Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of spacing on multi-aged interior douglas-fir stands in central British Columbia Bugnot, Jean-Loup 1998

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

Item Metadata

Download

Media
831-ubc_1999-0012.pdf [ 3.98MB ]
Metadata
JSON: 831-1.0075192.json
JSON-LD: 831-1.0075192-ld.json
RDF/XML (Pretty): 831-1.0075192-rdf.xml
RDF/JSON: 831-1.0075192-rdf.json
Turtle: 831-1.0075192-turtle.txt
N-Triples: 831-1.0075192-rdf-ntriples.txt
Original Record: 831-1.0075192-source.json
Full Text
831-1.0075192-fulltext.txt
Citation
831-1.0075192.ris

Full Text

EFFECTS OF SPACING ON MULTI-AGED INTERIOR DOUGLAS-FIR STANDS IN CENTRAL BRITISH COLUMBIA by Jean-Loup Bugnot B . S c . A . Universite Laval, 1995  A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF F O R E S T R Y in  T H E F A C U L T Y OF G R A D U A T E STUDIES F A C U L T Y OF FORESTRY Department o f Forest Resources Management  We accept this thesis as conforming to the required standard  THE U N I V E R S I T Y OF BRITISH C O L U M B I A December 1998 © Jean-Loup Bugnot, 1999  In  presenting  degree  this  thesis  in partial fulfilment  at the University of of this thesis for  department  or  by  his  or  the  requirements  for  scholarly her  I further agree that permission for  purposes  may be granted  representatives.  It  is  by  understood  that  permission.  T'Drc^f  VJA  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  advanced  ZtZ^d  ^  a  ^ ^  ^ e ^ r ~ l  e  extensive  the head of  publication of this thesis for financial gain shall not be allowed without  Department of  an  British Columbia, I agree that the Library shall make it  freely available for reference and study. copying  of  copying  my or  my written  ABSTRACT Re-measurement data over a period o f four years from 24 permanent plots were analyzed to determine the effects o f three different spacing regimes on growth o f uneven-aged interior Douglas-fir (Pseudotsuga menziesii var. glauca), near Williams Lake, British Columbia. Spacing was carried out during the fall and winter o f 1990-1991. Growth was analyzed at the stand and individual tree level between the beginning o f the 1993 growing season and the end o f the 1996 growing season. Variables o f interest at the stand level were growth and growth rate o f quadratic mean diameter, basal area per hectare, Lorey's height, volume per hectare and accumulated crown area per hectare. Variables o f interest for individual trees were growth i n diameter, basal area, total height, total stem volume, crown area and live crown length. Analysis o f variance was used to judge the significance o f treatment effects. A t the stand level, there were no significant differences among treatments for any o f the variables. However, treatment means o f quadratic mean diameter, basal area, volume, and accumulated crown area were consistently lower in the control plots than i n the spaced plots. In the spaced plots, the 5 m clumped spacing showed consistently higher growth rates i n relative terms. Stand diameter growth was 0.37 c m on the control, and ranged from 0.80 to 1.08 c m on the spaced plots. Net basal area growth averaged 2.34 m /ha on the control, and 3.23 to 3.52 m / h a on the spaced plots, while mortality represented 0.83 m /ha on the control, and 0.01 to 0.22 m /ha on the spaced plots. Spacing did not reduce net periodic annual volume increment which was 4.8 m /ha on the control, and 5.2 to 6.6 m /ha on the spaced plots. A t the individual tree level, small Douglas-fir tree (<15 c m dbh) growth responses significantly differed between control and spacing treatments for diameter, basal area, height, and volume. Large Douglas-fir tree (>15 c m dbh) diameter and basal area growth response differed significantly between the control and the 5 m clumped spacing. Despite similar trends, results were less distinct for nonDouglas-fir species. U p to 20-25 c m dbh, volume growth efficiency increased regularly with increasing dbh on both control and spaced plots. Trees seemed to be the most efficient for the 5 m clumped spacing and the least efficient for the control. The study also showed that age distribution was clearly distinct from diameter distribution. 2  2  2  2  3  3  TABLE OF CONTENTS  ABSTRACT  ii  T A B L E OF C O N T E N T S  iii  LIST OF T A B L E S  vi  LIST OF FIGURES  viii  ACKNOWLEDGEMENTS  ix  1. INTRODUCTION  1  2. BACKGROUND  3  2.1 S I L V I C S O F I N T E R I O R D O U G L A S - F I R  3  2.2 D Y N A M I C S O F M U L T I - A G E D S T A N D S  3  2.3 R E S P O N S E T O T H I N N I N G  5  2.4 M U L T I - A G E D S T O C K I N G P R O C E D U R E S  7  3. METHODS  10  3.1 S T U D Y A R E A  10  3.2 S T U D Y D E S I G N  11  3.3 T R E A T M E N T S  11  3.4 D A T A 3.4.1 Measurements 3.4.2 Derived Attributes 3.4.3 Error Checking 3.4.4 Outliers 3.4.5 Ecological Data  14 14 15 17 18 18  3.5 A N A L Y S I S 3.5.1 Variables o f Interest 3.5.2 Statistical Methods  18 18 19  4. RESULTS  21  4.1 P L O T G E N E R A L F E A T U R E S 4.1.1 Site Identification 4.1.2 A g e Structure 4.1.3 Plot Conditions  21 21 23 25 iii  4.2 S T A N D - L E V E L R E S P O N S E 4.2.1  32  Mortality  32  4.2.2 I n g r o w t h 4.2.3 T r e e S i z e D i s t r i b u t i o n  35 35  4.2.4 G r o w t h R e s p o n s e 4.2.4.1 O v e r v i e w  38 38  4.2.4.2 Q u a d r a t i c M e a n D i a m e t e r  39  4.2.4.3 B a s a l A r e a  39  4.2.4.4 L o r e y ' s H e i g h t 4.2.4.5 V o l u m e 4.2.4.6 A c c u m u l a t e d C r o w n A r e a  42 42 42  4.2.4.7 B i o m a s s  45  4.3 T R E E - L E V E L R E S P O N S E 4.3.1 O v e r v i e w  45 45  4.3.2 D i a m e t e r  46  4.3.2.1 S m a l l T r e e s  46  4.3.2.2 L a r g e T r e e s 4.3.3 B a s a l A r e a 4.3.3.1 S m a l l Trees  51 51 51  4.3.3.2 L a r g e T r e e s 4.3.4 H e i g h t  51 51  4.3.4.1 S m a l l T r e e s  51  4.3.4.2 L a r g e T r e e s  57  4.3.5 V o l u m e 4.3.5.1 S m a l l T r e e s  57 57  ,  4.3.5.2 L a r g e T r e e s  57  4.3.6 C r o w n A r e a  60  4.3.6.1 S m a l l Trees 4.3.6.2 L a r g e T r e e s 4.3.7 L i v e C r o w n L e n g t h  60 60 64  4.3.7.1 S m a l l Trees  64  4.3.7.2 L a r g e T r e e s  64  4.4 R E S U L T S S U M M A R Y F O R G R O W T H R E S P O N S E  5. DISCUSSION  68  5.1 S T A T I S T I C A L A N A L Y S I S 5.1.1  64  68  Analysis of Variance  68  5.1.2 A n a l y s i s o f C o v a r i a n c e  70  5.2 A G E S T R U C T U R E  71  5.3 G R O W T H  72  5.3.1 G r o w t h F a c t o r s  72  5.3.2 S t a n d G r o w t h  74  5.3.3 I n d i v i d u a l Tree G r o w t h  77  5.4 M A N A G E M E N T I M P L I C A T I O N S  81 iv  6. SUMMARY AND CONCLUSIONS  84  LITERATURE CITED  86  APPENDICES  94  A P P E N D I X I: D I A M E T E R D I S T R I B U T I O N O F T R E E S B Y S P E C I E S  95  LIST OF TABLES Table 1. Site identification Table 2. Average age (years) at stump height o f sample trees o f different species by size and by block Table 3. Summary o f plot conditions at the 1993 measurement Table 4. Summary o f plot conditions at the 1997 measurement Table 5. Summary o f accumulated crown area per ha in 1993 and 1997 Table 6. Biomass o f various tree components at the 1993 measurement (t/ha) Table 7. Biomass o f various tree components at the 1997 measurement (t/ha) Table 8. " K e y " SDI values for interior Douglas-fir (adapted from L o n g 1985) Table 9. Reineke's stand density index (SDI) by treatment in 1993 and 1997 Table 10. Summary o f mortality and ingrowth between 1993 and 1997 Table 11. De Liocourt's quotient values (q) by treatment i n 1993 and 1997 Table 12. M e a n o f stand-level growth response by treatment between 1993 and 1997 for five variables o f interest Table 13. Net biomass growth (%) o f various tree components between 1993 and 1997 Table 14. Comparison o f diameter growth among treatments, by diameter class and by species Table 15. Comparison o f diameter growth among diameter classes, by treatment and by species Table 16. Comparison o f basal area growth among treatments, by diameter class and by species Table 17. Comparison o f basal area growth among diameter classes, by treatment and by species Table 18. Comparison o f height growth among treatments, by diameter class and by species Table 19. Comparison o f height growth among diameter classes, by treatment and by species Table 20. Comparison o f volume growth among treatments, by diameter class and by species Table 21. Comparison o f volume growth among diameter classes, by treatment and by species Table 22. Comparison o f crown area growth among treatments, by diameter class and by species Table 23. Comparison o f crown area growth among diameter classes, by treatment and by species Table 24. Comparison o f life crown length growth among treatments, by diameter class and by species vi  22 24 26 27 28 30 31 33 33 34 38 40 44 47 50 53 54 55 56 58 59 62 63 65  Table 25. Comparison of live crown length growth among diameter classes, by treatment and by species 66  vii  LIST OF FIGURES Figure 1. V o l u m e distribution by dbh class and by treatment in 1997 Figure 2. Changes i n the number o f trees by size class on Control, between 1993 and 1997 Figure 3. Changes in the number o f trees by size class on treatment C 2 , between 1993 and 1997 Figure 4. Stand-level growth response in relation to Curtis' (1982) relative density for five variables o f interest between 1993 and 1997 Figure 5. Net volume growth between 1993 and 1997 in relation to 1993 growing stock Figure 6. M e a n o f individual-tree growth response by plot for trees less than 15 c m dbh, between 1993 and 1997 Figure 7. M e a n o f individual-tree growth response o f trees less than 15 c m dbh in relation to Curtis' (1982) relative density, by plot and for six variables o f interest between 1993 and 1997 Figure 8. M e a n o f individual-tree growth response o f trees larger than 15 c m dbh in relation to Curtis' (1982) relative density, by plot and for six variables o f interest between 1993 and 1997 Figure 9. Growth efficiency (stem volume growth / initial crown projection area) per dbh class between 1993 and 1997  viii  29 36 37 41 43 48  49  52 61  ACKNOWLEDGEMENTS The project on which this work is based was initiated by Dr. Peter Marshall. Funding for the project has been provided by a number o f sources over the years. The Challenge '89 Employment Program' and the Community Forestry Program (Williams Lake District) paid for the initial plot installation. The costs o f the spacing were covered by the Community Forestry Program. Plot re-establishment i n 1993 and subsequent analyses were funded by the Canada-British Columbia Partnership Agreement on Forest Resource Development ( F R D A I I ) . The 1997 plot re-measurements and this subsequent work are being supported by British Columbia's Forest Renewal Fund ( F R B C ) . I am highly appreciative o f this support. M a n y thanks are due to my major supervisor, D r . Peter Marshall, for his help, encouragement and careful review o f my work. I am also grateful to D r . Antal K o z a k , Dr. J i m Thrower, and Dr. A b d e l - A z i m Zumrawi for their support and valuable suggestions. Dr. Guillaume Therien provided considerable advice on S A S programing. Helpful review o f sections o f the manuscript was also provided by Celine Boisvenue, Robert Froese and A l e d Hoggett. Gordon Nienaber assisted with the field work. Special thanks to K e n Day and Claire Trethewey at the A l e x Fraser Research Forest. Finally I would like to thank my wife, Anna, for her support and encouragement.  JLB December 1998  ix  1. INTRODUCTION  Prior to European settlement, fires and insects were the primary disturbances i n the forests dominated by Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) Franco) i n the interior o f British Columbia (Vyse et al. 1990). Depending on the ecosystem, the fire return interval ranged from 8 to 50 years for surface fires, and from 150 to 250 years for stand replacing fires (Arno 1980, Province o f B C 1995). O n dry sites, low-intensity surface fires periodically destroyed the lower canopy, or reduced its density. Survival o f small trees depended mostly on irregularities i n the spatial patterns o f fires, whereas large trees with thick bark survived relatively unharmed (Hope et al. 1991). A s a result, stands frequently had open, multi-aged structures with cohorts o f trees in various age classes (Weaver 1967 cited by Kilgore 1981, Vyse etal. 1990, Province o f B C 1995). Since European settlement, fire control and partial cutting practices have altered the dynamics o f most existing interior Douglas-fir forests (Smith 1985, V y s e et al. 1990). Fire control has favoured regeneration as well as development o f dense pockets o f slow growing saplings and poles. Past unregulated cutting practices changed the structure o f the stands dramatically (Clark 1952) and may have had dysgenic effects (Howe 1995). In recent years, even though little work has been done on the dynamics o f these forests (Johnstone 1985), management practices have improved, with more consideration given to the control o f stand structure. A l s o , more attention is now given to special wildlife habitat concerns. In the Cariboo Forest Region, mule deer (Odocoileus hemionus hemionus Raf.) use interior Douglas-fir forests 1  as winter range. More specifically, a sufficient number o f old, large trees and a m i x o f multiaged stands, with various levels o f crown closure, seem to adequately meet food and shelter requirements o f mule deer during winter (Armleder et al. 1986, 1989, Armleder and Dawson 1992). A t the end o f the 1980's, a spacing project was initiated i n uneven-aged Douglas-fir stands on the A l e x Fraser Research Forest near Williams Lake, B . C . , to study the impact o f different spacing regimes on mule deer use o f the areas. This project took place in areas partially cut in the early 1960's, using a minimum diameter limit rule. When the project started, these areas were dominated by dense patches o f small trees. A stand dynamics study was able to take advantage o f the existing spacing design for the mule deer project (Marshall 1996, Marshall and Wang 1996). The main objective o f this thesis was to analyze and discuss stand and tree growth responses in this long-term experiment, six years after spacing was undertaken, and four years after the first complete post-spacing measurement. Specifically, the following null hypotheses were examined: 1.  Treatments do not affect stand growth i n quadratic mean diameter, basal area per  hectare, Lorey's height, volume per hectare and accumulated crown area per hectare. 2.  Treatments do not affect individual tree growth i n diameter at breast-height (dbh),  basal area, height, volume, crown area and live crown length. 3.  For each variable o f interest examined at the individual-tree level, treatment  response is not affected by initial measurement o f dbh. In the next chapter, work and pertinent information related to this study are presented. This is followed by a description o f methods i n Chapter 3. Results are presented i n Chapter 4 and discussed in Chapter 5, along with their management implications. 2  2. BACKGROUND 2.1 S I L V I C S O F I N T E R I O R D O U G L A S - F I R Interior Douglas-fir is wide ranging throughout western North America, and has the broadest ecological amplitude o f any tree species i n this part o f the continent (Arno 1990). This reflects its high genetic variability, which may also be considerable within populations (Rehfeld 1989). In British Columbia, the species occupies much o f the Interior Douglas-fir biogeoclimatic zone (IDF), which occurs mostly on low- to mid- elevations i n the south-central interior o f the Province. Over much o f the I D F , Douglas-fir is the climax tree species (Hope et al. 1991), and is moderately shade-tolerant (Daniel et al. 1979, p. 297). Damaging frosts can occur during every month o f the growing season (Environment Canada 1985, in Newsome et al. 1990), but water deficit, conditioned by soil drought and competing vegetation, is the main limiting factor o f growth and survival o f the tree (Lopushinski 1990). In the Fraser variant o f the dry cool interior Douglas-fir biogeoclimatic subzone (IDFdk3), i n which most o f the Knife Creek Unit is located, regeneration under a partial canopy is generally abundant (Steen and Coupe 1997, Day 1998).  2.2 D Y N A M I C S O F M U L T I - A G E D S T A N D S Multi-aged stands are stands with two or more cohorts, or groups o f trees, o f similar ages. Cohorts arise after a common disturbance, but age within a cohort may range from 1 year to 3  several decades (Oliver and Larson 1996, p. 147). Due to its specific origin, structure and environment, each cohort within a multi-aged stand has a unique development pattern (O'Hara 1996). Typically, the oldest cohort occupies the dominant canopy position (O'Hara 1996) and can adversely affect the growth o f younger, subordinate cohorts by limiting their light environment and possibly, their nutrient and water supply (Oliver and Larson 1996, p. 302). Investigating the soil water regime and growth in uneven-aged interior Douglas-fir stands near Kamloops, B . C . , K o r o l (1985) found that high canopy cover caused a reduction o f the soil water reserve by limiting snow throughfall and melt, as well as rainfall reaching the ground. A s a result, diameter and height growth were reduced in dense stands because o f inter-tree competition for limited moisture. In contrast, leaf canopies are often less dense on dry sites, allowing more light to reach the lower parts o f the stand and thus younger cohorts to be more vigorous (Grier and Running 1977). Lower canopy may also affect overstory trees, especially on dry sites where young cohorts may compete for moisture more effectively than older cohorts (Oliver 1995). Barrett (1965, 1970, 1982) found that understory vegetation, along with stand density, had marked effects on diameter, height and volume in a 40 to 70-year-old stand o f suppressed ponderosa pine (Pinus ponderosa Laws.) saplings i n central Oregon. In the Fraser variant o f the Interior Douglas-fir dry cool subzone (IDFdk3), where advance Douglas-fir regeneration is often dense, growth o f dominant trees may be reduced as a result o f below-ground competition for moisture and nutrients (Steen and Coupe 1997).  4  2.3 R E S P O N S E T O T H I N N I N G Excessive density results in reduced individual-tree growth and vigour, and increased mortality. L o w intensity disturbances, like thinning, make more growing space available to residual trees. A t the same time, they alter the availability o f water and nutrients by increasing the radiation flux and rainfall reaching the forest floor (Waring and Schlesinger 1985, Chapter 5, K i m m i n s 1997, Section C). Consequently, soil temperature increases, as well as microbial activity and mineralisation o f previously undecomposed organic matter (Assmann 1970, p. 232, Waring and Schlesinger 1985, Chapter 8, Oliver and Larson 1996, p. 25). Interior Douglas-fir i n the I D F zone are typically water-limited during late summer because o f long periods with little precipitation (Hope et al. 1991). Thinning may increase available water for residual trees and also reduce canopy interception. A s a result o f additional resources and better environmental conditions, tree growth rates should increase. In relatively similar conditions in the United States, response o f ponderosa pine to thinning release has been studied for a long time. Dunning (1922), cited by Smith (1985), studied the effect o f partial cutting on growth o f ponderosa pine. H e observed that promptness o f response to increased light, moisture and available space was closely related to crown size. It was later suggested that moisture was more important than light and that accelerated growth resulted more from an increased amount o f available water than from crown extension (Meyer 1931). Barrett (1963) observed that dominant ponderosa pine responded markedly to the removal o f adjacent subordinate trees. In parallel, Barrett (1981, 1982) also found that ponderosa pine trees, even i f long suppressed, can develop large diameters fairly rapidly i f enough growing space becomes available. Compared to ponderosa pine, little information is available on growth response o f interior Douglas-fir after thinning release. In the southern interior o f British Columbia, Clark (1952) was 5  the first to demonstrate growth increases o f remaining trees after logging. In the Cariboo, G l e w and Cinar (1966) studied Douglas-fir stand growth after seed tree marking. They found that radial growth i n the pole layer increased 67 % to average about one inch (2.54 cm) per decade following logging. Effects o f a heavy low thinning on a 45-year-old Douglas-fir stand near Quesnel, B . C . , were reported ten years later (Breadon 1981). Heavy thinning reduced basal area growth per ha but increased radial growth, crown width, and crown height o f the remaining trees. V y s e (1981) compared 50 healthy stems 1 to 10 m tall in an unlogged Douglas-fir stand near Williams Lake with 50 trees o f similar size in an area logged by diameter limit cutting i n the 1970/1971 winter. A l l trees showed increased diameter growth following logging. Both height and diameter growth i n the unlogged stand were strongly influenced by late spring precipitation. More recently, i n the Pothole Creek Demonstration Area near Merritt, B . C . , large diameter responses were observed after release, even in trees located at great distances from other trees removed in previous partial cuttings. In addition, it was also demonstrated that trees o f almost all sizes were capable o f responding to increased growing space (J.S. Thrower and Associates 1997a). Although radial growth is generally responsive to density changes that result from thinning, height growth o f dominant trees is normally not, within the range o f densities encountered i n managed stands (Smith 1986, p. 69). However, very close spacing on dry sites and poor soils may affect height growth o f some species (Sjolte-J0rgensen 1967). Short-term decreases in height growth following thinning (Staebler 1956, M i l l e r and Reukema 1977, C r o w n et al. 1977, Harrington and Reukema 1983) and long-term increases i n height growth with increasing spacing (Curtis and Reukema 1970, Reukema 1979) have been documented for evenaged stands o f coastal Douglas-fir. Reduced height growth i n dense stands on l o w site quality has  6  also been demonstrated for ponderosa pine (Lynch 1958) and lodgepole pine (Pinus contorta var. latifolia Engel.) (Alexander et al. 1967). In multi-aged stands, Barrett (1963) found that height growth o f ponderosa pine responded markedly to removal o f adjacent subordinate trees. Similar results were reported for interior Douglas-fir (Smith 1985). A l s o , substantial effects o f density on mean annual height growth were measured on this species (Korol 1985), as well as dramatic increases i n height growth after release for trees up to 100-year old i n the Merritt area (J.S.Thrower and Associates 1997b). This review clearly shows that manipulation o f forest cover can influence growth across size/age classes. However, even i f "...uneven-aged dry-belt Douglas-fir management has been practiced for over 50 years..." ( B C Ministry o f Forests 1992), it needs to be emphasized that "...probably all experience with selective cutting has been oriented towards harvesting, with little consideration given to how one controls stand structure over time and how one develops an entire system o f management..." (Helms and Lotan 1987). In order to achieve particular stand management objectives, controlling the total amount o f growing stock and finding an appropriate balance between cohorts are critical elements (O'Hara 1996). In addition, as suggested by several authors, distribution and arrangement o f tree crowns may be more important for stand growth and development than density (Mitscherlish 1963, Schiitz 1989, O'Hara 1989).  2.4 M U L T I - A G E D S T O C K I N G P R O C E D U R E S In North America, the conventional approach for regulating structure and growing stock in uneven-aged stands defines the stand structure with a negative exponential or reverse-J diameter distribution (de Liocourt 1898, Leak 1964). In practice, this approach is commonly called the "BDq" method because three parameters must be set to determine the stand structure: 7  basal area after harvest (B), diameter o f the largest trees (D) and ratio o f trees in a diameter size class to the number o f trees in the next larger diameter class (q) (Leak 1964, Alexander and Edminster 1977, Marquis 1978, Farrar 1981, Guldin 1991). B y defining a quotient "q", it is assumed that the diameter distribution is identical to the age distribution and that each age/size class occupies a similar amount o f growing space or horizontal crown area (Smith 1986, p. 17, N y l a n d 1996, p 201). However, i f multi-cohorts stands can have a reverse-J diameter distribution, the same distribution can also develop in single-cohort. Consequently, "...extrapolating from tree size distributions to age distributions i n stands can be very misleading.. ."(Oliver and Larson, p. 296). Furthermore, the method cannot be easily applied to stands with bimodal or multi-modal distribution o f diameters, which were common i n many presettlement stands o f the interior northwest (O'Hara and Valappil 1995). The method also requires a large number o f trees below commercial size (Long and Daniel 1990, Cochran 1992). "These trees take space that could be used by larger more valuable trees, may require pre-commercial thinning, and/or lead to excessive and unnecessary mortality" (O'Hara 1996). L o n g and Daniel (1990) proposed an alternative to the "BDc7" method. They suggested that Reineke's (1933) stand density index (SDI) can be used to define stocking levels for unevenaged stands. Based on the assumption that SDI is proportional to site utilization (Long and Dean 1986, L o n g and Smith 1990), they proposed apportioning the growing stock among diameter classes or group o f diameter classes on the basis o f SDI, with more growing space allocated to the merchantable diameter classes. Cochran (1992) supported L o n g and Daniel's SDI approach, but recommended taking into account stand structure before entry to determine the residual stocking level. More recently, O'Hara (1996) proposed a technique based on allocation o f growing space, as measured by diameter and leaf area, to the different cohorts o f multi-aged  8  ponderosa pine stands. A l s o , Day (1997) developed a stocking guide diagram adapted from Gingrich (1967) for interior Douglas-fir in the Cariboo region.  9  3. METHODS Sections 3.1 to 3.3 are adapted from Marshall (1996), i n which additional details may be found.  3.1 S T U D Y A R E A The Knife Creek Unit o f the University o f British Columbia A l e x Fraser Research Forest is located about 20 k m south east o f Williams Lake, British Columbia (52°05'N, 121°50'W). It is part o f the Fraser Plateau, which covers the majority o f the Cariboo Forest Region. The gently rolling landscape o f the study area is at an elevation o f about 1000 m , in a transitional location between the Interior Douglas-fir (IDF) Zone and the Sub-Boreal Pine-Spruce ( S B P S ) Zone. In 1989, three blocks (B, C and D) o f approximately 40 hectares each, were selected on the Knife Creek Unit, in areas selectively logged i n the early 1960's. In many places, the cut apparently removed much o f the standing sawtimber and released advance regeneration, most o f it Douglas-fir. When the project began, this regeneration had grown into a large number o f dense patches o f small-diameter trees (1-10 c m dbh). The plots were located i n these relatively uniform and dense patches, where spacing was expected to have a marked impact. Although all the plots were subjectively classified as dense before spacing, densities among the plots ranged between 4,260 and 13,720 stems per hectare. Douglas-fir was the leading species i n most plots, but lodgepole pine (Pinus contorta var. latifolia Engel.) and spruce (Picea glauca (Moench), Picea  10  engelmanni Parry and their crosses) were locally abundant. White birch (Betula papyrifera Marsh.) and aspen {Populus tremuloides Michx.) were present, but infrequent.  3.2 S T U D Y D E S I G N Each o f the three blocks was divided into quarters. Three o f the quarters received different, randomly assigned, spacing treatments, and the fourth was used as a control. T w o 0.05 ha plots were established i n each o f the quarters resulting in 8 plots per block and a total o f 24 plots for the study. The objective o f the study was to examine the impact o f the spacing regimes i n the locations most likely to show the largest response. A s a result, plots were not randomly located within experimental units, but were placed in dense patches that were relatively uniform i n composition and large enough to accommodate the plots. Although both dense patches within experimental units and plot centers within dense patches were not randomly chosen, it is assumed that plots are truly representative o f stands o f the area, within the range o f density indicated above.  3.3 T R E A T M E N T S Spacing was undertaken during the late fall and early winter o f 1990-91. The three spacing regimes were the 1992 standard spacing prescription o f the British Columbia Ministry o f Forests for these types o f stands and two clumped spacing regimes. Standard spacing: The focus o f the Standard prescription (S) was to allow at least 0.75 m between small (<12.5 cm) Douglas-fir and spruce, and either 2.5 or 2.8 m for small trees o f other species. The spacing and density rules were  11  1. A l l Douglas-fir and spruce greater than 25 c m diameter at breast height (dbh) were left uncut, regardless o f spacing, provided that they were i n good form. 2. A l l Douglas-fir and spruce between 12 and 25 c m dbh were left uncut, providing they were i n good form and were at least 0.75 m apart. 3. Trees less than 12 c m dbh were spaced without regard to trees 12 c m dbh and greater, providing they were at least 0.75 m apart and were not subject to crown competition. Crown competition was said to exist between trees closer than 1.5 m when the top o f the shorter tree exceeded the midpoint o f the live crown o f the taller tree. 4.  Spacing for trees less than 12 cm dbh depended on species. Adjacent lodgepole  pine was spaced to an optimum distance o f 2.8 m (1500 stems/ha), with an allowable variation from 1.5 to 4.0 m. Other species were spaced to an optimum distance o f 2.5 m (1800 stems/ha), with an allowable variation from 1.5 to 3.5 m. 5. Spacing around the edges o f openings with a diameter o f 5 m or greater was reduced to one half the optimum spacing detailed in Rule 4. 6. Douglas-fir and spruce less than 1.0 m i n height and lodgepole pine less than 0.5 m i n height were not cut, regardless o f spacing.  Clumped spacings: The objective of these spacing regimes was to produce clumps o f trees from one height class. A clump is defined as consisting o f 3-9 trees i n a circle 3 m i n radius. The two treatments differ i n terms o f the area in which trees less than 25 c m were removed around each clump. The 3 m Clumped spacing ( C l ) had a 3 m band o f spacing around each clump (1113 trees/ha left), and the 5 m Clumped spacing (C2) had a 5 m band o f spacing around each clump (668 trees/ha left). Additional spacing and density rules were 1. The height class o f the clump after spacing was the same as the height class o f the best formed and best stocked trees in the clump before spacing. 12  2. Height classes were considered to be: (i) Class 1 = 1-3 m ; (ii) Class 2 = 3-7 m ; (iii) Class 3 = 7-15 m ; (iv) Class 4 = greater than 15 m. 3. Douglas-fir, spruce and lodgepole pine less than 1.0 m i n height were not cut, regardless o f spacing. 4. A l l Douglas-fir, spruce and lodgepole pine greater than 25 c m dbh, in or out o f clumps, were not cut, regardless o f spacing, provided that they were o f good form. 5. Trees within a clump were spaced to an optimum distance o f 2.1 m , with an allowable variation from 0.5 to 2.5 m. The aim was to yield an average o f 7 trees per clump. 6. Trees o f a greater height class were left i n a clump, without regard for their spacing, provided that crown competition did not exist. Crown competition was assumed to exist when the leader and upper branches o f the shorter tree were suffering mechanical damage from the branches o f the taller tree. If crown competition existed, an alternative was selected for the shorter leave tree. 7. Deciduous trees within the clump were felled i f the crop trees were suffering crown competition from the deciduous tree. 8. Clumps could be left immediately adjacent to each other, provided that there was a height difference o f at least 3 m between the clumps. 9. Trees larger than 25 c m dbh outside o f clumps did not affect the inter-clump distance. Deciduous trees were also left outside the clumps and did affect the inter-clump distance.  13  3.4 D A T A  3.4.1 Measurements During the summers o f 1989 and 1990, plot locations were selected and living trees taller than 1.3 m within the plot boundary were tallied for species and dbh. Spacing was carried out during the fall and winter o f 1990-1991. During the spring and summer o f 1993, all living trees taller than 1.3 m within the confines o f the plots and those trees greater than 10 c m dbh within a 5 m distance o f the plot boundary were permanently tagged, and measured. The plots were remeasured during the spring and summer o f 1997. Diameters o f all living trees were measured i n the spring, before the end o f M a y , and other measurements did not include current year foliage, in order to exclude the 1997 growing season from the data. The measurements, standards and protocols used in previous measurements, and documented i n Marshall (1996), were followed. However, i n 1997, trees smaller than 5 cm dbh were measured using a calliper instead o f a diameter tape, and trees taller than 9 m were measured using an Impulse laser dendrometer instead o f a Suunto clinometer. The following measurements and observations were made for each tagged tree:  1 Species; 2 Diameter at breast height (dbh): diameter at 1.3 m along the stem, outside bark, measured using a calliper on the smaller trees (less than 5 cm) and a diameter tape on larger trees. D b h was recorded to the nearest 0.1 cm; 3  Total tree height: measured using a height pole on the shorter trees (less than 9  m) and an Impulse laser dendrometer on taller trees. The measurement was recorded to the nearest 0.1 m ; 4 Height to the base of the live crown: four heights were determined using either a height pole or an Impulse laser dendrometer and recorded to the nearest 0.1 m. The tree 14  was divided into quarters, proceeding around the tree i n a clockwise direction, with the first quarter to the left o f the tagged side o f the tree. The height to the lowest living, nonepicormic branch was measured for each quarter; 5  Crown diameter, recorded in two directions to the nearest 0.5m: first, parallel to  the direction o f the tag, and second, at a right angle to the tag; 6  Tree vigour; subjective code assigned to each tree based on the quantity and  quality o f its foliage, the development o f its crown, whether or not it was overtopped, etc. Four classes were assigned: 0 indicated that the tree was dead; 1 indicated that the tree was alive, but had little potential for future development; 2 indicated moderate potential for development; and 3 indicated good potential for development; 7 Ingrowth: all plots were examined for ingrowth trees. A n y such trees were assigned a unique tree number and tagged, and were measured for the same attributes as the other trees on the plot; 8 Age: age measurements focused on Douglas-fir. Lodgepole pine and spruce were also taken into account when they represented more than 20% o f the basal area o f the plot. For each o f the plots, trees outside the plot and the buffer strip were selected from up to 30 m as the first two nearest trees from the plot corner stake number 1, within each o f three dbh classes (0-15 cm, 15-30 cm, >30 cm). The age at stump height (0.3 m) was determined from an increment core. The ring count was made i n the field using a magnifying lens.  3.4.2 Derived Attributes Tree and stand summary statistics discussed i n this report were obtained by the following procedures: 1. Quadratic mean diameter ( Q M D ) is the diameter at breast height o f the tree o f arithmetic average basal area. 15  2. Lorey's height is a mean height, with the individual trees weighted proportionally to basal area (van Laar and A k c a 1997, p. 146): Lorey's height = (S (tree height) (tree basal area)) / (basal area/ha) 3. Crown length was obtained as the difference between the total height and the average o f the four heights to the base o f the live crown. 4. Crown area was estimated using the following formula: (El / 4) ( D i + D ) / 2 2  2  2  where D i and D2 are crown diameters measured in two perpendicular directions, parallel to the plot sides. 5. Total stem volume, inside bark, was estimated for each tree using the British Columbia provincial volume equations ( B . C . Forest Service 1976). 6. Biomass o f Douglas-fir (stump wood, stump bark, stem wood, stem bark, living branches, living needles, sum o f all six biomass components) was calculated using two different sets o f equations developed by Marshall and Wang (1996) for small and large trees. Stump wood and stem wood were then merged in one biomass component (wood), as well as stump bark and stem bark (bark). Biomass o f other species (wood, bark, branches, foliage, total o f all four biomass components) was predicted using equations provided by Standish et al. (1985). In a few cases, the estimated value o f some o f the biomass components for some small trees (less than 5 cm dbh) was negative. When this occurred, the negative value was replaced with zero. 7. De Liocourt's quotient (q) was calculated for each treatment using the leastsquares method described by Leak (1963). Numbers o f trees were listed by 5 c m dbh classes, and quotient values were computed using only those trees between 5 and 3 5-cm dbh,  16  inclusive, to avoid large variation i n the number o f small trees as w e l l as the erratic distribution o f larger trees. 8. Curtis' (1982) relative density index ( R D ) was determined for each plot. It was calculated as: R D = (basal area) / (square root o f quadratic mean diameter) This index is a variation o f Reineke's (1933) stand density index (SDI), which was also computed with the formula provided by L o n g (1985): SDI = (stems per ha) ( Q M D / 2 5 )  1 6  9. Tree growth efficiency Growth efficiency was expressed as the average ratio o f individual tree stem volume growth (m ) to crown projection area (m ), as defined by O'Hara (1988).  3.4.3 Error Checking In the 1993 measurements, the data included 3,709 tree observations. Five observations recorded at that time were removed from the data: three o f them concerned dead trees and the two others, trees barely alive due to girdling. These trees were considered as dead trees, felled during the spacing. The corrected data included 3,704 tree observations. During the 1997 measurement, data were checked i n the field for inconsistencies by comparison with the 1993 records. When an inconsistency occurred, the tree was re-measured and a check was put on the record sheet. Measurement data were entered into a spreadsheet format, within a few days o f being taken. Data were checked again before analysis. The second measurement was assumed more reliable than the first, and used to correct inconsistencies. The recorded species changed for two trees. Except for one diameter measurement, other inconsistencies were related to height measurements. Some o f them were due to calculation errors (5) or assumed to be recording 17  errors (12). After a last check in the field in 1998, the remaining inconsistencies (19) were corrected i n the 1993 data, using the height growth average o f the species i n the diameter class o f the tree.  3.4.4 Outliers Exceptional growth in diameter and height that was not due to obvious inconsistencies was not corrected. Example o f this type include 200 negative diameter growths (up to -6 mm) for small trees (less than 5 cm dbh), and 237 negative height growths (up to -6.9 m) due to broken tops or die back.  3.4.5 Ecological Data Each plot was classified according to the biogeoclimatic ecosystem classification ( B E C ) system within the Cariboo Forest Region (Steen and Coupe 1997). Site identification was based on careful observation and accurate description o f vegetation, soil and site features.  3.5 A N A L Y S I S  3.5.1 Variables of Interest Variables o f interest at the stand level were growth and growth rate o f quadratic mean diameter, basal area per hectare, Lorey's height, volume per hectare and accumulated crown area per hectare. Growth was calculated as the difference between the live stand at the start (1993 growing season) and at the end o f the growth period (1996 growing season). Variables o f interest for individual trees were growth i n diameter, basal area, total height, total stem volume, crown area and live crown length. Growth was the change i n the attribute during the same period o f time given above. The second and third null hypotheses were tested by species or group o f species (Douglas-fir, other species, all species combined) and by diameter class (0-5 cm; 5-10 cm; 10-15 18  c m and 0-15 cm; >15 cm). A t the tree level, statistical analysis was performed on arithmetic averages o f individual tree growth by diameter class and by species, or group o f species. A t the stand level, statistical analysis was performed on arithmetic averages o f stand growth attributes.  3.5.2 Statistical Methods Analysis o f variance ( A N O V A ) was used to judge the significance o f treatment effects. The analysis was carried out assuming a randomized complete block design with three blocks, four treatments and two plots (i.e, sampling units) for each o f the twelve experimental units. Treatments were fixed, and blocks were considered to be randomly chosen from a population o f blocks. The linear model with m observations per experimental unit is: Yiji = P- + Pi + Tj + COij + S (ij)i where Yiji  is the observation on the tree / from the block i on treatment j  \i  is the overall population average  Pi  is the effect o f block i  Tj  is the effect o f treatment j  coy  is the experimental error i n treatment j on block i  s (ij)i  is the sampling error within block / and treatment j  The response variables were assumed to be normally distributed. Homogeneity o f variance o f the response variable was tested using a Bartlett's test for samples o f unequal sizes (Snedecor and Cochran 1980, p. 252). Under the null hypothesis that the variances are the same, Bartlett's'test statistic has a chi-square (% ) distribution. When the 2  null hypothesis o f equality o f variance was rejected, the original variable was transformed, using 19  a power transformation. The transformed variable was tested the same way. When homogeneity of variances could not be established using a transformed variable, the analysis was carried out on the variable (original or transformed) that provided the lowest % value. 2  A n analysis o f variance was performed to examine growth response o f individual trees among treatments. A n analysis o f covariance ( A N A C O V A ) was performed to determine whether the diameter at first measurement affected growth response o f individual trees among diameter classes, with the initial diameter measurement as a covariate. When the effect o f initial diameter was significant, parallelism among the regression lines within diameter classes was tested. W h e n the assumption o f parallelism could not be rejected, treatment means were adjusted. A n A N O V A was performed on the adjusted means when initial diameter had a significant effect and regression lines were parallel. When these two conditions were not met, an A N O V A was performed on the unadjusted means. A t the stand level, an A N O V A was also performed to analyze growth in absolute and relative terms. When an A N O V A showed a significant treatment effect, a Scheffe's multiple range test was performed on the treatment means (Hicks 1993, p. 63). Scheffe's test was chosen because it can be used for samples o f unequal sizes. In addition, it allows the simultaneous comparison o f multiple treatment means while fixing the probability o f finding an erroneous significant result (Type I error) at a maximum o f 0.05 (Steel et al. 1997, Chapter 8). Statistical analyses were performed with SAS® (Version 6.12). Some plot attributes were also calculated for confirmation using an Excel® (Version 97) spreadsheet. For all tests, a level of 0.05 was used to judge statistical significance.  20  4. R E S U L T S  4.1 P L O T G E N E R A L F E A T U R E S  4.1.1 Site Identification A s would be expected, site conditions were homogeneous within each o f the three blocks (Table 1). Blocks B and C belonged to the Fraser variant o f the interior Douglas-fir dry cool subzone (IDFdk3), and block D belonged to the sub-boreal pine-spruce moist cool subzone (SBPSmk). O n each o f the 24 plots, sites series were identified as zonal, although some variation i n ecosystem attributes was recorded, both within blocks and within some o f the plots. For instance, plots 1, 3 and 4 from block B were slightly drier sites than zonal sites and tended towards 06 F d Feathermoss - Aster Site Series. Similarly, a gully running through plot 9 i n block D could have been identified as 06 S x w - Twinberry Site Series. However, variation i n the site features was minor, and did not justify any change in site identification. Zonal soils o f IDFdk3 and S B P S m k are predominantly Orthic Gray Luvisols, with a clayenriched horizon (Steen and Coupe 1997). In the study area, soil texture was generally silt loam in the eluvied horizon (Ae) and silt clay or silt clay loam i n the clay-enriched horizon (Bt). The latter, extremely dry and compact i n block B , was within 12-20 c m o f the soil surface, and 1  visibly restricted the penetration o f the roots. Both horizons were moderately gravelly. The surface organic layer was very thin (2-4 cm).  1  Ecological data were recorded at the end of May 1998.  21  Table 1. Site identification. Block  Treatment  Plot  B  Cl  1 2 3 4 5 6 7 8  IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl  9 10 11 12 13 14 15 16  SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica SBPSmk/01 PI - Pinegrass - Arnica  17 18 19 20 21 22 23 24  IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl lDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl IDFdk3/01 FdPl !DFdk3/01 FdPl  Control C2 S  D  S Cl C2 Control  C  C2 Control Cl S  Site series  22  - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss  - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss - Pinegrass - Feathermoss  The climate o f the I D F is characterized by cool, dry winters, and by warm, dry summers (Hope et al. 1990). More specifically, the annual precipitation in the IDFdk3 is about 433 m m (53% o f which falls as snow) and the daily temperature averages 3.3 °C, with a m i n i m u m o f 10.3 °C for the coldest month and a maximum o f 14.7 °C for the warmest (Steen and Coupe 1997). The I D F d k undergoes soil moisture deficit five months a year (Loyd et al. 1989, M i t c h e l l and Erickson 1989, in Bonnor 1990). The climate o f the S B P S is characterized by cold, dry winters and cool, dry summers (Steen and Coupe 1997). In the SBPSmk, the annual precipitation is about 506 m m and the daily temperature averages 3.2 °C, with a minimum o f -10.3 °C for the coldest month and a m a x i m u m of 13.7 °C for the warmest (Steen and Coupe 1997).  4.1.2 Age Structure In each o f the tree blocks, the sample suggested that small trees (<15 c m dbh) were on average markedly older than the disturbance created by the logging, 37 years ago (Table 2). 1  When logging took place, advance regeneration was already 29 to 44 years stump age (66 to 81 minus 37). W i t h a correction o f 7 to 15 years applied to age at stump height (Clark 1952), total age was then 36 to 59 years. The sample also suggested that age increased with increasing dbh. However, the age differences between small and medium-sized trees (15-30 c m dbh) were minor compared to those between medium-sized and large trees (>30 c m dbh). Most Douglas-fir less than 30 c m dbh appeared to belong to the same cohort, about 66-91 years old. A few trees o f all species belonged to a new cohort originating from the 1961 disturbance, which still allows seedlings to develop on  1  According to the forest cover map, the cutting occurred in 1961.  23  Table 2. Average age (years) at stump height of sample trees of different species by size and by block. Block  Species'  D b h Class  N  Mean  Min  Max  Stdev  (cm)  Fd  0-15 15-30 >30  16 16 14  80.6 90.5 160.4  47 64 89  102 113 >265  14.8 15.7  PI  0-15 15-30 >30  2  102.0  87  117  21.2  A l l Species  0-15 15-30 >30  16 18 14  80.6 91.8 160.4  47 64 89  102 117 >265  14.8 16.1  Fd  0-15 15-30 >30  16 16 13  66.3 78.8 135.9  28 55 69  88 113 201  14.5 17.4 39.8  PI  0-15 15-30 >30  2 4 1  80.5 90.3 107.0  74 65 107  87 110 107  9.2 23.2  A l l Species  0-15 15-30 >30  18 20 14  67.8 81.1 133.9  28 55 69  88 113 201  14.5 18.6 39.8  Fd  0-15 15-30 >30  16 16 13  67.5 75.8 132.8  43 42 85  139 154 195  21.7 23.5 28.8  PI  0-15 15-30 >30  2 6  29.0 90.5  26 72  32 118  4.2 19.7  Sx  0-15 15-30 >30  6 6 1  56.5 73.2 82.0  22 66 82  75 82 82  21.1 5.9  A l l Species  0-15 15-30 >30  24 28 14  61.5 78.4 132.8  22 42 82  139 154 195  23.0 20.7 30.8  Codes used in this document are B.C. Ministry of Forests inventory codes (At: trembling aspen; Et: white birch; Fd: Douglas-fir; PI: lodgepole pine; Sx: spruce). 1  24  former skid trails and other openings. So far, very few seedlings have established following the 1990 spacing. Trees greater than 30 c m dbh may possibly belong to more than two different older cohorts. Except for one tree 65.2 cm dbh, all large trees sampled (42) were less than 50 c m dbh. Furthermore, trees were older in the driest block (B), regardless o f their size. In block D , although few spruce trees were sampled, they appeared to be on average younger than Douglasfir.  4.1.3 Plot Conditions Plot statistics for live trees at each measurement (density, quadratic mean diameter, basal area, Curtis' (1982) relative density, volume, species composition) are summarized i n Tables 3 and 4. Accumulated crown area per ha is shown i n Table 5. In 1997, as w e l l as i n 1993, density, basal area per ha, Curtis' (1982) relative density and to a lesser extent, volume per ha and accumulated crown area per ha, were considerably higher on the control plots than on the spaced plots. Conversely, quadratic mean diameter and Lorey's height were lower on the control plots. Figure 1 shows marked differences in the volume distribution between control and unspaced plots. The greater growing stock in the controls is composed o f relatively small trees, while the spaced plots have larger volumes i n larger diameters. Plots that received treatment C 2 averaged lower density, Curtis' (1982) relative density index, basal area per ha, volume per ha, and accumulated crown area per ha, than plots in the other spacing regimes. The summary for biomass is shown in Tables 6 and 7. In 1997, total above ground biomass was the highest on the control (137.7 t/ha), and the lowest on treatment C 2 (92.9 t/ha). Treatment C l and S were similar with about 120 t/ha. Foliage biomass on the control plots averaged about 14 t/ha, while it appeared markedly below this level on spaced plots, especially  25  Table 3. Summary of plot conditions at the 1993 measurement. Treatment Block Plot  Cl  B D C  C2  B D C  S  B D C  Control  B D C  Density (trees/ha)  Basal Area (m /ha)  1 2 11 12 21 22  2,260 1,960 3,100 2,300 2,440 2,800  18.0 26.7 36.8 27.2 24.7 22.6  10.1 13.2 12.3 12.3 11.3 10.1  5.7 7.4 10.5 7.8 7.3 7.1  9.8 13.0 15.3 12.4 13.2 12.2  70.6 135.2 230.0 138.7 129.8 108.3  Avg  2,477  26.0  11.6  7.6  14.0  135.4  5 6 13 14 17 18  2,060 1,400 1,500 1,240 1,140 1,280  25.2 22.4 18.9 18.4 14.8 14.7  12.5 14.3 12.7 13.7 12.9 12.1  7.1 5.9 5.3 5.0 4.2 4.2  15.2 16.1 11.8 14.1 11.0 10.5  153.2 126.1 89.5 116.0 63.0 61.0  Avg  1,437  19.1  13.0  5.3  13.5  101.5  7 8 9 10 23 24  1,940 1,720 2,300 1,860 1,720 1,440  26.3 22.9 25.7 23.7 27.5 21.2  13.1 13.0 11.9 12.7 14.3 13.7  7.2 6.3 7.4 6.6 7.3 5.7  11.8 13.5 14.1 13.0 18.0 14.3  120.3 123.5 149.4 120.2 193.6 122.9  Avg  1,830  24.6  13.1  6.8  14.2  138.3  3 4 15 16 19 20  10,460 8,240 5,500 4,340 5,600 5,480  43.7 39.8 36.2 33.6 32.9 32.4  7.3 7.8 9.1 9.9 8.6 8.7  16.2 14.2 12.0 10.7 11.2 11.0  10.6 10.4 12.0 14.1 9.7 10.0  188.3 166.9 179.0 194.8 126.9 131.0  Avg  6,603  36.4  8.4  12.6  11.1  164.5  2  Q M D Relative Density (cm)  26  Lorey's Volume Height (m) (m /ha) 3  Species Composition (% Basal Area)  Fd Fd  100 100  Fd 4Pl SX43At9 4  4  Fd Sx FdsgPl^SXjAtoEpzi Fd 7Pl Sx Ep 6  13  61  39  4  16  Fd Pl At Fd FdsoPhoAto Fd Pl Sx 72  28  0  100  24  67  9  Fd7 Pli5Sx Ep 2  5  8  Fd oPli3Sx AtoEp 7  8  9  Fd Fd Pl Fd Pl Ato Fd Pl Ep Fd Pl SxoAtoEpo Fd Pl 100  84  76  16  77  23  86  14  0  76  24  24  Fd Pli Fd Pl FdesPlnSx^teEpu Fd iSx At Ep Fd Sx Ep Fd Pl At Ep 99  96  3  78  4  59  3  7  95  3  2  9  6  7  Table 4. Summary of plot conditions at the 1997 measurement. Treatment Block Plot  Cl  B D C  C2  B D C  S  B D C  Control  B D C  Density (trees/ha)  Basal Area (m /ha)  1 2 11 12 21 22  2,260 1,960 3,040 2,280 2,380 2,720  22.9 29.9 39.7 30.8 27.9 25.9  11.4 13.9 12.9 13.1 12.2 11.0  6.8 8.0 11.1 8.5 8.0 7.8  10.8 13.8 16.1 13.4 13.8 13.1  98.3 158.4 259.8 168.9 153.0 132.5  Avg  2,440  29.5  12.4  8.4  13.7  161.8  5 6 13 14 17 18  2,040 1,400 1,420 1,280 1,200 1,440  28.6 25.4 22.3 21.8 18.6 18.5  13.3 15.2 14.1 14.7 14.0 12.8  7.8 6.5 5.9 5.7 5.0 5.2  14.8 16.3 12.8 14.9 11.9 11.2  166.2 145.2 113.3 143.5 85.1 80.9  Avg  1,463  22.5  14.0  6.0  13.9  122.4  7 8 9 10 23 24  1,920 1,680 2,400 1,920 1,740 1,420  29.7 25.9 28.4 27.0 30.3 25.4  14.0 14.0 12.3 13.4 14.9 15.1  7.9 6.9 8.1 7.4 7.8 6.5  12.6 14.2 14.7 13.6 18.2 15.0  144.9 145.8 170.8 143.1 213.3 152.0  Avg  1,847  27.8  13.8  7.5  14.8  161.7  3 4 15 16 19 20  9,780 7,960 5,400 4,000 5,440 5,180  45.7 42.3 37.5 35.2 36.6 35.3  7.7 8.2 9.4 10.6 9.3 9.3  16.5 14.8 12.2 10.8 12.0 11.6  11.1 10.9 12.7 14.8 10.6 10.7  206.0 184.4 195.0 213.3 152.4 150.8  Avg  6,293  38.8  8.9  13.0  11.7  183.7  2  Q M D Relative Density (cm)  27  Lorey's Volume Height (m) (m /ha) 3  Species Composition (% Basal Area) Fd Fd  100 100  Fd4 Pl SX43At9 4  4  Fd Sx 62  38  Fd oPli5Sx AtoEp2o 6  5  Fd Pl Sx Ep 67  15  4  14  Fd Pl Ato Fd, Fd Pl oAto Fd Pl Sx Fd Pli Sx Ep FdyjPlnSxgAtoEpg 74  26  00  80  2  25  72  65  4  10  6  8  Fdioo Fd Pl Fd Pl At Fd Pl Ep Fd Pl Sx AtoEpo Fd Pl 85  77  87  77  23  23  15  0  13  0  77  23  0  Fd Pl, Fd Pl Fd Pl Sx At Ep Fd Sx At Ep Fd Sx Ep Fd Pl At Ep 99  97  68  4  32  59  3  6  95  3  2  9  4  7  80  3  3  12  13  Table 5. Summary of accumulated crown area per ha in 1993 and 1997. Treatment  Block  Plot  Total Crown Area 1993 (m /ha)  Total Crown Area 1997 (m /ha)  Crown Area Expansion (%)  1 2 11 12 21 22  6,189 8,164 11,274 8,341 10,814 10,375  11,292 10,229 14,447 15,881 12,069 10,300  82.5 25.3 28.1 90.4 11.6 -0.7  Average  9,193  12,370  34.6  5 6 13 14 17 18  7,622 5,409 5,289 4,392 5,069 5,713  13,230 7,971 7,563 5,535 6,972 7,202  73.6 47.4 43.0 26.0 37.5 26.0  Average  5,580  8,079  44.8  7 8 9 10 23 24  8,301 10,345 7,573 7,543 8,097 6,371  10,877 11,949 9,287 10,860 9,422 9,014  31.0 15.5 22.6 44.0 16.4 41.5  Average  8,038  10,235  27.3  3 4 15 16 19 20  17,997 14,665 14,349 9,721 12,868 15,279  18,633 16,748 15,399 10,960 14,924 17,082  3.5 14.2 7.3 12.7 16.0 11.8  Average  14,147  15,624  10.4  2  Cl  B D C  C2  B D C  S  B D C  Control  B D C  28  2  Figure 1. Volume distribution by dbh class and by treatment i n 1997. D b h classes are 5 c m wide (dbh class 1 = 0-5 cm, dbh class 11 = 50-55 cm).  29  Table 6. Biomass of various tree components at the 1993 measurement (t/ha). Treatment B l o c k  Cl  B D C  C2  B D C  S  B D C  Control  B D C  Plot  Wood  Bark  Branches  Foliage  Total  1 2 11 12 21 22  41.9 75.8 106.9 66.8 67.0 53.1  9.9 17.5 20.2 14.2 13.3 12.0  7.4 9.0 20.1 16.4 11.4 8.3  5.0 8.7 16.1 11.3 8.2 7.7  57.1 104.3 161.2 104.4 96.7 77.8  Avg  68.6  14.5  12.1  9.5  100.2  5 6 13 14 17 18  78.1 68.8 48.0 59.7 31.9 31.0  14.8 18.2 11.3 7.1 8.7 8.0  11.1 15.0 8.2 11.6 7.4 7.8  8.4 6.1 5.6 7.1 4.8 4.6  108.3 105.7 68.1 84.1 50.2 47.1  Avg  52.9  11.4  10.2  6.1  77.3  7 8 9 10 23 24  68.0 67.3 81.9 65.8 97.7 67.0  16.9 13.5 13.9 14.4 18.7 12.1  10.6 11.5 10.5 11.9 11.3 10.9  7.8 8.9 8.2 7.0 8.9 7.0  94.0 92.4 108.9 92.8 135.5 92.2  Avg  74.6  14.9  11.1  7.9  102.6  3 4 15 16 19 20  107.1 92.2 96.0 87.8 69.2 69.9  22.8 21.1 18.4 18.2 17.9 17.2  18.4 16.6 17.1 14.4 12.5 12.8  15.6 13.3 11.2 16.7 10.1 10.0  145.8 128.1 129.1 136.8 96.7 99.4  Avg  87.0  19.3  15.3  12.8  122.7  30  Table 7. Biomass of various tree components at the 1997 measurement (t/ha). Treatment B l o c k  Cl  B D C  C2  B D C  S  B D C  Control  B D C  Plot  Wood  Bark  Branches  Foliage  Total  1 2 11 12 21 22  57.7 88.1 120.2 81.8 80.1 66.2  13.5 20.0 22.2 16.7 15.8 14.7  9.0 12.0 23.2 23.6 15.0 12.0  8.2 10.7 18.0 15.7 9.5 8.6  79.3 121.3 179.7 126.6 117.0 96.6  Avg  82.3  17.2  15.8  11.8  120.1  5 6 13 14 17 18  87.3 78.7 59.8 73.0 41.9 40.0  17.3 20.8 13.8 8.7 11.4 10.7  18.2 17.7 7.5 14.3 9.1 9.5  11.4 8.0 7.8 8.3 6.1 6.3  123.4 120.7 85.2 102.5 64.9 60.5  Avg  63.4  13.8  12.7  8.0  92.9  7 8 9 10 23 24  81.3 79.1 93.1 77.6 108.5 82.1  19.6 15.8 16.0 17.0 21.0 15.4  12.1 15.3 11.9 13.5 15.3 13.9  10.2 10.5 10.0 9.7 10.2 9.4  112.7 108.9 124.2 109.8 151.1 114.4  Avg  86.9  17.5  13.7  10.0  120.1  3 4 15 16 19 20  118.6 102.9 106.1 96.4 83.0 81.5  24.7 23.1 19.7 19.3 20.8 19.4  20.1 17.6 17.1 18.8 13.8 15.2  16.5 13.8 12.6 17.6 12.2 11.4  161.2 141.4 141.6 151.2 116.1 114.6  Avg  98.1  21.2  17.1  14.0  137.7  31  on treatment C 2 . L o n g (1985) suggested that the lower limit o f self thinning in pure even-aged stands occurs at about 60% o f maximum Reineke's (1933) stand density index (SDI), and that the lower limit o f "full site occupancy" is at about 35%. Since some plots were pure or quasi-pure Douglas-fir stands with little age differentiation, SDI were also computed. " K e y " S D I values for interior Douglas-fir are shown in Table 8. Average SDI values for the four treatments are reported i n Table 9, while results at the plot level were computed but are not shown. The control plots averaged an initial SDI o f 1148, much higher than the lower limit o f self-thinning (870). In 1993, control plot 3 with a SDI o f 1459 was even above the maximum o f 1450 provided by L o n g (1985). In the spaced plots, SDI averaged that o f stands between the lower limit o f "full site occupancy" and the lower limit o f self-thinning, but some plots that received treatment C 2 (13, 14, 17 and 18) were below the level o f "full site occupancy".  4.2 S T A N D - L E V E L R E S P O N S E  4.2.1 Mortality Some mortality occurred on plots in all treatments (Table 10), but i n terms o f both stems per ha and basal area per ha, it was much higher on the control plots where it reached 5.0% and 2.3% o f the 1993 total, respectively. The least amount o f mortality occurred on plots that received the 5 m clumped spacing, representing on average 1.4 % o f the number o f stems per ha and 0.05 % o f the basal area per ha. Plots i n the 3 m clumped areas averaged about the same amount o f mortality as the plots i n the standard spacing areas. Tree mortality i n the controls was mainly confined to the smaller diameter trees and was presumably due to self-thinning. In the spaced plots, overbrowsing on young deciduous trees was the predominant cause o f mortality  32  Table 8. " K e y " SDI values for interior Douglas-fir (adapted from L o n g 1985). Percentage of maximum SDI  SDI  100 60 35 25  1450 870 510 360  Maximum Lower limit of self thinning Lower limit of "full site occupancy" On-set of competition  Table 9. Reineke's stand density index (SDI) by treatment i n 1993 and 1997.  1993 1997  Cl  C2  Treatment S  721 796  505 578  648 717  33  Control 1148 1197  Table 10. Summary of mortality and ingrowth between 1993 and 1997. Treatment  Cl  Block  B D C  C2  B D C  B D C  Control  B D C  Plot  Mortality  Ingrowth (trees/ha)  (tree/ha)  (nrVha)  (m /ha)  1 2 11 12 21 22  0 0 60 20 140 240  0.0 0.0 0.1 0.0 0.3 0.9  0.0 0.0 0.4 0.0 1.4 3.4  Avg  77  0.22  0.85  5 6 13 14 17 18  20 0 80 0 20 0  0.0 0.0 0.1 0.0 0.0 0.0  0.0 0.0 0.2 0.0 0.0 0.0  20 100  20  Avg  20  0.01  0.02  20  3  7 8 9 10 23 24  20 40 100 0 100 60  0.0 0.3 0.6 0.0 0.3 0.0  0.0 1.3 2.6 0.0 1.3 0.1  Avg  53  0.20  0.82  3 4 15 16 19 20  680 280 160 340 200 320  1.1 0.3 1.5 0.8 0.2 1.1  4.3 0.6 7.6 3.5 0.5 5.0  20 20  20  Avg  330  0.83  3.56  7  13  J  Ingrowth by species may not exactly sum to the totals because of rounding.  34  Fd  PI  At  Ep  All  40 140  40 20  0 0 0 0 80 160  30  10  40  20  20 60 40  0 0 0 40 80 160  3  20  47  20 40  0 0 200 60 120 40  10  70  200 60 60  40  10  7  43  60  0 0 60 0 40 20 20  among small trees. Most mortality among larger trees was the result o f wind and snow damage.  4.2.2 Ingrowth Some ingrowth occurred on plots in all treatments, including the control (Table 10). Ingrowth was mainly composed o f Douglas-fir, birch and aspen. Spruce seedlings were apparent in some places on blocks C and D , and could reach the minimum size for ingrowth inventory (1.3 m tall) in the near future. N o ingrowth occurred on the driest block (B) for any o f the treatments. Although there was high variability among plots, the control plots averaged the least amount.  4.2.3 Tree Size Distribution Figures 2 and 3 show at the same scale, the changes in the number o f trees by size class between 1993 and 1997, for the control and treatment C2. Treatments C l and S are not shown, but were similar to treatment C 2 . A s a result o f spacing, the frequency distribution o f number o f trees by dbh class i n the thinned plots presented the same non-symmetric bell-shaped pattern, despite different densities. O n these plots, a displacement o f the diameter distribution to the right was clearly visible from 5 up to 25 c m dbh. This displacement indicated a transfer o f trees from one dbh class to the other, as a result o f size increase (upgrowth). O n the control plots, the diameter distribution formed an inverted-J-shaped curve. Between 0 and 5 c m dbh, the reduction in the number o f trees caused by mortality was clearly distinct. Compared to treatment C 2 , a similar transfer o f trees occurred from 5 to 20 c m dbh. However, compared to the initial number o f trees, this transfer was limited, indicating a slower growth on control than on treatment C 2 . De Liocourt's quotient values (q) by treatment are shown i n Table 11. The q values for 1993 differed from those reported by Marshall (1996), who computed them using 2-cm dbh classes, between 5 and 29 c m dbh, inclusive. Between 1993 and 1997, quotient values decreased  35  Control  Control  200 100 0 ii  -100  i  m 3  i  1  4  l  5  6  7  8  9  10  11  -200 -300 -400 Dbh class  Figure 2. Changes in the number of trees by size class on Control, between 1993 and 1997. Dbh classes are 5 cm wide (dbh class 1 = 0-5 cm, dbh class 11 = 50-55 cm).  36  Treatment C 2  Treatment C 2 100  -100  2  10  11  > -200  -300 -400 Dbh class  Figure 3. Changes in the number of trees by size class on treatment C2, between 1993 and 1997. Dbh classes are 5 cm wide (dbh class 1 = 0-5 cm, dbh class 11 = 50-55 cm).  37  in all treatments, indicating a flatter diameter distribution, presumably as a result o f upgrowth. Appendices I A and I B give more details on the distribution o f trees by species, following the 1993 and 1997 measurements. Table 11. De Liocourt's quotient values (q) by treatment i n 1993 and 1997.  1993 1997  Cl  C2  1.43 1.41  1.40 1.38  Treatment S 1.38 1.32  Control 1.61 1.57  4.2.4 Growth Response 4.2.4.1 Overview Ten A N O V A s were performed to analyze growth response at the stand level. In one analysis, variances among treatments were found significantly different. There was a significant block-treatment interaction i n the analysis o f height growth and thus treatment means were not tested. For all other variables, growth responses were consistently lower i n the control plots than in the spaced plots. N o clear difference i n growth pattern could be seen among the spaced plots. Except for basal area, no significant difference was found among growth rates. However, growth rates were consistently lower i n the control plots than i n the spaced plots, for all variables but height. In the spaced plots, the 5 m clumped treatment showed consistently higher growth rates. Using Curtis' (1982) relative density index as a covariate, A N A C O V A s (not shown i n this report) demonstrated that density significantly affected Q M D and basal area growth, but had no effect on the growth response o f other variables.  38  4.2.4.2 Quadratic M e a n Diameter Q M D growth on C 2 was, on average, twice as much as on the control (Table 12). However, due to the difference between initial values, the difference i n a relative sense was considerably less. Q M D growth was similar for both C l and S. A s would be expected, Q M D growth declined with increasing relative density (Figure 4). Q M D growth was calculated as the difference between Q M D s at the start and end o f the growth period and hence, represents the combined effects o f actual biological growth, ingrowth and death o f trees (Curtis and Marshall 1989, Marshall and Tappeiner 1992). Mortality among small trees was markedly higher on control plots, widening the range o f Q M D s between start and end o f the growth period. Conversely, more ingrowth occurred on the spaced plots, narrowing their range o f Q M D s . Overall, both factors contributed to mask the actual biological growth. A s suggested by Curtis (1992), stand diameter growth can be described more accurately by the periodic diameter growth o f trees surviving at the end o f the growth period. Calculated this way, stand diameter growth was 0.83, 1.08, 0.80 and 0.37 cm, compared to 0.87, 1.01, 0.81 and 0.49 c m for Q M D growth on treatment C l , C 2 , S and Control, respectively. Biological growth differences between spaced and control plots were therefore larger than what was suggested by differences in Q M D , with stand diameter growth on treatment C 2 more than three times that o f the control. 4.2.4.3 Basal A r e a Similar to Q M D growth, basal area growth was negatively related to stand density (Figure 4), and 38 to 50% higher on spaced plots than on controls (Table 12). In an absolute sense, growth was fairly similar among spacing regimes, but growth rates were markedly lower in C l and S than i n C 2 . Basal area growth rates on C 2 and on the control were found to be  39  Table 12. Mean of stand-level growth response by treatment between 1993 and 1997 for five variables of interest . 1  Scale  Absolute  Variable  Variance  C2  S  Control  Quadratic Mean Diameter (cm)  0.87 a  1.01 a  0.81 a  0.49 a  Basal Area (m /ha)  3.52 a  3.43 a  3.23 a  2.34 a  Lorey's Height (m)  0.83  0.53  0.61  0.61  Volume (m /ha)  26.38 a  20.91 a  23.34 a  19.16 a  Accumulated crown area (m /ha)  3,176 a  2,496 a  2,196 a  1,477 a  Quadratic Mean Diameter  7.5 a  7.8 a  6.2 a  5.8 a  Basal Area  13.5 ab  18.0 b  13.2 ab  6.4 a  Lorey's Height  5.9  3.9  4.3  5.5  Volume  17.3 a  20.6 a  16.9 a  11.7 a  Accumulated Crown Area  34.6 a  44.8 a  27.3 a  10.4 a  2  3  2  Relative (%)  Treatment Cl  Identical letters under the means indicate no statistical difference at the 0.05 probability level. One star (*) indicates that the variances among treatments are significantly different at a 0.05 probability level. There was a significant block-treatment interaction in the analysis of Lorey's height growth and thus treatment means were not tested.  40  Quadratic mean diameter growth  a  1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0  Basal area growth  5 o  cn "ra" ro =5  <D CM  -I •  ro u> ro  •  m  5  10  15  5  20  Lorey's height growth  40  0.5 CD  o  •  •  egnowth 3/ha  1.0  •  •  E E. O  >  •_  -0.5 5  20  15  20  Volume growth  1.5  0.0  15  Relative density  Relative density  o  10  10  15  n  30 20 10 0-  20  5  Relative density  10 Relative density  Accumulated crown area growth  5 ro  o  £Z  O) csi  ro E £ o ro S  ii o  •  5  10  15  20  Relative density  Figure 4. Stand-level growth response in relation to Curtis' (1982) relative density for five variables of interest between 1993 and 1997.  41  significantly different. Extrapolating the results over a 10-year period, net basal area growth would be 5.8 m /ha on control plots, and would range from 8.0 to 8.8 m /ha i n the thinned stands. 4.2.4.4 Lorey's Height Outliers disrupted the analysis o f height growth. Due to broken tops or die back, 237 negative values were recorded, especially on spaced plots. Since some o f them were large, occurring on big trees, they had a marked impact i n the computation o f Lorey's height. A s a result, no consistent pattern o f growth appeared among treatments (Table 12). 4.2.4.5 Volume A s shown i n Table 12, volume production on the spaced plots was greater than on the unspaced controls, despite lower initial growing stocks (Table 3). O n the spaced plots, treatment C 2 , with the lowest initial growing stock, showed the lowest production, but the highest growth rate. Growth rates on treatment C l and S were similar. Overall, volume growth rates seemed closely related to basal area growth rates, and volume growth slightly related to Curtis' (1982) relative density (Figure 4). In addition, volume growth seemed to vary i n a similar way over a wide range o f growing stock (Figure 5). For the 4-year period between 1993 and 1997, net periodic annual volume increment averaged 5.2 to 6.6 m per ha on the spaced plots and 4.8 m 3  3  per ha on the controls. 4.2.4.6 Accumulated Crown Area Although there were no significant differences among treatments, growth rates were considerably higher on the spaced plots, especially on C 2 (Table 12). In an absolute sense, accumulated crown area growth was the greatest on C l , with considerable variation from plot to plot. These trends are consistent with foliage biomass increments (Tablel3).  42  Figure 5. Net volume growth between 1993 and 1997 in relation to 1993 growing stock.  43  Table 13. Net biomass growth (%) of various tree components between 1993 and 1997. Treatment  Block  Plot  Wood  Bark  Branch  Foliage  B  1 2 11 12 21 22  37.8 16.2 12.5 22.4 19.5 24.7  36.5 14.8 10.0 17.9 18.5 22.0  21.6 33.3 15.3 44.0 31.5 44.2  65.6 23.1 11.3 39.0 15.9 12.6  38.9 16.4 11.5 21.3 21.0 24.2  Average  20.1  18.2  30.5  24.2  19.8  5 6 13 14 17 18  11.8 14.4 24.7 22.1 31.2 29.3  17.2 14.1 22.1 22.4 30.2 34.0  63.4 18.2 -8.3 23.4 23.7 21.3  35.9 31.1 37.7 18.0 29.2 36.4  14.0 14.1 25.2 21.8 29.3 28.6  Average  19.9  21.4  24.9  31.1  20.2  7 8 9 10 23 24  19.6 17.6 13.5 18.0 11.0 22.5  16.1 17.7 14.8 18.2 12.2 27.4  14.0 33.8 13.1 12.8 36.0 26.9  31.5 18.7 22.1 39.3 15.2 35.1  19.9 17.9 14.1 18.4 11.5 24.1  Average  16.5  17.2  22.9  26.1  17.1  3 4 15 16 19 20  10.8 11.5 10.5 9.8 20.0 16.7  8.2 9.8 7.4 6.2 16.7 12.8  9.3 6.0 -0.2 30.2 10.4 18.8  5.5 3.3 12.2 5.3 21.0 13.9  10.6 10.4 9.8 10.6 20.1 15.3  Average  12.7  10.1  11.7  9.2  12.3  D C  B D C  B D C  B D C  44  .  Total  4.2.4.7 Biomass In absolute terms, average yearly above-ground biomass increment, calculated from Tables 6 and 7, was 3.7 t/ha in the control plots, and 3.9 to 4.9 t/ha on spaced plots. In relative terms, growth was faster on spaced plots, regardless o f the biomass component (Table 13). O n spaced plots, branches and foliage grew at a higher rate than wood and bark. 4.3 T R E E - L E V E L R E S P O N S E  4.3.1 Overview A t the individual tree level, 90 A N O V A s were performed to analyze growth response among treatments (five diameter classes, three groups o f species, six variables o f interest). Variances among treatments were found significantly different i n 41 o f these. In addition, 144 A N A C O V A s were performed to analyze growth response among diameter classes (two sets o f diameter classes, three groups o f species, four treatments, six variables o f interest). Variances among diameter classes were found significantly different in 68 o f these. The effect o f the covariate was significant i n 96 A N A C O V A s and regression lines could be assumed parallel in 38 of the 96. Therefore, results are shown for the adjusted means in 38 analyses. In all other cases, unadjusted treatment means are shown. For each o f the 6 variables, small tree (<15 c m dbh) growth response was negatively related to Curtis' (1982) relative density. For some variables (diameter, basal area, volume), this relationship was very strong. Above a value o f about 10 for Curtis' (1982) relative density, small tree growth in diameter, basal area, volume, crown area seemed to be really restricted. A strong negative relationship existed between large tree (>15 c m dbh) growth response in diameter and basal area, and Curtis' (1982) relative density. A similar negative relationship  45  existed for growth in height and volume, but was less distinct. N o clear relationship was apparent for growth in crown area and live crown length. For each o f the variables, although the treatment means were not always significantly different, small and large Douglas-fir tree growth responses were consistently higher on spaced plots than on controls. In addition, except for height and crown length, growth was consistently higher on the 5 m clumped spacing than on the other spacing regimes. Despite similar trends, results were less distinct for non-Douglas-fir species. When comparing diameter and basal area growth response among small and large Douglas-fir trees, there was a significant block/treatment interaction i n the control plots. In this case, a simple interpretation o f differences among treatments was not possible. Graphic modeling o f treatment responses was used. Since this method showed a weak interaction, with similar growth trends among treatments, treatment means were tested. In all other cases, no block/treatment interaction was found.  4.3.2 Diameter 4.3.2.1 Small Trees (<15 c m dbh) For all species together, small tree average dbh growth was significantly lower on the controls than on the spaced plots (Table 14), with little variation from plot to plot on the controls (Figure 6), and showed a strong negative exponential relationship with Curtis' (1982) relative density (Figure 7). Although growth responses o f non-Douglas-fir species were not significantly different, they were consistently higher on the spaced plots. Diameter growth o f Douglas-fir, along with diameter growth o f all species, was the highest on treatment C 2 i n each o f the smallest dbh classes (0-5, 5-10, 10-15 cm). In parallel, diameter growth increased significantly with increasing size between 0 and 15 cm dbh, regardless o f the treatment (Table 15).  46  T a b l e 14. C o m p a r i s o n o f d i a m e t e r g r o w t h a m o n g treatments ( m e a n o f i n d i v i d u a l tree g r o w t h r e s p o n s e ) , b y d i a m e t e r class a n d b y s p e c i e s . 1  Diameter class  0-5 cm  Species  Treatment  Fd Non-Fd All species  5-10 cm  Fd Non-Fd All species  10-15 cm  Fd Non-Fd All species  0-15 cm  Fd Non-Fd A l l species  >15cm  Diameter growth (cm)  Fd Non-Fd A l l species  Var  Cl  C2  0.25 a 0.17 a 0.22 a  0.31 a 0.15 a 0.28 a  0.19 ab 0.22 a 0.20 ab  0.06 b 0.10 a 0.06 b  0.77 a 0.54 a 0.73 a  1.01 a 1.52 a 1.03 a  0.69 a 0.55 a 0.69 a  0.32 b 0.32 a 0.32 a  1.10 a 0.82 a 1.02 a  1.39 a 1.15 a 1.35 a  1.06 a 1.13 a 1.06 a  0.69 b 0.38 a 0.62 b  0.78 a 0.53 a 0.72 a  1.03 a 0.86 a 1.00 a  0.80 a 0.45 a 0.77 a  0.27 b 0.28 a 0.27 b  1.21 ab 0.90 a 1.11 a  1.60 a 1.53 b 1.49 a  1.21 ab 1.01 a 1.16 a  0.90 b 0.65 c 0.80 a  Control  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  47  Figure 6. M e a n o f individual-tree growth response by plot for trees less than 15 c m dbh, between 1993 and 1997. Vertical lines represent two standard errors o f the mean around the plot mean.  48  Diameter growth  1-6  Basal area growth  3.0  n  1.4  2.5  1.2 -  •  i  1.0  CD  E co  b  ^  S <?  A  0.8 0.6 \  A  0.4 0.2 -  •  0.5  ^  10  5  •  n ]  0.8 0.7  15  Relative density  Height growth  1.0  10  5  15  Relative density  0.9  •  0.0  4  0.0  1.5 1.0  Volume growth  •  1  2  0.60.5 -  •  0.4 -  •  0.3 -  E  •  0.2 -  •  0.1 -  •  0.0 10  5  15  5  Relative density  Live crown length growth  3.5  1.2  3.0  •  0.8 0.6  •  15  20  Crown area growth  1.4  1.0  10  Relative density  I  •  •  N  2.5 2.0  •  0.4  C D ~— C  •  1  o  0.2  1.5 1.0 0.5  0.0 0.0 5  10  15  5  Relative density  10  15  Relative density  gure 7. Mean of individual-tree growth response of trees less than 15 cm dbh in relation to Curtis' (1982) relative density, by plot and for six variables of interest between 1993 and 1997.  49  T a b l e 15. C o m p a r i s o n o f d i a m e t e r g r o w t h a m o n g d i a m e t e r classes ( m e a n o r adjusted m e a n o f i n d i v i d u a l tree g r o w t h response), b y treatment a n d b y s p e c i e s . 1  Treat.  Species  Diameter growth (cm) Diameter class  Fd Non-Fd All species Fd Non-Fd All species Fd  Mean  Non-Fd  Mean  All species  Mean  Fd  Mean  Non-Fd All species  0.25  0.77  1.10  a  b  c  0.70 a 0.73 b  0.69 a 1.02 c  Mean  0.31  1.01  1.39  Mean  a  b  c  Adj. Mean 0.09 a Mean 0.28 a Mean  Diameter class  I0-15cm  Adj. Mean 0.31 a Mean 0.22 a Mean  Var  0-5cm 5-1 Ocm  1.49 b 1.03 b  1.29 ab 1.35 c  0.19  0.69  1.06  a  b  c  0.22 a 0.20 a  0.55 a 0.69 b  1.13 a 1.06 c  0.06  0.32  0.69  a  b  c  0.42 a 0.32 b  0.11 a 0.62 c  Adj. Mean 0.30 a Mean 0.06 a  Mean  Mean  Mean Mean Mean Mean Mean Mean *  Mean *  Mean *  Var  0-15cm  >15cm  0.78 a 0.53 a 0.72 a  1.21 b 0.90 a 1.11 b  1.03 a 0.86 a 1.00 a  1.60 b 1.53 a 1.49 b  0.80 a 0.45 a 0.77 a  1.21 b 1.01 a 1.16 b  0.27 a 0.28 a 0.27 a  0.90 b 0.65 b 0.80 b  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  50  4.3.2.2 Large Trees (> 15 c m dbh) Large tree average dbh growth was consistently higher than that o f smaller trees, with better growth on spaced than on control plots, and better growth on C 2 than on the other spaced plots (Tables 14 and 15). D b h growth o f large trees also showed a strong negative exponential relationship with Curtis' (1982) relative density (Figure 8).  4.3.3 Basal Area 4.3.3.1 Small Trees (<15 cm dbh) In parallel to dbh growth, small tree average basal area growth for all species was significantly lower on the control than on the spaced plots (Table 16), and showed a strong negative exponential relationship with Curtis' (1982) relative density (Figure 7). Basal area growth was also consistently higher on treatment C 2 than on the other spaced plots, where treatments C l and S showed similar growth patterns. Basal area growth o f Douglas-fir increased significantly with increasing dbh between 0 and 15 c m dbh, regardless o f the treatment (Table 17). 4.3.3.2 Large Trees (> 15 c m dbh) Large tree average basal area growth was consistently higher than that o f smaller trees, with better growth on spaced plots than on the controls, and better growth on C 2 than on the other spaced plots (Tables 16 and 17). Basal area growth o f large trees also showed a strong negative exponential relationship with Curtis' (1982) relative density (Figure 8).  4.3.4 Height 4.3.4.1 Small Trees t<15 c m dbh) Average height growth o f small trees o f all species was less on controls than on spaced plots; for Douglas-fir, this difference was significant (Table 18). O n thinned plots, this species 51  Basal area growth  Diameter growth  2.0 1.81.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 n  p  V E CO  6 •  5  10  5  15  Relative density  15  Relative density  Height growth  1.6  10  •  Volume growth  1  1.4 i 1.2 -  I  It  1.0 -  S  E  0.8  S  0.6  £  -5 x  0.4 0.2 0.0  5  10  15  20  5  Relative density  Live crown length growth  !  ,5  £  2.0  S  E.  1.5  •  15  Crown area growth  •  1.0  I o  0.5 0.0  5  10  Relative density  10  •  •  5  •  •  10  ^5  15  Relative density  Relative density  Figure 8. M e a n o f individual-tree growth response of trees larger than 15 c m dbh i n relation to Curtis' (1982) relative density, by plot and for six variables of interest between 1993 and 1997.  52  Table 16. Comparison of basal area growth among treatments (mean of individual tree growth response), by diameter class and by species . 1  Diameter class  Species  J  Cl 0-5 cm  0.18 a  0.10 ab  0.04 b  Non-Fd  0.06 a 0.11 a  0.05 a 0.14 a  0.09 a 0.10 a  0.05 a 0.04 a  Fd  1.04 a  1.38 a  0.95 a  0.40 b  *  Non-Fd  0.76 a 0.99 a  2.02 a 1.41 a  0.64 a 0.95 a  0.41 a 0.40 b  *  2.22 a  2.89 a  2.19 a  1.36 b  *  1.66 a 2.07 a  2.53 a 2.83 b  2.63 a 2.20 ab  0.78 a 1.22 c  *  1.24 a 0.87 a 1.15 a  1.76 a 1.64 a 1.72 a  1.42 a 0.72 a 1.37 a  0.40 b 0.44 a 0.40 b  3.82 ab 2.94 a 3.51 a  5.11 a 4.71 b 4.71 a  3.83 ab 3.72 c 3.75 a  2.81 b 1.78 d 2.42 b  Fd  Non-Fd All species  0-15 cm  Fd Non-Fd All species  >15cm  Var Control  0.13 a  All species  10-15 cm  z  Fd  All species  5-10 cm  Basal area growth (x 10" m ) Treatment C2 S  Fd Non-Fd All species  * * *  * *  *  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks Cand D.  53  Table 17. Comparison o f basal area growth among diameter classes (mean or adjusted mean o f individual tree growth response), by treatment and by species . 1  Treat.  Species  Basal area growth (x 10" m ) Diameter class Var Diameter class J  z  0-5cm 5-10cm 10-15cm  Cl  C2  S  Fd Non-Fd  Adj. Mean  All species  Mean  Fd  Mean  Non-Fd  Mean  All species  Mean  Fd  Mean  Non-Fd  Cont.  Mean  Mean  All species  Mean  Fd  Mean  Non-Fd  Adj. Mean  All species  Mean  0.13 a  1.04 b  2.22  0.52 a 0.11 a  1.00 a 0.99 b  1.33 a 2.07 c  0.18 a  1.38 b  2.89  0.05 a 0.14 a  2.02 b 1.41 b  2.53 b 2.83 c  0.10 a  0.95 b  2.19  0.09 a 0.10 a  0.64 a 0.95 b  2.63 a 2.20 c  0.04 a  0.40 b  1.36  0.60 a 0.04 a  0.57 a 0.40 b  0.08 a 1.22 c  c  c  c  c  Adj. Mean *  Adj. Mean Adj. Mean *  *  Mean Mean Adj. Mean  *  *  Mean Mean  *  Adj. Mean *  *  Mean Mean  *  Mean *  Var  0-15cm  >15cm  1.59 a 1.39 a 1.54 a  1.72 a 1.43 a 1.71 a  1.76 a 1.64 a 2.29 a  5.11 b 4.71 b 3.13 a  *  1.42 a 0.72 a 1.97 a  3.83 b 3.72 a 2.02 a  *  0.40 a 0.44 a 0.40 a  2.81 b 1.78 a 2.42 b  *  *  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  54  Table 18. Comparison o f height growth among treatments (mean o f individual tree growth response), by diameter class and by species . 1  Diameter class  0-5 cm  Species  Fd Non-Fd All species  5-10 cm  Fd Non-Fd All species  10-15 cm  Fd Non-Fd All species  0-15 cm  Fd  Non-Fd All species >15cm  Fd Non-Fd All species  Height growth (m) Treatment  Var  Cl  C2  Control  0.31 a  0.17 a  0.14 a  0.15 a  0.13 a 0.25 a  0.14 a 0.17 a  0.08 a 0.12 a  0.04 a 0.15 a  0.74 a  0.68  0.65  ab  ab  0.59 b  0.47 a 0.70 a  0.96 a 0.69 ab  -1.15 a 0.63 ab  0.44 a 0.57 b  0.99 a  1.00  0.89  a  a  0.70 a  0.72 a 0.92 a  0.97 a 1.00 a  0.52 a 0.88 a  0.36 a 0.62 a  0.74 a 0.46 a 0.67 a  0.71 a 0.69 a 0.70 a  0.70 a 0.10 a 0.65 ab  0.42 b 0.30 a 0.41 b  1.00 a 0.79 a 0.93 a  0.91 a 1.07 a 0.92 a  0.81 a 0.30 a 0.72 a  0.75 a 0.67 a 0.69 a  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks CandD.  55  Table 19. Comparison o f height growth among diameter classes (mean or adjusted mean o f individual tree growth response), by treatment and by species . 1  Treat.  Species Diameter class 0-5cm  Fd Non-Fd All species Fd  0.62 a  0.73 a  0.80 a  Mean  0.13 a 0.60 a  0.47 a 0.70 a  0.72 a 0.68 a  0.17 a  0.68 b  1.00  0.14 a 0.17 a  0.96 a 0.69 b  0.14 a  0.65 b  0.08 a 0.39 a  -1.15 a 0.70 b  c 0.52 a 0.74 b  0.15 a  0.59 b  0.70 b  0.04 a 0.15 a  0.44 a 0.57 b  0.36 a 0.62 b  Adj. Mean  Mean Mean  All species  Mean  Fd  Mean  Non-Fd  Mean  All species  Adj. Mean  Fd  Mean  Non-Fd  Mean Mean  Diameter class  5-10cm 10-15cm  Adj. Mean  Non-Fd  All species  Height growth (m) Var  c 0.97 a 1.00 c  Mean Mean *  Mean *  Mean Mean Mean *  Mean  0.89  Mean Mean Mean *  Mean *  Mean *  Var  0-15cm  >15cm  0.74 a 0.46 a 0.67 a  1.00 a 0.79 a 0.93 a  0.71 a 0.69 a 0.70 a  0.91 a 1.07 a 0.92 a  0.70 a 0.10 a 0.65 a  0.81 a 0.30 a 0.72 a  0.42 a 0.30 a 0.41 a  0.75 a 0.67 a 0.69 a  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D. 1  56  showed very little variation from treatment to treatment, except for the smallest dbh class (0-5 cm). Nevertheless, height increment tended to be slightly lower on treatment S than on the other types o f spacing, and this trend was the same for large trees. Average height growth o f small trees showed a distinct negative relationship with Curtis' (1982) relative density (Figure 7). 4.3.4.2 Large Trees (>15 c m dbh) Although the differences were not significant, height growth o f large trees was consistently higher than that o f smaller trees, regardless o f treatment and species (Table 19). In addition, height growth o f large trees (Douglas-fir and all species together) was consistently higher on spaced plots than on controls, with the highest growth on C l (Table 18). Unlike small trees, the height growth pattern o f large trees was not clearly related to Curtis' (1982) relative density (Figure 8).  4.3.5 Volume 4.3.5.1 Small Trees (<15 cm dbh) In parallel to basal area growth, small tree average volume growth for all species was significantly lower on the control than on the spaced plots (Table 20), and showed a strong negative exponential relationship with Curtis' (1982) relative density (Figure 7). A l s o , volume growth was slightly higher on treatment C 2 than on the other spaced plots. Except when outliers disrupted volume growth analysis, volume growth increased with increasing dbh between 0 and 15 c m dbh, regardless o f the treatment (Table 21). 4.3.5.2 Large Trees (> 15 c m dbh) Volume growth o f large trees was consistently higher on spaced plots than on controls, and despite little variation between spacing regimes, a slightly better growth was observed on  57  T a b l e 2 0 . C o m p a r i s o n o f v o l u m e g r o w t h a m o n g treatments ( m e a n o f i n d i v i d u a l tree g r o w t h r e s p o n s e ) , b y d i a m e t e r class a n d b y s p e c i e s . 1  Diameter class  Species  Volume growth (x 10" m ) J  Treatment Cl  0-5 cm  Fd Non-Fd All species  5-10 cm  Fd Non-Fd All species  10-15 cm  Fd Non-Fd All species  0-15 cm  Fd Non-Fd All species  >15cm  J  Fd Non-Fd A l l species  Var  C2  S  Control  0.41 a 0.25 a  0.42 a 0.15 a  0.27 a 0.29 a  0.19 a 0.22 a  0.35 a  0.34 a  0.28 a  0.20 a  5.17 a 4.89 a 5.12 a  5.88 a 10.01 a 6.05 a  5.00 a -3.41 a 4.90 a  2.69 b 2.83 a 2.71 b  14.32 a 13.59 a 14.12 a  16.14 a 20.99 a 16.91 a  14.31 a 22.26 a 14.61 a  9.00 b 6.87 a 8.44 b  7.10 a 6.65 ab 6.69 a  8.97 a 12.64 a 9.25 a  8.73 a 5.17 ab 8.46 a  2.60 b 3.51 b 2.70 b  32.05 a 32.27 a 32.13 a  33.73 a 44.38 b 32.28 a  29.75 a 34.92 a 30.27 a  22.71 a 17.80 c 20.11 b  *  *  *  *  *  *  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks CandD.  58  Table 21. Comparison o f volume growth among diameter classes (mean or adjusted mean o f individual tree growth response), by treatment and by species . 1  Treat.  Species  Volume growth (x 10" m ) Diameter class Var Diameter class J  0-5cm  Fd Non-Fd  Mean Adj. Mean  All species  Mean  Fd  Mean  Non-Fd  Mean  All species  Mean  Fd  Mean  Non-Fd  Mean  All species  Mean  Fd  Mean  Non-Fd All species  Adj. Mean Mean  J  5-10cm 10-15 cm  0.41 a 5.18 a 0.35 a  5.17 b 6.25 a 5.12 b  14.32 c 9.90 a 14.12 c  0.42 a 0.15 a 0.34 a  5.88 b 10.01 b 6.05 b  16.14 c 20.99 b 16.91 c  0.27 a 0.29 a 0.28 a  5.00 b -3.41 a 4.90 b  14.31 c 22.26 a 14.61 c  0.19 a 6.08 a 0.20 a  2.69 a 4.78 a 2.71 b  9.00 b -3.50 a 8.44 c  0-15cm  *  Mean Mean  *  Mean *  *  Adj. Mean Mean Mean  *  *  Mean Adj. Mean  *  Adj. Mean *  Mean * *  Mean Mean  *  Var >15cm  7.10 a 6.65 a 6.69 a  32.05 b 32.27 a 32.13 b  11.65 a 12.64 a 9.25 a  22.08 b 44.38 b 32.28 b  8.73 a 20.23 a 13.37 a  29.75 b 13.21 a 15.72 a  2.60 a 3.51 a 2.70 a  22.71 b 17.80 b 20.11 b  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  59  treatment C 2 (Table 20). Although some dispersion could be seen, the volume growth pattern o f large trees was negatively related to Curtis' (1982) relative density (Figure 8). Growth efficiency, expressed as the average ratio o f individual tree stem volume growth to crown projection area (O'Hara 1988), is shown in Figure 9. U p to 25 c m dbh on the spaced plots and up to 20 cm dbh on the control plots, volume growth efficiency regularly increased with increasing dbh, with the highest efficiency on treatment C 2 and the lowest on control. Beyond 20-25 c m dbh, results were less reliable because o f a limited number o f trees i n the larger dbh classes, and the huge effect o f die back or broken tops on averages (for instance, dbh class 7 i n treatment C2). However, results suggested that large trees did not grow more efficiently than trees less than 20-25 c m dbh.  4.3.6 Crown Area 4.3.6.1 Small Trees (<15 c m dbh) B e l o w a value o f about 10 for Curtis' index, no distinct relationship could be seen between crown area growth and relative density; above this value, growth was clearly restricted (Figure 7). Douglas-fir crown area growth was much higher on spaced plots than on controls, with again an advantage for treatment C 2 . O n average, crown growth increased with increasing dbh (Table 22). 4.3.6.2 Large Trees (>15 c m dbh) Although the relationship was not really distinct, average crown area growth o f large trees seemed to be negatively related to Curtis' (1982) relative density (Figure 8). Growth was less on controls than on spaced plots (Table 22), with the highest growth average on treatment C 2 . Not surprisingly, growth averages were better for big trees than for small trees (Table 23), with the largest difference on the control plots.  60  Dbh class  Figure 9. Growth efficiency (stem volume growth / initial crown projection area) per dbh class between 1993 and 1997. D b h classes are 5 c m wide (dbh class 1 = 0-5 cm, dbh class 11 = 50-55 cm).  61  T a b l e 2 2 . C o m p a r i s o n o f c r o w n area g r o w t h a m o n g treatments ( m e a n o f i n d i v i d u a l tree g r o w t h r e s p o n s e ) , b y d i a m e t e r class a n d b y s p e c i e s . 1  Diameter class  0-5 cm  Species  Fd  All species  Fd Non-Fd All species  10-15 cm  Fd Non-Fd All species  0-15 cm  Fd Non-Fd A l l species  >15cm  z  Treatment  Non-Fd  5-10 cm  Crown area growth (m )  Fd Non-Fd A l l species  Var  Cl  C2  S  Control  0.35 ab 0.08 a 0.26 a  0.57 a -0.02 a 0.42 a  0.30 ab 0.05 a 0.22 a  0.07 b -0.09 a 0.06 a  1.42 a 0.63 a 1.28 a  1.52 a 2.08 a 1.54 a  0.87 ab 0.88 a 0.87 a  0.27 b 0.05 a 0.25 b  2.20 a 0.70 a 1.79 a  2.07 a 0.48 a 1.82 a  1.69 a 0.59 a 1.65 a  0.67 b 0.67 a 0.67 b  1.46 a 0.49 a 1.22  1.56 a 0.52 a 1.40  1.18 ab 0.22 a 1.10  0.25 b 0.24 a 0.25  a  a  a  a  1.86 a 2.26 a 2.00  3.43 a 1.97 a 2.90  2.11 a 1.42 a 1.87  1.05 a 1.09 a 1.04  a  a  a  a  * *  *  *  * * *  * * *  * * *  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks Cand D.  62  Table 23. Comparison o f crown area growth among diameter classes (mean or adjusted mean o f individual tree growth response), by treatment and by species . 1  Treat.  Species  Crown area growth (m ) Diameter class O-Scm  Cl  Fd Non-Fd  C2  S  Cont.  Mean Mean  All species  Adj. Mean  Fd  Adj. Mean  Non-Fd  Mean  All species  Adj. Mean  Fd  Adj. Mean  Non-Fd  Mean  All species  Adj. Mean  Fd  Mean  Non-Fd  Adj. Mean  All species  Adj. Mean  Var  Diameter class  5-10cm 10-15cm  0.35 a 0.08 a 0.87 a  1.42 a 0.63 a 1.25 a  2.20 a 0.70 a 1.26 a  2.29 a -0.02 a 1.82 a  1.75 a 2.08 a 1.70 a  0.88 a 0.48 a 0.87 a  1.27 a 0.05 a 1.13 a  1.04 a 0.88 a 0.98 a  1.27 a 0.59 a 1.23 a  0.07 a 0.08 a 0.18 a  0.27 a 0.18 a 0.23 a  0.67 a 1.34 a 0.46 a  Mean *  Mean *  Mean *  Mean *  Mean Adj. Mean *  Mean *  Mean Mean *  *  Mean Mean  *  Mean *  0-15cm  >15cm  1.46 a 0.49 a 1.22 a  1.86 a 2.26 a 2.00 a  1.56 a 0.52 a 1.55 a  3.43 b 1.97 a 1.92 a  1.18 a 0.22 a 1.10 a  2.11 b 1.42 a 1.87 a  0.25 a 0.24 a 0.25 a  1.05 b 1.09 a 1.04 a  Var  *  *  *  *  *  *  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  63  4.3.7 Live Crown Length 4.3.7.1 Small Trees f<15 c m dbh) L i v e crown length growth o f small trees was negatively related to Curtis' (1982) relative density (Figure 7). Douglas-fir live crown length growth was higher on spaced plots than on controls, with an advantage for treatment S. O n average, live crown length growth increased with increasing dbh (Table 24). 4.3.7.2 Large Trees (>15 cm dbh) N o clear relationship existed between live crown length growth o f large trees and Curtis' (1982) relative density (Figure 8). In parallel, growth was similar among the control and spaced plots (Table 24).  4.4 R E S U L T S S U M M A R Y F O R G R O W T H R E S P O N S E A t the stand level, variables o f interest were growth and growth rate o f quadratic mean diameter, basal area per hectare, Lorey's height, volume per hectare and accumulated crown area per hectare. There were no significant differences among treatment means o f growth response in quadratic mean diameter, basal area, volume, and accumulated crown area. However, for these four variables, treatment means were consistently lower in the control plots than i n the spaced plots. In the spaced plots, there was no clear difference among spacing regimes i n an absolute sense, although treatment C 2 showed consistently higher growth rates. Due to a block-treatment interaction, Lorey's height growth was not tested. For the 4-year period between 1993 and 1997, stand diameter growth, calculated as the periodic diameter growth o f trees surviving at the end o f the growth period, was 0.37 c m on the control, and ranged from 0.80 to 1.08 cm on the spaced plots. Net basal area growth averaged  64  Table 24. Comparison of life crown length growth among treatments (mean of individual tree growth response), by diameter class and by species . 1  Diameter class  Species Cl  0-5 cm  0.46 a  0.32 a  0.39 a  0.38 a  Non-Fd  0.48 a 0.47 a  0.25 a 0.30 a  0.21 a 0.33 a  0.21 a 0.36 a  Fd  0.83 a  0.91 a  0.93 a  0.74 a  Non-Fd  0.67 a 0.80 a  1.08 a 0.91 a  -1.30 a 0.90 a  1.04 a 0.77 a  Fd  1.07 a  1.17 a  1.33 a  0.85 a  Non-Fd  1.02 a 1.06 a  1.57 a 1.23 a  0.98 a 1.31 a  0.82 a 0.85 a  0.83 ab 0.74 a 0.81 a  0.90 ab 1.07 a 0.91 a  1.05 a 0.28 a 0.99 a  0.60 b 0.72 a 0.62 a  1.45 a 0.67 a 1.18 a  1.36 a 1.67 a 1.32 a  1.21 a 1.19 a 1.20 a  1.38 a 1.37 a 1.31 a  All species  10-15 cm  All species  0-15 cm  Fd Non-Fd All species  >15cm  Var Control  Fd  All species  5-10 cm  Live crown length growth (m) Treatment C2 S  Fd Non-Fd All species  *  *  *  * * *  One star (*) indicates that the variances (var) among treatments are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks Cand D. 1  65  Table 25. Comparison o f live crown length growth among diameter classes (mean or adjusted mean o f individual tree growth response), by treatment and by species . 1  Treat.  Species  Live crown length growth (m) Var Diameter class  Var  Diameter class 0-5cm  Cl  C2  S  Fd Non-Fd  Mean  All species  Mean  Fd  Adj. Mean  Non-Fd  Mean  All species  Adj. Mean  Fd  Adj. Mean  Non-Fd  Cont.  Adj. Mean  Mean  All species  Adj. Mean  Fd  Adj. Mean  Non-Fd  Adj. Mean  All species  Adj. Mean  5-10cm 10-15cm  Adj. Mean  0.51 a  0.86 a  0.91 a  0.48 a 0.47 a  0.67 a 0.80 a  1.02 a 1.06 a  0.86 a  0.99 a  0.93 a  0.25 a 0.70 a  1.08 a 0.98 a  1.57 a 0.98 a  *  0.35 a  0.95 a  1.38 a  *  0.21 a 0.33 a  -1.30 a 0.93 a  0.98 a 1.37 a  0.63 a  0.79 a  0.57 a  0.76 a 0.64 a  1.26 a 0.81 a  0.76 a 0.58 a  Mean *  Adj. Mean  * Mean *  Mean Mean Adj. Mean Mean  *  Mean  * Adj. Mean * * *  Mean Mean  0-15cm  >15cm  0.98 a 0.74 a 0.96 a  1.02 a 0.67 a 0.62 a  0.90 a 1.07 a 0.91 a  1.36 a 1.67 b 1.32 b  1.24 a 0.28 a 0.99 a  0.69 b 1.19 b 1.20 a  0.70 a 0.72 a 0.62 a  0.77 a 1.37 a 1.31 b  *  *  *  One star (*) indicates that the variances (var) among diameter classes are significantly different at a 0.05 probability level. Identical letters under the means indicate no statistical difference at the 0.05 probability level. Growth of species other than Douglas-fir (non-Fd) were tested only on blocks C and D.  66  9  9  2.34 m /ha on the control, and ranged from 3.23 to 3.52 m /ha on the spaced plots, while 9  9  mortality represented 0.83 m /ha on the control, and 0.01 to 0.22 m /ha on the spaced plots. Net periodic annual volume increment averaged 5.2 to 6.6 m per ha on the spaced plots and 4.8 m per ha on the controls. A t the individual tree level, variables o f interest were growth i n diameter, basal area, total height, total stem volume, crown area, and live crown length. Growth was analyzed by species or group o f species (Douglas-fir, other species, all species combined) and by diameter class (0-5 cm; 5-10 cm; 10-15 c m and 0-15 cm; >15 cm). For each o f the 6 variables, small tree (<15 c m dbh) growth response was negatively related to Curtis' (1982) relative density. For some variables (diameter, basal area, volume), this relationship was very strong. A strong negative relationship existed between large tree (>15 c m dbh) growth response in diameter and basal area, and Curtis' (1982) relative density. A similar negative relationship existed for growth in height and volume, but was less distinct. N o clear relationship was apparent for growth in crown area and live crown length. For each o f the variables, although the treatment means were not always significantly different, small and large Douglas-fir tree growth responses were consistently higher on spaced plots than on controls. In addition, except for height and crown length, growth was consistently higher on the 5 m clumped spacing than on the other spacing regimes. Despite similar trends, results were less distinct for non-Douglas-flr species. U p to 20-25 cm, volume growth efficiency regularly increased with increasing dbh, on both control and spaced plots. Trees seemed to be the most efficient on treatment C 2 and the least efficient on control. 67  5. D I S C U S S I O N  5.1 S T A T I S T I C A L A N A L Y S I S  5.1.1 Analysis of Variance In analysis o f variance, valid application o f tests o f significance requires that the errors from the linear additive model be independently and normally distributed, with a common variance (Snedecor and Cochran 1980, Chapter 15; Hicks 1993, Chapter 4; Steel et al. 1997, Chapter 7). These assumptions imply that the observations on the dependent variable also are independently and normally distributed, with a common variance. Independence Randomization is used to make the observations independent o f each other, and to ensure an unbiased estimate o f treatment means. A s previously explained (Section 3.2, page 11), treatments were randomly assigned to the experimental units, but plots were not randomly located within experimental units. They were installed i n dense patches most likely to show the largest response to thinning, and large enough to accommodate the plots. Theoretically, such patches should have been mapped, as well as every possible plot location within each o f these patches, given the size, shape and orientation o f the plots. Afterwards, two o f these possible locations should have been randomly picked within each o f the experimental units. This randomization procedure was considered too time-consuming and costly, and was not done. Therefore, the results o f the sampling may be biased. Although the size o f the bias is not known, it can reasonably be assumed to be small, and thus conclusions drawn from the results can be 68  generalized to stands o f similar species and structural composition i n similar environmental conditions. Normality The assumption that experimental errors are normally distributed with a mean equal to zero ensures that the ratio o f the treatment mean square error to the error mean square follows an F distribution under the assumption o f no differences between treatment means. Normality is also required for the construction o f valid confidence intervals (Steel et al. 1997, p. 174, Rawlings et al. 1998, Chapter 10). Although exploratory tests showed that non-normality o f response variables was not rare, no transformation o f the scale o f measurement was carried out to try to reach approximate normality. Departures from normality were assumed to be moderate. Providing this assumption is valid, then departures from normality do not have serious consequences for the analysis (Bartlett 1947, in Rawlings et al. 1998, p. 409). Besides, as Rawlings et al. (1998) pointed out, transformations to stabilize variance (see next section) have the effect o f also improving normality. Homogeneity o f Variance Despite using power transformation whenever required, homogeneity o f variance was reached for just over half o f all A N O V A s and A N A C O V A s (125 out o f 234). In most cases, heterogeneity o f error was presumably due to the fact that growth response had a larger mean with a wider spread, and showed more variability on treated units than on the control, especially for the smallest dbh classes (<15 cm). Control plots also represented larger sample units, i n terms o f number o f trees, increasing the heterogeneity o f variance between spacing treatments and the unspaced controls. Homogeneity o f variance o f some variables was difficult to reach. Variance o f crown area growth was especially heterogeneous, presumably as a result o f the lack  69  of precision i n the measurement o f crown projection area. Essentially, heterogeneous variance results i n a loss o f precision i n the estimates (Snedecor and Cochran 1980, Chapter 15, Steel et al. 1997, p. 175, Rawlings et al. 1998, Chapter 10).  5.1.2 Analysis of Covariance A t the individual-tree level, covariance analysis was carried out because variation i n growth response was suspected to be partly attributable to dbh. Treatment means o f individualtree growth response were adjusted so as to remove the differences i n initial diameter, when the covariate was significant and the regression lines parallel. In most cases, the covariate was significant, but the assumption o f parallelism among regression lines was not met. Consequently, treatment means were not adjusted. When means were adjusted, the general trend o f the growth response was similar to those without adjustment; that is, an increasing growth response with increasing diameter. However, this trend was more erratic, especially on the control. The assumptions necessary for analysis o f variance are also required for analysis o f covariance. In addition, the regression o f the response variable on the independent variable must be linear, and the independent variable must be independent o f treatments (Hicks 1993, Chapter 16, Steel et al. 1997, Chapter 17). The latter assumption was not met, because tree dbh had been influenced by the treatment. Consequently, adjustment removed part o f the treatment effect (Steel et al. 1997, p. 431). In the long term, when the diameter distribution w i l l reflect more accurately the age distribution, this type o f analysis w i l l likely be better using age as covariate, since age is less dependent on treatment. Analysis o f variance and covariance were carried out even though the basic assumptions (independence, normality and common variance) required for the valid application o f tests o f significance were not strictly met. The above discussion assesses the importance o f departure from these assumptions, and its impact on the analyses. The results o f the study should be treated 70  cautiously, as suggested by Steel et al. (1998, p. 174),".. .in any case, most [biological] data do not exactly fulfill the requirements o f the mathematical model, and procedures for testing hypotheses should be considered approximate rather than exact".  5.2 A G E S T R U C T U R E Although age averages were not weighted by the number o f stems found i n each dbh class, it seemed clear that most small Douglas-fir trees (< 15 c m dbh) were established prior to the 1961 cut, and did not originate from a regeneration flush following logging, as previously suggested (Marshall 1996). In 1961, the total age o f these trees (36-59 years) was somewhat different from Clark's (1952) observations, o f 30 years as average age o f advance regeneration after logging. However, the results confirmed that advance regeneration o f Douglas-fir can respond well to release, even after decades o f suppression. The greater age average on the driest block (B) may be due to the fact that the lower canopy trees tend to live longer on droughty sites, compared to moister sites, because more light reaches the lower parts o f the stand through the thinner upper canopy (Oliver and Larson 1997, p. 62). The age distribution observed i n the stands strongly contrasted to the hypothetical reverse-J distribution. M e a n age o f small (<15 cm) and medium-sized (15-30 cm) Douglas-fir trees showed little difference, and were similar to those provided by G l e w and Cinar (1966), who found 80 years as average age o f the pole layer (3-11 inches) i n the Cariboo. Conversion to a reverse-J distribution would require decades o f constant commitment to uneven-aged management, along with an absence o f major damage to the stands.  71  5.3 G R O W T H  5.3.1 Growth Factors A s would be expected, results demonstrated that both stand and individual tree growth were stimulated by spacing. One can reasonably hypothesize that the positive effect o f thinning on growth is the compound result o f three different phenomena: 1. A greater availability o f site resources: •  Spacing increased light for tree crowns, contributing to the increase o f  both leaf area and the efficiency o f leaf area (Daniel et al. 1979, Chapter 5, Waring and Schlesinger 1985, Chapter 3, K i m m i n s 1997, p. 42). •  Spacing increased soil water availability for residual trees as a result o f  two combined effects. First, both stand transpiring surface area and live root density were reduced, thereby limiting water uptake (Kimmins 1997, Chapter 11, Dormer and Running 1986, Waring and Schlesinger 1985, Chapter 4, Kramer and K o z l o w s k i 1979, Chapter 13). This effect lasted for an undetermined period o f time and might have been over at the time o f the 1997 remeasurement, because the remaining trees could have extended their roots and crowns into spaces formerly occupied by the cut trees, and hence increased their water consumption. Second, canopy interception was reduced, allowing for a greater amount o f rainfall to reach the ground (Waring and Schlesinger 1985, Chapter 4, K i m m i n s 1997, Section C ) . In her study near Kamloops, K o r o l (1985) found that interception loss for a 100% canopy coverage was about 37% o f the total rainfall. She also found that canopy coverages o f between 25% and 65% resulted in the highest values o f the soil water reserve at the beginning o f the growing season, by increasing the duration o f the snow cover. I f similar results were demonstrated in this study, they might be attributed to 72  deeper snow accumulation i n thinned stands compared to unthinned controls, as a result o f wind eddies favoring the deposition o f snow i n small openings (Kimmins 1997, Chapter 11, Helvey 1975). They might also be attributed to greater evaporation losses i n the control plots as a result o f a greater snow interception by the compact cover o f vegetation (Korol 1985). Furthermore, spacing could have prolonged the time o f diameter growth later into the summer because water availability would not become a limiting factor as quickly as i n the unspaced stands (Zahner and Whitmore 1960, in Smith et al. 1997). •  B y providing better conditions for microbial activity, spacing can  accelerate the mineralization o f the organic matter present at the time o f the cut, and add readily decomposable material to the forest floor (green foliage, fine twigs and fine roots) (Kimmins 1997). Reduction in root density may also lead to a greater nutrient availability for residual trees. 2. Better environmental conditions Since more radiation reached the ground i n the thinned stands, soil temperature increased, facilitating water absorption, and thereby nutrient uptake (Lopushinsky 1990). Higher soil temperature is also known to favour microbial activity and to influence rates o f decomposition (Kimmins 1997, Chapter 8, Waring and Schlesinger 1985, Chapter 8). In addition, it is also possible that increased soil temperature led to longer growing seasons (Aussenac and Granier 1988). 3. Changes in the allocation o f photosynthates The energy obtained through photosynthesis is first used for respiration o f the living cells. Renewal o f fine roots and leaves is the second priority. Height growth and lateral branch growth, as well as root extension, come next. The remaining energy is used for diameter growth  (Oliver and Larson 1997, p. 75). O n droughty sites, a considerable proportion o f the photosynthates may be allocated below ground to develop and maintain fine roots, so that less is available for stem growth (Keyes and Grier 1981). This study did not examine the growth o f the entire tree, and thus above ground differences between spaced and unspaced plots might not reflect differences i n total biomass production. 5.3.2 Stand Growth A t the stand level, there were no significant differences among treatments for any o f the variables o f interest. Since spacing occurred only a few years ago, this was likely to happen. The high variability i n density and species that exists among plots could also explain the absence o f significant results, and could continue to mask differences i n subsequent analyses (Marshall 1996). Accumulated crown area: In the spaced plots, growth o f accumulated crown area considerably exceeded that o f the control plots (Table 5, page 28). Favorable changes i n the growth factors, previously described, can to a large extent explain such results. However, variation i n growth was considerable, especially for treatment C 1 . One reason for this variation might be the imprecision i n the measurement o f crown projection. Large crowns on taller trees, which comprised most o f the spaced plots, were more difficult to measure accurately than small crowns on short trees, which comprised most o f the control plots. In addition, the errors i n diameter measurement o f a large crown, compared to a small crown, led to a much greater error for the crown projection area. Quadratic mean diameter: A s would be expected, Q M D growth was higher on thinned plots (Table 12, page 40). Part o f this result was due to the slower growth on the many more small trees present i n the control plots (Marshall and Bugnot 1997), as well as to the arithmetic increase resulting from the death o f trees smaller than average (Omule 1988). Ingrowth also 74  influenced changes i n Q M D . Overall, growth clearly increased with decreasing relative density. The highest increments were obtained on treatment C 2 , but the results do not suggest growth could reach amounts as high as those reported by G l e w and Cinar (1966) i n their study o f Douglas-fir stand growth after seed tree marking in the Cariboo (one inch o f radial growth per decade). Basal area: Mortality strongly affected net basal area growth on the controls (Table 10, page 34). Mortality was much less on the spaced plots and thus, differences i n gross basal area growth between control and spaced plots were reduced. Marshall and Wang (1996) reported an average net basal area growth o f 3.4 m /ha per decade i n unregulated stands at Knife Creek, 2  while Day (1997) found 4.2 m /ha per decade. These results were markedly different from the results obtained by extrapolation on the control plots (5.8 m /ha), possibly because o f differences 2  in basal area and stand structure. Lorey's height: Lorey's height was chosen to estimate stand mean height because it seemed logical to weight individual tree height proportionally to basal area. In doing so, mean height growth i n the control plots was less affected by stunted height growth o f a great number of suppressed trees. O n the other hand, Lorey's height was considerably affected by broken tops and die back o f big trees, as well as by the species mixture. Consequently, this variable failed to bring to light growth differences revealed by arithmetic averages at the individual tree level. Volume: Stand volume growth is a function o f basal area growth, form change, and height growth. Volume growth differences between spaced and control plots were similar to basal area growth differences. However, the effect o f greater height growth o f both small and large Douglas-fir trees on spaced plots should be taken into account. Form changes could also have had an influence because thinning may produce more tapered tree trunks, leading to an overestimation o f volume production i n spaced plots (Hagburg 1942, in Spurr 1952, p. 294). 75  Although standing volume was less i n thinned plots, net volume growth per ha increased more rapidly than in unthinned stands (Table 12, page 40). This indicates that proper thinning does not necessarily lower net increment volume. Since old dominant trees maintained i n spaced plots likely had, and continue to have, a negative influence on the growth o f their subordinates, as well as a limited growth efficiency compared to younger trees (O'Hara 1988), volume production observed i n the spaced plots could have been even better. O n treatment C 2 , thinning below the level o f "full site occupancy" (Long 1985) may explain a lower gross and net volume production compared to treatments C l and S, as well as a lower gross production than the control plots (Oliver and Larson 1997, Chapter 15). Overall on thinned plots, crown closure had not yet been achieved, and enhanced growth could be expected until this condition is reached. O n control plots, net volume growth w i l l likely continue to be greatly reduced by the volume o f mortality. W i t h a net periodic annual increment o f 4.8 to 6.6 m /ha, the results were consistent with 3  yield capabilities o f 0.7 to 7 m /ha reported by Arno (1990), but somewhat different from observations o f others. A t Knife Creek, Marshall and Wang (1996) remeasured six permanent sample plots and reported volume growth ranging from 3.3 to 4.4 m /ha/year. Bonnor (1990) 3  calculated annual growth rates ranging from 1.9-5.3 m /ha in the IDFdk. In contrast, K o r o l 3  (1985) calculated growth rates as high as 9-11 m /ha/year at stand density o f 900-1200 trees/ha 3  and standing volumes o f 225 to 300 m /ha in the Kamloops Forest Region. 3  It should be noted that the plots i n this study were subjectively located i n well stocked areas, and had been diameter-limit logged in the past. The observed growth rates may not be indicative o f those that may occur i n unlogged stands i n this subzone, or i n managed stands with a balanced structure. Furthermore, i n the latter stands, Farron (1980) demonstrated that periodic  76  annual increment is a proper measure o f production only i f it is remeasured over a fairly long period o f time, so as to take into account variations i n climatic conditions. Biomass: W i t h an average o f 3.7 t/ha/year, total above-ground biomass growth on the control plots was within the range o f 2.5 to 4.4 t/ha/year provided by Marshall and Wang (1996) for various unregulated stands at Knife Creek. O n spaced plots, growth was consistently higher, ranging from 3.9 to 4.9 t/ha/year. However, limited growth o f total above-ground biomass on the control plots, compared to spaced plots, could be offset by higher below-ground biomass increment. In a study o f coastal Douglas-fir growing on poor (dry, l o w fertility) and good (moist, fertile) sites, Keyes and Grier (1981) demonstrated that on poor sites, the allocation o f biomass to the stem was reduced while root biomass was increased because a larger root system was required to satisfy the moisture and nutrient demands o f the above-ground components. W i t h an average o f 14.0 t/ha, foliage biomass on the control plots was above the range o f 11.0 to 13.4 t/ha calculated by Marshall and Wang (1996). This was also above the level o f 1213 t/ha provided by Turner and L o n g (1975) for fully closed coastal Douglas-fir stands on poor sites, and close to the upper range o f foliage biomass reported by Parde (1980) for evergreen coniferous forests. However, variation in growth rate was large (Table 13, page 44) and could be an indication o f the closeness to maximum leaf area. Control plots 3, 4 and 16 had the highest levels o f foliage biomass and the lowest growth rates, which is in agreement with Waring and Schlesinger (1985, Chapter 3) who reported decreasing stand growth efficiency with increasing leaf area i n Douglas-fir stands.  5.3.3 Individual Tree Growth Trees were initially sorted into two broad groups o f small (<15 cm) and large (>15 cm) stems. These categories were convenient for the analysis, but have little biological justification,  77  even though significant growth differences were detected for some variables. The subdivision o f the first group in three smaller subsets (0-5 cm; 5-10 cm; 10-15 cm) provided more accurate information. D b h can be readily and precisely measured, and dbh classes avoid the subjectivity o f crown classification. With increasing size differentiation, this approach might be very useful in subsequent analyses. Diameter: Except for the oldest trees, observations o f radial growth on increment cores showed a marked redistribution o f growth immediately after the 1961 cut. The 1990 spacing had less conspicuous effects. However, the results demonstrated that diameter growth was consistently higher on spaced plots, compared to control plots, regardless o f dbh (Table 14, page 47). Even the smallest trees (0-5 c m dbh) responded markedly to thinning. O n average, Douglas-fir trees i n this dbh class grew three to five times more on the spaced plots than their counterpart on the control plots. Since spacing was essentially a heavy l o w thinning, during which almost all trees removed were less than 15 c m dbh, large trees (>15 cm) were expected to be little affected by the cut, because their light environment was not dramatically altered. Large trees comprised old and relatively young trees, i n dominant and codominant positions. Each o f these categories had possibly experienced different growth patterns, but their overall response was consistently higher on the spaced plots, with a significant difference between control (0.9 cm) and treatment C2 (1.6 cm). Although changes i n the light environment o f the upper canopy might have been underestimated, especially changes for codominant trees, these results may be indicative o f the impact o f below-ground competition for limiting water resources (Marshall and Bugnot 1997). Diameter growth increased with increasing dbh. Since most trees greater than 15 c m dbh were less than 30 cm dbh (Appendix I), this growth pattern has to be confirmed for  78  larger trees. D a y (1997) observed than dbh growth o f Douglas-fir at Knife Creek slows down between 50 and 60 c m dbh. Diameter growth is closely related to crown size because, on the one hand, growth is accompanied by an increasing consumption o f photosynthates, necessary for the respiration o f increasing living tissues, and on the other hand, diameter growth is given the last priority i n the allocation o f photosynthates (Assmann 1970, Oliver and Larson 1997, Chapter 3). In parallel, crown development in multi-aged stands largely depends on vertical structure and spacing. In the selection forest o f Couvet i n the Swiss Jura, where Biolley first applied the "methode du controle" around 1890, Farron (1980) compiled 10 complete inventories for the 75-year period between 1893 and 1968, and demonstrated that diameter increment o f Norway spruce {Picea abies (L.) Karst.) and Silver fir {Abies alba A . ) , regularly increased over a wide range o f dbh between 17.5 and 82.5 cm. Schiitz (1989) demonstrated the same growth pattern for several selection forests i n different ecosystems throughout Switzerland, and attributed it to both improving crown position and increasing crown size with increasing age/diameter. In contrast, Becker (1995) reported relatively constant diameter growth across size classes i n managed mixed uneven-aged stands in Western Montana. Lundqvist (1993) found similar results i n selection stands o f Norway spruce i n Sweden. This is surprising because cohort age and crown position should logically result i n differences i n diameter growth among size classes. Height: Height growth o f small trees was significantly affected by spacing (Table 18, page 55). Trees grew faster on spaced plots, and there was a strong relationship with relative density (Figure 7, page 49). Improved light conditions in the lower canopy can to a large extent explain such differences, but enhanced water availability along with changes in photosynthate allocation should be considered as well (Oliver and Larson 1997, Chapter 3; K i m m i n s 1997, Chapter 11). 79  Although differences between control and spaced plots were not significant, height growth o f large trees also seemed to be positively affected by spacing. L i k e dbh growth, this suggests that high density o f trees i n the lower canopy could adversely affect height growth o f the upper canopy. This was i n agreement with K o r o l (1985) who observed a negative relationship between height growth and density o f interior Douglas-fir stands. This was also consistent with observations o f Steen and Coupe (1997) in the IDFdk. Barrett (1963) demonstrated a similar impact on ponderosa pine. Volume: A s suggested by O'Hara (1988), differences i n volume growth efficiency (Figure 9, page 61) could be related to variation i n photosynthate allocation to stem volume growth, as well as to variation in leaf area efficiency. In most plots, stand vertical structure was moderately differentiated. In the longer term, it w i l l be interesting to observe how increasing stand differentiation w i l l be reflected in differences i n volume growth efficiency for various categories o f trees, with what impact on stand volume growth, and what consequences on stand structure objectives. In this regard, since crown projection area imperfectly reflected crown size, more precise measurement o f leaf area would be required so as to provide a better estimate o f growth efficiency. C r o w n length: The results suggested that crown length o f small trees increased faster i n the spaced plots than i n the control plots (Table 24, page 65). In contrast, spacing had no impact on crown length o f large trees. Crown length was frequently difficult to measure since epicormic branches occurred on many trees. When such branches were well developed and likely to grow larger, they were considered to be part o f the crown, which sometimes led to considerable increases i n crown length. H o w much impact this had on means remains undetermined. V e r y likely, differences i n crown length growth were also due to the fact that residual trees grew faster in height on spaced plots, while their crown base did not recede (Aussenac et al. 1982). 80  Crown area: Not surprisingly, spacing favored crown area expansion o f both small and large trees, but there was apparently no differences between standard and clumped spacing (Table 22, page 62). Stiell (1982) reported interesting results o f a thinning experiment designed to compare growth o f trees in 4-tree clumps with that o f uniformly spaced trees at the same density (890 trees/ha) i n a 13-year-old red pine (Pinus resinosa Ait.). Ten years after thinning, clumped trees had smaller crowns, presumably because o f higher competition i n the inside o f the clumps, and smaller dbh. However, clumped trees had more cylindrical stems and overall, volume per ha was about the same in both treatments. Fifteen years after spacing, average crown size was about the same because uniformly spaced trees had started to experience competition from their neighbours.  5.4 M A N A G E M E N T I M P L I C A T I O N S The "BDq" approach is used and encouraged i n dry Douglas-fir stands i n the interior o f British Columbia (Vyse et al. 1990, B . C . Ministry o f Forests 1992, Steen and Coupe 1997). A t Knife Creek, this approach was chosen by Day (1998) to regulate stand structure. The results o f this study provide information on the possible evolution o f stands at Knife Creek, and elsewhere in the area. They also raise several issues regarding the application o f the " B D g " method, especially the use o f the q factor: •  W i t h the BD<jr approach, tree diameter is taken as the index o f age (Smith 1986, p.  436). Clearly, results showed that diameter distribution did not reflect age distribution. Since trees o f the same cohort constituted the bulk o f the stands, it w i l l be necessary sooner or later, to sacrifice some trees i n this cohort in order to provide growing space for a new age class. In addition, since there is a lack o f trees in the largest dbh classes, it w i l l be necessary to keep the oldest ones longer than might otherwise be desirable. Increment cores showed that those trees are 81  slow growing. After the heavy diameter limit cut o f 1961, they are also probably not the best phenotypes to rely on as a seed source (Howe 1995), whereas there is currently a need for a new cohort. •  A necessary assumption o f the B D q approach is a uniform rate o f diameter  increment across the regulated portion o f the diameter distribution (Fiedler 1995). Although data was not available for the upper part o f the diameter distribution, preliminary results suggested that it is not the case. •  A balanced uneven-aged stand with a negative exponential diameter distribution  has every age/size class covering an equal area (Smith 1986, p. 435, N y l a n d 1996, p. 201). A s advocated by several authors (Schiitz 1989, L o n g and Daniel 1990, Cochran 1992, O'Hara 1996), more growing space should be allocated to medium-sized and large trees wherever possible, because they increase i n merchantable volume rapidly. In the IDFdk3, regeneration is abundant (Steen and Coupe 1997). Small trees can be grown under large trees, and when they are released, can increase the size o f their crown after decades o f suppression. Therefore, they require less growing space than what would be necessary i f they were grown in the open (Smith 1986 p. 439). Only enough small trees should be kept to eventually provide replacements for merchantable trees. In addition, since one o f the forest management objectives at Knife Creek is the production o f "high quality, large diameter saw timber and veneer logs" (Day 1998), it seems critical that saplings and small poles grow slowly in the shade o f the overstory, prior to their release, in order to limit the development o f large branches, accelerate the recession o f the lower part o f the crown, and reduce the production o f juvenile wood. This influence o f shading on wood quality has been recognized for a long time, not only for tolerant species (Schiitz 1989). O'Hara (1996, 1998) suggested that other structures than those defined by a reverse-J diameter distribution may be feasible, sustainable, and even more productive. However, it may 82  be more appropriate simply to adopt a more flexible approach to the reverse-J regulation o f these forests. Smith et al. (1997, p. 379) provide some guidance by suggesting: " A less rigorous and reasonably successful solution to the problem [of diameter distribution] is to be content with maintaining uneven-aged stands that fluctuate rather widely around the diameter distributions that are tentatively deemed to be appropriate. This process is apparently followed i n long-continued management o f selection stands i n - Europe. During this process, changes in distribution and stand volume are monitored, and all reasonable efforts are made to readjust them at harvest times".  83  6. SUMMARY AND CONCLUSIONS  Although the results o f the study are based on only four years' growth following the first measurement, they already reveal interesting trends. The variables examined at the stand level did not differ significantly among the treatments (Hypothesis 1, page 2). However, treatment means o f quadratic mean diameter, basal area, volume, and accumulated crown area were consistently lower i n the control plots than i n the spaced plots. In the spaced plots, although there was no clear difference among spacing regimes i n absolute terms, treatment C 2 showed consistently higher growth rates i n relative terms. Stand diameter growth was 0.37 c m on the control, and ranged from 0.80 to 1.08 c m on the spaced plots. Net basal area growth averaged 2.34 m /ha on the control, and 3.23 to 3.52 2  m /ha on the spaced plots, while mortality represented 0.83 m /ha on the control, and 0.01 to 0.22 m /ha on the spaced plots. Spacing did not reduce net periodic annual volume increment which was 5.2 to 6.6 m /ha on the spaced plots and 4.8 m /ha on the control. A t the individual tree level, small (< 15 c m dbh) Douglas-fir tree growth responses significantly differed between spacing treatments and control for diameter, basal area, height, and volume. Large (>15 cm dbh) Douglas-fir tree diameter and basal area growth responses significantly differed between treatment C 2 and the control. The other variables examined did not differ significantly among the treatments (Hypothesis 2, page 2). For each o f the variables, 84  small and large Douglas-fir tree growth responses were consistently higher on spaced plots than on the controls. In addition, except for height and crown length, growth was consistently higher on treatment C 2 than on the other spacing regimes. In most analyses, the assumptions necessary to test the third hypothesis (page 2) were not met. For each o f the variables o f interest, small tree growth response was negatively related to Curtis' (1982) relative density. Diameter and basal area o f large trees were related to this index, whereas crown size was not. U p to 20-25 cm, volume growth efficiency regularly increased with increasing dbh, on both control and spaced plots. Trees seemed to be the most efficient on treatment C 2 and the least efficient on the controls. The study also showed that stand age structure was clearly distinct from diameter distribution. Most Douglas-fir were less than 30 c m dbh and appeared to belong to the same cohort, about 66 to 91 years old. A few trees were greater than 30 c m dbh and may possibly belong to more than two different older cohorts. Overall, the results stress the importance o f an appropriate stocking control that maintains both high stand volume growth and high individual tree diameter growth. Spacing allows betterquality stems to grow into merchantable classes more rapidly, and reduces losses from mortality. W i t h increasing stand structure differentiation, appropriate thinning w i l l become more and more challenging, and the results raise several issues regarding the use o f the q factor to regulate stand structure i n multi-aged interior Douglas-fir stands. The need for a large number o f small trees seems particularly questionable. Flexibility in the application o f the B D g method is recommended.  85  LITERATURE CITED Alexander, R . R . , D . Tackle, and W . G . Dahms. 1967. Site indexes for lodgepole pine, with correction for stand density; methodology. U S D A For.Ser. Gen. Tech. Rep. R M - 2 9 . 18 p. Alexander, R . R . and C . B . Edminster. 1977. Regulation and control o f cut under uneven-aged management. U S D A For.Serv. Res. Pap. R M - 1 8 2 . 7 p. Armleder, H . M . , R . J . Dawson, and R . N . Thompson. 1986. Handbook for timber and mule deer management coordination on winter range i n the Cariboo Forest Region. B . C . Ministry o f Forests. Victoria. B . C . Land M a n . Handb. 13. 98 p. Armleder, H . M . , D . A . Leckenby, D . J . Freddy, and L . L . Hicks. 1989. Integrated management o f timber and deer: interior forests o f western North America. U S D A For.Serv. Gen. Tech. Rep. P N W - 2 2 7 . 23 p. Armleder, H . M . and R.J. Dawson. 1992. Logging on mule deer winter range: an integrated management approach. The Forestry Chronicle, 68:132-137. Arno, S.F. 1980. Forest fire history in the northern Rockies. Journal o f Forestry. 78:460-465. Arno, S.F. 1990. Ecological relationships o f interior Douglas-fir. In Interior Douglas-fir: the species and its management. D . Baumgartner and J. Lotan, eds. Washington State University. Pullman, p.47-52. Assmann, E . 1970. The principles o f forest yield study. Pergamon Press. N e w Y o r k . 606 p. Aussenac, G . , A . Granier, and R. Naud. 1982. Influence d'une eclaircie sur la croissance et le bilan hydrique d'un jeune peuplement de Douglas (Pseudotsuga menziesii M i r b . Franco). Canadian Journal o f Forest Research 12:222-231. Aussenac, G . and A . Granier. 1988. Effects o f thinning on water stress and growth i n Douglasfir. Canadian Journal o f Forest Research 18:100-105. Barrett, J . W . 1963. Dominant ponderosa pine do respond to thinning. U S D A For. Serv. Res. Note P N W - 9 . 8 p. Barrett, J.W. 1965. Spacing and understory vegetation affect growth o f ponderosa pine saplings. U S D A For. Serv. Res. Note P N W - 2 7 . 8 p. 86  Barrett, J.W. 1970. Ponderosa pine saplings respond to control o f spacing and understory vegetation. U S D A For. Serv. Res. Pap. P N W - 1 0 6 . 16 p. Barrett, J.W. 1981. Twenty-year growth o f thinned and unthinned ponderosa pine in the M e t h o w Valley o f northern Washington. U S D A For. Serv. Res. Pap. P N W - 2 8 6 . 13 p. Barrett, J . W . 1982. Twenty-year growth response o f ponderosa pine saplings thinned to five spacings i n central Oregon. U S D A For. Serv. Res. Pap. P N W - 3 0 1 . 18 p. Becker, R. 1995. Operational considerations o f implementing uneven-aged management. In Uneven-aged management: opportunities, constraints, and methodologies. K . L . O'Hara, editor. Montana For. and Cons. E x p . Sta. School o f Forestry. University o f Montana. M i s c . Pub. N o 56. p.67-81. Bonnor, G . M . Growth and yield study o f interior Douglas-fir in British Columbia. In Interior Douglas-fir: the species and its management. D . Baumgartner and J. Lotan, eds. Washington State University. Pullman, p.269-274. Breadon, J . M . 1981. Heavy thinning response i n an interior Douglas-fir stand. Unpub. B . S . F . Thesis. Faculty o f Forestry. University o f British Columbia. Vancouver. B . C . British Columbia Forest Service. 1976. Whole stem cubic meter volume equations and tables. Forest Inventory Division, B . C . For. Serv., Victoria, B . C . British Columbia Ministry o f Forests. 1992. Correlated guidelines for management o f unevenaged drybelt Douglas-fir stands i n British Columbia. Victoria, B . C . 59 p. Clark, M . B . 1952. A preliminary study o f growth and development o f some Douglas-fir ponderosa pine forest types. B . C . Forest Service, Research Division, Victoria. B . C . Tech. Pub. T-38. 39 p. Cochran, P . H . 1992. Stocking levels and underlying assumptions for uneven-aged ponderosa pine stands. U S D A For. Serv. Res. Note P N W - R N - 5 0 9 . 10 p. Crown, M . , R . V . Quenet, and C . Layton. 1977. Fertilization and thinning effects on a Douglas-fir ecosystem at Shawnigan Lake: 3-year growth response. Can. For. Serv. Rep. B C - X - 1 5 2 . 36 p. Victoria. B . C . Curtis, R . O . 1982. A simple index o f stand density for Douglas-fir. Forest Science 28:92-94. Curtis, R . O . 1992. Levels-of-growing-stock cooperative study i n Douglas-fir: Report N o . 11Stampede Creek: a 20-year progress report. U S D A For. Ser. Res. Pap. P N W - R P - 4 4 2 . Curtis, R . O . and D . D . Marshall. 1989. O n the definition o f stand diameter growth for remeasured plots. Western Journal o f Applied Forestry 4(3): 102-103. 87  Curtis, R . O . and D . L . Reukema. 1970. Crown development and site estimates i n a Douglas-fir plantation spacing test. Forest Science 16:287-301. Daniel, T . W . , J . A . Helms, and F.S. Baker. 1979. Principles o f silviculture (second edition). M c G r a w - H i l l Book Company. N e w York. 500 p. Day, J . K . 1997. Stocking standards for uneven-aged interior Douglas-fir. Unpub. Directed Study Report. Faculty o f Forestry. University o f British Columbia. Vancouver. B . C . 29 p. Day, J . K . 1998. Selection management o f interior Douglas-fir for mule deer winter range. Master's Thesis. Faculty o f Forestry. University o f British Columbia. Vancouver. B . C . Dormer B . L . , and S.W. Running. 1986. Water stress response after thinning Pinus contorta stands i n Montana. Forest Science 3:614-625. Farrar, R . M . 1981. Regulation o f uneven-aged loblolly-shortleaf pine forests. U S D A For. Serv. . Gen. Tech. Rep. SO-34:294-304. Farron, P . E . 1980. Application de la methode du controle dans les forets communales des Verrieres de 1893 a 1968. Schweiz. Z . Forstwes. 131: 675-705. Fiedler, C . E . 1995. The basal area-maximum diameter-^ (BDq) approach to regulating unevenaged stands. In Uneven-aged management: opportunities, constraints, and methodologies. K . L . O'Hara, editor. Montana For. and Cons. E x p . Sta. School o f Forestry. University o f Montana. M i s c . Pub. N o 56. p. 94-109. Gingrich, S.F. 1967. Measuring and evaluating stocking and stand density i n upland hardwood forests i n the central States. Forest Science 13:38-53. Glew, D . R . and S.Z. Cinar. 1966. The results o f stand treatment following seed tree marking in drybelt fir types. B C Forest Service. For. Mgmt. Notes, N o . 5 . 17 p. Grier, C . C . and S.W. Running. 1977. L e a f area o f mature northwestern coniferous forests: relation to site water balance. Ecology 58:893-899. Guldin, J . M . 1991. Uneven-aged B D q regulation o f Sierra Nevada mixed conifers. Western Journal o f Applied Forestry 6:27-32. Harrington, C A . and D . L . Reukema. 1983. Initial shock and long-term stand development following thinning i n a Douglas-fir plantation. Forest Science 29:33-46. Helms, J . A . and J.E. Lotan. 1987. Selecting silvicultural systems for timber. In Ponderosa pine: the species and its management. D . Baumgartner and J. Lotan (ed.) Washington State University. Pullman, p.221-226. Helvey, J.D. 1975. Soil moisture depletion and growth rates after thinning ponderosa pine. U S D A For. Serv. Res. Note P N W - 2 4 3 . 9 p. 88  Hicks, C . R . 1993. Fundamental concepts i n the design o f experiments (fourth edition). Saunders College Publishing. N e w York. 509 p. Hope, G . D . , W . R . Mitchell, D . A . L l o y d , W . R . Erickson, W . L . Harper and B . M . Wikeem. 1991. Interior Douglas-fir zone. In Ecosystems o f British Columbia. D . Meidinger and J. Pojar (ed.). B . C . M i n . For., Victoria, B . C . Special Report Series 6. Chap. 10, p.153-166. Howe, G . E . 1995. Genetic effects o f uneven-aged management. In Uneven-aged management: opportunities, constraints, and methodologies. K . L . O'Hara (ed.) Montana For. and Cons. E x p . Sta. School o f Forestry. University o f Montana. M i s c . Pub. N o 56. p.27-32. J.S.Thrower and Associates Ltd. 1997 (a). Management and silviculture opportunities for interior Douglas-fir stands in the Pothole Creek Demonstration Area. Unpublished contract report to Weyerhaeuser Canada Ltd. and the Ministry o f Forests, Merritt Forest District. 75 p. J.S.Thrower and Associates L t d . 1997 (b). Growth and yield o f partially-cut Douglas-fir forest types in the Merritt T S A . Unpublished contract report to Weyerhaeuser Canada L t d . and the Ministry o f Forests, Merritt Forest District. 37 p. Johnstone, W . D . 1985. A n analysis o f problems in estimating the growth and yield o f interior dry-belt stands. B . C . Ministry o f Forests, Research Branch, Kalamalka Research Station, Vernon. B . C . Report RR85001-HQ. 52 p. Keyes, M . R . and C C . Grier. 1981. Above- and below-ground net production i n 40-year-old Douglas- fir stands on low and high productivity sites. Canadian Journal o f Forest Research 11: 599-605. Kilgore, B . M . 1981. Fire i n ecosystem distribution and structure: western forests and scrublands. Proc. Fire regimes and ecosystem properties. H . A . Mooney, T . M . Bonnickson, N . L . Christensen, J.E. Lotan and W . A . Reiners (Coord.). U S D A For. Serv. Gen. Tech. Rep. W O 26. K i m m i n s , J.P. 1997. Forest ecology (updated edition). M a c m i l l a n Publishing Company. N e w Y o r k . 531 p. K o r o l , R . L . 1985. The soil water regime and growth o f uneven-age interior Douglas-fir (Pseudotsuga menziesii var. glauca) stands. Master's Thesis. Faculty o f Forestry. University o f British Columbia. Vancouver. B . C . Kramer, P.J., and T . T . K o z l o w s k i . 1979. Physiology o f woody plants. Academic Press. N e w York. 811 p. Laar, A . van, and A . A k c a . 1997. Forest mensuration. Cuvillier Verlag. Gottingen, Germany. 418 p.  89  Leak, W . B . 1963. Calculation o f "q" by the least squares method. Journal o f Forestry 61:227228. Leak, W . B . 1964. A n expression o f diameter distribution o f unbalanced, uneven-aged stands and forests. Forest Science 10:39-50. Liocourt, F . de. 1898. De l'amenagement des sapinieres. Bulletin de la societe forestiere de Franche Comte et de Belfort. 4: 396-409. L o n g , J . N . 1985. A practical approach to density management. The Forestry Chronicle 23-27. L o n g , J.N., and T.J. Dean. 1986. Sapwood area o f Pinus contorta as a function o f mean size and density. Oecologia 68:410-412. L o n g , J.N., and T . W . Daniel. 1990. Assessment o f growing stock in uneven-aged stands. Western Journal o f Applied Forestry 5:93-96. Lopushinsky, W . 1990. Water relations o f interior Douglas-fir. In Interior Douglas-fir: the species and its management. D . Baumgartner and J. Lotan, eds. Washington State University. Pullman, p. 35-46. Lundqvist, L . 1993. Changes in the stand structure on permanent Picea abies plots managed with single-tree selection. Scandinavian Journal o f Forest Research 8:510-517. L y n c h , D . W . 1958. Effects o f stocking on site measurement and yield o f second growth ponderosa pine in the Inland Empire. U S D A For. Serv. Res. Pap. INT-56. 36 p. McCaughey,W.W.and D . E . Ferguson. 1988. Response o f advance regeneration to release i n the inland mountain west: a summary. In Future forests o f the mountain west: a stand culture symposium. U S D A For. Serv. Gen. Tech. Rep. INT-243, pp.255-266. Marquis, D . A . 1978. Application o f uneven-aged silviculture and management on public and private lands. In Uneven-aged silviculture and management i n the United States. U S D A For. Serv. Gen. Tech. Rep. W O - 2 4 . 234 p. Marshall, D . D . and J.C. Tappeiner. 1992. Levels-of-growing-stock cooperative study i n Douglas-fir: Report N o . 10-The Hoskins Study,1963-1983. U S D A For. Ser. Res. Pap. P N W - R P - 4 4 8 . 35p. Marshall, P . L . 1996. Response o f uneven-aged Douglas-fir to alternative spacing regimes: analysis o f the initial impact o f the spacing regimes. Canada-British Columbia Partnership Agreement on Forest Resource Development: F R D A - I I Report 242. 27 p. Marshall, P . L . and Y . Wang. 1996. Growth o f uneven-aged interior Douglas-fir as influenced by different stand structures. Canada-British Columbia Partnership Agreement on Forest Resource Development: F R D A - I I Report 267. 20 p. 90  Marshall, P . L . and J.-L. Bugnot. 1997. Response o f uneven-aged Douglas-fir stands to juvenile spacing. Presented to I U F R O Congress, Corvallis O R . Meyer, H . A . 1931. Effect o f release upon the form and volume o f western yellow pine. Journal o f Forestry 29:1127-1133. M i l l e r , R . E . and D . L . Reukema. 1977. Urea fertilizer increases growth o f a 20-year- old thinned Douglas-fir on a poor quality site. U S D A For. Serv. Res. Note P N W - 2 9 1 . 8 p. Mitscherlich, G . 1963. Untersuchungen in Schlag- und Plenterwaldern. Allgemeine-Forst-undJagdzeitung 134:1-12. Newsome, T . A . , D . C . Sutherland, and A . Vyse. 1990. Establishing Douglas-fir plantations in the dry belt o f Interior British Columbia. In Interior Douglas-fir: the species and its management. D . Baumgartner and J. Lotan, eds. Washington State University. Pullman. p.227-231. Nyland, R. D . 1996. Silviculture: concepts and applications. M c G r a w - H i l l . N e w Y o r k . 633 p. O'Hara, K . L . 1988. Stand structure and growing space efficiency following thinning i n an evenaged Douglas-fir stand. Canadian Journal o f Forest Research 18:859-866. O'Hara, K . L . 1989. Stand growth efficiency i n a Douglas-fir thinning trial. Forestry 62:409-418. O'Hara, K . L . 1996. Dynamics and stocking-level relationships o f multi-aged ponderosa pine stands. Forest Science 42 (4) Monograph 33. O'Hara, K . L . 1998. Silviculture for structural diversity: a new look at multiaged systems. Journal o f Forestry 7:4-10. O'Hara, K . L . , and N . I Valappil. 1995. A g e class division o f growing space to regulate stocking in uneven-aged stands. In Uneven-aged management: opportunities, constraints, and methodologies. K . L . O'Hara, editor. Montana For. and Cons. E x p . Sta. School o f Forestry. University o f Montana. M i s c . Pub. N o 56. p.123-143. Oliver, C D . 1995. Uneven-age stand dynamics. In Uneven-aged management: opportunities, constraints, and methodologies. K . L . O'Hara, editor. Montana For. and Cons. E x p . Sta. School o f Forestry. University o f Montana. M i s c . Pub. N o 56. pp. 82-93. Oliver, C D . and B . C . Larson. 1996. Forest stand dynamics (update edition). John W i l e y & Sons, N e w York. 520 p. Omule, S.A.I. 1988. Growth and yield 35 years after commercially thinning 50-year-old Douglas-fir. Canada-British Columbia Partnership Agreement on Forest Resource Development: F R D A Report 267. 15 p. Parde, J. 1980. Forest biomass. For. Abst. 41:343-362. 91  Province o f B . C . 1995. Biodiversity guide book. F P C M o F / M o E joint pub. Queens Printer Victoria. Rawlings, J.O., S . G . Pantula and D . A . Dickey. 1998. A p p l i e d regression analysis: a research tool (second edition). Springer. N e w York. 657 p. Rehfeldt, G . E . 1989. Ecological adaptations in Douglas-fir (Pseudotsuga menziesii var. glaucri): a synthesis. Forest Ecology and Management 28:203-215. Reineke, L . H . 1933. Perfecting a stand density index for even-aged forests. Journal o f Agriculture Research. 46:627-638. Reukema, D . L . 1979. Fifty-year development o f Douglas-fir stands planted at various spacing. U S D A For. Serv. Res. Pap. P N W - 2 5 3 . 21 p. Schiitz, J-P. 1989. L e regime du jardinage. Document autographique du cours de sylviculture 3. Chaire de sylviculture, E T H , Zurich. Switzerland. 55 p. Sjolte-J0rgensen, J. 1967. The influence o f spacing i n the growth and development o f coniferous plantations. In J . A . Romberger and P. M i k o l a (eds.) International Review o f Forestry Research. Academic Press. N e w york. Vol.2:43-94. Smith, D . M . 1986. The practice o f silviculture (eighth edition). John W i l e y & Sons, N e w Y o r k . 527 p. Smith, D . M . , B . C . Larson, M . J . Kelty, and P . M . S . Ashton. 1997. The practice o f silviculture: applied forest ecology (ninth edition). John W i l e y & Sons, N e w Y o r k . 537 p. Smith, J . H . G . 1985. Analysis o f growth o f rings, trees, and stands to improve yields and management o f interior Douglas-fir types. Faculty o f forestry. University o f British Columbia. Vancouver. B . C . 100 p. Snedecor, G . W . and W . G . Cochran. 1980. Statistical methods (seventh edition). The Iowa State University Press. Ames, Iowa, U S A . 507 p. Spurr, S . H . 1952. Forest inventory. The Ronald Press Company. N e w Y o r k . 476 p. Staebler, J.R. 1956. Evidence o f shock following thinning o f young Douglas-fir. Journal o f Forestry 54:339. Standish, J.T., G . H . Manning and J.P. Demaerschalk. 1985. Development o f biomass equations for British Columbia tree species. Can. For. Serv. Rep. Inform. Rep. B C - X - 2 6 4 . Steel, R . G . D . , J . H . Torrie and D . A . Dickey. 1997. Principles and procedures o f statistics: a biometrical approach (third edition). M c G r a w H i l l . N e w York. 666 p. 92  Steen, O . A . and R . A . Coupe. 1997. A field guide to forest site identification and interpretation for the Cariboo Forest Region. B . C . M i n . For., Victoria, B . C . , L a n d Manage. Handb. No.39. Stiell, W . M . 1982. Growth o f clumped vs. equally spaced trees. Forestry Chronicle 58:23-25. Vyse, A . 1981. The effect o f canopy removal by diameter limit cutting on plant moisture stress and foliage condition i n advanced regeneration o f Doulas-fir (E.P.841). In M o F , R B , For. Res. Rev. 1980-81:34-35. Vyse, A . , R . A . Smith and B . G . Bondar. 1990. Management o f interior Douglas-fir stands i n British Columbia: past, present and future. In Interior Douglas-fir: the species and its management. D . Baumgartner and J. Lotan, eds. Washington State University. Pullman. p.177-186. Waring, R . H . and W . H . Schlesinger. 1985. Forest ecosystems: concepts and management. Academic Press. N e w York. 340 p.  93  APPENDICES  94  A P P E N D I X I: D I A M E T E R D I S T R I B U T I O N O F T R E E S B Y S P E C I E S A . D I A M E T E R D I S T R I B U T I O N O F T R E E S B Y S P E C I E S I N 1993 Treatment Cl  Dbh (cm)  Fd  PI  0-5 5-10 10-15  283 773 463  10 33 43  53 103 87  24 3  110 27 43  1,520 257 13  87 20  243 97  27  180 30  2,057 404 16  1,790  107  210  2,477  0-15 15-30 >30 All C2  S  Control  777 380 373  7 50  0-15 15-30 >30  930 174 19  57 113 3  All  1,123  173  157 527 590  Ep  3  0-5 5-10 10-15  0-5 5-10 10-15  Density (trees/ha) Sx At  340  31  All species 481 940 637  47  33 10 7  257 397 443  13 17  47  50 13  1,097 318 22  30  47  63  1,437  47  20  13  37 3 23  3  260 533 613  0-15 15-30 >30  1,273 327 14  63 73 10  3  47  20  1,407 399 24  All  1,613  147  3  47  20  1,830  0-5 5-10 10-15  2,437 2,197 820  20 40 40  70 50 77  2,696 2,477 1,093  0-15 15-30 >30  5,453 193 23  100 33  417 57  100 10  197 20  6,266 314 23  All  5,670  133  473  110  217  6,603  43 13 43  127 177 113  Columns and rows may not exactly sum to the totals due to rounding.  95  B. D I A M E T E R DISTRIBUTION OF TREES B Y SPECIES IN 1997 . 1  Treatment Cl  C2  S  Control  Dbh (cm) 0-5 5-10 10-15  Fd  PI  Ep  All species  277  7  640 533  20 53  50 103 73  10 3  83 23 40  227 113  13  147 37  1,916 503 20  16  183  2,440  36  50 10 7  257 317 437  0-15 15-30 >30  1,450 333 17  80 20  All  1,800  100  0-5 5-10 10-15  167 307 383  3  0-15 15-30 >30 All 0-5 5-10 10-15  Density (trees/ha) Sx At  3 340  426 790 700  40  7  857 263 20  40 130 3  7 23  36  47 13  1,010 430 23  1,140  173  30  36  60  1,464  143 420 597  40 7 10  77  17  3  277 430 607  0-15 15-30 >30  1,160 424 13  57 80 16  3  77  17  1,314 504 29  All  1,597  154  3  34  17  1,847  0-5 5-10 10-15  2,174 2,110 897  23 17 37  93 157 107  10 27 30  70 43 73  2,370 2,353 1,143  0-15 15-30 >30  5,180 257 26  11  40  357 73  61 1  186 23  5,867 401 26  All  5,464  117  430  73  210  6,293  Columns and rows may not exactly sum to the totals due to rounding.  96  

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.831.1-0075192/manifest

Comment

Related Items