock was located in the plot centre.  A plot boundary was established by rotating a 1.26-m rope attached to the pig-tail stake around the plot centre.  Plots were cleared of all fecal samples, at the initial sampling time in early October 1998. To measure overwinter habitat use by deer, pellet-group counts commenced in the first two weeks of May 1999.  Similarly, relative habitat use in summer was measured by counting cowpies and pellet-groups in the first two weeks of October.  This same procedure was followed for five summer and five winter periods and all sample plots at a given study area were assessed by the same observers throughout the five years.  Fecal sample degradation was not likely an issue as only new samples deposited during a given summer or winter period were counted.  Pellet-groups and cowpies located on the edge of a sample plot had to have 50% or more of the group within the plot in order to be counted.   This technique, consistency of sampling personnel, and the relatively small edge to area ratio of our plots likely minimized potential inclusion bias.  Proportion of sample plots with cowpies was used to measure relative habitat use by cattle.  Density of pellet-groups was estimated per 5 m2 plot and then converted to a per ha value for relative habitat use by deer. 5.2.2.3 Habitat (overstory tree density)  During 1998 and 2003, all trees > 3 m were tallied within 20, fixed-radius (100 m2), non-permanent plots per treatment stand.  Sample plots were systematically located every 50 m along parallel transects separated by 50 m.  The mean number of trees per plot were converted (? 100) to per ha values.  Density of trees > 3 m for the four years between the two sample years (i.e., 1999 through 2002) was estimated by linear extrapolation.    145 5.2.2.4 Statistical analysis  Regional replicates functioned as blocks and were assigned as a random factor.  Three study areas were available for all attributes other than cattle habitat use.  Low numbers of cattle within the Kelowna study area lead to this replicate being omitted from analyses involving cattle.  The experimental design restricted the randomness of fertilizer treatment allocation (i.e., applied to one-half of each of the thinned stands), and hence a split-plot analysis of variance (ANOVA) was used to compare treatment means.  The density and fertilizer treatments were assigned as the main- and split-plots, respectively.  Time (year of sample) was assigned as a split-split plot factor.  Before performing any analyses, data not conforming to properties of normality and equal variance were subjected to transformations to better approximate the assumptions required by ANOVA (Zar 1999).  For count data, either logarithmic, square root, or an area hyperbolic sine function was used depending on the presence of zero counts and the magnitude of variance relative to the mean.  For percent data, an arcsine transformation was used (Fowler et al. 1998).  Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare mean values based on ANOVA results (Saville 1990).  In all analyses, the level of significance was ? = 0.05.  P-values ranging from 0.06 to 0.10 were reported as marginally significant.  The experimental design was unbalanced because of the lack of a fertilized unit for the unthinned density.  In order to preserve the ability to compare treatment means among managed (PCT and fertilized) and unmanaged stands (unthinned with no fertilizer), data were analyzed together.  The resulting degrees of freedom were decreased accordingly to account for the unbalanced design.  Ordinary least squares regression analyses were used to investigate relationships between relative habitat use of cattle and mule deer.  Relationships between habitat attributes and relative habitat use by mule deer were also investigated with regression analyses.  Regression models were limited to linear and second order polynomials.  Relationships were initially assessed for each study area and sample period separately and compared.  Pooling of data was not done if intercepts and/or slopes of regression models were significantly different.  The strength of a relationship (i.e., degree of    146 correlation) was described with the coefficient of determination and referred to as very weak (R2 = 0.00-0.03), weak (R2 = 0.04-0.14), modest (R2 = 0.15-0.47), strong (R2 = 0.48-0.78), and very strong (R2 = 0.79-1.00) as per Fowler et al. (1998). 5.3 Results 5.3.1 Relative habitat use Relative habitat use by cattle was significantly enhanced by fertilizer treatments and was greatest within the 500 stems/ha and least within the unthinned stands (Table 5.1).  Relative to the unthinned stands, the fertilized and unfertilized 500 stems/ha stands received 5.50 and 3.75 times more use, respectively (proportion of sample plots with cowpies = 0.44 ? 0.02 and 0.30 ? 0.04 vs. 0.08 ? 0.05) (Figure 5.1).  The overall density effect was significant (F4,4=194.98; P<0.01).   While habitat use appeared to be positively correlated with PCT intensity, this correlation broke down at very low densities of crop trees.  The 250 stems/ha stands received significantly less (DMRT; P=0.05) use than the 500 stems/ha stands, which received significantly greater (DMRT; P=0.05) use than any of the other stands (Table 5.1).  The overall fertilizer effect was also significant (F1,4=74.20; P<0.01), with fertilized stands receiving 2.07 times more use than unfertilized stands (proportion of sample plots with cowpies = 0.31 ? 0.03 vs. 0.15 ? 0.04) (Table 5.1).      147 Figure 5.1.  Mean index of relative habitat use for cattle (inferred from cowpie samples) among nine young, lodgepole pine stands.  Error bars represent SE and are based on two replicate study areas (n = 2).  Data were collected annually over six years.  00.050.10.150.20.250.30.350.40.450.5250 500 1000 2000 unthinnedprop. of plots with cowpies Target stand density (stems/ha) unfertilized fertilized148 Table 5.1.  Mean (n = 2 replicate study areas) (SE) indices of relative habitat use c by cattle among five levels of density and two levels of fertilization treatments sampled over six years.  Split-plot ANOVA results are also provided.  Within a row, mean values with different letters (lower and upper case letters for density and fertilization treatment means, respectively) are significantly different by Duncan?s multiple range test, adjusted, if necessary, for multiple contrasts.  Treatment Means (SE) Split-Split-Plot ANOVA Results Densitya Fertilizerb Density Fertilizer Density ? Fertilizer Time ? Density Time ? Fertilizer 250 500 1000 2000 Unthinned UF F 0.17c (0.03) 0.37a (0.03) 0.23b (0.04) 0.18c (0.03) 0.08d (0.05) 0.15B (0.04) 0.31A (0.03) F4,4 P F1,4 P F3,4 P F20,45 P F5,45 P 194.98 <0.01 74.20 <0.01 1.06 0.46 0.70 0.80 0.43 0.83 a  Density treatment: PCT target density = 250, 500, 1000, and 2000 stems/ha and unthinned b  Fertilizer treatment: UF = unfertilized; F = repeatedly fertilized c  Relative habitat use was inferred from the proportion of sample plots with cowpies present.     149 5.3.2 Cattle impact on mule deer  Relative habitat use by deer, based on number of pellet-groups per ha, during summer periods was not affected by density (F4,8=1.56; P=0.27), but was significantly (F1,8=17.74; P<0.01) higher in fertilized than unfertilized stands. This pattern in overall mean number of pellet-groups per ha was consistent across all four stand densities during summer months (Figure 5.2a).  In winter, relative habitat use by deer was not affected by density (F4,8=1.50; P=0.29) or fertilization (F1,8=2.08; P=0.19).  However, there was a similar pattern of higher numbers of pellet-groups in fertilized than unfertilized stands (Figure 5.2b).      150 Figure 5.2.  Mean index of relative habitat use for mule deer (inferred from pellet-group counts) among nine, young, lodgepole pine treatment stands during a) summer and b) winter periods.  Error bars represent SE and are based on three replicate study areas (n = 3).  Data were collected annually over five years.     Because the relationships between relative habitat use by cattle and mule deer differed among study areas and from year-to-year, regression analyses were carried out separately by study area and sample period.  No significant relationships were observed between concurrent (summer) habitat use by cattle and mule deer for either study area (Table 5.2).  A significant relationship was observed between habitat use by cattle in summer and habitat use by mule deer 0100200300400500600700800250 500 1000 2000 unthinnedpellet-groups/ha Treatment stands unfertilized fertilizeda) summer 0100200300400500600700800250 500 1000 2000 unthinnedpellet-groups/ha Target stand density (stems/ha) b) winter    151 during the following winter in one of five years at Summerland (Table 5.2).  Relative habitat use by cattle during the summer of 1999 was strongly and significantly (R2=0.60; F1,7=10.33; P=0.02) positively correlated with mule deer habitat use during the following winter (1999/2000) (Table 5.2).  A modest and marginally significant (R2=0.37; F1,7=4.02; P=0.09) positive relationship was indicated for the following winter (2000/2001) (Figure 5.3a; Table 5.2).  Positive relationships were also found for the other three sample years at Summerland, but were not significant (Table 5.2).   For the Cariboo study area, the strongest relationships between summer habitat use by cattle and mule deer habitat use the following winter were negative linear models; however, no relationships were significant.  The strongest relationship occurred during the 1999/2000 period and was only of modest strength and marginally significant (R2=0.37; F1,7=4.08; P=0.08) (Figure 5.3b; Table 5.2).  Table 5.2.  Relationship between relative habitat use by mule deer (y; inferred by number of pellet-groups/ha) and cattle (x; inferred by proportion of sample plots with cowpies present) among nine young, lodgepole pine treatment stands.  Cattle were present only during summers; therefore, the summer models represent concurrent use by these two herbivores, whereas the winter models represent the relationship between habitat use by cattle during the summer and the use of that same habitat by mule deer the following winter.  Linear (y = Const. + B1x) and quadratic (y = Const. + B1x + B2x2) models were explored; however, only the best fit (most significant) were presented.  Season Study area Year Model summary Parameter estimates Form R2 F P Const. B1 B2 Summer Summerland 1999 Linear 0.02 F1,7 = 0.11 0.75 259.05 320.94    2000 Linear 0.21 F1,7 = 1.90 0.21 436.08 -1102.55    2001 Quadratic 0.08 F2,6 = 0.27 0.77 62.78 1737.87 -3620.61   2002 Quadratic 0.08 F2,6 = 0.27 0.77 119.23 7188.09 -23932.28   2003 Linear 0.08 F1,7 = 0.57 0.48 151.54 1077.44   Cariboo 1999 Quadratic 0.24 F2,6 = 0.94 0.44 -26.70 317.68 -474.34   2000 Quadratic 0.02 F2,6 = 0.07 0.93 -9.38 185.63 -266.06   2001 Quadratic 0.27 F2,6 = 1.11 0.39 38.13 -274.87 627.28   2002 Linear 0.10 F1,7 = 0.80 0.40 92.13 -141.36    2003 Linear 0.07 F1,7 = 0.51 0.50 41.79 -109.25  Winter Summerland 1998/99 Quadratic 0.31 F2,6 = 1.32 0.34 -176.72 3325.89 -5118.36   1999/00 Linear 0.60 F1,7 = 10.33 0.02 54.40 1126.27    2000/01 Linear 0.37 F1,7 = 4.02 0.09 172.08 1716.37    2001/02 Linear 0.30 F1,7 = 3.01 0.13 245.28 1152.06    2002/03 Linear 0.25 F1,7 = 2.34 0.17 295.30 1105.94   Cariboo 1998/99 Quadratic 0.32 F2,6 = 1.42 0.31 -107.11 1026.51 -1381.16   1999/00 Linear 0.37 F1,7 = 4.08 0.08 228.58 -383.33    2000/01 Linear 0.25 F1,7 = 2.37 0.17 162.32 -307.29    2001/02 Quadratic 0.12 F2,6 = 0.42 0.68 288.31 -1364.15 2668.24   2002/03 Linear 0.16 F1,7 = 1.37 0.28 180.57 -440.92        152 Figure 5.3.  Relationship between habitat use by cattle in summer 1999 (inferred from proportion of sample plots with cowpies) with habitat use by mule deer (inferred from density of pellet-groups/ha) during the following winter (1999-2000) at a) Summerland and b) Cariboo study areas.     R? = 0.60 02004006008000.00 0.10 0.20 0.30 0.40 0.50Winter mule deer habitat use summer cattle habitat use a) Summerland R? = 0.37 01002003004000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70Winter mule deer habitat use Habitat use by cattle b) Cariboo    153  5.3.3 Overstory trees and winter habitat use by mule deer  The relationship between density of tall trees (> 3 m) and winter habitat use by mule deer at Summerland was consistently negative for all five winters (Table 5.3).  However, only during the winter of 2002/2003 was the relationship of modest strength and significant (R2 = 0.43; F1,7=5.35; P=0.05) (Figure 5.4).  No significant relationships were observed at Kelowna (Table 5.3).  At Cariboo, a strong and significant (R2 = 0.53; F2,6=15.44; P<0.01) U-shaped relationship was observed for the winter of 1999/2000 and a strong, significant (R2 = 0.56; F1,7=8.89; P=0.02), positive linear relationship was observed for the winter of 2000/2001 (Figure 5.5).  A modest and marginally significant (R2 = 0.37; F2,6=4.62; P=0.06) hump-shaped relationship was observed for the winter of 2001/2002 (Figure 5.5).    Table 5.3.  Relationship between density of trees taller than 3 m (x; trees/ha) and relative habitat use by mule deer (y; inferred by number of pellet-groups/ha) among nine young, lodgepole pine treatment stands over five winters.  Linear (y = Const. + B1x) and quadratic (y = Const. + B1x + B2x2) models were explored; however, only the best fit (most significant) were presented.  Study area Winter Model summary Parameter estimates Form R2 F P Const. B1 B2 Summerland 1998/99 Linear 0.15 F1,7 = 1.27 0.30 304.24 -0.03   1999/00 Linear 0.15 F1,7 = 1.20 0.31 317.10 -0.04   2000/01 Linear 0.03 F1,7 = 0.19 0.68 488.93 -0.03   2001/02 Linear 0.11 F1,7 = 0.82 0.40 543.45 -0.04   2002/03 Linear 0.43 F1,7 = 5.35 0.05 524.42 -0.04  Kelowna 1998/99 Quadratic 0.14 F2,6 = 0.50 0.63 121.22 0.24 0.00  1999/00 Linear 0.09 F1,7 = 0.71 0.43 375.16 -0.06   2000/01 Quadratic 0.37 F2,6 = 1.79 0.25 85.06 0.64 0.00  2001/02 Linear 0.05 F1,7 = 0.34 0.58 672.79 -0.10   2002/03 Linear 0.29 F1,7 = 2.81 0.14 850.17 -0.15  Cariboo 1998/99 Quadratic 0.07 F2,6 = 0.23 0.81 78.51 -0.07 0.00  1999/00 Quadratic 0.53 F2,6 = 15.44 <0.01 169.73 -0.22 0.00  2000/01 Linear 0.56 F1,7 = 8.89 0.02 -26.81 0.09   2001/02 Quadratic 0.37 F2,6 = 4.62 0.06 -165.59 0.55 0.00  2002/03 Linear 0.27 F1,7 = 2.55 0.15 209.88 -0.10         154 Figure 5.4.  Relationship between density of tall (> 3 m) trees and winter habitat use by deer during the winter of 2002/2003 (inferred from density of pellet-groups) at Summerland.    Figure 5.5.  Relationship between density of tall (> 3 m) trees and winter habitat use by deer during the winters of 1999/2000, 2000/2001, and 2001/2002 (inferred from density of pellet-groups) at Cariboo.  5.4 Discussion 5.4.1 Cattle and forest management Cattle tend to make greater use of open rather than closed habitats (Willms et al. 1980a; McLean 1983), and accordingly would be expected to favour more heavily thinned stands.  Therefore, it was surprising to note that the greatest relative habitat use consistently occurred within the 500 rather than the 250 stems/ha stands.  Chapter 3 reported that abundance of herbs R? = 0.43 02000400060008000100000 100 200 300 400 500 600 700 800density of >3 m tall trees (stems/ha) winter habitat use by deer (pellet-groups/ha) R? = 0.53 R? = 0.56 R? = 0.37 05001000150020002500300035000 50 100 150 200 250 300 350 400density of >3 m tall trees (stems/ha) winter habitat use by deer (pellet-groups/ha) 1999/2000 2000/2001 2001/2002   155 tended, on average, to be greater in the 250 stems/ha stands than all of the other treatment stands.  This trend supports the expectation that cattle would favour these stands for the enhanced forage opportunities.  When assessing forage abundance data from plots ungrazed by cattle (i.e., within fenced exclosures) and averaged over the same six sample years used to quantify cattle habitat use (1998-2003), the most important forage for cattle, grass (McLean et al. 1969), was 2.66 times more abundant within the 500 than the 250 stems/ha stands (21.74 ? 2.34 vs. 8.16 ? 0.01 m3/0.01 ha; unpublished data), which represented a significant difference (F4,4=35.31; P<0.01; DMRT; P=0.05).  As a result, it would appear that the greater use of the 500 than the 250 stems/ha stands by cattle could be explained by forage opportunities.  Interestingly, grass abundance within ungrazed plots was significantly (DMRT; P=0.05) and 1.84 times greater within the 1000 compared to the 500 stems/ha stands (39.99 ? 2.37 vs. 21.74 ? 2.34 m3/0.01 ha; unpublished data), yet the 1000 stems/ha stands received significantly less use by cattle than the 500 stems/ha stands.  The decreased use of the 1000 stems/ha stands, despite the superior grass forage opportunities provided there, may be explained by the avoidance that cattle often show for densely stocked stands (Willms et al. 1980a). An alternative explanation for the significant, yet nonlinear, response to density treatments could be that some extraneous environmental factors were influencing the habitat selection process.  For example, the proximity of treatment stands to roads, water, and salt licks could all have caused significant impacts to habitat use patterns.  Spatial analyses were not carried out; however, the consistent treatment effects among the two separate study areas strongly suggested that the habitat use patterns by cattle represented a treatment effect and influence from non-controlled factors were likely not significant.  The significantly enhanced use of fertilized stands by cattle can almost certainly be explained by both the greater forage opportunities (Chapter 3) and enhanced quality (crude protein fraction) of the pinegrass (Chapter 4) provided by these stands.  Wikeem et al. (1993b) noted that adequate forage production could be expected to last from 10 to 15 years post-harvest within lodgepole pine stands similar in ecology to the study areas.  PCT and initial fertilization treatments were applied 12 to 15 years post-harvest; the approximate time when forage conditions were likely becoming unfavorable due to overstory shading.  At the conclusion of this study (22 to 24 years post-harvest and 10 years after the onset    156 of treatments), cattle use of fertilized and heavily thinned stands remained enhanced relative to unfertilized and lightly or unthinned stands.  These results suggest that fertilization of PCT stands, particularly at 500 stems/ha, may result in enhanced forage production, a prolonged period of usable range, and greater use by cattle compared to lightly thinned or unmanaged stands. H1, that PCT and repeated fertilization of young lodgepole pine stands would increase relative habitat use by cattle was supported as both treatments did significantly enhance habitat use.  However, response to different levels of thinning was not linear as habitat use increased with thinning intensity to 500 stems/ha.  Thinning below this density resulted in decreased use by cattle. 5.4.2 Cattle impact on mule deer Competition is difficult to evaluate among large herbivores, as a true population-level measure of competition requires the manipulation of one species? numbers, followed by detailed measurements of how the other species responds (Kie et al. 1991; Stewart et al. 2002). However, habitat use patterns and exploration of relationships among these patterns can provide useful insight and suggest possible interactions among species.  While there is a general trend of increased mule deer use of stands with increasing thinning intensity during the summer, the use of the 500 stems/ha stands appeared to not fit well with this trend as it was noticeably less than that of the 1000 stems/ha stands (although the differences were not significant).  Because the 500 stems/ha stands also received significantly more cattle use than any of the other stands, the low mule deer use of these stands may have resulted from a negative interaction (e.g., competition) between these two herbivores.  However, correlation analyses failed to discover any significant relationships between cattle and mule deer habitat use suggesting that these animals were able to coexist within these managed stands without causing shifts in summer habitat use. At Summerland, relationships between habitat use by cattle and habitat use by mule deer the following winter were positive for all five summer-winter periods, although the relationships were only significant during the 1999/2000 period and marginally significant during the 2000/2001 period.  Without experimental manipulation it is not possible to determine whether or not these positive correlations represent neutral or commensal relationships (habitat use by mule    157 deer unaffected vs. enhanced by cattle grazing).  However, these observations strongly suggest that wintering mule deer were not negatively impacted by seasonal cattle grazing at Summerland.   While there were no significant relationships observed between habitat use by cattle and habitat use by mule deer the following winter at Cariboo, the three strongest relationships (R2=0.16 to 0.37) were all negative linear models.  This suggests that some negative interactions among these two herbivores may have occurred, which may represent a biologically important effect, although not statistically significant in this study.  Alternatively, wintering mule deer may be expected to make greater use of the more densely stocked stands that provided greater thermal, security, and snow interception cover relative to the more open stands preferred by cattle (Armleder et al. 1994) particularly during winters with heavy snowfall (Serrouya and D?Eon 2008).  Therefore, negative correlations may have resulted simply by cattle and mule deer having different habitat niches resulting in spatial separation.  The relationships between density of tall trees (> 3 m) and winter habitat use by mule deer generally supported this, as use of stands with higher densities of tall trees (better cover but poorer forage opportunities) was relatively high compared to the less densely stocked stands when winter conditions were severe and vice versa when conditions were more mild.  The two southern study areas generally receive less snow accumulations than the Cariboo study area (Meidinger and Pojar 1991).  Accordingly, winter habitat use by deer at Summerland and Kelowna generally did not indicate significant relationships with density of tall trees.  However, a significant negative relationship was observed at Summerland for the winter period of 2002/2003.  Interestingly, data from the nearest Environment Canada climate station (Peachland Greata Ranch; ID 1126078; Environment Canada 2013) indicated that total snowfall for the 2002/2003 winter was the least of the five winter periods assessed for habitat use, and only half of the average snow depth recorded for this five year period.  These trends suggest that mule deer made greater use of open forage habitat when winter conditions were mild enough to allow it.   The opposite relationship was observed at Cariboo, where three of the five models (two significant and one marginally significant) indicated a trend of increased use of habitats with greater densities of tall trees.  The only model that indicated a negative relationship was for the 2002/2003 winter, which was also the winter that received the least snowfall out of the five winter periods (Ochiltree Miocene Climate Station, ID 1095676; Environment Canada 2013).     158 Although this negative relationship between habitat use and tall tree density was not significant, it may suggest that tall tree cover was less important during winters with lower snow accumulations.  The suggestion that wintering mule deer select for open canopy, forage-rich habitats unless snow accumulations are deep enough that the deer must select for more densely stocked stands with snow-interception cover was supported by Serrouya and D?Eon (2008).  There were insufficient moose pellet-group data to investigate yearly interactions between cattle and moose habitat use.  However, mean data for winter habitat use by moose did not indicate less use of those stands that received the greatest cattle use (500 stems/ha stands) (Sullivan et al. 2006a). Consequently, hypothesis H2, that increased use of forested range by cattle would result in decreased use by mule deer was not supported.  Summer habitat use by mule deer appeared to be a function of forage opportunities and no significant relationships between cattle and mule deer habitat use during the summer were detected.  Negative relationships were better explained by the need for tree cover by deer during severe winter conditions than a negative response to cattle grazing. 5.5 Conclusions  PCT and repeated fertilization applied to young stands of lodgepole pine significantly enhanced habitat use by cattle that seasonally grazed in these managed stands.  In addition, the impact of PCT appears to be relatively long-term as the significant density effect, inferred by six years of fecal cowpie sampling data, was indicated from five to 10 years post-thinning.  Given that PCT took place at an approximate stand age when forage productivity was likely becoming limiting due to overstory tree shading (Wikeem et al. 1993b), PCT also appeared to be an effective means to prolong suitable range conditions for cattle.  While cattle appeared to benefit from PCT, there may be beneficial effects of some tree cover as habitat use was consistently greater within the 500 than 250 stems/ha stands.  The positive response of cattle to PCT and fertilization was likely a function of the enhanced forage opportunities and quality provided within these managed stands (Chapters 3 and 4).  The enhanced productivity of forage provided by fertilization and, to a lesser extent PCT, should, at least, minimize competition and, at best, enhance conditions for both cattle and mule    159 deer.  No competition between cattle and habitat use by mule deer was indicated during this six-year study.  On the contrary, positive relationships between these two herbivores were observed.  It is unclear whether these positive relationships were the result of positive interactions (e.g., commensalism), similar habitat selection criteria for both herbivores (i.e., both preferring forage-rich habitats), or a combination of both.         160 6 Response of plant community abundance and diversity during 10 years of cattle exclusion within managed stands of young lodgepole pine 6.1 Introduction Livestock grazing is the most widespread land management practice in western North America (NA) and is frequently conducted in silvopasture systems (Wikeem et al. 1993b; Fleischner 1994).  Silvopasture is forested range that is intentionally managed as an integrated system to provide both trees and forage for livestock.  Silvopasture systems have been shown to be both economically (Husak and Grado 2002) and ecologically feasible (Krzic et al. 2003).  Forested range is often described as ?transitory? because desired range conditions (adequate forage yields) in upland forests are available only for as long as the tree canopy remains open, allowing sunlight to support understory vegetation.  Following tree removal, range conditions throughout the Pacific Northwest (PNW) of NA typically remain favourable for ca. 10 to 15 years (Wikeem et al. 1993b). Bedunah et al. (1988) noted that the potential for forest range is vastly underutilized and that management could greatly increase productivity in these rangelands.  Silvicultural treatments such as thinning, by reducing shade and delaying canopy closure, may increase forage yields and prolong favourable range conditions (Lindgren et al. 2006).  Fertilization can also have direct beneficial effects on the quality and quantity of forage produced (Wikeem et al. 1993b).  Lodgepole pine (Pinus contorta  Dougl. ex Loud. var. latifolia Engelm.)  is a widespread tree species in western NA and often regenerates over-abundantly, resulting in excessive densities and reduced stand and tree growth (Koch 1996).  By concentrating growth on a smaller number of stems, precommercial thinning (PCT) offers the forest manager some control over the rotation, yield and value of the future crop (Johnstone 1985; Cole and Koch 1996).  In addition, extensive lodgepole pine forests are usually perpetuated through repeated fire disturbance, and therefore often occupy sites of low-N status (Brockley et al. 1992).  Not surprisingly, this species responds well to fertilization to enhance wood production (Lindgren et al. 2007), and potentially these silvicultural treatments could also benefit range condition.    161 Depending on the degree of cattle (Bos taurus L.) grazing, the integration of range and timber management can have both beneficial and detrimental impacts on crop tree production.  Beneficial impacts are primarily related to the reduced herb and shrub abundance often observed following cattle grazing.  The removal of vegetation that would otherwise compete with crop trees for light, nutrients, and moisture may enhance productivity of crop trees as well as decrease rodent feeding damage by reducing habitat quality for small mammals (Bedunah et al. 1988).  Cattle grazing may also enhance site conditions for tree regeneration (e.g., exposing mineral soil), and reduce potential forest fire hazard by the removal of flammable understory vegetation (Bedunah et al. 1988).  Detrimental impacts include browse and/or trampling damage to seedlings and young trees (Newman et al. 1997), soil compaction (Krzic et al. 2003), and increased soil erosion (Bedunah et al. 1988).  In European studies with variable densities of ungulates, grazing and trampling strongly reduced regeneration and life expectancy of saplings of various coniferous (Scott et al. 2000; Chauchard et al. 2006) and deciduous species (Ammer 1996; Kuiters and Slim 2002).   The ecological impacts of livestock grazing on biodiversity and productivity of ecosystems are variable and have been debated vigorously for decades (West 1993; Fleischner 1994).  Similarly, studies that have investigated the impact of livestock disturbance on plant community composition have reported a variety of responses, from an increase in exotic plant species richness (McIntyre et al. 2003), to no change (Stohlgren et al. 1999), to an increase in native species richness (Germano et al. 2001).  Ecosystem response to grazing cannot be generalized and is a function of multiple factors, including grazing intensity (West 1993), individual species? adaptation to grazing disturbance (D?az et al. 2001), as well as soil conditions and productivity of the grazed ecosystem (Bokdam and Gleichman 2000).  The focus of this study was on the impacts of cattle on plant community attributes within upland forest range, where lodgepole pine forests are managed primarily for timber production.  This study was designed to test two hypotheses, phrased as predictions: (H1), that due to the direct impacts of cattle grazing (including trampling, soil compaction, and spread of exotic plant species), the abundance of the understory vascular plant community (herbs and shrubs) will increase in response to cattle exclusion; and (H2), that due to reduced dominance of preferred    162 forage species, cattle grazing will result in an increase in species richness and species diversity of vascular plants. 6.2 Methods 6.2.1 Study areas Two study areas were chosen on the basis of having candidate stands of young (12 to 14 year old) lodgepole pine that had relatively uniform tree cover, comparable diameter, height, and density of trees prior to stand treatments.  The study areas are named after nearby town (Summerland) or forest region in which they are located (Cariboo).  The Summerland study areas is located within the Montane Spruce biogeoclimatic zone (dry and mild subzone; MSdm), whereas the Cariboo study area is within the Sub-Boreal Spruce biogeoclimatic zone (dry and warm subzone; SBSdw) (Meidinger and Pojar 1991). Range management of the 6050 ha Bald Range Summer Pasture, which includes the Summerland study area, was described as 125 cow/calf pairs, plus 6 bulls, grazed from June 8 to August 31, yielding 366 animal unit months (AUM).  Range management of the 3000 ha range unit, which includes the Cariboo study area, was described as 240 cow/calf pairs, plus 10 bulls, grazed from May 16 to June 15, and 50 cow/calf pairs, plus 5 bulls, grazed from September 1 to October 15, yielding 332 AUMs.   A complete description of these study areas is provided in Chapter 1. 6.2.2 Experimental design The two study areas acted as regional replicates (n = 2 blocks).  Within each replicate, there were five experimental units which had lodgepole pine stands treated (thinned) in the following randomized block design:  very low density (target 250 stems/ha), low density (target 500 stems/ha), medium density ( target 1000 stems/ha), high density (target 2000 stems/ha), and unthinned (> 3000 stems/ha).  Fertilization treatments were applied to one half of each of the thinned units, resulting in a total of nine treatment stands per replicate; each providing distinct conditions, as follows: 1) 250 stems/ha, 2) 250 stems/ha with fertilization, 3) 500 stems/ha, 4) 500 stems/ha with fertilization, 5) 1000 stems/ha, 6) 1000 stems/ha with fertilization, 7) 2000 stems/ha, 8) 2000 stems/ha with fertilization, and 9) unthinned (> 3000 stems/ha).  A fertilized    163 unthinned stand was not included in this design as this treatment combination would not be part of a management prescription.  A cattle exclosure was installed in each of the nine treatment stands at both study areas, either in the fall of 1993 or spring of 1994 (prior to annual vegetative growth).  Vegetation was sampled inside and outside of cattle exclosures annually (1994 to 2003) for 10 years following the onset of treatments.  As a result, the experimental design was a split-plot design, with the density treatment as the main-plot (five levels; 250, 500, 1000, 2000 stems/ha, and unthinned), fertilization treatment as a split-plot (two levels; fertilized or not), cattle grazing as a split-split-plot (two levels; seasonal cattle grazing or cattle excluded since 1994), and time as the split-split-split-plot (10 levels; annual sampling from 1994 to 2003). 6.2.2.1 Treatments  PCT to target densities was done at all study areas in the late summer-early fall of 1993, at which time stands were 12 to 14 years old.  Five years following thinning, heavily thinned stands (? 1000 stems/ha) were pruned to a height of approximately 3.0 m using manual pruning saws.  All thinning and pruning debris (slash) was left on site.  Fertilization treatments were initiated during the fall of 1994 and were repeated at two-year intervals for a total of five applications: spring 1997, fall 1998, fall 2000, and spring 2003.  These five treatments applied 100, 200, 150, 150, and 150 kg/ha N, respectively (total of 750 kg N/ha), in conjunction with other nutrients.  Complete descriptions of density and fertilization treatments are provided in Chapter 1 and Lindgren et al. (2007). 6.2.2.2 Cattle grazing  A cattle exclosure was installed in each of the nine treatment stands and provided a comparison of plant community attributes that had developed in forested range with a history of seasonal cattle grazing (hereafter referred to as open range) to those that had developed without cattle grazing (up to 10 years of cattle exclusion).  One exclosure was randomly located within each treatment stand and measured 0.03 ha in size (10 ? 30 m, ca. 1.5 m tall, enclosed by 6 strands of smooth wire).  Native ungulates such as mule deer (Odocoileus hemionus Raf.), moose (Alces alces L.), and elk (Cervus canadensis L.) may also have significant impacts on plant communities    164 (Alldredge et al. 2001).  However, grazing effects noted during this study can be attributed to cattle as exclosure design did not preclude the presence of these native ungulates.  Use of exclosure habitat by native ungulates was confirmed by the presence of deer and moose fecal pellets and bedding sites within exclosures.  All cattle exclosures were installed during the fall of 1993 or early spring of 1994 (prior to cattle release), and, while sampling of open range habitat began in 1993 (pre-treatment), sampling within exclosures did not begin until 1994, one year following PCT. 6.2.2.3 Vegetation sampling  Four vegetation transects were randomly located in each of the nine treatment stands of each study area; one in a cattle exclosure and the other three in open range.  The mean of the three open range transects was used to compare plant community attributes resulting from seasonal cattle grazing to those inside the exclosure (one transect), where cattle grazing had been excluded since 1994 (up to 10 years of no cattle grazing).  Vegetation transects measured 5 m wide by 25 m long and were made up of five contiguous 5-m ? 5-m plots. Each plot contained three sizes of nested sub-plots: the 5-m ? 5-m plot for sampling trees, a 3-m ? 3-m sub-plot for sampling shrubs; and a 1-m ? 1-m sub-plot for sampling herbs. Plants were classified as herb, shrub, or tree based on species, not size.  All plants were subdivided into six height classes: 0-0.25, 0.25-0.5, 0.5-1.0, 1.0-2.0, 2.0-3.0, and >3.0 m (Walmsley et al. 1980).  A visual estimate of the percentage of ground occupied by the vertical projection of a species crown was made for each species height class combination within the appropriate nested sub-plot.  A given plant would only be measured once within the height class in which the tallest part of the plant occurred.  These data were then used to calculate a crown volume index (m3/0.01 ha) for each plant species (Stickney 1985).  Crown volume index is the product of percent cover and corresponding height which gives the volume of a cylindroid, and represents the space occupied by the plant in the community.  Crown volume index values were then averaged by species for each plot size and converted to a 0.01-ha base to produce values for herb, shrub and tree layers.  Sampling was carried out annually during July-August and began in 1993 and 1994 in open range and exclosure plots, respectively.  The final sample took place in 2003.  Plant species were identified in accordance with Hitchcock and Cronquist    165 (1973).  Grasses were not identified to species.  To ensure consistency of any data collection bias, I conducted all of the vegetation sampling. 6.2.2.4 Diversity measures  Plant community diversity was described by both species richness and diversity, and was calculated separately for herb, shrub, and tree layers.  Species richness was the total number of species sampled.  Species diversity was described using the Shannon-Wiener index, which is based on information theory and the degree of difficulty in correctly predicting the next species sampled.  This diversity index is sensitive to changes in rare species, has good discriminant ability, and is well represented in the ecological literature (Burton et al. 1992; Magurran 2004).  Species diversity was calculated using the crown volume index for each plant species averaged across the three transects (or one transect for exclosures), each of which was an average of five sub-plots, in a given stand.  Species diversity was calculated separately for herbs, shrubs and trees. 6.2.2.5 Statistical analysis  In each of the nine treatment stands, plant community attributes that developed with seasonal cattle grazing were estimated as the mean of three vegetation transects located in open range.  Attributes observed without cattle grazing were estimated by a single vegetation transect located within a cattle exclosure.  A split-plot analysis of variance (ANOVA) was used to compare treatment means: density, fertilizer, and cattle grazing treatments were assigned as the main-, split-, and split-split-plots, respectively.  Time was assigned as a split-split-split-plot factor.  The two regional replicates functioned as blocks and were assigned as a random factor (n = 2).  Before performing any analyses, data not conforming to properties of normality and equal variance were subjected to transformations to better approximate assumptions required by any ANOVA (Zar 1999).  For count data (e.g., species richness) logarithmic, square root, or area hyperbolic sine function was used, depending on the presence of zero counts and the magnitude of variance relative to the mean.  For variables based on percent cover data (e.g., crown volume index), arcsine function was used (Fowler et al. 1998).  Duncan?s multiple range test (DMRT), adjusted for multiple contrasts, was used to compare mean values based on ANOVA results    166 (Saville 1990).  In all analyses, the level of significance was at least ? = 0.05.  P-values ranging from 0.06 to 0.10 were reported as marginally significant. 6.3 Results  There were no significant stand density ? cattle grazing interactions noted for any of the plant community attributes.  As a result, the impacts of cattle grazing were presented using means that were averaged across all stand densities. 6.3.1 Lack of pre-treatment exclosure data  Three randomly located open range transects were sampled during the pre-treatment year (1993); however, cattle exclosures were not installed until the first post-treatment year (1994).  Lack of pre-treatment data for the exclosure plots is unfortunate; however, statistical similarity (P > 0.10) in 1993 among all nine treatment stands for all plant community attributes (as indicated by the three open range transects in each stand) suggests strongly that significant differences between open range and exclosure plots can be attributed to exclosure treatments rather than variable site effects.   6.3.2 Herb layer  Prominent herb species included Achillea millefolium L., Antennaria microphylla Rydb., Antennaria neglecta Greene, Antennaria racemosa Hook., Arnica cordifolia Hook., Epilobium angustifolium L., Cornus canadensis L., Fragaria virginiana Duchesne, grasses, Hieracium albiflorum Hook., Lupinus arcticus S. Wats., and Taraxacum officinale Weber.  Additional important herb species at the Cariboo study area only included Aster ciliolatus Lindl., Galium boreale L., Lathyrus ochroleucus Hook., Trifolium pratense L., T. repens L., and Vicia americana Muhl.  Mean crown volume index of the herb layer was significantly (F1,4 = 228.30; P<0.01) greater in fertilized than unfertilized stands (Table 6.1).  Although there were considerable yearly and multi-year fluctuations, herb volume among all stands tended to increase with time until a peak was reached during 2000; volume tended to decrease in all stands thereafter.  Herb volume was noticeably greater within heavily thinned stands (1000 stems/ha or less) than higher density stands; however, the effect of density was not significant in either unfertilized (F4,4 = 0.34; P = 0.84) or fertilized (F3,3=3.50; P=0.16) stands.    167  Throughout this 10-year study, mean herb volume was consistently greater inside than outside cattle exclosures; however, within unfertilized stands, this difference was not significant (Table 6.1; Figure 6.1a).  In fertilized stands, a significant time ? cattle grazing interaction (F9,72=5.03; P<0.01) suggested that the impact of cattle grazing on herb volume had changed during the study.  A closer inspection revealed that, during the first three years after the onset of treatments (1994 to 1996), cattle exclusion was associated with only marginally (F1,4=4.79; P=0.09) greater herb volume compared to open range, whereas during the following seven years (1997 to 2003),  this effect became significant (F1,4=24.38; P=0.01) (Table 6.1; Figure 6.2a).      168 Table 6.1.  Summary of split-plot ANOVA results for plant community response (crown volume index, species richness, and Shannon-Wiener species diversity) to repeated fertilization and cattle grazing treatments (grazed vs. 10 years of cattle exclusion).  Unless otherwise specified, effects correspond to the entire 10-year period of this study (1994-2003).  Significant effects are indicated by bold text.  Fertilizer [unfertilized (U) vs. fertilized (F)]  Cattle grazing [grazed (G) vs. cattle excluded (E)]  Unfertilized  Fertilized Effect F1,4 P  Effect F1,5 P  Effect F1,4 P Herbs             Crown volume index U<F 228.30 <0.01  G=E 1.95 0.22  ?94-?96; G<E 4.79 0.09 ?97-?03; G<E 24.38 0.01             Species richness U>F 7.34 0.05  G=E 0.30 0.61  ?94-?98; G=E 1.42 0.30 ?99-?03; G>E 47.57 <0.01             Species diversity U>F 8.19 0.05  G=E 0.20 0.68  ?94-?99; G=E 1.99 0.23 ?00-?03; G>E 8.08 0.05 Shrubs             Crown volume index U=F 2.78 0.17  G=E 2.02 0.22  G<E 4.99 0.09             Species richness U>F 11.55 0.03  G<E 6.93 0.05  G=E 1.48 0.29             Species diversity U>F 9.54 0.04  G=E 3.66 0.11  G=E 0.60 0.48 Trees             Crown volume index U<F 32.12 <0.01  G=E 0.68 0.45  G=E <0.01 0.96             Species richness U=F 0.66 0.46  G=E 0.34 0.58  G=E 0.22 0.66             Species diversity U=F 0.01 0.91  G<E 5.25 0.07  G=E <0.01 0.97      169 Figure 6.1.  Mean (n = 2) plant community attributes for herbs, shrubs, and trees by fertilization and cattle grazing treatments (averaged across stand densities and years).  Attributes are a) crown volume index, b) species richness, and c) Shannon-Wiener species diversity index.  Differences among cattle grazing treatments (i.e., open range vs. cattle excluded) are indicated by asterisks; * or ** for marginally significant (0.05 < P < 0.10) and significant (P ? 0.05) effects, respectively.  Error bars represent SE.      020406080100120herbs shrubs trees herbs shrubs treesUnfertilized FertilizedCrown volume index (m3/0.01ha) a) Open range Cattle excluded Unfertilized  Fertilized ** ** ** ** * * 05101520herbs shrubs trees herbs shrubs treesUnfertilized FertilizedNumber of species b) 0123herbs shrubs trees herbs shrubs treesUnfertilized Fertilizedspecies diversity index c) ** **  **  *    170 Figure 6.2.  Mean (n = 2) attributes of the herb layer in repeatedly fertilized stands by sample year and cattle grazing treatments (averaged across stand densities). Attributes are a) crown volume index, b) species richness, and c) Shannon-Wiener species diversity index. Differences among cattle grazing treatments (i.e., open range vs. cattle excluded) for a given period are indicated by asterisks; * or ** for marginally significant (0.05 < P < 0.10) and significant (P ? 0.05) effects, respectively.  Error bars represent SE.        0204060801001201994 1995 1996 1997 1998 1999 2000 2001 2002 2003Crown volume index (m3/0.01 ha) a) Open range Cattle excluded* 051015201994 1995 1996 1997 1998 1999 2000 2001 2002 2003Number of species b) Year 01231994 1995 1996 1997 1998 1999 2000 2001 2002 2003Species diversity index c)    171  Fertilization treatments significantly (F1,4=7.34; P=0.05) decreased mean species richness of the herb layer (Table 6.1; Figure 6.1b).  Herb richness in unfertilized stands fluctuated significantly (F9,90=5.74; P<0.01) between ca. 10  and 15 species; however, cattle grazing did not have any significant impact (Table 6.1).  In fertilized stands, herb richness decreased significantly (F9,72=5.32; P<0.01) over time from a high of ca. 13 species at the beginning of the study, to a low of ca. 6 species during the final sample (Figure 6.2b).  In addition, a significant time ? cattle grazing interaction (F9,72=3.39; P<0.01) suggested that the impact of cattle grazing on herb richness had changed during the study.  Cattle grazing had no effect on herb richness during the first five years after the onset of treatments (1994 to 1998).  However, during the latter five years (1999 to 2003), exclusion of cattle caused a significant (F1,4=47.57; P<0.01) decrease in herb richness compared to open range conditions (Table 6.1; Figure 6.2b).    Mean species diversity of herbs was significantly (F1,4=8.19; P=0.05) less in fertilized than unfertilized stands (Table 6.1).  Although there were significant (F9,90=6.01; P<0.01) fluctuations over time, herb diversity in unfertilized stands was very similar at the end of the study (2003) as it was in the beginning (1993).  Cattle grazing in unfertilized stands did not have any significant effect on herb diversity (Table 6.1).  In fertilized stands, herb diversity decreased significantly (F9,72=6.66; P<0.01) over time (Figure 6.2c).  In addition, a significant time ? cattle grazing interaction (F9,72=2.14; P=0.04) suggested that the impact of cattle grazing on herb diversity had changed through time.  Cattle grazing had no effect on herb diversity during the first six years after the onset of treatments (1994 to 1999).  However, during the latter four years (2000 to 2003), cattle exclusion caused a significant (F1,4=8.08; P=0.05) decrease in herb diversity compared to open range (Table 6.1; Figure 6.2c).   6.3.3 Shrub layer  Prominent shrub species included Alnus viridis ssp. sinuata (Chaix) DC. (Regel) A. L?ve & D. L?ve, Linnaea borealis L., Lonicera involucrata (Rich.) Banks, L. utahensis Wats., Pachistima myrsinites (Pursh) Raf., Ribes lacustre (Pers.) Poir., Rubus idaeus L., Arctostaphylos uva-ursi (L.) Spreng., Salix spp. L., , Spiraea betulifolia Pall., and several species of Vaccinium L.  Other common shrub species in the Cariboo study area included Amelanchier alnifolia Nutt., Mahonia aquifolium Pursh, Rosa acicularis Lindl., and Shepherdia canadensis (L.) Nutt.    172  Although mean crown volume index of the shrub layer was noticeably greater in fertilized stands, the large variability (particularly among fertilized stands) resulted in a non-significant fertilization effect (Table 6.1).  Mean volume of shrubs increased significantly with time in both unfertilized (F9,90=11.01; P<0.01) and fertilized stands (F9,72=8.28; P<0.01). Compared to open range conditions, cattle exclusion consistently increased shrub volume (Figure 6.1a).  This cattle grazing effect was not significant for unfertilized stands; however, it was marginally significant (F1,4=4.99; P=0.09) for fertilized stands (Table 6.1; Figure 6.1a).  Mean species richness of the shrub layer was significantly less (F1,4=11.55; P=0.03) in fertilized than unfertilized stands (Table 6.1).  Shrub richness in unfertilized stands fluctuated significantly (F9,90=2.37; P=0.02) between ca. 8 to 10 species and was significantly (F1,5=6.93; P=0.05) greater with cattle exclusion compared to open range conditions (Figure 6.1b).  A marginally significant (F4,5=3.92; P=0.08) density ? cattle grazing interaction  revealed that the cattle grazing effect was primarily due to differences observed in stand densities of 1000 stems/ha and higher, and that shrub richness in the 250 and 500 stems/ha stands appeared to be relatively unaffected by cattle grazing.  Over time, shrub richness decreased significantly (F9,72=6.86; P<0.01) in fertilized stands, from a high of ca. 9 species at the beginning of the study, to a low of ca. 6 during the final sample.  While excluding cattle appeared to decrease shrub richness compared to open range (Figure 6.1b), this difference was not significant (Table 6.1).  Mean species diversity of shrubs was significantly (F1,4=9.54; P=0.04) less in fertilized than unfertilized stands (Table 6.1).  Shrub diversity did not change significantly with time in unfertilized stands (F9,90=0.85; P=0.57); however, it did decrease significantly in fertilized stands (F9,72=10.72; P<0.01).  6.3.4 Tree layer  In addition to lodgepole pine, prominent tree species included Douglas-fir, subalpine fir, hybrid interior spruce, and trembling aspen.  Mean crown volume index of the tree layer was significantly (F1,4=32.12; P<0.01) greater in fertilized stands (Table 6.1).  Tree volume generally decreased with increased levels of thinning.  However, the impact of density was only marginally significant for both unfertilized (F4,4=5.45; P=0.06) and fertilized stands (F3,3=8.39; P=0.06).     173 Cattle grazing did not have a significant impact on tree volume in either unfertilized or fertilized stands (Table 6.1).   Mean species richness of the tree layer was unaffected by fertilization or cattle grazing treatments (Table 6.1; Figure 6.1b).  Mean species diversity of trees was unaffected by fertilization treatments (Table 6.1).  In unfertilized stands, tree diversity tended to be greatest within the very low-density stands (250 stems/ha); however, differences were only marginally significant (F4,4=5.34; P=0.07).  Tree species diversity increased significantly (F9,90=7.85; P<0.01) with time in unfertilized stands.  Compared to open range conditions, exclusion of cattle appeared to increase tree species diversity in unfertilized stands (Figure 6.1c); however, differences were only marginally significant (F1,5=5.25; P=0.07) (Table 6.1).  In fertilized stands, tree species diversity was significantly different (F3,3=15.59; P=0.02) among stand densities, with the 250 and 500 stems/ha stands having greater (DMRT; P=0.05) diversity than both the 1000 and 2000 stems/ha stands.  Tree species diversity in fertilized stands increased significantly (F9,72=4.53; P<0.01) with time; however, cattle grazing had no effect (Table 6.1).  6.4 Discussion  To my knowledge, this study is the first large-scale and relatively long-term (10 years) investigation of plant community response to the exclusion of cattle in forested range in North America.  Plant community attributes were compared inside and outside of cattle exclosures across a wide range of conditions created by intensive silviculture treatments of PCT and repeated fertilization.   6.4.1 Abundance of vegetation  Lack of a significant density effect on herb volume was surprising given the numerous studies that have found very significant increases in herb and forage abundance following thinning (Burner and Brauer 2003; Lindgren et al. 2006).  Hall (1988) even suggested reducing tree density as the main method for enhancing forage for livestock.  Thinning treatments may not have had a significant impact on herb volume due to the young, open canopy conditions found within all stands, even high-density stands, resulting in high herb volumes regardless of thinning treatments.  Thinning effects may become significant with time as higher density stands begin to close canopy and shade out understory species.    174  Crude protein of an important forage species, pinegrass (Calamagrostis rubescens Buckley), was significantly enhanced by the repeated fertilization treatments (five years following the most recent fertilization), as was forage productivity (Chapter 4). Freyman and van Ryswyk (1969) found that N fertilizer (100 to 200 kg N/ha) improved palatability, nutritive value, and forage standing crop in forested rangelands which were very similar in ecology to my study areas.  Other studies also reported significant increases in forage abundance and quality following fertilization (Wikeem et al. 1993b; Kalmbacker and Martin 1996).  While Wikeem et al. (1993b) found that the effects of a one-time application of fertilizer (400 kg N/ha) diminished with each passing year and lasted only three years, the repeated fertilization treatments (100 to 200 kg N/ha every two years) continued to enhance herb and shrub volumes for the duration of the study.  In fertilized stands, initially, cattle grazing appeared to have little impact on herb volume compared to that in exclosures.  A possible explanation for this is that the initial overall low abundance of herbs was insufficient to increase the utilization of these stands by cattle.  As forage opportunities were enhanced by fertilization treatments (i.e., increased abundance, palatability, and nutritive value), cattle grazing pressure also appeared to increase, which eventually lead to significant reductions in herb volume compared to exclosure plots.  The data suggest that stands with mean volumes of herbs greater than ca. 50 to 60 m3/0.01 ha appeared to trigger selection by cattle, which then resulted in the significant reduction in herb volume observed after 1996.  Unfertilized stands, which never attained these herb volumes, had similar herb volume inside and outside of exclosures.  Selection for fertilized stands was also suggested by cattle activity (inferred from fecal cow-pie samples), which indicated that mean habitat use was 2.07 times greater within fertilized than unfertilized stands during the final six years of the study (Chapter 5).  In fertilized stands, shrub volume was also reduced by cattle grazing, although the difference between exclosures and open range was only marginally significant (P=0.09).  Reduced shrub volumes may have resulted from the direct effects of grazing.  However, because cattle generally prefer herbs to shrubs (Bedunah et al. 1988), the decrease in shrub volume may have been the indirect result of increased cattle disturbance, such as trampling and soil compaction, observed in fertilized stands.  The relatively large and increasing SE of shrub    175 volume in fertilized exclosures compared to fertilized open range suggested that cattle exclusion may have increased structural diversity of the shrub layer in these stands and that this effect increased with time.  Cattle grazing did not have any effect on the crown volume of the tree layer.  The hypothesis that exclusion of cattle grazing would increase volume of herbs and shrubs compared to open range, was only partially supported as this pattern occurred in fertilized, but not unfertilized stands. 6.4.2 Species richness and diversity  Fertilization decreased both species richness and diversity of herb and shrub layers.  For both herbs and shrubs, richness and diversity in fertilized stands decreased with time; however, they remained relatively unchanged for unfertilized stands.  Reduced species richness and diversity after fertilization was expected and was likely explained by principles of competitive exclusion.  The theory of competitive exclusion suggests that as the productivity of an ecosystem is increased, a few species tend to dominate and even exclude less competitive species, resulting in a community with possibly reduced richness, and almost certainly reduced diversity.  The response of species richness and diversity to fertilization can also be explained by the hump-back response model (Oba et al. 2001), which suggests that richness will be highest at an intermediate level of biomass and will decline as biomass increases, as is commonly observed following fertilization.  Other studies have also reported decreased richness and diversity after fertilization (Proulx and Mazunder 1998; Thomas et al. 1999).  Cattle grazing, by reducing the dominance of forage species that flourished following fertilization, appeared to slow the rate of competitive exclusion and increased herb richness and diversity relative to exclosure sites.  Before the impact of cattle grazing on herb species richness and diversity became significant, there was a period of up to five years when these measures were similar inside and outside of exclosures.  Cattle grazing has also been found to decrease litter cover (dead plant material) (Hayes and Holl 2003).  Excluding cattle, therefore, may lead to an accumulation of litter, which may impede the growth of some plants (particularly low-growing species) and would contribute to the lower species richness and diversity observed within cattle exclosures.  Another study also reported cattle to have a beneficial impact on species richness and diversity (Krzic et al. 2003).    176  In unfertilized stands, cattle grazing decreased shrub species richness and diversity compared to plots where cattle were excluded, although the effect on species diversity was not significant (P=0.11).  Shrub richness and diversity may have been greater within exclosures due to the presence of species sensitive to cattle grazing.  Moser and Witmer (2000) also reported significantly greater shrub species richness in ungrazed compared to grazed sites.   The effect of cattle grazing on species richness and diversity were reversed in fertilized and unfertilized stands.  Proulx and Mazumder (1998) and Harrison et al. (2003) also reported a similar interaction between grazing and nutrient status of a site.  Proulx and Mazumder (1998) found that grazing in nutrient-poor sites tended to decrease species richness due to lack of resources to support regrowth following grazing, while grazing within nutrient-rich sites tended to increase richness due to rapid regrowth as well as the establishment of inedible species.  The hypothesis that exclusion of cattle grazing would decrease plant species richness and diversity, seemed to be supported, but only in fertilized stands.  In fertilized stands, herb species richness and diversity were greater within open range compared to plots with cattle excluded.  Cattle grazing did not appear to have any impact on volume, richness, or diversity of the tree layer in these young lodgepole pine stands. 6.5 Conclusions  The impacts of cattle grazing on plant communities in young lodgepole pine silvopastures cannot be generalized as beneficial or detrimental, as the response is undoubtedly a function of grazing intensity and nutrient status of the ecosystem.  In productive sites, such as fertilized stands, impact of cattle grazing was beneficial, particularly for the herb layer, as grazing decreased volume and slowed the rate of competitive exclusion, resulting in increased species richness and diversity.  However, the beneficial impacts were not initially evident, as the combined effects of fertilization and cattle grazing took as long as five years to produce significant differences between exclosure and open range conditions.  Impacts of cattle grazing may be reversed in nutrient-poor sites, particularly for the shrub layer.  In unfertilized stands, cattle grazing did not significantly reduce herb or shrub volumes; however, richness and, to a lesser extent, diversity of the shrub layer declined.  Incremental silviculture treatments of PCT and repeated fertilization of young lodgepole pine stands    177 enhanced the range resource by increasing both the amount and quality of forage and, by delaying canopy closure, increasing the duration of this transient range.  Encouraging heterogeneity through a range of grazing pressures is a method of conserving biodiversity within rangelands throughout the world (Fuhlendorf and Engle 2001; Hayes and Holl 2003; McIntyre et al. 2003).  Consequently, to retain plant diversity within upland forested rangelands, managers should strive to include a diversity of grazing pressures, including both intensive grazing as well as areas with no grazing.  This study has clearly demonstrated that the influence of cattle grazing on plant community abundance and diversity may be directly impacted by forest enhancement treatments of repeated fertilization and, to a lesser extent, PCT.  Strategies for conservation of plant diversity within forested range should include a diversity of forest enhancement treatments.     178 7 General conclusions 7.1 Study strengths and limitations This study involved a major effort in terms of its operational scale, regional replication, and length of time that treatment effects have been monitored.  The combined area of the three study areas is over 188 ha, of which 167 ha were PCT and 126 ha were pruned.  Over 83 ha were fertilized five times during a 10 year period, amounting to 750 kg N/ha (a total of nearly 63 tonnes N).  These study areas represent the most intensively managed, operational-sized stands in all of Canada.  The intensive and long-term sampling regime further distinguishes this project.  The following is a summary of the total sampling effort that has taken place: ? Crop tree growth: more than 5000 permanently tagged crop trees (ca. 200 per stand) have had multiple measurements taken at five-year intervals for 15 years (1993, 1998, 2003, 2008);  ? Crop-tree foliar analysis: foliage samples from 270 randomly selected crop trees (10 per stand) were collected and assayed annually for nine years (1995-2003); ? Stand structure: more than 500 100-m2 plots (ca. 20 per stand) have had tree height distribution measured at five-year intervals for 10 years (1998, 2003, 2008);  ? Plant community: more than 100 125-m2 permanent vegetation plots (four per stand, including one located inside a fenced exclosure) have been sampled 12 times over 16 years (1993-2003 and 2008); ? Relative habitat use: more than 1250 5-m2 permanent pellet plots (ca. 100 per stand) were sampled semiannually for six years (1998-2003); and ? Forage quality and quantity: 180 1-m2 sample plots (10 per stand at six stands) were clipped and assayed in 2007. Few research projects have benefitted from such long-term and intensive sampling.  The resulting datasets have provided a wealth of information that only just begun to be interpreted.  In addition, the scope of inference from the results of this study is very large due to the fact that this study was conducted within zonal sites (Lloyd et al. 1990; Steen and Coupe 1997) of one of    179 the most widespread forest types of the Pacific Northwest, lodgepole pine (Lotan and Critchfield 1990).    However, there were some limitations associated with the experimental and sampling designs that restricted the effectiveness of the data.  Most notably, the experimental design was unbalanced due to the lack of a fertilized counterpart for the unthinned density.  This treatment was not included within my experimental design due to the well-documented lack of operational feasibility of such a treatment (Groot et al. 1984; Yang 1998, Vallinger et al. 2000).  However, there would have been statistical advantages to including this treatment.  Crop-tree sample plots were variable rather than fixed in area (i.e., based on selecting the 10 nearest crop trees to a plot centre), which made calculating area-based estimates of tree growth problematic, particularly for the unthinned stand.  Stand-level attributes were calculated by multiplying mean tree-level values by stand density; however, density estimates for the unthinned stands included all stems, whereas tree sampling focused on just the dominant and co-dominant stems.  Consequently, sample estimates of both tree- and stand-level growth overestimated the population means.  Consequently, the unthinned treatment was omitted from all statistical analyses of crop-tree growth, which limited the comparisons of treatment effects to the thinned stands.  Due to logistical and funding limitations, forage sampling was not conducted within the 250 stems/ha or unthinned stands.  Forage quantity and quality data from the 250 stems/stands, in particular, would have been interesting as it may have helped explain the surprising lack of cattle habitat use within these stands.  In addition, lack of measurements of weather and climatic variables, as well as soils, limited my ability to explain several of the observations made throughout this study.  Finally, a lack of spatial analysis to account for potential influences on habitat use of extraneous habitat features (e.g., water, roads, salt licks) and proximity of treatment stands to one another was as a limitation of this study.  The consistent habitat use trends noted for cattle and mule deer among the replicate study areas suggested that treatment effects were the primary factor influencing habitat use; however, spatial analyses may have provided further support for this statement.     180  7.2 Future opportunities First and foremost, future opportunities exist for continued monitoring of these research installations.  While specific treatment effects were observed during the initial 16 years of this project, several important questions remain unanswered.  For example, how long will a fertilizer effect be maintained for tree growth and forage quality?  What is the contribution of ingress to total stand productivity and value?  Will the treatment effects on plant community abundance, species diversity, and structural diversity be limited to early stand development, or will they remain, or perhaps reappear, as the stands mature?  How long will range conditions remain suitable for cattle?  How will plant communities continue to respond to cattle exclusion?  Studies have reported that recovery of grasslands following the removal of cattle may take as little as 20, or more than 65 years (Valone et al. 2001; Courtois et al. 2004).  Following long-term cattle grazing, some grasslands may develop into successionally stable communities that may never revert to the original pre-cattle grazing condition without management intervention (West et al. 1984).  Forested rangelands have also been shown to take 20 to 40 years to recover from cattle grazing (McLean and Tisdale 1972).  The value of future monitoring of these research installations cannot be overstated.  Although the Cariboo and Kelowna replicates have been affected by mountain pine beetle (MPB; 2005) and the Terrace Mountain wildfire (2009), respectively, continued monitoring of these study areas can still provide very important information.  For example, it is hypothesized that the contribution of understory trees (i.e., advanced regeneration or ingress) may mitigate losses resulting from overstory mortality caused by MPB (Vyse et al. 2009; Kayes and Tinker 2012).  The substantial density of ingress noted within heavily thinned stands may be an added benefit of such treatments that could minimize the need for costly replanting and accelerate the recovery of MPB-killed stands.  The primary recommendation that I would make for new research projects investigating the effects of PCT and repeated fertilization would be to include soil sampling.  The effects of fertilization on soil characteristics (e.g., chemistry, nutrient regime, soil flora and fauna) are undoubtedly substantial, as are the effects of cattle presence on nutrient cycling and possibly soil compaction.  A better understanding of the effects on below-ground attributes would greatly    181 enhance the ability to explain the above-ground effects noted within my study.  I would also recommend that destructive sampling of crop trees be carried out to enhance our understanding of the treatment effects on stem form (i.e., taper) and wood quality.   7.3 Contributions to forest and range management In general terms, the primary contribution of my thesis is to provide long-term scientific data supporting the application of PCT and fertilization for enhancing both silvicultural and range productivity within lodgepole pine stands.  Another important contribution is the data supporting the sustainability of these intensive management treatments. Specific results of my thesis that could contribute to the management of lodgepole pine stands are: ? Silvicultural losses of stand productivity caused by heavy thinning (? 1000 stems/ha) can be mitigated by enhanced tree growth within 15 years of thinning and sooner if considering contributions of ingress. ? Stand-level productivity of crop trees can be enhanced by nearly 20% with repeated fertilizer applications and this enhanced growth rate appears to last for several years after fertilization has stopped. ? Negative impacts to the plant community caused by repeated fertilization (decreased herb and shrub diversity) are short-lived.  ? Long-term impacts to the plant community caused by repeated fertilization include enhanced structural diversity of the herb, shrub, and total plant layers. ? Repeated fertilization increased crude protein levels of pinegrass by 18%, even five years after the most recent application. ? Range quality, yield, and longevity can be significantly enhanced by a combination of fertilization and aggressive thinning, particularly to 500 stems/ha.  ? Habitat use by cattle was not associated with any negative impacts on habitat use by mule deer within these managed stands.     182 These contributions have been made because of the unique experimental design.  This is the only operational scale and long-term study to investigate the effects of: ? a range of PCT intensities and repeated fertilization of lodgepole pine stands, ? PCT to the extremely low target density of 250 stems/ha, ? cattle exclusion in forested range. 7.4 General conclusions and management implications The benefits of intensive forest management have been clearly demonstrated at an operational scale in many parts of the world (e.g., Albaugh et al. 2004; Jokela et al. 2004; Fox et al. 2007; Nabuurs et al. 2007; Bergh et al. 2008), as have the benefits of integrating forest and range management (e.g., Knowles et al. 1973; Msika and Etienne 1989; Sibbald et al. 1989; Papanastasis et al. 1995).  However, these benefits have generally not been realized in Canada where conventional forest management remains extensive rather than intensive, and the management of trees and forage separate rather than integrated.  However, the need for enhanced productivity from both the forest (Sedjo 1999; Raunikar et al. 2010) and ranching industries (Watkinson and Ormerod 2001; Steinfeld et al. 2006) has garnered interest in both incremental silviculture (Lautenschlager 2000; Park and Wilson 2007) and silvopastoralism (Thevathasan et al. 2012).  Unfortunately, studies of such management strategies at northern latitudes have generally been conducted at small scales (< 0.1 ha) and/or over short terms (1-5 years), thereby limiting their scope of inference.  By investigating the effects of a wide range of PCT intensities (target densities of 2000, 1000, 500, and 250 stems/ha) and repeated fertilization treatments (five applications over 10 years) on young lodgepole pine stands, and monitoring responses to these treatments over a period of 15 years, my study begins to address this information gap. While increased individual tree growth possible with PCT is well documented (e.g., Johnstone 1985; Cole and Koch 1996), the effects on stand-level crop-tree productivity may be reduced as there are fewer trees contributing to total productivity.  Initial stand-level losses are expected, but, with time, such losses should be mitigated by the enhanced growth of the residual stems.  In my study, 10-year stand-level increments of lodgepole pine volume were significantly less within heavily compared to lightly thinned stands (Lindgren et al. 2007).  However, no differences in stand-level productivity were reported for 15-year increments.  These results    183 indicated that the enhanced growth of heavily thinned stands had compensated for the initial loss of stand productivity within 15 years post-PCT.  Initial stand-level losses of fiber production would be further mitigated if accounting for the contribution of non-crop trees (ingress), which significantly increased in density within heavily thinned and unfertilized stands.  The relative increase in stand productivity within heavily compared to lightly thinned stands occurred despite the pruning treatments which were applied to the former.  This was interesting as the significantly reduced crown size of pruned trees would have reduced the photosynthetic productivity of these trees, at least initially.  This suggested that pruning to promote clear wood and minimize the negative effects on stem form and wood quality properties associated with intensive thinning and fertilization (e.g., increased taper, decreased clear wood and wood density) could be carried out without any detectable negative impact on stand productivity within 10 years of pruning.  Fertilization treatments resulted in an 18% increase in 10- (Lindgren et al. 2007) and 15-year increments of stand volume, indicating the fertilizer effect found for the period when the five fertilizer treatments were applied was sustained five years after the final fertilization treatment.    PCT, even to very low densities (e.g., 250 stems/ha), had very little effect on long-term plant community attributes and significant effects were limited to enhanced structural diversity of the tree layer and a short-term (five years post-PCT) decrease in total structural diversity.  The enhanced tree layer structural diversity following heavy thinning became more pronounced with time as ingress of trees and understory herbs and shrubs were promoted by open canopy conditions.  The short-term decrease in total structural diversity likely resulted from the initial simplification of the tree layer by removing smaller trees.  By delaying canopy closure, heavy thinning appeared to prolong the enhanced structural diversity associated with younger seral stages compared with the closed canopy conditions of the more densely stocked stands.  Fertilization treatments had many more significant effects on plant community attributes; however, many of these effects had waned by the end of the 15-year study.  Significant short-term impacts of fertilization treatments included enhanced herb volume and decreased herb and shrub species diversity.  Structural richness of herbs, and structural diversity of herbs and shrubs, all remained significantly enhanced by fertilization throughout the study.  A decrease in tree layer volume and structural richness with fertilization resulted from suppression of ingress; however, total structural diversity was enhanced.  Negative impacts to biodiversity and habitat    184 structure by intensive management were not supported by this study.  In particular, no consistent treatment effects were detected for species richness, and negative impacts to other measures of species diversity were short-lived.  In addition, longer term impacts to habitat structure, such as increased total structural diversity, could be interpreted as beneficial for habitat.   Forage yield increased with decreasing canopy closure to a point; however, a threshold was observed where further reduced canopy closure had a negative effect on forage yield.  This threshold may represent the condition where resources other than light (e.g., nutrients and moisture) and/or where the loss of the beneficial influences of the tree layer (i.e., buffering effects) begin to limit productivity.  However, the benefits of delayed canopy closure, and therefore prolonged forage productivity, would still be realized following extreme PCT intensities.  Fertilization significantly enhanced forage yield; however, canopy closure occurred more rapidly, eventually shading out the understory forage.  Therefore, fertilization, for the objective of enhanced forage yield, should be accompanied by aggressive thinning.  Results also indicated significantly enhanced levels of crude protein in pinegrass (18% increase) following repeated fertilization, even five years after the most recent application. The enhanced forage opportunities provided by PCT and fertilization likely explained the significant increase in habitat use by cattle reported within fertilized and heavily thinned stands five to 10 years post-PCT.  While cattle appeared to benefit from PCT, there may be beneficial effects from some tree cover as habitat use was consistently greater within the 500 than the 250 stems/ha stands.  No negative impacts of cattle disturbance on habitat use by mule deer during either summer or winter were found over five years of study.  On the contrary, positive relationships between these two herbivores were observed.  It is unclear whether these positive relationships were the result of positive interactions (e.g., commensalism), similar habitat selection criteria for both herbivores (i.e., both preferring forage-rich habitats), or a combination of the two.  The impacts of cattle disturbance on the plant community were more complicated than those on mule deer and 10 years of cattle exclusion revealed several interesting results.  In fertilized stands, impact of cattle grazing was beneficial, particularly for the herb layer, as grazing decreased volume and slowed the rate of competitive exclusion, resulting in increased species richness and diversity.  These beneficial impacts were not initially evident, as the combined effects of fertilization and cattle grazing took as long as five years to produce    185 significant differences between exclosure and open range conditions.  The data suggest that stands with mean volumes of herbs greater than ca. 50 to 60 m3/0.01 ha appeared to trigger selection by cattle, which then resulted in the significant reduction in herb volume observed after 1996.  Herb volume in unfertilized stands did not attain these volumes and had similar volume inside and outside of exclosures.  The impacts were reversed in unfertilized stands, particularly for the shrub layer, where species richness was greater within exclosures.   Results of my study indicated the clear potential for long-term benefits associated with PCT and fertilization in terms of silvicultural productivity of lodgepole pine stands.  In addition, these benefits were achieved without any long-term negative impacts to the plant community diversity and with some benefits to habitat structure.  Fertilization, coupled with aggressive thinning, should provide for extended periods of suitable range condition, increased stocking densities of livestock, and an extended grazing season.  While these treatments increased habitat use by cattle, the increase was not associated with any negative impacts to mule deer habitat use, even during the critical overwinter period.  Within unfertilized stands, negative impacts of grazing to the plant community were limited to decreased shrub species richness.  This negative effect of grazing was not observed within fertilized stands, where grazing enhanced herb species richness and diversity.   Negative impacts to plant community diversity, habitat structure, and mule deer, which may lead to concerns regarding PCT and fertilization treatments, were not supported by my study.  My results indicated than these intensive management treatments can be applied without degrading environmental attributes and, in fact, can bring ecological benefits.  A conservative approach for maintaining natural levels of biodiversity is to maintain a broad spectrum of environmental conditions.  Therefore, the sustainability of regional-scale management strategies would be further enhanced by incorporating a variety of management intensities, ranging from intensive to extensive, as well as no management at all.       186 References Adler, P.B., Seabloom, E.W., Borer, E.T., et al.  2011.  Productivity is a poor predictor of plant species richness.  Science 23: 1750-1753. Agee, J.K.  1993.  Fire ecology of Pacific Northwest forests. Island Press. Washington, DC. USA. Albaugh, T.J., Allen, H.L., Dougherty, P.M., and Johnsen, K.H.  2004.  Long term growth responses of loblolly pine to optimal nutrient and water resource availability. Forest Ecol. Manage. 192: 3?19. Alldredge, M.W., Peek, J.M., and Wall, W.A.  2001.  Alterations of shrub communities in relation to herbivory in northern Idaho.  Northwest Sci. 75: 137-144. Allen, H.L., Dougherty, P.M., Campbell, R.G.  1990.  Manipulation of water and nutrients ? practice and opportunity in southern US pine forests. Forest Ecol. Manage. 30: 437-453. Ammer, C. 1996. Impact of ungulates on structure and dynamics of natural regeneration of mixed mountain forests in the Bavarian Alps. Forest Ecol. Manage. 88: 43?53. Amponsah, I.G., Lieffers, V.J., Comeau, P.G., and Brockley, R.P.  2004.  Growth response and sapwood hydraulic properties of young lodgepole pine following repeated fertilization. Tree Phys. 24: 1099-1108. Anderson, C.C. and Scherzinger, R.J.  1975.  Improving quality of winter forage for elk by cattle grazing.  J. Range Manage. 28: 120-125. Armleder, H.M., Waterhouse, M.J., Keisher, D.G., and Dawson, R.J.  1994.  Winter habitat use by mule deer in the central interior of British Columbia. Can. J. Zool. 72: 1721?1725. Austin, D.D. and Urness, P.J.  1986.  Effects of cattle grazing on mule deer diet and area selection.  J. Range. Manage. 39: 18-21. Bai, Y.F., Wu, J.G., Pan, Q.M., Huang, J.H., Wang, Q.B., Li, F.S., Buyantuyev, A., and Han, X.G.  2007.  Positive linear relationship between productivity and diversity: evidence from the Eurasian Steppe, J. Appl. Ecol. 44: 1023-1034.    187 Ballard, L.A. and Long, J.N.  1988.  Influence of stand density on log quality of lodgepole pine. Can. J. For. Res. 18: 911-916.  Basile, J.V. and Jensen, C.E.  1971.  Grazing potential on lodgepole pine clearcuts in Montana. Res. Pap. INT-98. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station. 11 pp. Battles, J.J., Shliski, A.J., Barrett, R.H., Heald, R.C., and Allen-Diaz, B.H.  2001.  The effect of forest management on plant species diversity in a Sierran conifer forest.  Forest Ecol. Manage. 146: 211-222 Bedunah, D., Pfingsten, W., Kennett, G. and Willard, E.  1988.  Influence of stand canopy density to forage production, pp. 99-107. In: Proceedings-Future Forests of the Mountain West: A Stand Culture Symposium. Gen. Tech, Rep. INT-243. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 402 pp. Berch, S.M., Brockley, R.P., Battigelli, J.P., Hagerman, S., and Holl, B.  2006.  Impacts of repeated fertilization on components of the soil biota under a young lodgepole pine stand in the interior of British Columbia.  Can. J. For. Res. 36: 1415-1426. Bergh, J., Nilsson, U., Grip, H., Hedwall P.O., Lundmark, T.  2008.  Effects of frequency of fertilization on production, foliar chemistry, and nutrient leaching in young Norway spruce stands in Sweden. Silva. Fenn. 42: 721-733. Birdsey, R.A.  1992.  Carbon storage and accumulation in United States forest ecosystems.  USDA Forest Service, General Technical Report WO-59. Blevins, D.P., Prescott, C.E., Allen, H.L., Newsome, T.A.  2005.  The effects of nutrition and density on growth, foliage biomass, and growth efficiency of high-density fire-origin lodgepole pine in central British Columbia.  Can. J. For. Res. 35: 2851-2859. Bock, C.E., Bock, J.H., and Smith, H.M.  1993.   Proposal for a system of federal livestock exclosures on public rangelands in the western United States.  Cons. Biol. 7:731-733. Bokdam J., and Gleichman, J.M.  2000.  Effects of grazing by free-ranging cattle on vegetation dynamics in a continental north-west European heathland.  J. Appl. Ecol. 37:415-431.    188 Brockley, R.P.  1996  Lodgepole pine nutrition and fertilization: a summary of British Columbia Ministry of Forests research results. FRDA Report 266.  Forest Resource Development Agreement, Canadian Forestry Service British Columbia Ministry of Forests, Victoria, Canada. Brockley, R.P.  2005.  Effects of post-thinning density and repeated fertilization on the growth and development of young lodgepole pine. Can. J. For. Res. 35: 1952-1964. Brockley, R.P.  2007.  Effects of 12 years of repeated fertilization on the foliar nutrition and growth of young lodgepole pine in the central interior of British Columbia.  Can. J. For. Res. 37: 2115-2129. Brockley, R.P.  2001.  Foliar sampling guidelines and nutrient interpretative criteria for lodgepole pine. B.C. Ministry of Forests, Victoria. Extension Note 52. Brockley, R.P., Trowbridge, R.L., Ballard, T.M. and Macadam, A.M.  1992.  Nutrient management in interior forest types. In: Proceedings  Forest Fertilization: Sustaining and Improving Nutrition and Growth of Western Forests, pp. 43-64. University of Washington Institute of Forest Resources Contribution 73, Seattle, WA. Brooks, D.J.  1997.  The outlook for demand and supply of wood: implications for policy and sustainable management.  Commonwealth Forestry Review 76: 31-36. Brown, J.R., Hemck, J., and Price, D.  1999.  Managing low-output agroecosystems sustainably:  the importance of ecological thresholds. Can. J. For. Res. 29: 1112-1119. Buergler, A.L., Fike, J.H., Burger, J.A., Feldhake, C.M., McKenna, J.R., and Teutsch, C.D.  2005.  Botanical composition and forage production in an emulated silovpasture.  Agron. J. 97:1141?1147. Buergler, A.L., Fike, J.H., Burger, J.A., Feldhake, C.M., McKenna, J.R., and Teutsch, C.D.  2006.  Forage nutritive value in an emulated silvopasture.  Agron. J. 98:1265?1273. Burner, D.M. and Brauer, D.K.  2003.  Herbage response to spacing of loblolly pine trees in a minimal management silvopasture in the southeastern USA.  Agrofor. Syst. 57:69?77.    189 Burton, P.J., Balisky, A.C., Coward, L.P., Cumming, S.G., and Kneeshaw. D.D.  1992.  The value of managing for biodiversity.  For. Chron. 68: 225-237. Campbell, J., Alberti, G., Martin, J., and Law, B.E.  2009.  Carbon dynamics of a ponderosa pine plantation following a thinning treatment in the northern Sierra Nevada.  Forest Ecol. Manage. 257: 453-463. Carey, A.B., Lippke, B.R., and Sessions, J.  1999.  Intentional systems management: managing forests for biodiversity.  J. Sust. For. 9: 83-125. Chauchard, S., Pille, G., and Carcaillet, C. 2006. Large herbivores control the invasive potential of non-native Austrian black pine in a mixed deciduous Mediterranean forest. Can. J. For. Res. 36: 1047-1053. Clark, P.E., Krueger, W.C., Bryant, L.D., and Thomas, D.R.  2000.  Livestock grazing effects on forage quality of elk winter range.  J. Range Manage. 53:97-105. Clason T.R. and Sharrow, S.H.  2000.  Silvopastoral practices. In: (H.E. Garrett, W.J. Rietveld, R.F. Fisher, ed?s) North American Agroforestry: An Integrated Science and Practice. American Society of Agronomy: Madison, WI. pp 119-147 Clason, A.J., Lindgren, P.M.F., and Sullivan, T.P.  2008.  Comparison of potential non-timber forest products in intensively managed young stands and mature/old-growth forests in south-central British Columbia.  Forest Ecol. Manage. 256: 1897-1909. Coe, P.K., Johnson, B.K., Kern, J.W., Findholt, S.L., Kie, J.G., and Wisdom, M.J.  2001.  Response of elk and mule deer to cattle in summer.  J. Range Manage. 54:A51-A76. Cole, D,M. and Koch, P.  1996.  Managing lodgepole pine to yield merchantable thinning products and attain sawtimber rotations.  Research Paper INT-RP-482.  USDA Forest Service Intermountain Research Station. Cole, E.C., Hanley, T.A., and Newton, M.  2010.  Influence of precommercial thinning on understory vegetation of young-growth Sitka spruce forests in southeastern Alaska.  Can. J. For. Res. 40: 619-628.     190 Courtois, D.R., Perryman, B.L., and Hussein, H.S.  2004.  Vegetation change after 65 years of grazing and grazing exclusion.  J. Range Manage. 57: 574-582. Damiran, D., DelCurto, T., Findholt, S. L., Pulsipher, G.D., and Johnson, B.K.  2003.  Influence of previous cattle and elk grazing on the subsequent quality and quantity of diets for cattle, deer and elk grazing late-summer mixed-conifer rangelands. In: Proceedings of the West. Sec. of Amer. Soc. Anim. Sci. 54:320-324. Darambazar, E., DelCurto, T., Ackerman, C., Pulsipher, G., and Damiran, D.  2003.  Changes in forage quantity and quality with continued cattle grazing in a mountain riparian pasture. In: Proceedings of the West. Sec. of Amer. Soc. Anim. Sci. 54:324-328. Dasci, M. and Comakli, B.  2011.  Effects of fertilization on forage yield and quality in range sites with different topographic  structure. Turkish Journal of Field Crops 16:15-22. D?az, S., Noy-Mier, I., and Cabido, M.  2001.  Can grazing response of herbaceous plants be predicted from simple vegetative traits?  J. Appl. Ecol. 38: 497-508. Edge, W.D. and Marcum, C.L.  1989.  Determining elk distribution with pellet-group and telemetry techniques.  J. Wildl. Manage. 53:621-624. Elliot, K.J., Boring, L.R., Swank, W.T., and Haines, B.R.  1997.  Successional changes in plant species diversity and composition after clearcutting a Southern Appalachian watershed.  Forest Ecol. Manage. 92: 67-85. Englund, S.R., O?Brien, J.J., and Clark, D.B.  2000.  Evaluation of digital and film hemispherical photography and spherical densiometry for measuring forest light environments.  Can. J. For. Res. 30:1999-2005 Environment Canada.  2013.  Government of Canada: National climate data and information (archive); URL- http://climate.weatheroffice.gc.ca/advanceSearch/searchHistoricData_e.html (accessed June 26, 2013) Farnden, C. and Herring, L.  2002.  Severely repressed lodgepole pine responds to thinning and fertilization: 19-year results. For. Chron. 78:404-414.    191 Farnum, P., Timmis, R., and Kulp, J.L.  1983.  The biotechnology of forest yield.  Science  219:694-702. Findholt, S.L., Damiran, D., Johnson, B.K., DelCurto, T., and Kie, J.G.  2005.  Diet composition, dry matter intake, and diet overlap of mule deer, elk, and cattle. March 20, 2004. Transactions of the North American Wildlife and Natural Resources Conference 69. Fleischner, T.L.  1994.  Ecological costs of livestock grazing in western North America.  Cons. Biol. 8:629-644. Fowler, J., Cohen, L., Jarvis, P.  1998.  Practical Statistics for Field Biology, second ed. John Wiley and Sons. 259 pp. Fox, T.R., Jokela, E.A., and Allen, H.L.  2007.  The development of pine plantation silviculture in the southern United States.  J. For. 105: 337-347. Freyman, S. and van Ryswyk, A.L.  1969.  Effect of fertilizer on pinegrass in southern British Columbia.  J. Range Manage. 6:390-395. Fridley, J.D., Grime, J.P., Huston, M.A., Pierce, S., Smart, S.M., Thompson, K., B?rger, L., Brooker, R.W., Cerabolini, B.E.L., Gross, N., Liancourt, P., Michalet, R., and Bagousse-Pinguet, Y.L.  2012.  Comment on ?Productivity is a poor predictor of plant species richness?.  Science 335: 1441. Fuhlendorf, S.D., and Engle, D.M.  2001.  Restoring heterogeneity on rangelands: ecosystem management based on evolutionary grazing patterns.  BioScience. 51: 625-632. Gaines, T.P. and G.A. Mitchell.  1979.  Boron determination in plant tissues by the azomethine H method.  Commun. Soil Sci. Plant anal. 10: 1099-1108. Galbraith, W.A, and Anderson, E.W.  1971.  Grazing History of the Northwest.  J. Range. Manage. 24:6-12. Germano, D.J., Rathbun, G.B., and Saslaw, L.R.  2001.  Managing exotic grasses and conserving declining species.  Wildl. Soc. Bull. 29: 551-559.    192 Gibbs, M.C., Jenks, J.A., Deperno, C.S., Sowell, B.F., and Jenkens, K.J.  2004.  Cervid forage utilization in noncommercially thinned ponderosa pine forests.  J. Range Manage. 57:435-441. Gilliam, F.S. and Roberts, M.R.  1995.  Impact of forest management on plant diversity.  Ecol. Appl. 5: 911-912. Gillman, L.N. and Wright, S.D.  2006.  The influence of productivity on the species richness of plants: a critical assessment.  Ecology 87: 1234-1243. Grace, J.B., Adler, P.B., Seabloom, E.W., et al.  2012.  Response to comments on ?Productivity is a poor predictor of plant species richness?.  Science 355: 1441. Groot, A., Brown, K.M., Morrison, I.K., and Barker, J.E.  1984.  A 10-year tree and stand response of jack pine to urea fertilization and low thinning.  Can. J. For. Res. 14: 44?50. Guo, Q.  2007.  The diversity-biomass-productivity relationships in grassland management and restoration.  Basic and Applied Ecology 8: 199-208. Hall, F.C.  1988.  Opportunities for enhancing livestock forage on forestland. In: Schmidt, Wyman C., compiler.  Preceedings ? future forests of the Mountain West: a stand culture symposium; 1986 September 29 ? October 3; Missoula, MT. Gen. Tech. Rep. INT-243. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station: 49-54. Halpern, C.B.  1989.  Early successional patterns of forest species: interactions of life history traits and disturbance.  Ecology 70: 704-720. Hanley, T.A.  1996.  Potential role of deer (Cervidae) as ecological indicators of forest management.  Forest Ecol. Manage. 88:199-204. Hansen, A.J., Spies, T.A., Swanson, F.J., and Ohmann, J.L.  1991.  Conserving biodiversity in managed forests, lessons from natural forests.  BioScience 41: 382-392. Harrison, S., Inouye, B.D., and Safford, H.D.  2003.  Ecological heterogeneity in the effects of grazing and fire on grassland diversity.  Conserv. Biol. 17: 837-845.    193 Hart, R.H., Hughes, R.H., Lewis, C.E., and Monson, W.G.  1970.  Effect of nitrogen and shading on yield and quality of grasses grown under young slash pines.  Agron. J. 62:285-287. Hartley, M.J.  2002.  Rationale and methods for conserving biodiversity in plantation forests. Forest Ecol. Manage. 155: 81-95. Hayes, G.F., and Holl, K.D.  2003.  Cattle grazing impacts on annual forbs and vegetation composition of mesic grasslands in California.  Conserv. Biol. 17: 1694-1702. He, F. and Barclay, H.J.  2000.  Long-term response of understory plant species to thinning and fertilization in a Douglas-fir plantation on southern Vancouver Island, British Columbia.  Can. J. For. Res. 30: 566-572. Hitchcock, C. L., and Cronquist A.  1973.  Flora of the Pacific Northwest.  University of Washington Press, Seattle, Washington.  730 pages. Hoen, H.F. and Solberg, B.  1994.  Potential and economic efficiency of carbon sequestration in forest biomass through management.  For. Sci. 40: 429-451. Hofmann, R. R.  1988.  Anatomy of the gastro-intestinal tract. In: The ruminant animal digestive physiology and nutrition, ed. D.C. Church, 14-43. New Jersey: Prentice Hall, Englewood Cliffs. Homyack, J.A., Harrisona, D.J., and Krohn, W.B.  2005.  Long-term effects of precommercial  thinning on small mammals in northern Maine.  Forest Ecol. Manage. 205: 43?57. Huettl, R.F. and Zoettl, H.W.  1992.  Forest fertilization: its potential to increase the CO2 storage capacity and to alleviate the decline of the global forests.  Water, Air, and Soil Pollution 64: 229-249. Humphrey, J.W. and Patterson, G.S.  2000.  Effects of late summer cattle grazing on the diversity of riparian pasture vegetation in an upland conifer forest.  J. Appl. Ecol. 37:986-996. Hunter, M.L., Jr., 1999.  Maintaining biodiversity in forest ecosystems.  Cambridge Univ. Press, Cambridge, UK.    194 Hunter, M.L., Jr., Schmiegelow, F.K.A.  2011.  Wildlife, Forests, and Forestry. Principles of Managing Forests for Biological Diversity. Second edition. Prentice Hall, New York, NY. 259 pp. Hurlbert, S.H.  1984.  Pseudoreplication and the design of ecological field experiments.  Ecol. Monogr. 54: 187-211. Husak, A.L., and Grado, S.C.  2002.  Monetary benefits in a southern silvopasture system.  South. J. Appl. For. 26:159-164. Jandl, R., Lindner, M., Vesterdal, L, Bauwens, B., Baritz, R., Hagedorn, F., Johnson, D.W., Minkkinen, K., and Byrne, K.A.  2007.  How strongly can forest management influence soil carbon sequestration?  Geoderma 137: 253-268. Jim?nez, E., Vega, J.A., Fern?ndez, C, and Fonturbel, T.  2011.  Is pre-commercial thinning compatible with carbon sequestration? A case study in a maritime pine stand in northwestern Spain.  Forestry 84: 149-157. Jim?nez, J.A., P?rez-Gorostiaga, V.P., Cui?as, P., Fonturbel, T., Fern?ndez, C., Madrigal, J., Hernando, C., and Guijarro, M.  2008.  Effects of pre-commercial thinning on transpiration in young post-fire maritime pine stands.  Forestry 81: 543-557. Johnstone, W.D.  1985.  Thinning lodgepole pine.  Lodgepole pine: the species and its management (eds D. M. Baumgartner, R. G. Krebill, J. T. Arnott and G. F. Weetman), pp. 253-262.  Washington State University Cooperative extension, Spokane, Washington, USA, and Vancouver, British Columbia, Canada. Jokela, E.J., Dougherty, P.M., and Martin, T.A.  2004.  Long-term production dynamics of loblolly pine stands in the southern United States: a synthesis of seven long-term experiments.  Forest Ecol. Manage. 192:117-130. Kalmbacher, R., and Martin, F.  1996.  Shifts in botanical composition of flatwoods range following fertilization.  J. Range Manage. 49: 530-534. Kayes, L.J. and Tinker, D.B.  2012.  Forest structure and regeneration following a mountain pine beetle epidemic in southeastern Wyoming.  Forest Ecol. Manage. 263: 57?66.    195 Kellner, O., Redbo-Torstensson, P., 1995. Effects of elevated nitrogen deposition on the field-layer vegetation in coniferous forests.  Ecol. Bull. 44: 227-237. Kephart, K.D. and Buxton, D.R.  1993.  Forage quality responses of C3 and C4 perennial grasses to shade.  Crop Sci. 33:831?837. Kering, M.K., Guretzky, J., Funderburg, E., and Mosali, J.  2011.  Effect of nitrogen fertilizer rate and harvest season on forage yield, quality, and macronutrient concentrations in midland Bermuda grass.  Comm. Soil Sci. Plant Anal. 42:1958?1971. Kie, J.E., Evans, C.J., Loft, E.R., and Menke, J.W.  1991.  Forage behavior by mule deer: the influence of cattle grazing.  J. Wildl. Manage. 55:665-674. Kishchuk, B.E., Weetman, G.F., Brockley, R.P. and Prescott, C.E.  2002.  14-year growth response of young lodgepole pine to repeated fertilization. Can. J. For. Res. 32: 153-160. Klinka, K., Koot, C., Kusbach, T., Macku, J., Rau, M., Trethewey, C., and Varga, P.  2004.  Site identification tools, field site maps and polygon summaries for the UBC Alex Fraser Research Forest. UBC Alex Fraser Research Forest, Williams Lake, BC. 70 pp. Knowles. R.L., Klomp, B.K., and Gillingham, A.  1973.  Trees and grass-an opportunity for the hill country farmer. Proc. of the Ruakura Farmer?s Conference, New Zealand For. Serv. reprint No. 705:108-121. Koch, P.  1996.  Lodgepole pine in North America.  For. Prod. Soc., Madison, WI, USA. Kozak, A.  1988.  A variable exponent taper equation. Can. J. For. Res. 18: 1363-1368. Kranabetter, J.M., Durall, D.M., and MacKenzie, W.H.  2009.  Diversity and species distribution of ectomycorrhizal fungi along productivity gradients of a southern boreal forest.  Mycorrhiza 19: 99-111. Krebs, C.J., Dale, M.R.T., Nams, V.O., Sinclair, A.R.E., O?Donoghue, M., 2001. Shrubs. In: Krebs, C.J., Boutin, S., Boonstra, R. (Eds.), Ecosystem Dynamics of the Boreal Forest. Oxford University Press, Oxford, UK, pp. 92-115. Kruegar, W.C.  1981.  How a forest affects a forage crop.  Rangelands 3:70-71.    196 Krzic, M., Newman, R.F., and Broersma, K.  2003.  Plant species diversity and soil quality in harvested and grazed boreal aspen stands of northeastern British Columbia.  Forest Ecol. Manage. 182: 315-325. Kuiters, A.T., and Slim, P.A. 2002. Regeneration of mixed deciduous forest in a Dutch forest-heathland, following a reduction of ungulate densities. Biol. Conserv. 105: 65?74. Lautenschlager, R.A.  2000.  Can intensive silviculture contribute to sustainable forest management in northern ecosystems?  Forestry Chronicle 76: 283-295. Leckie, S.E., Prescott, C.E., and Grayston, S.J.  2004.  Forest floor microbial community response to tree species and fertilization of regenerating coniferous forests.  Can J. For. Res. 34: 1426-1435. Lemmon, P.E.  1956.  A spherical densiometer for estimating forest overstory density.  For. Sci. 2:314?320. Leroux, S.J., Kerr, J.T., 2013. Land development in and around protected areas in the wilderness frontier. Conserv. Biol. 27: 166-176. Lin, C.H., McGraw, M.L., George, M.F., and Garrett, H.E.  2001.  Nutritive quality and morphological development under partial shade of some forage species with agroforestry potential.  Agrofor. Syst. 53:269-281. Linder, S.  1987.  Responses to water and nutrients in coniferous systems. Ecol. Studies 61: 180-202. Linder, S.  1995.  Foliar analysis for detecting and correcting nutrient imbalances in Norway spruce. Ecol. Bull. 44: 178-190. Lindgren, P.M.F. and Sullivan, T.P.  2012.  Response of plant community abundance and diversity during 10 years of cattle exclusion within silvopasture systems.  Can. J. For. Res. 42: 451-462.    197 Lindgren, P.M.F., Ransome, D.B., Sullivan, D.S., and Sullivan, T.P.  2006.  Plant community attributes 12 to 14 years following pre-commercial thinning in a young lodgepole pine forest.  Can. J. For. Res. 36:48-61. Lindgren, P.M.F., Sullivan, T.P., Sullivan, D.S., Brockley R.P., and Winter, R.  2007.  Growth response of young lodgepole pine to thinning and repeated fertilization treatments: 10-year results.  Forestry 80: 587-611. Lindh, B.C. and Muir, P.S.  2004.  Understory vegetation in young Douglas-fir forests: does thinning help restore old-growth composition?  Forest Ecol. Manage. 192: 285-296. Lloyd, D.A., Angove, K., Hope, G.D., and Thompson, C.  1990.  A guide to site identification and interpretation for the Kamloops Forest Region.  Land Management Handbook 23.  BC Ministry of Forests, Research Branch.  Victoria, BC. Loft, E.R. and Kie, J.G.  1988.  Comparison of pellet-group and radio triangulation methods for assessing deer habitat use.  J. Wildl. Manage. 52:524-527. Loft, E.R., Menke, J.W., Kie, J.G., and Bertman, R.C.  1987.  Influence of cattle stocking rate on the structural profile of deer hiding cover.  J. Wildl. Manage. 51:655-664. Lotan J.E., and W.B. Critchfield.  1990.  Pinus contorta (Dougl. ex. Loud) Lodgepole pine.  Pp. 302-315 In: Silvics of North America: 1. Conifers, Burns, R.M., and B.H. Honkala (tech. coords.). Agric. Handbk. 654. USDA For. Serv., Washington, DC. MacArthur, R.H., and MacArthur, J.W.  1961.  On bird species diversity.  Ecology 42: 594-598. Magurran, A.E.  2004.  Measuring biological diversity. Blackwell, Oxford, United Kingdom. Malkonen, E. and Kukkola, M.  1991.  Effect of long-term fertilization on the biomass production and nutrient status of Scots pine stands. Fertilizer Research 27: 113-127. McConnel, B.R. and Smith, J.R.  1965.  Understorey response three years after thinning pine. J. Range Manage. 18:129-132. McConnell, B.R. and Smith, J.G.  1970.  Response of understorey vegetation to ponderosa pine thinning in eastern Washington.  J. Range Manage. 23:208-212.    198 McDonald, G.T. and Lane, M.B.  2004.  Converging global indicators for sustainable forest management.  Forest Policy Econ. 6: 63-70. McDonald, P.M. and Fiddler, G.O.  1993.  Feasibility of alternatives to herbicides in young conifer plantations in California.  Can. J. For. Res. 23:2015-2022. McIntosh, R.  1982.  Effect of different forms and rates of nitrogen fertiliser on the growth of lodgepole pine.  Forestry 55: 61?68. McIntyre, S., Heard, K.M., and Martin, T.G.  2003.  The relative importance of cattle grazing in subtropical grasslands: does it reduce or enhance plant biodiversity?  J. Appl. Ecol. 40: 445-457. McLean, A.  1967.  Beef production on lodgepole pine-pinegrass range in southern British Columbia.  J. Range Manage. 20:214-216. McLean, A.  1983.  Producing forage for livestock on forest ranges.  In: Proc. International Hill Lands Symposium. Foothills for food and forests, Corvallis, OR.  pp. 175-183. McLean, A., and Tisdale, E.W.  1972.  Recovery rate of depleted range sites under protection from grazing.  J. Range Manage. 25: 178-184. McLean, A., Freyman, S., Miltimore, J.E., and Bowden, D.M.  1969.  Evaluation of pinegrass as range forage. Can. J. Plant Sci. 49:351-359. McNaughton, S.J.  1993.  Grasses and grazers, science and management.  Ecol. Appl. 3:17-20. Meidinger, D. and J. Pojar.  1991.  Ecosystems of British Columbia.  Research Branch, Ministry of Forests, Victoria, B.C. Special Report Series No. 6. Millar, C.I., Stephenson, N.L., and Stephens, S.L.  2007.  Climate change and forests of the future: managing in the face of uncertainty.  Ecol. Appl. 17: 2145-2151. Mittelbach, G.G., Steiner, C.F., Scheiner, S.M., Gross, K.L., Reynolds, H.L., Waide, R.B., Willig, M.R., Dodson, S.I., and Gough, L.  2001.  What is the observed relationship between species richness and productivity?  Ecology 82: 2381?2396.    199 Moore, M.M. and Deiter, D.A.  1992.  Stand density index as a predictor of forage production in northern Arizona pine forests.  J. Range Manage. 45:267-271. Moore, S.E., Allen, H.L.  1999.  Plantation forestry. In: Hunter, M.L., Jr. (Ed.), Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press, New York, NY, pp. 400-433. Moser, B.W., and Witmer, G.W.  2000.  The effects of elk and cattle foraging on the vegetation, birds, and small mammals of the Bridge Creek Wildlife Area, Oregon.  Intern. Biodeter. Biodegr. 45: 151-157. Moya, D., De las Heras, J., L?pez-Serrano, F.R., Condes, S. and Alberdi, I.  2009.  Structural patterns and biodiversity in burned and managed Aleppo pine stands.  Plant Ecol. 200: 217?228. Msika, B. and Etienne, M.  1989.  Modification des facteurs biotiqucs par la presence d?arbres en region mediterraneenne Francaisc: Effets sur la production herbagere. XVI Congres International des Herbages, 4- 11 October, Nice, France, Association Francaise pour la Production Fourragere, pp. 1619-1620. Nabuurs, G.J., Pussinen, A., van Brusselen, J., Schelhaas, M.J.  2007.  Future harvesting pressure on European forests. Eur. J. For. Res. 126: 391-400. Nakashizuka, T.  2007.  An interdisciplinary approach to sustainability and biodiversity of forest ecosystems: an introduction.  Ecol. Res. 22: 359?360. Nams, V.O., Folkard, N.F.G., Smith, J.N.M., 1993. Effects of nitrogen fertilization on several woody and non-woody boreal forest species. Can. J. Bot. 71: 93-97. Nave, L.E., Vance, E.D., Swanston, C.W., and Curtis, P.S.  2009.  Impacts of elevated N inputs on north temperate forest soil C storage, C/N, and net N-mineralization.  Geoderma 153: 231-240 Newman, R.F., Cameron, K.J., Wikeem, B.M., Youwe, P.L., Quinton, D.A., Pitt, M.D., and Powell, G.W.  1997.  Response of lodgepole pine seedlings to simulated cattle damage. B.C. Min. For. Res. Rep. 14, Victoria, BC.    200 Nohrstedt, H.O.  2001.  Response of coniferous forest ecosystems on mineral soils to nutrient additions: a review of Swedish experiences.  Scandinavian Journal of Forest Research 16: 555?573. Oba, G., Vetaas, O.R., and Stenseth, N.C.  2001.  Relationships between biomass and plant species richness in arid-zone grazing lands.  J. Appl. Ecol. 38: 836-845. Oelberg, K.  1956.  Factors affecting the nutritive value of range forage.  J. Range Manage. 9:220-225. Ogawa, Y., and Mitamura, T.  1984.  Studies on cessation of grazing on permanent pasture ecosystem: VI Effect of cessation of grazing on vegetational structures.  Grassland Science 30:74-78. Oliver, C.D.  1981.  Forest development in North America following major disturbances.  Forest Ecol. Manage. 3: 153-168. Oswald, B.P., Ott, B., Williams, H.M., Farrish, K.W., and Van Kley, J.E.  2009.  Understory vegetative diversity of post-thinned pine plantations treated with fertilizer, fire and herbicide in east Texas.  International Journal of Forestry Research 2009: 6 pages Pan, X., Liu, F., and Zhang, M. 2012.  Comment on ?Productivity is a poor predictor of plant species richness?.  Science 335: 1441. Papadopol, C.S. 2000.  Impacts of climate warming on forests in Ontario: options for adaptation and mitigation.  For. Chron. 76: 139-149. Papanastasis, V, Z. Koukoura, D. Alifragis, and Makedos, I.  1995.  Effects of thinning, fertilisation and sheep grazing on the understory vegetation of Pinus pinaster plantations.  Forest Ecol. Manage. 77:181-189. Park, A. and Wilson, E.R.  2007.  Beautiful plantations: con intensive silviculture help Canada to fulfill ecological and timber production objectives?  For. Chron. 83: 825-839. Parkinson, J.A. and S.E. Allen.  1975.  A wet oxidation procedure for the determination of nitrogen and mineral nutrients in biological material.  Comm. Soil Sci. Plant Anal. 6: 1-11.    201 Pase, C.P.  1958.  Herbage production and composition under immature ponderosa pine stands in the Black Hills. J. Range Manage. 11: 238-243. Paul, L.R., Chapman, B.K., and Chanway, C.P.  2007.  Nitrogen fixation associated with Suillus tomentosus tuberculate ectomycorrhizae on Pinus contorta var. latifolia.  Annals of Botany 99: 1101-1109. Peek, J.M., Korol, J.J., Gay, D., and Hershey, T.  2001.  Overstory-understory biomass changes over a 35-year period in southcentral Oregon.  Forest Ecol. Manage. 150:267-277. Peitz, D.G., Shelton, M.G., and Tappe, P.A.  2001.  Forage production after thinning a natural loblolly pine-hardwood stand to different basal areas.  Wild. Soc. Bull. 29: 697-705. Pielou, E.C.  1966.  The measurement of diversity in different types of biological collections.  J. Theoret. Biol. 13: 131-144. Power, H., LeMay, V., Berninger, F. Sattler,D., and Kneeshawe, D.  2012.  Differences in crown characteristics between black (Picea mariana) and white spruce (Picea glauca).  Can. J. For. Res. 42: 1733-1743.  Powers, R.F. and Reynolds, P.E.  1999.  Ten-year responses of ponderosa pine plantations to repeated vegetation and nutrient control along an environmental gradient. Can. J. For. Res. 29: 1027-1038. Prescott, C.E., Coward, L.E., Weetman, G.F., and Gessel, S.P.  1993.  Effects of repeated nitrogen fertilization on ericaceous shrub salal (Gaultheria shallon) in two coastal Douglas-fir forests.  Forest Ecol. Manage. 61: 45-60. Proulx, M, and Mazumder, A.  1998.  Reversal of grazing impact on plant species richness in nutrient-poor vs. nutrient-rich ecosystems.  Ecology 79: 2581-2592. Pyk?l?, J.  2000.  Mitigating human effects on European biodiversity through traditional animal husbandry.  Cons. Biol. 14:705-712. Quinton, D.A.  1984.  Cattle diets on seeded clearcut areas in central interior British Columbia.  J. Range Manage. 37:349-352.      202 Raison, R.J. and Myers, B.J.  1992.  The biology of forest growth experiment: linking water and nitrogen availability to the growth of Pinus radiata. Forest Ecol. Manage. 52: 279-308. Rajaniemi, T.K.  2002.  Why does fertilization reduce plant species diversity? Testing three competition-based hypotheses.  Journal of Ecology 90: 316-324. Raunikar, R., Buongiornoa, J., Turnerb, J.A., and Zhu, S.  2010.  Global outlook for wood and forests with the bioenergy demand implied by scenarios of the Intergovernmental Panel on Climate Change.  Forest Policy and Economics 12: 48-56. Rosenzweig, M.L. and Abramsky, Z.  1993.  How are diversity and productivity related? In: Species diversity in ecological communities: historical and geographical perspectives (eds. Ricklefs, R.E. and Schluter, D), pp. 52-65. University of Chicago Press, Chicago. Saarsalmi, A. and M?lk?nen, E.  2001.  Forest fertilization research in Finland: a literature review. Scandinavian Journal of Forest Research 16: 514?535. Sale, P.F.  1974.  Overlap in resource use and interspecific competition.  Oecol. 17:245-256. Saville DJ.  1990.  Multiple comparison procedures: the practical solution. Am. Stat. 44: 174-180. Scott, D., Welch, D., Thurlow, M., and Elston, D.A.  2000.  Regeneration of Pinus sylvestris in a natural pinewood in NE Scotland following reduction in grazing by Cervus elaphus. Forest Ecol. Manage. 130: 199?211. Scotter, G.W.  1980.  Management of wild ungulate habitat in the western United States and Canada: a review.  J. Range Manage. 33: 16-27. Sedjo, R.A.  1999.  The potential of high-yield plantation forestry for meeting timber needs ? Recent performance, future potentials, and environmental implications.  New Forests 17: 339-359. Sibbald, A.R., Griffiths, J.H., and Elston, D.A.  1989.  The effects of the presence of widely spaced conifers on understory herbage production in the U.K.  In: P.G. Jarvis (ed.), Agroforestry: Principles and Practice. Elsevier, Amsterdam. pp. 71-77.    203 Soucy, M., Lussier, J., and Lavoie, L.  2012.  Long-term effects of thinning on growth and yield of an upland black spruce stand.  Can. J. For. Res. 42: 1669-1677. SPSS Institute Inc.  2007.  Statistical Programs for the Social Sciences.  SPSS Institute Inc., Chicago, IL. Steen, O.A. and Coupe, R.A.  1997.  A guide to site identification and interpretation for the Cariboo Forest Region ? 2 part book with 2 supplements.  Land Management Handbook 39.  BC Ministry of Forests, Southern Interior Forest Region.  Victoria, BC. Steinfeld, H, Gerber, P., Wassenaar, T., Castel, V., Rosales, M., and de Haan, C.  2006.  Livestock?s long shadow: environmental issues and options.  Food and Agriculture Organization of the United Nations (FAO), Livestock, Environment and Development LEAD Initiative, Rome, 390 pp. Stewart, K.M., Bowyer, R.T., Kie, J.G., Cimon, N.J., and Johnson, B.K.  2002.  Temporospatial distributions of elk, mule deer, and cattle: resource partitioning and competitive displacement.  J. Mammal. 83: 229-244. Stickney, P.F.  1985.  Data base for early post-fire succession on the sundance burn, northern Idaho.  USDA For. Serv. Intermountain Res. Stat. Gen. Tech. Rep. INT-189. Stohlgren, T.J., Schell, L.D., and Vanden Heuvel, B.  1999.  How grazing and soil quality affect native and exotic plant diversity in Rocky Mountain grasslands.  Ecol. Appl. 9: 45-64. Stover, S.L.  1985.  Silviculture and grazing in the New Forest: rival land uses over nine centuries.  J. Forest Hist. 29: 32-42. Strengbom, J. and Nordin, A.  2008.  Commercial forest fertilization causes long-term residual effects in ground vegetation of boreal forests.  Forest Ecol. Manage. 256: 2175?2181. Stuth, J.W. and Winward, A.H.  1977.  Livestock-deer relations in the Lodgepole Pine-Pumice region of central Oregon.  J. Range. Manage. 31: 110-116. Sullivan, T.P., Sullivan, D.S., and Lindgren, P.M.F.  2001.  Stand structure and small mammals in young lodgepole pine forest: 10-year results after thinning. Ecol. Appl. 11: 1151?1173.    204 Sullivan, T.P., Sullivan, D.S., and Lindgren, P.M.F.  2012.  Influence of repeated fertilization and cattle grazing on forest ecosystems: Abundance and diversity of forest-floor small mammals.  Forest Ecol. Manage. 277: 180-195. Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., and Ransome, D.B.  2006a  Influence of repeated fertilization on forest ecosystems: relative habitat use by mule deer and moose.  Can. J. For. Res. 36: 1395?1406. Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., and Ransome, D.B.  2006b  Influence of repeated fertilization on forest ecosystems: relative habitat use by snowshoe hares (Lepus americanus).  Can. J. For. Res. 36: 2080-2089. Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., and Ransome, D.B.  2009.  Stand structure and the abundance and diversity of plants and small mammals in natural and intensively managed forests. Forest Ecol. Manage. 258S: S127-S141. Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., Ransome, D.B.  2006c.  Long-term responses of ecosystem components to stand thinning in young lodgepole pine forest: III. Growth of crop trees and coniferous stand structure. Forest Ecol. Manage. 228: 69-81. Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., Ransome, D.B.  2013.  Stand structure and small mammals in intensively managed forests: Scale, time, and testing extremes. Forest Ecol. Manage. (In press). Sutton, W.R.J.  1999.  The need for planted forests and the example of radiata pine.  New Forests 17: 95-109. Sword Sayer, M.A., Goelz, J.C.G., Chambers, J.L., Tang, Z., Dean, T.J., Haywood, J.D., and Leduc, D.J.  2004.  Long-term production dynamics of loblolly pine stands in the southern United States.  Forest Ecol. Manage. 192:71-96. Taki, H., Inoue, T., Tanaka, H., Makihara, H., Sueyoshi, M., Isono, M., and Okabe, K.  2010  Responses of community structure, diversity, and abundance of understory plants and insect assemblages to thinning in plantations. Forest Ecol. Manage. 259: 607-613.    205 Tamm, C.O., Aronsson, A., Popovic, B., and Flower-Ellis, J.  1999.  Optimum nutrition and nitrogen saturation in Scots pine stands. Stud. For. Suec. 206. Thevathasan, N.V., Gordon, A.M., Bradley, R., Cogliastro, A., Folkard, P., Grant, R., Kort, J., Liggins, L., Njenga, F., Olivier, A., Pharo, C., Powell, G., Rivest, D., Schiks, T., Trotter, D., Van Rees, K., Whalen, J., and Zabek, L.  2012.  Agroforestry research and development in Canada: the way forward.  In: Agroforestry ? The Future of Global Land Use. Volume 9, Advances in Agroforestry (eds. Nair, P.K.R. and Garrity, D).  pp. 247-283.  Springer, Netherlands. Thomas, S.C., Halpern, C.B., Falk, D.A., Liguori, D.A., and Austin, K.A.  1999.  Plant diversity in managed forests: understory responses to thinning and fertilization.  Ecol. Appl. 9: 864-879. Thomas, S.C., Halpern, C.B., Falk, D.A., Liguori, D.A., and Austin, K.A.  1999.  Plant diversity in managed forests: understorey responses to thinning and fertilization.  Ecol. Appl. 9:864-879. Thysell, D.R., and Carey, A.B.  2001.  Manipulation of density of Pseudotsuga menziesii canopies: preliminary effects on understory vegetation.  Can. J. For. Res. 31: 1513-1525. Turkington, R., John, E., Dale, M.R.T.  2001.  Plant Dynamics: Herbs and grasses. In Krebs, C.J., Boutin, S., Boonstra, R. (Eds.), Ecosystem Dynamics of the Boreal Forest. Oxford University Press, Oxford, UK, pp. 70-91. Turkington, R., John, E., Krebs, C.J. et al.,  1998.  The effects of NPK fertilization for nine years on boreal forest vegetation in northwestern Canada.  J. Veg. Science 9: 333-346. Valinger, E., Elfving, B., and M?rling, T.  2000.  Twelve-year growth response of Scots pine to thinning and nitrogen fertilization. Forest Ecol. Manage. 134: 45-53.  Valone, T.J., Meyer, M., Brown, J.H., and Chews, R.M.  2001.  Timescale of perennial grass recovery in desertified arid grasslands following livestock removal.  Conserv. Biol. 16: 995-1002.    206 VanderSchaaf, C.L., Moore, J.A., and Kingerly, J.L.  2000.  The Effect of multi-nutrient fertilization on understory plant diversity.  Northwest Science 74: 316-324. Vavra, M.  2005.  Livestock grazing and wildlife: developing compatibilities. Rangeland Ecol. and Manage. 58:128-134. Von Holle, B., Delcourt, H.R., and Simberloff, D.  2003.  The importance of biological inertia in plant community resistance to invasion.  J. Veg. Science 14: 425-432. Vose, J.M. and Allen, H.L.  1988.  Leaf area, stem wood growth, and nutrition relationships in loblolly pine.  For. Sci. 34: 547-563. Vyse, A., Ferguson, C., Huggard, D.J., Roach, J., and Zimonick, B.  2009.  Regeneration beneath lodgepole pine dominated stands attacked or threatened by the mountain pine beetle in the south central Interior, British Columbia. Forest Ecol. Manage. 2585: S36?S43. Waide, R.B., Willig, M.R., Steiner, C.F., Mittelbach, G., Gough, L., Dodson, S.I., Juday, G.P., and Parmenter, R.  1999.  The relationship between productivity and species richness.  Annual Review of Ecology and Systematics 30: 257?300. Waldern, D.E.  1971.  A rapid micro-digestion procedure for neutral and acid detergent fiber. Can. J. Anim. Sci. 51:67-69. Wallace, M.C. and Krausman, P.R.  1987.  Elk, mule deer, and cattle habitats in central Arizona.  J. Range Manage. 40:80-83. Walmsley, M.E., Utzig, G., Vold, T., and Van Barneveld, J.  1980.  Describing ecosystems in the field.  Ministry of Environment and Ministry of Forests, Victoria, B.C. Land Management Report No. 7. Walton, A., Hughes, J., Eng, M., Fall, A., Shore, T., Riel, B., and Hall, P.  2009.  Provincial-level projection of the current mountain pine beetle outbreak.  BC Ministry of Forest and Range, Victoria, BC. Wang, S. and Chen, H.Y.H.  2010.  Diversity of northern plantations peaks at intermediate management intensity.  Forest Ecol. Manage. 259: 360-366.    207 Watanabe, F.S. and S.R. Olson.  1965.  Test of an ascorbic acid method for determining phosphorus in water and NaHCO3 extracts from soil.  Soil Sci. Soc. Am. Proc. 49: 677-678. Watkinson, A.R. and Ormerod, S.J.  2001.  Grasslands, grazing and biodiversity: editor?s introduction.  J. Appl. Ecol. 38:233-237. Weetman, G.F.  1988.  Nutrition and fertilization of lodgepole pine. In Proceedings ? Future Forests of the Mountain West: a stand culture symposium .USDA Forest Service . Gen. Tech. Rep. INT-243.Ogden, Utah. Weetman, G.F., Dallaire, L.C., and Fournier, R.  1995.  Long-term effects of repeated N fertilization and straw application in a jack pine forest. 1. Twenty-two-year growth response. Can. J. For. Res. 25: 1978?1983. West, N.E.  1993.  Biodiversity of rangelands.  J. Range Manage. 46:2-13. West, N.E., Provenza, F.D., Johnson, P.S., and Owens, M.K.  1984.  Vegetation changes after 13 years of live-stock grazing exclusion on sagebrush semidesert in west central Utah.  Ecol. Appl. 37: 262-264. Widenfalk, O. and Weslien, J.  2009.  Plant species richness in managed boreal forests ? effects of stand succession and thinning.  Forest Ecol. Manage. 257: 1386-1394. Wikeem, B.M., McLean, A., Bawtree, A., and Quinton, D.  1993a.  An overview of the forage resource and beef production on Crown land in British Columbia. Can. J. Anim. Sci. 73: 779-794. Wikeem, B.M., Newman, R.F., and van Ryswyk, A.L.  1993b.  Forage response to N, P, and S fertilization on clearcut lodgepole pine sites.  Ecol. Appl. 46: 262-270. Willms, W., Bailey, A.W., McLean, A., and Tucker, R.  1981.  The effects of fall defoliation on the utilization of bluebunch wheatgrass and its influence on the distribution of deer in the spring. J. Range Manage. 34:16-18.    208 Willms, W., Bailey, W., and McLean, A.  1980a.  Effect of burning or clipping Agropyron spicatum in the autumn on the spring foraging behaviour of mule deer and cattle.  J. Appl. Ecol. 17:69-84. Willms, W., McLean, A., Tucker, R., and Ritcey, R.  1980b.  Deer and cattle diets on summer range in British Columbia.  J. Range Manage. 33:55-59. Wilson, D.S., Anderson, P.D., and Puettmann, K.J.  2009.  Evaluating the consistency of understorey vegetation response to forest thinning through synthetic analysis of operational-scale experiments.  Forestry 82: 583-596. Yang, R.C.  1998.  Foliage and stand growth responses of semimature lodgepole pine to thinning and fertilization.  Can. J. For. Res. 28: 1794?1804. Yeo, J.J., Peek, J.M., Wittinger, W.T., and Kvale, C.T.  1993.  Influence of rest-rotation cattle grazing on mule deer and elk habitat use in east-central Idaho.  J. Range Manage. 46: 245-250. Zar, J.H.  1999.  Biostatistical Analysis.  Prentice-Hall, Inc., Englewood Cliffs, N. J. 663 pp. Zhang, S.Y., Chauret, G., Swift, D.E., and Duchesne, I.  2006.  Effects of precommercial thinning on tree growth and lumber quality in a jack pine stand in New Brunswick, Canada.  Can. J. For. Res. 36: 945-952. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0165656/manifest

Comment

Related Items