Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A wood inventory and dendroecological analysis of large woody debris in small streams in the foothills… Powell, Sonya Rebecca 2006

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

Item Metadata

Download

Media
831-ubc_2006-0621.pdf [ 6.66MB ]
Metadata
JSON: 831-1.0092740.json
JSON-LD: 831-1.0092740-ld.json
RDF/XML (Pretty): 831-1.0092740-rdf.xml
RDF/JSON: 831-1.0092740-rdf.json
Turtle: 831-1.0092740-turtle.txt
N-Triples: 831-1.0092740-rdf-ntriples.txt
Original Record: 831-1.0092740-source.json
Full Text
831-1.0092740-fulltext.txt
Citation
831-1.0092740.ris

Full Text

A WOOD INVENTORY AND DENDROECOLOGICAL ANALYSIS OF L A R G E WOODY DEBRIS IN SMALL STREAMS IN T H E FOOTHILLS MODEL FOREST, HINTON, ALBERTA  By Sonya Rebecca Powell B. Sc. University of Toronto, 2002  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE In THE F A C U L T Y OF G R A D U A T E STUDIES (Geography)  UNIVERSITY OF BRITISH C O L U M B I A OCTOBER 2006 © Sonya Rebecca Powell, 2006  ABSTRACT This research explores large woody debris (LWD) dynamics in five 100-year old lodgepole pine-dominated and five > 100 year old spruce-dominated riparian forests of the Foothills Model Forest in west-central Alberta. The objectives of my research were (1) quantify the abundance and type of in-stream L W D in mature pine and spruce, forests, (2) determine ages of in-stream L W D using tree-ring methods, (3) compare ages of L W D to the age structure of canopy trees to determine how disturbance and stand dynamics contribute to recruitment, and (4) quantify rates of decay and residence times by testing for differences in L W D abundance and volume among sites and among decay and position classes. To identify spatial aspects of L W D generation, I conducted a census of L W D within a 50m reach at each study site. Individual logs ranged from 0.01 to 0.88m ; total log volumes were 6.62 to 41.35 m per ha. For both pine and spruce, log volume 3  decreased significantly between position classes and decay class for both species, largely due to change in the length of logs between classes. To identify temporal aspects of L W D generation, a dendroecological analysis was conducted. Two major disturbance types generated large woody debris in mature riparian forests. Lodgepole pine-dominated stands established ca. 1890 to 1900 after fires. During canopy closure, ca. 40 years after stand initiation, a pulse of tree mortality generated L W D . At the spruce-dominated sites, canopy trees established between 1730 and 1910. These stands were in later stages of stand development and within-stand disturbances generated L W D during the 1900s. I successfully estimated the year of death of 180 logs; 56% of estimates were high quality estimates from samples that included bark and/or sapwood. The age of L W D increased  significantly with decay and position classes. The oldest L W D of lodgepole pine, white spruce and black spruce were 82, 137 and 80 years, respectively. Given the longevity of L W D , I conclude that management decisions that alter the abundance and rate of recruitment of L W D into streams have long-term implications for the structure and dynamics of riparian environments as well as in-stream habitat and biodiversity.  iii  T A B L E OF CONTENTS  ABSTRACT  ii  T A B L E OF CONTENTS  ,  iv  LIST OF TABLES.  v  LIST OF FIGURES  vii  ACKNOWLEDGEMENTS  ix  1. INTRODUCTION  1  1.1 General Introduction and Thesis Overview  1  1.2 Literature Cited  7  2. A WOOD INVENTORY OF L A R G E WOODY DEBRIS IN SMALL STREAMS IN T H E FOOTHILLS MODEL FOREST, HINTON, A L B E R T A . , 9 2.1 Introduction  9  2.2 Methods  11  2.3 Results  22  2.4 Discussion  38  2.5 Literature Cited  44  3. A DENDROECOLOGICAL ANALYSIS OF L A R G E WOODY DEBRIS IN SMALL STREAMS IN T H E FOOTHILLS M O D E L FOREST, HINTON, ALBERTA 3.1 Introduction  ;  49 49  3.2 Methods  51  3.3 Results  64  3.4 Discussion  83  3.5 Literature Cited  92  4. THESIS SUMMARY AND MANAGEMENT IMPLICATIONS 4.1 Conclusions  97 .'  97  iv  LIST OF TABLES Table 2.1 Study area dimensions and volumes of large woody debris in the riparian zone (R. Zone) and stream channel at bankfull (In-stream) for five pine-dominated sties and five spruce-dominated sites 24 Table 2.2 Comparison of volume, length, and diameter of pieces of large woody debris among pine-dominated plots and position classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors 27 Table 2.3 Comparison of volume, length, and diameter of pieces of large woody debris among pine-dominated plots and decay classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors 29 Table 2.4 Comparison of volume, length, and diameter of pieces of large woody debris among spruce-dominated plots and position classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors 31 Table 2.5 Comparison of volume, length, and diameter of pieces of large woody debris among spruce-dominated plots and decay classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors 33 Table 3.1 Summary statistics for site-specific and regional standard chronologies for lodgepole pine (Pine), white spruce (WS), and black spruce (BS) for the Foothills Model Forest ^66 Table 3.2 Correlation matrix comparing site-specific and regional standard chronologies for lodgepole pine in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1907 to 2002 (n = 96 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when ° c = 0.05 67 Table 3.3 Correlation matrix comparing site-specific and regional standard chronologies for white spruce in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1897 to 2002 (n = 106 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when 0.05 68  v  Table 3.4 Correlation matrix comparing site-specific and regional standard chronologies for black spruce in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1890 to 2002 (n = 114 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when ° c = 0.05 68 Table 3.5 Number o f cores which intercepted the pith and which were corrected using Duncan's (1989) method. The correction factor statistics for cores corrected by Duncan's (1989) method, including average and range o f years added to cores at each site 71 Table 3.6 Comparison o f sample number, mean, standard deviation, and median o f pine seedling ages i n years at five lodgepole pine-dominated sites 72 Table 3.7 Comparison o f sample number, mean, and standard deviation o f white spruce seedling ages i n years at four white and black-spruce dominated sites. 73 Table 3.8 Comparison o f sample number, mean, and standard deviation o f black spruce seedling ages i n years at four white and black-spruce dominated sites 73  vi  LIST OF FIGURES Figure 2.1 Location of study sites 1 through 10 in the West Fraser Mills Ltd. Forest Management Area Allocation (FMAA) of the Foothills Model Forest, Alberta 13 Figure 2.2 Volume of individual logs in (a) the riparian zone and (b) the in-stream zone in the stream channel at bankfull at each study site. Note the differences in the scale of the y-axis 25 Figure 2.3 Frequency (%) of L W D in decay classes within position classes at lodgepole pine-dominated sites 35 Figure 2.4 Comparison of position by decay class at all lodgepole pine-dominated sites using contingency tables and chi-square goodness of fit tests 36 Figure 2.5 Frequency (%) of L W D decay classes within position classes at sprucedominated sites 37 Figure 2.6 Comparison of position by decay class at all spruce-dominated sites using contingency tables and chi-square goodness of fit tests 38 Figure 3.1 Location of study sites 1 through 10 in the West Fraser Mills Ltd. Forest Management Area Allocation (FMAA) of the Foothills Model Forest, Alberta 53 Figure 3.2 Standardized white spruce (a), black spruce (b), and lodgepole pine (c) regional chronologies. Sample depth through time is given for each chronology (d) 70 Figure 3.3 Age structure (az'to ei) and standard chronologies (aii to eii) for lodgepole pine-dominated sites. A l l age data were statistically and visually crossdated to the standard chronologies. Sample depth curves represent the sample sizes of the standard chronologies. Chronologies were horizontally standardized to remove growth-related trends. Tree death year was estimated from basal disks by correlation to the standardized chronologies 75 Figure 3.4 Age structure (a/ to ei) and standard chronologies (aii to eii) for sprucedominated sites. A l l age data were statistically and visually crossdated to the standard chronologies. Sample depth curves represent the sample sizes of the standard chronologies. Chronologies were horizontally standardized to remove growth-related trends. Tree death year was estimated from basal disks by correlation to the standardized chronologies 77 Figure 3.5 High quality (HQ) and best estimate (BE) large woody debris age data for lodgepole pine (LP), white spruce (WS), and black spruce (BS) 79  vn  Figure 3.6 Box plots of the year since death of lodgepole pine L W D by position class (a) and decay class (b). For each box, the horizontal line represents the median age. The box limits are the 25th and 75th percentiles of age. The error, bars are the 10th and 90th percentiles of age. The dots show the range in years since death 81 Figure 3.7 Box plots of the years since tree death of spruce L W D by position class (a) and decay class (b). For each box, the horizontal line represents the median age. The box limits are the 25th and 75th percentiles of age. The error bars are the 10th and 90th percentiles of age. The dots show the range in years since death 83  viii  ACKNOWLEDGEMENTS There are many people without whom this thesis would not have been possible. I must thank my advisor Lori Daniels. Thanks to Marwan Hassan for being a supportive member of my committee and to Rob Guy for being a committee member and great teacher. Thank you to the Foothills Model Forest Natural Disturbance Program for funding. A special thanks goes to Dave Andison, Rich McCleary, Rick Bonar and Chris Spytz for technical support and for liaising between my project and the Foothills Model Forest and Hinton Wood Products of West Fraser Mills Ltd. Thank you to all of my field and lab assistants for hard work well done, specifically to Kyle Terry and Jorg Trau for taking initiative on my project and helping me sort out concepts and methodologies and to Scott Wilson, Chantelle Bambrick, Bianka Sawiz, Victor Yan, Kari Werner, and Lou Parsons. Thanks to Vincent Kujala and Erik Schiefer for technical support. I must express my gratitude to my fellow grad students in the tree-ring lab, Amanda Stan, Shane McCloskey, and Jen Passmore for encouragement, support, advice, editorial efforts, and for making our lab a wonderful place to come to. I must thank my peers who led me through the complexities of grad student life. Thank you to Caroline Gravel, Dave Campbell, Josh Caulkins, Sarah Roberts, Chris Borstad, Sara Jackson, Aaron Barsky, and Vanessa Sparrow. Thanks to those friends who were recent graduates and gave me support and mentorship, specifically Bonnie Dong, Paula Vera, and Heather Shay. I must thank those "on the outside" who reminded me to have perspective: Anselmo Scholten, Rick Walls, Earl Morrison, and Troy Sitter. Thanks also to my girls back home for cheering for me and for reminding,me who I am: Amanda Smith, Amanda Stewert, Lauren Kilpatrick, Alana King, and Amy Torok.  M y final thanks goes the Powell clan, especially Mark Powell for financial assistance and to the only "Dr. Powell," Christopher.  x  1. INTRODUCTION 1.1 General Introduction and Thesis Overview Tree rings are often used as a source o f proxy environmental data because variations in tree ring widths can reflect environmental factors that directly or indirectly limit annual biological growth (Fritts, 1976). In ecologically good years, ring widths are wider than normal and in adverse years they are narrower. It is this principal that provides for the making o f paleoecological inferences. It is m y goal to apply this principle to large woody debris ( L W D ) found in small headwater streams o f the Foothills M o d e l Forest, near Hinton, Alberta. The premise o f this study is that the woody material within a stream today may have originated decades to centuries ago. M y main assumption is that when large woody, debris is generally larger than stream size, fluvial transport o f wood downstream transport is limited and wood decays and erodes in situ, linking Wood found i n streams to the immediate riparian zone (Webster et al., 1999). B y examining tree ring-width patterns using dendrochronological techniques, I can determine the year o f death ( Y O D ) o f trees which have fallen into the streams. Once dates o f tree death are known, the longevity and decay rates o f L W D w i l l be reconstructed. The assignment o f calendar years to pieces o f wood w i l l provide a timeframe for the recruitment o f L W D into streams and w i l l allow me to evaluate residence time o f L W D i n streams. B y examining wood i n different stages o f decomposition (decay classes) and i n different positions relative to the stream (position classes), I can assess how long a piece o f wood typically remains in each stage. B y examining changes in wood volume during each stage, I can assess L W D decay rates. I can compare L W D mortality data and L W D volume data to the radial growth information  1  from living canopy dominants to investigate the disturbance histories and forest stand developments o f the surrounding riparian forests. Large woody debris ( L W D ) influences hydrology and geomorphology in small, headwater streams. W o o d i n small streams provides channel structure and stability, specifically i n the creation o f step-pools and plunge-pools, which are integral to macroinvertebrate and fish habitat (Bisson et al. 1987). Studies have concluded that the removal o f L W D from headwater streams has resulted i n a simplification o f habitat and a loss o f biodiversity and fish biomass; the addition o f L W D has resulted i n an increase in plunge pool habitat and a corresponding increase i n fish biomass (Fausch and Northcote, 1991; Roni and Quinn, 2001). These findings indicate that L W D is primary i n creating fish habitat and in supporting fish food webs in small, first-order streams, hence, the recruitment o f L W D is a critical component o f headwater stream ecology (Hogan, 1987; Martin and Benda, 2001). Temporal processes o f large woody debris include L W D recruitment, residency, and decay rate. W e lack knowledge o f the links between stand dynamic processes, the mechanisms that create woody debris, and recruitment into the riparian zone. A recent review o f literature on the dynamics o f wood i n small forested streams by Hassan et al. (2005) lists wood inputs to a channel, yet does not include stand dynamics processes as a potential wood source. From a geomorphic perspective, structural development o f forest ecosystems does not play a large role i n the input o f wood to headwater streams. Alternately, a forest ecology perspective tells us that forest stand development phases are integral to tree mortality, hence the accumulation o f L W D . The competitive exclusion stage o f stand development is characterized by rapid biomass accumulation and  2  competitive exclusion of species and competitive thinning (Franklin et al, 2002). Intense intra-tree competition results in significant mortality, therefore a pulse of mortality should occur during this phase of stand development (Franklin et al, 2002). In evenaged lodgepole pine-dominated stands, I expect to see a decrease in annual incremental growth in the tree-ring chronologies coupled with a pulse of tree mortality, indicating the presence of a competitive exclusion. At mature, uneven-aged spruce-dominated sites, I expect to see incremental growth patterns responding to gap dynamics and small-scale disturbances and chronic L W D generation through time (Franklin et al, 2002). Dendrochronological techniques, specifically cross-dating of ring-width patterns, allows me to: interpret radial growth histories of individual trees, determine the year of initiation of tree growth, and determine the year of the outermost ring (used as an estimate of year of death when bark is missing) (Stokes and Smiley, 1968). L W D contribution to the streambed can be differentiated by (1) historic stand-level disturbances such as fire, (2) tree mortality due to self-thinning of post-disturbance stands, and (3) fine-scale disturbance by wind, flooding, or insects within the riparian forest. Tree deaths can be compared against the Franklin et al. (2002) models of forest stand dynamics and can represent episodic disturbances related to growth phases such as self-thinning or chronic disturbances related to gap dynamics. Dates of tree death can be compared to the dates of disturbances, such as known fires, floods, droughts, or insect infestations. Relatively little is known concerning the temporal variability of L W D in streams, but some work has been done. In a study by Webster et al. (1999), it was found that large i logs which fall into small streams are rarely transported and decay in situ. This study  3  helps justify my main assumption, that the surrounding forest is the sole source contributing to woody debris at each site. Since woody debris decays in situ, residency in headwater streams can only be expressed through the processes of decay and not through transport out of the system. The first influential published work on terrestrial wood decay classes was done by Thomas et al., in 1979. In this publication, decay stages of snags were classified according to a six-class system which later became widely used in the U.S. Pacific Northwest. This system was based on bark structure, branch system, bole fragmentation, and structural integrity. A five-class system of classifying the decay of logs was developed by Maser et al. (1979) in the same publication. This decay class system was the template for work published by Daniels et al., in 1997. In this work, Daniels et al. (1997) attempted to quantify the decay classes into discrete time periods. Some measure of success was realized from this study and conclusions indicated that decay classes as defined by Thomas et al. (1979) might require revisions depending on the biogeoclimatic zone sampled.  1  Hogan (1987) found that complete residency times ranges between 40 and 90 years for L W D in a stream. Webster et al. (1999) found that samples of submerged wood showed neither signs of decay nor a significant decrease in wood density after four to five years residency in the stream (1999). Decomposition is most often studied in laboratories. Hyatt and Naimon (2001) examined decay rates for coniferous wood and found that depletion rates followed an exponential decay curve where 80% of debris pieces were <50 years old, however, some pieces remained for up to 1400 years. Based on this data, a half-life of 20 years for conifer wood was calculated, meaning that almost  4  all conifer wood would disappear before 50 years. A pilot study in the Foothills Model Forest contest these results; pieces o f woody debris ranged from recent death to death 179 years before present, where the oldest debris pieces still maintained enough structural integrity as to distinguish and measure their tree rings (Daniels and Powell, 2003). The findings from Hyatt and Naimon's (2001) study were based on a series o f flume experiments and may have missed an essential component: that wood changes function as it moves towards an increasingly anaerobic environment and becomes incorporated into the stream banks and channel floor. If this latter idea is the case, then decay rates should not fit an exponential decay curve but rather some variation o f a normal curve. Laboratory work on decay rates most often model exponential decay (Murphy and K o s k i , 1989). Work b y Hogan (1987) challenges Hyatt and Naimon's model by stating that the decay rate and therefore residency time is related to the functionality o f L W D . Harmon et al. (2000) uses a technique which examines changes i n decomposition rates through wood age by using a combination o f repeated sampling and dating procedures. B y examining volume and density losses over time, the relationship o f wood decomposition to wood age can be determined (Harmon et al. 2000). Using a combination o f methods, baseline data can be gathered and true stream conditions can be successfully monitored. M y research goals in this thesis are to understand the link between natural disturbance, forest dynamics, and the recruitment and residency o f large woody debris to small, first order streams o f the Foothills M o d e l Forest. Understanding these links w i l l help me to identify historical trends in recruitment and residency o f in-stream wood. In Chapter Two, I quantify the abundance and volume o f L W D in riparian environments i n order to understand how quantity and quality o f L W D change through time. In Chapter  5  Three, I determine ages of in-stream large woody debris using tree-ring methods, compare ages of L W D to the age structure of canopy trees to determine how disturbance and stand dynamics contribute to the recruitment of wood into streams, and quantify rates of decay and residence times by comparing the relationship of wood decomposition to wood age in riparian zones. Understanding large woody debris dynamics has important implications for short-term and long-term management of riparian zones. M y final chapter will address management questions and will suggest sustainable management prescriptions and future work in L W D research.  6  1.2 Literature Cited Bisson, P. A., R. E. Bilby, M . D. Bryant, C. A . Dollof, G. B . Grette, R. A . House, M . L. Murphy, K . V . Koski and J. R. Sedell, 1987. Large Woody Debris in Forest Streams in the Pacific Northwest: Past, Present, and Future. In Stream Management: Forestry and Fisheries Interaction. Contribute 57, E. O. Salo and T. W. Cundy (Editors). Institute of Forest Resources, University of Washington, Seattle, Washington, pp. 143-190 Daniels, L. D., J. Dobry, K . Klinka, and M . C. Feller. 1997. Determining year of death of logs and snags of Thuja plicata in southwestern coastal British Columbia. Canadian Journal of Forest Research.  27: 1132-1141  Daniels, L . D., and S. Powell. 2003. Dendroecological Methods for Analysis of Large Woody Debris in Riparian Zones: A Pilot Study. Final Research Report Submitted to Natural Disturbance Program, Foothills Model Forest, Hinton, Alberta. October, 2003. Fausch, K . D. and T. G. Northcote. 1992. Large woody debris and salmonid habitat in a small coastal British Columbia stream. Canadian Journal of Fisheries and Aquatic Sciences. 49: 682-693  Franklin, J. F., T. A . Spies, R. Van Pelt, A . B. Carey, D. A . Thornburgh, D. R. Berg, D. B. Lindenmayer, M . E. Harmon, W. S. Keeton, D. C. Shaw, K . Bible, J. Chen. 2002. Disturbances and structural development of natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. Forest Ecology and Management.  155:399-423  Fritts, H . C. 1976. Tree Rings and Climate. Academic Press, 567 pp. Hassan, M . A., D. L . Hogan, S . A . Bird, C. L . May, T. Gomi, and D. Campbell, 2005. Spatial and temporal dynamics of wood in small streams. Journal of the American  Water Resources Association.  41:899-919  Hyatt, T. L. and R. Naimon. 2001. The residence time of large woody debris in the Queets River, Washington, U S A . Ecological Applications. 11(1): 191-202 Hogan, D. L . 1987. The influence of large organic debris on channel recovery in the Queen Charlotte Islands, British Columbia, Canada. In: Erosion and sedimentation in the Pacific Rim. Proceedings of the Corvallis Symposium, August, 1987. IAHS Publication No. 165 Martin, D. J. and L. E. Benda. 2001. Patterns of instream wood recruitment and transport at the watershed scale.. American Fisheries Society. 130:940-958  7  Maser, C , R. G . Anderson, K . Cromack Jr., J. T. Williams, and R. E . Martin. 1979. Dead and down woody material. In: Wildlife Habitats in Managed Forests - the Blue Mountains of Oregon and Washington. E d : J. W . Thomas. U S D A For. Ser. Agric. Handb. 553, Washington D . C . p 78-95 Murphy, M . L . A n d K . V . K o s k i . 1989. Input and depletion o f woody debris i n Alaska streams and implications for streamside management. North American Journal of Fisheries Management. 9: 427-436 Roni, P. and T. P. Quinn. 2001. Density and size of juvenile salmonids in response to placement o f large woody debris in western Oregon and Washington streams. Canadian Journal of Fisheries and Aquatic Sciences. 58: 282-292 Stokes, M . A . and T . L . Smiley. 1968. A n introduction to tree-ring dating. University o f Chicago Press, Chicago. Thomas, J. W . , Anderson, R. G . , Maser, C , and E . L . B u l l . 1979. Snags. In: Wildlife Habitats in Managed Forests - the Blue Mountains of Oregon and Washington. E d : J. W , Thomas. U S D A For. Ser. Agric. Handb. 553, Washington D . C . p 6-77 Webster, J. R., Benfield, E . F., Ehrman, T. P. Schaeffel, M . A . , Tank, J. L . , Hutchens, J. J. and D . J. D ' A n g e l o . 1999. What happened to allochthonous material that falls into streams? A synthesis o f new and published information from Coweeta. Freshwater Biology. 41:687-705  8  2. A WOOD INVENTORY OF L A R G E WOODY DEBRIS IN SMALL STREAMS IN THE FOOTHILLS MODEL FOREST, HINTON, ALBERTA 2.1 Introduction Large woody debris ( L W D ) behaves as an important geomorphic constraint in small, headwater streams. W o o d in small streams provides channel structure and stability, specifically in the creation o f step-pools and plunge-pools which are integral to macroinvertebrate and fish habitat (Swansen and Lienkamper, 1978; Bisson et al. 1987). Studies have concluded that the removal o f L W D from headwater streams has resulted in a simplification o f habitat and a loss o f fish biomass; the addition o f L W D has resulted i n an increase in plunge pool habitat and a corresponding increase i n fish biomass (Fausch and Northcote, 1992; Roni and Quinn, 2001). These findings indicate that L W D is primary in creating aquatic habitat conditions and in supporting fish food webs in small, first-order streams, hence, the recruitment o f L W D is a critical component o f headwater stream ecology (Hogan, 1987; Martin and Benda, 2001). Since wood in streams is integral to stream ecological function and stream biodiversity, it is important to quantify the abundance o f large woody debris ( L W D ) in riparian environments and to understand how quantity and quality o f L W D change through time. A n inventory o f wood qualifies the abundance and volume o f L W D in small, headwater streams. Changes in wood quantity through time is controlled by recruitment, storage, and transport processes which are expressed by the in-stream wood budget equation, A S = A l - A O where S = storage, I = input and O = output (Benda et al, 2003). A wood budget describes how wood is delivered to streams, where wood is stored, and how wood is transported out o f or depleted from a drainage basin or a single stream reach (Hassan et al, 2005; Benda et al, 2003). W o o d recruitment processes include fire, mass  9  wasting, snow avalanches, bank erosion, both coarse-scale (ie stand blow-down) and fine-scale (ie tree-fall gaps) wind disturbances, and chronic and episodic mortality related to stand development.(Nakamura and Swansen, 2003; Hassan et al, 2005). Once recruited to a stream, wood is stored in various environments relative to the stream, including within channel boundaries, both above and below bankfull, associated with the stream bed or channel sides, and on associated floodplains (Hassan et al, 2005). Wood depletion processes include transport, abrasion, and decay (Harmon et al, 1986). Bilby and Ward (1991) have reported that the abundance of wood is inversely related to stream size since downstream transport must facilitate the depletion of wood from a stream. When logs are shorter than bankfull width, log movement is most likely to occur (Bilby and Bisson, 1998). When woody debris is generally larger than bankfull width, fluvial transport of wood downstream is limited and wood decays and erodes in situ (Webster et al, 1999). Wood inventories and budgets have been documented in many streams in the Pacific Northwest region (PNW) of North America where stream channels are often coupled with steep hillslopes and where mass wasting, bank erosion, landslides and debris flows during large storm events and snow avalanches play a dominant role in supplying wood to streams (e.g., Nakamura and Swansen, 1993; Keller and Swanson, 1979; Schwab, 1998; Hogan et al, 1998; Hassan et al, 2005). In contrast to the PNW, headwater streams of the eastern foothills of the Rocky Mountains of Alberta are not confined by hillslopes and therefore wood recruitment processes are more likely related to stand-level dynamics and fine-scale disturbance than to geomorphic processes. Given the narrow channel widths of headwater streams combined with subdued topography of  10  the foothills, L W D is more likely to decay in situ. In this paper, I quantify the abundance and volume o f large woody debris in ten headwater stream study sites located in the foothills o f the Rocky Mountains i n west-central Alberta.  Specifically, I want to test i f  there are differences in L W D abundance and volume between sites, and between decay and position classes within sites. I expect to find that L W D abundance w i l l be similar between sites and that wood volume w i l l decrease with increasing decay and position class.  2.2 Methods 2.2.1 Study Areas I conducted this research in the Upper Foothills Natural Subregion in the foothills of the R o c k y Mountains in west-central Alberta (Figure 2.1, Beckingham et al., 1996). A l l study sites were located between 1041 and 1384 meters above sea level (m.a.s.l.) in the Hinton W o o d Products Forest Management Area o f the Foothills M o d e l Forest. Upland forests within this subregion are characterized by closed-canopy coniferous forests, most commonly including lodgepole pine {Pinus contorta Dougl. E x Loud), as well as white spruce (Picea glauca (Moench) Voss) and black spruce (Picea mariana (Mill.) B.S.P.). Frequent, stand initiating fires o f an historical average fire size o f 300 ha, with a mean return interval o f 100 years, have created a landscape mosaic o f successional lodgepole pine, mixed stands with lodgepole pine and white spruce, or mixed white and black spruce (Beckingham et al, 1996; McLeary, 2003). The study area encompasses the Forest Management Agreement Area managed by Hinton W o o d Products, a division o f West Fraser M i l l s Ltd. This area covers 10,000  11  2  k m and includes 20,000 k m o f mapped stream channels within the Athabasca and North Saskatchewan River watersheds. The channels are classified into reaches averaging 300 m in length with consistent stream order, drainage area, channel slope, and forest cover (McLeary et al, 2003) Study sites are found in.drainage networks within the Erith, Eunice, Moberly, Wroe, Fish, Wigwam, and Anderson watersheds which contribute to the larger Athabasca watershed. In this landscape, headwater streams can be defined as small (bankfull width <5.5 m) low-angle (<4% slope) channels that include first- and second-order streams where L W D is large relative to channel depth, resulting in a stable step-pool morphology (Church, 1992; G o m i at al, 2001; M c L e a r y et al, 2003; Hassan et al, 2005).  *  12  Figure 2.1. Location o f study sites 1 through 10 in the West Fraser M i l l s Ltd. Forest Management Area Allocation ( F M A A ) o f the Foothills M o d e l Forest, Alberta.  2.2.2 Site Selection Ten study sites including headwater streams were stratified according to the dominant tree species in the riparian forest surrounding each stream (Figure 2.1). Headwater streams were identified using a geographic information system (GIS) and  13  forest cover and stream inventory data for the Foothills Model Forest (McCleary et al, 2002). These databases were stratified to identify small streams and riparian forests with the following attributes: (1) The stream drainage basin measured <10 km . i  (2) Stream width at bankfull did not exceed 5.5 m. (3) Streams were not confined by hillslopes. (4) There was no evidence of log jams or movement of wood from upstream sources. (5) The surrounding riparian forests were mature (age > 100 years) and dominated either by lodgepole pine, white spruce, or a mixture of black and white spruce. (6) There was no evidence of harvesting of the riparian forest within two treelengths of the stream. Criteria 1 to 4 ensured that wood recruitment was more likely related to standlevel dynamics and fine-scale disturbance as opposed to debris torrents, landslides, or other geomorphic processes. They also controlled for wood transport. If wood is not transported downstream during peak flow events, then L W D decays in situ. In absence of wood transport, the wood found in the stream is directly associated with the adjacent riparian zone and an analysis of this wood can be linked to the spatio-temporal dynamics of the surrounding forest. Criterion 5 allowed comparison of L W D dynamics among and between mature pine- and spruce-dominated forests. Criterion 6 ensured that the study sites represented natural forest dynamics rather than modern anthropogenic effects such as timber harvest (Young et al, 2006).  14  2.2.3 Field  Sampling  At each study site, I surveyed a 50 m reach that included several step-pool geomorphic units to allow comparison to other published studies of stream geomorphology and L W D (Hauer et al, 1999; Berg et al, 1998; Richmond and Fausch 1995). The start points for the surveys were randomly determined. Survey transects moved upstream from the start point, along the thalweg. Stream width at bankfull was measured every five meters from 0 to 50 m. I censused all large woody debris (LWD) along the transect. L W D was defined as logs which exceeded 0.1 m in diameter and 1 m in length (Murphy and Koski, 1989; Fausch and Northcote, 1992; Nakamura and Swanson, 1994; Richmond and Fausch, 1995; Hauer et al, 1999; Marcus et al, 2002; Faustini and Jones, 2003). For each piece of L W D , I noted its location upstream from the start point (m) and orientation of the long-axis of the wood relative to the stream channel. When possible, I determined if the log was lodgepole pine or spruce, but I could not differentiate white and black spruce based on the morphology of the logs. For logs in advanced stages of decay, I indicated the species, was unknown. There is no standard procedure for measuring the dimensions of L W D ; several different methods have been reported in the literature, depending on the objectives of the study (Fausch and Northcote, 1992; Berg et al, 1998; Gomi et al, 2001). In this study, my goal was to estimate volume of L W D in the riparian zone, the area from which large woody debris was recruited, and the in-stream zone, the area within the stream channel at bankfull. Depending on the condition of the piece of wood, I measured length (m) and diameter (cm) as follows. For logs with a visible root wad, I measured diameter at breast height (dbh), 130 cm along the stem from the roots. I also  15  measured the distance from the roots to the centre of the stream channel, perpendicular to the channel, to determine the width of the riparian zone. For all other logs, diameter was measured at the midpoint of the length of the log. Logs that were embedded in the stream bank were excavated to estimate length and diameter. When excavation was not possible, a minimum length was recorded. For all logs I measured the length of the log directly associated with the channel at bankfull and the diameter at the midpoint within the channel. The presence or absence of soil and the percent cover of vegetation on the upper surface of each piece of wood was noted. Two classification systems were used to represent decreasing structural stability and decay of wood over time. Each classification system was assigned four categories so that I could directly compare the two systems with respect to wood volume and abundance. Thus, each piece of wood was assigned a position class (PC) and a decay class (DC). Position classes have been used in studies of L W D in aquatic systems to describe wood position, structure, and function relative to the stream (Hauer et al, 1999; Berg et al, 1998; Richmond and Fausch 1995). The decay class system was adapted from research of terrestrial coarse woody debris by Maser et al. (1979), Thomas et al (1979), Storaunet and Rolstad (2002) and Daniels et al (1997) and was used to classify wood based on wood structural characteristics and integrity. The position classification system highlights the relative position, and thereby function, of wood in a stream. Position classes have been used in studies of L W D in aquatic systems by Hauer et al, (1999), Berg et al, (1998), Nakamura and Swansen (1994) and Richmond and Fausch (1995) and describe the position of wood relative to the channel. In the classification system I used, L W D was divided into four position classes  16  (PC): (PCI) Bridge - log spans channel, touching both banks and resting on the floodplain, (PC2) Partial bridge - spanned log has broken in one or more places within the stream channel, (PC3) Loose - log is no longer associated with the floodplain and is now associated with the streambed; log is fully submerged during bankfull, (PC4) Buried - log has become incorporated into the streambed or the sides of the streambank and has stored sediment and become at least partially buried. With respect to function of L W D in the stream, L W D position classes 1 and 2 provide overhead cover, shade, and protection from predators for salmonids (Richmond and Fausch, 1995). L W D in classes 3 and 4 is integral for stream geomorphology and habitat diversity by forming steps and pools, storing sediment, and providing bank stability (Richmond and Fausch, 1995). The four categories of decay include the following attributes: Decay Class 1 (DC1) wood has > 75% bark still intact, bark adheres tightly; branches have fine (3  ld  order) branchlets; sapwood is sound, and log retains structural integrity (DC2) L W D has 25% - 75% bark intact which, in places, is loosely attached to the bole; 1 order branches st  have a solid connection to the bole; wood is solid with evidence of decay on outer sections of sapwood only, (DC3) wood has 0% to 25% bark present, adhering loosely to the sapwood; 1 order branches and branch nubs are present and sit loosely in the bole; st  along some parts of the bole, wood shows significant signs of decay to depths of 5 to 10 cms, (DC4) bark is no longer attached; branch nubs only are present; along some parts of the bole, wood is soft, crumbly or fibrous, and decay can penetrate nearly through the sapwood.  17  2.2.4 Data Analysis 2.2.4.1 Calculation of L WD volumes  I used two different equations to calculate total volume of each log, depending on the condition of the piece of wood. For tapered logs with a visible root wad, for which dbh (m) and length (m) were measured, I calculated volume (m ) as: 3  Total log volume tapered = (7t)(dbh/2) (length) 2  [1]  3 For segments of downed trees that were cylindrical and for which length (m) was measured but dbh was not discernable, I used the following equation:  Total log volume cylindrical  =  (7i)(diameter/2) (length) 2  [2]  where, diameter (m) was measured at the midpoint along the length of the log. For all logs, I identified the segment of the log directly associated with the stream channel at bankfull and used the following equation to calculate its volume (m ):  In-stream log volume = (71) (diameter/2) (length)  [3]  where, diameter (m) was measured at the midpoint of the segment of the log in the channel and length (m) was measured along the log within the channel from one bankfull boundary to the other or from bankfull to the end of the log. I used analysis of variance ( A N O V A ) to test for differences in mean total log volume and mean in-stream log volume among sites (Anonymous, 1997). Both variables  18  were tested for normality and equal variances, which are primary assumptions of A N O V A (Zar, 1984). They failed the assumption of normality; however, A N O V A is robust with respect to population normality when sample size is large (Box and Anderson, 1955; Srivastava, 1959; Tiku, 1971; Zar, 1984). Since my sample size was 326 logs, I proceeded with the A N O V A s . For significant variables, I used Tukey's pairwise multiple comparison procedure to compare log volumes among sites. At the spatial scale of the stream reach, I calculated the total volume of L W D for the riparian zones by summing the total volumes for tapered and cylindrical logs in each reach. I calculated the total in-stream volume of L W D by summing the in-stream log volumes for each reach. These values represent the absolute amount of L W D in the study sites, but are not directly comparable because the width of the riparian zones and stream channels varied among the 10 reaches. To allow direct comparison of the volumes among sites, I scaled the total volumes of L W D according to the area of each riparian zone and stream channel as follows:  Scaled L W D volume in riparian zone = total volume of L W D (m ) area of riparian zone (m ) 3  [4]  where, the area of the riparian zone was the length of the study reach (50m) multiplied by the width of the riparian zone. The width of the riparian zone was defined as two times the distance between the roots and the centre of the stream channel for the L W D that was furthest from the channel. In this study, the width of the riparian zone demarcated the area in which trees can fall and contribute to in-stream L W D .  I used the following equation to scale the total in-stream volumes of L W D :  19  Scaled in-stream volume of L W D = total instream volume of L W D (m ) area of stream channel (m )  [5]  where, the area of the channel was the length of the study reach (50m) multiplied by the average width (m) of the channel. 2.2.4.2 Statistical Analysis — Lodgepole Pine-dominated  * Sites  I tested for differences in the mean volume, diameter and length of individual logs among lodgepole pine-dominated study sites and among position classes and decay classes using two-way analysis of variance (ANOVA) (Anonymous, 2004). A l l variables were tested for normality and equal variances, which are primary assumptions of A N O V A (Zar, 1984). I applied the natural logarithmic transformation since all variables were not distributed normally. Neither logarithmic nor square root transformations of diameter met the assumption of normality; however the natural logarithmic transformation passed an equal variance test (Anonymous, 1997). Since A N O V A is robust with respect to population normality when sample size is large (Box and Anderson, 1955; Srivastava, 1959; Tiku, 1971; Zar, 1984) and my sample size was 154 logs, I proceeded with the test. For the decay class analysis, the natural logarithm of all three variables passed tests of normality and equal variances. For significant interactions and factors, I conducted a Tukey's pairwise multiple comparison procedure. A l l tests were considered significant at the «= = 0.05 level. A l l results were reported in the original units.  20  2.2.4.3 Statistical Analysis - Spruce-dominated  Sites  I similarly tested for differences in abundance of L W D , measured as volume, among spruce-dominated study sites, position classes and decay classes using analysis of variance (ANOVA) (Anonymous, 2004). Differences in diameter and length were also tested to determine if they explained changes in volume. A l l variables were tested for normality and equal variances, which are primary assumptions of A N O V A (Zar, 1984). Since all variables were not distributed normally, I applied the natural logarithmic transformation. Since A N O V A is robust with respect to normality when sample sizes are large, and my sample size was 155 logs, the natural logarithms were used in multiple comparison procedures (Box and Anderson, 1955; Srivastava, 1959; Tiku, 1971; Zar, 1984). Using the transformed variables, I tested for differences in mean volume, diameter, and length of wood pieces between study sites, position and decay classes within study sites, and for interactions between sites and classes, using two-way A N O V A (Anonymous, 2004). For interactions and for factors which were significant at the °= = 0.05 level, I conducted a Tukey's pairwise multiple comparison procedure. A l l results are reported in the original units. 2.2.4.4 Comparison of position and decay classes  I calculated the frequency of logs in each position and decay class, for each site, for the five pine sites combined and for the five spruce sites combined. I tested for a relationship between decay classes and position classes using a contingency table and chi-square (y ) analysis (Zar, 1984). Tests were conducted separately for the five pine2  dominated sites and five spruce-dominated sites. A l l statistical tests were considered significant at the °= = 0.05 level.  21  2.3 Results 2.3.1 Volume of LWD Within each 50m reach, there were 16 to 48 logs (32.5 ± 8.7) (Table 2.1, Figure 2.2). The mean number o f logs in sites dominated by lodgepole pine (33.8 ± 11.6, n = 5) was similar to the number o f logs in spruce-dominated sites (31.2 ± 5.7, n = 5). I was unable to determine the genera o f all pieces o f L W D . M y assumption was that most L W D in lodgepole pine-dominated forests was lodgepole pine, and L W D in sprucedominated forests was largely spruce. The volume o f individual logs ranged from 0.01 to 0.88 m (Figure 2a). W i t h i n 3  sites, mean log volumes ranged from 0.06 to 0.16 m , but were not significantly different 3  (p < 0.001). Most pieces o f wood (75%) were less than 0.2 m , but a few logs were very 3  large, measuring 0.88 m (Site 2), and 0.87 m (Site 9). 3  3  The largest range o f volumes  occurred at sites 2, where volumes range from 0.01 to 0.88 m , and site 9, where volumes 3  ranged from 0.01 to 0.87 m . Site 4 had the smallest range o f volumes, between 0.006 3  and 0.18 m . 3  The in-stream volumes o f individual logs represented a section o f most pieces o f L W D and were smaller than the total volumes, ranging from 0.0001 m to 0.53 m 3  3  (Figure 2.2b). Within sites, mean in-stream volumes ranged from 0.06 to 0.16 m , but were not significantly different (p < 0.001). Most pieces o f wood (75%) were less than 0.17 m (site 9), but a few logs were very large, measuring 0.53 m (site 8) and 0.54 m (site 10). The largest range o f volumes occurred at site 8, where volumes range from 0.01 m to 0.53 m . Site 4 has the smallest range o f volumes, between 0.006 m and 0.1 3  3  3  m . 3  22  At the scale of the reach, the absolute volume of L W D in the riparian zones varied from 0.98 to 6.81 m (Table 2.1). Site 4 had the lowest volume of wood at lodgepole 3  3  3  pine-dominated sites, 0.98 m , and Site 3 had the largest total volume, 4.96 m . At the spruce-dominated sites, Site 6 had the lowest total volume, 1.76 m , and Site 9 had the 3  largest volume of wood of all sites, 6.45 m . The volume of in-stream L W D was 30.1 to 3  68.4 % of the total volumes and ranged from 0.67 to 4.23 m . At the site level, the in3  stream volumes followed similar trends as the total volumes; Site 9 had the largest volume of in-stream wood, 4.23 m . The widths and areas of the different riparian zones studied were variable. Average in-stream bankfull widths ranged from 0.8 m (Site 10) to 3.5 m (Site 4) (Table 2.1). As a result, the smallest in-stream area was at Site 10 and the largest area was at Site 4. The largest riparian-zone area (Site 1 =0.175 ha) differed from the largest instream area since this measure was calculated using not stream channel width, but the longest pieces of wood on either side of the bank extending into the riparian zone. Thus riparian area and in-stream area were not linked at the site level. The smallest riparian area was at Site 8 where riparian area equaled 0.081 ha. With respect to scaled volume in lodgepole pine-dominated riparian zones, Site 2 had the largest volume of L W D per hectare, 32.37 m /ha, and Site 4 contained the smallest volume, 6.62 m /ha. At the 3  3  spruce-dominated study sites, Site 9 had the largest volume of wood per hectare, 41.35 m /ha, and Site 6 had the smallest, 14.38 m /ha. The scaled volume of in-stream L W D at 3  3  pine-dominated sites ranged from 38.24 (Site 4) to 246.36 m /ha (Site 3). In-stream 3  L W D at spruce dominated sites ranged from 110.30 (Site 7) to 272.90 m /ha (Site 9). 3  23  Site 1 2 •3 4 5 6 7 8 9 10  Table 2.1 Studyarea dimensions and volumes of large woody debris in the riparian zone (R. Zone) and stream channel at bankfull (Instream) for five pine-dominated sties and five spruce-dominated sites. Width ( m) Area (ha) Volume (m ) Scaled Volume (m /ha) N Length (m) Species In-stream In-stream In-stream R. Zone In-stream R. Zone R. Zone R. Zone Channel J  0.175 2.15 4.56 2.9 35.0 0.015 148.28 Pine 33 50 21.4 0.107 1.45 3.47 161.11 Pine 34 50 0.009 1.8 2.2 34.0 0.011 0.170 2.71 4.96 246.36 Pine 48 50 0.67 0.98 38.29 50 3.5 29.6 0.018 0.148 Pine 16 0.015 0.167 4.78 100.00' Pine 38. 50 3.0 33.5 1.50 0.122 1.32 1.76 114.78 2.3 24.5 0.012 Spruce 30 50 16.2 0.081 1.82 2.77 110.30 50 3.3 0.017 Spruce 33 0.012 0.107 1.78 2.65 148.33 50 2.4 21.3 Spruce 25 31.2 " 0.156 4.23 6.45 272.90 50 3.1 0.016 Spruce 40 0.004 0.153 0.93 3.09 232.50 0.8 30.5 Spruce 28 50 In-stream includes the in-stream portion of the logs, where the logs were measured for volume between channel margins. R . Zone (Riparian Zone) includes the instream wood as well as wood resting on the floodplain where the piece of wood extending furthest from the stream, but still crosses the channel margin, denotes the riparian zone boundary.  3  26.057 32.37 29.18 6.62 28.57 14.38 34.20 24.88 41.35 20.26  2  24  1.0  0.8  0.6  _  0.4  o  > 0.2  0.0 n = 33 n = 34 n = 48 n = 1 7 n = 38 n = 30 n = 33 n = 25 n = 40 n = 28  0.6  0.5  0.4  S  0.3  H  ©  E -§ >  0.2  0.1  •  •  ± ±  •  0.0 n = 32 n = 34 n = 48 n = 17 n = 38 n = 30 n = 33 n = 25 n = 40 n = 28 -i 1 1 1 1 r T r 1 2 3 4 5 6 7 8 9 10  S t u d y Site  Figure 2.2 V o l u m e o f individual logs i n (a) the riparian zone and (b) the in-stream zone in the stream channel at bankfull at each study site. Note the differences i n the scale o f the y-axes.  25  2.3.2 Comparison of Logs Among Position and Decay Classes  For the five lodgepole pine-dominated sites combined, only 31 logs were in PCI and more than 40 logs were in each of the other position classes. Within sites, the smallest number of logs was in PCI at three of five sites. The largest number of logs was in classes 2 and 3 at all sites combined and at three of five individual sites. The volume, length, and diameter of lodgepole pine L W D varied significantly among position classes (p = < 0.001) (Table 2.2). For volume and length, position classes 1 and 2 were significantly greater than classes 3 and 4, but classes 1 and 2 were not significantly different from each other, nor were classes 3 and 4. For diameter, position classes 1 and 2 were significantly greater than PC3, and PC4 overlapped with these two groups.  26  Table 2.2 Comparison o f volume, length, and diameter o f pieces o f large woody debris among pine-dominated plots and position classes. For each characteristic, mean values are followed b y standard deviations i n parentheses. Means i n each column followed b y the same superscript are not significantly different. Probability values (p-values) <0.05 are i n bold, indicating significant factors. . Diameter (cm) Length (m) N Volume (m ) Factor m (sd) m (sd) m (sd) 17.66 (6.00) 7.80 (6.34) 33 0.14(0.16) Plot 1 15.26 (7.06) 6.04(5.08) 2 34 0.10(0.17) 14.69 (5.91) 0.10(0.14) 7.73 (7.17) 3 48 15.63 (4.27) 5.39(5.32) 0.06 (0.05) 4 16 20.71 (23.03) 0.12(0.09) 9.01 (5.65) 5 39 22.53 (26.1,4) 14.17 (5.46) 31 0.18 (0.18) PC 1 18.56 (6.03) 11.05 (5.38) 0.17(0.14) 2 49 13.46 (4.32) 0.06 (0.06) 3.78 (2.66) 3 49 14.55 (4.29) 0.05 (0.08) 2.56 (1.76)" 4 41 Plot x Position Class 21.95 (9.15) 13.67 (7.35) 1 x1 6 0.31 (0.27) 17.75 (4.83) 9.86 (5.64) 1x2 14 0.14(0.10) 15.04 (4.92) 2.92(1.42) 0.54(0.03) 1x3 9 16.80 (4.56) 2.78(1.17) 1x4 4 0.06 (0.03) 13.05 (3.02) 4 0.06 (0.27) 12.21 (1.35) 2x1 22.07 (11.85) 10.31 (5.66) 0.28 (0.29) 2x2 7 13.35 (4.45) 4.55 (3.81) 2x3 12 0.05 (0.05) 13.83 (3.80) 2.71 (2.54) 11 0.06 (0.11) 2x4 18.75 (6.23) 16.15 (5.57) 3 x1 13 0.21 (0.19) 17.55 (5.64) 0.13 (0.13) 10.18(7.28) 3x2 8 11.12(3.88) 3.41 (2.13) 13 0.04 (0.05) 3x3 12.59 (4.55) 14 0.05 (0.11) 2.51 (1.33) 3x4 15.45 (7.14) 11.16(7.41) 4x1 2 0.12 (0.09) 18.43 (2.94) 0.11 (0.05) 10.26 (5.34) 4x2 4 11.63 (2.95) 3.46(2.39) 4 0.03 (0.02) 4x3 16.48 (3.75) 0.03 (0.02) 1.52 (0.47) 4x4 6 40.00 (58.40) 12.72 (4.33) 5x1 6 0.11 (0.05) 18.28 (4.06) . 0.17(0.09) 13.05 (3.96) 5x2 16 15.72 (3.39) 4.22 (2.73) 5x3 11 0.09 (0.11) 17.03 (3.19) 3.31 (2.03) 5x4 0.06 (0.02) 6 3  a  a  a  a  a  a  b  b  b  ab  b  P-values for A N O V A Plot Position Plot x Position  0.409  0.051  <0001  <0.001  <0.001  0.265  0.889  0.439  0.500  27  For the five lodgepole pine-dominated sites combined, 7 pine logs were in decay class ( D C ) 1 and over than 30 logs were i n each o f decay classes 2, 3, and 4. D C 3 contained the largest number o f logs, n = 74. Within each site, the largest number o f logs per class was in D C 3 at all sites but Site 4. The smallest number o f logs at each site was in D C 1 ; Sites 4 and 5 had no logs i n D C 1 . For the lodgepole-pine dominated sites, volume per plot (p = 0.025) and volume per decay class (p < 0.001) were significantly different and interacted significantly (Table 2.3). . Although significant differences occurred, there was tremendous overlap between plots and decay classes with respect to In volume. Decay classes 1 and 3 had significantly higher values than decay classes 3 and 4 although there was overlap among all classes. Specifically, at site 1, within D C 2 , and Site 3, D C 1 , volume was significantly greater than all other decay classes at all other sites. This result can be attributed to large average lengths and diameters. The length o f L W D was statistically different among decay classes; classes 1 and 2 were longer than class 3 and 4. N o significant differences in diameter were found among plots or decay classes and they did not interact.  28  Table 2.3 Comparison of volume, length, and diameter of pieces of large woody debris among pine-dominated plots and decay classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors. Diameter (cm) Length (m) Factor N Volume (m ) m (sd) m (sd) m (sd) 7.80 (6.34) 17.66 (6.00) Plot 1 33 0.14(0.16) 15.26 (7.06) 0.10(0.17) 6.04 (5.08) 2 34 7.73 (7.17) 14.69 (5.91) 0.10(0.14) 3 48 15.63 (4.27) 5.39(5.32) 4 0.06 (0.05) 16 20.71 (23.02) 9.01 (5.65) 0.12(0.09) 5 39 21.90 (9.34) 17.07 (9.45) 0.33 (0.25) DC 1 8 21.62 (24.03) 0.15 (0.12) 13.90 (4.21) 2 36 15.59 (5.83) 0.10(0.14) 6.17(4.72) 74 3 3.42 (2.62) 14.55 (4.59) 4 52 0.06 (0.06) Plot x Decay Class 19.58 (9.18) 14.94(10.33) 1x 1 4 0.24 (0.26) 22.25 (9.50) 14.88 (5.48) 1 x2 4 0.32 (0.27) 5.80 (4.25) 16.79 (5.29) 15 0.10(0.09) 1 x3 16.37(3.50) 5.11 (3.07) 1 x4 0.09(0.04) 10 9.40 5.15 2x 1 4 0.04 11.93 (32.38) 10.34(3.16) 2x2 4 0.05 (0.02) 17.06 (8.62) 0.15 (0.22) 6.62 (5.61) 2x3 18 14.08 (4.58) 3.62 (3.74) 2x4 11 0.05 (0.05) 29.17(3.33) 23.88 (1.68) 3x 1 3 0.53 (0.09) 16.55 (3.44) 15.62 (4.02) 3x2 10 0.13 (0.10) 4.82 (4.21) 13.69 (4.19) 0.06 (0.06) 3x3 19 11.98 (5.16) 3x4 0.05 (0.10) 3.21 (1.88) 16 4x1 0 18.80 (2.13) 13.67 (2.60) 4x2 4 0.13 (0.05) 3.95 (2.34) 11.38 (2.61) 0.03 (0.02) 4x3 5 16.84 (3.55) 0.04 (0.02) 1.70 (0.63) '4x4 7 5x1 0 28.64 (37.66) 13.48 (4.35) 5x2 14 0.16(0.07) 16.35 (3.97) 0.13 (0.12) 8.18 (4.78) 5x3 17 2.95 (1.62) 16.08 (3.46) 5x4 0.05 (0.03) 8 P-values for A N O V A 0.235 0.146 Plot 0.025 0.282 <0.001 Decay <0.001 0.482 0.97 Plot x Decay 0.15 J  a  a  b  b  abc  c  ab  ab  a  ab  ab  a  c  abc  a  a  a  a  a  abc  abc  ab  29  A trend similar to the distribution of lodgepole pine logs among position classes occurred at the spruce-dominated sites, where the fewest logs were in P C I . Twenty-nine spruce logs were in P C I , 36 logs were in PC2, and over 40 logs were found in classes 3 and 4. At three of five sites, position class 1 contained the least number of logs. At Site 6, only one log was found in P C I . The greatest numbers of logs were in classes 3 or 4 for all sites, except Site 8, where the greatest number of logs was in P C I . For all sites combined, the greatest number of logs was in PC4. At the spruce-dominated sites, L W D volume and length were significantly different among plots (p = 0.034 and p < 0.001, respectively) and volume, length, and diameter, were significantly different among position classes (p < 0.001 for all three variables, Table 2.4). Tukey pairwise comparisons did not differentiate L W D volume among the plots, but they did differentiate L W D length among plots (Table 2.4). L W D length divided into two groups in which lengths in plots 8 and 9 were greater than length in plot 10 and lengths in plots 6 and 7 overlapped with both groups. With respect to volume, position class 1 and 2 were greater than 3 and 4. Length of logs in position classes 1 and 2 was greater than class 3 and 4. Log diameters were divided into two groups whereby mean diameter in position class 1 and 2 was greater than position class 3, but position class 4 overlapped with these two groups.  30  Table 2.4 Comparison of volume, length, and diameter of pieces of large woody debris among spruce-dominated plots and position classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors. Factor Length (m) Diameter (cm) n Volume (m ) m (sd) m (sd) m (sd) 14.97 (5.04) Plot 6 3.47(4.02)" 30 0.06 (0.07) 17.00 (7.64) 7 33 0.08 (0.12) 5.30 (4.50) 8 0.11 (0.12) 7.03 (5.03) 17.11 (6.01) 25 7.26 (6.00) 19.96 (7.89) 9 39 0.16(0.17) 10 0.11 (0.15) 5.03 (6.57)" 20.62 (7.58) 28 PC 1 0.33 (0.51) 3.12 (12.87) 24.17 (13.40) 29 2 36 0.17(0.24)" 2.91 (8.25) 20.14 (10.29) 14.66 (6.14) 3 41 2.59 (2.51)" 0.05 (0.05) 4 0.05 (0.07) 2.86 (2.06)" 18.45 (7.29) 49 Plot x Position Class 6x1 1 0.31 17.04 26.50 6x2 7.32 (4.92) 17.07 (6.16) 6 0.12(0.07) 12.35 (4.14) 6x3 8 0.03 (0.03) 2.04 (0.67) 6x4 15 0.03 (0.03) 1.80 (0.94) 14.75 (3.89) 7x1 22.66 (5.90) 5 .0.23 (0.11) 12.03 (3.98) 7x2 21.46 ( 12.57) 7 0.13 (0.20) 9.00 (4.60) 13.97 (4.21) 7x3 14 0.03 (0.02) 2.27 (1.08) 7x4 4.56 (3.72) 7 0.03 (0.03) 2.83 (4.10) 8x 1 10.84 (2.44) 17.78 (4.03) 8 0.16(0.16) 8x2 0.14(0.12) 9.79 (4.92) 15.58 (6.04) 6 11.83 (2.10) 8x3 4 0.04 (0.02) 4.00 (3.05) 20.67 (7.52) 8x4 7 0.05 (0.02) 1.59(0.75) 9x 1 15.01 (6.85) 23.27 (9.42) 7 0.33 (0.31) 19.16(6.46) 9x2 11 0.17(0.10) . 9.83 (3.38) 12 17.09 (5.76) 9x3 0.09 (0.07) 3.13 (1.51) 9x4 3.61 (4.04) 22.18 (10.17) 9 0.12(0.13) 24.92 (7.88) lOx 1 8 0.35 (0.32) 13.05 (5.78) 10x2 18.82 (4.89) 6 0.03 (0.10) 3.89 (5.91) 10x3 3 0.09 (0.09) 0.95 (0.09) 9.53 (4.81) 10x4 11 0.04 (0.03) 0.94 (0.17) 19.78 (7.85) P values for A N O V A 0.064 Plot 0.034 O.001 Position O.001 <0.001 <0.001 Plot x Position 0.175 0.739 0.151 3  ab a  a  a  3  a  a  3  c  b  c  a  31  Similar to the distribution of logs among decay classes at the lodgepole pine sites, the fewest logs were found in DC1 at the spruce-dominated site. For all spruce-dominated sites combined, eight logs were in DC1, and 39 to 56 logs were in decay classes 2 through 4. At all individual sites, DC1 contained the fewest number of logs; Site 7 had no logs in decay class 1. The greatest number of logs was in classes 3 or 4 for all sites, except Site 8, where the greatest number of logs was in DC2. At the spruce-dominated sites, L W D volume (p < 0.001) and length (p < 0.001) were significantly different among decay classes (Table 2.5). Decay classes 1 and 2 had significantly greater volume values than decay classes 3 and 4. L W D lengths were grouped into three statistically different groups in which class 1 was longer than DC2. Both classes 1 and 2 were longer than decay classes 3 and 4. Diameter was not significantly different among decay classes but it varied significantly among plots. The diameter of L W D in plot 10 was greater than diameter in plot 6, and diameters in plots 7, 8, and 9 overlapped with all plots (Table 2.5).  32  Table 2.5 Comparison o f volume, length, and diameter o f pieces o f large woody debris among spruce-dominated plots and decay classes. For each characteristic, mean values are followed by standard deviations in parentheses. Means in each column followed by the same superscript are not significantly different. Probability values (p-values) <0.05 are in bold, indicating significant factors. Diameter (cm) Length (m) Volume (m ) Factor n m (sd) m (sd) M(sd) 14.97 (5.04) 3.47 (4.01) 30 0.06 (0.07) Plot 6 17.00 (7.64) 5.30 (5.00) 0.08 (0.12) 7 33 17.12 (6.01) 7.03 (5.03) 25 0.11 (0.11) 8 7.26 (6.00 19.96 (7.89) 0.16(0.17) 9 39 20.62 (7.58) 5.03 (6.57) 10 28 0.11 (0.15) 28.06 (23.54 14.30 (5.04) 0.58(0.91) DC 1 8 20.24 (9.03) 10.78 (4.42) 2 0.20 (0.25) 39 18.08 (8.34) 52 4.19(4.57) ' 3 0.09 (0.13) 2.16(2.34) 17.45 (7.37) 4 56 0.05 (0.07) Plot x Decay Class 26.50 17.04 6x1 1 0.31 18.88 (4.77) 8.41 (4.62) 6x2 0.14(0.05) 5 13.50 (5.97) 2.31 (0.40) 5 0.04 (0.03) 6x3 13.72 (3.75) 0.03 (0.03) 6x4 19 1.18 (0.88) 7x1 20.37 (6.53) 11.94 (3.05) 7x2 0.17(0.13) 7 16.66 (8.83) 3.54 (3.79) 0.07 (0.13) 7x3 19 3.4 (3.94) 14.53 (3.76) 7x4 7 0.04 (0.03) 18.35 (10.68) 2 0.28 (0.35) 10.70 (0.99) 8x1 16.13 (4.80) 10.24 (4.05) 8x2 11 0.13 (0.09) 13.70 (2.62) 5.2 (4.12) 0.06 (0.05) 8x3 5 20.73 (7.43) 0.05 (0.02) 1.43 (0.66) 8x4 7 10.38 (4.42) 15.80(1.41) 9x1 2 0.11 (0.23) 20.67 (8.21) 12.00 (5.09) 9x2 11 0.24 (0.17) 19.42 (7.00) 6.68 (6.37) 9x3 13 0.15 (0.23) 20.53 (9.37) 3.36 (3.36) 9x4 13 0.11 (0.19) 24.38 (10.40) 18.40 (4.62) lOx 1 3 0.35 (0.32) 21.64 (7.30) 10.05 (5.46) 10x2 5 0.03 (0.10) 19.3 (6.67) 2.63 (3.57) 10x3 10 0.09 (0.09) 20.64 (8.54) 0.92 (0.16) 10x4 0.04 (0.03) 10 P values for A N O V A 0.698 <0.001 Plot 0.470 0.055 Decay <0.001 <0.001 0.122 0.159 Plot x Decay 0.477 J  a  ab  ab  ab  b  a  a  a  b  b  c  b  c  33  2.3.3 Comparison of Position and Decay Classes  In the lodgepole pine-dominated riparian forests, the position and decay classes were significantly related (p = <0.001) (Figures 2.3 and 2.4), although the two classification systems varied relative to one another. For example, pine logs in PCI were mostly in DC2 and included a greater than expected number of logs in DC2. Logs in PC2 included mostly logs in decay classes 2 and 3. The number of observations in decay classes 2 and 3 were greater than expected. In PC3, the majority of logs were in DC3, which was the only decay class to have more observations than expected. Logs in PC4 were mostly in DC4, but included some logs in decay class 3. Only DC4 included more observations than expected. PC4 contained the least number of decay classes and was the least variable. Position classes 2 and 3 were the most variable and contained all decay classes.  34  25 _• _• ZZI ZZ  Decay Decay Decay Decay  Class 1 Class 2 Class 3 Class 4  I  j  i 2  1 4  Position Class  Figure 2.3 Frequency (%) of LWD in decay classes within position classes at lodgepole pine-dominated sites.  35  35 x = 116.18 (df=9, p < 0.001) z  30 4 Observed Expected  25  |  cr Q)  20  H  15  H  10  la  l U  JL  T 1 i i i I i 1 1 i i i i 1 i r P1xD1 P1xD2 P1xD3 P1xD4 P2xD1 P2xD2 P2xD3 P2xD4 P3xD1 P3xD2 P3xD3 P3XD4 P4XD1 P4xD2 P4xD3 P4xD4  D e c a y C l a s s e s Nested in Position C l a s s e s Figure 2.4 Comparison of position by decay class at all lodgepole pine-dominated sites using contingency tables and chi-square goodness of fit tests.  Similarly, the decay classes and position classes at all spruce-dominated sites were significantly related (p = <0.001) (Figures 2.5 and 2.6). Position class one was dominated by logs in decay classes 1 and 2 but also contained logs in decay class 3. Position class 2 was dominated by logs in decay classes 2 and 3 and was the most variable class, containing all decay classes. Position class 3 was dominated by logs in decay class 3 and 4 and had a small percentage of logs in decay class 2. Position class 4 was dominated by decay class 4 but contained a small percentage of decay classes 2 and 3. The number of observations in decay classes land 2 within position classes 1 were  36  greater than expected by chance. Observations in decay class 2 within position class 2, decay class 3 within position class 3, and decay class 4 within position class 4 were greater than expected by chance. Clearly, these two classification systems are linked.  35 I  30  Decay Class 1 I Decay Class 2 Decay Class 3 1 Decay Class 4  25  o cz  20 4  CD  =3 CT  15 H  10  H  i  i l 1  2  JZL  i  J3J  3  4  Position C l a s s  F i g u r e 2.5 Frequency (%) o f L W D decay classes within position classes at sprucedominated sites.  37  50 x"= 146 (df=9,p<0.001)  40  Observed Expected  c CD  cr CD  20  10 A  -i i i i I HP i p HP HP P P1xD1 P1xD2 P1xD3 P1xD4 P2xD1 P2xD2 P2xD3 P2xD4 P3xD1 P3xD2 P3xD3 P3XD4 P4XD1 P4xD2 P4xD3 P4xD4  Decay C l a s s e s Nested in Position C l a s s e s Figure 2.6 Comparison o f position by decay class at all spruce-dominated sites using contingency tables and chi-square goodness o f fit tests.  2.4 Discussion Abundance and volume o f large woody debris varied among and within the ten headwater streams i n the foothills o f Alberta. The number o f logs at each site ranged from 16 to 48, total volumes o f L W D i n riparian areas ranged from 0.98 to 6.45 m , and 3  in in-stream environments ranged between 0.67 and 4.23 m . Scaled volumes range 3  between 6.62 and 34.2 m /ha in riparian zones and 38.29 and 272.9 m /ha in in-stream 3  3  environments. Most literature on L W D volume reports in m / m , and these ranged 3  between 0.18 and 10.0 m / m , with 75% o f volumes below 4.6 m / m , in studies done 3  3  38  which is reported for the Pacific Northwest (Keller and Swanson, 1979; Lienkaemper and Swanson, 1987; Robison and Beschta, 1990; McHenry et al., 1998; Hogan et al. 1998; Hyatt and Naiman, 2001; Martin and Benda, 2001). For comparison, instream volumes in this study ranged between 0.01 and 0.08 m /m and riparian area volumes ranged 3  between 0.02 and 0.13 m /m. The relatively low volumes at my study sites are explained 3  by the differences between coastal forests and sub-boreal forests. In the forests of the Foothills Model Forest, trees are smaller om diameter and shorter than trees in coastal forests. Within sites, individual logs varied in size, but average volume of logs did not differ among sites, despite differences in species composition. A l l sites were sampled within mature forests such that all riparian forests were at a similar stage of development and thus trees and L W D were of a similar size. For lodgepole pine and spruce riparian forests, both the position and decay classification systems represent progressive changes to L W D in terms of volume, which is driven largely by changes in length due to increasing structural instability over time. In this study, I measured variation in L W D size within sites using three variables: volume, length, and diameter. Significant groupings occurred by position class within both pinedominated and spruce-dominated sites. As expected, variables generally decreased in mean size with increasing position class. Volume at spruce-dominated sites had the clearest trend with three distinct groups where PCI was larger.in volume than PC2, which was larger than PC3 and PC4. Volume and length at pine-dominated sites followed this trend, but were differentiated into two distinct groups where classes 1 and 2 were larger than classes 3 and 4. Diameter in both forest types followed this trend, but there was more overlap between groups and mean diameter was generally less consistent  39  than length and volume. As such, I suggest that changes in volume with respect to position class are more strongly influenced by L W D becoming shorter, than by diameter decreasing over time due to mechanical and biological erosion. By definition, L W D in PC3 must be shorter than position classes 1 and 2 to fit between bankfull widths and rest on the stream bed. These logs originate as either bridges (PC 1) or partial bridges (PC 2) but have become a smaller, broken segment of the previously larger log. Despite significant overlap in volume among all decay classes at lodgepole pinedominated sites, it is evident that L W D in decay classes 1 and 2 are the largest. In general, mean volume decreases as decay class increases, despite much overlap between classes. Similarly, mean length decreases as decay class increases, but diameter did not vary significantly among classes. Variation in L W D volumes among sites was likely due to differences in forest productivity. For example, I found that L W D in DC2 at Site 1 and in DC1 at Site3 were significantly greater in mean volume than the L W D at other sites and in other classes. It must be noted that spatially, these two sites are closer together than any other two sites meaning there are similarly-sized and comparatively large trees at these two sites. Similar to the lodgepole pine-dominated sites, the volume and length L W D at spruce-dominated sites decreased in mean size with increasing decay class. L W D length in spruce forests differentiated into three groups and formed the clearest trend whereby length decreases with increasing decay class. Volume in spruce forests differentiated into two groups following this trend, but mean diameter was not significantly different among decay classes.  40  The biotic (biological decay) and abiotic (mechanical erosion) processes which govern changes in position and decay classes are similar and interact, resulting in increasing structural instability of wood in riparian zones over time. As a result, the position and decay classes are significantly related in both lodgepole pine- and sprucedominated riparian forests (Figures 2.3-2.6). For both species, the abundance of logs among position and decay classes varied among and within sites. The apparent trend among sites is that the fewest logs are found in position and decay classes 1. The greatest number of logs is found in PC3 and PC4 for lodgepole pine-dominated and sprucedominated sites, respectively. Class 3 and 4 logs originate from either bridges (PCI) or partial bridges (PC2) and therefore represent a section of a larger log. Since multiple sections of previously whole logs may be represented in PC3, then abundance of L W D in PC3 would be higher than PCI or PC2. Buried logs (PC4) are essentially PC3 logs which have become embedded in the stream bank by storing sediment and in some cases by changing meander bends. Since PC4 logs are buried by definition, their abundance could have been underestimated in this study as abundance in this case is related to exposure along the current streambank and subsequent excavation effort by the censurers. Similarly, the greatest number of logs in is found in DC3 for pine-dominated sites, and DC4 for spruce-dominated sites. If logs changed classes linearly and at a constant rate, then logs should be evenly distributed among classes, however this is not the case. Since abundance and volume of L W D in small, headwater streams is controlled by recruitment, storage, and transport processes then variation in abundance and volume of logs in different position and decay classes must represent variation in input to and output from each class (Benda et al., 2003). Output is represented by decay and  41  transport; however log transport is negligible in these systems. Therefore output is by decay processes only. Input is governed by tree-fall into the stream which varies through time and is controlled by chronic and episodic disturbances such as windthrow and forest stand dynamics. Further, the decay class or position class into which logs enter is important. A snag may fall into any decay class, depending on the delay between time since death and time since fall which ultimately results in variation in abundance of logs in each of position and decay classes. The results of this study show that a greater abundance of L W D can be found in position and decay classes 3 and 4. These results are also explained by rates of change through classes where wood resides in different classes for different lengths of time. For example, rapid loss of needles and fine branchlets in decay class 1 relative to slower loss of bark and secondary branches in decay class 2. Ultimately, rates of progression through decay classes and position classes will vary among logs depending on species, size, time since death to time since fall, and factors contributing to tree death. From a management, perspective, it is simpler and more cost-effective to classify wood into position classes than into decay classes since wood position can be determined from the floodplain. As a forest manager walks downstream along a stream reach, a census can be conducted fairly rapidly. Since decay classifications are based on the integrity of the wood itself, a manager would have to enter the stream and handle each piece of wood. Further, decay classes were developed for terrestrial coarse woody debris whereby log conditions and position relative to the forest floor were important attributes for understanding wildlife habitat and biodiversity (Maser et al. 1979; Thomas et al. 1979; Daniels et al. 1997; Storaunet and Rolstad 2002, DeLong et al, 2005). In streams  42  and riparian zones, the function of dead wood differs; logs located within a channel will alter flow hydraulics, store sediment, and influence channel morphology (e.g., Swanson and Lienkaemper, 1978; Hogan, 1987; Bisson et al., 1987). In riparian environments, it is more effective to categorize logs based on functional classifications such as position classes. L W D position classes 1 and 2 provide overhead cover, shade, and protection from predators for salmonids; L W D in classes 3 and 4 is integral for stream geomorphology and habitat diversity by forming steps and pools, storing sediment, and providing bank stability (Richmond and Fausch, 1995). The evidence in this study suggests that position and decay classes are linked. Volumes of individual pieces of L W D , both in-stream and within riparian zones are consistent among sites. Changes in volume are related to decay or position class such that as the class number increases, volume decreases, and this is driven largely by changes in mean length from position class to position class and decay class to decay class. Since position classes at all sites combined differentiated into distinct groups with respect to all variables better than decay classes, I propose that the position class model is better able to represent L W D volume. I recommend the use of position classes for studies of large woody debris that aim to understand the interactions between forest dynamics, stream hydrology and geomorphology, and position of wood relative to the streambed. In riparian environments, these are more important attributes for study than stage of decay for habitat and diversity.  43  2.5 Literature Cited Anonymous, 1997. Sigma Stat for Windows release 2.0 standard version.  SPSS Inc.  Chicago, IL, USA. Anonymous, 2004. SPSS for Windows release 12.0.2 standard version.  SPSS Inc.  Chicago, IL,USA. Beckingham, J. D., I. G. W. Corns, and J. H . Archibald. Field Guide to Ecosites of West-Central Alberta. Special Report 9, Canadian Forest Service, Northwest Region. 1996. Benda, L., D. Miller, J. Sias, D. Martin, R. Bilby, C. Veldhuisen, and T. Dunne, 2003. Wood Recruitment Processes and Wood Budgeting. In: The Ecology and management of Wood in World Rivers, S.V. Gregory, K . L . Boyer, and A . M . Gurnell (Editors). American Fisheries Society, Symposium 37, Bethesda, Maryland, pp. 49-74. Berg, N . , A . Carlson, and D. Azuma. 1998. Function and dynamics of woody debris in stream reaches in the central Sierra Nevada, California. Canadian Journal of Forest Resources 32: 1460-1477.  Bilby, R. E. and P. A . Bisson, 1998. Function and Distribution of Large Woody Debris. In: River Ecology and Management, R. J. Naiman and R. E. Bilby (Editors). Springer, New York, 324-346. Bilby, R. E., and J. W. Ward. 1991. Characteristics and function of large woody debris in streams draining old-growth, clear-cut, and second-growth forests in SW Washington. Canadian Journal of Fisheries and Aquatic Sciences 48: 2499-2508  Bisson, P. A., R. E. Bilby, M . D. Bryant, C. A . Dollof, G. B . Grette, R. A . House, M . L. Murphy, K . V . Koski and J. R. Sedell, 1987. Large Woody Debris in Forest Streams in the Pacific Northwest: Past, Present, and Future. In: Stream Management: Forestry and Fisheries Interaction. Contribute 57, E. O. Salo and T. W. Cundy (Editors). Institute of Forest Box, G. E. P. and S. L. Anderson. 1955. Permutation theory in the derivation of robust criteria and the study of departures from assumption. Journal of the Royal Statistical Society B17: 1 -34  Church, M . , 1992. Channel Morphology and Topology. In: The Rivers Handbook, C. Calow and G. Petts (Editors). Blackwell, Oxford, vol. 2, pp. 126-143.  44  Daniels, L . D., J. Dobry, K . Klinka, and M . C. Feller. 1997. Determining year of death of logs and snags of Thuja plicata in southwestern coastal British Columbia. Canadian Journal of Forest Research 27: 1132-1141 Delong, S.C., L.D. Daniels, B. Heemskerk, and K.O. Storaunet. 2005."Temporal development of downed wood habitats in wet spruce-fir stand in east central British Columbia. Canadian Journal of Forest Research 35: 2841-2850. Fausch, K . D. and T. G. Northcote. 1992. Large woody debris and salmonid habitat in a small coastal British Columbia stream. Canadian Journal of Fisheries and Aquatic Sciences 49: 682-693. Faustini, J. M . and J. A . Jones, 2003. Influence of large woody debris on channel morphology and dynamics on steep, boulder-rich mountain streams, Western Cascade, Oregon. Geomorphology 51: 187-206. Gomi, T., R. C. Sidle, M . D. Bryant and R. D. Woodsmith. 2001. The characteristics of woody debris and sediment distribution in headwater systems, southeastern Alaska. Canadian Journal of Forest Research 31: 1386-1399 Harmon, M . E., J. F. Franklin, F. J. Swanson, P. Sollins, S.V. Gregory, J. D. Lattin, Anderson, N . H., S. P. Cline, N . G. Aumen, J. R. Sedell, G. W. Lienkaemper, K . Cromack, and K. W. Cummins, 1986. Ecology of Coarse Woody Debris in Temperate Ecosystems. Advances in Ecological Research 15:133-302. Hassan, M . A., D. L . Hogan, S. A . Bird, C. L . May, T. Gomi, and D. Campbell, 2005. Spatial and temporal dynamics of wood in small streams. Journal of the American Water Resources Association 41: 899-919 Hauer F.R., G.C. Poole, J.T. Gangemi and C V . Baxter, 1999. Large woody debris in bull trout (Salvelinus confluentus) spawning streams of logged and wilderness watersheds in northwest Montana. Canadian Journal of Fisheries and Aquatic Science 56: 915-924. Hogan, D. L . 1987. The influence of large organic debris on channel recovery in the Queen Charlotte Islands, British Columbia, Canada. In: Erosion and sedimentation in the Pacific Rim. Proceedings of the Corvallis Symposium, August, 1987. IAHS Publication No. 165 Hogan, D. L., S. A . Bird, and M . A . Hassan, 1998. Spatial and Temporal Evolution of Small Coastal Gravel-Bed Streams: The Influence of Forest Management on Channel Morphology and Fish Habitat. In: Gravel-Bed Rivers in the Environment, P. C. Klingeman, R. L. Beschta, P. D. Komar, and J. B . Bradley (Editors). Water Resources Publications, L L C , Highland Ranch, Colorado, U S A , pp. 365-392.  45  Hyatt T.L. and R.J. Naiman, 2001. The Residence Time of Large Woody Debris in the Queets River, Washington. Ecological Applications ll(l):191-202 Keller, E. A. and F. J. Swanson, 1979. Effects of large woody organic material on channel form and fluvial processes. Earth Surface Processes and Landforms 4: 361-380. Lienkaemper G.W. and F.J. Swanson, 1987. Dynamics of Large Woody Debris in Streams in Old-Growth Douglas-Fir Forest. Canadian Journal of Forest Research 17:150-156. Martin, D. J. and L. E. Benda. 2001. Patterns of instream wood recruitment and transport at the watershed scale. American Fisheries Society 130: 940-958 Marcus W.A., R.A. Marston, C R . Colvard Jr., and R.D. Gray, 2002. Mapping the spatial and temporal distribution of woody debris in streams of the greater yellowstone ecosystem, U S A . Geomorphology 44:323-335. Maser, C , R. G. Anderson, K . Cromack Jr., J. T. Williams, and R. E. Martin. 1979. Dead and down woody material. In: Wildlife Habitats in Managed Forests - the Blue Mountains of Oregon and Washington. Ed: J. W. Thomas. U S D A For. Ser. Agric. Handb. 553, Washington D.C. p 78-95 McHenry, M . L . , E. Shott, R.H. Conrad, and G.B. Grette, 1998. Changes in the quantity and characteristics of large woody debris in streams of the Olympic Peninsula, Washington, U S A (1982-1993). Canadian Journal of Fisheries and Aquatic Sciences 55:1395-1407. McCleary, R., C. Bambrick, C. Sherburne, and S. Wilson. 2003. Report 2.4.1b. Level 1 classification: GIS - based stream reach characteristics. Fish and Watershed Program, Foothills Model Forest, Hinton Alberta. March 12, 2003. 55 pp. McCleary, R., C. Widk, and J. Blackburn. 2002. Comparison between field and GIS derived descriptors of small streams within the west-central foothills of Alberta. Fish and Watershed Program, Foothills Model Forest, Hinton Alberta. April 2, 2002. 40pp. Murphy, M . L. And K. V . Koski. 1989. Input and depletion of woody debris in Alaska streams and implications for streamside management. North American Journal of Fisheries Management 9: 427-436 Nakamura, F. and F. J. Swanson, 1993. Effects of coarse woody debris on morphology and sediment storage in mountain stream system in western Oregon. Earth Surface Processes and Landforms 18: 43-61 Nakamura, F. and F. J. Swanson, 1994. Distribution of coarse woody debris in a  46  mountain stream, western Cascade Range, Oregon. Canadian Journal of Forest Research 24:2395-2403. Nakamura, F. and F.J. Swanson, 2003. Dynamics of Wood in Rivers in the Context of Ecological Disturbance. In: The Ecology and Management of Wood in World Rivers, S.V. Gregory, K . L. Boyer, and A . M . Gurnell (Editors). American Fisheries Society, Symposium 37, Bethesda, Maryland, pp. 279-298. Richmond, A . D. and K. D. Fausch, 1995. Characteristics and function of large woody debris in subalpine Rocky Mountain streams in northern Colorado. Canadian Journal of Fisheries and Aquatic Sciences 52:1789-1802. Robison E.G. and R.L. Beschta, 1990. Coarse Woody Debris and Channel Morphology Interactions for Undisturbed Streams in Southeast Alaska, U.S.A. Earth Surface Processes and landforms 15: 149-156. Roni, P. and T.P. Quinn. 2001. Density and size of juvenile salmonids in response to placement of large woody debris in western Oregon and Washington streams. Canadian Journal of Fisheries and Aquatic Sciences 58: 282-292 Schwab, J. W., 1998. Landslides on the Queen Charlotte Islands: Processes, Rates, and Climatic Events. In: Carnation Creek and Queen Charlotte Islands Fish/Forestry Workshop: Applying 20 Years of Coastal Research to Management Solutions. D. L. Hogan, P. J. Tschaplinski, and S.Chatwin (Editors). B C Ministry of Forests, Research Branch, Victoria, B C . Land Management Handbook No. 41. Srivastava, A . B. L . 1959. Effects of non-normality on the power of the analysis of variance. Biometrika 46: 114-122 Storaunet, K . O., and J. Rolstad. 2002. Time since death and fall of Norway spruce logs in old-growth and selectively cut boreal forest. Canadian Journal of Forest Research 32: 1801-1812 Swanson F. J. and G. W. Lienkaemper, 1978. Physical Consequences of Large Organic Debris in Pacific Northwest Streams. U S D A Forest Service, General Technical Report GTR-PNW-69. Thomas, J. W., R. G. Anderson, C. Maser, and E. L. Bull. 1979. Snags. In: Wildlife Habitats in Managed Forests - the Blue Mountains of Oregon and Washington. Ed: J. W. Thomas. U S D A For. Ser. Agric. Handb. 553, Washington D.C. p 6-77 Tiku, M . L. 1971. Power function of the F-test. Journal of the American Statistical Association 62: 525-539 Webster, J. R., E. F. Benfield, T.P. Ehrman, M . A . Schaeffel, J. L . Tank, J. Hutchens,  47  and D. J. D'Angelo. 1999. What happened to allochthonous material that falls into streams? A synthesis of new and published information from Coweeta. Freshwater Biology 41: 687-705  Young, M . K., E. A . Mace, E. T. Ziegler, and E. K . Sutherland. 2006. Characterizing and contrasting instream and riparian coarse wood in western Montana basins. Forest Ecology and Management 226: 26-40 1  Zar, J. H . 1984. Biostatistical Analysis. 2 USA  nd  ed. Prentice Hall, Engelwood Cliffs, NJ,  48  3. A DENDROECOLOGICAL ANALYSIS OF L A R G E WOODY DEBRIS IN SMALL STREAMS IN T H E FOOTHILLS MODEL FOREST, HINTON, ALBERTA  3.1 Introduction Large woody debris (LWD) is biologically and geomorphically integral to small stream riparian environments (Harmon et al, 1986; Bisson et al, 1987). Wood in small streams provides channel structure and stability, specifically in the creation of step-pools and plunge-pools (Swansen and Lienkamper, 1978; Harmon et al, 1986; Bisson et al, 1987; Montgomery et al, 1995). L W D is integral to fish habitat and for supporting fish food webs in small streams; hence, the recruitment of L W D is a critical component of headwater stream ecology (Hogan, 1987; Martin and Benda, 2001). Therefore, understanding processes that generate L W D and decomposition or loss of L W D is important for conservation of biodiversity and sustainable forest management. Much research of wood abundance has been done in the Pacific Northwest region (PNW) of North America, where stream channels are often coupled with steep hillslopes and where mass wasting, bank erosion, landslides and debris flows during large storm events and snow avalanches play a dominant role in wood supply to streams. In contrast to the PNW, headwater streams of the eastern foothills of the Rocky Mountains of Alberta are uncoupled to hillslopes as the landscape is more subdued and hillslope geomorphology processes as contributors to L W D generation, such as landslides and debris flows, are less important. In these landscapes, I hypothesize L W D is created by fine-scale disturbances such as wind, bank erosion, and insect disturbance, and post-fire stand development processes. However, we lack knowledge of the specific stand dynamics mechanisms that creates woody debris (e.g., Keller and Swanson, 1979;  49  Schwab, 1998; Nakamura and Swansen, 1993; Hogan et al, 1998, Hassan 2005). Temporal processes of large woody debris such as L W D recruitment, residence time, and decay rate have rarely been quantified at high levels of resolution. To date, little is known about recruitment or residency of wood in small streams both in general and in the F M F in particular.  3.1.1 Research  Objectives  M y research goals are to understand the link between natural disturbance, forest dynamics, and the recruitment and residency of large woody debris to small, first order streams of the Foothills Model Forest. Understanding these links will help us to identify historical trends in recruitment and residency of in-stream wood. Specifically, my research objectives are threefold: 1.  Determine ages of in-stream large woody debris using tree-ring methods.  2.  Compare ages of L W D to the age structure of canopy trees to determine how disturbance and stand dynamics contribute to the recruitment of wood into streams.  3.  Quantify rates of decay and residence times by comparing the relationship of wood decomposition to wood age in riparian zones.  To meet these objectives I used dendrochronological techniques. B y examining tree ring-width patterns I reconstructed forest development and determined the year of death of trees which had fallen into the streams. I then used the mortality data to investigate the linkages between L W D generation and riparian forest disturbance history. Using classification systems to represent decreasing structural integrity and decay of  50  wood, I assigned each L W D to one o f four categories that demarcate key decay and structural attributes in order to quantify decay rates and residence times, thus linking wood decomposition to wood age.  3.2 Methods 3.2.1 Study Area I conducted this research i n the Upper Foothills Natural Subregion in the foothills of the R o c k y Mountains i n west-central Alberta (Figure 3.1, Beckingham et al., 1996). A l l study sites were located between 1041 and 1384 meters above sea level (masl) in the Hinton W o o d Products Forest Management Area o f the Foothills M o d e l Forest. Upland forests within this subregion are characterized by closed-canopy coniferous forests, most commonly including lodgepole pine (Pinus contorta Dougl. E x Loud), as well as white spruce (Picea glauca (Moench) Voss) and black spruce (Picea mariana (Mill.) B.S.P.). Frequent, stand initiating fires o f an historical average fire size o f 300 ha, with a mean return interval o f 100 years, have created a landscape mosaic o f successional lodgepole pine, mixed stands with lodgepole pine and white spruce, or mixed white and black spruce (Beckingham et al, 1996; McLeary, 2003). The study area encompasses the Forest Management Agreement Area managed by Hinton W o o d Products, a division o f West Fraser M i l l s Ltd. This area covers 10,000 k m and includes 20,000 k m o f mapped stream channels within the Athabasca and North 2  Saskatchewan River watersheds. The channels are classified into reaches averaging 300 m i n length with consistent stream order, drainage area, channel slope, and forest cover (McLeary et al, 2003) Study sites are found i n drainage networks within the Erith,  51  Eunice, Moberly, Wroe, Fish, Wigwam, and Anderson watersheds which contribute to the larger Athabasca watershed. In this landscape, headwater streams can be defined as small (bankfull width <5.5 m) low-angle (<4% slope) channels that include first- and second-order streams where L W D is large relative to channel depth, resulting in a stable step-pool morphology (Church, 1992; Gomi at al, 2001; McLeary et al, 2003; Hassan et al, 2005).  52  W e s t Fraser M i l l s L t d F M A A  ^  Figure 3.1 Location o f study sites 1 through 10 in the West Fraser M i l l s Ltd. Forest Management Area Allocation ( F M A A ) o f the Foothills Model Forest, Alberta.  3.2.2 Site Selection Ten study areas including headwater streams were stratified according to the dominant tree species in the riparian forest surrounding each stream (Figure 1, Table 1). Headwater streams were identified using a geographic information system (GIS) and  53  forest cover and stream inventory data for the Foothills Model Forest (McCleary et al, 2002). These databases were stratified to identify small streams and riparian forests with the following attributes: (1) The stream drainage basin measured <10km . (2) Stream width at bankfull did not exceed 5.5m. (3) Streams were not confined by hillslopes. (4) There was no evidence of log jams dr movement of wood from upstream sources. (5) The surrounding riparian forests were mature (age > 100 years) and dominated either by lodgepole pine, white spruce, or a mixture of black and white spruce. (6) There was no evidence of harvesting of the riparian forest within two treelengths of the stream. Criteria 1 to 4 ensured that wood recruitment was more likely related to standlevel dynamics and fine-scale disturbance as opposed to debris torrents, landslides, or other geomorphic processes. They also controlled for wood transport. If wood is not transported downstream during peak flow events, then L W D decays in situ. In absence of wood transport, the wood found in the stream is directly associated with the adjacent riparian zone and an analysis of this wood can be linked to the spatio-temporal dynamics of the surrounding forest. Criterion 5 allowed comparison of L W D dynamics among and between mature pine- and spruce-dominated forests. Criterion 6 ensured that the study sites represented natural forest dynamics rather than modern anthropogenic effects such as timber harvest (Young et al, 2006).  54  3.2.2 Field  Sampling  3.2.2.1 Sampling of Large Woody Debris  Each study site consisted of a 50 m reach of stream that included several step-pool geomorphic units to allow comparison to other published studies of stream geomorphology and L W D (e.g., Fausch and Northcote, 1992; Nakamura and Swansen, 1994; Berg et al, 1998; Potts and Anderson, 1990). The start points for the stream reaches were randomly determined. Sampling of L W D moved upstream from the start point, along the thalweg. Stream width at bankfull was measured every five meters from 0 to 50 m to ensure our second criterion of site selection was fulfilled. Two classification systems were used to represent decreasing structural stability and decay of wood over time. Each classification system was assigned four categories so that I could directly compare the two systems with respect to time since tree death. Thus, each piece of wood was assigned a decay class and a position class. The decay "class system was adapted from research of terrestrial coarse woody debris by Maser et al. (1979) and was used to classify wood based on wood structural characteristics and integrity (Table 1). The four categories of each decay class include the following attributes: Decay Class 1 (DC1) wood has > 75% bark still intact, bark adheres tightly; branches have fine (3 order) branchlets; sapwood is sound, and log retains structural rd  integrity (DC2) L W D has 25% - 7 5 % bark intact which, in places, is loosely attached to the bole; 1 order branches have a solid connection to the bole; wood is solid with st  evidence of decay on outer sections of sapwood only, (DC3) wood has 0% to 25% bark present, adhering loosely to the sapwood; 1 order branches and branch nubs are present st  and sit loosely in the bole; along some parts of the bole, wood shows significant signs of  55  decay to depths of 5 to 10 cms, (DC4) bark is no longer attached; branch nubs only are present; along some parts of the bole, wood is soft, crumbly or fibrous, and decay can penetrate nearly through the sapwood. The second classification system is used to highlight the relative position, and thereby function, of wood in a stream. Position classes have been used in studies of L W D in aquatic systems by Hauer et al, (1999), Berg et al, (1998), Nakamura and Swansen (1994) and Richmond and Fausch (1995) and describe the position of wood relative to the channel. In the position classification system I used, L W D was divided into four position classes: Position Class 1 (PCI) Bridge (BR) - log spans channel, touching both banks and resting on the floodplain, (PC2) Partial bridge (PB) - spanned log has broken in one or more places within the stream channel, (PC3) Loose (L) - log is no longer associated with the floodplain and is now associated with the streambed; log is fully submerged during bankfull, (PC4) Buried (B) - log has become incorporated into the streambed or the sides of the streambank and has stored sediment and become at least partially buried. With respect to function of L W D in the stream, L W D PCI and PC2 provide overhead cover, shade, and protection from predators for salmonids (Richmond and Fausch, 1995). L W D in PC3 and PC4 is integral for stream geomorphology and habitat diversity by forming steps and pools, storing sediment, and providing bank stability (Richmond and Fausch, 1995). At each study site, a subset of L W D was sampled by cutting cross-sectional disks to estimate the year of death using dendroecological methods. Moving upstream from a random start point, I systematically sampled the first five L W D in each of the four position classes. Twenty samples were taken per site, except site 6 where there was only  56  1  one bridged L W D available to be sampled. Each cross-sectional disk was sampled from the bole using either a chainsaw or hand saw. Before cutting, the sampled section was bound tightly with duct tape so that the section would maintain its.structural integrity during cutting, transportation and processing (B. Luckman, Pers comm.). When possible, the disks were taken 30 cm from the roots to include a long ring-width series and to estimate tree age when it died. 3.2.2.2 Sampling Canopy Dominant Trees  To estimate tree ages and to build species- and site-specific tree-ring chronologies, increment cores were extracted from living canopy-dominant trees at all sites following standard dendrochronological protocols (Josza, 1988). Trees were sampled on either side of the stream, and within three tree lengths from the stream. At spruce-dominated sites, 20 black and 20 white spruces were cored at 30 cm above the estimated point of germination, the point at which shoots and roots diverge. At lodgepole pine-dominant sites, 30 lodgepole pine trees were cored 30 cm above the germination point. At sites 2 and 5, which were dominated by lodgepole pine, white spruce was present in the canopy and 6 and 13 trees were cored, respectively. Similarly, site 8 was dominated by spruce, but lodgepole pine was present in the canopy and 9 individuals were cored. Some canopy dominant trees were cored more than once in an attempt to include the pith of the tree in the sample core or to avoid decayed heartwood. For each tree I recorded species, diameter at breast height (dbh, 1.3 m above germination point) and diameter at core height (dch, 30 cm above germination point). To estimate the number of years required for the trees to grow to coring height, six to twelve seedlings of each species were sampled at each site. Six lodgepole pine  57  seedlings were sampled from clearings adjacent to sites 1 to 5. Seedlings were sampled in clearings as opposed to the understory of the stand in order to represent open growing conditions similar to those experienced by the canopy dominants when they established. Most white and black spruce trees had narrow rings close to the pith; therefore, seedlings were sampled from the understory of the stands. Not all white and black sprucedominated sites had both white and black spruce seedlings. Site 6 had only black spruce seedlings present, and site 10 had only white spruce seedlings. I selected seedlings germinated from seed rather than those that had reproduced asexually.  3.2.3 Laboratory  Methods  3.2.3.1 Chronology  Development  A l l increment cores were mounted and dried on wooden supports and sanded with a belt sander using successively finer sand paper to 400 grit (Stokes and Smiley, 1968). Rings were visually crossdated and the tree was assigned a calendar year for the innermost ring or the pith (Swetnam et al., 1985). Ring widths were measured to the nearest 0.001 mm using a Velmex bench interfaced with a computer. To ensure that calendar years were accurately assigned to each ring, the resulting ring-width series were statistically crossdated using the program C O F E C H A (Holmes, 1986, Grissino^Mayer 2001). C O F E C H A compares each tree-ring series against all other series to identify errors in tree ring dates, specifically to identify missing or false rings. Missing rings occur when a tree-ring does not form due to poor growing conditions or i f ring formation is not complete around the circumference of the tree and the ring is not present in the section of the tree sampled (Phipps, 1985). Alternately, false rings occur when growth  58  rates decrease during the growing season and cell formation transitions from earlywood to latewood cells. If environmental conditions become favourable to growth after formation of these latewood cells, then earlywood formation can reinitiate, causing an apparent false tree-ring (Phipps, 1985). Properly dated and highly correlated ring-width series are combined into speciesand site-specific master ring-width chronologies. Two types of standardization were applied using the computer program A R S T A N (Cook, 1985). (1) Horizontal standardization was used to assess the dynamics of each stand. Using the horizontal standardization method, each ring-width series was normalized to show departures from the mean growth rate and all series were averaged to represent long-term and short-term trends in tree growth at the site. (2) To compare chronologies among stands of different age, ring-width series were standardized with a negative exponential curve or sloping line to remove the age- or size-related growth trend. The detrended ring-width series were then averaged to produce a standard chronology of dimensionless tree-ring indices. Regional chronologies of lodgepole pine, white spruce, and black spruce were constructed by comparing the crossdated ring-width series from all sites. The regional chronologies included the most highly inter-correlated series from individual sites. Similar to the site-specific chronologies, the regional chronologies were standardized to remove age- or size-related trends and to produce standard chronologies. I used Pearson product moment correlation analysis to compare the site-specific and regional chronologies for each species. Correlation matrices were calculated using the standard chronologies for the period common to all chronologies.  59  To assess the age structure of canopy trees growing in the riparian forest, tree ages were calculated using the following equation (Wong and Lertzman, 2001): Age ai = Age tot  core height  ^estimate of missing/false rings ^estimate of years to pith ' K  [l]  Where: ^estimate of missing/false rings  =  the estimated number of false and missing rings found by  crossdating procedures. ^estimate of years to pith  = the estimated number of years from the last full measured  ring to the pith. i  h = number of years required to reach core height c  In this study,  estimate of missing/fakeringswas  zero since only cores that were properly  crossdated against site- or regional chronologies were included in the analysis. For cores that intercepted the pith,  ^estimate of years to pith  was also zero and age was calculated using the  following equation: Age  c o r e  height  = 2003 - pith year + 1  [2]  For tree cores that do not contain the pith, Duncan's (1989) method was used to approximate the number of missing rings using the following equation: Number of rings to pith = e stimate of years to pith e  =  rL /8H1 + H/2 X 2  [3 ]  Where: L = length of the innermost incomplete ring on the core H = height of the innermost incomplete ring on the core X = mean ring width of three rings adjacent to the innermost incomplete ring  60  This equation estimates the number of rings from the last complete ring observed on the increment core to the pith. The calculated number of missed rings was added to the number of rings visible to estimate the age of the tree at core height. A l l tree ages were corrected to account for the number of years required for the tree to grow to the core height of 30 cm above the ground. Seedlings were used to estimate a height correction factor (h ), the number of years it takes a seedling to grow to c  30 cm or coring height (Veblen et al., 1991). The difference in the number of rings between the root collar and coring height above the ground was calculated for all sampled seedlings. I calculated the average years to coring height for each species at each site. For white spruce, I used one-way A N O V A and Tukey's multiple comparisons method to test for significant differences in mean age among sites. Seedling ages of lodgepole pine and black spruce were not normally distributed so I used a non-parametric Kruskal Wallace one-way A N O V A on ranks and Dunn's multiple comparisons method to test for differences in mean age. Sites that were not significantly different were grouped and one age correction was calculated for each species. Once Age ai was estimated for each increment core, these data were used to tot  create age frequency histograms for each site in order to interpret stand dynamics.  3.2.3.2 Crossdating Large Woody Debris  Disks from large woody debris in decay classes 1 and 2 were air dried and then sanded following standard dendroecological protocols (Stokes and Smiley, 1968). For wood samples that were at more advanced stages of decay (classes 3 and 4), the samples  61  were sanded to 400 grit while they were frozen. After sanding, these disks were air dried and some were re-sanded, to ensure that ring boundaries were clearly visible. L W D from pines versus spruce were distinguished based on wood anatomy. Pines have large-diameter resin canals, thin-walled epithelial cells, and earlywood makes an abrupt transition to latewood (Hoadley, 1933; I. Hartley, Pers. comm.). Conversely, spruces have relatively small resin canals and earlywood makes a gradual transition to latewood (Hoadley, 1933; I. Hartley, Pers. comm.). Spruces at the species level are indistinguishable with regards to wood anatomy alone. Therefore, I used the results from statistical crossdating to differentiate white and black spruce.. The ring-width series from two radii on each piece of woody debris were measured and statistically crossdated against the species- and site-specific standard chronologies to determine the calendar year of the outermost ring of the wood sample and the pith date. L W D from spruce were crossdated against both white spruce and black spruce chronologies. The regional chronologies were used to crossdate L W D at sites where there were few canopy trees of the same species in the riparian forest so that a sitespecific chronology had not been developed. Using C O F E C H A , 50-year segments lagged by 20 years were correlated against the standard chronologies. The calendar years corresponding to the highest correlations for the largest number of 50-year segments were assigned to each ring-width series. For samples with eroded rings on part of the circumference, the two ring-width series yielded different outer-ring dates which corresponded to the same pith date. For these samples, the most recent outer-ring date represented the year of death. Once an outer-ring date was determined using C O F E C H A , the ring-width series was plotted against the appropriate site- and species-specific  62  standard chronology to visually verify the marker rings were properly aligned and growth trends o f the individual tree matched those o f the surrounding canopy trees. The statistical crossdating method assumes that the outermost ring o f the wood sample is the last year that formed when the tree was alive and is considered a best estimate o f the year o f tree death (Daniels et al, 1997). For L W D samples with bark attached to the sapwood or with bark beetle galleries i n the cambium or with intact" sapwood, the outermost ring dates had a higher probability o f representing the actual year of death. When rings are lost from a sample due to erosion or decay, the crossdating results do not accurately depict the year o f tree death. For eroded or decayed samples, the year o f tree death under-estimated and the number o f years since death is over-estimated. To evaluate the quality o f our estimates o f years o f death, each wood sample was assessed. Samples with bark, beetle galleries and/or intact sapwood were classified as "high-quality" (HQ) estimates o f year o f death; samples without one o f these features were classified as "best-estimates" ( B E ) o f year o f death. For each species, I tested time since death o f L W D among position classes and decay classes for normality and equal variances, which are primary assumptions o f A N O V A (Zar, 1984). I tested for differences i n time since death among position classes and decay classes using a one-way analysis o f variance ( A N O V A ) for lodgepole pine and two-way A N O V A for black and white spruce (Anonymous, 1997). For factors which were significant, I conducted a Tukey's pairwise multiple comparison to differentiate among classes. A l l statistical tests were considered significant at the  = 0.05 level.  63  3.3 Results 3.3.1 Ring-width chronologies I developed fourteen new species- and site-specific chronologies and three new species-specific regional chronologies for lodgepole pine, white spruce and black spruce in the Foothills Model Forest (Table 3.2). A total o f 107 lodgepole pine increment cores were used to create five stand-level master chronologies. Each o f the five site-specific lodgepole pine chronologies included increment cores from 16 to 27 trees and ranged, from 92 to 135 years in length. The regional lodgepole pine chronology included cores from each o f the five sites for a total o f 55 cores and extended from 1856 to 2004 (Table 3.2, Figure 3.2). The chronologies for white and black spruce were longer than the chronologies for lodgepole pine (Table 3.2). For white spruce, four o f five site-specific chronologies began in the m i d - to late-1700s; at site 8, the chronology started in 1895. The site-specific white spruce chronologies included 10 to 19 cores. The regional chronology included 53 cores and extended from 1764 to 2004. Four black spruce sitespecific chronologies included 8 to 18 cores. Three chronologies began i n the late- 1700s and early- 1800s; at site 8, canopy dominants were younger and the chronology started in 1888. The black spruce regional chronology included 50 cores from all four sites and extended from 1806 to 2004. Summary statistics for the site-level and regional chronologies indicated common variation to environmental factors among trees within each study site, as well as common variation through time (Table 3.2). The series inter correlation (IC) quantifies the amount o f common variation among tree-ring series within a chronology (Fritts 1976). IC values ranged from 0.39 to 0.61, with the majority o f values between 0.50 and 0.56 for all three  64  species at the site and regional spatial scales. These values indicated'that tree growth responded simultaneously to coarse-scale climate as well as stand-level disturbance agents. Mean sensitivity (MS) measures ring-width variability between successive years within a chronology (Fritts 1976). M S values ranged between 0.10 and 0.14 which indicated relatively low variation in ring widths from one year to the next (Fritts 1976). Standard deviations (SD) ranged from 0.16 to 0.38, indicating more variation in ring width over the lifespan of the trees relative to the M S values. First-order autocorrelation (AC(1)) expresses the statistical association of each value in a time series with the previous value and these values ranged from 0.61 to 0.96 (Fritts, 1976).  65  Table 3.1 Summary statistics for site-specific and regional standard chronologies for lodgepole pine (Pine), white spruce (WS), and black spruce (BS) for the Foothills Model Forest. Site Name 1 2 3 4 5 6 7 8 9 10 6 7 8 9 Regional Regional Regional 1  2  3  4  Species Pine Pine Pine Pine Pine WS WS WS WS WS BS BS BS BS Pine WS BS  Plot Location Longitude Latitude 53°0'N 122°53'W 53°43'N 123°52'W 53°9'N 122°52'W 123°13'W 53°7'N 124°3'W 53°33'N 124°39'W 53°29'N, 53°16'N, 124°33'W 53°17'N, 123°22'W 53°16'N, 124°31'W 53°28'N, 124°28'W 53°29'N, 124°39'W 124°33'W 53°16'N, 53°17'N, 123°22'W 124°31'W 53°16'N, X X X X X X  Elevation (masl) 1196 1368 1041 1049 1384 1320 1413 1309 1369 1298 1320 1413 1309 1369 X X ' X  No. Trees 18 22 16 24 27 15 16 12 10 19 15 18 16 8 55 - 53 50  Mean Series Length (years) 92 102 99 100 135 171 183 92 175 141 138 141 100 155 149 240 201  Chronology Start End i.e.1 2002 0.541 1905 1893 2003 . 0.568 " 0.394 1901 2003 2003 0.505 1896 2004 0.608 1856 2003 0.540 1775 1764 2003 ' 0.581 0.562 1895 2003 1758 2003 0.471 2002 0.504 1786 0.483 1785 2003 2003 0.550 1786 •0.527 1888 2003 1806 2003 0.539 0.522 2004 1856 1764 . 2003 0.535 2004 0.509 1806  Series Statistics S.D. M.S. 0.13 0.26 0.14 0.26 0.10 0.16 0.11 0.18 0.12 0.17 0.11 0.19 0.12 0.18 0.12 0.22 0.11 0.28 0.12 0.38 0.14 0.23 0.14 0.19 0.12 0.20 0.14 0.32 0.16 0.11 0.10 0.18 0.12 0.19  AC(1) 0.80 0.80 0.65 0.67 0.61 0.74 0.72 0.80 0.87 0.96 0.78 0.67 0.73 0.87 0.61 0.78 0.79  4  IC Intercorrelation quantifies the amount of common variation among tree-ring series within a chronology (Fritts, 1976). M S Mean sensitivity expresses mean percent change from one year to the next (Fritts, 1976). SD Standard deviation expresses the square root of the variance of a population (Zar, 1984). AC(1) First-order autocorrelation expresses the statistical association of each value in a time series with the previous value (Fritts,  1976).  The correlations among the site- and regional-chronologies allow comparison of tree growth at a broad spatial scale. For all three species, the site-specific and regional standard chronologies were significantly correlated (Tables 3.3, 3.4 and 3.5), except the lodgepole pine chronologies for sites 1 and 5 (Table 3.2). This discrepancy may be explained, in part, by the large geographical distance between sites 1 and 5 (Figure 3.1). However, the standard chronology for site 5 had relatively low correlations to all other site-specific chronologies. For sites with low correlations, trees may be responding to climate differently or within-stand variation may influence the chronology more than the regional climate. The mean series length of the chronology for site 5 was 135 years, about 35 years longer than the other lodgepole pine chronologies. This difference in stand age may have contributed to the low correlations. Similarly, the standard chronology for black spruce at site 9 had relatively low correlations to all other site-specific chronologies, although the correlations were statistically significant. The black spruce chronology at this site included only 8 cores and the rings near the pith were wide, which is a departure from the growth patterns of black spruce at other sites that had narrow rings near the pith. Table 3.2 Correlation matrix comparing site-specific and regional standard chronologies for lodgepole pine in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1907 to 2002 (n = 96 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when ° c = 0.05. 1 1 2 3 4 .5  Regional  <0.01 <0.01 <0.01  2 0.74  <0.01 <0.01 0.13 <0.01 <0.0T <0.01  3 0.63 0.69  <0.01 <0.01 <0.01  4 0.50 0.56 0.57  <0.01 <0.01  5 0.16  6.26 0.39 0.39  Regional 0.70 0.81 0.73 0.73 0.70  <0.01  67  Table 3.3 Correlation matrix comparing site-specific and regional standard chronologies for white spruce in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1897 to 2002 (n = 106 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when ° c = 0.05. . Site 0  —  t  6 7 8 9 10 Regional  Site  - — 0.60  <0.01 <0.01 <0.01 <0.01 <0.01  <0.01 <0.01 <0.01 <0.01  8  9  10  Regional  0.60 0.48  0.57 0.61 0.78  0.59 0.58 0.61 0.68  0.64 0.76 0.67 0.78 0.92  <0.01 <0.01 <0.01  <0.01 <0.01  <0.01  Table 3.4 Correlation matrix comparing site-specific and regional standard chronologies for black spruce in the Foothills Model Forest. Correlation coefficients (upper right) were calculated for 1890 to 2002 (n = 114 years), the period common to all chronologies. P-values (lower left) in bold font indicate statistically significant correlations when o c = 0.05. Site 6 7 8 9 Regional  6  7 0.529  <0.01 <0.01 <0.01 <0.01  <0.01 <0.01 <0.01  Site 8 0.458 0.422  <0.01 <0.01  9  Regional  0.329 0.277 0.363  0.73 0.699 0.698 0.671  <0.01  The white spruce, black spruce, and lodgepole pine regional, standard chronologies were significantly correlated between 1856 and 2003, the period common to all three chronologies (Figure 3.2). The correlation between white and black spruce (r = 0.80, p<0.01) was greater than the correlations between lodgepole pine and the spruces (r = 0.15, p<0.01 and r = 0.16, p<0.01 for white and black spruce, respectively). Evidently, white and black spruce responded similarly to climate and environmental variation at the regional spatial scale. The lodgepole pine regional chronology shows a different species response pattern to variables which include stand dynamics and interactions. The lodgepole pine regional chronology starts with rings at the pith wider than the tree-ring  68  index of 1.0, whereas the white and black spruce chronologies generally had narrow rings near the pith with tree-ring indices <1.0. Overall, the trends in ring variation were similar, except between the 1940s and the early 1950s, when the lodgepole pine ringwidth indices decreased then increased, while the spruce indices decreased gradually.  69  1750  1800  1850  1900  1950  2000  Calendar Year White Spruce Black Spruce Lodgepole Pine  Figure 3.2 Standardized white spruce (a), black spruce (b), and lodgepole pine (c) regional chronologies. Sample depth through time is given for each chronology (d).  70  3.3.2 Canopy Tree Ages in the Riparian Forest •  1  3.3.2.1 Correction for Missed Piths ("estimate f years to pith) 0  A total of 296 crossdated cores were used to contribute to age data at ten sites. (Table 3.6). Fifty-eight cores intercepted the pith and the number of rings to the pith was estimated for 238 trees. Corrections for missed piths ranged from 1 to 21 years at the lodgepole pine-dominated sites, with mean correction factors between 3.3 and 6.3 years. The mean correction factors for spruce-dominated sites were higher and ranged between 4.2 and 20 years. The difference in correction factors was due to the difference in initial growth patterns observed among the species. Compared to lodgepole pine, white and black spruce had slow initial growth and narrow rings so that cores that did not intercept the pith had missed a relatively large number of rings. Table 3.5 Number of cores which intercepted the pith and which were corrected using Duncan's (1989) method. The correction factor statistics for cores corrected by Duncan's (1989) method, including average and range of years added to cores at each site. Site Species Total Pith Corrected Correction Factor No. Cores Intercepted No. (Years) No. No. Mean Range 4.5 13 1-10 1 Pine 18 5 4.7 11 22 2-10 2 Pine 33 1-12 3.3 Pine 9 20 3 29 2-21 6.3 4 3 25 Pine 28 6.0 11 • 29 1-20 5 Pine 40 4-41 18.1 1 18 6 WS 19 2-44 15.1 18 2 16 BS 14 2-38 11.6 7 WS 15 1 1 3 0 .11.7 BS 17 5 12 2 2-16 WS 16 . 7.1 8 18 4.2 4 13 1 - 11 BS 17 2 7 1-8 4.3 LP 9 16.6 1 12 3-58 9 WS 13 8.6 1 7 2-38 . BS 8 20.0 14 0 14 2-35 10 WS Total 296 58 238  71  3.3.2.2 Seedling Age to Core Height Analysis  (h ) c  The age correction for coring height of 30cm was less for lodgepole pine than for black and white spruce (Tables 3.7 to 3.9). Median age of lodgepole pine seedlings at 30cm tall ranged from 3 to 10 years and varied significantly among sites (p = 0.004). Seedlings at site 5 were significantly older than seedlings at site 3 (Table 3.7). Tree rings near the pith of seedlings from site 5 were narrow indicating slow initial growth, which was not representative of the conditions under which the canopy trees established. Therefore, seedlings at site five were not used to calculate the average seedling age at 30cm tall. Of the remaining four sites, 32 samples contributed to an average age correction of 5 ± 1.6 (mean ± standard deviation) years to core height. Table 3.6 Comparison of sample number, mean, standard deviation, and median of pine seedling ages in years at five lodgepole pine-dominated sites. N mean standard deviation median Site# 1.4 3.5 * 1 6 3.8 ab 2 5.4 1.1 7 0.4 • 3 3 7 3.1 ab 4 12 5.3 1.7 6.4 5 7 9.7 4 32 1 to 4 5 1.6 fl  5  a  5  Age of white spruce seedlings at 30cm tall ranged from 2 to 20 years and varied significantly among sites (p = 0.001). Seedlings at site 8 were significantly younger than seedlings at sites 7, 9, and 10. Thus, data from site 8 were excluded from the calculation of the average age correction for white spruce. Mean seedling age at 30cm tall for sites 7, 9 and 10 was 12 + 4.1.  72  Table 3.7 Comparison o f sample number, mean, and standard deviation o f white spruce seedling ages i n years at four white and black-spruce dominated sites. Site # N mean standard deviation 7 8 9 10 7,9 and 10  9 8 9 7 25  12.6 6.1* 12.1° 12* 12  3.3 2.6 2.4 4.7 4.2  a  .  Black spruce seedlings were not significantly different was between sites (p = 0.16). Table 3.9 shows variation amongst medians between sites but this variation is still within one standard deviation (sd = 5.7) thus, there was not a statistically significant difference i n median values between the study sites. A g e o f black spruce seedlings at 30 cm tall ranged from 4 to 23 years and average age across all four sites was equal to 12 ± 5.7 years. Table 3.8 Comparison o f sample number, mean, and standard deviation o f black spruce seedling ages i n years at four white and black-spruce dominated sites. Site # TV mean standard deviation median 6 7 8 9 6 to 9  6 7 7 7 27  15.5 10.1 13.7 10.6 12  7.4 5.2 4.6 5.3 5.8  15 10 13 10 11  3.3.2.3 Canopy Tree Ages and Growth Rates Tree ages and age structures varied among species and sites. A t sites 1 to 5, lodgepole pine established between 1850 and 1905 and formed distinct cohorts that spanned three to four twenty-year age classes (Figure 3.3). W i d e rings near the pith indicated these trees established i n relatively open sites, with little inter-tree competition. Approximately 40-50 years after stand initiation, ring widths narrowed and indices decrease to values <1.0. White spruce was present i n the overstory at sites 2 and 5  73  (Figure 3.3). Trees at site 2 established in 1890, 1894, 1936, 1949 and 1967; those at site five (n = 11) established from 1855 to 1885.  74  .10 -  25 >. o  20  OJ 3 CT CD  t5  Site 1  at  c  LL  (n = 20)  10 5 0 • 30 25 -  a c  OJ Z3 CX QJ  LL  Site 2  bi  20 -  (n = 21)  15 10 5 0 30 •  Site 3 25 • • [ ::> =3 CT  £  LL  20  (n = 20)  15 • 10 -  25  Site 4 H d/  (n = 19)  30 25  5  20 -  0) =>  15 •  c  cr  CD LL  Site 5 e/  e/V  (n = 21)  TT  10 •  QLL  5 -  1850  1875  1900  1925  1950  1975  2000  1850  - Lodgepole Pine Sample Depth n Lodgepole Pine Age Data n White Spruce Age Data  1875  1900  1925  1950  1975  2000  Calendar Year  Calendar Year ^ — T V  Lodgpole Pine Standard Chronology Tree Death - High Quality Tree Death - Best Estimate Index = 1  Figure 3.3 Age structure (ai to ei) and standard chronologies (aii to en) for lodgepole pine-dominated sites. A l l age data were statistically and visually crossdated to the standard chronologies. Sample depth curves represent the sample sizes of the standard chronologies. Chronologies were horizontally standardized to remove growth-related trends. Tree death year was estimated from basal disks by correlation to the standardized chronologies.  75  Black spruce and white spruce ages were more variable within and between sites. Four of five spruce sites (6, 7, 9, and 10) were uneven aged with no distinct cohorts. Canopy trees established between 1736 and 1924, spanning up to twelve 20-year age classes at individual sites (Figure 3.4). At site 8, the canopy included black and white spruce and lodgepole pine. A l l three species formed a distinct cohort that established between 1880 and 1910, likely following a stand-replacing disturbance such as fire. At site 8, black spruce, white spruce, and lodgepole pine had narrow rings at the pith, wider rings during the middle portion of growth, and narrowing towards the bark. \  76  2.5  Site 6  ai  3-  2.0  BS (n = 9) WS (n = 7)  if  BS (n = 7) WS (n = 4)  LP (n = 9) BS (n = 2) WS (n = 4)  LP (n = 7) BS (n = 2) WS (n = 9)  n n Site 10  el  BS (n = 2) WS (n = 16)  . 10  n 1750  nHnn n n  1800  1850 1900 Calendar Year  1950  Black Spruce Sample Depth White Spruce Sample Depth • • • Black Spruce Age Data i i White Spruce Age Data i i Lodgepole Age Data  2000  1750  1800  • v  1850 1900 Calendar Year  1950  2000  Black Spruce Standard Chronology White Spruce Standard Chronology Tree Death - Ugh Quality Estimate Tree Death - Best Estimate Index = 1  Figure 3.4 Age structure (ai to ei) and standard chronologies (aii to eii) for sprucedominated sites. A l l age data were statistically and visually crossdated to the standard chronologies. Sample depth curves represent the sample sizes of the standard chronologies. Chronologies were horizontally standardized to remove growth-related trends. Tree death year was estimated from basal disks by correlation to the standardized chronologies.  77  3.3.3 Year of Death of LWD by Statistical Crossdating I successfully crossdated 180 of 190 pieces of L W D to determine year of death (YOD) and age (time since death) of 116 lodgepole pine, 41 white spruce, and 23 black spruce. Fifty-six percent (n = 100) of L W D were high-quality (HQ) estimates of Y O D since bark and/or sapwood were present on these samples, indicating that the outer-ring date represented the last year the tree was alive. High quality estimates of Y O D ranged from 7 to 82 years for lodgepole pine, 10 to 137 for white spruce, and 2 to 80 for black spruce. Best estimates (BE) of Y O D were made for the remaining 80 samples (44%) for which outermost ring dates could not be confirmed as the last year of growth since bark and insect galleries were not present on the outermost portion of the wood (Figure 3.5). Best estimates ranged from 13 to 86 years since tree death for lodgepole pine, 6 to 133 years for white spruce, and 5 to 143 years for black spruce.  78  160 140 CO CD  Q  120  CD CD  100  .E  cn co CO CD  CO  80 60 40  -4—'  CO  Q  20  CD <  (n = 6 6 )  (n = 50)  (n = 2 3 )  LP-HQ  LP-BE  WS-HQ  (n = 18)  (n = 11)  (n = 12)  W S - BE  BS - HQ  BS - BE  Figure 3.5 High quality (HQ) and best estimate (BE) large woody debris age data for lodgepole pine (LP), white spruce (WS), and black spruce (BS). Ten samples could not be crossdated to the living tree chronologies. Four of these ten were confirmed to be Abies sp. by wood anatomy analysis (Hartley, Pers. Comm.). Of the remaining six samples that were not crossdated, two were in decay class 3 and six were in decay class 4. These samples were rejected since an acceptable date for the outermost ring could not be determined due to spurious statistical results and advanced wood decay.  79  3.3.4 Time-series of wood decay The age of L W D of lodgepole pine varied significantly among position classes (p = <0.001, Figure 6a) and decay classes (p = <0.001, Figure 6b). L W D in position class 1 (33.6 ± 19.9 years) was younger than'LWD in position classes 3 and 4 (50.3 ± 16.9 and 54.3 ± 1 6 . 1 , respectively). The age of L W D in position class 2 (44.3 + 21.9) was not different from the other classes. There were two distinct age groups of L W D when classified by stage of decay. L W D in decay classes 1 and 2 (16.5 ± 12.2 and 34.6 ± 19.1) was significantly younger than L W D in decay classes 3 and 4 (52.4 ± 1.7.3 and 50.9 ± 17.8).  80  Figure 3.6 Box plots of the year since death of lodgepole pine L W D by position class (a) and decay class (b). For each box, the horizontal line represents the median age. The box limits are the 25th and 75th percentiles of age. The error bars are the 10th and 90th percentiles of age. The dots show the range in years since death.  A two-way A N O V A for black spruce and white spruce using position class and year since death was performed. Position class by year since tree death (Figure 3.7a) and decay class by years since tree death (Figure 3.7b) for black and white spruce failed the assumption of normality required by a two-way A N O V A (p = <0.001 and p = 0.001, respectively), but passed the tests of equal variance. Since A N O V A is robust with respect to the assumption of population normality, and the validity of the analysis is compromised only slightly by even large deviations in normality, these A N O V A results  81  were considered valid due to the large sample size (n — 64) (Box and Anderson, 1955; Srivastava, 1959; Tiku, 1971; Zar, 1984). There was no significant difference between black spruce and white spruce (p = >0.05) and no significant interaction between species and decay (p = 0.918) and species and position classes (p = 0.714), but a significant difference among decay classes and position classes (p = 0.003, Figure 3.7a and 3.7b). A pairwise multiple comparison procedure (Tukey Test) grouped L W D in position class 1 (38.8 ± 34.7) as younger than L W D in position class 3 (77.5 ± 32.9). The age of L W D in position class 2 (43.8 ±31.8) was younger than the ages of L W D in position class 3 but not different in age from other classes. L W D ages in position class 4 (76.1 ± 42.8) were older than L W D age in position class 1 but not different from other classes. L W D in decay class 4 (82.8 ± 38) was older than L W D in decay classes 1 and 2 (21 ± 24 and 38.9 ± 33, respectively). The age of decay class 3 (62.1 ± 33.8) was not different from other classes.  82  A  AB  BC  X = 38.8 X = 43.8 X = 77.5 X=„76.1  (n = 17)  (n = 20)  (n = 16)  (n = 11)  Position Class  A  AB  B  Figure 3.7 Box plots of the years since tree death of spruce L W D by position class (a)  and decay class (b). For each box, the horizontal line represents the median age. The box limits are the 25th and 75th percentiles of age. The error bars are the 10th and 90th percentiles of age. The dots show the range in years since death.  3.4  Discussion  3.4.1 Age of LWD In the headwater streams of the Foothills Model Forest, L W D persists for many decades to more than a century. The oldest L W D for which I was able to estimate year of death (YOD) using dendroecological techniques was a white spruce log that had been  83  dead for 137 years. The oldest black spruce and lodgepole pine were 80 and 83 years old, respectively. Although I was able to successfully crossdate 95% of L W D that I sampled, some logs were too decayed to sample for tree-ring analysis and may have been older than the dated L W D . In this study, the oldest logs of each species were high quality (HQ) samples have retained their bark and/or included evidence that the sapwood and cambium were present and therefore I am confident that their outermost ring dates are accurate. These high quality samples make up 54% of my entire sampled set. For the remaining 46% of samples, there is some uncertainty in my estimates of Y O D . Sources of error include the loss of outermost tree rings due to mechanical erosion and wood decay resulting in best estimates (BE) of outermost rings, as indicated by the open triangles in Figures 3.3 and 3.4. Because the B E samples may have lost outermost rings, the crossdated results may underestimate Y O D and overestimate the age of the log. Logs in decay classes 3 and 4 were more likely to be missing bark and outer rings than logs in decay classes 1 and 2. Similarly, logs that were loose on the stream bed (position class (PC) 3) or embedded in the banks (PC 4) were more likely to be missing bark and outer rings than bridges and partial bridges (PCI and 2). Among species, the bark was missing from black and white spruce L W D more commonly than lodgepole pine L W D . In this study I attempted to use two approaches to reduce the error due to erosion and decay of outer rings; the first approach was more successful than the second approach. (1) Log decay can be asymmetrical, where the top sides of logs are susceptible to pedogenic processes. Therefore, I cut cross-sectional disks from L W D in decay classes 2, 3 and 4 and measured the ring-width series on the least decayed radii of the wood to  84  provide the best possible estimates of the year of death. (2) I attempted to estimate the number of missing sapwood rings for lodgepole pine L W D that was missing the bark; sapwood is not discernable from heartwood in spruces (Hoadley, 1990). I assumed that i f I could quantify the amount of sapwood in living trees and measure the sapwood remaining on L W D that was missing bark, I could estimate how much sapwood may have been lost from the L W D . In 104 cores from live/lodgepole pine canopy trees, the sapwood included 49.8 ± 13.9 rings. However, the outermost sapwood rings of the living trees were much narrower than the outermost rings of most L W D for which a correction was needed (Figure 3.3) and the resulting "corrections" were too variable for these results to be useful. Therefore, I did not apply a "sapwood correction" to account for outer rings missing from the L W D ; instead, I differentiated the "high quality" estimates from the "best estimates" of the year of death for individual logs.- Additional research is needed to quantify rates of erosion and loss of outer rings from L W D of lodgepole pine and spruce. For example, the number of rings missing from the outside of logs can be determined through flume experiments and species-specific long-term permanent plot studies (e.g. Hyatt and Naimon, 2001).  3.4.2 Recruitment of Large Woody Debris  By examining tree-ring width patterns using dendrochronological techniques, I determined the year of death of trees which have fallen into streams. B y examining radial growth of canopy dominant trees at my sites, I then used the tree-ring history and mortality data to investigate the temporal dynamics of disturbance history and forest stand development in the surrounding riparian zone. I had expected that some L W D  85  would be older than the surrounding canopy trees and would have originated from disturbances that occurred before the last stand-initiation fire. However, all crossdated wood was younger than canopy trees, had been generated from the current stand, and recruited into the stream since stand initiation. 3.4.2.1 Lodgepole Pine  The age histograms for lodgepole pine-dominated sites indicate that all five stands were even-aged cohorts which initiated between 92 and 135 years ago (Figure 3.3). A l l sites were dominated by lodgepole pine, although sites 2 and 5 contained white spruce (Figure 3.3). A l l canopy dominant trees had wide; above-average growth rings near the pith. Even-aged structure and fast initial growth suggests that these stands initiated following disturbance, most likely, fire. When stands were ca. 40 years old, there was a sharp decrease in ring width coupled with the generation of L W D at four of five sites. This decrease in ring-width at several sites suggests a high degree of inter-tree competition within sites 40 years after stand initiation. At that time, L W D was being generated by inter-tree competition due to canopy closure during the competitive exclusion stage of stand development (Franklin et al, 2002). At site 3, six L W D were generated approximately 30 years following stand initiation, as radial growth of the surrounding canopy trees began to decrease but prior to the series decreasing below the long-term average ring-width. It must be noted that these six individuals were only bestestimates of year of tree death, meaning they are potentially missing some outermost rings. It is likely that this L W D had died more recently than indicated by the crossdated Y O D . Thus, the Y O D would be shifted to the right on Figure 3 if the number of missing rings were known and tree death would correspond with canopy closure rather than  86  preceding it. Tree mortality due to competition and within-stand disturbances contributed to recruitment of wood into streams up to and including the 1990's at all five sites. Tree death can occur abruptly, in the case of windthrow, but more frequently, it is a gradual process which involves numerous interacting and synergistic abiotic and biotic processes (Franklin et al, 1987). 3.4.2.2 White and Black Spruce  In general, spruce-dominated riparian forest stand dynamics showed a different pattern from lodgepole pine-dominated sites. Four of five spruce sites (6, 7, 9, and 10) were uneven aged with no dominant cohorts, where canopy trees established between 1730 and 1910 (Figure 3.4). At these sites, incremental growth of the canopy trees was slow near the pith since seedlings initiated beneath an existing canopy. Tree rings grew wider over time and then narrowed again towards the bark. One exception is site 8 (Figure 3.4cz) which is mixed-species, broadly even-aged with three age-classes and initiated later than other sites, where lodgepole pine, black spruce and white spruce cohorts established between 1888 and 1895, likely following a stand-replacing fire. Black spruce, white spruce, and lodgepole pine had narrow rings at the pith, wider rings during the middle portion of growth, and narrowing towards the bark. Year of death of L W D at site 8 was concurrent with the decline in growth rates following the peak in the black spruce series and continued to the present. L W D species included lodgepole pine, white spruce, black spruce and Abies sp., although no Abies were found in the overstory. At site 9, L W D consisted of lodgepole pine as well as white spruce and black spruce, although lodgepole pine was not present in the overstory. The site 9 black spruce chronology correlated poorly with sites dominated by black spruce and white spruce with  87  no lodgepole L W D present. It is likely that a mixed-species stand including lodgepole pine, white spruce, and black spruce existed in the past. Furthermore, the black spruce chronology at site 9 begins with wide growth rings near the pith, as opposed to narrow growth rings, which is a departure from the growth patterns of black spruce at other sites. I suggest that the black spruce chronology initiated as a post-fire stand with lodgepole pine ca. 1750. Lodgepole has since been out-competed in the canopy resulting in no canopy dominants, but evidence of a previous lodgepole canopy is recorded in the L W D . At sites 6, 7, and 10, black spruce and white spruce L W D was generated through time. The oldest crossdated L W D with intact outermost rings died in 1866 and was 137 years old. Most L W D was generated during the 1900's due to fine-scale disturbances and continued to the present. This type of L W D generation reflects the stage of forest stand development of these stands. The primary developmental process through which mature stands are transformed into old-growth is "patchy mortality" of canopy trees due to small scale disturbance (Wells et al, 1998). In the absence of large-scale disturbances, multi-cohort and uneven-aged stands develop due to gap dynamics as individual tree deaths cause understory trees to recruit to the canopy (Wells et al, 1998). In these spruce-dominated sites, we see uneven-aged stands with chronic or continuous generation of L W D through time and the loss of the pioneer cohort of lodgepole pine at two sites, indicating that these stands are transiting between the maturation and old-growth stages of forest stand development (Franklin et al, 2001).  88  3.4.3 Time Series of Wood Decay and Residence Time  I have assigned each L W D to one of four categories which demarcate key decay and structural attributes in order to quantify decay rates and residence time, and determined their years since death, thus linking wood decomposition to wood age. It is expected that as a piece of L W D decays over time, it transits through position and decay classes one through four, expressing increasing decay characteristics and structural instability. While wood decays rapidly in terrestrial environments due to contact with the forest floor and its associated decomposers and detritivores, L W D which enter aquatic systems often decays slowly in submerged or buried environments (Guyette et al, 2002). By linking key decay and structural characteristics (expressed by decay and position classes, respectively) with wood age, my goal was to understand the temporal dynamics and distribution of L W D in riparian zones of small, headwater streams in the foothills of the Rocky Mountains. 3.4.3.1 Lodgepole  Pine  Lodgepole pine were analyzed to quantify rates of decay and residence times by quantifying wood age and sample size in each of four position classes and each of four decay classes (Figure 3.6). As expected, the age of L W D generally increases with increasing position and decay class; however, the mean age in class 3 is less than class 4 in both the decay classification system and the position classification system. This anomaly is partly explained by the uncertainty in the age of L W D in decay classes 3 and 4 is greater'than L W D in classes 1 and 2 for which the bark and sapwood are more commonly present. It also suggests that L W D does not always progress through a linear series of decay and position classes at a constant rate. For example, the scarcity of L W D  89  in decay class 1 (n = 6) suggests that wood moves through the initial decay class quickly. In fact, most logs moved through decay class 1 within 25 years. Although the mean ages of decay classes 3 and 4 were similar (50.9 and 52.4 years, respectively), logs in decay class 3 were more abundant than logs in decay class 4 (Chapter 2). It appears that decay slows and logs accumulate in decay class 3, but wood decays rapidly in class 4 and/or the wood is incorporated into stream banks and beds by the deposition of sediment over time (Hauer et al, 1999). 3.4.3.2 White and Black Spruce  Despite significant overlap in the age of white and black spruce among position classes, it is evident that bridges and partial bridges (position classes 1 and 2) were younger than logs that were loose on the streambed (position class 3) and embedded logs (position class 4)(Figure 3.7a). On average, logs that were loose on the streambed were older than embedded logs, suggesting sediment accumulation within the streambed does not affect all logs equally. Similar to lodgepole pine, spruce L W D in decay classes 1 and 2 were younger than L W D in decay classes 3 and 4, although the classes overlapped (Figure 3.7b). Fewer spruce logs were in decay classes 1 (n = 4) and 4 (n = 14) than in classes 2 (n = 21) and 3 (n = 25), indicating that L W D may spend more time in the middle classes whereas the initial and final stages of decay are more rapid. 3.4.3.3 Comparison of White Spruce, Black Spruce, and Lodgepole Pine  Regardless of species, L W D age increases with increasing position and decay class. The oldest logs are in position classes and decay classes 3 and 4. Both lodgepole pine and spruce-dominated sites have the greatest abundance of logs in decay class 3. Recent literature suggests that in similar forest types, class 3 is the most common stage of  90  decay in terrestrial environments (eg. DeLong et al, 2005). Compared to lodgepole pine, all spruce decay and position class mean ages are older, indicating that decay of lodgepole pine is faster than that of spruces. Despite these differences, both species can be grouped into two groups consisting of position or decay classes 1 and 2 (younger) and position or decay classes 3 and 4 (older), where the cutoff between the younger and older groups is 40 years since death. 3.4.4  Conclusion  In conclusion, as a piece of L W D decays over time, it transits through position and decay classes, expressing increasing decay characteristics and structural instability. This transition occurs faster for lodgepole pine than for spruces since spruce L W D in all classes is older than pine L W D . Logs spend varying amounts of time in each decay class so that class 1 has the smallest abundance of logs and class 3 has the highest abundance. As a log transits from position class 2 to class 3 at approximately 40 years after tree death, a major change in its in-stream function occurs whereby the log becomes integral for stream geomorphology and habitat diversity by forming steps and pools, storing sediment, and providing bank stability (Richmond and Fausch, 1995). Once it has entered the stream, L W D can persist 80 to more than 130 years, depending on species and stand history. Therefore, management decisions that alter the abundance and rate of recruitment of L W D into streams have long-term implications for the structure and dynamics of riparian environments as well as in-stream habitat and biodiversity.  91  3.5 Literature Cited Anonymous, 1997. Sigma Stat for Windows release 2.0 standard version. SPSS Inc. Chicago, IL, USA. Alexander, K . 2002. The invertebrates of living and decaying timber in Britain and Ireland - a provisional annotated checklist. EnglishNature Research Reports. No. 467. Beckingham, J. D., I. G. W. Corns, and J. H . Archibald. Field Guide to Ecosites of West-Central Alberta. Special Report 9, Canadian Forest Service, Northwest Region. 1996. Berg, N . , A. Carlson, and D. Azuma. 1998. Function and dynamics of woody debris in stream reaches in the central Sierra Nevada, California. Candaian Journal of Forest Resources 32: 1460-1477 Bisson, P. A . R. E. Bilby, M . D. Bryant, C. A . Dollof, G. B. Grette, R. A . House, M . L . Murphy, K . V . Koski, and J. R. Sedell. 1987. Large wood in forested streams in the pacific northwest, past, present, and future. In: Streamside Management; Forestry and Fisheries Interactions. University of Washington Press, Seattle, Washington. Pp. 143-190 Boddy, L. 2001. Fungal community ecology and wood decomposition processes in angiosperms: from standing tree to complete decay of coarse woody debris. Ecological Bulletins 49: 43-56 Box, G. E. P. and S. L. Anderson. 1955. Permutation theory in the derivation of robust criteria and the study of departures from assumption. Journal of the Royal Statistical Society B I T ' : 1-34 Church, M . 1992. Channel Morphology and Topology. In: The Rivers Handbook, C. Calow and G. Petts (Eds). Blackwell, Oxford. 2: 126-143 Cook, E. R. 1985. A Time Series Analysis Approach to Tree-Ring Standardization. Unpublished Ph D. Dissertation, University of Arizona, Tucson, Arizona, U S A . Daniels, L. D., J. Dobry, K. Klinka, and M . C. Feller. 1997. Determining year of death of logs and snags of Thuja plicata in southwestern coastal British Columbia. Canadian Journal of Forest Resources 27: 1132-1141 Delong, S.C., L.D. Daniels, B . Heemskerk, and K.O. Storaunet. 2005. Temporal development of downed wood habitats in wet spruce-fir stand in east central British Columbia. Canadian Journal of Forest Research 35: 2841-2850.  Duncan, R. P. 1989. A n evaluation of errors in tree age estimates based on increment cores of Kahikatea (Dacrycarpus  dacrydiodes).  16:31-37  New Zealand Natural Sciences  •  •  Edmonds, R. L. and A. Eglitis. 1989. The role of Doulas-fir beetle and wood borers in the decomposition and nutrient release from Douglas-fir logs. Canadian Journal of Forestry 19(7): 853-859 Fausch, K . D. and T. G. Northcote. 1992. Large woody debris and salmonid habitat in a small coastal British Columbia stream. Canadian Journal of Fisheries and Aquatic Sciences 49: 682-693  Franklin, J., Spies, T., Van Pelt, R., Carey, A . B., Thornburgh, D. A., Berg, D. A . , Lindenmeyer, D. B., Harmon, M . E., Keeton, W. S., Shaw, D. C , Bible, K., and J. Chen. 2002. Disturbances and structural development of natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. Forest Ecology and Management.  155:399-423  Fritts, H. C. 1976. Tree Rings and Climate. Academic Press, 567 pp. Gomi, T., R. C. Sidle, M . D. Bryant and R. D. Woodsmith, 2001. The characteristics of woody debris and sediment distribution in headwater streams, southeastern Alaska. Canadian Journal of Forest Research 31: 1386-1399  Grissino-Mayer, H. D. 2001. Evaluating Crossdating Accuracy: A Manual and Tutorial for the Computer Program C O F E C H A . Tree-Ring Research 57(2): 205-221 Guyette, R.P., Cole, W.G., Dey, D . C , Muzika, R 2002. Perspectives on the age and distribution of large wood in riparian carbon pools. Canadian Journal of Fisheries and Aquatic Sciences 59: 578-585.  Harmon, M . E., J. F. Franklin, F. J. Swanson, P. Sollins, S.V. Gregory, J. D. Lattin, Anderson, N . H., S. P. Cline, N . G. Aumen, J. R. Sedell, G. W. Lienkaemper, K . Cromack, and K. W. Cummins, 1986. Ecology of Coarse Woody Debris in Temperate Ecosystems. Advances in Ecological Research 15: 133-302. Hassan, M . A., D. L . Hogan, S. A . Bird, C. L . May, T. Gomi, and D. Campbell, 2005. Spatial and temporal dynamics of wood in small streams. Journal of the American  Water Resources Association  41: 899-919  Hartley, I. 2006. Personal Communication. Ecosystem Science and Management Program. University of Northern British Columbia. Prince George, B C . Hauer, F. R., Poole, G. C , Gangemi, J. T., and C. V . Baxter, 1999. L W D in bull trout spawning streams of logged and wilderness watersheds in northwest Montana. Canadian Journal of Fisheries and Aquatic Sciences 56(6): 915-924  93  Hoadley, R. Bruce. 1990. Identifying Wood. The Taunton Press Inc., 223 pp. Hogan, D. L . 1987. The influence of large organic debris dn channel recovery in the Queen Charlotte Islands, British Columbia, Canada. In: Erosion and sedimentation in the Pacific Rim. Proceedings of the Corvallis Symposium, August, 1987. IAHS Publication No. 165  1  Hogan, D. L., S. A . Bird, and M . A . Hassan, 1998. Spatial and Temporal Evolution of Small Coastal Gravel-Bed Streams: The Influence of Forest Management on Channel Morphology and Fish Habitat. In: Gravel-Bed Rivers in the Environment, P. C. Klingeman, R. L . Beschta, P. D. Komar, and J. B. Bradley (Editors). Water Resources Publications, L L C , Highland Ranch, Colorado, USA, pp. 365-392. Holmes, R. L. 1986. Quality control of crossdating and measuring: a user's manual for program C O F E C H A . Pp 41-49 in Holmes, R. L., R. K . Adams and H . C. Fritts (eds.) Tree-ring Chronologies of Western North America: California,  Eastern  Oregon and Northern Great Basin. University of Arizona Press, Tucson, Arizona, USA. Hyatt, Timothy, L . and Robert Naimon. 2001. The residence time of large woody debris in the Queets River, Washington, USA. Ecological Applications 11(1): 191-202 Jozsa, L . 1988. Increment Sampling Techniques for High Quality Cores. Wood Science Department. Labetoire Vancouver Laboratory. Special Publication No. S P - 3 : ISSN #0824-2199 Keller, E. A . and F. J. Swanson, 1979. Effects of large woody organic material on channel form and fluvial processes. Earth Surface Processes and Landforms 4: 361-380. Loman, A . A . 1970. The effect of Heartwood fungi of lodgepole pine on its phenolic heartwood extractives. In: Interaction of Organisms in the Process of Decay of Forest Trees. Symposium under the chairmanship of Dr. Alex L. Shi go, Bulletin No. 13. 43 pp. Luckman, B. 2003. Personal Communication. Department of Geography. University of Western Ontario. London, O N . Martin, D. J. and L. E. Benda. 2001. Patterns of instream wood recruitment and transport at the watershed scale. American Fisheries Society 130: 940-958 Montogmery, D. R., J. M . Buffington, R.D Smith, K. M . Schmidt, and G. Pess. 1995. Pool spacing in forest channels. Water Resource Research 31: 1097-1105 McCleary, R., C. Widk, and J. Blackburn. 2002. Comparison between field and GIS derived descriptors of small streams within the west-central foothills of Alberta.  94  Fish and Watershed Program, Foothills Model Forest, Hinton Alberta. April 2, 2002. 40pp. McCleary, R., C. Bambrick, C. Sherburne, and S. Wilson. 2003. Report 2.4.1b. Level 1 classification: GIS - based stream reach characteristics. Fish and Watershed Program, Foothills Model Forest, Hinton Alberta. M a r c h ! 2, 2003. 55 pp. Nakamura, F. and F. J. Swansen. 1994. Distribution of C W D in a mountain stream western Cascade Range, Oregon. Earth Surface Processes and Landforms 18(1): 43-61 Phipps, R. L. 1985. Collecting, preparing, crossdating, and measuring tree increment cores. U.S. Geological Survey, Water-Resources Investigations Report 85 - 4148 Potts, D. F. and B. K . M . Anderson, 1990. Organic debris and the management of small stream channels. Western Journal of Applied Forestry 5(1): 25-28 Richmond, A . D. and K. D. Fausch, 1995. Characteristics and function of large woody debris in subalpine Rocky Mountain streams in northern Colorado. Canadian Journal of Fisheries and Aquatic Sciences 52:1789-1802. Schwab, J. W., 1998. Landslides on the Queen Charlotte Islands: Processes, Rates, and Climatic Events. In: Carnation Creek and Queen Charlotte Islands Fish/Forestry Workshop: Applying 20 Years of Coastal Research to Management Solutions. D. L. Hogan, P. J. Tschaplinski, and S. Chatwin (Editors). B C Ministry of Forests, Research Branch, Victoria, B C . Land Management Handbook No. 41. Speight, M . C. D. 1989. Saproxylic Invertebrates and their Conservation. Nature and Environment Series, No. 42. Strasbourg: Council of Europe. Srivastava, A . B. L . 1959. Effects of non-normality on the power of the analysis of variance. Biometrika 46: 114-122 Stokes, M . A . and T. L. Smiley. 1968. A n Introduction to Tree Ring Dating. University of Chicago Press. Chicago, Illinois, U S A . Swanson, F. J. and G. W. Lienkamper. 1978. Physical Consequences of Large Organic Debris in Pacific Northwest Streams. U S D A Forest Service General Technical Report GTR-PNW-69. Swetnam, T. W., Thompson, M . A., and E. K . Sutherland. 1985. Using dendrochronology to measure radial growth of defoliated trees. U.S. Dept. of Agriculture, Forest Service, Cooperative State Research Service, Wasington, D.C. Tiku, M . L. 1971. Power function of the F-test. Journal of the American Statistical Association 62: 525-539  95  Veblen, T. T., Hadley, K . S., Reid, M . S., and A. J. Rebertus. 1991. The response of subalpine forests to spruce beetle outbreak in Colorodo. Ecology 72: 213-231 Wells, R. W., K. P. Lertzman, and S. C. Saunders. 1998. Old-growth definitions for the forests of British Columbia, Canada. Natural Areas Journal 18: 279-292 Wong, C. M . and K. P. Lertzman. 2001. Errors in estimating tree age: Implications for studies of stand dynamics. Canadian Journal of Forest Research 31: 1262-1271 Young, M . K., E. A. Mace, E. T. Ziegler, and E. K . Sutherland. 2006. Characterizing and contrasting instream and riparian coarse wood in western Montana basins. Forest Ecology and Management 226: 26-40. Zar, J. H . 1984. Biostatistical Analysis. 2 USA  nd  ed. Prentice Hall, Engelwood Cliffs, N J ,  96  4. THESIS SUMMARY AND MANAGEMENT IMPLICATIONS  4.1 Conclusions In this study I identified two major disturbance types that generated large woody debris in mature lodgepole pine and spruce forests in the Foothills Model Forest. After stand-replacing fires, even-aged lodgepole pine and mixed species lodgepole pine and spruce forests regenerate. During canopy closure ca. 40 years after stand initiation, a pulse of tree mortality results in the recruitment of large woody debris (LWD). In older stands in later stages of development, within stand disturbances occur as a result of stand dynamics. L W D inputs vary in different forest types and under different forest histories; however, recruitment is an ongoing process as evidenced by the presence of wood in all decay classes and position classes at my study sites. Given that spruce L W D persists fori 50 years and lodgepole pine for 90 years, forest management must account for impacts on the amount and type of woody debris in riparian forests on the order of a century. Volumes of individual pieces of L W D , both in-stream and within riparian zones were consistent among sites. Position and decay classifications represented progressive changes to L W D in terms of volume, length and diameter. Changes in volume were related to decay or position class such that as the class number increases, volume decreases, and this was driven largely by changes in mean length from position class to position class and decay class to decay class. Changes in L W D age were directly related to decay and position class such that as class number increases, age increases. Since position classes differentiated into distinct groups with respect to volume and age at all sites combined, I propose that the position class model should be used to  97  represent L W D volume and L W D age. I recommend the use of position classes for studies of large woody debris that aim to understand the interactions between forest dynamics, stream hydrology and geomorphology, and position of wood relative to the streambed. Future work on large woody debris processes should address recruitment and residency in other forest types and in forests of all ages and stages of development. Specifically, a chronosequence study used in conjunction with permanent study plots would contribute to baseline data of natural rates of input and decay of L W D in riparian forests. Then we can compare baseline data with the impacts of management choices such as fire prevention, prescribed burning, buffer zones, and harvest on L W D dynamics in riparian zones. Ultimately, I recommend linking research of forest dynamics with research of stream geomorphology and bio-diversity so that we can make positive management choices for the future.  98  

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.831.1-0092740/manifest

Comment

Related Items