Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Regeneration patterns on some old-growth and clearcut sites in the Mountain Hemlock zone of southern… Brett, Robert B. 1997

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

Item Metadata

Download

Media
831-ubc_1997-0301.pdf [ 5.41MB ]
Metadata
JSON: 831-1.0075274.json
JSON-LD: 831-1.0075274-ld.json
RDF/XML (Pretty): 831-1.0075274-rdf.xml
RDF/JSON: 831-1.0075274-rdf.json
Turtle: 831-1.0075274-turtle.txt
N-Triples: 831-1.0075274-rdf-ntriples.txt
Original Record: 831-1.0075274-source.json
Full Text
831-1.0075274-fulltext.txt
Citation
831-1.0075274.ris

Full Text

R E G E N E R A T I O N P A T T E R N S O N S O M E O L D - G R O W T H A N D C L E A R C U T S I T E S IN T H E M O U N T A I N H E M L O C K Z O N E O F S O U T H E R N B R I T I S H C O L U M B I A by R O B E R T B. B R E T T B.A. Universi ty of Wes te rn Ontar io 1981 A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in F A C U L T Y O F G R A D U A T E S T U D I E S Department of Forestry (Forest Sc iences ) W e accept this thesis as conforming to the required s tandard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A M a y 1997 © Robert Bruce Brett In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Patterns of tree regenerat ion change with elevat ion in old-growth forest s tands on Brit ish Co lumb ia ' s southern coast . At lower e levat ions, where snow is infrequent, regenerat ion tends to be cent red in canopy gaps c a u s e d by the death of one or more trees (the gap model ) . At higher e levat ions, where snow usual ly remains until early summer , regenerat ion is restr icted to e levated microsi tes that emerge earl iest from the snow (the t ree- is land model) . Little is yet known about regenerat ion patterns in old-growth forest s tands between these two s y s t e m s , those within the forested Mounta in Hemlock (MH) biogeocl imat ic s u b z o n e . O u r lack of knowledge has b e c o m e more of a concern a s logging p rog resses into higher e levat ions of the s u b z o n e where there are ever -deeper snowpacks . To investigate these patterns, I es tab l ished 12 study s i tes in the Tet rahedron R a n g e near Sechel t , 50 km northwest of Vancouve r . S ix s i tes were in old-growth forest s tands, and six were in natural ly-regenerated c learcuts that had been logged 1.1-12 years prior to sampl ing . S ix si tes were s teep ( - 5 0 % slope) and six were flat (~ 2 5 % s lope) . E levat ions were slightly higher for old-growth s i tes (1080-1195 m) than c learcut s i tes (1060-1100 m). Old-Growth Sites: T rees were very s low-growing and took an average of a lmost 500 years to enter the canopy layer. Regenerat ion w a s most success fu l on mounds and near to a canopy tree. It w a s unaffected by overhead canopy cover (that is, the p resence or a b s e n c e of a canopy gap), apparent ly b e c a u s e of the preva lence of low-angle, diffuse light. In contrast to most forested e c o s y s t e m s , a lmost all t rees were growing on the undisturbed forest f loor rather than on decay ing wood or mineral soi l . Overa l l , regenerat ion patterns were more consis tent with the t ree- is land model than the gap model of regenerat ion s ince trees were most likely to surv ive on mounds and c lose to a canopy tree. Stil l , the p resence of s o m e regenerat ion in gaps , espec ia l ly on s teep s i tes, showed that the study si tes occup ied a transit ion be tween the gap and t ree- is land mode ls . The tree- is land model w a s best exp ressed on late-snowmelt s i tes that were most similar to high-elevat ion s i tes. It w a s a lso more apparent in the regenerat ion patterns of Chamaecyparis nootkatensis (A laska yel low-cedar) , a spec ies near the upper limit of its e levat ional range, than those of Tsuga mertensiana (mountain hemlock) , a s p e c i e s in the middle of its range. Clearcut Sites: A lmost all t rees >150 cm tall were A. amabilis (Paci f ic si lver fir) which had been present in the previous old-growth stand before cutting. A surprisingly high proport ion of t rees (45%) estab l ished within a 3-year window from 1 year before logging through 1 year after logging, more than half of which were C. nootkatensis. On ly 2 0 % of regenerat ion es tab l ished more than one year after logging, and none estab l ished >8 years after logging. Th is l imited ingress likely resulted from the a b s e n c e due to clearcutt ing of nearby seed-p roduc ing t rees. There w a s much more friable forest floor and coa rse woody debr is (from logging s lash) than in adjacent o ld-s tands, but a lmost all regenerat ion w a s still found on undisturbed forest floor. Regenera t ion w a s less c o m m o n on mounds in c learcuts than on mounds in adjacent o ld-growth s tands , apparent ly b e c a u s e mounds were disturbed during logging more than other micros i tes. There w a s no ev idence that Vaccinium spp . (blueberr ies and huckleberr ies) impeded regenerat ion s ince 8 4 % of t rees and seed l ings were growing below or amidst Vaccinium and establ ishment and survival w a s higher where it w a s present. T h e s tands that deve lop on these c learcuts will remain for many centur ies dramat ical ly different f rom the old-growth forest s tands they rep laced. W h e r e cutting is appropr iate, s u c h negat ive features could be avo ided by leaving an adequate s e e d source , retaining live and d e a d canopy t rees, and protecting sub-canopy trees during cutt ing. S i tes wou ld then a lso retain many of the old-growth character ist ics required by wildlife and other, non-t imber va lues . Resu l t s from old-growth si tes highlight the si te-speci f ic nature of regenerat ion patterns and the abrup tness of the transition to t ree- is land patterns. Ye t low-elevat ion cutting methods, espec ia l ly c learcutt ing, are still used within this transition even where regenerat ion requires the protect ion of an overhead canopy. Any p resence of regenerat ion patterns match ing the tree-is land mode l shou ld warn forest managers of potential regenerat ion prob lems. In s u c h a reas , the dec is ion to cut shou ld not be automatic, especia l ly g iven the s low growth and high non-t imber va lues of these forests. W h e r e cutting does occur , it shou ld leave a s much of the sub -canopy and canopy layers as poss ib le . A s snow inc reases further and there is a greater p resence of regenerat ion patterns matching the tree- is land mode l , any cutting is inappropriate. T h e relat ionship between regenerat ion patterns and s n o w depths could provide an eco log ica l bas is for manag ing forests within the M H zone . Speci f ical ly , the p resence of regenerat ion patterns that match the tree- is land model (even if d iscrete tree is lands are not present) is a reflection of severe growing condit ions and potential regenerat ion prob lems. S imp le m e a s u r e s of the relative abundance of t ree- is land patterns could be a d d e d during s tandard site d iagnos is to determine the severi ty of growing condi t ions, e.g. , the proport ion of understory and sub -canopy trees that are growing near a canopy tree or on mounds . S u c h a c lassi f icat ion wou ld be appl icable regard less of management object ive. V Table of Contents Abst ract ii Tab le of Con ten ts v List of Tab les vii List of F igures viii A c k n o w l e d g e m e n t s ix Chap te r 1. Introduction 1 1.1 T h e Mounta in Hemlock (MH) Zone 1 1.2 Regenera t ion Pat terns in the M H Zone 3 1.3 Objec t ives and Thes i s Outl ine 5 Chap te r 2. Regenera t ion Pat terns on Old-Growth S i tes 6 2.1 Introduction 6 2.2 Me thods 7 Study A r e a 7 Study Des ign 9 Data Co l lec ted 12 Da ta Ana lys i s 18 2.3 Resu l t s 20 Structure, Compos i t i on , and A g e 20 T h e Microsi te Envi ronment 24 Subst ra tes 24 Microtopography 25 C a n o p y C o v e r and Dis tance to the Nearest C a n o p y Tree 25 S n o w - Microsi te Relat ionships 27 Tree-Micros i te Rela t ionsh ips 33 Subst ra tes 33 Microtopography and Mound Partners 34 C a n o p y C o v e r and Dis tance to the Nearest C a n o p y Tree 36 Interactions Be tween Microsi te Factors 42 Growth Ra tes and Micros i tes 45 Growth Form Anoma l i es 45 2.4 D iscuss ion 46 A re micros i tes, snow, and regenerat ion patterns related? 46 Do regenerat ion patterns reflect the gap or t ree- is land mode l? 49 2.5 C o n c l u s i o n s 51 Chap te r 3. Natural Regenera t ion on Clearcut S i tes 5 3 3.1 Introduction 5 3 3.2 Me thods 53 Study A r e a 53 Study Des ign 54 Data Co l lec ted 55 Data Ana lys i s 56 vi 3.3 Resu l t s 59 Structure, Compos i t i on , and A g e 59 Subst ra tes 65 Microtopography and Logging S l a s h 67 Vaccinium 68 Stock ing 70 Growth Form Anoma l i es 72 3.4 D iscuss ion 74 W a s natural regenerat ion success fu l? 74 Did most natural regenerat ion establ ish before or after logging? 75 W h i c h subst rates favoured natural regenerat ion? 77 W h i c h microtopographic locat ions favoured natural regenerat ion? 78 Did compet i t ion with Vaccinium impede natural regenerat ion? 79 3.5 C o n c l u s i o n s 80 Chap te r 4. S u m m a r y and Conc lus ions 84 Literature C i ted 86 Append i x A . Non-tree spec ies by life form, occur rence, and percent cover 94 Append i x B. Class i f icat ion of growth form anomal ies 95 Append i x C . Regenera t ion status of c learcut study locat ions in 1992 96 vii List of Tables 2.1 Old-growth study site descr ipt ions 10 2.2 Definit ion of terms used in Chapte r 2 13 2.3 D e c a y classi f icat ion for dead trees 14 2.4 S p e c i e s composi t ion by height c lass and site 21 2.5 Percent cover of substrates by site 25 2.6 C a n o p y cover by site and by high and low shade 26 2.7 S n o w depths by site and s lope type 28 2.8 Root ing substrate of understory t rees, sap l ings, and seed l ings 34 2.9 L ive mound partners of live and dead canopy trees 37 3.1 C learcu t study site descr ipt ions 54 3.2 Definit ion of terms used in Chapte r 3 58 3.3 Percent cover of substrates by site 65 3.4 Growth form anomal ies by spec ies 73 viii List of Figures 1.1 Distribution of the Mounta in Hemlock biogeocl imat ic z o n e 2 1.2 Elevat ional s e q u e n c e of b iogeocl imat ic z o n e s on B .C . ' s southern coast 2 2.1 Locat ion of Tet rahedron Provincia l Park and study a rea 8 2.2 Si te descr ipt ion and sampl ing des ign 11 2.3 C a n o p y cover c lassi f icat ion 13 2.4 Definit ion of high s h a d e and low shade for understory t rees and quadrats 15 2.5 Deve lopmenta l s tages of seed l ings 16 2.6 Height and diameter distributions of t rees by spec ies 22 2.7 Growth rate to a d iameter of 40 c m (at s tump or core height) 2 3 2.8 Quadra ts by canopy cover c lass and d is tance to nearest canopy tree 27 2.9 S n o w depths compared to microtopography 30 2.10 S n o w depths compared to canopy cover 31 2.11 S n o w depths compared to d is tance to the nearest canopy tree 32 2.12 Observed- to -expec ted f requency of t rees by microtopography 35 2 .13 F requency and observed- to-expected f requency of non-canopy t rees by s lope 38 2.14 F requency and observed- to-expected f requency of sub -canopy t rees by canopy cover , d is tance to the nearest canopy tree, and s lope 39 2 .15 F requency and observed- to-expected f requency by high and low s h a d e 41 2.16 S u m m a r y of dead trees by condit ion, spec ies , and canopy layer 43 2.17 F requency and observed- to-expected f requency of sub -canopy trees by microtopography, canopy c lass , and d is tance to the nearest canopy tree 44 3.1 Jitter plot of heights of t rees and seed l ings on their year of estab l ishment 57 3.2 Height distributions of t rees and seed l ings by spec ies 59 3.3 Height distribution of t rees and seedl ings by spec ies and site 60 3.4 Height growth of A. amabilis on clearcut and old-growth si tes 62 3.5 Es t imated age c l ass distribution by spec ies and height c lass 63 3.6 Es t imated age c lass distribution by height c lass and site 64 3.7 Observed- to -expec ted ratio on avai lable substrates by spec ies 66 3.8 T r e e s and seed l ings on undisturbed forest floor by spec ies and height 66 3.9 Observed- to -expec ted ratio of t rees by microtopography 67 3.10 Height of t rees relative to Vaccinium 68 3.11 Pro jected f requency and spec ies composi t ion of f ree-growing t rees 69 3.12 S tock ing during the year of sampl ing and projected stocking after 10 years 71 Acknowledgements M y greatest debt is to Dr. Kare l K l inka whose genuine interest and encouragement first attracted me to U B C Forest S c i e n c e s and suppor ted me throughout my studies. A n y o n e who has had the privi lege of learning with Dr. K l inka knows that a day in the field with him is better (and more fun) than a year in any c lass room. I a lso benefitted from the ass i s tance of my excel lent commit tee. Eve r s ince Dr. Ken Ler tzman first invited me to help him count seed l ings at C y p r e s s , he has been a constant help organiz ing my thinking and writing. Dr. Peter Marsha l l is the best statist ician an ecologist could ask for, though he may not apprec iate the adver t isement . Bob G r e e n appl ied his character ist ic thoroughness to a subject he knows thoroughly, the M H zone , and I value his comments and suggest ions highly. I a m grateful for f inancial support from the B . C . Ministry of Fores ts , a Universi ty Gradua te Fe l lowship from U B C , and scho la rsh ips from the family of Dona ld S . M c P h e e . It w a s an honour to work with Dr. Ja ry Dobry, master dendrochronologis t . T h a n k s a lso to three of my excel lent professors : Dr. Lee G a s s and Dr. Bill R e e s , both of whom contr ibuted to a 180-degree shift in my thinking, and Dr. L e s Lavku l ich, who taught me the di f ference between soil and dirt. Don MacLau r i n , R P F , helped fuel my interest in forest managemen t and encou raged my return to schoo l . I ran through a lot of f ield ass is tants and enjoyed working with each of them: Lori Dan ie ls , Murray D e e , Mart in G o o s e n s , Terry Higgins, Dawn Mann ing , J a n e Mi l len, A n n e m e i k e Smi ts , Pa l V a r g a , and Chr is toph Wi ld Burger. Pa l V a r g a w a s especia l ly helpful early on , when I only knew sampl ing as someth ing done at the buffet table. T h a n k s to Chr is t ine C h o u r m o u z i s and C h a s We l l s for their wizardry in computer graphics. A m o n g the people who generous ly he lped me get my bear ings in the Tetrahedron were S ig L e h m a n of the Sunsh ine C o a s t Reg iona l District, Br ian Smar t and the staff of the Seche l t F ie ld Off ice, and Mike Scot t of Interior. D a n B o u m a n , silviculturalist and friend of the Tet rahedron, a l lowed me to use his beautiful home. Thanks to the Tetrahedron Ski C lub whose backcountry cab ins I v is i ted with great p leasure . Congra ts to all on a great accompl ishment : Tet rahedron Provinc ia l Park! M y mother, Phyl l is Brett, endured my trials of s tudenthood for a longer t ime than most yet she mainta ined her s e n s e of humour through it all - thanks M o m . A s for the rest of my family, okay , I lost the bet. Thanks to those who kept me s a n e , espec ia l ly D a v e , Koerner 's , Lor i , the Kl inko ids, A lex , Pau l , Murray, Rick, and the P e a k Cha i r . A n d , most important, thanks to J a n e for her love, support , and ever-wi l l ingness to take off to the M H zone . Bob Brett M a y 5, 1997 Chapter 1. Introduction 1.1 The Mountain Hemlock (MH) Zone The Mounta in Hemlock (MH) zone , one of B .C . ' s 14 biogeocl imat ic z o n e s , is located at suba lp ine e levat ions from - 9 0 0 - 1 6 0 0 m in southern B . C . and 300-1000 m in northern B . C . (Kraj ina 1969; Pojar e r a / . 1991; Figure 1.1). It is div ided into two s u b z o n e s : forested at lower e levat ions, and park land at higher elevat ions (Brooke e r a / . 1970). T h e c losed -canopy forests of the forested M H subzone , the focus of this study, occupy an elevat ional band between montane forests of the Coas ta l Wes te rn Hemlock ( C W H ) zone and tree is lands of the park land M H s u b z o n e (Orloci 1965; Brooke e r a / . 1970; Figure 1.2). Frequent s torms and heavy precipitation ( -1700-5000 m m annual ly) result f rom the orographic uplift of moist Paci f ic air m a s s e s forced to rise by the coasta l mountain barrier (Brooke e r a / . 1970; Pojar e r a / . 1991). F rom 2 0 - 7 0 % of precipitation falls a s snow which results in deep , heavy snowpacks (2-3+ m) and a snow-f ree per iod of 3-5 months (Brooke et al. 1970). T h e transit ion from the C W H to M H zone , def ined by the elevat ion at wh ich 7". mertensiana (Bong.) Car r . (mountain hemlock) first outnumbers Tsuga heterophylla (Raf.) S a r g . (western hemlock) , is often abrupt and co inc ides with greatly inc reased snow (Peterson 1964, 1969; Or loc i 1965). In addit ion to 7". mertensiana, the other character ist ic spec ies of the M H z o n e are Abies amabilis Doug l . ex Fo rbes (Paci f ic si lver fir) and Chamaecyparis nootkatensis (D. Don) S p a c h (A laska ye l low-cedar ; Brooke e r a / . 1970; Pojar e r a / . 1991). E a c h of these spec ies is very shade-to lerant , long-l ived, and meets criteria for la te -success iona l or c l imax spec ies (Minore 1979; Kraj ina e r a / . 1982; Burns and Honka la 1990). They a lso grow well in full light and share character is t ics with what would common ly be cons idered p ioneer spec ies (Herr ing and Ether idge 1976; Se ide l 1985; An tos and Zobe l 1986; Arnott e r a / . 1995). The M H z o n e is unusua l in that there is usual ly no change in spec ies composi t ion after d is turbance a s only these three spec ies are c o m m o n in all s tages of s tand development . Figure 1.1 Distribution of the Mountain Hemlock biogeoclimatic zone. Figure 1.2. Elevational sequence of biogeoclimatic zones on British Columbia's southern coast. Approximate elevation above sea level 3 A l s o unusua l , due to the coo l , wet cl imate and deep snowpacks , is the inf requency of major d is turbances. S ince catastrophic d isturbance by fire, insects, or d i sease is rare (Brink 1959; B rooke e r a / . 1970; Ler tzman 1989), s tanding mortality and sma l l - sca le windthrows are the dominant forces of canopy turnover (Ler tzman 1992). Though fires do occur , a s shown by charcoa l ev idence , their role in long-term stand dynamics is minor (Brink 1959; B rooke e r a / . 1970; Kruml ik 1979; K. Ler tzman and L. Brubaker , unpubl . data). E a c h of the three main s p e c i e s common ly live to very old ages with recorded max imums of 750 yea rs for A. amabilis, 1824 yea rs for C. nootkatensis, and >1000 years for T. mertensiana (Pojar and M a c K i n n o n 1994). There have been no addit ions or l osses of tree spec ies in the past 5,000 yea rs (Hansen 1947 in Brink 1959; K. Ler tzman and L. Brubaker , unpubl . data). A base l ine study of un logged ecosys tems in the M H zone w a s publ ished by Brooke er al. (1970) a s a compi lat ion of two P h . D . theses (Peterson 1964; B rooke 1966). It is the only comprehens i ve work deal ing with environment-vegetat ion relat ionships in the M H z o n e , though its study a rea w a s restricted to the southern coast of B .C . ' s main land. Other contr ibutors to our understanding of the M H zone include Brink (1959, 1964), Krumlik (1979), Pojar e r a / . (1991), Kl inka e r a / . (1992+V+), and Ler tzman and his fellow researchers (Ler tzman 1989, 1992, 1995; Ler tzman and Krebs 1991; Ler tzman etal. 1996). 1.2 Regeneration Patterns in the MH Zone T h e regenerat ion pattern of t rees in any forested ecosys tem is the product of many p r o c e s s e s that interact at a variety of sca les . Regenerat ion patterns are affected by s u c h c o a r s e - s c a l e factors a s regional cl imate, e levat ion, aspect , and s lope , but a lso by f ine-sca le or microsi te factors whose inf luence is restricted to an a rea a s smal l as 1-5 m 2 (Ol iver and Larson 1990; S ta thers e r a / . 1990). The mosa ic of microsi tes helps determine the pattern of regenerat ion on a site (Grubb 1977; Harper 1977) and a m o n g the microsi te factors known to contr ibute to this pattern are substrate (Christy and M a c k 1984; Ha rmon and Frankl in 1989); 4 microtopography (Beatty 1984; Pe te rson and Campbe l l 1993), and canopy cover , i.e., c l osed canopy vs . canopy gaps (Piatt and Strong 1989). S n o w , where abundant , is known to compound the impact of microsi tes on regenerat ion patterns (Brink 1959, F o n d a and Bl iss 1969; Brooke et al. 1970; Frankl in era/. 1971 ; Lowery 1972) and there are few forested ecosys tems in the world with more s n o w than the M H z o n e (Brooke et al. 1970; Pojar era/. 1991). F ine-sca le patterns of s n o w accumula t ion and melt are affected by both microtopography (Brooke era/. 1970; Lowery 1972; Beatty 1984) and canopy cover (Gold ing and S w a n s o n 1978, 1986; Hares tad and Bunnel l 1981 ; Berry and Rothwel l 1992). S n o w accumu la tes least and melts earl iest on mounds below canopy trees b e c a u s e of the comb ined effects of canopy interception, black body radiation from the tree bole, and s tem drip (Brooke era/. 1970). The growing s e a s o n near canopy t rees can be >1 month longer than microsi tes only 5 m away (Brooke era/. 1970). In many forests where snow is rare, canopy gaps offer regenerat ion the greatest a c c e s s to light and other resources and so recruitment to the canopy is highest in them. W h e r e s n o w is abundant , however , regenerat ion patterns are inverted and t rees c lump together on e levated is lands s ince only these microsi tes offer a long-enough growing s e a s o n . G a p p r o c e s s e s are known to dominate at e levat ions up to the C W H / M H transit ion (Ler tzman and K rebs 1991; Arsenaul t 1995; Ler tzman et al. 1996) while t ree- is land p r o c e s s e s def ine the park land M H s u b z o n e (Brooke et al. 1970; Frankl in and Dy rness 1988). In the forested M H s u b z o n e , at e levat ions between these two sys tems , it is still unknown whether t rees are more likely to regenerate far from other t rees or c lose to other t rees. T h e regenerat ion of t rees in the M H zone has b e c o m e a press ing managemen t i ssue s ince the introduction of logging in the 1960's. M u c h of the concern has f ocussed on whether low-elevat ion logging pract ices are appropriate at these higher e levat ions, espec ia l ly s ince early regenerat ion prob lems were l inked to s lashburn ing and the planting of s p e c i e s unsui ted to a snowy cl imate (Frankl in 1964; Reuter 1973; Utzig and Herr ing 1974; K l inka and Pend l 5 1976). Fores ters now avoid s lashburn ing and planting and instead rely a lmost exc lus ive ly on natural regenerat ion. In spite of better results, concerns remain that natural regenerat ion is not entirely success fu l s ince it is often c lumped and dominated by one spec ies , A. amabilis (Koppenaa l and Mitchel l 1992; Kl inka e r a / . 1992). There are a lso conce rns about the role of a d v a n c e regenerat ion (understory t rees from the previous forest stand), and how natural regenerat ion is af fected by such microsite factors a s substrate, microtopography, and compet i t ion with Vaccinium spp . (blueberries and huckleberr ies; K l inka et al. 1992). 1.3 Objectives and Thesis Outline My object ive is to advance our understanding of regenerat ion patterns within the M H z o n e by compar ing and contrast ing these patterns in old-growth s tands and c learcuts . M y implicit assumpt ion is that learning how regenerat ion patterns in intact forests change with elevat ion will help guide forest management on si tes throughout the M H z o n e , whether they are to be cut or not. To facilitate compar isons between old-growth s tands and c learcuts , my sampl ing des ign is as similar as poss ib le for the two sys tems . In Chap te r 2, I examine the relationship between microsi tes and regenerat ion patterns in old-growth s tands by address ing two quest ions: (1) Are microsi tes, snow, and regenerat ion patterns re lated?; and (2) Do regenerat ion patterns reflect the gap or t ree- is land mode l? In Chap te r 3, I add ress five quest ions based on concerns ra ised by Kl inka e r a / . (1992): (1) W a s natural regenerat ion success fu l ? ; (2) Did most natural regenerat ion establ ish before or after logg ing?; (3) Wh ich substrates favoured natural regenerat ion?; (4) W h i c h microtopographic locat ions favoured natural regenerat ion?; and (5) Did competi t ion with Vaccinium spp . impede natural regenerat ion? Chap te r 4 summar i zes my f indings and conc lus ions . 6 Chapter 2. Regeneration Patterns on Old-Growth Sites 2.1 Introduction Patterns of s n o w accumulat ion and melt magnify the impact of microsi tes on regenerat ion in subalp ine forests (Brink 1959; F o n d a and Bl iss 1969; Brooke e r a / . 1970; Frankl in etal. 1971; Barbour etal. 1991). F ine-sca le variat ions in microtopography, light levels, and substrate combine with snow to form a cont inuum of microsite condi t ions that affect tree estab l ishment , surv ival , and eventual recruitment into sub-canopy and canopy layers. The forested M H s u b z o n e , the focus of this study, l ies above the Coas ta l Wes te rn Hemlock ( C W H ) z o n e and below the park land M H subzone (Pojar etal. 1991). In the C W H z o n e , where s n o w is infrequent, regenerat ion is primarily through gap dynamics (the gap model) and is f o c u s s e d on canopy gaps c a u s e d by the death of one or many trees (Ler tzman 1992; Arsenau l t 1995; Le r tzman et al. 1996). In contrast, regenerat ion in the very snowy park land M H s u b z o n e fol lows the t ree- is land model and is limited mostly to ra ised tree is lands where snowmel t is earl iest (Brink 1959; Brooke etal. 1970; Frankl in and Dyrness 1988). T h e gap and tree- is land mode ls predict oppos ing patterns. T h e gap mode l predicts that most regenerat ion will be centred in canopy gaps s ince light and other sca rce resources are most avai lab le far from other t rees. The tree- is land model predicts that regenerat ion will be success fu l only near other t rees s ince the main limiting resource is length of growing s e a s o n , not light. Though research has establ ished the preva lence of the gap model at lower e levat ions and the t ree- is land model at higher e levat ions, it has yet to add ress regenerat ion patterns at intermediate e levat ions, i.e., within the forested M H subzone . T o improve our understanding of regenerat ion patterns in the forested M H s u b z o n e , I will examine two general quest ions: (1) Are microsi tes, snow, and regenerat ion patterns re la ted?; and (2) Do regenerat ion patterns reflect the gap or t ree- is land mode l? I begin by descr ib ing the structure and composi t ion of the study s tands as well a s the microsi te 7 envi ronment . I then examine patterns of snow accumulat ion and melt relating to these micros i tes. Final ly, I compare the distribution of regenerat ion and microsi tes. 2.2 Methods Study Area T h e study a rea is on the southwestern edge of Tet rahedron Provinc ia l Park , uphill of Seche l t and - 5 0 km northwest of Vancouve r (49° 35 ' N, 123° 38' W ; Figure 2.1). T h e a rea l ies within the Windward Moist Mari t ime ( M H m m l ) variant of the forested M H s u b z o n e (Green and Kl inka 1994). Loca l topography is formed by rolling hills with rocky r idges, s teep mid -s lopes , and flat val ley bot toms often punctuated by lakes. S n o w typically covers the ground from N o v e m b e r to M a y or J u n e and reaches an average max imum depth of - 2 - 3 m in late Apri l ( B . C . M in . Env . 1985; 1993, 1994, 1995). S n o w p a c k s are very dense , averag ing - 4 0 % snow-water equivalent ( S W E ) in January and February and >50% S W E in late spr ing. S n o w depths vary abruptly due to slight var iat ions in elevat ion, s lope, aspect , and canopy cover and the length of the growing s e a s o n can differ by more than 1 month within 5 m (Brooke e r a / . 1970). Wi th increas ing elevat ion and snow depths, Tsuga mertensiana rep laces Tsuga heterophylla, and this change in spec ies dominance marks the lower boundary of the M H z o n e (Brooke e r a / . 1970). S ince 7". mertensiana is character ist ic of co ld , snowy c l imates whi le T. heterophylla is more likely to be d a m a g e d by snow and frost (Minore 1979; S c a g e l etal. 1989), this transit ion is a lso a proxy for greatly increased snow depths. In addit ion to T. mertensiana, the two other major spec ies in the study a rea are Abies amabilis Doug l . ex Fo rbes (Paci f ic s i lver fir) and Chamaecyparis nootkatensis (D. Don) S p a c h (A laska ye l low-cedar ; B rooke et al. 1970; Pojar etal. 1991). T. heterophylla is only c o m m o n on steeper , south-aspect s l opes and at lower e levat ions. 8 Figure 2.1 Location of Tetrahedron Provincial Park and study area 9 E r i caceous shrubs , especia l ly Vaccinium spp. , m o s s e s , and s o m e herbs form the majority of understory vegetat ion. V . alaskaense (A laska blueberry) is most c o m m o n , whi le V. ovalifolium (oval - leaved blueberry) and V. membranaceum (black huckleberry) are more patchi ly distr ibuted. Other character ist ic e r i caceous shrubs include Rhododendron albiflorum (white-f lowered rhododendron), Menziesia ferruginea (false aza lea) , and Cladothamnus pyroliflorus (copperbush) . M o s s e s can form an a lmost cont inuous carpet on the forest f loor and include Rhytidiopsis robusta (p ipecleaner moss) , Dicranum spp . , Pleurozium schreberi ( red-s temmed feathermoss) , and Rhizomnium glabrescens (fan moss) . Herbs growing on the m o s s layer include Rubus pedatus ( f ive-leaved bramble) and , less common ly , Clintonia uniflora (queen's cup) and Orthilia secunda (one-s ided wintergreen). Blechnum spicant (deer fern) and herbs such a s Tiarella trifoliata ( three- leaved foam flower) and Streptopus spp . (twistedstalk) are abundant on s e e p a g e si tes and a long s t ream edges . S p e c i e s found in late-snowmel t a reas include Sphagnum girgensohnii (white-toothed peat moss ) , Phyllodoce empetriformis (pink mountain-heather), and Luetkea pectinata (partridgefoot). A comple te list of s p e c i e s is inc luded as Append ix A . Study Design R e s e a r c h began in May , 1993. I es tab l ished three study locat ions in old-growth forest s tands (Figure 2.1). E a c h study location cons is ted of one 'flat' site (17 -25% slope) and one s teep site (45-49% slope) for a total of s ix si tes (Figure 2 .2a ; Tab le 2.1). T h e three flat s i tes were actual ly gent ly-s loping and I use the term only to convey that they were flat relative to the three s teep s i tes. T w o study locat ions, L e s s e r and Stee le , were south- facing and one , Edwards , w a s north-facing. Elevat ions above s e a level ranged from 1080-1195 m. Si te se lect ion criteria inc luded: (1) constant south or north aspec t ; (2) s teep site directly above or near flat si te; (3) l imited edaph ic variat ion; and (4) p resence of A. amabilis, C. nootkatensis, and T. mertensiana. 10 Tab le 2 .1 . Old-growth study site descr ipt ions. Site codes combine the first initials of the study locat ion and s lope type, e.g. , E F signif ies Edwards Flat. Soi l moisture reg imes ( S M R ) and soi l nutrient reg imes ( S N R ) follow Kl inka etal. (1989). S M R abbrev iat ions: F = f resh; M = moist; V M = very moist. S N R abbreviat ions: P = poor; M = med ium. S tudy S lope Si te Elevat ion S lope Aspec t S M R S N R locat ion type C o d e (m) (%) (deg. azim.) E d w a r d s Flat E F 1080 17 60 (NE) M / V M P S teep E S 1140 45 64 (NE) M P L e s s e r Flat L F 1195 22 222 (SW) M / V M P / M S teep L S 1195 46 220 (SW) F P S tee le Flat S F 1135 25 211 (SW) M P S teep S S 1160 49 213 (SW) F P I located two 50 m t ransects on each site to avoid large deviat ions in s lope and aspec t , with a horizontal t ransect a long the contour and a vert ical t ransect b isect ing at right ang les (on the s lope line) to form a c ross (Figure 2.2b). I used this c ross - shaped des ign b e c a u s e early tests s h o w e d that it captured more microsite variabil ity than a single 100 m transect. T ransec ts were marked by p lacing a cloth measur ing tape on the ground for the length of the t ransect and locat ing f lags every 5 m. Three types of sampl ing a reas were then es tab l i shed: microsite quadrats, stand structure plots, and age plots. Microsite Quadrats: S ince I wanted to quantify microsite gradients, I s a m p l e d cont iguous 1 m 2 quadrats centred on each site's two 50 m t ransects (total samp l ing a r e a = 0.01 ha per site). A folding 4 m levell ing rod del ineated the quadrats with the cloth tape a s the midpoint. Microsi te sampl ing included all tree and non-tree vegetat ion. Stand Structure Plots: To examine the effect of microsi tes on taller t rees and to determine s tand structure, I needed a larger sampl ing a rea than the microsi te plots and therefore used 10 m by 50 m plots centred on the midpoint of each transect. S i n c e the middle 10 m of the two plots on each site over lapped, the total a rea s a m p l e d w a s 0.09 ha per si te. On ly t rees >1.3 m tall were included in s tand structure sampl ing . 11 Figure 2.2. Si te descr ipt ion and sampl ing des ign . (a) S l ope types for the six plots. O n e flat and one steep site were es tab l ished at e a c h of three study locat ions (Edwards , Lesser , and Steele) . Site codes combine the first initials of the study locat ion and of the s lope type (e.g., E F = Edwards Flat). T h e terms 'flat' and 's teep' are relative and flat s i tes, though more gently s loping than s teep si tes, had s lopes of up to 2 5 % . Range of slopes: for flat sites = 17-25%; for steep slopes = 45-49%. y s Edwards Steep (ES) -^^Le^s^r^te^p^ - ^ o \ v a r t s ^ l a U E F ^ '"'Lesser Flat (LF) '"Steele Flat (SF) (b) T r a n s e c t d i m e n s i o n s a n d layout. O n e ver t ica l a n d o n e hor izonta l t ransec t w e r e l oca ted on e a c h site with 5 0 con t i guous 1m x 1m quad ra t s cen t red on t hem. S t a n d s t ruc ture plots w e r e next cen t red ove r the midd le of the mic ros i te quad ra t s . e n m 49 -• --39 36 35 34 33 32 •• • 29 -1 | 2 | 3 | 4 | S | 6 | 7 | 8 | 9 110| 11112113| 14| 15116117| 18| 19 202122123 1 24 1 25 | 26 j 27 [ 281291 30131 32 1 33 1 34 1 35136 1 37 1 38 1 39 1 40141142143144 1 45146 1 47 1 48149 1 50 :-horizontal transect : 19 -17 16 15 14 13 12 11 10 9 8 7 ~r 5 4 T~ 2 1 — 1m x 1m microsite quadrats 10m x 50m stand structure plots (includes microsite quadrats) vertical transect |< 10 m >| 12 Age Plots: Al l co res were taken from the E S (Edwards Steep) site due to the accessib i l i ty of v igorous A. amabilis and C. nootkatensis. S tumps on an ecological ly-equivalent c learcut study locat ion, Batchelor (Chapter 3), were used for Tsuga and addit ional C. nootkatensis samp les . S a m p l e s for est imating age to breast-height were a lso taken from this c learcut . Data Collected Site Data: E levat ion, s lope, and aspect (in degrees azimuth) were recorded at each site. S n o w depths were measu red (±0.1 m) using a graduated pole at 0.5 m intervals a long each t ransect on two study locat ions (Edwards and Steele) on M a y 15-16, 1994 and all three study locat ions on Apri l 3-5, 1995. Ave rage max imum snow depths were est imated from the height (±0.1 m) of l ichens on the uphill s ide of twenty randomly c h o s e n trees on e a c h site (Brooke etal. 1970; Long 1976). Microsite (Quadrat) Data: I c lassi f ied the canopy cover above each quadrat as fo l lows: c l osed canopy (between two or more canopy trees with the crown of a canopy tree directly overhead) ; canopy gap (open sky directly overhead) ; and expanded gap (between a canopy gap and the bole of a canopy tree; Runk le 1982, 1992; Figure 2.3). T h e relatively open canopy cover of the study a rea made del ineat ing canopy gaps difficult s o I set the min imum a rea at 25 m 2 .1 cons ide red a canopy opening to be fil led, i.e., no longer a gap, when it inc luded a canopy tree (Table 2.2). Th is criterion meant that the height limit of t rees within gaps w a s - 1 5 m and w a s simi lar to other gap studies (Runkle 1981; Y a m a m o t o 1993). M a n y researchers limit their definit ion of gaps to canopy open ings c a u s e d by the death of a branch or at least one tree (Runk le 1992), but s ince my goal w a s to descr ibe the growing condi t ions for individual t rees, I inc luded all gaps that met my other criteria. 13 Figure 2 .3 . C a n o p y cover classi f icat ion (Runkle 1982, 1992). Relat ive canopy openness d e c r e a s e s from canopy gap (eg), to expanded gap (eg), to c losed canopy (cc). On ly canopy gaps are directly below open sky. A n expanded gap is between a canopy gap and the bo les of t rees at the edge of that gap. Remain ing a reas are in c losed canopy . C a n o p y drip microsi tes are directly below the edge of a canopy tree's c rown. e x p a n d e d g a p (eg) c a n o p y g a p (eg) e x p a n d e d g a p (eg) c a n o p y d r i p c a n o p y d r ip Tab le 2 .2 . Definit ion of terms used in Chapter 2. Term Definition Seedling Sapling Tree Understory trees Sub-canopy trees Non-canopy trees Canopy trees Mound partners Any spec ies of tree <10 cm tall. Any spec ies of tree 10-129 cm tall. Any spec ies of tree >1.3 m tall. Shorter trees (most <6 m but all <8 m tall) whose height growth rate w a s usual ly limited due to lack of light; lateral growth of these trees often exceeded height growth (umbrel la growth form). T rees whose heights were intermediate between canopy and understory trees (-6-17 m tall in the study area) . T rees in sub-canopy and understory layers. The tallest trees on a site whose upper crowns were unshaded by other t rees (-15-36 m tall in the study area) . Al l t rees growing on the s a m e mound. 14 For each gap, I recorded its s ize and the number of gapmakers (Runkle 1981 , 1992; Ler tzman 1989), def ined here as dead canopy trees >40 c m base d iameter (the approx imate lower limit of live canopy trees). S ince gap s i zes were est imated from ground level , they were g rouped into 50 m 2 c l a s s e s , plus a 25 m 2 c l ass . The fol lowing data were recorded for g a p m a k e r s : spec ies (if known); condit ion (standing dead , s n a p p e d >2 m above the ground, s tump <2 m tall, or uprooted; Ler tzman and Krebs 1991; Runk le 1992), and d e c a y c l a s s (Table 2.3). T o explore whether non-canopy trees were d a m a g e d by snow s loughing down from canopy t rees, quadrats directly be low the outer projection of a canopy tree's c rown were c lass i f ied as canopy drip (Figure 2.3). Tab le 2 .3 . D e c a y classi f icat ion for dead trees (following M a s e r etal. 1979, T r i ska and C r o m a c k 1979, and Sol l ins 1982). D e c a y c l a s s Structural integrity Texture Most ly intact S o u n d Hard and dry Part ly rotten Hear twood sound , supports own weight Hard , large p ieces Most ly rotten Hear twood rotten, does not support own weight Soft, b locky p ieces Total ly rotten None Soft, powdery when dry I a d d e d two categor ies to more fully descr ibe the understory light envi ronment, low s h a d e and high s h a d e (Figure 2.4), but note that my use of these terms differs from that of Ol iver and Larson (1990). Quadra ts or trees were cons idered to be in high s h a d e if another tree or a north-facing s lope b locked the sun 's rays above - 6 0 ° , the approx imate elevat ion of the sun at s u m m e r solst ice (Brooke etal. 1970). Quadra ts or t rees were cons ide red to be in low s h a d e if the quadrat or tree's leader were shaded by a tree w h o s e b ranches were <5 m above the ground. I chose this 5 m limit b e c a u s e the darkest microsi tes were s h a d e d by non-canopy t rees and the lowest b ranches on most canopy trees were >5 m above the ground. 15 Figure 2.4. Definition of high shade and low shade for understory t rees (<5 m tall) and quadrats . High and low shade were determined in relation to the approximate elevat ion of the sun at summer solst ice ( -60° as shown by diagonal arrows). A n understory tree or quadrat w a s c l assed as being in high shade (HS) if there was a canopy tree or north-aspec t s lope between the sun and the tree's leader or the quadrat sur face. A n understory tree or quadrat was c l assed as being in low s h a d e (LS) if the. b ranches of another tree were between the sun and the understory tree's leader or the quadrat sur face (<5 m above the ground). Note that these definit ions are different than those u s e d by Ol iver and Larson (1990). noHS, noHS, no HS, no HS, HS, HS, north LS no LS LS LS LS LS south I ca tegor ized each quadrat by the fol lowing microtopography c l a s s e s : depress ion micros i tes , s lope microsi tes (i.e., no mound or depress ion) , smal l mounds (<25 c m off the ave rage s lope sur face) , med ium mounds (25-75 cm) , and large mounds (>75 cm) . W h e r e there w a s a mound , I a lso recorded the posit ion of the quadrat relative to it (top, downhi l l , uphil l , or s ide) and its aspec t (flat, east, or west) . Subs t ra tes were samp led by expos ing the top 10 cm of the forest f loor with a garden trowel and est imat ing percent cover to the nearest 5%. At least 3 s a m p l e s were e x p o s e d on e a c h quadrat with more samp les where there w a s more variability. I c lass i f ied subst ra tes a s undis turbed forest floor, friable forest floor, decay ing wood, e x p o s e d decay ing w o o d , coa rse w o o d y debr is , mineral so i l , rock, or tree. Fr iable forest floor had a crumbly texture whi le undis turbed forest f loor w a s more difficult to penetrate with the trowel; these c l a s s e s co r respond to friable Mormoder and compac ted Mor humus forms, respect ively (Green et al. 16 1993). Decay ing wood, which was bright orange, was subdiv ided into two c l a s s e s : e x p o s e d decay ing wood if on the ground sur face, and decay ing wood if below the sur face. C o a r s e woody debr is consisted of logs, b ranches, or s tumps that were above the ground sur face. I recorded the spec ies , percent cover, and height (except for bryophytes and l ichens) of all non-tree vegetat ion. Tree seedl ings were identified by spec ies and c lass i f ied into one of five deve lopmenta l s tages: new germinant, green coty ledons, dead coty ledons, no coty ledons, and lateral ly-branched (Figure 2.5). Due to difficulties in differentiating the two s p e c i e s of Tsuga seed l ings , I grouped them together. F igure 2 .5 . Developmental s tages of seedl ings (after K o h y a m a 1983). new germinant green cot lyedons dead coty ledons no co ty ledons lateral ly-branched g r e e n c o t y l e d o n s , g r e e n co t y l edons , d e a d co t y l edons , no c o t y l e d o n s , no c o t y l e d o n s , no ep ico ty l with ep ico ty l with ep ico ty l wi th ep ico ty l 1 or m o r e b r a n c h e s I recorded the following data for all sapl ings and t rees: s p e c i e s ; canopy layer (Table 2.2); height (using a steel tape measure , a 4 m levell ing rod, or a cloth tape, c l inometer , and tr igonometry); vigour (from 0 - 5, where 0 is dead and 5 is most v igorous; after Lut tmerding et al. 1990 and Carter and Kl inka 1992); s ize of mound (if any) and posit ion of tree relative to it (top, uphil l , downhil l , or edge); and any growth anomal ies such a s pistol butt (snow crook) or umbre l la growth forms (Appendix B) . I recorded substrates only for sap l ings and those understory t rees where the substrate was obv ious. Height increments for the prev ious year were only recorded for sapl ings and trees whose leaders were visible and where the incrementa l growth was distinct. Diameter at breast height w a s recorded for all t rees. 17 Stand Structure Data: The s a m e data were col lected on s tand structure plots a s microsi tes quadrats except that only t rees (i.e., >1.3 m tall) were inc luded, and d is tance to the nearest canopy tree w a s recorded while substrate w a s not. To add ress the composi t ion and structure of individual mounds , I c e n s u s e d all mounds that suppor ted at least one tree and recorded all t rees located on the s a m e mound (all of which I termed mound partners). W h e n mounds ex tended outs ide plot boundar ies, the spec ies and canopy layer of any addit ional mound partners were a lso recorded. Age Data: Obtain ing high quality cores for tree ring ana lys is in old forests can be difficult due to large d iameters (Brooke etal. 1970), frequent heart rot (Burns and Honka la 1990), frost c racks (K. Ler tzman and J . Dobry, pers. comm.) , and asymmet r ic growth rings (Lor imer 1985), espec ia l ly for Tsuga t rees. I therefore used a combinat ion of methods for samp l ing a g e s : cor ing for A. amabilis, count ing s tump rings for Tsuga (not identif ied to spec ies ) , and both methods for C. nootkatensis. I s a m p l e d canopy trees with an increment core fol lowing the methods desc r ibed by J o z s a (1988). To avoid heartrot and miss ing rings, only trees with good vigour were c h o s e n , though this criterion likely b iased the samp le in favour of faster-growing, younger t rees. T o avo id compress ion w o o d , two cores were removed at breast height from opposi te s ides of each tree and at ang les approximately perpendicular to the s lope. C o r e s were mounted on a w o o d e n f rame, s a n d e d to better differentiate rings, then counted using a 40x binocular m ic roscope . On ly the core that w a s c losest to the pith and provided the c learest rings w a s used in further data ana lys is . Tsuga and an addit ional 5 C. nootkatensis were a g e d by count ing s tump r ings. I dec ided to include a smal l samp le of C. nootkatensis s tumps s ince reaching the pith with an increment corer w a s imposs ib le for many of the largest t rees. I cut a v-notch on the s tump from pith to outer edge with a large utility knife and a ruler. The rings then e x p o s e d on the 18 s tump were wetted and/or cha lked to help differentiate rings, and counted with a 10x hand lens. I excava ted 20 dead understory t rees to establ ish approx imate a g e s to s tump and cor ing heights. I cut d isks at the ground sur face and 1.3 m above the root collar, s a n d e d them, and then counted rings using a 40x binocular mic roscope. Seed l ing a g e s were b a s e d on a random samp le of at least 5 seed l ings (where possible) for each deve lopmenta l s tage and s p e c i e s . I cut the seed l ings with a razor b lade at their root collar, b rushed them with chalk and/or water to help differentiate rings, then counted the rings under a 40x binocular m ic roscope . Data Analysis M e a n s were compared with t-tests and ana lys is of var iance using S igmaSta t (Kuo etal. 1987). I first plotted data to sc reen for extreme s k e w n e s s , bi-modality, and unequal var iance , and t ransformed them if necessa ry and appropr iate. The data were then tested for normality (using the Ko lmogoroy -Smi rnov test) and equal var iance (using the Levene med ian test). W h e n they p a s s e d both tests at a = 0.05, I used the intended parametr ic test. W h e n da ta were inappropriate for parametr ic test ing, e.g. , the many zero va lues for height increments and M a y s n o w depths, I used the non-parametr ic equivalent to the t-test (Mann-Whi tney test) or ana lys is of var iance (Kruskal -Wal l is test) or, in extreme c a s e s , s imply descr ibed the da ta without test ing. Mult iple compar i sons after parametr ic and non-parametr ic ana lys is of var iance were per formed using the S tuden t -Neuman-Keu ls (SNK) test only when there were signif icant d i f ferences between samp les (Zar 1984). W h e n sample s i zes were unequal in non-parametr ic multiple compar i sons , I appl ied Dunn 's test. I u s e d ch i -square (x2) ana lys is to test the observed- to-expected f requenc ies of t rees, sap l ings , and seed l ings in relation to microsite factors. Expec ted f requenc ies were b a s e d on the null hypothes is that regenerat ion w a s randomly distributed and its occur rence on a part icular microsi te w a s proport ional to the availabil ity of that microsi te. For examp le , if 19 decay ing wood covered 2 0 % of the ground sur face, we would expect it to support 2 0 % of regenerat ion. Cont ingency tables were used to compare observed f requenc ies on different microsi tes to each other. The Ya tes continuity correct ion w a s appl ied to all 2 x 2 ch i -square and cont ingency table tests to better match the x2 distribution (Zar 1984). To further prevent b ias , I comb ined c l a s s e s s o that no expected counts were less than 1.0 and no more than 2 0 % of expec ted counts were less than 5 (Zar 1984). T o prevent over -emphas iz ing the importance of rare microsi tes, I s h o w both actual f requenc ies and the ratio of observed- to-expected f requenc ies used in ch i -square ana lys is (F igures 2 .13, 2.14, and 2.17). For example , a rare microsite could support twice a s many t rees a s expec ted yet affect regenerat ion patterns only slightly. Before correlat ions were performed in S Y S Y A T for W indows (Wilk inson e r a / . 1992), da ta were tested and t ransformed as descr ibed above . The overal l error rate of multiple correlat ions w a s control led by a Bonferonni correct ion (Glantz 1992). S p e a r m a n rank correlat ions were used when proxy va lues were ass igned to microtopography and canopy cover c l a s s e s (F igures 2.9 and 2.10); these are ordinal va lues based on predict ions of s n o w depth relative to other microsi tes and so are inappropriate for P e a r s o n correlat ions (Glantz 1992). To al low compar i sons with F igures 2.9 and 2.10, I reported S p e a r m a n rank correlat ions in Figure 2.11 (a compar ison between actual d is tance to the nearest canopy tree and s n o w depth) where P e a r s o n correlat ions would otherwise be preferable. M y ana lys is of height growth rates w a s restricted to A. amabilis s ince it w a s the only s p e c i e s with a determinate growth form that a l lowed an accurate measurement of the prev ious year 's height increment, especia l ly for supp ressed t rees. Al l tests were conducted with a = 0.05. The type of t ransformat ion, when u s e d , is • spec i f ied with test results and all data are presented in their original units. Al l m e a n s are p resen ted with s tandard deviat ions, e.g. , 83 ±46 years . 20 2.3 Results Structure, Composition, and Age Most canopy t rees were either T. mertensiana or C. nootkatensis, whi le A. amabilis dominated all shorter height c l a s s e s except seed l ings (Table 2.4 and Figure 2.6a). T h e height distribution of A. amabilis fo l lowed a s teep inverse-J , or negat ive exponent ia l , curve. C. nootkatensis and T. mertensiana had flatter, slightly b imodal distr ibutions result ing from fewer individuals in shorter, and more in taller, height c l a s s e s . T. heterophylla canopy t rees were present on only three si tes (LS , S F , and S S ) and virtually absent a m o n g sub -canopy t rees. D iameter distr ibutions showed similar patterns to height distributions except that C. nootkatensis had a longer tail than other spec ies and T. mertensiana s h o w e d more of an inverse-J s h a p e d curve (Figure 2.6b). C a n o p y trees of each spec ies were taller and had larger d iameters on s teep than flat s i tes (Mann-Whi tney test; T > 6227; p < 0.029). C a n o p y t rees were very old and s low-growing as it took them an ave rage of - 4 8 8 years to reach the canopy layer. Th is min imum est imate inc ludes 8 3 ±46 years to grow to breast height (1.3 m) and a further 405 ± 1 1 2 years to grow to a d iameter of 40 c m (the lower limit of most canopy trees). Est imated ages ranged from 370-901 years for A. amabilis (n = 22), 395 -1404 years for C. nootkatensis (n = 22), and 319-897 for Tsuga (n = 12). S o m e t rees grew faster than others, but the overal l trend w a s constant, s low growth (Figure 2.7). Th is trend cont rasted with that expec ted with the gap mode l , where understory and sub -canopy t rees respond to the formation of a new gap by growing much faster and maintaining that faster growth rate a s they grow into the canopy layer. Of the three spec ies , only Tsuga tended to grow faster with inc reased s i ze , especia l ly at d iameters >10 c m . 21 Table 2.4. Spec ies composition (frequency per hectare) by height c lass and site. Tsuga seedlings were not identified to species . Height Class Site Species E F E S LF L S S F S S Mean (%) Canopy trees A. amabilis 67 33 111 78 33 33 59 (20.6) C. nootkatensis 133 67 111 122 44 78 93 (32.3) T. mertensiana 111 78 133 111 133 122 115 (40.0) T. heterophylla 0 0 0 33 44 44 20 (7.1) Total 311 178 356 344 256 278 287 (100.0) Sub-canopy trees A. amabilis 44 89 156 67 33 144 89 (47.1) C. nootkatensis 44 22 78 0 0 11 26 (13.7) T. mertensiana 67 33 44 133 67 67 69 (36.3) T. heterophylla 11 0 0 11 11 0 6 (2.9) Total 167 144 278 211 111 222 189 (100.0) Understory trees A. amabilis 200 511 1078 1000 444 944 696 (58.3) C. nootkatensis 456 378 433 78 0 0 224 (18.8) T. mertensiana 422 322 356 233 67 78 246 (20.6) T. heterophylla 11 0 11 11 78 56 28 (2.3) Total 1089 1211 1878 1322 589 1078 1194 (100.0) Saplings A. amabilis 4000 4600 2400 5700 15400 8100 6700 (79.0) C. nootkatensis 600 900 400 700 1700 200 750 (8.8) T. mertensiana 1600 200 100 0 2200 100 700 (8.3) T. heterophylla 0 300 0 0 1700 0 333 (3.9) Total 6200 6000 2900 6400 21000 8400 8483 (100.0) Seedlings A. amabilis 25400 35300 38700 119200 116300 38900 62300 (33.4) C. nootkatensis 29400 45000 117400 101200 71100 74000 73017 (39.1) Tsuga 78600 32700 71700 30600 48700 45600 51317 (27.5) Total 133400 113000 227800 251000 236100 158500 186633 (100.0) 22 Figure 2.6. Height (a) and diameter (b) distributions of t rees >1.3 m tall by spec ies . T h e shortest height c l ass inc ludes trees from 1.3-3.9 m tall. T h e y-ax is sca le d o e s not show full f requenc ies for A. amabilis in the smal lest height and d iameter c l a s s e s ; actual f requenc ies are included at the top of the histogram bar. D iameters were m e a s u r e d at breast height. A. amabilis T. mertensiana upper limit of height class (m) 250 C. nootkatensis T. heterophylla upper limit of height class (m) (a) Height distribution by spec ies . (b) Diameter (at breast height) distribution by spec ies. 2 3 Figure 2.7. Growth rate to a diameter of 40 c m (at s tump or core height). S i n c e the d iameter at breast-height of most canopy trees w a s >40 c m , these charts reflect growth rates for understory and sub-canopy trees (each line represents one tree). A g e s do not include the number of years it took trees to reach breast height. Note that flatter s l opes indicate faster growth rates. / 24 There were many seed l ings of each spec ies , but A. amabilis w a s most likely to surv ive to the lateral ly-branched s tage: there were 16 lateral ly-branched A. amabilis seed l ings for every 2 C. nootkatensis and one Tsuga. C. nootkatensis w a s most c o m m o n a m o n g new germinants and seed l ings with green coty ledons. Tsuga w a s most c o m m o n a m o n g seed l ings with no coty ledons, though this result may the difficulty in differentiating s tages for Tsuga. A. amabilis were strangely rare a m o n g new germinants as only 9 were found on all quadrats (= 150/ha) . Seed l i ng ages for each spec ies inc reased with their s tage of deve lopment (two-way ana lys is of var iance on natural log-transformed data; F > 17.1; p < 0.0001), and A. amabilis were o ldest at e a c h s tage. Latera l ly-branched seed l ings averaged 6.4 ±1.6 c m in height and 12.4 ±7 .4 years of age , and the oldest w a s a 27 year-o ld A. amabilis. T h e annua l height growth of seed l ings and sap l ings, particularly A. amabilis, w a s offset by burying of the lower s tem through soi l c reep. For example , one A. amabilis sapl ing that w a s 60 c m tall had >60 growth rings bur ied below the ground sur face. The Microsite Environment Substrates Undis turbed forest floor covered 8 0 % of the ground sur face and decay ing w o o d or coa rse woody debr is covered most of the rest (Table 2.5). Most coa rse woody debr is cons is ted of the s tems of fal len trees. Mineral soil w a s rare except where run-off f rom a new logging road dra ined onto the E F site. Though not quanti f ied, most undisturbed forest floor w a s very compac ted and cons is ted of a greasy, very acid ic humus with abundant fungal hyphae (Hemimor and Humimor humus forms, G r e e n etal. 1993). M o s s covered 4 2 % of the ground. 25 Tab le 2 .5 . Percent cover of substrates by site. Abbrev iat ions: uff = undisturbed forest floor; fff = friable forest floor; dw = decay ing wood ; edw = exposed decay ing w o o d ; c w d = coa rse woody debr is ; ms = mineral soi l . substrate Si te uff fff dw edw cwd ms rock tree Total E F 76 .7 0.2 5.6 1.9 9.6 2.4 0.5 3.1 100.0 E S 88.9 0.0 0.3 3.1 3.1 0.0 2.2 2.6 100.0 L F 75 .5 0.2 9.2 0.3 11.0 0.7 0.4 2.8 100.0. L S 80 .5 0.0 9.5 0.0 4.6 0.0 1.6 4.0 100.0 S F 83.2 0.4 9.2 0.2 3.9 0.2 0.0 3.1 100.0 S S 75 .3 0.1 8.8 0.0 10.8 0.0 0.0 5.2 100.0 M e a n 80 .0 0.2 7.1 0.9 7.1 0.6 0.8 3.4 100.0 Microtopography M o u n d s covered 2 8 % of the ground with a lmost the s a m e proport ion on flat and s teep s i tes. They were a lmost equal ly divided among smal l , med ium, and large s i z e s . S l ope micros i tes were the most c o m m o n and covered 7 1 % of s teep and 4 7 % of flat s i tes. Mos t of the di f ference between site types w a s due to the p resence of more depress ion microsi tes on flat (23%) c o m p a r e d to s teep (3%) s i tes. Canopy Cover and Distance to the Nearest Canopy Tree O n average , 3 5 % of quadrats were in canopy gap, 3 8 % were in expanded gap , and 2 7 % were in c losed canopy (Table 2.6). But these ave rages mask the great var iat ion in the "gapp iness " of individual s i tes s ince the proportion of quadrats in c losed canopy ranged from 1 1 % on the E F site to 4 4 % on the S S site. Cons is tent with their defining character is t ic , all c l osed canopy quadrats were in high shade but there w a s a lso high s h a d e on 4 1 % of e x p a n d e d gap and 3 4 % of canopy gap quadrats. There w a s low s h a d e on 3 4 % of quadrats , with the lowest proportion on c losed canopy quadrats. Quadra ts in c losed canopy were most likely to be heavi ly s h a d e d (both high and low shade) . Quadra ts in canopy gaps were most l ikely to be unshaded (no high or low shade) , though the majority (53%) had s o m e sort of 26 shad ing . T h e main conc lus ion to draw from these results is that canopy cover a lone d o e s not fully descr ibe the light environment of understory t rees. G a p s i z e s ave raged 99 ±106 m 2 in a distribution that w a s s k e w e d to lower va lues . A l though the med ian gap s ize w a s 50 m 2 on both flat and s teep s i tes, m e a n s were higher on the flat site of each study location ( E F > L F > S F > E S > L S > S S ; s k e w e d distr ibutions and low samp le s i z e s prevented testing), due to the few, but very large gaps on these s i tes. E a c h flat si te, but no s teep site, had one gap with an a rea >400 m 2 . The proport ion of gaps that had no gapmake r fo l lowed the s a m e ranking as mean gap s ize and ranged from 1 0 0 % on the north-facing E F site to 5 0 % on the south-facing S S site. Tab le 2.6. C a n o p y cover (a) by site; and (b) by high s h a d e and low s h a d e (in percent) . S e e Figure 2.3 for definit ions of canopy cover c l a s s e s and Figure 2.4 for definit ions of high and low shade . Abbrev iat ions: cc = c losed canopy; eg - expanded gap; eg = canopy gap . Canopy cover c lass Total High/low Canopy cover c lass Total Site cc eg eg shade cc eg eg E F 11.0 45.0 44.0 100.0 H S 26.5 15.5 12.0 54.0 E S 26.0 36.0 38.0 100.0 no H S 0.3 22.3 23.0 45.7 LF 24.0 34.0 42.0 100.0 L S 7.8 13.5 13.0 34.3 L S 23.0 44.0 33.0 100.0 no L S 19.2 24.5 22.0 65.7 S F 34.0 40.0 26.0 100.0 H S , L S 7.7 4.0 5.0 16.7 S S 44.0 29.0 27.0 100.0 H S , no L S 18.8 11.5 7.0 37.3 Flat 23.0 39.7 37.3 100.0 no H S , L S 0.2 9.5 8.0 17.7 Steep 31.0 36.3 32.7 100.0 no H S , no L S 0.2 12.8 15.0 28.0 Total 27.0 38.0 35.0 100.0 Total 27.0 38.0 35.0 100.0 (a) By site (b) By high shade (HS) and low shade (LS). G a p terminology impl ies a relationship between canopy cover and d is tance to the nearest canopy tree s o that c losed canopy microsi tes should be c losest to, and canopy gap micros i tes farthest f rom, the nearest canopy tree. A s expec ted , most quadrats in c l osed canopy were <2 m from a canopy tree; most in expanded gap were 2-3.9 m from a canopy tree; and most in canopy gap were >4 m from a canopy tree (Figure 2.8). T h e main over lap in this c lassi f icat ion w a s c a u s e d by the high proportion of quadrats in expanded gap that were <2 27 m from a canopy tree. C a n o p y drip w a s present on 2 4 % of quadrats and most (68%) canopy drip quadrats were 3-4.9 m from the nearest canopy tree. F igure 2.8. Quadra ts by canopy cover c lass and dis tance to the nearest canopy tree (in percent) . 40 -I closed canopy 30 • —•— quadrats (°/ 20 quadrats (°/ 10 expanded gap n canopy gap • i —• "~" ~~ • U '1 <1 1.0-1.9 2.0-2.9 3.0-3.9 4.0-4.9 distance to the nearest canopy tree (m) 5+ Snow - Microsite Relationships Apri l s n o w ave raged 1.7 ±0.5 m in depth, with a max imum of 3.3 m, and it d e c r e a s e d in m e a n depth by study location as fol lows: L e s s e r > Edwards > Stee le (Table 2.7). The re w a s comple te s n o w cover in Apri l and almost identical depths on flat and s teep si tes of e a c h study locat ion. Major di f ferences only b e c a m e apparent as snow mel ted. By mid -May , s teep and south- fac ing s i tes had melted more than flat and north-facing s i tes and the proport ion of bare ground ranged from 8 7 % on the S S site to 1 6 % on the E F site. T h e s e site rankings were simi lar to those for canopy cover , mean gap s ize , and proportion of canopy gaps with no gapmake r (previous sect ion). The average max imum snow depth, est imated from l ichen height, w a s 2.9 ±0 .3 m, - 7 0 % higher than mean Apri l snow depths. 28 Tab le 2.7. S n o w depths (in metres) by site and s lope type. M a y s n o w depths were not m e a s u r e d on the L F and L S study s i tes. Site Lichen height (m) April 3-5, 1995 May 15-16, 1994 Mean (s.d.) Range Mean (s.d.) Range Mean (s.d.) Range Bare (%) E F 2.9 (0.2) 2.5-3.2 1.6 (0.4) 0.5-2.7 0.4 (0.4) 0.0-1.5 16.0 E S 3.2 (0.4) 2.7-3.8 1.8 (0.4) 0.6-2.4 0.2 (0.2) 0.0-0.7 41.6 LF 3.2 (0.3) 2.6-3.6 2.1 (0.5) 0.5-3.3 Not sampled L S 3.0 (0.3) 2.5-3.6 1.9 (0.4) 0.4-3.0 Not sampled S F 2.3 (0.2) 1.9-2.6 1.4 (0.4) 0.1-2.3 0.1 (0.2) 0.0-0.7 56.4 S S 2.8 (0.3) 2.4-3.2 1.4 (0.3) 0.2-2.2 0.0 (0.1) 0.0-0.6 87.1 Flat 2.8 (0.2). 1.9-3.6 1.7 (0.5) 0.1-3.3 0.3 (0.3) 0.0-1.5 36.4 Steep 3.0 (0.3) 2.4-3.8 1.7 (0.4) 0.2-3.0 0.1 (0.2) 0.0-0.7 64.4 Total 2.9 (0.3) 1.9-3.8 1.7 (0.5) 0.1-3.3 0.2 (0.3) 0.0-1.5 50.4 Three factors were expec ted to affect snow depths: microtopography, canopy cover , and d is tance to the nearest canopy tree. A s expec ted , Apri l and M a y snow packs were general ly deepe r on non-mound microsi tes, in canopy gaps , and farther from canopy t rees (analys is of var iance and Kruska l -Wal l is tests; p < 0.0001 for all tests). T o further examine the co r respondence between snow depths and these microsite factors on individual samp l ing points, I plotted snow depths a long each transect (Figures 2.9 to 2.11). E a c h of these three f igures cons is ts of a pair of charts for each of eight t ransects from the E d w a r d s and Stee le study locat ions (the L e s s e r study location is not inc luded b e c a u s e M a y s n o w depths were not recorded). The upper chart in each pair plots Apri l (solid line) and M a y (dashed line) snow depths. The lower chart plots proxy va lues for microtopography c l ass (Figure 2.9), proxy va lues for canopy cover (Figure 2.10), and actual va lues for d is tance to the nearest canopy tree (Figure 2.11). The proxy va lues in these first two f igures reflect expec ted snow depths s o that 1 = shal lowest and 3 = deepest . Proxy va lues for microtopographic c l a s s e s in Figure 2.9 are: mounds = 1, s lope microsi tes = 2, and depress ion micros i tes = 3. Proxy va lues for canopy cover c l a s s e s in Figure 2.10 are: c losed canopy = 1; expanded gap = 2; and canopy gap = 3. Pat terns of the lower line can be compared to Apri l and M a y s n o w depths to determine the co r respondence between expec ted and actual snow depths . 29 S p e a r m a n correlat ion va lues (r s ) between microsite factors and Apri l s n o w depths are inc luded with e a c h transect. S n o w depths showed strikingly similar trends in May 1994 and Apri l 1995 in spite of being m e a s u r e d in different years . The co r respondence w a s c learest on the site with the deepes t s n o w in May , E F (Pearson correlat ions for the two t ransects: r = 0.94 and 0.84; F igures 2.9 to 2.11), but microsi tes with the least snow in Apri l mel ted first on all t ransects . T h e three microsi te factors were, however, imperfectly related to snow depths on individual sampl ing points. The poorest predictor of Apri l snow depths on individual sampl ing points w a s microtopography (Figure 2.9) while d is tance to the nearest canopy tree w a s best (Figure 2.11), espec ia l ly in the near perfect relationship between microsi tes c lose to a canopy tree and early snowmel t . The relationship between the three factors and snow depths w a s posit ive on all except one transect, S F V , where there w a s a weakly negat ive relat ionship be tween "gapp iness" and snow depths ( r s = -0.44, p < 0.002). O n e of the poss ib le reasons for the relatively low co r respondence between microsi te factors and s n o w depths on individual sampl ing points (Figures 2.9 to 2.11) w a s that these f igures did not take into account the comb ined effect of more than one microsi te factor. The proport ion of microsi tes that were bare of snow in May il lustrates this comb ined effect. Us ing the three microsi te factors individually, the fol lowing proportion of microsi tes had mel ted out by M a y : 6 3 % of mounds compared to 2 3 % of depress ion microsi tes; 6 5 % in c losed canopy c o m p a r e d to 4 1 % in canopy gap; and 6 7 % that were <2 m from the nearest canopy tree c o m p a r e d to 3 8 % that were >4 m from the nearest canopy tree. Comb in ing these factors inc reased di f ferences in melt ing patterns, e.g. , 8 3 % of mounds in c losed canopy and <2 m from the nearest canopy tree were bare compared to 8 % of depress ion microsi tes in canopy gap that were >4 m from the nearest canopy tree. 30 F igures 2.9. S n o w depths compared to microtopography. E a c h t ransect cons is ts of two charts. The upper chart plots Apri l (upper, sol id line) and M a y (lower, dashed line) snow depths a long the transect. The lower chart appl ies proxy va lues by microtopography: mounds (mound) = 1; s lope microsi tes (slope) = 2; and depress ion microsi tes (dep.) = 3. Note that each site has two transects, one horizontal (e.g., E F H ) and one vert ical (e.g., E F V ) . S p e a r m a n rank correlation va lues (r s) for Apri l snow depths and microtopography proxy va lues are included at the lower right of each transect pair. A n aster isk indicates p < 0.05 using a Bonferonni correct ion. dep.=3 slope=2 mound=1 a) E F H dep.=3 slope=2 mound=1 c) E S H dep.=3 slope=2 mound=1 300 e) S F H 200 dep.=3 slope=2 mound=1 30 10 50 0 10 r s (Apri l ) = 0.21 b) E F V r s (Apr i l ) = 0.21 30 40 50 0 10 r s (Apr i l ) = 0.24 d) E S V r s (Apr i l ) = 0.29 r s (Apr i l ) = 0.28 f)SFV r s (Apr i l ) = 0.02 g) S S H r s (Apr i l ) = 0.48* t ransect d is tance (m) h) S S V r s (Apr i l ) = 0.06 t ransect d is tance (m) 31 F igures 2 .10. S n o w depths compared to canopy cover . E a c h transect cons is ts of two charts. T h e upper chart plots Apri l (upper, sol id line) and May (lower, dashed line) snow depths a long the transect. The lower chart appl ies proxy va lues by canopy cover: c losed canopy (cc) = 1; expanded gap (eg) = 2; and canopy gap (eg) = 3. Note that each site has two transects, one horizontal (e.g., E F H ) and one vertical (e.g., E F V ) . S p e a r m a n rank correlation va lues (r s) for Apri l snow depths and canopy cover proxy va lues are inc luded at the lower right of each transect pair. A n aster isk indicates p < 0.05 using a Bonferonni correct ion. 300 1 200 £ •8 100 -1 1 1 1 1 , L^ i— a) E F H e) S F H r s (Apr i l ) = 0.57* b) E F V r s (Apri l ) = 0.23 r s (Apr i l ) = -0 .07 f) S F V r s (Apr i l ) = -0.44* g) S S H r s (Apr i l ) = 0.36 t ransect d is tance (m) h ) S S V r s (Apr i l ) = 0.14 t ransect d is tance (m) 32 F igures 2 .11 . S n o w depths compared to d is tance to the nearest canopy tree. E a c h transect cons is ts of two charts. The upper chart plots Apri l (upper, sol id line) and May (lower, d a s h e d line) snow depths a long the transect. The lower chart appl ies actual d is tances to the nearest canopy tree (in m). Note that each site has two transects, one horizontal (e.g. , E F H ) and one vert ical (e.g., E F V ) . S p e a r m a n rank correlation va lues (r s) for Apri l s n o w depths and d is tance tb the nearest canopy tree are included at the lower right of e a c h t ransect pair. A n aster isk indicates p < 0.05 using a Bonferonni correct ion. 300 0 10 20 30 40 50 0 10 20 30 40 50 a) E F H r s (Apr i l ) = 0 .71* b) E F V r s (Apr i l ) = 0.35* 300 0 10 20 30 40 50 0 10 20 30 40 50 c) E S H r s (Apr i l ) = 0.32* d) E S V r s (Apr i l ) = 0.56* 300 e ) S F H r s (Apr i l ) = 0 .35* f) S F V r s (Apr i l ) = -0.23 300 i , 1 1 1 1 | 1 1 ' ' g ) S S H r s (Apr i l ) = 0 .43* h) S S V r s (Apr i l ) = 0.29 t ransect d is tance (m) t ransect d is tance (m) 33 It is a lso difficult to use a s imple correlat ion to determine the strength of the relat ionship be tween microsi te factors and snow depths s ince smal l shifts between var iab les reduce the r s va lue. For examp le , peaks and troughs in snow depths and d is tance to the nearest canopy tree shown for the E F V and E S H transects were likely related more strongly than shown statist ical ly ( r s = 0.35 and 0.32, respectively) b e c a u s e they were slightly, but inconsistent ly, offset (Figure 2.11). There w a s no consistent east -west or uphil l-downhil l shift a s wou ld be expec ted if factors such as aspec t or the direction of prevai l ing winds determined s n o w patterns at this sca le . G a p s i ze w a s related to snow depths only on the north-facing study locat ion, Edwards , where the med ian Apri l depth in gaps >50 m 2 (the median size) w a s 1.8 m compa red to 1.4 m in smal le r gaps , and median M a y depths were 0.2 m and 0.0 m, respect ively (lack of repl icat ion prevented testing). Med ian depths on the south- facing study locat ion, S tee le , were the s a m e regard less of gap s ize in both Apri l (1.4 m) and M a y (0.0 m). T h e s e t rends were simi lar on both flat and s teep si tes within each study locat ion. S n o w on canopy drip quadrats w a s not apprec iab ly deeper in either Apri l or May (Mann-Whi tney test; T < 49519; p > 0.11). Tree-Microsite Relationships Substrates Undis turbed forest floor suppor ted more seed l ings , sap l ings, and understory t rees than expec ted (x2 > 12.0; p < 0.001; Tab le 2.8). Surv ival w a s apparent ly higher on undisturbed forest f loor than other substrates as it accounted for 80% of ground cover , but suppor ted 82% of seed l ings , 92% of sap l ings, and 99% of understory t rees. Most other regenerat ion w a s cent red on decay ing wood which suppor ted 9% of seed l ings but only 6% of sap l ings and 1% of understory t rees. Decay ing wood suppor ted 20% of T. mertensiana sap l ings but no T. 34 mertensiana understory t rees. C o a r s e woody debris suppor ted 7 % of all seed l ings , most ly new germinants , but no regenerat ion >30 cm tall. There w a s little or no relationship between m o s s cover and seed l ings or sap l ings . T h e highest correlat ion va lue of nine tests w a s that for m o s s cover and A. amabilis seed l ings , yet even this relat ionship w a s weak ( r s = 0.37; p < 0 .0001; Bonferonni correct ion appl ied) . Tab le 2.8. Root ing substrate of understory t rees, sap l ings, and seed l ings (in percent) . Tsuga seed l ings were not identified to spec ies . The samp le s ize of understory t rees w a s limited by the difficulty in identifying substrates for taller t rees. S e e Tab le 2.5 for a key to abbreviat ions. Height c l a s s Subst rate S p e c i e s n uff fff dw edw cwd ms Tota l Unders tory t rees A. amabilis 42 97.6 0.0 2.4 0.0 0.0 0.0 100.0 C. nootkatensis 12 100.0 0.0 0.0 0.0 0.0 0.0 100.0 T. mertensiana 17 100.0 0.0 0.0 0.0 0.0 0.0 100.0 T. heterophylla 3 100.0 0.0 0.0 0.0 0.0 0.0 100.0 Total 74 98.6 0.0 1.4 0.0 0.0 0.0 100.0 Sap l i ngs A. amabilis 402 95.3 0.0 3.7 0.2 0.5 0.2 100.0 C. nootkatensis 45 82.2 2.2 8.9 6.7 0.0 0.0 100.0 T. mertensiana 40 70 .0 10.0 20.0 0.0 0.0 0.0 100.0 T. heterophylla 20 85.0 0.0 10.0 0.0 0.0 5.0 100.0 Total 507 91.7 1.0 5.7 0.8 0.4 0.4 100.0 S e e d l i n g s A. amabilis 3738 88.0 0.0 8.1 0.7 3.1 0.0 100.0 C. nootkatensis 4381 81.8 0.3 8.6 1.2 7.7 0.4 100.0 Tsuga 3079 74.0 0.3 12.2 1.4 11.4 0.7 100.0 Total 11198 81.7 0.2 9.4 1.1 7.2 0.4 100.0 Microtopography and Mound Partners T r e e s were disproport ionately c o m m o n on mounds , especia l ly med ium-s i zed mounds , and this pattern b e c a m e more pronounced among taller t rees (Figure 2.12). Of five height c l a s s e s , only seed l ings were not over- represented on mounds (56.70 < x2 < 486 .43 ; p < 35 0.001). M o u n d s occup ied only 2 8 % of the ground sur face but suppor ted 4 1 % of understory, 6 1 % of sub -canopy , and 9 7 % of canopy t rees. For each sub-canopy tree, there were 2088 seed l ings on non-mound microsi tes compared to only 328 seed l ings on mounds . Both of these results provide ev idence that t rees were most likely to survive on mounds . C. nootkatensis w a s most l ikely and T. mertensiana w a s least likely to be located on mounds . T h e proport ion of sub -canopy trees on mounds w a s highest on the late-snowmelt s i tes, E F (87%) and E S (77%) and lowest on the two ear ly-snowmelt s i tes, S F (60%) and S S (53%). Figure 2.12. Observed- to -expec ted f requency of t rees by microtopography. T h e x-ax is c r o s s e s the y-ax is at an observed- to-expected ratio of 1.0, i.e., where the f requency of t rees w a s equal to what w a s expec ted . 5 T Non-canopy t rees on mounds were more c o m m o n on the downhil l than uphill s ide , espec ia l ly a m o n g sub -canopy trees and on s teep si tes. O n s teep s i tes, 4 3 % of t rees were downhil l compa red to 8 % uphill; on flat s i tes, 2 5 % of t rees were downhil l compa red to 1 3 % uphil l . T h e apparent survival of trees w a s greater on the downhil l s ide of mounds a s there w a s a higher proport ion of sub -canopy than understory t rees located on these microsi tes on both s lope types. Non -canopy trees were more likely to be on the top of mounds on flat (21%) than 36 s teep (11%) s i tes. There w a s no relationship between mound aspec t (i.e., east - c o m p a r e d to west- facing) and tree locat ion. A total of 240 mounds that suppor ted at least one tree were c e n s u s e d . T h e m e a n number of mound partners (trees on the s a m e mound) by mound s ize was : 1.4 ±0 .6 (small), 2.8 ±2 .2 (medium), and 4.3 ±2.6 (large). At least one live canopy tree occup ied 128 of s a m p l e d mounds but only four mounds suppor ted >3 canopy t rees. More than one canopy tree of the s a m e spec ies on a mound w a s uncommon except for C. nootkatensis. The re w a s no ev idence of la rge-sca le mortality on mounds as 1 5 % of mounds inc luded one d e a d canopy tree, but none had more. There w a s at least one live tree ( regardless of canopy layer) on 9 2 % of mounds and at least one dead tree on 2 3 % of mounds . M o u n d s occup ied by live C. nootkatensis canopy trees suppor ted more mound partners (4.8 ±3 .3 ; n = 46) than those occup ied by A. amabilis (3.4 ±2.9 ; n = 29), T. heterophylla (3.1 ±2 .0 ; n = 12), or T. mertensiana (2.6 ±2.3 ; n = 60; Tab le 2.9). C. nootkatensis w a s a lso the most c o m m o n canopy tree shar ing a mound with one or more other canopy tree, whi le A. amabilis w a s the sub -canopy and understory spec ies most c o m m o n on canopy tree mounds . T. mertensiana w a s most likely and C. nootkatensis w a s least likely to be the only canopy tree on a mound . S u b - c a n o p y trees were present on 2 4 % of all c e n s u s e d mounds and 2 5 % of mounds with at least one live or dead canopy tree. T h e s e results may indicate that mounds are a more important advantage to regenerat ion than the p resence of a canopy tree, espec ia l ly for T. mertensiana. Canopy Cover and Distance to the Nearest Canopy Tree Regenera t ion patterns were affected by proximity to the nearest canopy but not ove rhead canopy cover , especia l ly among sub-canopy trees and on flat s i tes (Figure 2.13). With increas ing height, regenerat ion w a s increasingly likely to be on microsi tes that were <2 m from the nearest canopy tree than farther away. There were fewer seed l ings and sap l ings than 37 expec ted on these c loser microsi tes (x2 > 25 .82 ; p < 0.01), but 1 4 % more understory t rees and 5 6 % more sub -canopy trees than expected (x2 > 7.86; p < 0.025). C a n o p y cover , in contrast, w a s statist ically unrelated to the location of understory or sub -canopy trees (x2 < 4 .60 ; p > 0.10). Sti l l , there were many sub-canopy trees in canopy gaps (Figure 2.13a), espec ia l ly T. mertensiana on s teep, south- facing s i tes, which indicated that regenerat ion w a s not restr icted only to tree is lands. Tab le 2.9. Live mound partners of live and dead canopy t rees. On ly mounds that conta ined at least one tree are l isted (= 5 3 % of all mounds sampled) . C o l u m n s list the number of live and d e a d canopy trees occupy ing a mound . R o w s list the number of live mound partners occupy ing the s a m e mound. Abbreviat ions: Aa = Abies amabilis; Cn -Chamaecyparis nootkatensis; Tm = Tsuga mertensiana; Th = Tsuga heterophylla; unk. = unknown. C a n o p y layer Spec ies of live canopy tree Spec ies of dead canopy tree Mound partner Aa Cn Tm Th Total Aa Cn Tsuqa unk. Total Live canopy trees None 20 27 50 8 105 1 5 5 11 22 Aa 1 7 2 0 10 0 1 0 4 5 Cn 6 7 8 2 23 0 1 3 3 7 Tm 2 8 2 2 14 0 1 0 0 1 Th 0 2 2 1 5 0 0 0 3 3 Live sub-canopy trees None 21 22 49 10 102 1 6 7 16 30 Aa 5 15 6 1 27 0 2 0 2 4 Cn 1 3 2 0 6 0 0 0 2 2 Tm 2 7 4 1 14 0 0 1 0 1 Th 0 0 0 1 1 0 0 0 0 0 Live understory trees None 15 12 30 6 63 0 3 5 12 20 Aa 10 30 26 4 70 1 5 3 6 15 Cn 3 6 4 0 13 0 3 0 1 4 Tm 8 14 3 3 28 0 1 0 4 5 Th 2 1 0 1 4 0 1 0 3 4 Total no. of mounds 29 46 60 12 147 1 8 8 20 37 38 Figure 2 .13 . F requency and observed- to-expected f requency of non-canopy t rees by s lope type. Lef t -hand f igures (a, c) present results by canopy cover and r ight-hand f igures (b, d) present results by d is tance to the nearest canopy tree. Upper f igures (a, b) present actual f requenc ies per hectare and lower f igures (c, d) present ratios of obse rved f requenc ies to those expec ted by the availabil ity of each canopy cover or d is tance c l a s s . Symbo l s : understory t rees = c i rc les; sub -canopy trees = squa res ; flat s i tes = open symbo ls ; s teep si tes = sol id symbo ls . Note that there are two s c a l e s on the y-ax is sca le in F igures (a) and (b). They are presented this way to highlight t rends a m o n g sub -canopy trees and to al low compar isons of s lopes between sub -canopy and understory t rees: a similar s lope means that the proport ional (not absolute) change in f requency is similar. canopy cover class distance to the nearest canopy tree (a) Frequency by canopy cover (b) Frequency by distance to the nearest canopy tree (c) Observed-to-expected ratio by canopy cover (d) Observed-to-expected ratio by distance to the nearest canopy tree 39 Figure 2.14. Frequency and observed-to-expected frequency of sub-canopy trees by canopy cover, distance to the nearest canopy tree, and slope. Upper figures (a) present actual frequencies per hectare and lower figures (b) present ratios of observed frequencies to those expected by the availability of each combination of canopy cover and distance class. • flat sites; • steep sites. expanded gap 60 40 20 canopy g a p . <2m 2-3.9m 4+m <2m 2-3.9m 4+m distance to the nearest canopy tree <2m 2-3.9m 4+m (a) Frequency per hectare < 2 m 2-3.9 m 4+ m <2 m 2-3.9 m 4+m <2 m 2-3.9 m 4+m distance to the nearest canopy tree Most sub-canopy trees in closed canopy and expanded gap were <2 m from a canopy tree while most in canopy gap were >4 m from a canopy tree (Figure 2.14). On their own, these results are not surprising since most closed canopy microsites were close to a canopy tree and most canopy gap microsites were far from a canopy tree (Figure 2.8). What was not anticipated, however, was that microsites <2 m from a canopy tree supported more sub-canopy trees than expected, regardless of canopy cover (x2 = 13.88; p < 0.001). This trend was strongest on flat sites, where observed-to-expected ratios declined consistently with 40 increasing distance from a canopy tree. The only exception was in canopy gaps on steep sites where both actual frequencies and observed-to-expected ratios were higher >4 m from a canopy tree than 2-3.9 m from a canopy tree; most of these trees (82%) were T. mertensiana. These results provide evidence that tree-island processes predominate on flat sites while both gap and tree-island processes are present on steep sites. In stands where both processes are active, regeneration might be expected to be most successful on microsites close to a canopy tree and also in canopy gap, and least successful on microsites far from a canopy tree but also in closed canopy. Though this expectation was supported by the data (Figure 2.14), the rarity of these microsites (7 of each, out of a total of 600 quadrats) limited their possible impact on regeneration. Apparent survival, expressed as the ratio of sub-canopy to understory trees, was greatest for all species on closed canopy microsites on both flat and steep sites. T. mertensiana was the species most likely, and C. nootkatensis least likely, to survive to the sub-canopy layer in canopy gaps. T. mertensiana and C. nootkatensis understory trees were the only species present in the large, late-snowmelt canopy gaps that were especially common on flat sites. Survival on these microsites was apparently poor as they included few sub-canopy trees of any species. High and low shade affected regeneration patterns differently by species and height class (Figure 2.15). Seedlings of all species were over-represented on microsites that were in high shade but not low shade (HS, no LS); these were also the microsites most likely to be close to a canopy tree (Table 2.6) and thus receive the greatest seed rain. Two patterns emerged among saplings and understory trees (<5 m tall), one for A. amabilis and one for C. nootkatensis and T. mertensiana. The two patterns were most clearly displayed by the different effect of high and low shade on understory trees. High shade had no effect on the location of A. amabilis (x2 = 0.00; p > 0.95) but was associated with fewer C. nootkatensis and T. mertensiana than expected (x2 > 17.51; p < 0.001). In contrast, low shade had no effect on the 41 latter two spec ies (x2 < 1.37; p > 0.10) but w a s assoc ia ted with fewer A. amabilis than expec ted (x2 = 4.15; p < 0.05). The abundance of C: nootkatensis and T. mertensiana in unshaded microsi tes (i.e., no H S and no LS) w a s consistent with their ability to surv ive to the understory layer jn late-snowmelt gaps . Understory t rees of these two s p e c i e s thus matched the regenerat ion patterns predicted by the gap model better than did A. amabilis. Figure 2.15. F requency and observed- to-expected f requency of seed l ings , sap l ings , and understory t rees <5 m tall by high s h a d e (HS) and low s h a d e (LS) . Upper f igures (a) present actual f requenc ies per hectare and lower f igures (b) present ratios of obse rved f requenc ies to those expec ted by the availabil ity of each combinat ion of high s h a d e and low s h a d e . O A. amabilis; • C. nootkatensis; • T. mertensiana. HS, LS noHS, HS, noHS, HS, noHS, HS, noHS, HS, noHS, HS, noHS, LS noLS noLS LS LS noLS noLS LS LS noLS noLS (a) Frequency by high and low shade (b) Ratio of observed-to-expected frequencies by high and low shade 42 Microsites under canopy drip supported more saplings than expected (x2 = 15.60; p < 0.005), but lower apparent survival removed this advantage among understory and sub-canopy layers (x2 < 5.97; 0.05 < p < 0.10). Thirty-six percent of saplings, but only 22% of understory and sub-canopy trees and 24% of quadrats, were located below canopy drip. A total of 40 gapmakers (dead trees or stumps >40 cm diameter) were present on the study plots, of which all but one uprooted tree were in various stages of decay consistent with a standing death (Figure 2.16a). Eighty-five percent of stumps were evidently long dead since they were either mostly or totally rotten (Figure 2.16b). Most gapmakers could not be identified to species due to decay, though these were likely to be A. amabilis or Tsuga since C. nootkatensis was more easily identified (Figure 2.16c). The species mix of dead and live trees was similar except that there were no dead sub-canopy or understory C. nootkatensis. Interactions Between Microsite Factors How did microtopography combine with canopy cover and distance to the nearest canopy tree to affect the location of sub-canopy trees? Microtopography was the over-riding factor as the highest frequencies and the only observed-to-expected ratios >1.0 were on mounds, regardless of canopy cover or distance class (Figure 2.17). The microsites with the most sub-canopy trees were those on mounds and most related to canopy trees, i.e., in closed canopy or <2 m from a canopy tree. C. nootkatensis was especially likely to be restricted to these microsites. The notable exception to these trends was on steep sites where there were many sub-canopy trees, mostly T. mertensiana, on non-mound microsites in canopy gap or >2 m from a canopy tree. 4 3 Figure 2.16. S u m m a r y of the 40 dead trees present on the study plots: (a) d e a d canopy t rees by condi t ion; (b) canopy tree s tumps by decay c lass ; and (c) dead t rees by s p e c i e s and canopy layer. S e e Tab le 2.3 for definit ions of decay c l a s s e s . 30 20 10 30 standing broken stump uprooted mostly partly intact rotten mostly rotten totally rotten (a) D e a d canopy t rees by condit ion (b) C a n o p y tree s tumps by d e c a y c l ass Aa • Cn Tsuga M unknown canopy trees sub-canopy trees understory trees (c) D e a d t rees by s p e c i e s and canopy layer 44 Figure 2 .17. F requency and observed- to-expected f requency of sub -canopy t rees by microtopography, canopy cover, and d is tance to the nearest canopy tree. Lef t -hand f igures (a, c) present results by canopy cover and right-hand f igures (b, d) present results by d is tance to the nearest canopy tree. Upper f igures (a, b) present actual f requenc ies per hectare and lower f igures (c, d) present ratios of obse rved f requenc ies to those expec ted by the availabil ity of each canopy cover or d is tance c l ass . Abbrev ia t ions: cc = c losed canopy ; eg = expanded gap; eg = canopy gap . • flat s i tes; • s teep s i tes. 80 60 40 20 mound, mound, mound, no no no cc eg eg mound, mound, mound, cc eg eg 80 60 40 20 - t - - t -mound, mound, mound, no no no <2m 2-3.9m 4+m mound, mound, mound, <2m 2-3.9m 4+m (a) Frequency by canopy cover (b) Frequency by distance to the nearest canopy tree 3 -• V. 2 £ 1 no no no mound mound mound cc eg eg -+- -+-mound, mound, mound, cc eg eg o 2 • £ 1 A / \ no no no mound mound mound \ <2m 2-3.9m 4+m mound, mound, mound, <2m 2-3.9m 4+m (c) Ratio of observed-to-expected frequencies by canopy cover (d) Ratio of observed-to-expected frequencies by distance to the nearest canopy tree 45 Growth Rates and Microsites Most non-canopy A. amabilis t rees grew slowly: med ian height increments were 1 cm/yr for understory t rees and 10 cm/yr for sub-canopy trees. Med ian relative growth rates, e x p r e s s e d a s the annual height growth per cm of height, were higher for sub -canopy t rees (0.010 cm/cm/yr , or 1% per yr) than understory trees (0.006 cm/cm/yr , or 0 . 6% per yr; M a n n -Whi tney test; T = 10114; p < 0.002). Relat ive growth rates were 2 0 % higher for understory t rees and 4 5 % higher for sub -canopy trees on flat s i tes compared to s teep si tes (t-test for sub -canopy t rees; t = -2 .57; p =0.015; Mann-Whi tney test for understory t rees; T = 31614 ; p = 0.021), an unexpec ted result given that growing s e a s o n s were shorter on flat s i tes. T h e only statistically signif icant relationship between microsi te factors and relative growth rates w a s that understory trees grew 2 9 % faster on mounds (Mann-Whi tney test; T = 39398 ; p = 0.006). Though low samp le s i zes prevented further statistical test ing, contrast ing t rends for understory and sub-canopy trees were evident. A m o n g understory t rees, relative growth rates were highest in canopy gap and >2 m from a canopy tree. S u b - c a n o p y t rees, however , grew a s fast in c losed canopy as canopy gap (and 1 0 % faster than in e x p a n d e d gap), and >14% faster on microsi tes <2 m from a canopy tree than those farther away . Relat ive growth rates were a lso recorded by high and low s h a d e condi t ions for A. amabilis understory t rees <5 m tall. Low shade w a s assoc ia ted with 4 3 % s lower growth (Mann-Whi tney test; T = 14523 ; p < 0.0001), while high shade had no effect (Mann-Whi tney test; T = 28045 ; p = 0.553). Growth Form Anomalies Most understory trees had more than one growth form anomaly , though few were ex t reme. T h e majority were s tem anomal ies , especia l ly pistol butts, s tem s w e e p s , and dog legs (definitions of these terms are included as Append ix B) . Umbre l la growth forms were much more c o m m o n a m o n g A. amabilis (80%) than other spec ies (13%). Part of this 46 dif ference w a s likely b e c a u s e the determinate growth form of A. amabilis made the umbrel la s h a p e more obv ious . Tsuga and C. nootkatensis were more likely to be bent in s tem s w e e p or prostrate growth forms, especia l ly in late-snowmelt gaps . More ext reme anoma l ies , like broken tops, dog legs, pistol butts, and s tem s w e e p s , were especia l ly c o m m o n on late-snowmelt and rock wal l microsi tes and so were likely c a u s e d by the weight or movement of snow. A lmos t all understory tree s tems had many ca l luses that indicated numerous b reaks of the terminal leader, likely a lso due to snow. The f requency and severi ty of these ca l luses w a s not quant i f ied. There w a s no ev idence to show that canopy cover , p resence of canopy drip, or microtopography were related to growth form anomal ies . Tsuga, espec ia l ly in the sub-canopy layer, w a s most likely to have a pistol but t -shaped lower s tem. Pistol-butted s tems apparent ly straightened with height growth as they were least c o m m o n and severe in the canopy layer. C a n o p y trees had severe pistol butt (i.e., >50 c m horizontal d is tance) in <10% of c a s e s . Sub -canopy T. mertensiana and C. nootkatensis had the highest inc idence of severe pistol butt with >30% of t rees af fected. L e s s than 1% of all t rees had pistol butts with a horizontal d is tance of >100 c m . S lope did not affect the inc idence or sever i ty of pistol butt. 2.4 Discussion Are microsites, snow, and regeneration patterns related? Though mean snow depths were related to microsite factors, snow on individual samp l ing points (every 0.5 m) w a s surprisingly var iable. G i ven this variabil ity, t ree -based rather than quadra t -based measuremen ts would have been preferable. Never the less , the st rong co r respondence between microsite factors and snow, and between microsi te factors and regenerat ion, s h o w s the impact of snow on regenerat ion patterns. Cons is ten t with the tree-is land model of regenerat ion, survival w a s highest on mounds and microsi tes c lose to a canopy tree, both of which were in turn related to longer growing s e a s o n s . 47 Whi le site and microsite factors affected patterns of snow accumula t ion (measured in Apri l) , their impact w a s most apparent in how they affected snowmel t (measured in mid-May) . T h e o p e n n e s s of the canopy on s o m e si tes (especial ly E F ) w a s assoc ia ted with, and apparent ly c a u s e d by, deep and late-melting snow that affected regenerat ion patterns. T h e s e gaps were up to one tree height in diameter, which is the s ize that retains s n o w longest (Gold ing and S w a n s o n 1978; Berry and Rothwel l 1992). T h e most obv ious effect of s n o w in late-snowmelt gaps w a s the rarity of canopy tree s tumps or live sub -canopy t rees. S u c h gaps have not suppor ted canopy trees for many centur ies nor are they likely to in the near future. Late snowmel t in gaps therefore contr ibutes to the greater o p e n n e s s of these s tands c o m p a r e d to those in lower-elevat ion, gap-dr iven sys tems (reviewed in Ler tzman 1989). O v e r h e a d canopy cover did not prove to be a useful predictor of regenerat ion patterns in these s tands . The distinction between microsi tes that are in gaps and those that are not b e c o m e s less c lear with increasing elevat ion as overal l canopy o p e n n e s s (gappiness) i nc reases . Though I know of no quantif ication for this asser t ion, light levels be low the canopy are a lmost certainly higher and more homogeneous in these s tands than in old-growth s tands at lower e levat ions. T rees are not as tall, their b ranches are shorter, there is a l imited sub -canopy layer, and gaps remain uninhabited for longer. The result, comb ined with the low angle of growing s e a s o n sunlight at these latitudes (Brooke etal. 1970; C a n h a m etal. 1990; van Pelt 1995), is an inc reased importance of diffuse light and sunf lecks with a predictably lesser inf luence of gap p r o c e s s e s . At s o m e point in the elevat ional transit ion from gap to t ree- is land p r o c e s s e s , the relat ionship between canopy cover and regenerat ion must therefore d e c r e a s e s o much a s to be unimportant. Due to the overwhelming cover and importance of undisturbed forest f loor, and in contrast to other sys tems shar ing one or all of the s a m e genera , regenerat ion patterns were not strongly affected by substrate heterogeneity. In other sys tems , regenerat ion is disproport ionately c o m m o n on 'disturbed' subst rates, those that are products of d is tu rbances 48 too recent to al low the formation of a thick humus layer, e.g. , decay ing w o o d , coa rse woody debr is , and mineral soil (Minore 1972; Chr is ty and M a c k 1984; An tos and Zobe l 1986, Harmon and Frankl in 1989; Ler tzman 1992; Danie ls 1994). Disturbed subst rates are important in more cont inental subalp ine forests of western North A m e r i c a as wel l , espec ia l ly for the regenerat ion of Picea (e.g., Knapp and Smith 1982; Harvey etal. 1987; V a r g a 1997). W h y were d isturbed subst ra tes s o unimportant here? T h e c o m m o n explanat ion for the abundance of regenerat ion on coa rse woody debr is in other s y s t e m s is that its elevat ion off the forest floor reduces competi t ion (Harmon and Frankl in 1989; Ler tzman 1989) or prevents burying by litter (Thornburgh 1969; Chr is ty and M a c k 1984). E leva ted microsi tes were assoc ia ted with success fu l regenerat ion here, a s demonst ra ted by the abundance of t rees on mounds (mostly fo rmed over boulders) . But a lmost all s u c h mounds were covered by humus to a depth of at least 10 c m which s h o w e d that many years had p a s s e d s ince the last d is turbance. T h e s e factors argue against an important role for anything but the most decayed wood , i.e., that which is being incorporated into the forest floor. They a lso suggest that where s tump remains are found beneath a canopy tree, a s previously reported in the M H zone (Peterson 1964; Ler tzman 1989), the estab l ishment of the canopy tree w a s likely poss ib le only after humus had fo rmed over the s tump. T w o poss ib le reasons for the lesser importance of woody subst rates found here are: (1) native M H spec ies may be less adapted to these substrates than the Paci f ic Northwest s p e c i e s usual ly ci ted as coloniz ing them, T. heterophylla and Picea spp . ; and (2) regenerat ion may be rooted insecure ly in woody substrates, especia l ly coa rse woody debr is , and therefore more suscept ib le to uprooting through snow creep. A third reason may be that many s tud ies , espec ia l ly those b a s e d on exper imental data, rely on results from only the first few yea rs after germinat ion (e.g., Chr is ty and M a c k 1984; Harmon and Frankl in 1989; Harvey etal. 1987; N a k a m u r a 1992). A s Harmon and Frankl in (1989) caut ion, a focus on estab l ishment rather 49 than survival may over-est imate the importance of s o m e subst rates espec ia l ly if, a s in the c a s e of coa rse woody debr is , future mortality may be higher than on the forest floor. In contrast to woody substrates, the lack of regenerat ion on mineral soi l l ikely resul ted from the rarity of so i l -expos ing windthrows and its result ing unavailabil i ty as a subst rate. T. mertensiana, A. amabilis, and C. nootkatensis are known to regenerate success fu l l y on mineral soi l (Burns and Honka la 1990), and this w a s shown by the abundant regenerat ion of these s p e c i e s on the one windthrow microsite within the study plots. T h e neutral or weak ly posit ive relationship between m o s s cover and regenerat ion patterns contrasts with two other s tudies. Harmon and Frankl in (1989) explain the poor regenerat ion on m o s s in their study a s due to competi t ion between trees and the m o s s , but primari ly when m o s s mats are much thicker (>10 cm) than s e e n a m o n g the m o s s s p e c i e s desc r ibed here. Though N a k a m u r a (1992) found lower mortality for 2 year-o ld seed l ings on m o s s than on other s e e d b e d s , he did not compare mortality rates on m o s s to those on undisturbed forest floor nor did he track later surv ival . S n o w c reep can exert extreme downhil l p ressures in high snow a reas (Huz ioka et al. 1967; M a c k a y and Mathews 1967; Brooke e r a / . 1970; Lowery 1972), but its impact on regenerat ion patterns here can only be inferred. The strong tendency of non-canopy t rees to be downhi l l of mounds , especia l ly on s teep si tes, sugges ts that snow creep reduced survival on the uphill s ide (Wil l iams 1966; Brooke etal. 1970; Leaphart etal. 1972; Lowery 1972; M e g a h a n and Stee le 1987, 1988). That the proportion of individuals downhil l of mounds w a s greater a m o n g sub -canopy than understory t rees supports this explanat ion. Do regeneration patterns reflect the gap or tree-island model? Regenera t ion patterns consistent with those predicted by the t ree- is land mode l and , to a lesser deg ree , those predicted by the gap mode l , show that the study s tands occupy a transit ion be tween high- and low-elevat ion ecosys tems . A s expec ted in s u c h a transit ion, regenerat ion patterns were not consistent among spec ies or si tes and were apparent ly related 50 to the impact of snow. For example , regenerat ion on late-snowmelt s i tes w a s more. l ikely to match the t ree- is land model than regenerat ion on ear ly-snowmelt s i tes. Whi le I u s e d non-canopy t rees to reach this conc lus ion , it is a lso reflected in the composi t ion of canopy t rees. T h e s p e c i e s best adapted to heavy snow, C. nootkatensis and T. mertensiana (Brooke et al. 1970), ou tnumbered A. amabilis in each canopy layer on only one site, E F , the site with the latest snowmel t . A n d the spec ies best adapted to lower e levat ions, 7". heterophylla, reached the canopy layer on only three s i tes: both south-facing s teep si tes ( S S and LS ) and the flat site with the earl iest snowmel t (SF) . S p e c i e s could a lso be differentiated by their regenerat ion patterns. C. nootkatensis regenerat ion matched the tree- is land model most c losely as it w a s most c o m m o n on mounds and near to a canopy tree. T. mertensiana showed more of a mixed pattern s ince it w a s able to regenerate both c lose to a canopy tree and , especia l ly on s teep, south- fac ing s i tes, in canopy gaps . Its use of mounds w a s a lso less than the other spec ies . The dif ference be tween these s p e c i e s may be related to their ecologica l n iches. That is, the regenerat ion of a s p e c i e s near the upper limit of its elevat ional range (such as C. nootkatensis) may be more restr icted by s n o w than a s p e c i e s in the middle of its range such as T. mertensiana. W h e r e regenerat ion fol lows the gap mode l , the death of one or more canopy t rees shou ld lead to the faster growth of non-canopy trees in the new gap. However , most gaps were not c a u s e d by the recent death of a canopy tree (as shown by the a b s e n c e or a d v a n c e d decay of gapmakers ) . Nor w a s there any ev idence that t rees entered the canopy layer by sudden ly growing faster (releasing) as would be expec ted under the gap mode l : most canopy t rees had a history of growing slowly and steadi ly. The only poss ib le except ion w a s T. mertensiana which tended to grow faster once it reached a breast-height d iameter of - 1 0 c m . Th is result w a s based on 12 samp les that were not in c lose proximity to each other, and would likely have been subject to different overhead canopy condi t ions. T h e similarity of the d iameter at wh ich they began to grow faster sugges ts that the c a u s e w a s growth above the s n o w p a c k 51 rather than the formation of new gaps (though there may be a relat ionship be tween the two). Other researchers have a lso hypothes ized that trees larger than about 10 c m at breast height avo id be ing t rapped under snow during the winter and so can take advantage of a longer growing s e a s o n (Brink 1959; Brooke e r a / . 1970; Ler tzman 1989). Though the growth patterns of understory A. amabilis were simi lar to those predicted by the gap mode l (faster growth in canopy gaps and far from canopy trees), the contrast ing trends for sub -canopy trees more c lose ly matched the tree- is land mode l . Sma l l samp le s i z e s a m o n g sub -canopy t rees and the a b s e n c e of growth rate data for T. mertensiana and C. nootkatensis al low only tentative conc lus ions to be made , but these are trends support ing other results a l ready p resented , i.e., that the location and growth rates of understory t rees are not necessar i l y good indicators of the future development of a s tand. T h e implicit assumpt ion throughout this study has been that the present growth envi ronment is representat ive of past condi t ions. From one perspect ive, this assumpt ion may be reasonab le . T h e preva lence of very old t rees, single snags , and a thick humus layer (also reported by Brooke etal. 1970 and Ler tzman 1989) sugges t that the study s tands have been at a simi lar s tage of deve lopment (true old-growth, sensu Ol iver and Larson 1990) for many centur ies. However , the cyc le from germinat ion to s e n e s c e n c e is long enough to s p a n major cl imatic f luctuations that undoubtedly alter the proportion of precipitation fall ing a s snow. Regenera t ion patterns, so strongly affected by snow, would therefore by changed and the elevat ional band that conta ins the transition from gap to t ree- is land p r o c e s s e s would accord ing ly move higher or lower. The forested M H subzone , the express ion of this transit ion, is therefore not f ixed at a certain elevat ion. 2.5 Conclusions Regenera t ion patterns in the study s tands represent a transit ion from the gap model to the t ree- is land model and are related to snow, particularly patterns of snowmel t . T h e tree-52 is land mode l is exp ressed most clearly on late-snowmelt s i tes and by C. nootkatensis, a s p e c i e s near the upper limit of its range. T. mertensiana, a spec ies in the middle of its range, is capab le of regenerat ing near canopy trees and , especia l ly on ear ly-snowmel t s i tes, in canopy gaps . Regenera t ion patterns are important beyond their theoretical interest. Th is study s h o w s that i nc reased snow restricts regenerat ion to microsi tes protected by an ove rhead canopy , even where d iscrete tree is lands are not present. Regenera t ion patterns are therefore eco log ica l indicators of site severi ty and should be cons idered when managemen t dec is ions are m a d e . For examp le , they could be used to help del ineate the three managemen t c l a s s e s p roposed by K l inka etal. (1992) for s tands in the forested M H subzone : (1) s tands where regenerat ion after cutting is likely to be problem-free, though s low; (2) s tands that are marginal for t imber product ion due to potential regenerat ion prob lems and the importance of other, non-t imber va lues ; and (3) s tands unsuitable for cutting. I bel ieve that these c l a s s e s co inc ide (at least roughly) with s tands where regenerat ion patterns match: (1) the gap mode l ; (2) a transit ion be tween gap and tree- is land mode ls ; and (3) the tree- is land mode l . Regenera t ion patterns could be included in site reconna issance to quantify the p resence of patterns matching the tree- is land model , and therefore to est imate site sever i ty. Speci f ica l ly , s u c h a reconna issance should quantify the proportion of sub -canopy t rees on mounds and c lose to canopy trees compared to those in canopy gaps . T h e s e indicators cou ld then be bols tered by s tandard measu res of site classi f icat ion (Green and Kl inka 1994) including: height of the canopy layer, spec ies composi t ion (especial ly the proport ion of T. heterophylla compa red to T. mertensiana), percent canopy c losure, indicator plant ana lys is (Kl inka etal. 1989), and date of snowmelt . S u c h a classi f icat ion would be val id whether the managemen t focus w a s timber, water, wildlife, recreat ion, or w i lderness. 53 Chapter 3. Natural Regeneration on Clearcut Sites 3.1 Introduction T h e best method for regenerat ing logged si tes in the M H zone remains unc lear after approx imate ly 30 years of logging history. Ever s ince early regenerat ion fai lures were l inked to s lashburn ing and the planting of unsuitable, low-elevat ion spec ies (Reuter 1973; Utz ig and Herr ing 1974; K l inka and Pend l 1976), most si tes have been left to regenerate naturally. Though this change has improved results, K l inka et al. (1992) identified a number of si lvicultural conce rns within the zone , many of which centred on the role of natural regenerat ion. B a s e d on these concerns , I undertook a survey of natural ly-regenerat ing s i tes to answer the fol lowing quest ions: 1. W a s natural regenerat ion success fu l? 2. Did most natural regenerat ion establ ish before or after logging? 3. W h i c h subst rates favoured natural regenerat ion? 4. W h i c h microtopographic locat ions favoured natural regenerat ion? 5. Did competi t ion with Vaccinium impede natural regenerat ion? Th is chapter comp lements the study of regenerat ion patterns in old-growth s tands in Chap te r 2 and , where appropr iate, I compare results from clearcut and old-growth s i tes. 3.2 Methods Study Area T h e study a rea and its old-growth s tands are descr ibed in Sect ion 2.2. S imi lar s tands at e levat ions <1100 m were clearcut between 1975 and 1985 then left to regenerate naturally without s lashburn ing or planting. Vaccinium alaskaense (A laska blueberry) domina tes the non-tree vegetat ion on these c learcuts. Other shrubs that are locally abundant include V. ovalifolium (oval - leaved blueberry), V. membranaceum (black huckleberry) , Rhododendron 54 albiflorum (white-f lowered rhododendron), and Menziesia ferruginea (false aza lea ) . Ve ry little light r eaches the ground under these shrubs so little e lse grows other than a minor cover of Cornus canadensis (bunchberry), Rubus pedatus ( f ive-leaved bramble) , and m o s s e s such as Rhytidiopsis robusta (p ipecleaner moss ) , Polytrichum juniperinum (juniper ha i rcap moss ) , and Dicranum spp . Epilobium angustifolium (f ireweed) is c o m m o n on microsi tes without Vaccinium. Drier s i tes may support Linnaea borealis (twinflower) and Hieracium albiflorum (white-f lowered hawkweed) , whi le wetter s i tes support Lysichitum americanum (skunk cabbage) , Veratrum viride (Indian hel lebore), Coptis asplenifolia ( fern-leaved goldthread), and Sphagnum spp . (peat m o s s e s ) . A complete list of spec ies is included as Append ix A . Study Design I es tab l ished three study locat ions in c learcuts that had been logged 11-12 yea rs prior to samp l ing (Figure 2.1). E a c h study location cons is ted of one 'flat' site (22-24% slope) and one s teep site (49 -53% slope) for a total of s ix si tes (Figure 2 .2a ; Tab le 3.1). E levat ions above s e a level ranged from 1060-1100 m. I located t ransects and 1 m 2 microsi te quadrats a s fully desc r ibed in Sec t ion 2.2 and shown in Figure 2.2b. Tab le 3 .1 . C learcu t study site descr ipt ions. Site codes combine the first initial of the study locat ion and the s lope type, e.g. , B F signif ies Batchelor Flat. So i l moisture reg imes ( S M R ) and soil nutrient reg imes ( S N R ) follow Kl inka etal. (1989). S M R abbrev iat ions: F = f resh; M = moist; V M = very moist. S N R abbreviat ions: P = poor; M = med ium. Study S lope Site Elevat ion S lope Aspec t Y e a r of S M R S N R locat ion type code (m) (%) (deg. azim.) logging Batche lor Flat B F 1070 24 24 (NE) 1981 M / V M P S teep B S 1090 50 24 (NE) 1981 M P M a y n e Flat M F 1080 22 198 (S) 1981 M / V M P / M S teep M S 1100 49 190 (S) 1981 F P Tann is Flat T F 1060 23 191 (S) 1982 M P S teep T S 1090 53 192 (S) 1982 F P 55 Study locat ions were se lec ted to meet the fol lowing criteria: (1) constant south or north aspec t ; (2) s teep site directly above or near flat site; (3) limited edaph ic var iat ion; (4) s imi lar date of logging with enough delay to d isplay regenerat ion; (5) not burned or p lanted; (6) no sk id roads or other severe logging d is turbance; (7) abundant Vaccinium; and (8) l imited representat ion of T. heterophylla. To al low compar isons , I at tempted to match the aspec t s and s lopes of s tudy locat ions in c learcuts with those of the old-growth s tands desc r ibed in Chap te r 2 (Table 2.1). However , progress ive logging left old-growth s tands only at e levat ions higher than c learcuts for all except the north-aspect pairing of the Batchelor and E d w a r d s study locat ions. I a l so intended to test the effect of forest edges on natural regenerat ion, but the few forest e d g e s left in the study a rea did not meet my select ion criteria. T h e Batche lor and Mayne study locat ions were at lower e levat ions of the W indward Moist Mar i t ime ( M H m m l ) biogeocl imat ic variant of the M H z o n e (Green and K l inka 1994). M a p s p repared by the B . C . Ministry of Forests (1992) show the Tann is study locat ion be low the M H boundary, within the Montane Very Wet Mari t ime ( C W H v m 2 ) variant of the C W H zone . However , these maps were prepared at a sca le that was too coarse to dist inguish the var iable transit ion be tween z o n e s in the study a rea (Brett 1996). A c loser inspect ion of the Tann is study locat ion s h o w e d a slightly greater p resence of T. mertensiana than T. heterophylla, espec ia l ly on flatter port ions, which met the definition of the M H rather than the C W H z o n e . Data Collected Site and microsite sampl ing methods were the s a m e as descr ibed for old-growth s i tes (Sect ion 2.2), except that canopy cover classi f icat ion w a s unnecessa ry and a g e s were s a m p l e d differently. T o relate tree a g e s and heights, I randomly samp led two trees of each s p e c i e s and from e a c h 10 c m height c lass (10-149 cm) from microsite quadrats at each study locat ion. S i n c e t rees >150 c m tall were almost a lways much older than the age of the c learcut (as shown by vis ible height re lease and conf i rmed by limited destruct ive sampl ing) , further 56 destruct ive sampl ing w a s restricted to trees <150 cm tall. F ive seed l ings , of each deve lopmenta l s tage (Figure 2.5) and spec ies , and from each study locat ion, were a lso randomly s a m p l e d . W h e n tree or seedl ing samp les were unavai lable within microsi te quadrat boundar ies , I s a m p l e d adjacent individuals (if present). Al l samp les were excava ted or cut to include their root col lar. S a m p l e s were sanded or re-cut with a razor b lade, cha lked and/or wetted to better differentiate r ings, then rings were counted using a 40x binocular m ic roscope . Due to incomplete or miss ing rings, reported ages should be cons idered min ima. W h e n s a m p l e s s h o w e d a dramat ic increase in the rate of d iameter growth (diameter re lease) , I es t imated the number of years s ince re lease by count ing the number of rings outward from the point at wh ich ring widths not iceably inc reased. Addi t ional data were recorded in the field for t rees that showed ev idence of a dramat ic inc rease in the rate of height growth (height re lease) . Height re lease w a s clear ly d isp layed in the growth whor ls of A. amabilis and somewhat less clearly in the annua l increments of C. nootkatensis and Tsuga. I measured re lease height to the nearest 0.1 m and recorded the number of yea rs s ince height re lease. W h e r e poss ib le , these data inc luded two t rees from e a c h 10 c m height c l ass >150 cm tall. Data Analysis Analyt ica l methods are descr ibed fully in Chapte r 2. The fol lowing only exp la ins deviat ions from that descr ipt ion. Ages: M y main goal in sampl ing ages w a s to determine the proport ion of t rees that were a l ready estab l ished at the time of logging (residuals, or advance regenerat ion), and those that es tab l ished after logging ( ingress). A g e s were samp led over a range of heights to der ive a regress ion line and predict the age distribution of unsamp led trees but, s ince a g e s were poorly related to heights (Figure 3.1), t rees were instead grouped into age c l a s s e s . In addit ion to age c l a s s e s that I te rmed residuals and ingress, I identified a third age c lass (germinants) b e c a u s e there w a s a distinct group of t rees that germinated between 1 year before and 1 year after 57 logging (Figure 3.1; Table 3.2). The age distribution w a s then est imated within each study location by multiplying the proportion of res iduals, germinants, and ingress s a m p l e d in each 10 cm height c lass by the f requency of all t rees in that height c l ass . Substrate: To focus on substrates inhabited by t rees, "unavai lab le" subst ra tes including logging s l a s h , rock, and s tumps were removed from ch i -square and cont ingency table ana l yses (only 1 tree, 11 cm tall, was growing on any of these sur faces) . A s a result, x 2 va lues for the remaining "avai lable" substrates are more conservat ive (because expec ted va lues are higher) but are a lso more likely to differentiate between these subst ra tes. Figure 3 .1 . Jitter plot of heights of samp led trees and seed l ings on the year of their establ ishment (n = 210). Data points are offset from each other (jittered) s o that ail can be s e e n . The year of establ ishment is reported relative to the year of logging s o that negat ive va lues indicate trees and seed l ings that es tab l ished before logging, posit ive va lues indicate establ ishment after logging, and zero va lues indicate es tab l ishment during the year of logging. No trees were found that es tab l ished > 8 yea rs after logging. T h e -5 c lass includes all t rees that es tab l ished 5 or more years before logging. 500 400 E 300 o ffi 200 100 0 I I I I — « res iduals o I I I germinants l I i i i i i ingress nP o - ?> „ o */. 0 * 0 o % « *<> o« o <# • 0 0 a -2 + 1 +2 +4 +5 +6 +7 +8 establishment year (relative to logging) 58 Tab le 3.2. Definit ion of terms used in Chapter 3. Note that the definition for t rees in Chap te r 3 inc ludes what were cal led sapl ings in Chapte r 2. Te rm Definition Seedling Any tree spec ies <10 c m tall. Tree Any tree spec ies > 10 cm tall. Ingress T r e e s and seed l ings estab l ished >2 years after logging. Germinant Trees and seed l ings estab l ished within ±1 year of logging. Residual T r e e s and seed l ings estab l ished >2 years before logging (advance regenerat ion). Vaccinium: A tree is cons idered free-growing in B . C . if it meets min imum height requi rements or, when growing amidst other vegetat ion (e.g. Vaccinium) is 1.25 t imes the height of that vegetat ion ( B . C . Ministry of Forests 1995). To project the number of yea rs it wou ld take t rees growing below Vaccinium to b e c o m e free-growing, I used m e a n A. amabilis growth rates within each 10 cm height c lass (the height growth of the other s p e c i e s w a s too difficult to d iscern , especia l ly when they were suppressed) . I projected future growth by adding the increment that co r responded to a tree's height c lass to that tree's height, repeat ing until it reached the next c lass , then adding the appropriate increment from this next c l ass . I repeated this p rocess until the tree w a s 125 c m tall (using the assumpt ion that Vaccinium se l dom grew taller than 100 cm). Stocking: S tock ing is a measure of the distribution of f ree-growing trees (see above) that are a lso we l l - spaced . The B . C . Ministry of Fores ts (1995) appl ies a min imum inter-tree spac ing of 2 m, and only one free-growing tree within that spac ing contr ibutes to s tock ing. T o determine s tock ing, I comb ined two microsite quadrats to create one 2 m 2 cel l then ca lcu la ted the p resence or a b s e n c e of at least one free-growing tree within that cel l . To project future s tock ing, I used growth projections a s descr ibed in the paragraph above , but with one modif icat ion. For t rees growing without surrounding vegetat ion, but be low min imum height requi rements, I projected the number of years to exceed that min imum height. 59 3.3 Results Structure, Composition, and Age A. amabilis w a s the most c o m m o n spec ies overal l and for t rees >150 c m tall (Figure 3.2). C. nootkatensis w a s as abundant as A. amabilis in shorter height c l a s s e s whi le Tsuga w a s least c o m m o n throughout the height range. The tallest recorded tree w a s 4 .35 m in height. S i n c e only one Pinus monticola Dougl . e x D . Don (western white pine) w a s recorded it w a s exc luded from further ana lys is . W h e n there is a constant ingress of seed l ings , as on the old-growth s i tes of Chap te r 2, height distr ibutions follow an inverse-J shape with f requenc ies dec reas ing exponent ia l ly from shorter to taller height c l a s s e s . The relative scarci ty of seed l ings on clearcut s i tes, however , s h o w e d that ingress w a s not constant, especia l ly for A. amabilis. There were only 1233 seed l i ngs /ha on clearcut si tes (compared to 187,000/ha on old-growth si tes), and none es tab l ished in the 4 years before sampl ing . Most seedl ings (64%) were restricted to one site, B F , and 5 0 % of all seed l ings were C. nootkatensis (F igures 3.2 and 3.3). Within e a c h study locat ion, flat s i tes had more seed l ings than s teep si tes. F igure 3.2. Height distribution of trees and seedl ings by spec ies . 1000 1000 to JZ CD 7 5 0 CL B- 500 CD CT 2 5 0 CD 0 A. amabilis T. mertensiana fiffififyrrr^ f / y / ywjS^ i i 1 C M ^ f C O C O O O U T l O C O upper limit of height class (cm) CO CD 1000 750 o 500 c CD CT 250 CD 0 C. nootkatensis T. heterophylla 0 1 0 > 0 > 0 0 > 0 ) 0 > C J > 0 } 0 > + C M ^ C D C O O C M ^ C O C O O 1- T- T- T- O CM upper limit of height class (cm) 60 Figure 3.3. Height distribution of trees and seedlings by species and site. Key to species abbreviations: Aa = Abies amabilis; Cn = Chamaecyparis nootkatensis; Tm = Tsuga mertensiana; Th = Tsuga heterophylla. 6000 4000 2000 Batchelor Flat (BF) 6000 4000 2000 6000 Batchelor Steep (BS) 4000 2000 Tannis Flat (TF) 6000 _ 4000 2000 6000 Tannis Steep (TS) 4000 2000 Mayne Flat (MF) 6000 4000 2000 Mayne Steep (MS) upper limit of height class (cm) 61 S i n c e s i tes had a similar number of trees >150 cm tall, the large d i f ferences in their height distr ibutions were c a u s e d mainly by different f requenc ies of shorter t rees and seed l ings (Figure 3.3). Three si tes (BF , B S , and TF) had many more seed l ings and t rees <150 c m tall than the remaining three si tes (TS, M F , and M S ) . The B F site suppor ted the most regenerat ion <150 c m tall (31,300 s tems/ha) , fol lowed by the B S and T S si tes (10,800 and 9,800 s t ems /ha , respect ively) . Dens i t ies on the three less-populated si tes were 2500-4600 s t ems /ha . T h e magni tude of these di f ferences w a s not apparent when select ing s i tes s ince most shorter t rees were h idden under Vaccinium. A. amabilis w a s present throughout the height range on all s i tes but w a s most abundant on the two north-facing s i tes, B F and B S . C. nootkatensis w a s more c o m m o n on the flat than s teep site of each study locat ion. T. mertensiana w a s much more abundant than T. heterophylla on north-facing si tes and slightly more abundant on three of four south- fac ing s i tes (all except the T S site). The T S site had few T. heterophylla (500/ha) and no T. mertensiana. A. amabilis t rees grew faster on clearcut than old-growth s i tes (Mann-Whi tney test of relative growth rates; T = 149955; p < 0.0001) and height growth on clearcut s i tes w a s posit ively related to height (Pearson correlation on natural log-transformed data ; r = 0 .83; p < 0 .0001 ; F igure 3.4). The mean height of residuals at re lease w a s 47.7 ±34.2 c m (n = 62) and ranged from 42.1 ±21.5 c m for C. nootkatensis to 60.0 ±36.1 c m for T. heterophylla. M a x i m u m height at re lease w a s 170 c m . There w a s a delay of 2.9 ±1.6 years before d iameter re lease and height re lease occur red a lmost 2 years later, 4.8 ±1.3 years after logging. M e a n height growth rate of t rees before and after re lease w a s 3.2 cm/yr and 18.6 cm/yr, respect ively, and their m e a n age w a s 48.5 ±14.6 years . On ly 2 0 % of all t rees and seed l ings estab l ished more than one year after logging (Figure 3.5a). Of those that establ ished earlier, 3 5 % were residuals and 4 5 % were germinants . There w a s no ingress of A. amabilis >5 years after logging and no ingress of any spec ies >8 62 years after logging. The majority (76%) of residuals were A. amabilis while approximately half of all germinants and ingress were C. nootkatensis. T. mertensiana w a s almost evenly distr ibuted among age c l a s s e s but showed relatively strong ingress. T. heterophylla had the fewest residuals. The number of residuals was similar ac ross height c l a s s e s but their proport ion of regenerat ion inc reased with height (Figure 3.5b). A lmost all t rees >150 c m tall were res iduals. A lmost all (94%) ingress and 6 4 % of germinants were <50 cm tall. Destruct ive sampl ing of t rees revealed that more were residuals than w a s originally recorded during non-destruct ive sampl ing . Excava t ions revealed very thick root b a s e s , sprout ing, and/or re-orienting of c rushed branches and s tems. For example , what appea red to be severa l separate, 2 m tall and sexual ly mature C. nootkatensis turned out to be b ranches of a s ingle tree crushed during logging. Figure 3.4. Height growth of A. amabilis on clearcut s i tes (n=295) and old-growth s i tes (n=434). No seed l ings (height <10 cm) are inc luded. 50 40 E 30 c 20 10 © © © f t © © © © ©• ©© « • © ooo© ©© © o© © « 9 * © « © « 100 200 300 height (cm) 400 500 100 200 300 height (cm) 500 Clearcu t si tes Old-growth s i tes 63 Figure 3.5. Est imated age c lass distribution: (a) by spec ies , and (b) by height c l a s s . 4000 A. amabilis CD C. nootkatensis H T. mertensiana H T. heterophylla residual germinant ingress (a) Es t imated a g e c lass distribution by s p e c i e s 4000 3000 >> 2000 p 1000 <50 residual germinant 3 ingress Ml 50-99 100-149 height class (cm) 150+ (b) Es t imated a g e c lass distribution by height c lass Di f ferences in the abundance of germinants and ingress c a u s e d most of the variat ion in height distr ibutions a m o n g si tes s ince the f requency of res iduals, especia l ly those >150 c m tall, w a s simi lar (Figure 3.6). Three si tes (BF , B S , and TF) had many more ingress and germinants than the remain ing three s i tes (TS, M F , and M S ) , and there were more ingress and germinants on the flat site of each study location pair ing. The M S site w a s especia l ly underpopula ted by germinants and ingress as A. amabilis residuals made up 9 0 % of regenerat ion. Within e a c h study locat ion, the flat site had more germinants and ingress than the s teep site. 64 Figure 3.6. Est imated age c lass distribution by height c lass and site. Note that the sca le for the Batche lor Flat (BF) site is slightly different than for other s i tes. Tannis Flat (TF) 12000 8000 4000 <50 50-99 100-149 height class (cm) 150+ Tannis Steep (TS) 12000 8000 4000 <50 50-99 100-149 height class (cm) 150+ Mayne Flat (MF) Mayne Steep (MS) 12000 5s 8000 4000 12000 8000 4000 <50 50-99 100-149 height class (cm) 150+ <50 50-99 100-149 height class (cm) 150+ 65 Substrates Relative to adjacent old-growth stands (Chapter 2), there was less undisturbed forest floor and much more friable forest floor (Table 3.3). Logging slash more than tripled the proportion of coarse woody debris on clearcut sites compared to old-growth sites. Undisturbed forest floor on clearcut sites covered 55% of the ground surface on flat sites compared to 27% on steep sites. Steep sites had more friable forest floor (especially on the TS site), logging slash, and exposed rock than flat sites. Friable forest floor consisted of a Mormoder humus form (Green etal. 1993) that indicated greater biological activity than the Mor humus form of undisturbed forest floor (Green and Bernardy 1991). There was virtually no mineral soil. Unavailable substrates (logging slash, rock, and stumps) covered 32% of the ground surface compared to 11% on old-growth sites. Moss cover was lower on clearcut (5%) than old-growth (42%) sites. Table 3.3. Percent cover of substrates by site. Key: uff = undisturbed forest floor; fff = friable forest floor; dw = decaying wood; edw = exposed decaying wood; ms = mineral soil. Slash includes both naturally-occurring coarse woody debris (= CWD in Table 2.5) and logging slash. The stump class also includes standing trees on old-growth sites. Available substrates Unavailable substrates Site uff fff dw edw ms slash rock stump Total BF 59.4 8.2 7.5 0.7 0.0 20.8 0.9 2.6 100.0 BS 41.5 11.5 8.7 2.5 0.0 28.4 0.2 7.2 100.0 MF 57.1 9.4 8.6 0.9 0.2 22.3 0.0 1.6 100.0 MS 38.3 12.3 7.5 3.8 0.0 31.4 4.7 2.2 100.0 TF 49.4 11.6 4.6 1.6 0.0 25.8 0.4 6.7 100.0 TS 27.4 30.2 5.3 1.3 0.0 21.5 8.9 5.7 100.1 clearcut total 45.5 13.8 7.0 1.8 0.0 25.0 2.5 4.3 100.0 old-growth total 80.0 0.2 7.1 0.9 0.6 7.1 0.8 3.4 100.0 Undisturbed forest floor supported more trees and seedlings of each species than expected (4.84 < x 2 < 71.81; 0.0001 < p < 0.05; Figure 3.7), and even more than on old-growth sites (28% compared to 4% more than expected on clearcut and old-growth sites, respectively). Friable forest floor and exposed decaying wood supported the least regeneration 66 relative to their abundance . Tsuga were the only trees and seed l ings more c o m m o n than expec ted on a substrate other than undisturbed forest floor (decaying wood) , but they were still >4 t imes more c o m m o n on undisturbed forest floor. S ince taller t rees were more likely to be on undisturbed forest floor than were shorter trees and seed l ings , survival w a s likely greater on that substrate for all spec ies except T. heterophylla (Figure 3.8). Undis turbed forest floor suppor ted >90% of regenerat ion regard less of age c lass . F igure 3.7. Observed- to -expec ted ratio on avai lable substrates by spec ies . The f requency of t rees and seed l ings on a substrate (observed) is compared to the percent cover of that substrate (expected). Key to spec ies abbreviat ions: Aa = Abies amabilis; Cn = Chamaecyparis nootkatensis; Tm = Tsuga mertensiana; Th = Tsuga heterophylla. I undisturbed forest floor d friable forest floor ED decaying wood 2.00 2 T3 B o CD Q. <j> 1.00 0.00 1.73 1.34 1.29 E3 1.14 0.28 0.4: Es3 0.51 E2 0.29 I 0.28 exposed decaying wood 1.38 1.14 I i i 1 0.41 r 0.12 0.10 0.14 0.26 Figure 3.8. T r e e s and seed l ings on undisturbed forest floor by spec ies and height c l a s s (in percent) . Note that the y-axis d o e s not extend to zero . 100 T <50 50-99 100+ height class (cm) 67 Microtopography and Logging Slash T r e e s were not over- represented on mounds in c learcuts a s they were in old-growth s tands (Figure 3.9). O n flat s i tes, trees were over- represented on depress ion micros i tes and under - represented on s lope microsi tes (x2 > 21 .43; p < 0.001), but unrelated to mounds (x2 = 0.014; p > 0.90). Most t rees on depress ion microsi tes were C. nootkatensis and T. mertensiana. O n s teep si tes, trees were over- represented on s lope microsi tes and under-represented on mounds (x2 > 5.22; p < 0.025), but unrelated to depress ion micros i tes (x2 = 3.26; 0.05 < p < 0.10). On ly 1 2 % of ingress were located on mounds compared to 2 2 % of res iduals and germinants . It is unclear why ingress were unable to take advantage of these less-popu la ted micros i tes on either flat or s teep si tes, but it may have been related to the lower proport ion of undisturbed forest floor on mound (36%) than non-mound (49%) microsi tes. A s on old-growth s i tes, more trees growing on mounds were located downhi l l of the mound on s teep si tes (53%) than flat s i tes (31%; x2 = 4 .28; p = 0.039); few were located uphill of mounds on either s teep (6%) or flat (10%) s i tes. Fewer t rees than expec ted were located bes ide logging s lash (fallen logs) on flat s i tes (x2 = 31.59; p < 0.001), but on s teep s i tes there w a s no relat ionship (x2 = 1.85; p > 0.10). A m o n g trees that were growing bes ide logging s l a s h , more were located downhil l than uphil l: 4 5 % compared to 2 5 % on flat s i tes, and 6 3 % c o m p a r e d to 2 0 % on s teep si tes, respect ively. F igure 3.9. Observed- to -expec ted ratio of t rees by microtopographic c l ass on (a) c learcut s i tes, and (b) old-growth s i tes. C learcut data include trees >10 c m tall, but old-growth da ta include only understory t rees >130 c m tall. O flat s i tes; • s teep si tes. mound slope depression mound slope depression 68 Vaccinium Though at least some Vaccinium was found on 88% of clearcut and 90% of old-growth quadrats, it was taller and had greater cover on clearcut sites (0.70 m and 53%, respectively) than old-growth sites (0.51 m and 28%, respectively). Though not quantified, Vaccinium on clearcuts was also denser (had a higher leaf area index) than on old-growth sites, so microsites below Vaccinium were far darker. The percent cover of Vaccinium was higher on flat (59%) than steep (46%) sites, but heights were similar between slope types. Vaccinium cover was positively, though weakly, related to the cover of undisturbed forest floor (Pearson correlation on arcsine square root-transformed data; r = 0.54; p < 0.0001). Of all trees, 84% were on microsites with Vaccinium and 52% were shorter than the surrounding Vaccinium (Figure 3.10). A. amabilis was more common than expected on microsites with Vaccinium while other species, particularly T. heterophylla, were less common than expected, (x2 > 5.10; p < 0.025). Germinants were under-represented on Vaccinium microsites (x2 = 5.50; p < 0.025), though most (78%) were still associated with it. Vaccinium was unrelated to the location of either ingress (90%) or residuals (92%, x 2 < 0.73; p > 0.75). Figure 3.10. Height of trees relative to Vaccinium. A. amabilis C nootkatensis T. mertensiana T. heterophylla Did Vaccinium slow height growth or reduce vigour? A. amabilis grew fastest amidst but above Vaccinium (14.8 ±11.2 cm/year), less quickly without Vaccinium (7.9 ±8.3 cm/year), and slowest below Vaccinium (2.0 ±2.3 cm/year). This range of growth rates, however, was 69 related to a tree's absolute height (F = 183.6; p < 0.0001) rather than its height relative to Vaccinium (F = 0.734; p = 0.533), e.g. , shorter t rees grew more slowly regard less of Vaccinium ( A N O V A of natural log-transformed height increment covar ied with natural log- t ransformed height). Vaccinium w a s similarly unrelated to tree vigour a s trees growing on micros i tes with Vaccinium were equal ly v igorous as those growing without. There were no d e a d t rees or t rees likely to die within one year (vigour <2), regard less of Vaccinium cover . Dur ing the year of sampl ing , 3583 t rees/ha (34% of all trees) were f ree-growing accord ing to B . C . Ministry of Fores ts (1995) guidel ines, i.e., either growing without Vaccinium or >1.25 of Vaccinium height (Figure 3.11). With cont inued growth, an addit ional 2417 and 2 2 0 0 t rees /ha will be added after 10 and 20 years , respect ively. Al l t rees and seed l ings currently growing on the study si tes will be >1.25 the height of Vaccinium within 45 yea rs . Without addit ional ingress, the spec ies mix of f ree-growing trees will remain simi lar over the next 45 yea rs . A. amabilis and C. nootkatensis will increase from 43 to 4 7 % and 33 to 3 5 % , respect ively, whi le T. mertensiana and T. heterophylla will dec rease from 15 to 1 2 % and 9 to 6%, respect ively. F igure 3 .11 . Projected f requency and spec ies composi t ion of t rees >1.25 t imes the height of Vaccinium, the height cons idered by the B . C . Ministry of Fores ts (1995) to be f ree-growing. Height growth rates are based on the annual height increment of A. amabilis within each 10 c m height c lass , which were added iteratively to each tree's height. T h e year of sampl ing is year 0. I A. amabilis O C nootkatensis d T. mertensiana H T. heterophylla .-•••lllll 0 5 10 15 20 25 30 35 40 45 years after sampling 9000 I § CD J= CD 6000 3000 70 Stocking T h e B . C . Ministry of Fores ts (1995) a s s e s s e s post- logging regenerat ion in the M H z o n e in two s tages : (1) a stock ing survey determines whether there is an adequate distribution of we l l - spaced t rees that are ecological ly appropriate for the site; and (2) a f ree-growing survey 8 years later determines whether a min imum number of we l l - spaced t rees are >1.25 the height of non-crop vegetat ion or taller than a minimum height. 1 By 1992, 10-11 years after logging and one year before sampl ing , all s i tes had sat isf ied the min imum stock ing s tandard ( M S S ) of 500 s t ems /ha , though none w a s above the target stocking s tandard ( T S S ) of 900 s t e m s / h a ( B . C . Ministry of Fores ts 1992; Append ix C ) . S ince regenerat ion can be patchy even on s i tes that e x c e e d the M S S (Kl inka et al. 1992), I used a f iner-scale method for a s s e s s i n g its distribution in the year of sampl ing , then est imated growth rates to create a projected distribution 10 years later. The stocking percentages presented here are therefore not directly comparab le to Ministry of Forest calculat ions but rather indicate the relative pa tch iness of regenerat ion on the six s i tes. T h e legacy of taller residuals (>150 cm) on the six si tes limited d i f ferences in s tock ing at the t ime of sampl ing as stocking ranged from 1 8 % (MS) to 3 8 % (BF ; Figure 3.12). Di f ferences between s i tes, however , will increase after 10 years of cont inued growth due to di f ferences in the abundance of shorter t rees. Ave rage stocking on the three s i tes with more short t rees (BF , B S , and TF) will increase from 3 5 % to 6 3 % compared to an inc rease from 2 4 % to only 3 3 % on the three si tes with fewer short t rees ( M F , M S , and T S ) . 1Stocking guidelines for the MHmml variant (B.C. Ministry of Forests 1995) are: • Minimum heights by species: 0.60 m for A. amabilis ; 1.0 m for C. nootkatensis and 7. mertensiana; and (CWHvm2 variant only) 1.75 m for T. heterophylla. • Minimum height relative to neighboring vegetation (free-growing height): 1.25 times the height of neighboring vegetation within a 1 m radius. • Well-spaced guidelines: 2 m spacing (3.14 m diameter) between free-growing trees. 71 Figure 3.12. S tock ing during the year of sampl ing and projected s tock ing after 10 years . B lack bars indicate the p resence of at least one free-growing tree meet ing min imum height guidel ines ( B . C . Ministry of Forests 1995). S i tes include one horizontal t ransect (charts on left) and one vert ical transect (charts on right). Within each transect, s tock ing 0 years (upper) and 10 years (lower) after sampl ing are p resented . Percent s tock ing after 0 and 10 years is listed at the lower right of each site. Pro jected s tock ing is b a s e d on A. amabilis growth rates. 0 yrs 10yrs 0 yrs 10 yrs III II I Batchelor Flat (BF) II l l l l I I II I II Hill Mill III III III I III II m i l l i m u m i i site stocking after 0 and 10 years: 38%, 68% I I II II II Batchelor Steep (BS) site stocking after 0 and 10 years: 30%,, 64% O) _c Q. E a (0 i_ d) (B (0 i_ <TJ O 0 yrs 10 yrs 0 yrs 10 yrs 0 yrs 10 yrs Oyrs 10 yrs I II I II l l l l I i mini i i IIII i Tannis Flat (TF) I I HI M l Tannis Steep (TS) II III II II II 1IU II II Mayne Flat (MF) II I I M i l l I M i l I I i IIII urn i I I 1 site stocking after 0 and 10 years: 36%, 56% 1 II I I II II I site stocking after 0 and 10 years: 26%, 34% I I II I III site stocking after 0 and 10 years: 28%, 36% I H I II Mayne Steep (MS) site stocking after 0 and 10 years: 18%, 28% horizontal transect distance (2m quadrats) vertical transect distance (2m quadrats) 72 Th is growth of shorter t rees to free-growing height in the 10 years after samp l ing will reduce the number of t ree less patches, but mostly on the three more-popula ted s i tes (Figure 3.12). Frequent pa tches of >10 m will remain on the less-populated s i tes, whi le no patch >8 m is likely to remain on the more-populated s i tes. S o m e of the patches obv ious ly resul ted from d is turbance by logging a s shown by large accumula t ions of s lash and friable forest floor. S t u m p s of canopy trees (>40 cm base diameter) and sub-canopy trees (20-39.9 c m b a s e diameter) were found on s o m e t reeless patches which showed that these microsi tes suppor ted t rees in the former old-growth s tand. T h e addit ion of spac ing and min imum height requirements to f ree-growing requi rements (previous sect ion) favoured A. amabilis over other spec ies . At the t ime of samp l ing , A. amabilis compr ised 5 8 % of stocking compared to 4 3 % of f ree-growing t rees. S tock ing contr ibuted by other spec ies included C. nootkatensis (27%), T. mertensiana (14%), and T. heterophylla (1%). After 10 years , the proportion of A. amabilis will d e c r e a s e slightly to 5 0 % of s tock ing and C. nootkatensis will increase to 3 4 % . Proport ions of T. mertensiana (13%) and T. heterophylla (3%) will remain approximately the s a m e . Seed l i ngs were unequal ly distributed a m o n g the 600 quadrats a s 11 c lumps of >2 individuals conta ined 53 of 74 seed l ings . S ince no relationship between seed l ing locat ion and microsi te factors could be identif ied, the likely c a u s e of c lumping w a s the distribution of s e e d s . S u c h infrequent and c lumped ingress means that current t ree less patches are unlikely to be fi l led by seed ing from outside the s i tes. Growth Form Anomalies T r e e s >1.3 m tall on clearcut si tes were more likely to be chlorotic or c rushed by logging s l ash and less likely to d isplay umbrel la or s tem s w e e p growth forms than understory t rees on old-growth s i tes (Table 3.4). The inc idence of ch loros is , however , w a s still low. Of 20 chlorot ic t rees, 18 were A. amabilis and 14 of these were located on the flat, sou th -aspec t s i tes, T F and M F . The annual height growth of chlorotic t rees s lowed to 5 4 % that of other t rees 73 after approx imate ly 6-7 years of strong re lease, but still grew almost 5 % per year (mean relative height growth rate = 0.05 ±0.03 cm/cm/yr) . Ch lo ros is w a s unrelated to Vaccinium or subst rate. T r e e s growing amidst but shorter than Vaccinium were slightly less likely to have growth form anoma l ies than trees taller than Vaccinium or on microsi tes without Vaccinium. Pistol-butted trees were slightly more c o m m o n on clearcut than old-growth s i tes but were less seve re : only 4 % of pistol butts had a horizontal d is tance >50 c m . Tab le 3.4. Growth form anomal ies by spec ies (expressed as a ratio of t rees with anoma l ies to those without). Old-growth data include only understory t rees >130 c m tall. T o make compar i sons easier , c learcut data are a lso limited to t rees >130 cm tall. Growth form anoma l ies are def ined in Append ix B. S p e c i e s on clearcut s i tes total for Growth form Abies C. nootka- T. mert- T. hetero- old-growth anoma ly amabilis tensis ensiana phylla total s i tes B roken top 0.05 0.05 0.00 0.14 0.05 0.06 Chloro t ic 0.31 0.11 0.00 0.00 0.21 0 .03 C r u s h e d 0.14 0.16 0.25 0.14 0.16 0.01 D a m a g e d leader 0.03 0.00 0.00 0.14 0.03 0.08 D o g leg 0.12 0.11 0.17 0.14 0.13 0.16 Mult iple leaders 0.02 0.37 0.00 0.00 0.08 0.06 Pisto l butt 0 .83 0.63 0.92 0.71 0.79 0.62 Prostrate 0.00 0.00 0.00 0.00 0.00 0.01 Sh rub (bush) 0.00 0.00 0.00 0.00 0.00 0.00 S t e m s w e e p 0.07 0.32 0.00 0.14 0.11 0.34 Umbre l la 0.00 0.05 0.00 0.00 0.01 0.52 S a m p l e s i ze (n) 58 19 12 7 96 6 4 5 Overa l l ratio 1.57 1.79 1.33 1.43 1.57 1.90 74 3.4 Discussion Was natural regeneration successful? T h e determinat ion of regenerat ion s u c c e s s depends on the criteria used . Acco rd ing to Ministry of Fores ts criteria, the study si tes are success fu l l y regenerated: there are not only enough t rees to produce a future crop, there are s o many that future thinning will be needed ( B . C . Ministry of Fores ts 1992; Append ix C ) . There were a number of conce rns identif ied in this study, however , especia l ly regarding the spec ies mix and structure of the future s tand , a s wel l a s the distribution of regenerat ion. Natural regenerat ion w a s unsuccess fu l from the standpoint of maintaining the s a m e proport ion of spec ies that w a s in the previous old-growth s tand. Clearcut t ing in the M H z o n e usual ly results in a spec ies shift to A. amabilis due to its abundance a s advance regenerat ion in old-growth s tands (Kl inka etal. 1992; Koppenaa l and Mitchel l 1992; Arnott etal. 1995; Chap te r 2). T h e strong component of C. nootkatensis w a s less expec ted and w a s due to st rong estab l ishment in the years just before and after logging. There w a s likely a large bank of germinants in the previous old-growth s tand (Chapter 2), and this may have been bols tered by a heavy s e e d year that co inc ided with the year of logging. Whether the strong p resence of C. nootkatensis will remain throughout s tand deve lopment is unclear s ince there is s o m e doubt about its ability to survive into upper canopy layers under low light condi t ions (Kl inka et al. 1992). T h e minor p resence of T. mertensiana contrasts with its dominance a m o n g canopy t rees in adjacent old-growth s tands (Chapter 2) and w a s likely due to a lack of s e e d s . D o e s the d u m p i n e s s of t rees and the many t reeless pa tches , espec ia l ly on the three less-popu la ted s i tes, constitute unsuccess fu l regenerat ion? It is unreal ist ic (and probably undesi rable) to expect full site occupancy s ince t reeless patches (gaps) are c o m m o n in the M H z o n e (Ler tzman etal. 1996; Chapte r 2), especia l ly on flat s i tes. However , 2 of the 3 s i tes that had many t ree less patches were s teep and many of those patches conta ined s tumps which s h o w e d that they had been occup ied in the previous s tand. 75 Regenera t ion w a s most success fu l on flat and north-facing s i tes likely a s a result of lower mortality from heat and moisture s t ress compared to south- facing s i tes, espec ia l ly those that are s teep (Reuter 1973; Se ide l and Coo ley 1974; Bal lard etal. 1977; E m m i n g h a m and Ha lverson 1981). E v e n within the relatively cool and wet M H zone , the removal of the forest canopy can result in lethal condi t ions, particularly when it a lso c a u s e s faster snowmel t (Reuter 1974; pers . obs. ) . B e s i d e s being relatively drier and warmer than flat s i tes, s teep s i tes had more disturbed (friable) forest floor and were more likely to be covered by logging s l ash and e x p o s e d rock. T h e combinat ion of higher temperatures and unsui table, easi ly dried-out subst ra tes apparent ly reduced survival and/or establ ishment on these s i tes. Mois ture s t ress may have affected spec ies composi t ion a s wel l . A. amabilis and 7". mertensiana were less c o m m o n on south-facing si tes than were C. nootkatensis and T. heterophylla. A. amabilis is known to have high moisture requirements (Herring and Ether idge 1976; Kotar 1977, 1978; Kraj ina e r a / . 1982) and partial sunlight near s tand edges improves survival and growth for both A. amabilis (Herring and Ether idge 1976; W a g n e r 1980; van Pel t 1995) and T. mertensiana (Burns and Honka la 1990). The re is a concern that snow can ser iously d a m a g e regenerat ion in the M H z o n e , espec ia l l y C. nootkatensis and T. mertensiana (Scage l etal. 1989; K l inka etal. 1992; Arnott et al. 1995). I bel ieve there are three reasons why such d a m a g e w a s not evident here: (1) natural ly- regenerated trees are less suscept ib le to d a m a g e than planted trees (Arnott et al. 1995); (2) most regenerat ion w a s short enough to be protected under the winter s n o w p a c k and not yet subject to inc reased d a m a g e with growth above the snowpack (Arnott etal. 1995; R. G r e e n , pers . comm.) ; and (3) the location of the study si tes at the lowest e levat ions of the M H z o n e meant snow loads were not great enough to cause ser ious d a m a g e . Did most natural regeneration establish before or after logging? Whi le the importance of residuals in regenerat ing s i tes has been previously reported (e.g., Herr ing and Ether idge 1976; Vogt etal. 1989; G r e e n and Bernardy 1991 ; K l inka etal. 76 1992), the role of germinants has not, likely b e c a u s e such studies limited their focus to two age c l a s s e s : pre- and post- logging. In this study, germinants establ ishing within a 3-year w indow from one year before logging to one year after logging accounted for a lmost half of all regenerat ion. Separa t ing this age c lass from post- logging regenerat ion highlighted the lack of ingress that might have remained hidden otherwise. S o m e might argue that they are the result of post- logging growing condi t ions, but I bel ieve that there is a lso an aspec t of legacy in that most were from the seed l ing or germinant bank (sensu K o h y a m a 1983) of the prev ious o ld-growth s tand and most germinated on its undisturbed forest floor, even if their subsequen t growth w a s in a clearcut. T h e relative unimportance of ingress w a s surpr is ing, especia l ly the lack of ingress more than a few years after logging. S ince ingress can be an important source of regenerat ion (Minore and Dubras ich 1981; Arnott etal. 1995), its scarci ty here indicates that the availabil i ty of s e e d s w a s inadequate or that there were problems with germinat ion and surv iva l . T h e rapid d e c r e a s e in establ ishment after logging sugges ts that off-site s e e d s were never important contr ibutors to regenerat ion and that the dec rease co inc ided with the deplet ion of the s e e d bank (Kl inka and Pend l 1976) Ingress w a s clear ly not limited only by the lack of growing s p a c e (sensu O l iver and La rson 1990) s ince there were numerous patches where no t rees and virtually no other vegetat ion were growing, especia l ly on stump mounds . Instead, the main limitation to ingress w a s a lmost certainly the a b s e n c e of nearby, seed-produc ing t rees. The c losest remain ing canopy t rees to any study location were in a narrow fringe, 1-2 trees deep , at a lake edge - 1 0 0 - 2 0 0 m from the Batchelor study locat ion. The canopy trees c losest to the M a y n e and Tann is study locat ions were - 2 0 0 - 1 4 0 0 m distant. S ince s e e d densi t ies of the native s p e c i e s d e c r e a s e rapidly away from a stand edge (Franklin and Smith 1974; Cark in etal. 1978; Burns and Honka la 1990), there would be few s e e d s avai lable on the study locat ions. Further ingress will therefore be de layed until regenerat ing trees reach sexua l maturity. T h e relative lack of T. 77 mertensiana compared to adjacent old-growth s tands w a s likely a lso c a u s e d by the unavailabi l i ty of s e e d s s ince it can v igorously co lon ize logged a reas when there is adequate s e e d (Frankl in and Smi th 1974; Se ide l and Coo ley 1974). Which substrates favoured natural regeneration? A s d i s c u s s e d in Chapte r 2, success fu l regenerat ion in other sys tems is common l y related to disturbed substrates, especia l ly mineral soi l , decay ing wood , and coa rse woody debr is (e.g., Vogt 1989; Spi t t lehouse and Stathers 1990; C o a t e s etal. 1991). In contrast , most regenerat ion in both clearcut and old-growth si tes of this study, and c learcut s i tes of at least one other study in the M H zone (Green and Bernardy 1991), w a s on undisturbed forest floor. T h e relat ionship between undisturbed microsi tes and residuals can be partially exp la ined by higher surv ival rates during logging, but the overwhelming majority of germinants and ingress were a lso on undisturbed forest floor. Poo r regenerat ion on friable forest floor has been noted in other s tudies and exp la ined as the result of the adaptat ion of native spec ies , particularly A. amabilis, to regenerat ion on thick, matted Mor humus forms (Green and Bernardy 1991; K l inka etal. 1992; Chap te r 2). T h e transit ion to a fr iable Mormoder or Moder humus form may a lso have inc reased the porosity of the forest f loor and , together with the lack of shad ing on these microsi tes, resul ted in greater moisture s t ress . No t rees were growing on the many fal len logs left after logging and t ree less pa tches were especia l ly evident where logging s lash had been pi led near landings. T h e bark w a s still intact on many of these fal len trees and , based on observat ions from adjacent o ld-growth s tands (Chapter 2), it may be centur ies before they have d e c a y e d enough to be appropr iate s e e d b e d s for regenerat ion. In agreement with warn ings against ex t reme soi l d is turbance (Green and Bernardy 1991; K l inka etal. 1992; Banner etal. 1993), t ree less pa tches on adjacent s i tes were often assoc ia ted with the exposed mineral soi l on sk idder trails. 78 Which microtopographic locations favoured natural regeneration? T h e neutral or negat ive assoc ia t ion between mounds and natural regenerat ion cont rasted with that s e e n in old-growth s tands (Chapter 2) and w a s likely c a u s e d by logging. Logg ing may have changed the relationship between regenerat ion patterns and mounds in at least three ways . First, direct and indirect mortality from logging would be highest on mounds s ince they suppor ted a lmost all canopy trees in the previous old-growth s tand. S e c o n d , forest f loor d is turbance w a s greatest on mounds and this d is turbance apparent ly reduced the ability of s e e d s to germinate. Third, t rees establ ishing on mounds after logging might not have had the s a m e microcl imat ic advantage as trees on mounds in old-growth s tands. S n o w mel ted severa l w e e k s earl ier on clearcut than old-growth si tes (unpubl. data) s o the advan tage of the longer growing s e a s o n on mounds would be less and , without the shad ing of canopy t rees, the moisture s t ress would be greater. Th is is not to say , however , that microtopography will not affect regenerat ion patterns in the future. In old-growth s tands, mounds were assoc ia ted more with greater surv ival than establ ishment so , as the canopy c loses on these c learcuts , they may yet affect s tand deve lopment . If s n o w w a s less of a factor on clearcut than old-growth s i tes, why w a s there a simi lar propensi ty for t rees on mounds to be on the downhil l s ide , not to ment ion a tendency for t rees to be on the downhil l s ide of logging s lash? A n obv ious explanat ion for these patterns is that they were s imply artifacts from the previous old-growth s tand, but this would not expla in why t rees downhil l of logging s lash surv ived better. S n o w creep may have been a factor, but it is a lso poss ib le that b e c a u s e uphill microsi tes had less snow, there w a s more d a m a g e f rom frost and ice. Simi lar ly, uphill microsi tes would be subject to greater moisture s t ress dur ing the growing s e a s o n and this may a lso have reduced survival . Final ly, soil c reep limited the height growth and likely the survival of seed l ings on old-growth si tes (Chapter 2), s o it cannot be ruled out a s a factor on clearcut s i tes, especia l ly s ince the forest floor d is turbance and s lash c a u s e d by logging inc reased the material prone to moving downhi l l . 79 Did competition with Vaccinium impede natural regeneration? Compet i t ion can be def ined as : "...the negative effects which one organ ism has upon another by consuming , or controll ing a c c e s s to, a resource that is limited in availabil i ty" (Keddy 1989, p.2). Sh rubs and herbs can compete in this s e n s e with t rees for such resources a s light, water, and nutrients (e.g., Ol iver and Larson 1990; Burton 1993; G r e e n and Kl inka 1994). T h e negat ive impacts of e r i caceous shrubs on tree survival and growth are we l l -documented (e.g., R a d o s e v i c h 1984; M e s s i e r and K immins 1990; de Mont igny 1992; G . W e e t m a n in Koppenaa l and Mitchel l 1992; Mall ik 1995). L e s s reported is a posit ive relat ionship, or facil itation (Connel l and Sla tyer 1977; Berkowi tz etal. 1995), between trees and shrubs such a s Vaccinium. Vaccinium can protect t rees from sun-sca ld (Kotar 1978; Minore 1986) a s well a s frost d a m a g e and s n o w p ress (Scage l etal. 1989). Natural regenerat ion is a lso more success fu l on those c learcut s i tes where Vaccinium w a s present in the previous old-growth s tand (Green and Bernardy 1991). How then did Vaccinium affect natural regenerat ion in this s tudy? M y results show no ev idence that Vaccinium impeded regenerat ion s ince most regenerat ion, regard less of age c lass , was assoc ia ted with it. O n e surpr ise w a s that over half of all regenerat ion w a s below the Vaccinium canopy and therefore only not iceable under c lose examinat ion. A s predicted by Kl inka etal. (1992), this bank of otherwise invisible regenerat ion shou ld cont inue to add to site s tock ing as it eventual ly overtops the Vaccinium. T h e assoc ia t ion of res iduals, Vaccinium, and undisturbed forest floor is further ev idence that lack of logging d is turbance inc reased the survival of advance regenerat ion from the prev ious o ld-growth s tand. In addit ion, microsi tes with Vaccinium microsi tes suppor ted the majority of ingress wh ich s h o w e d that they provided more favourable condit ions for tree estab l ishment than micros i tes without Vaccinium. Chlo ros i s and growth check have been reported on M H si tes with abundant Vaccinium and an undisturbed, compac ted , and thick forest floor (G . W e e t m a n in Koppenaa l and Mitchel l 1992; K l inka etal. 1992). However , not only were ch loros is and growth check u n c o m m o n on the study s i tes, they were a lso unrelated to substrate or Vaccinium cover , and limited a lmost 80 exc lus ive ly to A. amabilis on flat, south-aspect s i tes, notably the T F site. S ince healthy C. nootkatensis ou tnumbered chlorotic A. amabilis on this site, ch loros is w a s likely c a u s e d by site rather than microsi te factors. For example , A. amabilis may be less wel l -adapted than C. nootkatensis to the ex t remes in soi l moisture and temperature on a flat, warm-aspec t site that is no longer s h a d e d by a forest canopy. 3.5 Conclusions M a n y of the regenerat ion problems identified here were predictable c o n s e q u e n c e s of ex tens ive clearcutt ing at lower elevat ions of the M H zone . The shift to A. amabilis, the lack of ingress and structural diversity, and the c lumped regenerat ion pattern have all been reported before. Ye t despi te cal ls for al ternat ives spann ing more than 30 years (Frankl in 1964; Brooke etal. 1970; Reuter 1973; Utzig and Herr ing 1974, Vogt etal. 1989; Banner etal. 1993), tradit ional c learcutt ing remains the dominant cutting method in the M H zone . C o s t -ef fect iveness is undoubtedly an important reason for the overwhelming use of c learcutt ing, but it is a lso b a s e d on the premise that max imum growth rates will be ach ieved with the total removal of the canopy layer. In the M H zone , however, tree establ ishment , surv iva l , and growth form are likely more important considerat ions than growth rates (Scage l etal. 1 9 8 9 ; . K l inka etal. 1992; C h a p t e r 2 ) . Clearcut t ing d o e s not take advantage of the many non-canopy t rees a l ready es tab l ished in an old-growth stand that, if p reserved, cou ld jump-start regenerat ion and retain a better s p e c i e s mix and more structural diversity. Nor is it intended to leave a sou rce of s e e d s to ensure a cont inuous ingress of regenerat ion. A s an example of a different app roach to regenerat ing s i tes in the M H zone , a partial cut near the study s i tes (on the C h a p m a n Plateau) left a variety of snags , live but low vigour canopy t rees, and much of the sub -canopy layer. A s a result, there is much higher structural diversity, more abundant C. nootkatensis and T. mertensiana, and greater ingress than on any of the study s i tes. In spite of the h ighgrading of 81 canopy t rees and excess i ve d isturbance c a u s e d by sk id trails, regenerat ion is v igorous and the site retains many character ist ics of the previous old-growth s tand known to be required by wildlife (Hopwood 1991). Retain ing taller advance regenerat ion would a lso reduce rotation a g e s and increase y ie lds. For example , one study of advance regenerat ion in Q u e b e c found that an inc rease of 3 m in the height of advance regenerat ion is equivalent to an inc rease of 3 m in site index (Pothier e r a / . 1995). In old-growth s tands adjacent to these c learcuts, there w a s a pool of >300 T. mertensiana and C. nootkatensis understory and sub-canopy trees per ha that were >2 m tall and had good vigour (i.e., v igour >3; Chapter 2), yet no such trees surv ived logging on these c learcuts . Preserv ing taller t rees would have been especia l ly advan tageous on the three less -populated s i tes where ingress w a s so poor. It is especia l ly striking to note that 7 % of t ree less pa tches conta ined the s tumps of sub-canopy trees (i.e., 20-40 cm base diameter) f rom the prev ious old-growth s tand which need not have been cut. Preserv ing advance regenerat ion is known to improve s u c c e s s after cutting in the M H z o n e (e.g., Frankl in 1964; K l inka and Pend l 1976; E m m i n g h a m and Ha lverson 1981 ; Bu rns and Honka la 1990; G r e e n and Bernardy 1991; K l inka etal. 1992; Banner etal. 1993) but the preservat ion of taller t rees is actively restricted by current pol icy. Not only have safety conce rns led to a 3 m "knockdown" rule, where all t rees >3 m tall must be cut, but current s tock ing s tandards only accept advance regenerat ion shorter than 1.0 m (C . nootkatensis and T. mertensiana) or 1.5 m (A. amabilis; B . C . Ministry of Fores ts 1995). Th is preference for short advance regenerat ion apparent ly s tems from Herr ing and Ether idge 's (1976) study that eva luated A. amabilis advance regenerat ion on 11 c learcuts s i tes located in the C W H z o n e . T h e authors r ecommended against preserv ing advance regenerat ion >2 m tall due to an inc reased risk of d a m a g e during close-ut i l izat ion logging, i.e., where all s tems over a min imum s ize must be removed from the site. 82 There are severa l reasons to revisit the lack of accep tance for taller a d v a n c e regenerat ion in the M H zone . Close-ut i l izat ion clearcutt ing is inappropriate on many si tes in the M H z o n e for the reasons given above (and in Chapte r 2), and a lso b e c a u s e c lose util ization inc reases the inc idence of d a m a g e from logging (Herring and Ether idge 1976). Though Herr ing and Ether idge show that d a m a g e is more frequently incurred by taller A. amabilis, they found a lmost no inc idence of decay result ing from damage and , in fact, noted that d a m a g e w a s a lmost unnot iceable in many trees after 20-30 years . Part of the bias against taller regenerat ion may be b a s e d on a fear that it is not able to surv ive the abrupt change in growing condi t ions after logging. But the s u c c e s s of taller regenerat ion on the partial cut desc r ibed above s h o w s that taller t rees can thrive, at least where s o m e canopy cover is left intact. A n d finally, the interdictions against taller advance regenerat ion were implemented at a t ime when conce rns for spec ies and structural diversity were not a s important as they are today (Hopwood 1991). K l inka etal. (1992) l inked chlorosis and growth check to the tying-up of nutrients in thick and compac ted Mor humus forms, especia l ly where decay ing wood is a major component . They therefore sugges ted that s o m e dis turbance of the forest floor wou ld improve nutrient availabil i ty. However , growth check is not a major problem here and any advan tages of greater nutrient availabil ity would likely be outweighed by the loss of res iduals with greater d is turbance during logging. Both Chap te r 2 and Chapte r 3 showed the great variability of regenerat ion patterns a m o n g adjacent s i tes. Th is variability s t resses the need for all aspec ts of managemen t to be s i te-speci f ic , including B . C . Ministry of Forests surveys . For example , su rveys on these s i tes wou ld have prov ided a better a s s e s s m e n t of regenerat ion if s i tes were properly stratif ied by edaph ic condi t ions. Instead, the survey a reas ranged in s ize from 29-82 ha and s p a n n e d a large edaph ic range. T h e survey that conta ined the two Tann is s i tes, for examp le , inc luded south and north aspec ts , flat and steep s lopes , and ridge tops and bogs . A s a result, it cou ld 83 not differentiate the reasonably good regenerat ion on the T F site from the poor regenerat ion on the T S site. Chap te r 2 a lso showed that there is no abrupt eco log ica l boundary be tween the two Tsuga s p e c i e s as implied by the current stocking s tandards which al low only one s p e c i e s of Tsuga on s i tes in the study a rea : T. heterophylla on C W H v m 2 si tes (except on one very moist site se r ies at higher elevat ions) and T. mertensiana on M H m m l si tes ( B . C . Ministry of Fores ts 1995). T h e accep tance of T. mertensiana just be low the C W H / M H boundary might reduce growth rates, but would a lso reduce the risk of d a m a g e from frost and snow. Al l s tud ies of post- logging regenerat ion in the M H zone , including this one , have been at its lowest limits. Regenera t ion s u c c e s s and growth rates will only dec rease a s higher e levat ion s i tes are logged (Kl inka etal. 1992; Banner etal. 1993; K l inka 1996; B. Sp lech tna unpubl . data). Chapte r 2 showed that the forested M H subzone occup ies a transit ion be tween gap and t ree- is land sys tems and it is likely that regenerat ion b e c o m e s unfeasib le s o m e w h e r e within this transit ion (Kl inka etal. 1992). 84 Chapter 4. Summary and Conclusions Chap te r 2 desc r ibes regenerat ion patterns within old-growth s tands that are transit ional be tween lower-elevat ion, gap-dr iven forests and tree is lands at higher e levat ions. Cons is ten t with predict ions of the tree- is land mode l , regenerat ion in these s tands is most success fu l c l ose to a canopy tree and on mounds rather than in canopy gaps . Late-melt ing snow inc reases the p reva lence of regenerat ion patterns matching the tree- is land mode l . Though regenerat ion w a s present in canopy gaps , the formation of gaps does not drive regenerat ion patterns a s at lower e levat ions where s n o w is infrequent. Whi le Chap te r 2 demonst ra tes that microsi tes and snow are related to regenerat ion patterns in old-growth s tands, Chapte r 3 shows that the pattern of natural regenerat ion after c learcutt ing is most ly an artifact of residuals and germinants from the prev ious old-growth s tand . Mos t natural regenerat ion is thus related to microsi tes undisturbed by logging, i.e., those with an intact forest floor and abundant Vaccinium cover . It a rgues that c lass i fy ing the age of natural regenerat ion a s either pre- or post- logging m a s k s the fact that post- logging ingress from s e e d s originating off-site w a s negligible. Though we tend to picture biogeocl imat ic units as discrete l ines on a map , the reality is much different. Within the study a rea , for example , t ree- is land e c o s y s t e m s that are simi lar to those in the park land M H subzone in both vegetat ion and phys iognomy inhabit flat s i tes a s low a s 1000 m, whi le forest s tands dominated by T. heterophylla can be on s teep s i tes at e levat ions a s high a s 1150 m (Brett 1996). Though such variabil ity d o e s not affect forest p lanning on a broad sca le , it must be add ressed when deal ing at the s tand level , part icularly within a transit ional a rea such as the forested M H subzone . O u r current pol icy v iews the forested M H subzone a s an extension of lower-elevat ion forests, albeit with different spec ies . A s such , much of the subzone has been c learcut or is schedu led for clearcutt ing in the near future. But the high non-t imber va lues in these forests, a s wel l a s their s low growth, a rgues against their automatic inclusion as sou rces of t imber. If 8 5 the dec is ion to cut is made , methods should be heavi ly modif ied. S ince regenerat ion requires the protect ion from snow provided by an overhead canopy, any cutting at lower e levat ions of the M H z o n e shou ld retain a s much of the canopy and sub-canopy layers a s poss ib le . At higher e levat ions and on late-snowmelt s i tes, any cutting is inappropriate. Chap te r 2 conc ludes with a suggest ion to use regenerat ion patterns a s an eco log ica l bas i s for the managemen t c l a s s e s first descr ibed by Kl inka etal. (1992). A s the cutting of M H s tands p rog resses up to elevat ions where tree- is land p r o c e s s e s dominate, it is t ime to revisit the w i sdom of cutting these forests, especia l ly by clearcutt ing. 86 Literature Cited Antos , J . A . and D.B. Zobe l . 1986. Habitat relat ionships of Chamaecyparis nootkatensis in southern Wash ing ton , O regon , and Cal i forn ia. C a n . J . Bot. 64(9): 1898-1909 . Arnott, J .T . , R .K . S c a g e l , R . C . E v a n s , and F.T. Pend l . 1995. High elevat ion regenerat ion strategies for subalp ine and montane forests of coasta l Brit ish Co lumb ia . C a n . For . Serv . and B . C . M in . For. , Victor ia, B . C . F R D A R e p . No . 229. 30 pp. Arsenau l t , A . 1995. Pattern and p rocess in old-growth temperate rainforests of southern Brit ish C o l u m b i a . P h . D . thesis, Univ. Brit. Co lumb ia . 186 pp. B . C . Ministry of Fores ts . 1992. Regenerat ion surveys for cutb locks 150 and 57. Seche l t F ie ld Off ice, Seche l t , B . C . B . C . Ministry of Fores ts . 1995. Estab l ishment to free growing guidebook, V a n c o u v e r Forest Reg ion . B . C . M in . For. and B . C . Env. , Victor ia, B . C . 130 pp. B . C . Ministry of the Envi ronment. 1985. S u m m a r y of snow survey measu remen ts in Brit ish C o l u m b i a 1935-1985. B . C . M in . Env. , Water M a n a g e . B ranch , Victor ia, B . C . B . C . Ministry of the Envi ronment . 1993, 1994, and 1995. S n o w survey bulletin. Month ly reports for Ma rch (inc. J a n . and Feb. ) , Apr i l , May , and J u n e . Water M a n a g e . B ranch , B . C . M in . Env . , Lands , and P a r k s (before 1994) and B . C . M in . Env . (1994-95), V ic tor ia , B C . Ba l la rd , T . M . , T .A . B lack, and K . G . McNaugh ton . 1977. S u m m e r energy ba lance and temperatures in a forest c learcut in southeastern Brit ish C o l u m b i a . In Brit. C o l u m b i a So i l S c i . W o r k s h o p Report , Meet ing No . 6, B . C . M in . A g r i c , Victor ia, B . C . pp. 74-86 . Banner , B., W . M a c K e n z i e , S . Haeuss le r , S . T h o m s o n , J . Pojar, and R. Trowbr idge. 1993 . A field guide to site identification and interpretation for the Pr ince Rupert Forest Reg ion . B . C . M in . For . R e s . P rog . , Land M a n a g e . Handbook No . 26. Barbour , M . G . , N .H . Be rg , T . G . F . Kittel, and M . E . Kunz . 1991. S n o w p a c k and the distribution of a major vegetat ion ecotone in the S ier ra N e v a d a of Cal i forn ia. J . B i ogeog . 18: 141-149. Beatty, S . W . 1984. Influence of microtopography and canopy spec ies on spat ial patterns of forest understory plants. Eco logy 65(5): 1406-1419. Berkowi tz , A . R . , C D . C a n h a m , and V . R . Kel ly. 1995. Compet i t ion vs . facil itation of tree seed l ing growth and survival in early success iona l communi t ies . Eco logy 76(4): 1156-1168. Berry, G . L . and R.L. Rothwel l . 1992. S n o w ablat ion in smal l forest open ings in southwest A lber ta . C a n . J . For. R e s . 22 : 1326-1331. 87 Brett, R . B . 1996. Del ineat ion of M H forested and park land s u b z o n e s in the C h a p m a n A s s e s s m e n t Unit. Contract report for the V a n . For. Reg ion , B . C . M in . For. , Nana imo , B . C . 1 1 p p . Brink, V . C . 1959. A directional change in the subalp ine forest-heath ecotone in Gar iba ld i Park , Brit ish Co lumb ia . Eco logy 40 : 10-16. Brink, V . C . 1964. Plant establ ishment in the high snowfal l a lpine and subalp ine regions of Brit ish Co lumb ia . Eco logy 45(3): 431-438. B rooke , R . C . 1966. Vegetat ion-envi ronment relat ionships of subalp ine Mounta in Hemlock zone e c o s y s t e m s . P h . D . thesis, University of British Co lumb ia , Vancouve r , B . C . 225 pp, A p p . 110 pp. B rooke , R . C , E . B . Pe te rson , and V . J . Kraj ina. 1970. The subalp ine Mounta in Hemlock zone . E c o l . Wes te rn N. Amer . 2: 148-349. Burns , R . M . and B . H . Honka la . 1990. Si lv ics of North Amer i ca . Vo l . 1 Con i fe rs . U S D A For. Se rv . Agr i . Handbook 445 , Wash ing ton , D .C . 675 pp. Bur ton, P . J . 1993. S o m e limitations inherent to static indices of plant compet i t ion. C a n . J . For . R e s . 23 : 2141-2152 . C a n h a m , C D . , J . S . Dens low, W . J . Piatt, J . R . Runk le , T .A . S p i e s and P . S . Whi te . 1990. Light reg imes beneath c losed canop ies and tree-fall gaps in temperate and tropical forests. C a n . J . For . R e s . 20: 620 -631 . Ca rk i n , R . E . , J . F . Frankl in, J . Booth , and C E . Smi th . 1978. Seed ing habits of upper -s lope tree spec ies . IV. S e e d flight of noble fir and Paci f ic si lver fir. U S D A For. Serv . R e s . Note P N W - 3 1 2 , Paci f ic Northwest Forest and Range Exper iment Stat ion, Por t land, O R . 10 pp. Car ter , R . E . and K. K l inka. 1992. Variat ion in shade to lerance of Douglas-f i r , western hemlock, and western redcedar in coasta l British Co lumb ia . For. Eco l . M a n a g e . 55 : 87 -105 . Chr is ty , E . J . and R . N . Mack . 1984. Variat ion in demography of juveni le Tsuga heterophylla a c r o s s the substratum mosa ic . Journa l of Eco logy 72 : 75 -91 . C o a t e s , K .D. , W . H . E m m i n g h a m , and S . R . Radosev i ch . 1991. Con i fer -seed l ing s u c c e s s and microcl imate at different levels of herb and shrub cover in a Rhododendron -Vaccinium - Menziesia communi ty of south central Brit ish C o l u m b i a . C a n . J . For . R e s . 2 1 : 858-866 . Conne l l , J . H . and R . O . Slatyer. 1977. M e c h a n i s m s of s u c c e s s i o n in natural communi t ies and their role in communi ty stability and organizat ion. Amer . Natur. 111 : 1119-1144. Dan ie ls , L .D. 1994. Structure and regenerat ion of old growth Thuja plicata s tands near Vancouve r , Brit ish Co lumb ia . M . S c . thesis, Univ. Brit. C o l u m b i a . 98 pp. 88 de Mont igny, L . E . 1992. A n investigation into the factors contributing to the growth-check of conifer regenerat ion on northern Vancouve r Island. P h . D . thes is . Univ. Brit ish C o l u m b i a , Vancouver , B . C . 191 pp. E m m i n g h a m , W . H . and N . M . Ha lverson. 1981. Commun i t y types, productivity, and reforestat ion: management implicat ions for the Paci f ic Si lver Fir Z o n e of the C a s c a d e Mounta ins . In C D . Ol iver and R . M . Kenady (eds.). P roceed ings of the biology and managemen t of true fir in the Paci f ic Northwest sympos ium. Contr ibut ion 45 , Universi ty of Wash ing ton Co l lege of Forest R e s o u r c e s , Seatt le, W A . pp. 291-303 . F o n d a , R .W. and L . C Bl iss . 1969. Forest vegetat ion of the Montane and Suba lp ine Z o n e s , O lymp ic Mounta ins , Wash ing ton . Eco l . Monog . 39: 271 -301 . Frankl in , J . F . 1964. Eco logy and silviculture of the true f i r-hemlock forests of the Pac i f ic Northwest . P roc . S o c . A m . Foresters 1964: 28-32. Frankl in , J . F . , W . H . Moir, G . W . Doug las , and C . Wiberg . 1971. Invasion of suba lp ine m e a d o w s by t rees in the C a s c a d e R a n g e , Wash ing ton and Oregon . Arct ic and Alp ine R e s . 3(3), pp. 215-224 . Frank l in , J . F . and C E . Smi th . 1974. Seed ing habits of upper-s lope tree s p e c i e s . II. D ispersa l of a mountain hemlock seedc rop on a clearcut. U S D A For. Serv . R e s . Note P N W - 2 1 4 , Pac i f i c Northwest Forest and Range Exper iment Stat ion, Por t land, O R . 9 pp.. Frankl in , J . F . and C T . Dy rness . 1988. Natural vegetat ion of O regon and Wash ing ton . O r e g o n State Univ. P r e s s , Corval l is , O R . 452 pp. G lan tz , S . A . 1992. Pr imer of Biostat ist ics. Third Edit ion. McGraw-H i l l , N e w York , N Y . 440 pp. Go ld ing , D.L. and R . H . S w a n s o n . 1978. S n o w accumulat ion and melt in smal l forest open ings in A lber ta . C a n . J . For. R e s . 8(4): 380-388. Go ld ing , D.L. and R . H . S w a n s o n . 1986. S n o w distribution patterns in c lear ings and the adjacent forest. Water R e s o u r c e s R e s e a r c h , 22(13): 1931-1940. G r e e n , R . N . and P. Bernardy. 1991. Natural regenerat ion failure at high e levat ions in the Ch ipmunk C r e e k dra inage, Chi l l iwack Forest District. R e s . M e m o 59, B . C . M in . For . V ic tor ia , B . C . 3 pp. G r e e n , R . N . , R.L . Trowbr idge, and K. Kl inka. 1993. Towards a taxonomic c lassi f icat ion of humus forms. For . S c i . M o n o g . 29 (supplement to For. S c i . 39). G r e e n , R . N . and K. K l inka. 1994. A field guide to site identif ication and interpretation for the V a n c o u v e r Forest Reg ion . B . C . Min . For. , Victor ia, B . C . 285 pp. G rubb , P . J . 1977. The main tenance of spec ies - r i chness in plant communi t ies : T h e impor tance of the regenerat ion n iche. Bio l . Rev . 52 : 107-145. Ha res tad , A . S . and F.L. Bunnel l . 1981. Predict ion of snow-water equivalents in coni ferous forests. C a n . J . For. R e s . 11: 854-857. 89 H a r m o n , M . E . and J . F . Frankl in. 1989. T ree seed l ings on logs in Picea-Tsuga forests of O r e g o n and Wash ing ton . Eco logy 70(1): 48-59. Harper , J . L . 1977. Populat ion biology of plants. A c a d e m i c P r e s s , London . 892 pp. Harvey , A . E . , M .F . J u r g e n s e n , M . J . La rsen , and R.T. G r a h a m . 1987. Rela t ionsh ips a m o n g soi l microsi te, ec tomycor rh izae, and natural conifer regenerat ion of old-growth forests in western Mon tana . C a n . J . For. R e s . 17: 58-62. Herr ing, L . J . and D .E . Ether idge. 1976. A d v a n c e amabi l is fir regenerat ion in the V a n c o u v e r Fores t District. B . C . For. Serv . and C a n . For. Serv . Joint R e p . 5. Pac i f ic Forest R e s e a r c h Cent re , Victor ia, B . C . 23 pp. H o p w o o d , D. 1991. Pr inc ip les and pract ices of New Forestry: a guide for Brit ish C o l u m b i a n s . Land M a n a g e . R e p . No . 7 1 , B . C . M in . For. , Victor ia, B . C . 95 pp. H u z i o k a , T., H. S h i m u z u , E. Ak i taya, and H. Nari ta. 1967. Observat ion of c reep rate of s n o w on mountain s lopes , Tesh io District, Hokka ido. In H. d u r a (ed.) P h y s i c s of s n o w and ice. P roc . Int. Conf . Low T e m p . S c i . , V . 1 . , Pt. 2. Inst. Low T e m p . S c i . , Hokka ido Univ. , S a p p o r o , J a p a n , pp. 1177-1183. J o s z a , L. 1988. Increment core sampl ing for high quality co res . Spec ia l Pub . No . S P - 3 0 . Forintek C a n a d a Corp . , Vancouver , B . C . 26 pp. Keddy , P .A . 1989. Compet i t ion. C h a p m a n and Hal l , London . 202 pp. K l inka, K. and F. Pend l . 1976. Prob lem analys is of regenerat ion in high elevat ion in the V a n c o u v e r Forest Reg ion . B . C . For. Serv . R e s . Div. Land M a n a g e . Se r ies R e p . N o . 2, Vic tor ia, B . C . 90 pp. K l inka, K., V . J . Kraj ina, A . C e s k a , and A . M . S c a g e l . 1989. Indicator plants of coasta l Brit ish C o l u m b i a . Universi ty of Brit ish Co lumb ia P r e s s , Vancouver , B . C . 288 pp. K l inka, K., R . E . Carter , G . F . W e e t m a n , and M. Jul l . 1992. Silvicultural ana lys is of the subalp ine Mounta in Hemlock zone . B . C . Min . For. , Burnaby, B . C . 46 pp. K n a p p , A . K . and W . K . Smi th . 1982. Factors inf luencing understory seed l ing estab l ishment of Enge lmann spruce [Picea engelmannii) and subalp ine fir (Abies lasiocarpa) in southeast W y o m i n g . C a n . J . Bot. 60: 2753-2761 . K o h y a m a , T. 1983. Seed l ing s tage of two subalp ine Abies spec ies in distinction from sapl ing s tage : a mat ter -economic analys is . Bot M a g . (Tokyo) 96: 49 -65 . K o p p e n a a l , R . S . and A . K . Mitchel l . 1992. Regenerat ion of montane forests in the C o a s t a l Wes te rn Hemlock zone of Brit ish Co lumb ia : a literature review. For . C a n . and B . C . M in . For. , Victor ia, B . C . F R D A R e p . No . 192. 22 pp. Kotar , J . 1977. Altitudinal distribution of Ab ies amabi l is a s a function of moisture s t ress . In R .D . A n d r e w s III (ed.). P roc . S y m p . on Terrestr ial and Aquat ic Eco log ica l S tud ies of the Northwest . Dept. Bio l . , Eastern W a s h . State Co l lege , C h e n e y , W a s h . pp. 9 -21 . 90 Kotar , J . 1978. Relat ionship of early seedl ing deve lopment to altitudinal distribution of Abies amabilis. Bul l . Torrey Bot. C lub 105(4): 289-295 . Kraj ina, V . J . 1969. Eco logy of forest t rees in Brit ish Co lumb ia . Eco l . Wes te rn N. Amer . 2: 1-146. Kraj ina, V . J . , K. K l inka, and J . Worra l l . 1982. Distribution and ecolog ica l character is t ics of t rees and shrubs of Brit ish Co lumb ia . Facul ty of Forestry, Univ. Brit. C o l u m b i a , Vancouve r , B . C . 131 pp. Krumlik, G . J . 1979. Compara t i ve ana lys is of nutrient cycl ing in the subalp ine Mounta in Hemlock Z o n e . P h . D . thesis, Univ. Brit. Co lumb ia , Vancouver . 196 pp. Kuo , J . , E. Fox , S . A . G lan tz , and S . M c D o n a l d . 1987. S igmaSta t for W i n d o w s J a n d e l Scient i f ic, S a n Ra fae l , C A . Leaphar t , C D . , R .D . Hungerford, and H .E . J o h n s o n . 1972. S tem deformit ies in young t rees c a u s e d by snowpack and its movement . U S D A For. Serv . R e s . Note INT-158. Intermountain Forest and Range Exper iment Stat ion, O g d e n , UT . 10 pp. Le r t zman , K . P . 1989. G a p - p h a s e dynamics in a subalp ine old-growth forest. P h . D . thes is , Univ. Brit ish Co lumb ia , Vancouver , B . C . 167 pp. Le r t zman , K . P . 1992. Pat terns of gap -phase replacement in a suba lp ine, old-growth forest. Eco logy 73(2): 657-669. Le r t zman , K . P . 1995. Forest dynamics , differential mortality and var iable recruitment patterns. J . V e g . S c i . 6: 191-204. Le r t zman , K . P . and C . J . K rebs . 1991. G a p - p h a s e structure of a suba lp ine old-growth forest. C a n . J . For . R e s . 2 1 : 1730-1741. Le r t zman , K . P . , G . D . Suther land, A . Inselberg, and S . Saunde rs . 1996. C a n o p y gaps and the l andscape mosa ic in a coasta l temperate rainforest. Eco logy 77(4): 1254-1270. L o n g , J . N . 1976. Forest vegetat ion dynamics within the Abies amabilis Z o n e of a western C a s c a d e s wate rshed . P h . D. thesis, University of Wash ing ton , Seat t le . 174 pp. Lor imer, C . G . 1985. Methodolog ica l considerat ions in the ana lys is of forest d is turbance history. C a n . J . For . R e s . 15: 200-213. Lowery , R . F . 1972. Eco logy of subalp ine zone tree c lumps in the North C a s c a d e Mounta ins of Wash ing ton . P h . D . dissertat ion. Univ. Wash ing ton . 138 pp. Lut tmerding, H.A., D.A. Demarch i , E . C L e a , D.V. Meid inger, and T. Vo id . 1990. Descr ib ing e c o s y s t e m s in the f ield. 2nd E d . M O E Manua l 11, B . C . Ministry of the Env i ronment , Vic tor ia, B . C . 213 pp. M a c k a y , J . R . and W . H . Ma thews . 1967. Observa t ions on p ressures exer ted by c reep ing snow, Mount Seymour , Brit ish Co lumb ia , C a n a d a . In H. d u r a (ed.) P h y s i c s of s n o w and ice. 91 P roc . Int. Conf . Low T e m p . S c i . , V . 1 . , Pt. 2. Inst. Low T e m p . S c i . , Hokka ido Univ. , Sappo ro , J a p a n , pp. 1185-1197. Mall ik, A . U . 1995. Convers ion of temperate forests into heaths: role of ecosys tem dis turbance and e r i caceous plants. Env . M a n a g e . 19(5): 675-684. M a s e r , C , R . G . A n d e r s o n , K. C r o m a c k Jr. , J .T . Wi l l iams, and R . E . Mart in. 1979. D e a d and down woody mater ial . In J . W . T h o m a s (ed.). Wildlife habitats in m a n a g e d forests - t h e B lue Mounta ins of Oregon and Wash ing ton . U S D A For. Serv . Agr ic . Handbook 553 , Pac i f ic Northwest Forest and R a n g e Exper iment Stat ion, Por t land, O R . pp. 78 -95 . M e g a h a n , W . F . and R. Stee le . 1987. A n approach for predict ing snow d a m a g e to p o n d e r o s a pine plantat ions. For . S c i . 33(2): 485-503 . M e g a h a n , W . F . and R. Stee le . 1988. A field guide for predict ing snow d a m a g e to ponde rosa pine plantat ions. U S D A For. Serv . R e s . Note. INT-385. Intermountain R e s e a r c h Stat ion, O g d e n , UT . 9 pp. Mess ie r , C . and J . P . K immins . 1990. Factors limiting coni ferous seed l ing growth in recently c learcut s i tes dominated by Gaultheria shallon in the C W H v m s u b z o n e . F R D A R e p . 149, B . C . M in . For. , Victor ia, B . C . 30 pp. Minore , D. 1972. Germinat ion and early growth of coasta l tree spec ies on organic s e e d b e d s . U S D A For. Serv . R e s . P a p . P N W - 1 3 5 , Paci f ic Northwest Forest and R a n g e Exper iment Stat ion, Por t land, O R . 18 pp. Minore , F. 1979. Compara t i ve autecological character ist ics of northwestern tree s p e c i e s : a literature review. U S D A For. Serv . , G e n . T e c h . R e p . P N W - 8 7 . Pac i f ic Northwest Forest and R a n g e Exper iment Stat ion, Por t land, O R . 72 pp. Minore , D. 1986. Germinat ion , survival and early growth of conifer seed l ings in two habitat types. U S D A For. Serv . R e s . P a p . P N W - 3 4 8 , Paci f ic Northwest Forest and R a n g e Exper iment Stat ion, Port land, O R . 25 pp. Minore , D. and M . E . Dubras ich . 1981. Regenerat ion after clearcutt ing in suba lp ine s tands near Wind igo P a s s , O regon . J . Forestry 79(9) :619-621. N a k a m u r a , T. 1992. Effect of bryophytes on survival of conifer seed l ings in suba lp ine forests of central J a p a n . E c o l . R e s . 7: 155-162. Ol iver , C D . and B . C . La rson . 1990. Forest s tand dynamics . McGraw-H i l l , Inc., N e w York . 467 PP-Or loc i , L. 1965. The Coas ta l Wes te rn Hemlock Zone on the southwestern Brit ish C o l u m b i a main land. / n V . J . Kraj ina (ed.). Eco logy of Wes te rn North A m e r i c a , V o l . 1. Univ. Brit. C o l . Dept. Bot., Vancouver , pp. 18-34. Pe te r son , C . J . and J . E . Campbe l l . 1993. Microsi te di f ferences and temporal c h a n g e in plant communi t ies of treefall pits and mounds in an old-growth forest. Bul l . Tor rey Bot. C lub 120(4): 451-460 . 92 Pe te r son , E . B . 1964. Plant assoc ia t ions in the subalp ine Mounta in Hemlock z o n e of southern Brit ish Co lumb ia . P h . D . Thes i s . Depar tments of Bio logy and Botany, Univ. of Brit ish C o l u m b i a , Vancouver , B . C . 171 pp. Pe te rson , E . B . 1969. Rad iosonde data for character izat ion of a mountain env i ronment in Brit ish C o l u m b i a . Eco logy 50(2): 200-205 . Piatt, W . J . and D.R. St rong (eds.). 1989. Spec ia l feature - treefall gaps and forest dynamics . Eco logy 70(3): 535-576 . Pojar , J . , K. K l inka, and D.A. Demarch i . 1991. Mounta in Hemlock zone . In D. Meid inger and J . Pojar (eds.). S p e c . R e p . Ser ies 6, B . C . M in . For. , Victor ia, B . C . , pp. 113-124. Pojar, J . and A . M a c K i n n o n . 1994. Plants of coasta l Brit ish C o l u m b i a including Wash ing ton , O r e g o n , and A l a s k a . Lone P ine Publ ish ing, Vancouver , B . C . 528 pp. Pothier, D., R. Doucet , and J . Bolly. 1995. The effect of advance regenerat ion height on future yield of b lack spruce s tands. C a n . J . For. R e s . 25(4): 536-544 . R a d o s e v i c h , S . R . 1984. Interference between greenleaf manzan i ta (Arctostaphylos patula) and ponde rosa pine (Pinus ponderosa). In M.L. Duryea and G . N G r o w n (eds.) . Seed l i ng phys io logy and reforestation s u c c e s s . Nijhoff/Junk Pub l ishers , Dordrecht, T h e Nether lands, pp. 259-270. Reuter , F. 1973. High elevat ion reforestation problems in the Vancouve r Forest District: a problem ana lys is . B . C . For. Serv . , Victor ia, B . C . 46 pp. Runk le , J . R . 1981 . G a p regenerat ion in s o m e old-growth forests of the eastern Uni ted Sta tes . Eco logy 62(4): 1041-1051. Runk le , J . R . 1982. Pat terns of d is turbance in s o m e old-growth mes ic forests of eas tern North A m e r i c a . Eco logy 63(5): 1533-1546. Runk le , J . R . 1992. Gu ide l ines and samp le protocol for sampl ing forest gaps . G e n . T e c h . R e p . P N W - G T R - 2 8 3 . U S D A For. Serv . Paci f ic Northwest R e s e a r c h Stat ion, Por t land, O r e g o n . 44 pp. S c a g e l , R., R. G r e e n , H. von Hahn , and R. E v a n s . 1989. Exploratory high-elevat ion regenerat ion trials in the Vancouve r Forest Reg ion : 10-year spec ies per fo rmance of p lanted stock. F R D A R e p . 098, B . C . Min . For. , Victor ia, B . C . S e i d e l , K .W. 1985. Growth response of supp ressed true fir and mountain hemlock after re lease . U S D A For. Serv . R e s . Note P N W - 3 4 4 . Paci f ic Northwest Forest and R a n g e Exper iment Stat ion, Port land, O R . 22 pp. Se ide l , K .W. and R. Coo ley . 1974. Natural reproduct ion of grand fir and mountain hemlock after she l terwood cutting in central O regon . U S D A For. Serv . R e s . Note P N W - 2 2 9 . Pac i f i c Northwest Forest and R a n g e Exper iment Stat ion, Por t land, O R . 10 pp. 93 Sol l ins , P. 1982. Input and decay of coa rse woody debris in coni ferous s tands in western O r e g o n and Wash ing ton . C a n . J . For. R e s . 12(1): 18-28. Spi t t lehouse, D.L. and R . J . Stathers. 1990. Seed l ing microcl imate. B . C . M in . For. , L a n d M a n a g e . R e p . No . 65 , Victor ia, B . C . 28 pp. Sta thers , R . J . , R. Trowbr idge, D.L. Spi t t lehouse, A . M a c a d a m , and J . P . K immins . 1990. Eco log ica l pr inciples: bas ic concepts . In Regenera t ing British C o l u m b i a forests. D .P . Lavender , R. Par i sh , C M . J o h n s o n , G . Montgomery, A . V y s e , R.A. Wi l l is , and D. Wins ton (eds.). Univ. Brit ish Co lumb ia P r e s s , Vancouver , B . C . pp. 45-54 . Thornburgh, D.A. 1969. Dynamics of the true-fir hemlock forests of the west s lope of the Wash ing ton C a s c a d e R a n g e . P h . D . thesis. Universi ty of Wash ing ton , Seat t le . 210 pp. T r i ska , F . J . and K. C r o m a c k Jr . 1979. The role of wood debris in forests and s t reams. In Fores ts : f resh perspect ives from ecosys tem ana lys is . R . H . War ing (ed.). P roc . 40th Bio logy Co l loqu ium, Apr . 1979, Corval l is . Oregon State Universi ty P r e s s , Corva l l i s , O R . pp. 171-190. Utz ig , G . and L. Herr ing. 1974. Factors signif icant to high elevat ion forest management : a report emphas iz ing soil stability and stand regenerat ion. B . C . For . Serv . R e s . Div., V ic tor ia, B . C . 52 pp. V a n Pelt , R. 1995. Understory tree response to canopy gaps in old-growth Douglas- f i r forests of the Pac i f ic Northwest. P h . D . dissertat ion. Univ. Wash ing ton . 232 pp. V a r g a , P. 1997. Structure and regenerat ion pattern of old-growth s tands in the Moist C o l d Enge lmann Sp ruce - Suba lp ine Fir S u b z o n e of central Brit ish C o l u m b i a . M . S c . thes is . Universi ty of Brit ish Co lumb ia , Vancouver , B . C . 111 pp. Vogt , K., E . Moore , S . Gower , D. Vogt, D. Spruge l , and D. Gr ier . 1989. Productivi ty of upper s lopes forests in the Paci f ic Northwest. In D.A. Perry, R. Meur i sse , R. Mil ler, J . Boy le , J . M e a n s , C . R . Perry, and R . F . Powers (eds.). Maintain ing the long-term productivity of Pac i f ic Northwest ecosys tems . T imber P r e s s , Por t land, O regon , pp. 137-163 . W a g n e r , R. 1980. Natural regenerat ion at the edge of an Abies amabilis Z o n e c learcut on the west s lope of the central Wash ing ton C a s c a d e s . M . S c . thes is , Univ. Wash ing ton , Seat t le , Wash ing ton . 72 pp. Wi l k inson , L , M. Hil l , S . Mice l i , G . B i rkenbeuel , and E. V a n g . 1992. S Y S T A T for W i n d o w s . Evans ton II: S Y S T A T Inc. Wi l l iams, C . B . 1966. S n o w d a m a g e to coni ferous seed l ings and sap l ings . U S D A For . Se rv . R e s . Note P N W - 4 0 . Paci f ic Northwest Forest and R a n g e Exper iment Stat ion, Por t land, O R . 10 pp. Y a m a m o t o , S . - l . 1993. G a p character ist ics and gap regenerat ion in a subalp ine coni ferous forest on Mt Ontake , central Honshu , J a p a n . E c o l . R e s . 8: 277-285 . Zar , J . H . 1984. Biostat ist ical ana lys is . Prent ice-Hal l . Eng lewood Clif fs, N J , U S A . 718 pp. 94 Append ix A . Non-t ree spec ies by life-form (Kl inka etal. 1989), occur rence , and percent cover . Nomenc la tu re fol lows Kl inka etal. (1989) and Pojar and M a c K i n n o n (1994). Clearcut sites Old-growth sites Life form Species Common name Quadrats Cover (%) Quadrats Cover (%) Cladothamnus pyroliflorus Copperbush 0 0.00 55 1.57 Menziesia ferruginea False azalea 67 1.70 69 1.31 Rhododendron albiflorum White-flowered rhododendron 23 0.67 39 0.49 Deciduous Rubus spectabilis Salmonberry 2 0.03 0 0.00 shrubs Salix scouleriana Scouler's willow 3 0.37 0 0.00 Sambucus racemosa Red elderberry 6 0.18 0 0.00 Sorbus sitchensis Sitka mountain ash 1 0.01 1 0.06 Vaccinium spp. Huckleberry (blueberry) 53 52.68 54 27.57 Chimaphila menziesii Menzies' pipsissewa 0 0.00 6 0.01 Evergreen Gaultheria humifusa Alpine wintergreen 4 0.02 6 0.13 shrubs Gaultheria ovatifolia Western tea-berry 13 0.42 15 0.18 Linnaea borealis Twinflower 3 0.46 0 0.00 Phyllodoce empetriformis Pink mountain-heather 1 0.00 14 0.03 Athyrium filix-femina Lady fern 18 0.12 0 0.00 Blechnum spicant Deer fern 33 0.47 26 0.96 Ferns and Dryopteris expansa Spiny wood fern 23 0.15 0 0.00 fern allies Gymnocarpium dryopteris Oak fern 1 0.00 4 0.04 Lycopodium clavatum Running club moss 12 0.03 5 0.01 Polystichum munitum Sword fern 6 0.01 0 0.00 Anaphalis margaritacea Pearly everlasting 4 0.01 0 0.00 Arnica cordifolia Heart-leaved arnica 0 0.00 18 0.25 Caltha biflora Two-flowered marsh marigold 11 0.67 0 0.00 Clintonia uniflora Queen's cup 38 0.28 3 0.01 Coptis asplenifolia Fern-leaved goldthread 6 0.18 7 0.01 Cornus canadensis Bunchberry 257 3.12 45 0.15 Dicentra formosa Bleeding heart 3 0.05 0 0.00 Epilobium angustifolium Fireweed 347 3.08 1 0.00 Goodyera oblongifolia Rattlesnake-plantain 1 0.00 7 0.03 Hieracium albiflorum White-flowered hawkweed 12 0.05 0 0.00 Listera caurina Northwestern twayblade 0 0.00 1 0.02 Herbs Listera cordata Heart-leaved twayblade 2 0.00 11 0.02 (forbs) Luetkea pectinata Partridgefoot 0 0.00 4 0.01 Lysichitum americanum Skunk cabbage 6 0.14 17 0.40 Mycelis muralis Wall-lettuce 6 0.03 0 0.00 Orthilia secunda One-sided wintergreen 1 0.00 57 0.13 Parnassia fimbriata Fringed grass-of-Parnassus 0 0.00 7 0.03 Pedicularis bracteosa Bracted lousewort 0 0.00 8 0.11 Rubus pedatus Five-leaved bramble 358 5.52 424 2.34 Streptopus amplexifolius Clasping twistedstalk 5 0.02 19 0.05 Streptopus roseus Rosy twistedstalk 3 0.01 61 0.27 Streptopus streptopoides Small twistedstalk 1 0.00 36 0.10 Tiarella trifoliata Three-leaved foam flower 2 0.00 8 0.01 Tiarella unifoliata One-leaved foam flower 0 0.00 43 0.09 Veratrum viride Indian false hellebore 13 0.17 65 0.57 Lichens Cladina spp. Shrub lichens 36 0.11 28 0.06 Barbilophozia floerkei Snow-mat liverwort 3 0.01 1 0.03 Liverworts Bazzania tricentra Three-toothed whip liverwort 0 0.00 56 1.11 Calypogeia trichomanis n/a 0 0.00 14 0.17 Pellia neesiana Ring pellia 0 0.00 6 0.04 Dicranum spp. Dusky fork moss, broom moss 174 1.43 318 6.18 Mnium spinulosum Menzies' red-mouthed mnium 0 0.00 13 0.12 Pleurozium schreberi Red-stemmed feathermoss 33 0.34 53 0.87 Polytrichum juniperinum Juniper haircap moss 55 0.24 3 0.06 Mosses Pseudotaxiphyllum elegans Small flat moss 0 0.00 16 0.31 Rhizomnium glabrescens Fan moss 1 0.00 15 1.10 Rhytidiadelphus loreus Lanky moss 72 0.45 188 3.30 Rhytidiopsis robusta Pipecleaner moss 242 2.16 477 27.28 Sphagnum fuscum/rubellum Red peat mosses 6 0.27 0 0.00 Sphagnum girgensohnii White-toothed peat moss 9 0.15 66 2.78 Saprophytes Corallorhiza mertensiana Western coralroot 0 0.00 4 0.02 Total number of species (% cover) 46 (75.83) 47 (80.35) 95 Append i x B. Class i f icat ion of growth form anomal ies . C l a s s e s fol low Leaphar t etal. (1972), S c a g e l etal. (1989), and personal observat ions. Growth Fo rm A n o m a l y Definition Broken top (BT) A b s e n c e of leader and ev idence of b reakage. Chlorotic (CL) Ye l lowish discolourat ion of fol iage. Crushed (CR) Part of s tem or whole tree p inned under coa rse woody debr is , common ly logging s lash in c learcuts. Damaged leader (DL) Leader miss ing or deformed. Dog leg (DG) Horizontal growth above the tree base of >5 c m . Hor izontal d is tance and height to resumpt ion of vert ical growth were measu red to descr ibe the severity of deformity. Direct ion of dog leg w a s a lso noted (i.e., uphil l, downhi l l , east , west) . Multiple leaders (ML) More than one leader and no clear apical dominance by any. Multiple leaders were general ly found originating from the top half of s tems. No distinction w a s made for basa l forking a s descr ibed in S c a g e l etal. (1989). Pistol butt (PB) Horizontal growth from the tree base of >5 c m . Hor izontal and vert ical d is tances and direction of deformity were a s noted for "dog leg" above . Prostrate (PR) Ang le of s tem from base to tip <45 degrees . Shrub (SH) No c lear apical dominance shown by many leaders originating from near the tree base . Spire (SP) D ieback of tree top character ist ic espec ia l ly of canopy C. nootkatensis. S f e m sweep (SW) C u r v e d growth form resembl ing the shape of a hockey st ick where the angle of s tem from base to tip >45 deg rees . Umbrella (UM) Horizontal growth greater than vert ical growth result ing in an umbrel la or palm tree appearance . 96 Append ix C . Regenera t ion status of c learcut study locat ions in 1992 ( B . C . Ministry of Fores ts ) , one year before sampl ing . Study locat ion Batche lor M a y n e Tann is Su rvey a rea 39 ha 29 ha 82 ha Tota l s t e m s / h a 12567 6400 3 5 6 0 M in imum stock ing s tandard (stems/ha) 500 500 5 0 0 Target s tock ing s tandard (stems/ha) 900 900 900 W e l l - s p a c e d t rees (stem/ha) 883 900 8 6 0 Free-growing t rees (stems/ha) 700 320 680 Success fu l l y s tocked? y e s y e s y e s F ree-g row ing? yes no y e s W e l l - s p a c e d s tems: A. amabilis 6 7 % 5 8 % 5 6 % C. nootkatensis 1 0 % 1 2 % 1 8 % T. mertensiana 1 7 % 2 7 % 0 % T. heterophylla 6 % 3 % 2 6 % R e c o m m e n d e d future act ion by study location ( B . C . Ministry of Fores ts 1992): Batchelor : 1997 pre-stand tending survey for poss ib le 1998 incremental juveni le spac ing . M a y n e : 1997 free-growing survey in 1997; predicted to be free of brush in 4-5 years . Tann is : 1997 pre-stand tending survey for poss ib le 1998 incremental juveni le spac ing . 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075274/manifest

Comment

Related Items