UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Post-harvest floor changes and nitrogen mobilization in an Engelmann spruce-subalpine fir forest David, Clive Addison 1987

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

Item Metadata

Download

Media
831-UBC_1987_A1 D38.pdf [ 15.32MB ]
Metadata
JSON: 831-1.0097136.json
JSON-LD: 831-1.0097136-ld.json
RDF/XML (Pretty): 831-1.0097136-rdf.xml
RDF/JSON: 831-1.0097136-rdf.json
Turtle: 831-1.0097136-turtle.txt
N-Triples: 831-1.0097136-rdf-ntriples.txt
Original Record: 831-1.0097136-source.json
Full Text
831-1.0097136-fulltext.txt
Citation
831-1.0097136.ris

Full Text

POST-HARVEST FOREST FLOOR CHANGES AND NITROGEN MOBILIZATION IN AN ENGELMANN SPRUCE-SUBALPINE FIR FOREST by C L I V E A D D I S O N D A V I D B . Sc. F . , The Unive r s i ty of N e w Brunswick , 1975 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I 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 D O C T O R O F P H I L O S O P H Y in T H E 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 We accept this thesis as conforming to the required standard 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 J a n u a r y 1987 © C L I V E A D D I S O N D A V I D , 1987 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 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6G/81) ABSTRACT Engelmann spruce-subalpine fir [Picea engelmannil Parry ex Engelm.-A6fes lasiocarpa (Hook.) Nutt.] (ESSF) forests occupy large portions of western North America, and of British Columbia (B.C.) in particular. These areas represent a harsh environment for plant growth. The ESSF forests of B.C. have presented serious problems of regeneration following harvesting; several factors stimulated speculation as to whether N supply limitations were involved. This study was intended to highlight the post-harvest N dynamics of an ESSF forest, and the implications of the latter for silvicultural practices. Its general objectives included characterization of the post-harvest assart effect, and investigation of the N status and growth of advance regeneration. These were achieved by means of a comparative study of an age sequence of harvested sites. The assart effect -lasted for at least eight years after harvesting, with a peak of change between years three and six. There were no major physical changes in the forest floor. Low C/N ratios between 19 and 32 were believed to have contributed to increased N availability. ESSF forests may have a generally higher level of N availability than previously supposed. The advance regeneration benefited from the assart effect. Nutrient uptake appeared to increase generally from at least three years after harvesting; increases of up to 78% were noted for N. There appeared to be no general macronutrient or micronutrient limitation to growth. However, evidence of S deficiencies was encountered in some trees. Moreover, the critical levels used for N may be in need of revision. A revised critical level of 1.40% for foliar N concentrations is proposed for subalpine fir n advance regeneration. I f this is accurate, regeneration m a y have been at least temporarily N- l imi t ed from year eight after harvest ing. A more rigorous investigation of these possibilities is needed. The cutt ing method applied to the sites approximated a one-cut shelterwood method. The method as encountered in this study should not be considered a viable s i lv icul tural option for s imi lar E S S F forests. Its successful application would involve some degree of forest floor manipulat ion to improve seedbed conditions and soil microclimatic regimes. The findings of this study demonstrate that the environmental and biological characteristics of E S S F forests make high levels of planning and care a prerequisite for the success of s i lv i -cultural practices. The question of what comprises realistic growth and yield expectations of second-rotation stands in the E S S F zone needs to be addressed urgently. i i i T A B L E OF CONTENTS A B S T R A C T i i L I S T O F T A B L E S v i i i L I S T O F F I G U R E S ix A C K N O W L E D G E M E N T S x i Dedication x i i Chapter 1 I N T R O D U C T I O N 1 1.1 Introduction 1 1.2 Developing a Research Strategy and Study Objectives 2 1.2.1 Ni t rogen Ava i l ab i l i t y in Northern Forests 2 1.2.2 The A s s a r t Effect 7 1.2.3 The Impact of Harves t ing 11 1.2.4 The Subalpine F i r Question 14 1.3 The Objectives of the Study 16 1.4 The Approach i n Concept 17 1.4.1 A n Age Sequence of Uni fo rm Sites 18 1.4.2 Indicative Var iables and Relationships 23 1.4.3 The Scientific Hypotheses Under Consideration . 27 1.4.3.1 Some Methodological Considerations . 27 1.4.3.2 The Scientific Hypotheses 32 IAA The Organizat ion of the Thesis 35 Chapter 2 S I T E D E S C R I P T I O N S A N D B A S I C F I E L D L A Y O U T 37 2.1 The Research A r e a 37 2.1.1 Locat ion 37 2.1.2 Geology 40 2.1.3 Genera l Cl imate 42 2.1.4 Soils 45 2.1.5 Vegetat ion 46 2.1.5.1 Original Vegetation 46 2.1.5.2 The Harvested Sites 49 2.2 F i e ld L a y o u t and Sampl ing Scheme 53 2.2.1 F i e ld L a y o u t 53 2.2.2 Sampl ing Scheme 60 2.2.2.1 Field Phase 60 2.2.2.2 Preparatory Laboratory Phase 63 Chapter 3 S E A S O N A L T E M P E R A T U R E A N D M O I S T U R E T R E N D S 66 3.1 Review of Issues and Concepts 66 3.1.1 Temperature 67 3.1.2 Mois ture 72 3.1.3 Objectives and Hypotheses 78 3.2 Methods 81 3.2.1 Temperature 81 - iv 3.2.2 Mois ture 82 3.3 Results and Discussion 84 3.3.1 Temperature 84 3.3.2 Mois ture 88 3.4 Conclusions • 97 Chapter 4 P O S T - H A R V E S T F O R E S T F L O O R P H Y S I C A L C H A R A C T E R I S T I C S .. 99 4.1 Introduction 99 4.2 Objectives and Hypotheses 101 4.3 Methods 102 4.4 Results and Discussion 103 4.4.1 To ta l Depths 103 4.4.2 Weights 105 4.4.3 B u l k Densities 107 4.4.4 M o r H u m u s Groups 108 4.5 Conclusions I l l Chapter 5 P O S T - H A R V E S T S O I L C H E M I C A L C H A R A C T E R I S T I C S 112 5.1 Introduction 112 5.2 Objectives and Hypotheses 114 5.3 Methods 116 5.3.1 p H Values 116 5.3.2 Ni t rogen and Phosphorus 117 5.3.3 Carbon and C / N Ratios 119 5.3.3.1 Carbon 119 5.3.3.2 Carbon.-Nitrogen Ratios 121 5.3.4 Potass ium, Ca lc ium, and Magnes ium 121 5.3.5 U n i t - A r e a Elementa l Weights 122 5.4 Results and Discussion 124 5.4.1 p H Values 124 5.4.2 Ni t rogen Concentrations 128 5.4.3 Phosphorus Concentrations 131 5.4.4 Exchangeable Potassium, Ca lc ium, and M a g n e s i u m 135 5.4.5 Carbon Concentrations and CarbonrNitrogen Ratios 139 5.4.6 U n i t - A r e a Elementa l Weights 145 5.4.6.1 Nitrogen Content 149 5.4.6.2 Phosphorus Content 150 5.4.6.3 Exchangeable Potassium, Calcium, and Magnesium Contents 150 5.4.6.4 Carbon Contents 152 5.5 Conclusions 153 Chapter 6 P O S T - H A R V E S T N I T R O G E N A V A I L A B I L I T Y P A T T E R N S 156 6.1 Introduction 156 6.2 Objectives and Hypotheses 161 6.3 Methods 162 v 6.3.1 Incubation Studies 162 6.3.2 Chemical Extractions 166 6.3.3 Ion Exchange Resins 167 6.4 Results and Discussion 169 6.4.1 Incubation Studies 169 6.4.1.1 Forest Floor Materials 169 6.4.1.2 Mineral Soil Materials 175 6.4.1.3 Unit-Area Weights 179 6.4.2 Chemical Extract ions 181 6.4.2.1 Forest Floor Materials 181 6.4.2.2 Mineral Soil Materials 185 6.4.2.3 Unit-Area Weights 187 6.4.3 Ion Exchange Resins 187 6.5 Conclusions 192 Chapter 7 P R E - A N D P O S T - H A R V E S T T R E E G R O W T H 195 7.1 Introduction 195 7.2 Objectives and Hypotheses 198 7.3 Methods 199 7.3.1 Height growth 199 7.3.2 Diameter Growth and Age 200 7.4 Results and Discussion 201 7.4.1 Height Growth 207 7.4.2 Diameter Growth 209 7.4.3 Tree Age Considerations 210 7.5 Conclusions 213 Chapter 8 P O S T - H A R V E S T F O L I A R C H E M I S T R Y 215 8.1 Introduction 215 8.2 Objectives and Hypotheses 217 8.3 Methods 221 8.3.1 M a i n Foliage Analys i s 222 8.3.2 Screening T r i a l Approach 224 8.4 Results and Discussion 226 8.4.1 M a i n Foliage Ana lys i s 226 8.4.1.1 Needle weights and Nitrogen 226 8.4.1.2 Phosphorus, Potassium, Calcium, and Magnesium 235 8.4.1.3 NIP, K/Ca, and Ca/Mg Ratios 239 8.4.2 Micronutr ients and Sulphur 241 8.4.3 Screening T r i a l Analyses : 245 8.4.3.1 Directional Relationships 245 8.4.3.2 Overall Analyses 248 8.5 Conclusions 248 Chapter 9 G R O W T H I N R E L A T I O N T O S O I L F A C T O R S A N D F O L I A R C H E M I S T R Y 252 9.1 Introduction 252 9.2 Objectives and Hypotheses 253 vi 9.3 Methods 253 9.4 Results and Discussion 256 9.5 Conclusions 260 Chapter 10 S Y N T H E S I S : H A R V E S T I N G E F F E C T S , G R O W T H , A N D S I L V I C U L T U R E 261 10.1 Recapitulat ion of Pr inc ipa l Results 261 10.1.1 The Soil 262 10.1.1.1 General Chemical Changes 262 10.1.1.2 Nitrogen Availability Changes 263 10.1.2 The Advance Regeneration 264 10.1.2.1 Post-Harvest Growth 264 10.1.2.2 Foliar Chemistry 265 10.1.2.3 Influences on Tree Growth 266 10.1.3 Microcl imat ic Regimes 266 10.2 Overv iew of the Post-Harvest Si tuat ion 267 10.3 S i lv icu l tu ra l Consequences and Implications 270 10.3.1 Background Highl ights 271 10.3.2 The Present Sys tem 275 10.3.2.1 Theory Versus Reality 276 10.3.2.2 Silvicultural Consequences 281 . 10.3.3 Prognosis 284 10.3.3.1 The Untreated Cutovers 285 10.3.3.2 Future Harvests 287 10.4 Conclusions 290 Chapter 11 C O N C L U S I O N S 292 L I T E R A T U R E C I T E D 297 v i i List of Tables Table 2.1 Phys ica l characteristics of the selected sites 41 Table 2.2 Cl imat ic characteristics of the E S S F biogeoclimatic zone ......44 Table 2.3 S u m m a r y of the basic soil characteristics of the sites 47 Table 2.4 Ranges of mensurat ional values for Wet-Belt E S S F stands on well - and moderately well-drained podzolic soils wi th south-eastern aspects 50 Table 2.5 Mensura t iona l characteristics of regeneration on the three oldest cutovers in 1982 52 Table 4.1 Means and 95% confidence limits for forest floor total depths, F / H bulk densities, and weights 104 Table 4.2 Dis t r ibut ion of mor humus groups on the sites 109 Table 5.1a Means and 95% confidence limits of the elemental unit-area weights for the forest floor fraction 146 Table 5.1b Means and 95% confidence limits of the elemental unit-area weights for the minera l soil fraction 147 Table 6.1 Means and 95% confidence limits of the estimated unit-area weights of mineral ized N given by the four measures 180 Table 6.2 Means and 95% confidence limits of the estimated unit-area weights of inorganic N given by KC1 extraction 188 Table 7.1 S u m m a r y of response delay estimates 206 Table 7.2 M e a n release ages for advance regeneration on the cutovers 211 Table 8.1 Means and 95% confidence limits of the micronutr ient concentrations, S concentrations, and N / S ratios in the foliage sub-samples 242 Table 8.2 Means and 95% confidence l imits of the needle weights, foliar N concentrations, and foliar N contents for the grouped fertilizer plots .... 249 Table 9.1 Means and 95% confidence limits of absolute diameter growth for each soil attribute considered 257 vi i i List of Figures Figure 1.1 The interactions among several processes and their effects on N avai labi l i ty 5 Figure 2.1a Locat ion of the study area in B r i t i s h Columbia 38 Figure 2.1b Locat ion of the study sites in the Kamloops Forest Region 39 Figure 2.2 Height and diameter class frequency distributions for regeneration on the oldest cutovers 51 Figure 2.3 V i e w s of the mature stand i n summer, 1983: (a) Exter ior , and (b) Interior 54 Figure 2.4 The three-year-old cutover in summer, 1983 55 Figure 2.5 V i e w of the six-year-old cutover in summer, 1983 56 Figure 2.6 Subalpine fir advance growth on the eight-year-old cutover, summer 1983 57 Figure 2.7 V i e w of the eleven-year-old cutover in summer, 1983 58 Figure 3.1 Means and 95% confidence l imits for the air and forest floor temperatures of the sites, J u l y to Augus t 1983 85 Figure 3.2 Means and 95% confidence l imits for the forest floor moisture contents of the sites, J u l y to Augus t 1983 89 Figure 3.3 Mois ture retention curves for the mature stand and six-year-old cutover 91 Figure 3.4 Means and 95% confidence l imits for the forest floor moisture contents, J u l y to Augus t 1983 92 Figure 3.5 Idealized moisture retention curve and extrapolation (dashed line) for the combined sites 95 Figure 5.1 Means and 95% confidence l imits of the p H values across the age sequence 125 Figure 5.2 Means and 95% confidence l imits of the N concentrations of forest floor and minera l soil materials across the age sequence 129 Figure 5.3 Means and 95% confidence l imits of the forest floor and minera l soil P concentrations, including resin-P concentrations 132 Figure 5.4 Means and 95% confidence l imits of the forest floor and minera l soil exchangeable K , C a , and M g concentrations 136 ix Figure 5.5 Means and 95% confidence limits of the forest floor and mineral soil C concentrations and C / N ratios 140 F igure 6.1 Means and 95% confidence limits of the N mineralization concentrations given by the four measures 170 F igure 6.2 Means and 95% confidence l imits of the inorganic N concentrations given by KC1 extraction 182 F igure 6.3 Means and 95% confidence l imits of the inorganic N concentrations adsorbed by the ion exchange resins 190 F igure 7.1 Pat terns of absolute annual height growth of advance regeneration 202 F igure 7.2 Pat terns of absolute annual diameter growth of advance regeneration 203 F igure 7.3 Pat terns of relative height growth of advance regeneration 204 Figure 7.4 Patterns of relative diameter growth of advance regeneration 205 F igure 8.1 Means and 95% confidence limits of the foliar needle weights, N concentrations, and N contents 227 Figure 8.2 Means and 95% confidence limits of the foliar P , K , C a , and M g concentrations and contents 228 F igure 8.3 Means and 95% confidence limits of the foliar N / P , K / C a , and C a / M g ratios 229 Figure 8.4 Direct ional relationships among needle weights, N concentrations, and N contents 246 x ACKNOWLEDGEMENTS The assistance and support of many persons was needed in the execution of the various phases of this study. It is impract ical to name everyone; nevertheless, I wish to express sincere thanks to them a l l . The greatest thanks are due to D r . G . F . Weetman, m y graduate supervisor; his contribution to m y efforts generally would be difficult to overestimate. I w i sh to thank the members of m y supervisory committee and also Dr . M . Novak for their advice, assistance, and comments; Mess rs . R. Fournier , B . von Spindler, B . Wong, M . Tsze and M r s . E . Tsze for their assistance; and M r . K . C r y s t a l for his enthusiastic support and assistance in the field and subsequent phases. Thanks are also due to M r . D . L l o y d of the B . C. M i n i s t r y of Forests (Kamloops) for his invaluable assistance; also to past and present staff members of Clearwater Timber Products, L t d . —notably Messrs . E . R. Swanson and A . Kokoshke . I cannot thank m y wife, Beverley, sufficiently. H e r patience, encouragement, and assistance in typing the manuscript were v i ta l . F i n a l l y , I would like to thank the Univers i ty of B r i t i s h Columbia, the Depar tment of Forest Sciences, and the I. W . K i l l a m Foundat ion for personal financial support. The project was funded through research grants to D r . G . F . Weetman. xi DEDICATION This effort is dedicated to memories of my family , and the future of our children. x i i CHAPTER 1 INTRODUCTION 1.1 INTRODUCTION Enge lmann spruce-subalpine fir [Picea engelmannii P a r r y ex Engelm.t-A&z'es lasiocarpa (Hook.) Nutt . ] (ESSF) forests occupy large portions of western N o r t h A m e r i c a , and of B r i t i s h Columbia (B.C.) in par t icular . These areas represent a ha rsh environment for plant growth; growing seasons are generally short, while cold soils and low levels of microbial activity contribute to slow nutrient cycl ing rates. More specifically, the E S S F forests of B . C . have exhibited serious problems of regeneration following harvesting—many of these being associated wi th the conditions left by the latter process. N a t u r a l regeneration from seed or by plantat ion establishment has proven very difficult, while advance regeneration often exhibits low post-harvest growth rates. There are also potential problems associated wi th non-crop vegetative competition. The foregoing characteristics hold serious implications for the E S S F forests and their second-rotation yields, especially in light of a growing dependence of the forest industry on the same. Such factors stimulated speculation as to whether problems of nitrogen (N) supply for tree growth existed generally in this setting. This study was init iated p r imar i ly to examine whether N supply problems existed in a second-rotation crop in an E S S F forest. The intention was to shed l ight on the N dynamics of such forests following harvesting, and to relate these findings to possible s i lv icul tural t Scientific nomenclature follows Taylor and M a c B r y d e (1977); common names follow the latter as well as Ceska (1979) and Angove (1981). Where common name differences existed Angove (1981) was given precedence. 1 ) 2 al ternatives. 1.2 DEVELOPING A RESEARCH STRATEGY AND STUDY OBJECTIVES A review of the main considerations in the development of a research strategy and study objectives is appropriate here. The discussion includes overviews of the problem of N avai labi l i ty i n northern forests, a phenomenon termed the "assart effect", the impact of harvest ing on nutrient availabil i ty, and the role of advance regeneration wi th in the specific context of this investigation. 1.2.1 Nitrogen Availability in Northern Forests Northern coniferous forests have long been considered as having N avai labi l i ty as one of their p r ima ry growth- l imit ing characteristics (Tamra, 1950; Weetman , 1958 and 1980a; T a m m , 1964 and 1982; Weetman and N y k v i s t , 1963; Pritchett, 1979; Cole and Rapp, 1981). This lack of plant-available N has been attributed to several factors. F o r example, short growing seasons, low mean growing season and annual temperatures (both soil and air), fungal domination of decomposition processes, mycorrh iza l influences, and the unfavourable chemical composition of the organic substrate (itself largely derived from the trees which stand to benefit from its decomposition) have a l l been noted as interactively contributing to N shortages (Romell, 1935; T a m m , 1950; M . Alexander, 1977; Meentemeyer , 1978; Pritchett, 1979; Swif t et al, 1979; Gosz, 1981; V a n Cleve et al, 1981; T a m m , 1982; Vitousek, 1982; Alexander , 1983). 3 The forest floor and humus (along wi th their associated microflora and fauna) p lay extremely important roles in the dynamics of the forest environment ( M . Alexander , 1977; Bormann and L ikens , 1979; Pritchett, 1979; Gosz, 1981; Anderson et al, 1983; Spiers et at, 1984); their impact on N avai labi l i ty to trees merits some elaboration. According to Pri tchett (1979), the te rm "forest floor" is generally used to refer to a l l organic matter, including litter and decomposing organic layers, resting on the minera l soil surface in forested ecosystems. The forest floor may be considered as part of the accumulated organic mater ia l in the forest soil sub-system, which m a y be divided into that which occurs at the soil surface (ectorganic) and that which is incorporated into the minera l horizons (endorganic) (Lowe and K l i n k a , 1981). St r ic t ly defined, the te rm "humus" refers only to that component in which decomposition has proceeded to an extent where residues are no longer recognizable; however, the te rm as used by many foresters and soil scientists often means any organic portion of the soil profile (Russell, 1973; Pritchett, 1979). Throughout this presentation, "humus" refers exclusively to endorganic mater ial while "forest floor" is used for surface organic layers. It is widely accepted that almost a l l of the N in surface soils is held in organic forms largely unavailable to plants; this seems part icular ly true in forest soils, where N is found principally in the forest floor and uppermost minera l horizon (Pritchett, 1979). Though there is some evidence that soluble organic N can be absorbed directly by trees through their mycorrhiza (Heal et al, 1982), the bulk of N uptake by trees in acid forest soil rooting environments is apparent ly in ammonium N ( N H J -N) form. Trees therefore have to depend 4 heavi ly on decomposition and concomitant processes, notably N mineral izat ion. Swif t et al. (1979) identified three categories of factors influencing decomposition (and thus N availabil i ty) — the physico-chemical environment (e.g. temperature, moisture, p H , etc.), the decomposer organisms involved (e.g. types, numbers), and the quanti ty and quali ty of the organic substrate (e.g. chemical composition, physical nature, etc.). These factors appear to be highly interactive in the forest environment. The humus forms (mor, moder, or mull) resulting from such influences can generally be taken as a qualitative indication of both the mode of decomposition and N avai labi l i ty levels wi th in the system. Figure 1.1 from Gosz (1981) illustrates some of the interactions and their resultant effects on N avai labi l i ty . The influence of the forest humus form on the growth of trees has been a subject of investigation for more" than a century. W i t h origins in Europe and Scandinavia, arguments over the perceived merits of mul l and undesirabil i ty of mor humus in terms of N supply and tree growth have generated a body of theory and empirical knowledge of great practical value to silviculturists (Romell, 1935; T a m m , 1950; Weetman, 1962; Russel l , 1973; Pritchett, 1979). One of the results of the foregoing has been a long-standing interest to v a r y i n g degrees in the conservation and management of humus. Several investigators have proposed and reported concerning manipulation of the forest floor for the enhancement of tree growth and site productivity (Romell, 1935; Weetman, 1962; Bormann and L ikens , 1979; Lowe and K l i n k a , 1981). Lowe and K l i n k a (1981) expressed the view that since it is the uppermost soil layer which can be most influenced by s i lv icul tura l practices, humus form management to this end should be an integral par t of s i lv icul tura l management in Br i t i sh Columbia . LOW N AVAILABILITY t LOW N UPTAKE / x HIGH N AVAILABILITY t HIGH N UPTAKE LOW MX IN PLANT TISSUE HIGH N WITHDRAWAL FROM OLO TISSUE HIGH POLYPHENOL, ORGANIC ACID PRODUCTION V LOW N % HIGH N % IN PLANT TISSUE LOW N WITHDRAWAL FROM OLD TISSUE LOW POLYPHENOL, ORGANIC ACIO PRODUCTION IN LITTERFALL, V STABLE POLYPHENOL-\\ PROTEIN COMPLEXES \ LOW DECOt \ MINER ALIZA MPOSITION, TION RATES t LOW NITRIFICATION HIGH N % IN LITTERFALL, REDUCED OR UNSTABLE POLY PHENOL-PROTEIN COMPLEXES HIGH DECOMPOSITION, MINERALIZATION RATES t ' HIGH NITRIFICATION (M O R HUMUs\ F O R M A T I O N ) / M U L L H U M U S \ ( F O R M A T I O N j Figure 1.1 The interactions among several processes and their effects on N avai labi l i ty (F rom Gosz, 1981). 6 Manipula t ion measures can include fertilizer applications (Viro, 1963; Mahendrappa, 1978; Mahendrappa and Salonius, 1982), the inclusion of minor tree and other vegetative components wi th crop trees (Tamm, 1950; K i m m i n s and Hawkes , 1978; M i l e s , 1981; B r o w n and Har r i son , 1983), the introduction of suitable soil organisms, and/or promotion of conditions under which they can flourish (Brown and Har r i son , 1983). Fert i l izat ion has an additional advantage in that it can also serve as a rapid diagnostic measure of the nutr i t ional status and response potential of trees (Timmer and Stone, 1978; Weetman and Fournier , 1982). However , forest floor management in northern coniferous forests is not without difficulties. F o r both forest regeneration by natural means (seeding-in) and plantation establishment, some form of forest floor disturbance or removal is usually desirable for the creation of proper seedbeds or planting spots at the least. In the case of plantat ion establishment at high elevations, site preparat ion principal ly aimed at removing organic layers has been seen as a necessity (Dobbs and M c M i n n , 1977; Burdet t et at, 1984; Potts, 1985). However , removal of forest floor layers can lead to drastic reductions in growth (Herr ing and M c M i n n , 1980; Weber et al, 1985)—a fact recognized for approximately a century in Europe (Pritchett, 1979). Prescribed burning is another frequently-used forest management/s i lvicul tural tool which can be in conflict wi th the concept of humus form management. There are also such questions as the impacts on watershed values, par t icular ly i n higher elevation sites (Hi l lman and Golding, 1981). It can therefore be seen that a general understanding of the effects of intended forest floor manipulations or other treatments—including harvesting—on nutrient avai labi l i ty and tree growth is essential. 7 1.2.2 The Assart Effect Following major disturbance of the forest, there is usual ly a period of increased nutrient avai labi l i ty attributable to several factors (Romell, 1935 and 1938; T a m m , 1964; T i m m e r and Weetman, 1969; Stone, 1975; Salonius et al, 1977; Bormann and Likens , 1979; Vitousek, 1981 and 1983). These include: 1. Inputs of fresh organic matter (including roots) to decomposer organisms; 2. Changes in temperature and moisture regimes which often enhance microbia l activity and decomposition; 3. Decreased vegetative competition for moisture and nutrients; and 4. Relaxation of mycorrhiza l suppression of decomposition—the effects noted by Gadgi l and Gadgi l (1975 and 1978), and also Berg and Lindberg (1980). This liberation of nutrients has been referred to as the "assart effect" (Romell, 1957; T a m m , 1964 and 1979; Weetman, 1980a; Demontigny and Auc l a i r , 1982); where fire has been the agent of disturbance, the term "ash-bed effect" (or "ash bed effect") has also been applied—particularly in Aus t r a l i a , where it has long been employed in forestry practice (Kessell and Stoate, 1938; H a t c h , 1960; Humphreys and Lamber t , 1965). The magnitude and duration of the assart nutrient flush are of great pract ical importance to man's attempts to modify the forest environment to his advantage. For centuries, its benefits were realized through systems of shifting cult ivation in Europe (Romell, 1957; Noirfal ise and 8 Th i l l , 1960); s imi la r ly , in parts of the Tropics, it has been the mains tay of traditional agr icul tura l production. In terms of forest management, the potential for channell ing nutrients so mobilized into growth of desirable trees has been stressed (Tamm, 1950; Weetman, 1980a). W h a t was the origin of the term "assart effect"? It is simple to visualize why the ash-bed te rm might have been coined; however, the origins of the assart phrase are more obscure. T a m m (1964) gave a definition of the term "assart effect", and attributed its coining to L . - G . Romell . However , a survey of the available l i terature as far back as the 1930's yielded no English-language publication by Romel l which could be definitely construed as its origin. Romell (1935 and 1938) described the processes involved (inter alia) but did not use the term; Romel l (1938) employed other terms for the phenomenon (e.g. th inning effect, soil act ivat ing effect, s t imulat ion effect), implying that the assart term had not yet been conceived. Romel l (1957) was the earliest publication found in which direct reference to the assart effect was made, and indeed T a m m and Pettersson (1969) impl ied that this was one of its first uses. However , the context (Romell, 1957) indicated that the term had been earlier defined and was comparat ively well-known to his audience. In his detailed survey, T a m m (1950) made no explicit use of the term, though he discussed the phenomenon itself and its implications. Therefore, it m a y be said that the actual origin of the term "assart effect" remains vague, but its first use can be dated as dur ing the 1950's. Regardless of its precise or igin, the term undoubtedly had its roots i n the word "assart". The latter i n turn has its use in Engl ish dating back to the sixteenth century (well before that in French and Lat in) , and was a legal t e rm associated 9 mainly with both the act and the result of the conversion of forest land to arable land by the grubbing up of trees and bushes (Oxford Eng l i sh Dict ionary, 1933; Week, 1966; James , 1981). According to James (1981), dur ing the late Middle Ages, the creation of an assart was the most serious offence against the king's forests. Its seriousness derived as much from the unauthorized felling of trees as from the removal of the stumps and roots; that is, the permanence of the conversion away from forest cover. F r o m the purist 's viewpoint, this latter characteristic might be seen as rendering the te rm "assart effect" misleading in terms of its application to forest management practices. Nevertheless, this wri ter feels that it presents a very succinct way of describing a complex phenomenon; the term and its derivatives have been used to refer to the same throughout this report. The potential for having the benefits of the assart effect realized in crop tree growth should be of special value in considering those northern coniferous forests where mor humus accumulation seems to contribute to N deficiencies (Weetman, 1958 and 1980a; Weetman and N y k v i s t , 1963). More specifically, the magnitude and durat ion of the assart f lush are of importance to the establishment and growth of regeneration in those high-elevation areas of Br i t i sh Columbia where spruce [Picea engelmannii P a r r y ex Enge lm. , P. glauca (Moench) Voss, and their hybrids] plantations have not been very successful [Vyse, 1981; Vyse and LeLacheur , 1979; Burdet t et al, 1984; B r i t i s h Columbia M i n i s t r y of Forests ( B . C . M . O . F . ) , 1986a]. The same applies to cutovers in the latter areas where spruce exhibits low rates of ingress—even wi th site treatments—and/or subalpine fir advance regeneration often appears unresponsive to release by 10 logging (Herr ing, 1977; Her r ing and M c M i n n , 1980; Monchak , 1982). If regeneration is delayed or growth retarded sufficiently for the trees to miss the benefits of the assart flush, serious backlog regeneration problems m a y develop on such sites (Tamm, 1950; Weetman, 1980a). Weetman (1980a) indicated that little information is available concerning the pattern of the assart effect following harvest ing of northern coniferous forests. V e r y recently, two such studies were completed in B . C . —those of M a r t i n (1985) on the coast and H a s k i n (1985) in the dry interior. The Weetman (1980a) statement is st i l l true for the interior high-elevation forests of B . C . ; the H a s k i n (1985) study appears to be the only one to have investigated assart aspects in the latter. Th is is surpr is ing given the increasing importance of these forests to the provincial industry. Fo r the past 20 years, the forests of the interior have yielded the major portion of B . C . ' s lumber production; since 1972 the total harvest in the interior has been consistently greater than that of the coast ( B . C . M . O . F . , 1980). Spruce is the dominant species in interior forests; at higher elevations it occurs w i th subalpine fir, the latter contributing almost entirely to any advance regeneration present. Such spruce-firt forests comprised some 50% to 70% of the productive volume in the southern interior (Bickerstaff et al, 1981). A s easily accessible and/or more productive forests become exhausted, the more marginal spruce-fir types are assuming increasing importance —especially in view of an impending shortage of operable timber wi th in the near future ( B . C . M . O . F . , 1980). The E S S F biogeoclimatic zone (Krajina, 1965) is the th i rd largest continuously t " F i r " in B . C . usual ly refers to Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] , while "ba lsam" has been the traditional term for the Abies spp. However , throughout this presentation "f i r" refers to the Abies spp., and principal ly to subalpine fir. 11 forested zone i n B . C . , covering some 12% of the area of the province (Coupe, 1983). In the southern interior, it is easi ly the largest of the three types in which spruce (and fir) dominates. Its reputation of being a harsh environment for tree growth is wel l known. It has become apparent that the E S S F type presents severe problems in terms of its regeneration after harvest ing, wi th thousands of hectares inadequately regenerated.! W i t h increasing pressure on these areas, and in light of the dearth of information concerning disturbance effects, such problems threaten to become even more acute. 1.2.3 The Impact of Harvesting In general, much has been wr i t ten concerning the effects of disturbances—including harvesting—on forest ecosystems and sites (e.g. Leaf, 1979; Bormann and L ikens , 1979; Covington, 1981; Gosz, 1981; Vitousek, 1981 and 1983; Vitousek et al, 1979; Vitousek and Mel i l lo , 1979; T a m m , 1982). The most well-known of these is perhaps the Hubba rd Brook study in N e w Hampshire , which has provided much valuable information (e.g. L ikens and Bormann , 1974; Bormann and L ikens , 1979). This extensive body of literature wi l l not be reviewed i n detail here; rather, only a selection of items germane to the subject of this report w i l l be highlighted. Harves t ing initiates a complex set of reactions and processes in the forest ecosystem. This is par t icular ly true of the forest floor layers. Decomposition rates are generally enhanced; however, for nutrients such as N—the major component t A . V y s e , B . C . M . O . F . (Kamloops), pers. comm. November 18, 1985. of which is held i n organic forms —the key to their avai labi l i ty to trees lies in their net mineral izat ion rates. A s stated earlier, mineral izat ion is governed pr incipal ly by environmental influences, the quant i ty and quali ty of the organic substrate, and the types and numbers of microorganisms—these working interactively. B a a t h (1980) observed a decrease in fungal biomass over time (years) following cutting in Sweden, while Sundman et al. (1978) s imi la r ly observed a strong increase in both invertebrate fauna and bacterial populations following cutt ing in F in land . Matson and Vitousek (1981)- also concluded that clearcutt ing can lead to higher bacterial populations—particularly nitrifiers. Studies such as these suggest that the very nature of the decomposition processes is radical ly changed wi th such disturbances. The forest floor thus plays an important role in these and other changes following harvesting. One effect that has been noted is a decrease in forest floor thickness and weights following cutt ing or disturbance (Bormann and Likens , 1979; Covington, 1981; M a r t i n , 1985). Th i s decrease can continue far beyond the point at which leaf area (and thus perhaps leaf fall) appears to have recovered substantial ly. In the Hubbard Brook study, it continued for 10 to 20 years after cutting, while leaf area had recovered after three to four years (Bormann and L ikens , 1979). Bormann and L ikens (1979) hypothesized that the latter represented part of an effort by the ecosystem to reorganize itself following disturbance, w i th its nutrient capital as the m a i n vehicle. Gosz (1981) suggested that the forest system moves from the disturbed state, wi th relatively high levels of plant-available N , towards a more conservative use of that element by means of rapid ingrowth, competition, and immobil izat ion of nutrients in tissues. Vitousek (1982) also discussed possible mechanisms and their implications in this context. F o r N , the effect is therefore 13 that of a gradual reduction in its avai labi l i ty over time, a slowing down in its movement, and its maintenance in either organic or cationic form. Losses are therefore effectively minimized (Gosz, 1981; Vitousek, 1982). M o r humus formation and perpetuation in northern coniferous forests can be viewed in this context as the na tura l tendency of the forest in dynamic equil ibr ium. Al though net mineralization levels have been established as the key to N avai labi l i ty , many other factors influence the actual uptake of the latter by trees. Env i ronmen ta l and site factors such as soil, temperature and moisture regimes can have major effects here. Seasonal fluctuations in fine root biomass have been noted (Kimmins and Hawkes , 1978; Whi tney and T immer , 1983), and have been observed to restrict N uptake in spruce (Whitney and T immer , 1983). Monchak (1982) believed that in the E S S F zone of the Kamloops Forest Region, B . C . , shallow rooting, high stem densities, and high water stress were contributing factors to observed N deficiencies in advance regeneration on cutovers. A pattern of nutr i t ional stages recognized by M i l l e r (1981) is also of importance in considering the post-harvest situation. He observed that before canopy closure, growth is determined principal ly by the soil supply. Tree demands on the latter are low, since trees do not fully occupy the site; however, the ini t ial demand from competing minor vegetation is h igh (Cole and Rapp, 1981; Mi l l e r , 1981; H e a l et al., 1982). Even feather mosses can constitute a significant source of competition wi th trees for available N , and/or form an important control in this regard (Bernier and Roberge, 1962; Wee tman and T immer , 1967; Weber and V a n Cleve, 1981 and 1984). Thus, while before canopy closure the total demand by vegetation for N may be relat ively high, most of this can be from non-crop 14 components. The implication of the above is that while the assart effect m a y be operational following the harvest, other factors m a y militate against the tree crop being the principal beneficiary. The foregoing considerations indicate that there is a need to investigate and understand the mechanisms, processes, and tree responses operating following harvesting—in short, the assart effect in relat ion to the establishment and growth of regeneration. This may be part icular ly true for the E S S F forest areas of B . C . referred to earlier; these aspects assume even greater importance where second-growth management options involve decisions as to whether subalpine fir advance regeneration should be retained as at least a major component. 1.2.4 The Subalpine Fir Question Advance growth has played an indispensable role in the regeneration of N o r t h Amer i can spruce-firt forests (Smith, 1962); in the E S S F and s imi lar higher-elevation spruce-fir types of B . C . , this option would involve acceptance of subalpine fir as a major proportion of the second crop. The latter species has the widest distribution of any true fir i n N o r t h A m e r i c a (Henderson, 1982; Y o u n g , 1985), and is part icularly abundant in the B . C . interior (Kraj ina et al, 1982; Wat ts , 1983). However, in B . C . subalpine fir is regarded wi th some contempt (Haddock, 1982; Handley, 1982); in the southern interior, there is a trend toward ignoring or eradicating the species (Tozer, 1984).? The m a i n objections to its use include alleged slow growth (Watts, 1983; Alexander et al., t In this case this term embraces a l l Picea spp.-Abies spp. forest type combinations in Nor th Amer ica . t A l so A . V y s e , B . C . M . O . F . , pers. comm., November 18, 1985. 1984), poor post-logging form and damage characteristics (Herring, 1977), susceptibility to decay (Kra j ina et al, 1982) and to the balsam wooly aphid {Adelges piceae Ratz.) (Gara , 1982; Haddock, 1982). While the disdain is not entirely unfounded, the bases of some of its aspects are questionable. Fo r example, the slow-growth label has been unfair ly applied, though this perception of the species' performance seems most difficult to modify. Ten years ago, Crossley (1976) argued convincingly against such a blanket judgement; this was later supported—though not unconditionally—by the data of several investigators (Herr ing, 1977; H e r r i n g and M c M i n n , 1980; Johnstone, 1978; Monchak, 1982; Bergstrom, 1983; E is and Craigdal l ie , 1983). Perhaps one of the main origins of the slow-growth label lies i n the fact that assessments of growth were (inadvertently) made for trees under widely differing physiological and other conditions and also perhaps wi th inadequate indicator variables. A s an example, McCaughey and Schmidt (1982) reported a ten-year release periodic annual height increment of approximately eight to eleven centimetres. However, this would include (and be downwardly biased by) the period of response delay—a well-established phenomenon i n true fir release response. B y contrast, other investigators have reported release current height increments of up to 34 c m y r " 1 (Herr ing and M c M i n n , 1980; Monchak, 1982). Moreover, absolute annual height increment is perhaps not the best indicator of performance of advance regeneration, since its assessment on a stand basis includes trees of different ini t ia l sizes; in such cases, relative growth measures are better indicators (Hunt, 1982). The second-growth alternative to subalpine fir advance regeneration on 16 many such spruce-fir sites is to attempt to establish spruce plantations. However , these have not been very successful (Vyse, 1981; V y s e and LeLacheu r , 1979; B . C . M . O . F . , 1986a); moreover, there is an additional cost of site preparat ion as wel l as the risks involved in the effects the latter might have on the forest floor and growth as outlined earlier. It is therefore unrealistic to assume that the plantation approach wi l l be able to satisfy adequately the regeneration needs of present backlog and future cutover sites in B . C . ' s spruce-fir forests. The logical extension of this is that it would be s i lvicul tural ly unwise to ignore totally the potential of the advance regeneration option—especially i f this is done on the basis of traditional biases. In most cases, subalpine fir advance growth is already i n place to form part of the second crop, and is fully adapted to the harsh E S S F conditions. However, this alternative is not without its disadvantages; to circumvent these, an understanding of the species' post-harvest growth response in relation to the assart pattern and earlier considerations becomes desirable. Th is information would improve the confidence and efficacy of s i lv icul tura l decision-making. While M a r t i n (1985) investigated such aspects for the Coasta l Western Hemlock biogeoclimatic zone, to this wri ter 's knowledge there have been no studies of this type in the E S S F zone in B . C . 1.3 THE OBJECTIVES OF THE STUDY The general objectives of this study were as follows: 1. To investigate and describe the magnitude, duration and dynamic pattern of the assart effect following harvesting in an E S S F forest; 17 2. To verify the occurrence of N deficiencies reported i n advance regeneration i n the E S S F zone; i f found to occur, to identify possible reasons for the same; 3. To describe patterns of growth of advance regeneration following harvest ing in an E S S F forest, and their relationship to the assart pattern; and 4. To discuss the implications of (1) to (3) above for s i lv icul tura l practices in the E S S F zone. In the chapters which follow, specific jobjectives are developed for a stated context; al l fall wi th in the general ambit of the statements above. 1.4 T H E A P P R O A C H I N C O N C E P T There are m a n y ways to approach an investigation wi th the objectives stated above. W h a t follows is an outline of the principal formative ideas which governed the approach employed in this instance. It should be stressed that the latter was formulated during the 1982-83 period, and thus could not benefit from subsequent findings by other investigators. The assumptions and cri teria governing site selection are first discussed. Next , the variables and relationships examined and the general stat ist ical approach employed are presented. F ina l ly , the main experimental hypotheses which were tested (and their alternates) are considered. 18 1.4.1 An Age Sequence of Uniform Sites Ideally, any study of the assart flush in relation to growth should take place on sites the characteristics of which are tracked individual ly from pre-disturbance until some point; for example, unti l leaf area reached its original level, or perhaps unt i l canopy closure. Such an approach was used in the Hubbard Brook study (Bormann and L ikens , 1979) and by investigators such as Sollins and McCor ison (1981). Unfortunately, constraints of available resources (e.g. personnel, financial support, time, etc.) frequently preclude the employment of this approach. One alternative has been to use a sequence of different-aged (i.e. years after disturbance) sites wi th in a given forest type on the assumption that any pattern over time w i l l be reflected by the variables measured on these sites—the "chronosequence" approach (Lang et al, 1981; M a r t i n , 1985; Wallace and Freedman, 1986). Tha t is, by "point-in-time" measurements on the different-aged sites, the pattern over the span of ages sequences can be constructed. This approach has been applied in various forms (and wi th various objectives) by such investigators as Covington (1981), L a n g et al. (1981), Nicolson et al. (1982), Matson and Boone (1984), H a s k i n (1985), M a r t i n (1985), and Wallace and Freedman (1986). There are several problems associated wi th using an age sequence approach in evaluation of disturbance effects over time. A n immediately obvious l imitat ion is the difficulty of adequately representing mobile nutr ient elements. Site uniformity is a pr incipal concern; it is difficult to dist inguish between observed effects attributable to natural between-site var iab i l i ty and those 19 attributable to the disturbance and time. In this study, it was believed that the latter difficulty could be circumvented (or at least the effects of natural between-site var iab i l i ty minimized) i f sites were selected to be as uniform as possible in their basic physical characteristics (e.g. biogeoclimatic subzone, elevation, aspect, slope, harvest ing method, post-harvest treatments, etc.)(Covington, 1981). O f course, use of a characteristic as a site selection criterion conceptually precluded it from inclusion i n any assessment of time effects. F r o m the viewpoint of logistics, available equipment, and intended procedure, it was desirable to have the sites geographically close together. The span of the age sequence sought was also of importance. A t Hubbard Brook the effects of disturbance were noted for some ten to twenty years subseqently (Bormann and Likens , 1979). However, the studies of Sundman et al. (1978) and B a a t h (1980) indicated that at least from the microbiological viewpoint such effects in northern coniferous forests might be somewhat shorter in duration. The well-documented (for example, see Vitousek et ai, 1979) appearance of increased nitrate N (NO^ -N) levels after disturbance also enters the picture here. In acid soils under northern coniferous forests, ni tr if ication can usually only proceed at elevated levels where conditions have been markedly ameliorated (thus enhancing decomposition and net mineral izat ion processes, increases in bacterial populations —especially nitrifiers, etc.). In addition, according to T a m m (1950), a group of plants which he termed a "nitrate-loving flora" become temporari ly prominent in conjunction with increased N O 3 - N avai labi l i ty; the group included Epilobium angustifolium L . (fireweed) and Rubus idaeus L . (American red raspberry). T a m m (1964) also mentioned Senecio spp. (ragworts 20 and groundsel) as belonging to this group. A substantial presence of members of the "nitrate flora" does not by itself conclusively indicate that N O 3 - N concentrations have increased markedly on a site (Tamm, 1964). Nevertheless, T a m m (1964) believed that such presence can at least be interpreted as indicating a good N supply. U s i n g data from a long-term opt imum nutri t ion study in Sweden (Tamm, 1974), Bormann and L ikens (1979) provided some support fpr the latter belief; they noted that increases in A m e r i c a n red raspberry cover were associated wi th up to a twenty-fold increase i n soil available N . Moreover, at Hubbard Brook the latter species was among the most prominent invaders within the first three years following clearcutting (Bormann and Likens , 1979). Fireweed is a wel l -known prolific invader of disturbed sites in m a y parts of Nor th America—including interior B . C . —and over a wide range of elevations (Endean and Johnstone, 1974; Lyons , 1976; Bo rmann and L ikens , 1979; Angove, 1981). Its association wi th post-disturbance elevated N O 3 - N levels in several instances is noteworthy (Matson and Vitousek, 1981; Sollins and McCor i son , 1981; Gordon and V a n Cleve, 1983). Moreover , i n the Sollins and McCor i son (1981) study Senecio spp. also dominated the site (along wi th fireweed) two years after cutting. Grasses also increased in frequency following clearfelling (Tamm, 1950). The time of appearance and also the durat ion of increased N O 3 - N concentrations can thus be construed as defining a period when any l ikely assart effect is in full operation—perhaps at its peak where N avai labi l i ty to plants is concerned. It is fully recognized that the t iming and rates of nitrification observed following disturbance are site-dependent to a large degree (Vitousek et al, 1979; Vitousek, 1981; Wiklander , 1981). However , for a "uniform" group of sites (as discussed above) differences in inherent fert i l i ty should not enter the picture. Cole and Ba l l a rd (1968) demonstrated that the greatest increases in N fluxes can take place wi th in the first few months following harvest ing. Increased fluxes of N O 3 - N were observed by Sollins and McCor i son (1981) about seven months after logging, but only at a depth of two meters in the soil. A t lesser depths such increases did not take place unt i l 18 months had elapsed. Unfortunately, i t was unclear as to when (if at a l l w i th in the study period) these levels declined. Whi le appearing highly variable w i t h obvious seasonal trends, the increased NO5 - N concentrations lasted for at least two years in the upper soil horizons. Gordon and V a n Cleve (1983) noted extremely high and variable NO5 - N levels for nine to eleven months following harvest ing; the levels declined subsequently to relatively low values, but remained higher than a corresponding mature forest. Contrary to a mere lag period before increases, Nicolson et al (1982) actually observed significantly lower NO5 - N concentrations in s t ream samples for as much as two years after harvest ing and scarification. However , by four years after the cut annual N O 3 - N losses were elevated and continuing. The investigators attributed these losses to dramatic increases in water production following cutting. F r o m the above observations, it was thought that any post-harvest assart effect under northern coniferous forest conditions could last for a m a x i m u m of ten to fifteen years. If the intensity of disturbance was lower than that of a full clearcut (as is the case when advance regeneration and residuals are retained), this period would be more in the vicini ty of ten years . In any case, 22 the period of m a x i m u m N availabil i ty (as indicated by heightened N O 3 - N levels) could be anywhere from a few months to possibly six years following logging. A n age sequence spanning as much as twenty years would have been highly desirable; however, should this not be available, a ten-year sequence was deemed at least adequate for a definition of the commencement and peak period of the assart f lush. Thus , ideally what was required was an age sequence of sites in the E S S F spanning at least ten (and preferably 15 to 20) years. In addition, insofar as possible, the sites had to be relat ively close together wi th in the same biogeoclimatic subzone and have had the same in i t ia l forest cover. They had to lie at approximately the same elevation, and have s imi lar slopes and aspects. Fur ther , it was thought desirable to have a mature stand meeting these criteria close to each harvested site for comparative purposes. F i n a l l y , ease of access in terms of distance from a Univers i ty of B r i t i s h Columbia ( U . B . C . ) base was also a consideration. The age sequence of sites used in real i ty fell somewhat short of the ideal described above, but was considered the best available. In 1982, five cutovers wi th a ten-year span were found; a l l were essentially on the same mountainside. This sequence was later modified sl ightly to yield a f inal set of four cutovers wi th the same span; further details w i l l be dealt w i th in Chapter Two. The main l imitations of the age sequence examined lay in three facets. The first was that only one cutover area was available at each age point; thus, no replication of sites was possible. The effect of this was to prevent the obtaining of any direct measure of the between-site var ia t ion in the characteristics of interest. Uni formi ty therefore had to be assumed on the basis of physical characteristics. 23 Secondly, comparable mature stands were not available i n close proximity to any of the cutovers. A single mature stand at some distance away—but satisfying the relevant criteria—had to be used as a comparison base. The final point was that harves t ing was by clearcut wi th protection of residuals ( B . C . M . O . F . , 1979). Thus , there could have been (and appeared to be) differences among the sites in terms of the density of retained advance regeneration and intensity of disturbance—though in the latter case these seemed to be relat ively minor. 1.4.2 Indicative Variables and Relationships The pr incipal assumption underlying this study is that comparisons among a carefully selected age sequence of cutovers and a representative mature stand can be used to demonstrate patterns of change wi th time derived from harvest ing. Therefore, the variables and relationships chosen to indicate these patterns needed careful consideration. Because of its overwhelming importance and relat ive unavai labi l i ty in the context of the growth of northern coniferous forests, a strong emphasis was placed on the various aspects and forms of N in this study. The phenomenon of increased NOj - N levels following harvest ing has been observed i n sites as far apart geographically as A l a s k a (Gordon and V a n Cleve, 1983), Vancouver Island, B . C . (Binkley and Packee, 1982), the Oregon Cascades (Sollins and McCor i son , 1981), Indiana (Matson and Vitousek, 1981), and northwestern Ontario (Nicolson et al., 1982). Perhaps the most famous example of this was in the Hubbard Brook study (Likens and B o r m a n n , 1974; Bormann and L ikens , 1979). 24 A s earlier indicated, the magnitude and t iming of the NO3 - N formation is dependent on several processes and site factors. However , for the uniform age sequence, the N O 5 - N phenomenon suggested a conceptually simple inferential approach to an assessment of whether crop trees are benefiting from the assart flush: If, in a harvested area, mineral iza t ion rates and N O 3 - N levels were observed to have increased while crop tree foliage weights, N concentrations, and N contents remained at deficient or cr i t ical levels, i t could be concluded that the assart effect was in operation but the trees were not the beneficiaries to any marked degree. G i v e n the foregoing considerations, the following characteristics were chosen for study: 1. Thicknesses and weights of forest floor layers; 2. Levels of N in the forest floor and upper minera l soil, part icularly wi th respect to mineral izat ion and any presence of NO5 - N ; 3. Amounts of other macronutrients [phosphorus (P), potassium (K), calc ium (Ca), and magnesium (Mg)] and also carbon (C) and p H levels in the forest floor and upper mineral soil; 4. Forest floor and minera l soil temperature and moisture trends over the growing season; 5. Height and diameter growth of advance regeneration before and after harvest ing; 6. Macronutr ient (N , P , K , C a , and Mg) levels and 25 needle weights of foliage of advance regeneration; and 7. On selected sites in the age sequence (youngest, middle, and oldest sites), the response of advance regeneration to N fertilizer treatments. In addition, basic assessments of forest floor and soil classifications as well as the mensurational characteristics of the stands/cutovers in question were considered desirable. Using the primary data generated by the above, the principal trends and relationships to be investigated can be considered for convenience to fall into four basic—but overlapping—comparison groups as given below: 1. Group I - Trends of the following against time after harvest: a. Forest floor weight changes; b. Forest floor and air temperature and moisture characteristics; c. Forest floor and mineral soil N forms and availability using indices (anaerobic incubations, chemical extractions, and ion exchange resins); d. Forest floor and mineral soil macronutrient (as earlier listed) levels; e. Forest floor and mineral soil pH values; f. Forest floor and mineral soil C/N ratios; g. Foliar macronutrient and needle weight values; and h. Height and diameter growth of advance regeneration. Group II - Trends of the following against height and diameter growth of advance regeneration since harvesting: a. Forest floor weight changes; b. Forest floor and mineral soil N forms and availability as above; c. Height, diameter, and age of advance regeneration at the time of harvest; and d. Height and diameter growth prior to harvest. Group III - Trends of the following with fertilizer treatment effects: a. Forest floor weight changes; b. Forest floor and mineral soil N mineralization levels; and c. Foliar macronutrient levels and nutrient status assessments. Group IV - Miscellaneous specific aspects: a. Possible refinement of N mineralization trends 27 using temperature and moisture values; and b. Comparison of soil moisture and foliar N levels in light of observed mineral izat ion and relative nitrification (if present) levels. 1.4.3 The Scientific Hypotheses Under Consideration A consideration of the overal l formulation of the scientific hypotheses under examinat ion is necessary. A brief review of some of the more important methodological issues is first presented. This is followed by the development of the m a i n hypotheses themselves. Subsid iary hypotheses stated in the chapters which follow are proposed wi th in the context of these main ones. 1.4.3.1 Some Methodological Considerations The views of K . Popper are considered dominant in terms of current methodological approaches to scientific research. H i s cornerstone requirement was that any scientific theory must be potentially falsifiable; tests of hypotheses should therefore be aimed at their falsification, since hypotheses cannot be verified (Dolby, 1982; Magee, 1982; Bunne l l , 1984). Wi th in this context, P ia t t (1964) urged that rigorous application of a formal set of steps—his "strong inference"—to a system of multiple hypotheses was the one t ruly effective approach. However, Quinn and D u n h a m (1983) have been among those who have argued that the Piat t (1964) approach is difficult to apply s tr ic t ly in m a n y 28 ecological situations. F o r that matter, investigators such as Roughgarden (1983) appeared to be against the use of such formal rules—and especially the Popperian criterion—in the na tura l sciences. Between these two extremes, Simberloff (1983) was of the view that par t icular ly in community ecology Popper's prescription should be moderated. He considered that both the complexity of communities and the difficulty of manipulat ing them dictated that refutation by a single observation be held insufficient to discredit . a theory immediately; several repetitions of such an event were needed. Another area of controversy on the methodological plane stems from the question of whether "nul l models" and nul l hypotheses have any place in ecological research. N u l l models may be defined as those models which are aimed at the el imination of the effects of some part icular ecological process (Harvey et al, 1983). One of their main values seems to be in forming a basis from which departures m a y be measured (Quinn and D u n h a m , 1983). The complicated arguments involved here m a y be followed in publications such' as those of H a r v e y et al. (1983), Quinn and D u n h a m (1983), Roughgarden (1983), and Simberloff (1983). One very recent addition to this series which gives a clear and succinct picture is that of Connor and Simberloff (1986). It is far beyond the scope of this study to attempt a resolution of the issues presented above. The position was taken that insofar as possible the Popperian prescription would be followed. Piat t ' s (1964) approach was attempted in principle; however, constraints of resources prevented a rigorous evaluation of al l possible alternatives. Moreover , problems of a lack of mutual exclusivi ty of alternatives (Quinn and D u n h a m , 1983) were encountered. A further consideration was the precise form of the hypotheses to be 29 tested or otherwise examined. Dolby (1982), and later Connor and Simberloff (1986), outlined the links and differences among scientific theories, scientific hypotheses, and statistical hypotheses: Scientific theories concern processes, while scientific hypotheses are verbal forms of such theories. Stat is t ical hypotheses are precise formulations of the consequences of scientific theories, and require data from clearly defined populations; therefore, they are quite distinct from scientific hypotheses. Toft and Shea (1983) noted that in ecology hypotheses are typical ly statist ical ones and not t ru ly Popperian (scientific). In both the scientific and statistical cases, testing can have two possible correct conclusions and two kinds of errors; in the statistical case, the latter are widely known as Type I (rejection of a true nul l hypothesis) and Type II (failure to reject a false nul l hypothesis) errors (Toft and Shea, 1983; Connor and Simberloff, 1986). N u l l models and their concomitant nu l l hypotheses have been said to have the advantage of being conservative. Tha t is , since Type II errors are committed more often than Type I by researchers, the tendency would be towards finding no evidence of any experimental effect (Toft and Shea, 1983; Connor and Simberloff, 1986). While there are notable exceptions (e.g. i n examining the effects of pesticides), in basic ecological research the costs of Type II error have been considered to be far less dramatic than those of Type I errors. Otherwise stated, it has been deemed a worse error to jump prematurely to a false conclusion than to fail to detect an exist ing effect (Toft and Shea, 1983; Connor and Simberloff, 1986). Fur thermore , i f the probability of Type II error is sma l l , the degree of departure from the nul l hypothesis can indicate the strength of the effect under consideration (Toft and Shea, 1983; Quinn and D u n h a m , 1983; Connor and Simberloff, 1986). In light of the above considerations, the use of 30 the nul l approach to scientific hypothesis testing was adopted in this study. The results of the testing of stat ist ical hypotheses based on the consequences of these nul l hypotheses were accepted as adequate cri ter ia for falsification. The final methodological issue was the extent to which this s tudy could be considered experimental . Th is has some bearing on whether statist ical techniques could be val id ly applied. Piat t ' s (1964) "strong inference" steps would be ideally applicable to a fully-controlled laboratory situation, presumably wi th as elegant a design as feasible or necessary. Results would be obtained from specific treatments or manipulat ions, and compared to those obtained in the absence of any treatment (the "control") (Quinn and Dunham, 1983; Connor and Simberloff, 1986). However , much ecological work lies at the other end of the spectrum, and m a y be termed non-experimental research (Connor and Simberloff, 1986). The latter entails the collection and evaluation of data for their consistency wi th specific theories or w i th the operation of specific causal processes; there is no knowledge of wha t values the data might yield in the absence of such processes (Connor and Simberloff, 1986). This study was perceived as ly ing between these two extremes. It could not be considered as fully controlled, pr incipal ly because of its field-sampling orientation, its after-the-fact character, and other problems outlined earlier (see section 1.4.1). O n the other hand, the different sites in the sequence could be perceived as having each resulted from a "time treatment". Their variables could thus be analyzed statistically using techniques such as the analysis of variance ( A N O V A ) — a s s u m i n g that the assumptions of the latter were met—with the mature stand's values taking the place of a control. A posteriori tests could be used to determine individual differences. Where the assumptions of the A N O V A could not be reasonably satisfied, the non-parametric equivalents of these tests could be applied (Hollander and Wolfe, 1973; Soka l and Rohlf, 1981). The A N O V A designs envisioned were principally simple one-way and two-way Model I analyses; in the two-way case, where imbalance in cells occurred these could be legitimately analyzed using appropriate modifications (Gilbert, 1973; Gr ieg and Bjerring, 1980; L i , 1982). Hurlber t (1984) recognized two basic classes of experiments—the "mensurative" and the "manipulat ive" . In mensurative types, either space or time is the only factor considered. Tests of significance are not usual ly involved, neither does the investigator apply any specific treatments or manipulat ions. In contrast, the latter are the defining features of the manipulat ive class of experiments. However , there are also studies which measure the properties of a system at points wi th in it, then examine whether real differences exist among them. These were called "comparative mensurative experiments" (Hurlbert , 1984). This study could therefore be termed a comparative mensurat ive experiment in which the application of statist ical design techniques was considered appropriate. This is in accord wi th the view expressed by Connor and Simberloff (1986). The latter stated that m a n y important ecological questions lie beyond the scope of feasible experimentation; methods should therefore include approaches other than controlled experiments. The pseudo-experimental nature of this study further emphasized the advisabi l i ty of employing the nul l model approach. Quinn and D u n h a m (1983) noted that in non-experimental systems appropriately constructed nul l models are used to describe the system in the absence of the effects of the postulated causal process. Connor and Simberloff (1986) viewed such models as 32 approximat ing the role of the control in the non-experimental case, and also as attempts to place non-experimental data in an experimental context. 1.4.3.2 The Scientific Hypotheses The scientific hypotheses can now be considered in l ight of the previously stated objectives and the points raised in the foregoing section. The foremost of these concerns the theory of the assart effect itself. This theory is on the level of a global conjecture sensu Dolby (1982); using the nul l approach discussed earlier, the pr incipal hypothesis under test becomes: H 0 : There was no assart effect operational on the harvested sites. The simplest mutua l ly exclusive alternate is therefore: y There was indeed an assart effect operational on the harvested sites. If H 0 is true, there should be no detectable differences among sites (including the mature stand) over time where the soil-related variables of Groups I and I V (see section 1.4.2) are concerned. The results of tests on the concomitant statist ical hypotheses derived from this premise therefore become the cri ter ia of falsification. Should the tests fail to reject H 0 ^, there should be no further testing. The most that could be done would be a reporting of the variable values observed; i t might also be instructive to speculate on the reasons underlying the failure to reject the nu l l hypothesis (Toft and Shea, 1983). Rejection of H 0 33 would lead us to consider and report on the magnitudes, durations, and other characteristics of the observed effects; it would also allow us to proceed to the next global conjecture. O n acceptance of the idea of an operational assart effect, the next logical step in the context of the stated objectives would be to examine whether its benefits accrue to the advance regeneration present. In s imi lar fashion to the presentation above, the relevant hypothesis under test becomes: H 0 : There was no benefit from the assart effect accruing to the advance regeneration. Its alternate is: H-i 2: There was some benefit accruing to the advance regeneration from the assart effect. If H 0 ^ is true, there should be no detectable differences in the advance regeneration's foliar macronutrient values, needle weight values, and growth in the Group I set; neither should there be any indication of any relationship between growth of advance regeneration since the harvest and the soil-related variable values of Group II. Test results on related statist ical hypotheses once again become the cri ter ia for falsification. Rejection of H 0 ^ would lead to an examination of the t iming and extent of such benefits wi th in the observed assart pattern. Here , any noticeable limitations on the magnitudes of such benefits might also be considered. No apparent limitations of any kind would lead us to cease further testing; however, any apparent l imitat ions to such benefits and/or 34 acceptance of H 0 would require some elucidation of possible reasons. In the same format as for those foregoing, the global conjectures involved could be stated as: H 0 ^: There was no macronutrient limitation to the ability of advance regeneration to benefit from the assart flush. H , There was indeed a macronutrient limitation to the ability of advance regeneration to benefit from the assart flush. H 0 ^: There was no micronutrient limitation to the ability of advance regeneration to benefit from the assart flush. H , .: There was indeed a micronutrient limitation to ' 4 the .ability of advance regeneration to benefit from the assart flush. H 0 ^: There was no growing season soil temperature limitation to the ability of advance regeneration to benefit from the assart flush. H , j.: There was indeed a growing season soil temperature limitation to the ability of advance regeneration to benefit from the assart flush. H 0 g: There was no growing season soil moisture limitation to the ability of advance regeneration to benefit from the assart flush. 35 : There was indeed a growing season soil moisture limitation to the ability of advance regeneration to benefit from the assart flush. It should be noted that this is the point at which non-exclusivity of alternatives (mentioned earlier) was encountered; therefore, testing may not result in as clear a result as might be desirable. Moreover, available resources constrained the collection of data for possible testing of H 0 and H 0 g above; as stated earlier, strong emphasis was placed on N variables as indicators—thus, in this context, H 0 rather than the latter two null hypotheses. As before, the test bases should be provided by derived statistical hypotheses and tests. Rejection of each of the. above null hypotheses would be followed by an examination and reporting of the values involved; regardless of the test results on the others, each of the null hypotheses in this latter group would have to be considered in turn. Should none of the null hypotheses be rejected, no further testing will be carried out in this study. Rather, the values observed for the variables would be discussed, and speculation as to further causes of non-benefit to advance regeneration attempted. 1.4.4 The Organization of the Thesis In the foregoing sections, a background to the problem under consideration was presented; this was followed by a discussion of the objectives involved and conceptual approach employed. For the purpose of clarity, the material which follows has been organized into 11 chapters. In Chapter Two, the basic site descriptions and field sampling (and other) layout are presented. Observed 36 seasonal temperature and moisture trends are reported in Chapter Three, but a few of the implications of these are discussed as necessary in subsequent chapters. Chapter Four deals with post-harvest physical aspects of the forest floor layers, while the parallel chemical characteristics of both forest floor and mineral soil are presented in Chapter Five. The perceived importance of post-harvest soil N availability patterns in this study is reflected in its separate treatment (Chapter Six). Having considered the environmental and soil changes, attention is next focussed on the observed response of the advance regeneration in Chapters Seven through Nine. Pre- and post-harvest growth patterns are the subject in Chapter Seven. In Chapter Eight a consideration of post-harvest nutrient status is presented. Chapter Nine represents an attempt to identify possible associations and/or relationships among observed growth characteristics and both soil variables and foliar chemistry. Finally, a synthesis of the cumulative results and their implications is presented in Chapter Ten. This includes considerations of post-harvest growth and nutrient dynamics in relation to silvicultural practices in the E S S F forests. The overall conclusions of the study are presented in Chapter 11. CHAPTER 2 SITE DESCRIPTIONS AND BASIC FIELD LAYOUT In this Chapter, the site descriptions and field layout are presented. Fo r the layout and sampling scheme, only general aspects affecting all phases of the study have been discussed. Fur the r details specific to each phase may be found in subsequent Chapters. 2.1 THE RESEARCH AREA 2.1.1 Location The general research area is in the Kamloops Forest Region, near its north-western boundary wi th the Cariboo Forest Region (Watt et al, 1979; L l o y d , 1983). It is situated approximately 26 k m directly north-west of the town of Clearwater , and is part of Tree F a r m Licence (TFL) No . 18 of Clearwater T imber Products, L t d . (CTP) in the N o r t h Nehal l i s ton Forest. The sites all lie on the southern slopes of Swayback Ridge in the M a n n Creek /Moi ra Lake drainage basin (Figures 2.1a,b), the latter being part of the Nor th Thompson r iver system. Thei r latitude and longitude are approximately 5 1 ° 49 ' N and 120° 15' W respectively. A l l sites lie between 1480 m and 1600 m elevation wi th in the Shuswap Highlands Mois t Cen t r a l Enge lmann Spruce Subalpine F i r ( E S S F m l ) va r i an t (Lloyd, 1983); however, two of them lie v i r tua l ly on its lower boundary (Lloyd, 1982). The cutovers are re la t ively close together, while the mature stand lies north-east of them at a distance of approximately two kilometres. 37 38 Figure 2.1a Location of the study area in British Columbia. Figure 2.1b Location of the study sites in the Kamloops Forest Region. 40 A prel iminary survey of the area was carried out i n the fall of 1982. F ive harvested sites were selected for examination, corresponding at that time to a freshly-cut (four months old), three-, five-, seven-, and ten-year-old cutovers; a mature stand was also examined. This ini t ia l sequence provided basic mensurational and other data on the sites. In addition, the freshly-cut, five-, and ten-year sites were fertilized at that time (see Chapter Eight) . The final age sequence used was modified from the above for the m a i n investigations, which were carried out during the 1983 field season. The final sequence retained only the five-, seven-, and ten-year-old cutovers (which had become six, eight, and eleven years old respectively by 1983) from the earlier set, but included a three-year-old (in 1983) site and a more representative mature stand. The salient physical characteristics of these sites are summarized in Table 2.1. 2.1.2 Geology The study area is on the south-eastern portion of the interior Pla teau. It lies very near to the boundary between the Quesnel and Shuswap Highlands , but forms part of the latter [Holland, 1964; Geological Survey of Canada (G.S .C. ) , 1971; Baker , 1978; D . L l o y d f ] . A detailed account of the physiographic history and features of these regions was given by Hol land (1964). These Highlands were formed by the uplift and dissection in the late Pliocene epoch (some two to five mil l ion years ago) of an erosion surface formed earlier in the Ter t i a ry Period. In the Shuswap Highlands , the land surface was elevated over two thousand metres. The Pleistocene glaciations which followed lasted well over one t Regional Research Ecologist, B . C . M . O . F . (Kamloops); pers. comm. October, 1982 and J u l y , 1983. 41 Table 2.1 Phys ica l characteristics of the selected sites. S I T E A G E Y E A R O F C O D E I N C U T 1983 (yr) A R E A E L E V A - S L O P E % A S P E C T S L O P E (ha) T I O N (m) (deg) P O S I T I O N (macro) T M T3 T 6 T8 T i l 3 6 8 11 1980 1977 1975 1972 15.0 2.7 6.2 9.0 29.4 1595 1580 1495 1485 1556 25 0-2 50 36 5 140 176 142 165 upper upper middle middle upper 42 mil l ion years, ending some ten thousand years ago. These glaciations dominated the formation of the landscapes observable at present (Holland, 1964; Ryder , 1978). According to the G . S . C . ' s (1971) mapped information, the sites appear to fall within an area of ra ther complicated geology. The oldest bedrock is comprised principally of P re -Cambr ian (at least 500 mil l ion years ago) and later metamorphic rocks such as feldspathic quartz-mica and other schists. However , in much of the area, this is overlain by Cretaceous (ca. 70 to 120 mi l l ion years ago) granitic rocks such as quartz monzonite and granodiorite. Indeed, samples taken from rocks observed to be frequent and prominent on the sites were identified as being of this group, t Pleistocene and Recent (Holocene) glacial deposits form the th i rd part of this complex (G .S .C . , 1971; Ryder , 1978). Mora ina l materials form the pr incipal soil parent mater ia l over much of the Shuswap Highlands at such elevations (Lloyd, 1983). 2.1.3 General Climate Kra j ina (1965) outlined the major climatic characteristics of the E S S F biogeoclimatic zone in general. In the southern interior of B . C . , this zone m a y be found at elevations between 1220 m and 2288 m . L l o y d (1983) implied that in the Nor th Thompson-Shuswap basin its lower l imi t is about 1350 m . In the Koppen system, the E S S F zone is classified as hav ing a Dfc climate; that is, a continental cold humid climate wi th cool, short summers (Kraj ina , 1965; t D r . E . P . Meagher , Professor, Dept. of Geological Sciences, U . B . C . ; pers. comm., M a r c h 14, 1986. 43 Trewar tha , 1968; Money , 1978). Table 2.2 illustrates some of these characteristics; from these, the rather harsh character of the zone as an environment for plant growth becomes evident. L l o y d (1983) recognized the E S S F zone of the Kamloops Fores t Region as having six forested subzones. Four of these occur in the N o r t h Thompson-Shuswap drainage basins; in terms of an increasingly wet climate, they are the D r y Southern (ESSFe) , Mois t Cent ra l (ESSFm) , Wet Centra l ( E S S F w ) , and Wet Upper (ESSFu) forested subzones. Some climatic characteristics for the Kamloops Region's E S S F zone as a whole were given by L l o y d (1985) (Table 2.2). There were no operational climate stations near the research area for the elevations in question. The closest station for which long-term data were available was at Hemp Creek (Table 2.2). Th is station, no longer operational, fell i n the Nor thern Shuswap Highlands Mois t Centra l Interior Cedar Hemlock ( I C H m l ) var iant ; this occurs at much lower elevations, and represents a markedly wa rmer and drier environment than the E S S F zone (Lloyd, 1983 and 1985). Complete normals (and data for 1983) were also available for a similar low-elevation station at Vavenby (Table 2.2). However , of the data available, those obtained from Boss Mounta in in the Cariboo Fores t Region (Table 2.2) were believed most representative of the research area's conditions (D. L l o y d , pers. comm., October 1982 and A p r i l 17, 1986). The values in Table 2.2 imply that the research sites should be in the upper precipitation and lower temperature ranges of Kra j ina ' s (1965) E S S F zone figures. The bulk of the precipitation should be as snow, while the growing seasons should be relat ively short and with lower temperatures compared to the E S S F zone as a whole. Table 2.2 Climatic characteristics of the ESSF biogeoclimatic zone. L O C A T I O N SOURCE PRECIPITATION (cm) T E M P E R A T U R E (°C) G R O W I N G F R O S T - F R E E D E G R E E - PERIOD D A Y S Annual Total Annual Growing Mean Annual January Mean July Mean Snowfall Season (May Monthly Monthly - Sept.) ESSF General Kamloops Region ESSF Boss Mt." ESSF (1532 m) Hemp Ck. 6 (640 m, ICHml) Vavenby' (445 in) Krajina (1965) Lloyd (1985) Atmos. Envir. Serv. (1982a,b) Atmos. Envir. Serv. (1982a) Atmos. Envir. Serv. (1982a,b) 41 - 183 175 - 1016 11002 118 56 43 700 782 211 116 40 26 20 1 - 4 1 U)1 (4) (6) -18 -29 12 16 -10 (-10) (-8) 12 (15) (18) < 1000 2000' 720' 672] 1670' 35 'Above 6°C. !Bclieved unadjusted, hence difference in magnitude from others. 'Above 5°C. 'In Cariboo Forest Region; latitude 52° 6' N, longitude 120° 53' W; approx. 46 km north of research area; data are complete 30-year normals. 'Bracketed figures derived from mean daily normal data. 'latitude 51° 55' N, longitude 120° 3' W; approx. 18 km north-east of research area; data are 30-year adjusted normals. 7Latitude 511 35' N, longitude 119° 47' W; approx. 43 km south-east of research area; data are complete 30-year normals. 45 2.1.4 Soils Basic classifications of the soils were carr ied out us ing three 50 cm by 50 cm pits in a l l but the 1980 (T3) cutover. Because of its smal l size, only two pits were used in the latter case. The pits were dug at upper, middle, and lower slope positions wi th in the portions of the sites covered by the general sampl ing scheme (see section 2.2). Soil texture was determined by hand in the field. Samples for minera l soil bulk density calculations were obtained using a core sampler of length 7.5 cm and internal diameter 7.2 cm to a depth of 15 cm. Temperature readings were taken at depth intervals down to the 50 cm level or as near to this as was possible. Neither forest floor classifications nor chemical analyses were performed for these profiles, since such work was to be executed i n greater detail on the sampled plots. However , mor humus forms (Bernier, 1968; K l i n k a et ai, 1981) were predominant on' a l l sites. W i t h the exception of two profiles in the 1977 (T6) cutover, al l profiles were classified as Orthic Humo-Ferr ic Podzols ( O . H F P ) . F o r the T 6 site, the upper-slope profile was considered an Eluviated Dys t r i c Brun iso l ( E . D Y B ) with disturbed horizons, while the lower-slope profile was classified as an Orthic Fe r ro -Humic Podzol ( O . F H P ) [Canada Soil Survey Committee (C .S .S .C. ) , 1978]. Soi l textures ranged from clayey to loamy-skeletal; s i l ty loams (SL), si l ty clay loams ( S i C L ) , and, to a lesser extent, loams (L) and sandy loams (SL) were the most common. The soils were shallow (Dumanski , 1978); although as a general rule C horizon thicknesses were not determined, lithic contact or an impermeable layer was invar iab ly encountered at less than one metre of depth. The depth to 46 the C horizon in the profiles ranged from 12 cm to 41 cm. Roots were often at least plentiful down to this level; in the majority of cases, fine roots were observed down to the C horizon. A s u m m a r y of these soil characteristics is presented in Table 2.3. Charcoal was observed at the H/minera l interface on all sites, indicating a rather severe and widespread fire very many years in the past. The soil characteristics of the sites fal l wel l wi th in the range noted for comparable plots i n the E S S F m and E S S F w subzones - the "Wet Bel t" (D. L loyd , in litt., A p r i l 18, 1986). In the latter, the trend was towards slightly coarser and deeper soils. The depth to the C horizon tended to be more uniform (30 -40 cm). However , root abundance and absolute depth of penetration were very s imi lar to those of the Swayback Ridge sites. Shallow, coarse-textured Humo-Fer r ic Podzols were the observed soils under s imi lar conditions near Sock Lake , approximately 14 k m east of the research area and also wi th in T F L No. 18 (Herr ing and M c M i n n , 1980). 2.1.5 Vegetation 2.1.5.1 Original Vegetation A l l sites were considered to be medium i n quali ty. The original stands of the cutovers were comprised of mature, uneven-aged (140 to 250 years) Enge lmann spruce and subalpine fir trees 28 m to 38 m in height ( M . E . Monte i th , Dis t r ic t Manager , Clearwater Forest Distr ic t ; in litt., J u l y 15, 1982; E . R. Swanson , t in litt., M a r c h 15, 1984). Merchantable volumes in the vicinity of t Former P lann ing Forester, C T P , Clearwater . 47 Table 2.3 S u m m a r y of the basic soil characteristics of the sites. S I T E C O D E S U B - G R O U P T E X T U R E S 2 D E P T H T O R O O T P E N E T R A T I O N C O D E 1 (< 2 mm) C H O R I Z O N (cm) (cm) Plentiful or Fine Roots Abundant T M O . H F P S i C L , S i L , S L 31 • • 41 12 - 33 31 -• 41 + T 3 O . H F P S iC , S i C L 20 • • 30 5 - 20 20 -• 30 + T 6 E . D Y B , O . H F P , O . F H P S i C , S i C L , S i L , L 12 • • 21 9 - 21 + 12 • • 21 + T 8 O . H F P S i C L , S i L , S L , L 26 • • 36 26 - 28 26 • • 36 + T i l O . H F P S i C , S i L , 13 • • 17 13 - 17 + 13 • • 17 + S L , L 1 See text for explanation of codes. 2 S i C = sil ty clay; S i C L = silty clay loam; S i L -- sandy loam. sil ty loam; L = loam; S L 48 280 m 3 ha" 1 have been commonly obtained from such sites (E. R. Swanson, pers. comm., September, 1983). The "medium" rat ing implies a site index (100 yr) of 13 to 20 for both Enge lmann spruce and subalpine fir (Watts, 1983). Measurements taken in the mature stand supported the above figures. Tota l basal area was estimated as 51 m 3 ha" 1 ; 68% of this was accounted for by spruce wi th its larger sizes, 27% by subalpine fir, and 5% by lodgepole pine (Pinus contortd Dougl . ex Loud . var . latifolia Engelm.) . However , the understory was comprised overwhelmingly of subalpine fir advance regeneration, wi th relat ively few spruce. The stand averaged 4030 stems ha" 1 of trees wi th measurable breast height diameters (Dbh). O f this, subalpine fir accounted for 3570 stems with an average D b h of 4.8 cm, spruce for 431 stems wi th an average D b h of 30.3 cm, and lodgepole pine for 29 stems wi th an average Dbh of 33.6 cm. The main components of the minor vegetation included Rhododendron albiflorum Hook, (white-flowered rhododendron), Vaccinium membranaceum Dougl. ex Hook, (big or blue huckleberry), Rubus pedatus Smi th (five-leaved creeping raspberry), Paxistima myrsinites (Pursh) Raf. (falsebox), as wel l as Pleurozium schreberi (Brid.) M i t t . t . No occurrences of Menziesia ferruginea Smi th (false azalea) or Gymnocarpium dryopteris (L.) N e w m . (oak-fern) were noted on any of the sites. The characteristics of the major and minor vegetation mentioned above were those of a site' wi th a submesic soil moisture regime (hygrotope) and a poor to medium soil nutrient regime (trophotope) in the E S S F m l variant (Lloyd, t Nomenclature for mosses follows C r u m et al. (1973). 49 1983). The observed mensurational values are wi thin the ranges observed by L l o y d (in litt., A p r i l 18, 1986) for comparable Wet-Bel t E S S F stands (Table 2.4). It should be noted that the sites on which the latter occurred were of better quali ty; site index values place them more in the "good" category (Watts, 1983). Th i s is reflected in the higher basal areas and volumes recorded. 2.1.5.2 The Harvested Sites The present stands on the cutovers resulted from clearcuts wi th protection of residuals ( B . C . M . O . F . , 1979); they are comprised pr imar i ly of subalpine fir advance regeneration. The sites had a l l been skidder-logged during the summer or fal l ; a diameter l imit of 18 cm had been used. Other than slashing (at the time of logging or shortly thereafter) to remove some of the poorer-quality trees, no treatments had been undertaken since the harvest (E. R. Swanson, pers. comm., J u l y , 1983; in litt., M a r c h 15, 1984). Bas ic mensurational characteristics for regeneration on the three oldest cutovers are presented in Figure 2.2 and Table 2.5. These data were obtained during the 1982 pre l iminary survey. Of note are the predominance of the subalpine fir component, and the apparently low rates of ingress of spruce. Stem densities should be taken as indicative only, since subalpine fir advance growth usual ly has a clumped spatial distribution after logging (Herring, 1977; Monchak, 1982). The densities observed were in the lower portion of the range observed by H e r r i n g (1977). The minor vegetation component on each site had no doubt changed wi th disturbance and time from its original condition to 1983. The species mentioned 50 Table 2.4 Ranges of mensurat ional values for Wet-Bel t E S S F stands on well- and moderately well-drained podzolic soils w i th south-eastern aspects.! C H A R A C T E R I S T I C R A N G E Age (yr) 166 - 300 Site index (100 yr) - subalpine fir 18 - 26 - Engelmann spruce 22 - 26 Rooting depth (cm) 21 - 36 Stems ha" 1 500 - 2970 M e a n Dbh (cm) 18 - 45 B a s a l a rea t ( m 2 ha" 1 ) 44 - 81 Volume ( m 3 ha" 1 ) 123 -677 t F r o m L l o y d (in litt., A p r i l 18, 1986). t L i m i t e d data; includes aspects between 9 0 ° and 2 7 0 ° . 51 O - C 6 0 0 - i 500-400-300 E (D 200 tn 100 0 600 500 D JC \ 400 -300 E <D -•— 200 CO 100 0 1 T6 T8 JrzL o JC 6 0 0 - i 500-400-V) 300 E ® 200-CO 100-0 1 V n 1—T 0.40.60.8 1 1.2 1.4 T11 ~rn r y y y ,—,— 1 2 3 4 5 6 7 8 9 10 Height C l a s s (m) Dbh C l a s s (cm) • i Spruce EZ2 Fir Figure 2.2 Height and diameter class frequency distr ibutions for regeneration on the oldest cutovers. 52 Table 2.5 Mensura t iona l cutovers in 1982. characteristics of regeneration on the three oldest C H A R A C T E R I S T I C S I T E A N D S P E C I E S T 6 T 8 T i l E S t S F E S S F E S S F Stems ha" 1 52 (0)t 1638 (391) 103 (0) 2006 (745) 52 (0) 2552 (1638) M e a n height (m) 0.46 1.21 0.61 1.43 0.74 1.98 M e a n D b h § (cm) - 2.2 - 2.4 - 3.0 M e a n basal a r e a § ( m 2 ha" ' ) - 0.2 - 0.4 - 1.5 t E S = Enge lmann spruce; S F = subalpine Fir. ^Numbers in brackets indicate stems wi th Dbh greater than 1.0 cm. § S t e m s less than 1.0 cm D b h omitted. 53 earlier were st i l l evident on each site; however, others such as fireweed had also become very prominent. These changes also made it more difficult to place the sites within an edatopic grid. However , they a l l appeared to correspond to the submesic to mesic hygrotope positions of the E S S F m l var iant (D. L l o y d , pers. comm., Ju ly , 1983). It was assumed that, like the mature stand, they fell wi th in the poor to medium trophotope positions. Figures 2.3 to 2.7 i l lustrate various characteristics of the mature stand and harvested sites. 2.2 FIELD LAYOUT AND SAMPLING SCHEME 2.2.1 Field Layout In the pre l iminary survey of 1982, twelve 0.01 ha circular plots located along an open traverse were used for each site. Single samples of forest floor material , foliage samples from ten dominant subalpine fir trees (excluding residuals) nearest the plot centre, and mensurational data were collected from each plot. In the cases of the then freshly-cut, five-, and ten-year-old sites, fertilization was also done. These prel iminary data allowed calculation of the sample numbers required to yield estimates of the means of selected variables under conditions of random sampling. A formula from Husch et al. (1972) for an Figure 2.3 Views of the mature stand in summer , 1983: (a) Exter ior , and (b) Interior. 55 Figure 2.4 The three-year-old cutover in summer, 1983. (Note lush herbaceous cover.) 56 F igure 2.5 V i e w of the six-year-old cutover in summer, 1983. (The bulk of the minor vegetation in the photograph is fireweed.) 57 Figure 2.6 Subalpine fir advance growth on the eight-year-old cutover, summer 1983. Figure 2.7 V i e w of the eleven-year-old cutover in summer , 1983. 59 infinite population was used (Quesnel and Lavkulich, 1980): n = t 2 ( C V ) 2 / ( A E % ) 2 where: n = required number of sampling units; t = value from the Student's t distribution with n — 1 degrees of freedom and probability (p) = 0.05; C V = coefficient of variation from the preliminary sample; and A E % = allowable sampling error in percentage points (10% was used). The calculated number of sample units required for a given variable differed greatly from site to site. The final numbers applied to the final age sequence represented a compromise between the calculated requirement and the numbers allowed by the constraints of available resources. Thus, 30 sample units were planned (and collected) for the T M and T3 sites, 100 for the T6, 90 for the T8, and 70 for the T i l . As explained later in this Chapter, each soil sample unit was comprised of a composite of three excavations. The size of the T i l cutover enabled the complete separation of the fertilized area from other activities. Fertilization was limited to the northern one-third of the area, while the remaining portion was used otherwise. The main activities were carried out during July and August, 1983. All 60 were restricted to that continuous portion of each site which appeared to represent its overal l condition best. These portions were confirmed as being generally submesic to mesic in their moisture regimes for the E S S F m l (D. L l o y d , pers. comm., Ju ly , 1983). Wi th in them, a 400 m 2 square area was la id out paral le l to the slope on a centrally-located section. A m a x i m u m - m i n i m u m thermometer was installed facing north in open shade at the centre of each square area. After Ba l l a rd (1972), these were placed with their bulbs five centimetres above the ground. La te r in the season (during August) , tensiometers were also installed at these centres. The 400 m 2 areas were used in the gravimetr ic determination of forest floor moisture trends over the 1983 study period. They were excluded from other sampl ing activities; however, their centres were used to locate the sampling points as wel l as soil temperature sensors. Sample points were located at random along randomly selected lines radia t ing from the above centres. S imi la r ly , four thermistor probes were randomly located at the F / H interface of the forest floor on each site. 2.2.2 Sampling Scheme 2.2.2.1 Field Phase The ma in intent of the sampling scheme was to allow examination of the characteristics of and inter-relationships among forest floor samples, the related minera l soil below them, and an individual of the subalpine fir advance growth l inked to the latter two. This was to be achieved for each site, thereby enabling comparisons among sites. Thus, for the advance regeneration, a single subalpine 61 fir representative formed the basic sample unit. This tree was in turn used to orient forest floor and minera l soil sample collections. The "acceptable" tree closest to each sample point was chosen. F o r a tree to be selected it had to fulfil l the following criteria: 1. Its current height had to be 0.30 m or greater; 2. Its height at the time of harvest (as determined from branch whorls and internodal lengths) must have been three metres or less [Herring's (1977) standard]; 3. It must have been present at the time of logging [determined s imi la r ly to (2) above, and in consideration of the time taken for trees to at tain 0.30 m—at least six years, and often as much as seventeen (Watts, 1983)]; and 4. Its form and condition had to warran t its consideration as a crop tree. Sample trees had to be at least four metres apart, and wi thin a short (less than five metres) off-line distance from the sample point. If no qualifying tree was near the sample point, the first acceptable tree encountered further along the line was selected. Trees were tagged wi th orange flagging and a luminum tags. F o r each tree, total height (nearest 0.1 m up to the end of 1982 growth), D b h , and diameter at s tump height (Dsn) — 0.30 m—were recorded where possible. In addition, annual height growth (as given by internodal lengths) was recorded going back to the time of harvest, and, where possible, five years previous to the latter. In the fall of 1983, the current year 's foliage of each tree was 62 sampled for nutrient analysis. S tump height discs were taken in September, 1984 for the examination of diameter growth patterns. The forest floor and minera l soil were sampled in the vic in i ty of the sample trees. For these components, a given sample was made up of three sub-samples taken from equidistant points at a distance from the tree of twice its crown radius. This approach was taken since the use of composite samples taken in such a fashion can reduce sampl ing intensity requirements considerably (Powers, 1980 and 1984a). A m i n i m u m distance of one metre was used for trees wi th crown radi i less than 0.5 m . A t the sub-sampling points, the thicknesses of L , F , and H layers were recorded, and then the L - l a y e r was removed and discarded. The sub-samples (F and H layers) were carefully removed using 225 c m 2 square templates. Roots wi th diameters of approximately one centimetre and greater were excluded from the samples; these were encountered relatively infrequently. The corresponding minera l soil sub-sample was comprised of an undefined quanti ty of the soil underlying the organic sub-sample; the minera l soil was sampled to a depth of 15 cm. in each case. In the fall of 1983, cation and anion exchange resin bags were placed separately at the F / H interface near 30 sample trees on each site. Random selection was used for the sample trees on the T 6 , T8 , and T i l sites. The resins were for the monitoring of N H J - N , N O 3 - N , and phosphate-P ( H 2 P O ^ ) levels over an extended period, and were retrieved after approximately 12 months in situ (see Chapter Six). 63 F r o m the above, i t should be apparent that in the final age sequence of sites each sample point had a "sample plot" associated with it. E a c h "plot" was comprised of a tagged, measured tree the foliage of which was later sampled, a composite F / H forest floor sample, and a composite mineral soil sample. In addition, selected "plots" had anion and cation exchange resins buried at the F / H interface. The numbers of such "plots" per site were therefore the same as those of the sample points given earlier. 2.2.2.2 Preparatory Laboratory Phase The ma in field sampl ing process yielded a total of 960 samples (320 composites each of forest floor and minera l soil, and 320 foliage samples) for processing. Both the number of chemical analyses planned and constraints of available resources dictated a second stage of compositing. A brief description of the procedures followed to derive the final samples for analytical work is given below. Forest floor samples were air-dried in an environment the temperature of which varied between 20°C and 25°C. Because of the limited facilities available, this step took several months. The air-dry weights were recorded, then smal l sub-samples were oven-dried at 105 °C for periods of between 24 and 36 hr . These were used to calculate oven-dry weights of materials. A i r - d r y bulk densities were used to eliminate suspect samples—those which seemed overly dense, possibly because of adherence of minera l particles. Samples which yielded values greater than 0.25 M g m " 3 [near Armson ' s (1977) upper limit] were discarded; 64 in total, 17 samples were so eliminated. Most of the samples had at least one vi r tual ly intact square of F / H mater ia l . Ten sub-samples were taken from a random selection of such squares from the T M and T 6 sites. These were used to develop moisture retention curves for these two sites. The remain ing original samples were crushed manua l ly , then passed through a 2.0 m m sieve; mater ia l fail ing to pass was discarded. The T M and T3 samples were retained i n their original form for the second stage; samples for each of the other sites were randomly composited on a depth-weighted (F/H) basis after Car te r and Lowe (1982a,b). There was now a total of 135 forest floor samples—29 from the T M , 25 from the T 3 , 31 from the T 6 , 27 from the T 8 , and 23 from the T i l sites. A l l further analyses were performed on these second-stage samples. Mine ra l soil and foliage samples underwent a related but shorter process. M i n e r a l soil samples were s imply air-dried then sieved as described above. Second-stage^ composites were then created using the same random selections as for the forest floor, but this t ime in equal proportions. W i t h one exception (see Chapter Six), the numbers used in the analyses were the same as for the forest floor samples. Samples of foliage were first oven-dried at 7 0 ° C for 24 to 36 hours, then the needles carefully stripped from the twigs. Composites were again created according to forest floor selections for subsequent analyses. Preparatory processing thus yielded a system of 405 second-stage samples—135 each of the forest floor (F/H) , minera l soil, and foliage categories. Or ig ina l samples from the T M and T3 sites formed their own second stages, while composites were created for the remaining sites. A key feature was that a 65 given second-stage sample in each category had its counterparts in the other two. Subsequent analyses were performed on these second-stage samples, rather than on the originals collected. CHAPTER 3 SEASONAL TEMPERATURE AND MOISTURE TRENDS Climatic factors exert a tremendous influence on processes such as organic matter decomposition and the growth of organisms in ecosystems. Therefore, in any study such as this, some examination of trends of key climatic variables is highly desirable. There being no climate station in close proximi ty to the research area, such information was gathered in a rather rudimentary fashion for each site in the age sequence. These data are presented in this chapter. In addition, although somewhat out of their logical places in the testing sequence, it was convenient to include here considerations pertaining to ma in hypotheses H 0 and 5 H 0 6 . 3.1 REVIEW OF ISSUES AND CONCEPTS Temperature and moisture levels, together wi th substrate qual i ty, are among the prime controls of decomposition and other microbia l processes, and thus N avai labi l i ty ( M . Alexander , 1977; Meentemeyer, 1978; Swift et al., 1979). Several studies have demonstrated these effects under various northern coniferous forest conditions; K r a u s e et al. (1978), Soderstrom (1979), Edmonds (1980), V a n Cleve et al. (1981), and Weber and V a n Cleve (1981;1984) are examples of these. Such influences m a y be assumed to operate generally in a s imi lar fashion in the E S S F forests. However , as earlier noted, harvest ing and other disturbances usually affect the temperature and moisture regimes of given sites, thereby contributing directly and/or indirectly to an assart effect. W i t h i n the context of 66 67 this study, questions therefore arise as to the nature, magnitudes, and consequences of any such changes derived from harvest ing of the chosen sites. Whi le microclimatic changes have been noted in complete clearcuts (Timmer and Weetman, 1969; B a l l a r d et ai, 1977; Black, 1982), the degree to which these might occur wi th the "incomplete" cuts of the research area was considered important to the investigation of a possible assart pattern and post-harvest tree response. 3.1.1 Temperature F o r this study, the p r imary interest in temperature effects on the sites is in relation to their influences (direct and indirect) on decomposition/mineralization processes as wel l as nutrient uptake and growth of trees. The Boss Mounta in climate normals [Atmospheric Environment Service ( A . E . S . ) , 1982a,b; also see Table 2.2] indicate that the soil temperature of the sites should fall clearly wi th in the middle to lower range of the Cold Cryoboreal class. Soils of the latter exhibit mean annual temperatures ( M A S T ) of between 2 ° C and less than 8 ° C ; the 'growing season (days over 5°C) is between 120 and 180 days, and degree-days over 5 ° C between 555 and 1110 (Lavkulich and Valent ine , 1978). Fo r such cold sites, temperature can be the major factor controlling the movement of N in the forest floor layers (Weber and V a n Cleve, 1984). Moreover , for subalpine fir advance regeneration, Monchak (1982) found that growth was l imited below 450 growing degree-days (base of 5 ° C air temperature). The extent to which harvest ing might have modified the temperature regimes of the sites should have had a major impact on microbial activity and N availabi l i ty from the forest floor 68 as well as on growth. S imi la r ly , the duration of these enhanced conditions in the post-harvest period should also be influential . Because of inevitable differences among the cutovers in characteristics such as in i t ia l post-harvest stem density, temperature data obtained therefrom might not be directly comparable on an individual-site basis. However , differences between the cutovers as a whole compared to the mature stand should be indicative of change; in addition, s t r iking differences among cutovers might nevertheless be informative. Severa l studies have indicated what soil (and air) temperature values might be expected under different conditions of vegetative cover and/or intensity of disturbance. T i m m e r and Weetman (1969) observed that thinning and clearfelling of upland black spruce [Picea mariana (Mill . ) B . S. P.] increased mean summer soil temperatures at al l depths in the organic layers. A t the -5 cm level in the clearcut, mean summer temperatures were increased by 3 ° C to 4 ° C . In August , the m a x i m u m growing season mean weekly temperature at the H/minera l interface was 1 1 ° C for an uncut stand; by contrast, in the clearcut it was already 2 0 ° C at the -5 cm level near the end of J u l y . The organic layers of the clearcut were warmer in the summer (but cooler in the winter) than those of the uncut stand; the overall effect was to prolong the growing season for roots (Timmer and Weetman, 1969). Salonius et al. (1977) considered temperature relationships in spruce-fir forests of the Mar i t imes . Spacing of an "extremely dense" 22-year-old ba lsam fir [Abies balsamea (L.) M i l l . ] stand in Cape Breton appeared to raise growing season soil temperatures (at -2.5 cm) by about 2 ° C on the average. Both spaced and control growing season values varied between 5 ° C and 1 7 ° C approximately. Soil temperatures closely paralleled air 69 temperatures for both the spaced and control conditions, enabling the development of significant l inear regression relationships between the two variables (Salonius et at, 1977). Studies i n B . C . have produced s imilar results. B a l l a r d (1972) demonstrated that quantitative and qualitative differences in vegetative cover on high-elevation sites affect seasonal temperature values at the -60 cm level. Since the -50 cm temperature trends are indicative of the general soil cl imate (Black, 1982), these findings suggest that disturbances and vegetative differences can greatly modify the overal l effects on growth of the relatively harsh subalpine climates—perhaps even to the extent of changing (if only in the short term) the soil climate classification which might be applicable. In decreasing order of growing season soil temperature levels, bare ground exhibited the highest, followed equally by herbaceous and shrub types; next was single tree cover, and finally tree clumps. F o r the vegetated sites, d iurnal temperature variat ions (expressed in terms of that observed wi th bare ground) also followed this pattern; they were greatest under herbaceous cover and least under tree clumps, w i th the evergreen shrub site between these two (Ballard, 1972). Taken as a whole, the vegetative pattern of sites was that which might have been expected of a secondary succession after a major disturbance. The related pattern of temperature values and variat ions observed by Ba l l a rd (1972) might therefore be indicative of what might be expected from the harvested sites of this study. In spruce-fir forests near Prince George, B . C . , growing season temperatures at the -5 cm level were as much as 5 ° C higher on sites with manual ly removed vegetation than on untreated sites. These temperatures were between 1 0 ° C and 1 5 ° C on the 70 untreated site, and between 1 2 ° C and 2 1 ° C on the clipped site. The pattern was repeated at the -20 cm level, but wi th lower temperatures ( 7 ° C to 15°C) for both conditions and narrower differences over the season (Dobbs and M c M i n n , 1977). Studies at lower elevations on the B . C . coast have yielded similar patterns and results to the foregoing (Black, 1982). The influence of temperature and temperature regime changes on tree growth and other physiological processes is also important in this context. Changes enhancing microbial activity might not necessarily lead to enhanced tree uptake and growth. Fo r example, marked microbial act ivi ty and decomposition have been observed at sub-zero temperatures (Soderstrom, 1979) and under snow cover (Edmonds, 1980) in northern coniferous forest soils. B y contrast, according to Pri tchet t (1979), the min imum range for root growth is from above freezing to 7 ° C , wi th optima from 1 0 ° C to 2 5 ° C . Root growth apparently ceases once the soil freezes (Vogt and Gr ie r , 1982). Species from cooler environments tend to have lower ranges and optima than those from warmer ones (Pritchett, 1979; K r a m e r , 1983). According to T i m m e r and Weetman (1969), shoot growth can occur even when roots are frozen. L o w air temperatures would constrain such growth; Rumney (1968) stated that vegetative growth seldom takes place below approximately 6 ° C . T i m m e r and Wee tman (1969) noted that 7 . 2 °C was usually taken as the threshold temperature for the ini t iat ion of root act ivi ty. The activity of ectomycorrhizal fungi (which are so v i t a l to the efficient functioning of northern coniferous trees) may also be constrained by relat ively low temperatures. A n in vitro study of 47 ectomycorrhizal species by Dennis (1985) placed their opt imum temperature range as between 1 5 ° C and 3 0 ° C ; none grew below 5 ° C . 71 However , for Pacific silver fir [Abies amabilis (Dougl. ex Loud.) Forbes], Vogt and Gr i e r (1982) observed high levels of active mycor rh iza l root biomass even at 1 ° C under a snowpack. There is also the question of reduced water uptake by and flow wi th in plants at low soil temperatures. Cr i t i ca l temperatures of 1°C to 3 ° C have been observed. Fo r Engelmann spruce i n its na tura l environment, 7 . 5 ° C was the cr i t ical temperature below which increased flow resistance was noted (Whitehead and J a r v i s , 1981). Dobbs and M c M i n n (1977) suggested that the opt imum soil temperature for this species may be 2 0 ° C . Increased flow resistance in Pacific silver fir—also an upper-elevation species—occurred at soil temperatures below 5 ° C (Hinckley et al, 1982). A recent study has shown that in subalpine fir, root resistances to water uptake increase significantly below 10° C. A t a given temperature, these increases were greater than those observed in Enge lmann spruce (Sowell, 1985). The above notwithstanding, it was expected that the subalpine fir" advance regeneration should be at least potentially able to benefit from temperature regime enhancements. While a general threshold temperature for shoot growth of 5 ° C is frequently applied, Wor ra l l (1983) showed that high-elevation provenances of subalpine fir can have low threshold temperatures—less than 3 ° C — a s wel l as low heat sums for bud burst. F r o m their own and other studies, Dobbs and M c M i n n (1977) concluded that treatments which could increase mean rooting zone growing season temperatures to within the 10 - 2 0 ° C bracket would be beneficial for planted spruce seedlings. This generalization could probably be applied to 72 subalpine fir advance regeneration also. The foregoing discussion has centered on the possible beneficial effects of increased temperatures through harvest ing. However , it should be noted for completeness that adverse effects can also occur—especially in regard to seedling growth. Ba l l a rd (1972), B a l l a r d et al. (1977), and Black (1982) al l noted that detr imental surface temperatures can be encountered on exposed or clearcut surfaces under certain conditions. In part icular , Ba l l a rd et al. (1977) explored the mechanisms responsible for this. Ea r l i e r , T i m m e r and Weetman (1969) discussed s imi la r issues, especially in terms of adverse effects of high temperatures on microbial populations. 3.1.2 Moisture A s wi th temperature, the p r i m a r y interest in moisture regime influences on the sites was in relation to their possible effects on decomposition/ mineral izat ion processes, nutr ient uptake, and tree growth. The location of the sites wi th in a region of higher precipitation would point towards a conclusion that moisture should not be a l imi t ing factor for tree growth (Lloyd, 1982b). Th is is supported by the consideration that forest floors of E S S F forests have much higher moisture retention capabilities than those of other cover types (Hi l lman and Golding, 1981). However , Monchak (1982) observed that under certain conditions i n spruce-fir zones of the Kamloops Region, water availabil i ty could be the most important l imi t ing factor to post-harvest growth of subalpine fir advance regeneration. O n cutovers, water stress was usual ly associated wi th one or more 73 of (a) growing season precipitation below 250 m m , (b) coarse-textured soils, (c) shallow rooting depths, and (d) low-elevation southern aspects. Wa te r stresses were not lowered sufficiently to permit good growth unti l g rowing season precipitation levels were greater than 350 m m . Moreover, high water stress was found to be one of two major causes of N shortages in the advance regeneration (Monchak, 1982). Though both the Boss Moun ta in and Hemp Creek precipitation normals ( A . E . S . , 1982a) indicate that the sites receive more than 250 m m , it is plausible that harvesting effects could induce moisture stress conditions—if only temporari ly. Thus, even wi th the decreased post-harvest demand for water (by vegetation removal), the complete absence of growing season moisture stresses on the trees cannot be routinely assumed. There is also the further consideration that forest floors—where the majority of the fine roots and mycorrhizae might be encountered (Kimmins and Hawkes , 1978; Vogt and Grier , 1982)—might experience extreme drying in addition to the elevated temperatures discussed earlier (Potts, 1985). This might promote morta l i ty of such roots, which in turn has been observed to cause N deficiencies (Whitney and T i m m e r , 1983). Mahendrappa and Salonius (1982) speculated that dry periods might contribute to enhanced nitrification through the reduction of competition for N H J - N by roots and heterotrophs. Wha t soil water content values are associated wi th l imited microbial act ivi ty and growth? According to Bol len (1974), the opt imum for microbial act ivi ty i n forest soils is approximately 50% of soil water-holding capacity, wi th a m i n i m u m and m a x i m u m at five and eighty per cent, respectively. Fo r fungi, hypha l act ivi ty appears more strongly influenced by moisture fluctuations than by 74 temperature, and high water contents often do not have major effects (Soderstrom, 1979). Short- term fluctuations in bacterial populations can be significant; increases have been associated more wi th rainfall events than wi th actual moisture contents (Clarholm and Rosswal l , 1980). In terms of the soil water matric potential, sources such as Griff in (1972), Russel l (1973), M . Alexander (1977), and Swif t et al. (1979) aided development of the generalizations which follow. Fung i can exhibit vegetative growth over a surprising range of values, principal ly from almost water-logged conditions down to a potential of -15 M P a ; some genera go as low as -50 M P a . Thei r opt imum range is between -5 M P a and -10 M P a . Bacter ia l growth can occur over a slightly narrower range, down to -10 M P a ; however, the opt imum range occurs between -0.1 M P a and -0.01 M P a . Decomposition has been observed wi th in the range zero M P a to -5.0 M P a , w i th an opt imum between -0.05 M P a and -0.01 M P a . Ammonificat ion has been noted over as wide a range as fungal growth, though at increasingly slower rates at potentials below -1.5 M P a . The opt imum range for this process appears to be between -0.09 M P a and -0.05 M P a - generally between 50% and 75% of moisture holding capacity. Ni t r i f ica t ion has a very narrow range, from almost zero M P a to only -0.5 M P a approximately; optima generally fall near field capacity—usually wi th in the range 50% to 65% of moisture holding capacity (Griffin, 1972; Russel l , 1973; M . Alexander , 1977; Swift et al., 1979). Whi le the matr ic potential at which permanent wi l t ing occurs is known to va ry wi th the species involved, the above generalizations indicate clearly that decomposition and ammonification can proceed under conditions unfavourable to growth and uptake 75 by higher plants. The lower l imi t of soil and plant water contents at which growth becomes l imi t ing is of p r imary importance in this context. B u c k m a n and B r a d y (1969) implied that plant growth opt ima are generally found between 15% and 50% of available water storage capacity ( A . W . S . C . ) . t Hase and T i m m e r (1982) found that opt imum growth of black spruce in the Ontario clay belt occurred wi th in the soil-moisture range of 20% to 30% by volume. It is generally accepted that the tension with which the water is retained in the soil (i.e. matr ic potential) is more important to plant growth than actual water contents. Moreover , while the permanent wil t ing point (P .W.P. ) differs among species, the -1.5 M P a level is accepted as imposing at least severe stress on most plants (Buckman and B r a d y , 1969; Russell , 1973; Pritchett , 1979; Black and Spittlehouse, 1982; Nobel , 1983). Fo r forestry applications, the -0.2 M P a level m a y be considered as imposing moderate stress (Black, 1982; B lack and Spittlehouse, 1982). Black (1982) also considered the period dur ing which the soil volumetric water content remained below 40% of A . W . S . C . to be indicative. The water potentials wi th in the plants involved and their reactions to stress are key factors here. Accord ing the Whitehead and J a r v i s (1981), for conifers there is a non-linear—often exponential—relationship between leaf and soil water potentials wi th dry ing . Chaney (1981) i l lustrated this wi th data from southern pines. General ly, as the soil dries out without replenishment, its water potential wi l l converge wi th those of the leaves/needles. Fur the r d ry ing of the soil t A . W . S . C . is considered to be moisture held between field capacity and the permanent wil t ing point (Pritchett, 1979; Ba l l a rd , 1981). 76 induces permanent wi l t ing; this usual ly occurs in the range -1.2 M P a to -1.5 M P a (Russell, 1973; K r a m e r , 1983; Nobel , 1983), but can be below -2.0 M P a in several tree species (Bal lard , 1981). Throughout the dry ing process down to the P . W . P . , it appears that leaf and soil water potential values usual ly remain within -0.5 M P a of each other (Chaney, 1981; K r amer , 1983; Nobel , 1983). A t what plant water potential values might subalpine fir representatives be considered under water stress? A n y water flow resistance changes would of course be influential here. Pur i t ch (1973) examined such aspects for four true fir species including subalpine fir. The patterns of photosynthesis, respirat ion, and transpirat ion wi th increasing stress were a l l very similar . O f the four species, subalpine fir was the first to show declines in photosynthesis, but its rate of decline was more gradual than a l l except grand fir [Abies grandis (Dougl. ex D . Don) L ind l . ] . The decline started at a needle/twig water potential of approximately -0.9 M P a , and reached 50% of the init ial well-watered rate at approximately -2.9 M P a . Respirat ion declines started earlier at -0.7 M P a ; a constant m i n i m u m level of approximately 60% of the ini t ia l rate was attained at approximately -2.0 M P a . Transpira t ion losses commenced at the same time, but followed a slightly steeper path than photosynthesis. A l l species continued their moisture losses at between 10% and 30% of their ini t ia l rates even under severe stresses. In light of the earlier considerations, Puri tch 's (1973) data suggest that subalpine fir might be considered under moderate water stress at soil water matric potentials of about -0.1 M P a , and definitely under severe stress by -1.5 M P a . The observations of Hinck ley et al. (1982) support this view. These authors concluded that in the Pacific North-West , true firs tend to be poorer drought 77 resisters than their tree associates. Subalpine fir was ranked as having a medium capability for drought avoidance and a relat ively low drought tolerance (Hinckley et al, 1982). In southeastern B . C . , the species' drought avoidance strategy, moisture stress sensit ivi ty, and resultant growth losses add to the disdain wi th which it is regarded. Relatively few studies have investigated the moisture retention characteristics of forest floor layers, especially at higher elevations and/or on harvested sites. Plamondon (1972) reported concerning the characteristics of an undisturbed forest floor in coastal B . C . ; the L , F , and H layers were one, seven, and nine centimetres thick respectively. The water retention characteristics varied wi th depth and the degree of decomposition of the layers . The lower portion of the H layer retained more water than any other, seemingly because it contained a higher proportion of micropores. General ly, very little additional water drained between matric potentials of -0.1 M P a and -1.5 M P a . The A . W . S . C . was in the vicini ty of 12% and 29% in the upper F and lower H horizons respectively (Plamondon, 1972). Plamondon (1972) concluded that forest floor water retention characteristics can have a significant effect on plant growth. This was clearly demonstrated by Potts (1985) for a harvested site wi th in a subalpine fir forest type in western Montana . He observed extreme water potential values (e.g. -100 M P a ) in organic layers which averaged five centimetres i n thickness. In the foregoing presentation, the major influences of temperature and moisture on decomposition processes and tree growth were discussed. Whi le these two factors were considered separately, it should be noted that their effects are 78 more often interactive. The increased flow resistance wi th decreasing temperatures discussed earlier was an example of this. Schlentner and V a n Cleve (1985) noted unique peaks in soil respiration in the 10 - 1 5 ° C temperature and 100 - 150% (by weight) moisture ranges at the 15 cm depth wi th in the forest floor. (If a bulk density of 0.13 M g m " 3 is assumed, the moisture range would be approximately 13% to 20% by volume.) The activity of fungi has been shown to be highly influenced by such interactions (Swift et al, 1979). H a s k i n (1985) concluded that i n her study the interaction between temperature and soil moisture m a y have imposed additional N availabil i ty constraints on trees. 3.1.3 Objectives and Hypotheses The ma in objective of this portion of the study was to attempt to provide answers to questions which arose from the issues discussed in the previous section. N a m e l y , what general growing season temperature and moisture content values actual ly occur on the sites? H a d harvesting induced marked changes in the growing season temperature and moisture regimes (including moisture retention characteristics)? W h a t differences, i f any, existed in the temperature and moisture regimes among the sites? Were observable differences l ikely to result in differences in N avai labi l i ty and tree growth? Were there any marked and/or sustained growing season water stresses on the advance regeneration on any of the sites? The resources available allowed only a rudimentary approach to this. Thus, it was decided that an examination of forest floor temperatures, moisture contents and retention characteristics, as well as air temperatures during the peak period to the growing season should provide some valuable insights. The 79 Boss Mounta in cl imatic data ( A . E . S . , 1982a,b) indicated that J u l y and Augus t were normal ly the only months wi th air temperatures greater than 1 0 ° C ; 60% of the annual growing-degree-day total occurred dur ing these two months. Moreover, J u l y and Augus t were normal ly the growing season months wi th the lowest precipitation; M a y was actually lower, but its low growing degree-day value was believed to exclude it from the actual growing season. The J u l y - Augus t period was therefore assumed to cover the peak of the growing season on the sites. The hypotheses under test here are subsidiaries of ma in hypotheses H 0 o and H 0 g (see Chapter One). They are as follows: H 0 : There was no change in the growing season temperature regime after harvest ing. H , y There was indeed some change in growing season temperature regime after harvest ing. H 0 ^: There was no major change in forest floor moisture retention characteristics after harvest ing. H t There was a major change in forest floor moisture retention characteristics after harvest ing. H 0 : There was no change in the growing season soil moisture regime after harvest ing. H , ^: There was some change i n the growing season soil moisture regime after harvest ing. 80 H 0 : There were no sustained growing season water stresses on subalpine fir advance regeneration after harvest ing. H , ^: There were sustained growing season water stresses on subalpine fir advance regeneration after harvest ing. I f H 0 is true, there should be no differences in the mean air and forest floor temperatures of the cutovers compared to those of the mature stand during the growing season. Coincidence or otherwise of 95% confidence intervals was taken as the cri terion of falsification. If H 0 ^ is correct, there should be no marked differences in moisture retention curves constructed for the forest floors of the mature stand and the cutover sites. Statist ical tests for differences in mean values generated at each common pressure point along the curves were accepted as the general falsification criteria. However, it was accepted that a weak stat is t ical significance might have no practical relevance. Hypothesis H 0 was O considered in s imi lar fashion to hypothesis H 0 ^, wi th a s imi lar criterion of falsification. Hypothesis H 0 was l ikely to be true only i f there was no profound period of the growing season during which soil water matr ic potentials were at or below -1.5 M P a . This was therefore accepted as the criterion of falsification. N o formal hypotheses concerning direct temperature l imitat ions on growth could be satisfactorily tested in this study. Nevertheless, it was felt that temperatures fal l ing below the opt imum ranges indicated earlier would provide some valuable insights. For example, mean air temperatures of less than 5 ° C and/or forest floor temperatures below 1 0 ° C (and especially below 7 ° C ) could be construed as 81 probably imposing some limitations (direct or indirect). The resources available l imited pr inc ipal data collection to one growing season only. The results therefore have to be interpreted in light of the degree to which that period could be considered c l imat ica l ly "normal" for the sites. 3.2 METHODS The basic site layout was presented in Chapter Two. Measurements were taken p r imar i ly dur ing the Ju ly -Augus t period of 1983. Addi t ional measurements of forest floor temperatures for comparative purposes were done i n mid-October of that year . 3.2.1 Temperature Soil temperature measurement methods are discussed by Taylor and Jackson (1965) and Black and Spittlehouse (1982). Forest floor temperatures were measured using an Atk ins Model 44000-C thermistor thermometer. Four thermistor probes were randomly installed at the F / H interface of the forest floor on each site. This corresponded to average actual installation depths of approximately four to seven centimetres. The F / H interface was chosen as a level of s tandard biological conditions wi th in the highly variable forest floors. The thermistors were a l l 5 K beads potted in A r m s t r o n g C-7 epoxy. Each was mounted on approximately two metres of Belden two-conductor, shielded, stranded wire (A. W . G . 22). Most of the cable was buried, leaving only approximately 82 0.30 m exposed for taking readings. Before instal lat ion, the probes were a l l checked for l inear i ty and accuracy against a F isher mercury-in-glass thermometer w i t h a precision of 0 . 2 5 ° C . No calibration was found necessary. Probe readings were obtained to the nearest 0 . 1 ° C . A i r temperatures were taken from the central ly located max imum-min imum thermometers. These were read to the nearest 0 . 5 ° C , and at the same time as the forest floor temperatures. The mean temperature was taken as the average of the m a x i m u m and min imum readings for the measurement interval . Both forest floor and air temperatures were measured at i rregular intervals; with a few exceptions, the intervals were two to three days on the average. 3.2.2 Moisture Samples from the F / H interface of the forest floor wi th in the 400 m 2 central area of each site were taken for gravimetr ic determination of moisture contents (L. A . Richards, 1965). Moisture cans wi th volumes of 325 c m 3 and 174 c m 3 were used. A t each sampling event, 12 samples (each fil l ing the can) were taken systematical ly at approximately one-metre intervals in the down-slope direction. The first such line started approximately 1.5 m from the upper left-hand corner of the area. Each subsequent line was positioned paral lel to the previous one at a distance of one metre. Sampl ing was at irregular intervals , but was at least weekly. Since Ju ly , 1983 was an extremely wet period on the sites, it was felt that moisture sampling could be relaxed in favour of more pressing field requirements. The samples were oven-dried at 1 0 5 ° C for approximately 24 hours commencing in the evening of the collection day. 83 Tensiometer measurements (S. J . Richards , 1965; Black and Spittlehouse, 1982) were also taken on the sites. E igh t tensiometers of column lengths 15 cm and 30 cm were installed wi th their cups at the H/mine ra l interface or as near below that level as was feasible. E a c h instrument was equipped wi th a vacuum guage graduated in centibars (0.001 M P a ) . Two tensiometers were installed on the T M site, and one on each cutover. In addition, two others were installed lower on the same slope outside of the cut areas. Installation was within the 400 m 2 area where applicable. The tensiometer network thus covered the entire research area. Unfortunately, two events great ly detracted from the network's usefulness. F i r s t l y , supply problems delayed instal lat ion unt i l the second week of Augus t . Secondly, four of the instruments were subsequently destroyed—apparently by wildlife; these included both of those i n the mature stand wi th in 24 hours of instal lat ion. Readings were taken to the nearest 0.001 M P a , and with a s imilar frequency to those of the gravimetr ic sampl ing and temperature measurements. The instruments were lightly recharged after each reading. Mois ture retention curves were constructed for the lower forest floor portions using a pressure plate procedure[L. A . Richards, 1965; Soilmoisture Equipment Corporation (n.d.)a,b]. It was unfeasible to produce curves for al l the sites; the T 6 site was chosen for sampl ing as the most disturbed [on the basis of v i sua l ly evaluated mineral soil exposure (Bockheim et al., 1975)], and also as the chronologically central site. Its comparison wi th the mature stand would indicate the most extreme differences possible wi th in the chosen sequence. Fo r each of the T M and T6 sites, ten samples were selected randomly from those points which yielded intact forest floor F / H squares (see Chapter Four) . 84 Cyl indr i ca l metal rings of approximate in ternal volume 68 c m 3 were used to extract "undisturbed" samples for placement on the pressure plates. The procedure was such that H-layer mater ia l was in contact wi th the plate, w i th the F- layer material (if included) above. Samples were allowed to imbibe water for 24 hours prior to the start of a pressure sequence. The -0.01, -0.033, -0.10, -0.20, -0.40, and -1.50 M P a points were used to define the curves. The Students's f-test and the non-parametric median test were used to check for differences between the two curves for these points (Gibbons, 1971; Hollander and Wolfe, 1973; Sokal and Rohlf, 1981). The Mich igan Interactive D a t a A n a l y s i s Sys tem ( M I D A S ) statistical programme (Fox and Guire , 1976) at U . B . C . was used. 3.3 RESULTS AND DISCUSSION 3.3.1 Temperature Black and Spittlehouse (1982) advised that hourly readings are necessary to determine soil temperature ranges at the 15 cm depth. Therefore, the results of the rudimentary approach taken in this study should be taken as merely indicative rather than authoritative. The means and 95% confidence l imits of air and forest floor temperatures on the five sites for the measurement period are presented in Figure 3.1. The mean air temperature under the mature stand was 1 1 . 6 ° C , while those of the cutovers ranged from 1 4 . 4 ° C to 1 7 . 6 ° C . The three youngest cutovers clearly have significantly higher air temperature means than the mature stand, but they do not appear to differ from each other. The mean 85 20n CD 15 10 0 T t 0 t 0 v « r t l c o l b a r s o r « 95% c o n f i d a n c * limits f 5 J MATURE STAND CUT AGE = 3 CUT CUT AGE AGE = 6 =8 CUT AGE = 11 • For»s1 rioor (J/H) O Air (+5 cm) SITE / CUTOVER Figure 3.1 Means and 95% confidence limits for the air and forest floor temperatures of the sites, July to August 1983. 86 of the oldest (Ti l ) cutover is intermediate, and does not appear to differ significantly from the means of either the mature stand or the youngest two cutovers. The lower mean of the T i l may be related to its higher stem density compared to the other cutovers (see Table 2.5); however, increased canopy height and crown closure are also possible important causes. Forest floor temperatures also follow this general pattern, but at a lower level than the air temperatures (between two and four degrees below). However, in this case, even the mean of the T i l cutover appears to differ significantly from that of the mature stand. The mean forest floor temperature was 9 . 0 ° C for the mature stand, and ranged from 10.7°C to 14 .4°C for the cutovers. Unlike the air temperatures , the T i l mean seems well displaced from those of the T6 and T8 sites. For both air and forest floor temperatures, the trend is toward a definite increase of values after harvesting for approximately eight years. Following this, temperatures are likely to decline gradually as the stand develops. The range of observed air temperatures indicate that this variable is not likely to be seriously limiting growth on any of the sites. However, in this context growth might be expected to be less than optimal, since the highest mean was less than 2 0 ° C ; it is .unlikely that values greater than this were attained. Records of the A.E.S . for the Boss Mountain station indicate that the July-August period was not markedly divergent from the normals (R. McLaren, Climatologist, A.E.S.; pers. comm. July 3, 1986). Because of differences in measurement technique the actual figures from that station are not directly comparable with those observed on the research area. The mean air temperature for the period at Boss Mountain was approximately 1 2 ° C . It can be seen that 87 the observed values of the sites are acceptably close to this; for the harvested sites, they are s imi lar to air temperatures observed after clearcutt ing on Vancouver Island (Black, 1982) and after spacing in Cape Breton (Salonius et at, 1977). Forest floor mean temperatures provide an interesting picture i n terms of their effects as discussed earlier. Growth of advance regeneration in the mature stand may have been sl ight ly constrained by the temperature. However , cutover means were a l l apparently raised by approximately two to five degrees Celsius to wi th in the 1 0 - 2 0 ° C bracket. Thus , there should be some benefits accruing to tree growth as wel l as microbia l act ivi ty. Nevertheless, as with air temperatures, the means fell in the lower end of the optima, and thus growth might be correspondingly less. Interestingly, even the highest observed mean falls s l ightly below the lower boundary for opt imal growth of ectomycorrhizal fungi noted by Dennis (1985). The possibili ty exists that the activity of mycorrh iza l roots might have been strongly temperature-l imited on a l l the sites. While tree growth and uptake might have been rendered sub-optimal by forest floor (soil) temperatures, these would have retarded decomposition/mineralization processes to a much lesser degree. Since there has been evidence that nitrifiers can adapt to low temperatures (Nakos, 1984), neither would nitrification processes have been limited. It is therefore possible that a substantial amount of the nutrients released by these processes was not taken up by the advance regeneration—especially given the low densities and lack of canopy closure. The above considerations must be tempered by the fact that observations were l imited to the forest floor, which m a y not have been the main zone of root function. 88 Moreover, measurements covered only a portion of the growing season. The marked differences among the forest floor temperatures indicates that there should be corresponding differences in N mineralization rates. Th is needs to be taken into account where a laboratory incubation method at a constant temperature is used to assess N avai labi l i ty . The forest floor temperatures observed on the sites were s imi lar to those for mineral soils discussed earlier (Dobbs and M c M i n n , 1977; Salonius et al, 1977). They were sl ight ly lower and wi th a narrower range than those observed by Timmer and Weetman (1969). The magnitude and range of the temperature increase after disturbance appeared s imilar to those observed in the latter studies. F ina l ly , the trends of the forest floor and air temperature means were closely parallel . This implies that wi th a more rigorous (but s t i l l simple) data collection system, very significant simple l inear regressions might be successfully developed between these two variables for the growing season. This would serve to simplify studies of this nature in a given locality. 3.3.2 Moisture Because of the extreme var iabi l i ty of soil moisture even over a relat ively smal l area, values were reported rounded to the nearest whole number. Fur ther , the sampling intensities were insufficient to fully reflect the observed var iabi l i ty . The means and 95% confidence l imits of the forest floor (F/H) volumetric water content are presented i n F igure 3.2. It was assumed that no adjustment for coarse fragments was necessary in this case. The bulk densities used for the 89 40 i Vertical b a r s are 95% conf idence limits 30 H 20 H 10 H M*an Content MATURE STAND CUT AGE = 3 CUT AGE = 6 CUT AGE = 8 CUT AGE = 11 o - J SITE / CUTOVER Figure 3.2 Means and 95% confidence limits for the forest floor moisture contents of the sites, July to August 1983. 90 conversion from gravimetr ic to volumetric water contents (Armson , 1977; Black, 1982; H i l l e l , 1982) were the forest floor (F/H) mean bulk densities obtained during another phase of the study (see Chapter Four) . F igure 3.3 contains the moisture retention curves developed for the T M and T 6 sites. F i n a l l y , the means and 95% confidence l imits derived from the individual sampl ing events over the measurement period are given in Figure 3.4. In addition, stress cr i ter ia (the equivalents of the -0.2 M P a and -1.5 M P a values) suggested by Black (1982) and Black and Spittlehouse (1982) have been incorporated in the latter. The mean volumetric water contents of the sites did not differ significantly; they ranged from 20% to 27%. This moisture content is wi th in the lower portion of the broad optimal ranges for tree growth discussed earlier. However , the estimated forest floor matr ic potentials (see later) belied this. The lack of significance implies that there were no differences in the soil moisture regimes of the sites over the period. Thus, differential effects on growth and/or such processes as decomposition, mineral izat ion, and nitrification were unlikely. However, non-significance could also be the result of high within-site var iabi l i ty with low sampling intensity. The mean value on the T6 retention curve (Figure 3.3) corresponding to -0.01 M P a was omitted because it was sl ightly greater than 100% (101%). This could have been due to errors associated wi th the volume and bulk density estimates of the retention samples wi th in the metal r ing. Indeed, the bulk densities calculated for the r ing samples of both sites tended to be higher (range 0.14 - 0.24 M g m~ 3 ) than might be expected for mor humus layers. More l ikely, it could have been from mater ial lost in the removal of the samples from 91 0.001-0 20 40 60 80 100 SOIL MOISTURE CONTENT (% by volume) Figure 3.3 Moisture retention curves for the mature stand and six-year-old cutover. 60-, 0 H 1 1 1 1 1 1 1 r 1 8 15 22 29 5 12 19 26 JULY AUGUST 1983 Figure 3.4 Means and 95% confidence limits for the forest floor moisture contents, July to August 1983. 93 the plates for weighing and/or final drying. Such losses were almost inevitable, par t icular ly after the higher pressure runs. In either case, there would also be a slight systematic upward bias present in the other data points for that curve—and perhaps for the T M curve as well . Th is is not believed to have altered the results or interpretations to any consequential degree. The f-tests showed only the -0.40 M P a and -1.5 M P a points as being significantly different (p — 0.02 and 0.04 respectively). Together wi th possible sources of error inherent in the retention curve construction procedure, such weak statistical significances were not taken as indicating the existence of practical differences. This result differs from Page (1974), who observed decreased moisture retention capabilities in surface layers following clearcutt ing of balsam fir stands i n Newfoundland. The two retention curves of this study were assumed to be v i r tua l ly the same; averages of their two values (at a given matric potential value) were used subsequently to determine volumetric water content values. U s i n g the volumetric water content at -0.01 M P a of the T M site for both cases, the A . W . S . C . of the forest floor layers was calculated as approximately 59%. Rela t ive ly little water loss occurred when matr ic potentials were lowered below -0.20 M P a . Fo r that matter, 78% of the A . W . S . C . was lost between -0.01 M P a and -0.10 M P a . Plamondon (1972) observed s imi lar water loss patterns in forest floor mater ia ls . F r o m representative figures given by Ba l l a rd (1981), the forest floor A . W . S . C . was approximately 1.7 times that which might be expected of a sil t loam—possibly the minera l soil texture exhibit ing the highest A . W . S . C . value. The moisture retention curves were used to convert calculated field volumetric water contents to forest floor matr ic potential values. F r o m Figure 94 3.4, it would appear that forest floor water potential values which might be considered as being stressful were indeed encountered. For al l sites, extreme stresses occurred part icularly dur ing Augus t . F o r the cutovers, stresses were extreme for the entire month; the mature stand had some relief from such stress, but dur ing the second week only. The T 6 site appeared to be the only one which encountered extreme stresses in J u l y ; the three oldest cutovers a l l showed at least moderate stress values for that month. Assuming no difference in moisture retention characteristics, the data from the two sites were combined to give volumetric water content and matr ic potential values for a single retention curve. Fol lowing Campbel l (1974), a least squares linear regression fit of these values on a log-log scale was used to produce an idealized moisture retention curve (Figure 3.5). The equation so derived was: ln (*) = 1 1 . 7 2 - 3 . 5 8 ln (0) where: •!» = forest floor matr ic suction, M P a ; and 0 = forest floor volumetric water content, %. The r 2 value for this equation was 0 .91; however, this is somewhat suspect because of the transformation of the dependent variable. The curve was extrapolated (dashed line) i n order to gain some idea of the forest floor matric potential values at the driest points. The lowest volumetric water contents observed were approximately 17% for the mature stand, 16% for the T3 and T 8 sites, 9% for the T 6 site, and 13% for the T i l site. These corresponded to estimated matric potentials of -5 M P a for the mature stand, -6 M P a for the T3 95 0.001-1 i i i 1—i—i—i—i—i 1 1 1 1— 10 SOIL MOISTURE CONTENT (% by vo lume) 100 Figure 3.5 Idealized moisture retention curve and extrapolation (dashed line) for the combined sites. (See text for regression equation.) 96 and T 8 sites, -47 M P a for the T 6 site, and -13 M P a for the T i l site. The records of the A . E . S . for the Boss Mounta in station show that the Ju ly -Augus t period of 1983 had approximately 1.4 times more precipitation than might be expected normal ly . However , while J u l y had approximately 2.5 times more than normal , Augus t actually had much less—only 0.4 times what might have been expected (R. M c L a r e n , pers. comm., J u l y 3, 1986). This no doubt contributed to the somewhat unexpected result of sustained extreme moisture stress levels in a "Wet Bel t " subzone. The data from surviving tensiometers provide only weak support for the above. Tensiometers are ineffective at soil moisture potentials below -0.08 M p a (S .J . Richards, 1965; Black and Spittlehouse, 1982). This l imi t was only fleetingly reached by one of the two lower-slope instruments outside of the cutovers. If the data are accurate, the moisture stress levels may have been. l imited to the organic layers only. However, it m a y also be that recharging of the tensiometers was excessive, causing an " i r r igat ion" effect which would invalidate the data so obtained (Black and Spittlehouse, 1982). The general finding of stressful moisture values occurring in the forest floor after harvest ing is in agreement wi th the results of Potts (1985) for a drier subalpine site. The latter investigator noted the occurrence of stressful conditions from J u l y on into September. Moreover , as noted earlier, extremely low potentials were observed. It is interesting that w i t h less than half of the samples taken per sampl ing event reported by Potts (1985), the range of variat ion of moisture content values in this study was very s imi lar to that one. The possibility exists that on harvested sites w i th more or less intact forest floors, dry organic surface layers might act as an insulator preventing evaporative loss from the minera l 97 soil. I f so, moisture stress might not have as important a role to play in post-harvest growth as was indicated by this study and that of Potts (1985). A t the observed levels of forest floor moisture, it can be safely assumed that water availabil i ty imposed no constraints on decomposition/mineralization processes. However, nitrification would have been slowed considerably during the Augus t period. Thus, the situation on the cutovers was s imi lar to that noted earlier under temperature. Insofar as they depended on the forest floor layers for a supply of moisture and nutrients, the growth of advance regeneration would have been severely constrained by the Augus t stresses. W i t h no such constraints on the bulk of microbial act ivi ty, i t m a y be that the bulk of nutrients released in the forest floor was not taken up by the advance regeneration. 3.4 CONCLUSIONS W i t h i n the limitations imposed by the rudimentary methods employed, some general conclusions can now be stated in terms of the subsidiary hypotheses under consideration. The results indicate that hypothesis H 0 should be rejected. Therefore, there was indeed a change in growing season temperature levels following harvesting. A i r and forest floor temperatures were collectively increased by some 3 ° C to 6 ° C for at least eight years. Differences existed among the temperature regimes of the sites. There was nothing in the data to suggest that major changes in forest floor moisture retention characteristics and/or soil moisture regime occurred after harvest ing. There appeared to be no differences among the sites in this regard. Hypotheses H 0 9 and H 0 „ were not 98 rejected. However , H 0 can be rejected to the extent that advance regeneration was dependent on forest floor layers . There was clear evidence of extreme and sustained growing season water stresses on subalpine fir advance growth after harvesting. A t the most extreme points, estimated matric potentials between -6 M P a and -47 M P a were encountered in the forest floors of the cutovers. The fact that the organic layers of the mature stand also exhibited low matr ic potentials suggests that low forest floor moisture levels may have been a characteristic of the sites, rather than an effect of harvesting. Concerning possible temperature l imitations on tree growth, it is l ikely that these existed to a smal l degree; growth m a y therefore have been sub-optimal. The results have demonstrated that there can be relat ively minor temperature and major moisture avai labi l i ty constraints to post-harvest growth of advance regeneration—even on such "Wet Bel t" E S S F sites. B y contrast, decomposition and mineral izat ion processes did not appear to be so affected. Therefore, i t is unl ike ly that the subalpine fir advance growth utilized a l l of the nutrients so released. F i n a l l y , one implication of the forest floor temperature differences on the sites is that the results of laboratory incubation procedures should be modified to reflect this in studies of this nature. CHAPTER 4 POST-HARVEST FOREST FLOOR PHYSICAL CHARACTERISTICS The importance of the forest floor and the role which it appears to play after major disturbances was discussed in Chapter One. In this chapter, the post-harvest physical state of the organic layers of the Swayback Ridge sites is examined. Emphas is was placed on humus form classifications and forest floor bulk densities and weights. 4.1 INTRODUCTION A s discussed in Chapter One, a general decline in forest floor depths and weights following disturbance has been documented. A n y investigation of the assart effect after harvest ing of the E S S F forests should thus logically examine such changes. Moreover, an assessment of the weights of the organic layers is necessary to the calculation of nutrient levels on a mass-per-unit-area basis. This section of the study was thus intended to shed light on the dynamics of the assart effect both directly (physical changes) and indirectly (contributions to other phases of this study). The question of humus form classification is also of importance to this study. The term "humus form" is used to indicate a concept, s imi lar to that of the soil profile, to designate natural ly-occurring, biologically active units comprised of the forest floor horizons. The humus form is part of the soil; it m a y comprise a single horizon, but is more usually a profile composed of successive 99 100 horizons l inked together genetically (Dumanski , 1978). It has generally been recognized that the composition, morphological characteristics, and sequence of horizons reflect a var ie ty of biological processes and life forms (and thus a part icular micro-environment). These facets have been used as a basis for humus form classification (Hoover and Lun t , 1952; Bernier, 1968; D u m a n s k i , 1978; K l i n k a et al, 1981). M o r humus forms are generally associated wi th coniferous or heath cover on soils poor in bases and nutrients. Decomposition through fungal action appears predominant; protozoa, collembola, and mites are the m a i n faunal agents. The horizons tend to be strongly acid, wi th slower decomposition and less complete humification; they are sharply delineated from minera l horizons (Bernier, 1968; Russel l , 1973; Pritchett , 1979; K l i n k a et al., 1981). The pre l iminary survey had indicated that mor humus forms were dominant on a l l the sites—as might be expected under coniferous forests in the E S S F zone. Bernier (1968) proposed that mor humus forms be differentiated into four groups—fibri-mors, humi-fibrimors, fibri-humimors, and humimors . K l i n k a et al. (1981) later proposed a differentiation into five groups—velomors, xeromors, hemimors, hemihumimors , and humimors. Both systems use the relative proportions (thicknesses) of the horizons as the principal basis for differentiation; in the case of xeromors, K l i n k a et al. (1981) also applied a moisture deficiency criterion. The view has been expressed that a given humus form would fall into the same general category in either of the two systems (Watts, 1983). However , it can be demonstrated that insofar as the development of an H layer is indicative of differences in the degree of humification, the K l i n k a et al. (1981) system should be more sensitive than Bernier 's (1968) in differentiating among 101 profiles wi th lower humification levels (an expected condition in the E S S F zone); the reverse should hold true for those wi th higher humification levels. It was believed useful to assess which groups (Bernier, 1968; K l i n k a et al., 1981) were dominant on each site. Th is helped to further characterize the sites; i n addition, major departures from a common trend would imply different conditions of microbial activity and N avai labi l i ty . 4.2 OBJECTIVES AND HYPOTHESES The main objectives of this phase of the study were to provide information on forest floor characteristics (bulk densities, weights, and humus forms), and to assess whether any major changes had occurred in the same over time since the harvest. Investigation of physical changes falls wi th in the ambit of main hypothesis H 0 (see Chapter One). The subsidiary hypothesis under test and its alternate m a y be stated as follows: H 0 y N o major changes have occurred in the physical characteristics of the forest floor dur ing the post-harvest period. H , : Major changes have indeed occurred in the phys ica l characteristics of the forest floor dur ing the post-harvest period. If H 0 ^ is true, there should be no major differences in the forest floor (F and H layers) bulk densities and weights of the sites. Moreover , a single humus form group should be the common dominant one on a l l sites. Stat is t ical tests were accepted as the cr i ter ia for falsification of cases. A simple inspection of what mor proportion of sampled points was accepted form differences. 4.3 METHODS 102 H 0 for the bulk density and weight humus group comprised the largest as the cri terion for assessing humus The basic field layout and sampling scheme employed were presented i n Chapter Two . The thicknesses of L , F , and H layers were recorded to the nearest 0.1 cm at the middle of the up-slope side of each of the three forest floor excavations made per sampling point. For each layer , the thickness was taken to be the average of the three readings. The thicknesses were subsequently used to determine the mor humus group classification at each point by the systems of both Bernier (1968) and K l i n k a et al. (1981). A t each point, the presence or absence of rotting wood in the humus form was noted. The air- and oven-dry weights (see Chapter Two) of the tri-sample first stage F / H composite from each sampling point were used to provide estimates of the corresponding air- and oven-dry bulk densities, as wel l as per-hectare weights of the air- and oven-dry F / H horizons. The data for total forest floor depths, bulk densities, and weights thus generated were analyzed as a two-way Model I A N O V A , w i th the individual sites and the rotting wood categories as the treatments. Rot t ing wood was included to aid separation of significant sources of var ia t ion; dur ing the pre l iminary survey, it had been observed that its presence was often associated wi th deeper organic profiles. Because of the imbalance in 103 cells, the A N O V A was carried out using the G E N L I N computer programme available at U . B . C . (Greig and Bjerr ing, 1980). The data were checked, using G E N L I N and M I D A S , as to whether they met the assumptions of a valid A N O V A (Fox and Guire , 1976; Anonymous , 1976). Transformations were necessary to achieve homogeneity of the variances for a l l except the air-dry bulk densities. Natura l - logar i thm transformations were used for bulk densities and weights, while a power function was used for the total depths. F i n a l l y , since the L- layer was excluded from the bulk density and weight analyses, single-classification non-parametric tests for differences among its depths were carried out. The K r u s k a l - W a l l i s and median test procedures (Gibbons, 1971; Hollander and Wolfe, 1973; Sokal and Rohlf, 1981) wi th in the M I D A S programme were employed; in these cases, the sites only formed the treatments. 4.4 RESULTS AND DISCUSSION The means and 95% confidence l imits of the measured and derived variables are presented in Table 4.1. There were no significant differences (p > 0.05) among the means of any of the variables of the five sites. 4.4.1 Total Depths There were no significant differences among the mean depths of the sites [p > 0.05, F = 1.4, degrees of freedom (df) = 4, 308]. Means ranged from 5.7 cm to 6.8 cm. The mean depth of a l l samples wi th rotting wood present was 8.6 cm, whereas in its absence the mean depth was 5.4 cm. The observed 104 Table 4.1 Means and 95% confidence limits for forest floor total depths, F / H bulk densities, and weights. I T E M Depth U N I T S cm S I T E / V A L U E S T M T3 T 6 T 8 T i l (—values rounded; figures in brackets are 95% limits—) 6.8 (8.4,5.7) 5.8 (7.1,4.9) 6.7 (7.5,6.1) 5.7 (6.4,5.2) 6.1 (7.0,5.4) B u l k Densi ty air-dry oven-dry M g / m 3 0.16 (0.17,0.15) 0.14 (0.16,0.13) 0.17 (0.18,0.15) 0.14 (0.16,0.13) 0.15 (0.16,0.15) 0.13 (0.14,0.13) 0.15 (0.16,0.14) 0.13 (0.14,0.12) 0.16 (0.17,0.15) 0.14 (0.15,0.13) Weight air-dry oven-dry M g / h a 90.9 79.7 (103.4,79.9) (91.6,69.4) 83.0 72.8 (94.5,73.0) (83.7,63.4) 87.0 (93.5,81.0) 78.6 (84.5,73.2) 79.4 (85.8,73.5) 71.5 (77.3,66.1) 78.1 (84.9,71.8) 69.7 (75.8,64.1) 105 depths are wi th in the range of depths observed by L l o y d (in litt., A p r i l 18, 1986) for comparable Wet-Belt E S S F stands, and also close to those of spruce-fir forest further north in B . C . (Kimmins , 1974). They are much lower than those reported for B . C . coastal forests (Plamondon, 1972; M a r t i n , 1985), but within the range of those observed wi thin the E S S F forests of the coast/interior transition ( K l i n k a et at, 1982). The values are very close to those of mor humus forms under a var ie ty of other forest types (Wooldridge, 1970; Mahendrappa and Kings ton , 1980). The lack of significant differences among the mean depths of the sites represents a departure from the general trends discussed in Chapter One. Fo r example, organic layers declined wi th in three to seven years after both spacing and clearcutt ing in eastern spruce-fir forests (Page, 1974; Salonius et al, 1977; Piene, 1978). Fol lowing clearcutting on the B . C . coast, M a r t i n (1985) noted a decline in forest floor depths of approximately 33% of the original depth. The possibil i ty exists that high within-site var iab i l i ty m a y have served to mask any "true" differences. Al ternat ively , it m a y be that in that relatively cold environment, the part ia l nature of the harvest resulted in a disturbance which was insufficient to trigger major post-harvest depth declines. 4.4.2 Weights The calculated weights are necessarily overestimates, since no attempt was made estimate the actual proportion of the area covered by floor materials. There were no significant differences among the mean F / H dry weights, whether 106 air-dry or oven-dry values were considered (for both: p > 0.05, df = 4, 287; F = 1.8 and 2.2 for air- and oven-dry cases respectively). A i r - d r y mean weights ranged from 74.2 M g ha" 1 to 92.8 M g ha" 1 , while their oven-dry counterparts ranged from 69.7 M g ha" 1 to 83.0 M g ha" 1 . There appears to be little published information on forest floor weights for B . C . ' s E S S F forests. K i m m i n s (1974) observed a value of approximately 56 M g ha" 1 for forest floors under spruce-fir cover near Prince George; this value was believed high. However , the values of forest floor weights obtaining in spruce and/or fir types v a r y tremendously (Kimmins et al, 1985). In general , it m a y be said that the weights observed in this study are on the higher end of the range found for comparable types in the compendium of K i m m i n s et al. (1985). Especial ly given the possible methodological differences, they are acceptably within the latter range. They are also close to values noted by L a n g et al. (1981) in eastern subalpine balsam fir forests. However, they are far below the weights reported by M a r t i n (1985) and H a s k i n (1985) for B . C . coastal and dry interior conditions respectively. Concerning the apparent lack of differences among the sites in their mean F / H weights, what was said above for depths is also applicable here. However , it should be noted that other studies have reported either slight or no differences after disturbances. Fo r example, L a n g et al. (1981) found no pattern or trend in forest floor weight wi th stand age. Fo r hardwood forests in Nova Scotia, Wallace and Freedman (1986) pointed out that they did not observe the large weight losses and pattern of recovery so exemplified by the Hubbard Brook study. The authors attributed their observed differences more to changes in litter inputs than 107 to decomposition differences; this was s imi la r to earlier observations by Piene (1978) and Covington (1981). Since, in m y study, the L- layer was excluded from weight considerations, it might be argued that the "miss ing" weight losses could have occurred there. Piene (1978) noticed a substantial decrease in L- l aye r weights three years after spacing; F and H layers were not so affected. However , such an event was unl ike ly in this study given that no depth differences occurred. Fur ther , on the assumption that oven-dry bulk densities of L- l aye r materials were not markedly different, any weight losses would have been reflected in L- layer depth differences. The non-parametric tests (used because of an insoluble heterogeneity of variances) found no evidence of such differences in the L- layer alone. 4.4.3 Bulk Densities A s wi th the other two related variables, no significant differences were found among the mean F / H bulk density of the sites, whether air- or oven-dry values were considered (for both: p > 0.05, df = 4, 287; F = 1.9 and 1.5 for air- and oven-dry cases respectively). M e a n air-dry bulk densities ranged between 0.15 M g m " 3 and 0.17 M g m " 3 ; the mean oven-dry values ranged from 0.13 M g m " 3 to 0.14 M g m " 3 . The oven-dry values are very close to the general bulk density values found in mor humus forms (Pritchett, 1979). In addition, they fal l at the centre of the range observed by K i m m i n s (1974) under more northerly spruce-fir forests. A s wi th the other two related variables, the lack of differences among the 108 mean F / H bulk densities represents a departure from the general trend. The comments made earlier also apply here. However , the earlier result (Chapter Three) which implied that moisture retention characteristics were unchanged lends some support to the contention that harvest ing may not have sufficiently disturbed the site to trigger major changes. Changes in bulk densities would have been accompanied by changes in retention characteristics. The relat ively narrow range of variat ion in the values also implies that high within-site var ia t ion did not play an important "masking" role wi th respect to any "true" differences. The 95% confidence intervals of the sites w i th the greater numbers of samples are very close to that reported by Plamondon (1972); the latter's in terval was based on a sample size more than double that of the highest in this study. 4.4.4 Mor Humus Groups The velomors of K l i n k a et al. (1981) did not occur„ on the Swayback Ridge sites. Moreover, the results of Chapter Three notwithstanding, no profiles were found which corresponded to the xeromors (K l inka et al, 1981). S i m i l a r l y , no profiles corresponding to Bernier 's (1968) humimors were noted. The results are presented in Table 4.2. For the K l i n k a et al (1981) system, hemihumimors are by far the dominant group on a l l sites. W i t h the exception of the T i l site, hemimors occupied a greater proportion of sampled points than humimors. There was a noticeable decline in the percentage of hemi-humimors found on the T3 site as compared to the others, wi th a corresponding increase in the other two groups; for the latter two, the increase was more 109 Table 4.2 Distr ibut ion of mor humus groups on the sites. G R O U P S I T E / P E R C E N T A G E S T M (n = 30t ) T3 (n = 30) T 6 (n=100) T8 (n = 90) T i l (n = 70) K L I N K A et al. hemimors 16.7 26.7 8.0 12.2 4.3 hemihumimors 80.0 63.3 87.0 82.3 88.6 humimors 3.3 10.0 5.0 2.2 7.1 unclassified - - - 3.3 -B E R N I E R fibrimors - - - 1.1 -humi-fibrimors 86.7 86.7 82.0 76.7 88.6 fibri-humimors 13.3 13.3 18.0 20.0 11.4 unclassified - - - 2.2 -IDenotes sample size. 110 marked in the hemimor group. This pattern implies that the T 3 site had a relatively lower degree of humificat ion overal l than the others; conceivably, this could be explained by the fresh inputs of slash and other organic matter w i th felling. If this is true, then the re turn to the general pattern later in the age sequence might imply that this "anomaly" disappeared wi th in six years after harvesting. In addition, the values from the T i l site could then be taken as indicating a trend toward more complete humification of the mater ia ls . These observations are without statist ical support, since no such tests were carr ied out on the data. Moreover, Federer (1982) pointed out that results of such forest floor studies can be seriously affected by the sampling procedures employed and the subjectivity of the investigator—even when only one individual is involved. However , if the T 3 site indeed differed from the others as indicated, this is l ikely to show up at the chemical and/or microbially-affected levels (e.g. in N mineralization studies). The same m a y also be true of the T i l site, though the differences here appeared lesser than for the T3 cutover. The results using the Bernier (1968) system are very consistent, w i th no apparent differences in the classification of the T3 or T i l sites as w i th the K l i n k a et al. (1981) system. The humi-fibrimors comprised the dominant group, followed by the fibri-humimors. O n l y one sample point was classified as hav ing a fibrimor. The dominant groups of both systems may be seen as ly ing at the lower end of the humification scale. It is therefore believed that the differentiation among sites is a result of the higher sensit ivi ty of the K l i n k a et al. (1981) system to such conditions, as pointed out earlier. I l l 4.5 CONCLUSIONS Some general conclusions can now be made in terms of the hypothesis under test. F r o m the statist ical cri teria, there was nothing i n the data to indicate that H 0 should be rejected. It is therefore concluded that no major changes occurred in the physical characteristics of the forest floor layers of the sites during the post-harvest period. It is possible that high within-site variation-masked any "true" differences; however, it is believed that the par t ia l nature of the harvest constituted a disturbance which was insufficient in that biogeoclimatic subzone to trigger the marked forest floor declines noted elsewhere. A n y decline would arise from decreased litter inputs simultaneous wi th increased decomposition rates, both compared to a mature stand condition. Li t ter inputs would not have been totally cut off by the harvest, and m a y have been augmented by those of herbs and other pioneer vegetation. One implication of the above conclusion is that decomposition rates did not increase so much over those of the mature stand that they could outstrip litter input rates. The humus form classifications yielded some evidence (though not supported statistically) of minor physical changes, internal to the forest floor, w i th in the first six years following harvest ing. Ini t ia l ly , inputs from felling appeared to car ry the organic layers to a somewhat lower overal l stage of humification; however, this effect disappeared wi th in six years of harvesting. If this is indeed the case, differences in the forest floor chemistry and mineralization levels of the sites should also emerge in the subsequent phases of this study. CHAPTER 5 POST-HARVEST SOIL CHEMICAL CHARACTERISTICS The physical characteristics of the forest floors were examined i n Chapter Four. We now turn our attention to what post-harvest changes might have occurred in both the forest floor and minera l soil in terms of their chemistry. Emphasis was placed on those aspects considered most important to tree growth and nutrient avai labi l i ty—soil reaction, macro-nutrient levels and C / N ratios. 5.1 INTRODUCTION The theory of the assart effect involves increases in the availabili t ies of nutrients—especially those wi th large fractions held in organic forms. Thus , it was necessary to examine the basic chemistry of the organic and minera l layers. The aspects of greatest interest involved those variables and processes which affect tree growth and nutrient avai labi l i ty directly or indirectly. Th is phase of the study was therefore intended to provide information on general macronutrient (N , P, K , C a , Mg) levels, soil reaction, and C / N ratios, and also to investigate the existence and dynamics of an assart pattern in terms of these quantities. The special case of N avai labi l i ty is dealt w i th in detail in Chapter S ix ; therefore, only "total" N is considered here. Fo r the other macronutrients, some emphasis was placed on assessing levels of those forms which might have the greatest influence on tree growth. Fo r soil K , C a , and M g , it is generally accepted that their exchangeable pools are the most important ones for plant 112 113 growth (Buckman and Brady , 1969; Russell , 1973; Pritchett , 1979). Therefore, only extractable forms of these nutrient elements were considered. The forms of P present and their avai labi l i ty to and uptake by plants are influenced by a variety of factors, of which soil reaction is one of the more important . Under acid forest soil conditions, the ma in mineral form of P involved should be the p r imary orthophosphate ion, H 2 P O i ; ; this is the form in which plants are believed to absorb the bulk of their P requirements (Russell , 1973; Tisdale and Nelson, 1975; Pri tchett , 1979). Measures of both "total" and plant-available P were included. The plant-available measure used was an in situ ion-exchange resin procedure, and was therefore a cumulative index. Ion exchange resins have a long history of use in the measurement of available P in soils (Amer et al, 1955; His lop and Cooke, 1968; Sibbesen, 1977). Krause (Professor, Un ive r s i ty of N e w Brunswick , pers. comm. June 6, 1986) indicated that excellent results were so obtained in forest soils. The organic matter contents of soils influence many of their physical and chemical characteristics, and is thus of some importance to their complete characterization. Organic carbon (C) content determinations are closely related to this; conversion factors have frequently been used to obtain one quanti ty from the other (MacDonald , 1977; Nelson and Sommers, 1982). Because of methodological problems and the concomitant high var ia t ion in results obtained, Nelson and Sommers (1982) strongly advocated the reporting of organic C as a measure of soil organic matter content. Such an index was employed here. F i n a l l y , C / N ratios have been accepted as strongly influencing net N mineral izat ion rates. Th is st i l l appears to be generally true for forest soils, 114 though there has been accumulating evidence that the latter situation may present a more complex picture than that in i t ia l ly proposed based on ideas from agricultural soils (Berg and Staaf, 1981; Bosat ta and Staaf, 1982; Berg and Ekbohm, 1983; Bosat ta and Berendse, 1984). The ratios have therefore been included as an important characteristic of both the forest floors and minera l soils of the study sites. 5.2 OBJECTIVES AND HYPOTHESES The first objective of this phase of the investigation was to provide information on the levels (concentrations and per-hectare contents) of "total" N and P , "avai lable" P , exchangeable K , C a , and M g , and f inal ly p H and C / N ratio values. The materials of interest were the F / H layers of the forest floor and the 0-15 cm portion of the under lying minera l soil. Secondly, this information was to be used to assess whether there were differences among the sites in terms of their general post-harvest soil chemistry. Investigation of changes in the chemical characteristics of the sites comprises a part of the testing of ma in hypothesis H 0 . M a i n hypothesis H 0 1 o is also indirectly related; however, subsequent chapters deal more directly and adequately wi th that aspect. The subsidiary hypotheses under test and their alternates m a y be stated as follows: H 0 y No major changes have occurred in the general chemical characteristics of the forest floor dur ing the post-harvest period. 115 H , : Major changes have indeed occurred in the general chemical characteristics of the forest floor dur ing the post-harvest period. H 0 No major changes have occurred in the general chemical characteristics of the minera l soil surface fraction during the post-harvest period. H 1 : Major changes have indeed occurred in the general chemical characteristics of the minera l soil surface fraction during the post-harvest period. I f H 0 ^ and H 0 ^ are accurate representations of the post-harvest situation, there should be no marked differences among the indicated organic and minera l fractions of the sites in terms of the chemical characteristics stated earlier. The p H , "total" N and P , C, and C / N ratios were considered to be of greater importance here, owing to their comparatively low susceptibil i ty to change on a given site over the short term (e.g. one growing season). The same can be assumed to hold for the orthophosphate-P measures, since the in situ approach takes a cumulat ive measure. B y contrast, concentrations of exchangeable bases were given a lower emphasis because of their ephemeral nature. Statist ical tests for the differences among the sites with respect to these variables were accepted as the falsification cr i ter ia of the null hypotheses. 116 5.3 METHODS The basic field layout, sampling scheme, and system of creating composite samples were detailed in Chapter Two. For both forest floor and mineral soil fractions, a l l analyses and tests in this phase involved the second-stage composites only. 5.3.1 pH Values A randomly selected subset of ten samples from each site was chosen for p H determinations. The same selections applied to both organic and surface minera l soil samples. The weight of mater ia l used was five grams for both organic and minera l subsamples. Determinat ion was done in 0.0 I M C a C l 2 in soil-to-matrix ratios of 1:8 and 1:2 for forest floor and minera l soil materials respectively. Readings were taken on a Radiometer 29 p H meter to the nearest 0.1 p H units. The data were analyzed in each case as a single-classification Model I A N O V A (Sokal and Rohlf, 1981) wi th the sites as the treatments. A posteriori comparisons among the means employed the Student-Newman-Keuls method (Sokal and Rohlf, 1969; Dowdy and Wearden, 1983). The data were checked for violations of the assumptions of A N O V A using the M I D A S (Fox and Guire , 1976; Anonymous , 1976) and G E N L I N (Greig and Bjerr ing, 1980) programmes. No transformations were necessary. The A N O V A and subsequent comparisons were performed using the G E N L I N programme. Pa i rwise f-tests (Sokal and Rohlf, 117 1981) were also performed between the corresponding forest floor and minera l soil values using the M I D A S programme. 5.3.2 Nitrogen and Phosphorus "Tota l " N and P concentrations were determined simultaneously by colourimetric methods on a Technicon A u t o A n a l y s e r following a semimicro-Kjeldahl digestion procedure. In these, N is determined by the indophenol blue method, which makes use of the Berthelot reaction; P is determined from the formation of a phosphomolybdate complex and its subsequent reduction to molybdenum blue (Twine and Wi l l i ams , 1971; Technicon Instruments Corporation, 1978; Bremner and M u l v a n e y , 1982; Keeney and Nelson, 1982; Olsen and Sommers, 1982). Duplicate analyses were carried out on all forest floor samples. The forest floor data were analyzed as a two-level nested A N O V A wi th unequal sample sizes (Sokal and Rohlf, 1981); the individual sites formed the first level, while the samples formed the second. Tukey 's test (Sokal and Rohlf, 1981; Dowdy and Wearden, 1983) was employed for the a posteriori comparisons. Both A N O V A and comparisons were performed using the A N O V A R statistical programme available at U . B . C . (Greig and Osterl in, 1978). M i n e r a l soil data were analyzed as a single-classification Model I A N O V A , w i th the sites as the treatments. A posteriori comparisons were performed using Tukey ' s test. These analyses were carr ied out using the G E N L I N package. Conformity of al l data to the assumptions of a val id A N O V A was checked using the M I D A S and G E N L I N programmes. Arcsine and natural-log transformations (Sokal and Rohlf, 1981) were necessary to achieve homogeneity of variances in the minera l soil N and P cases 118 respectively. No transformations were necessary for the forest floor data. The monitoring of H 2 PO« avai labi l i ty by ion exchange resins was in conjunction wi th the use of the latter for - assessing changes in N H J - N and N O 5 - N levels (see Chapter Six) . In the fall of 1983, 30 nylon mesh bags containing 11 g oven-dry equivalent of F i sher R e x y n 201 (OH) strong base organic anion exchange beads were placed at the F / H interface 30-50 cm from the bases of selected sample trees on each site. A l l "plots" of the T M and T3 sites had resins; selections for resin placement on other sites was by random selection among sample "plots". The mesh size of the beads was between 16 and 50. A l l bags were treated wi th a mercuric chloride solution to prevent microbial interference wi th collected ions; approximately five per cent of the resin exchange sites were so saturated. The bags were retrieved after approximately 12 months in situ. The resins i n their bags were then extracted with 100 m l 2 M KC1 solution after H a r t and B ink ley (1984). The bags were shaken in the solution for approximately 30 minutes, then left to equilibrate for approximately 24 hours. The solution was then gravity-fi l tered through W h a t m a n No . 41 paper. Ext rac ts were kept in cold storage (approximately 1.7°C) unti l analyses could be performed. Aliquots were used to determine H 2 PO« concentrations colourimetrically by means of a Technicon A u t o A n a l y s e r (Technicon Instruments Corporation, 1971a). Resin data were processed stat ist ically in s imilar fashion to the minera l soil P data discussed above. However , the transformed data exhibited a slight anomaly in the tests for homogeneity. The variances were homogeneous by Bart le t t ' s test (Sokal and Rohlf, 1981), but weakly heterogeneous (p = 0.02) according to Laya rd ' s test (Laya rd , 1973; Gre ig and Bjerring, 1980). To 119 circumvent any weak heterogeneity, the nominal significance threshold for the A N O V A and subsequent comparisons was taken to be at p = 0.01 instead i f the customary p = 0.05. 5.3.3 Carbon and C/N Ratios 5.3.3.1 Carbon A subsample of between two and three grams of each forest floor second-stage composite was ground to pass a 0.5 m m sieve. Mine ra l soils were not ground; subsamples were obtained by passing approximately five grams of mater ia l through the 0.5 m m sieve. O f these subsampled fractions, weights approximating 0.05 g and 0.2 g of forest floor and minera l soil respectively were used in the C determinations. O n the assumption that al l C in the samples could be considered organic C , the latter was determined using total C procedures with a L E C O Model 521 high-temperature induction furnace (Nelson and Sommers, 1982). Approx imate ly 0.75 g t in and 1.50 g iron accelerators were added to each sample. The unit employed K O H to absorb the C 0 2 generated; thus, the measure of C concentrations was volumetrically based. B lanks (accelerators only) were used to condition the system before each run. In addition, L E C O one-gram standard samples were included at the beginning and end of each run—each run consisting of approximately 20 samples. Duplicate analyses were performed on a systematically chosen 20% of each run, plus on any samples yielding apparently 120 unusual readings. In cases of noticeable discrepancies, averages of the two readings were taken as the data. Readings were taken on the 1.0 g scale of the instrument, and a correction for atmospheric temperature and pressure applied in the calculation of C concentrations. This calculation was done at the time of the analyses, so that unusual or anomalous readings were easily and immediately checked. The forest floor data were analyzed (using the G E N L I N programme) as a single-classification A N O V A wi th the sites as treatments. Tukey 's test was employed for the comparisons among means. A power transformation was necessary to achieve homogeneity among the variances; the assumptions of A N O V A were acceptably met. However , for mineral soil samples, no transformation sufficed to provide homogeneity of the variances. The data were therefore analyzed us ing the non-parametric Kruska l -Wal l i s and median tests w i th sites as the treatments. There is some conflict regarding the methodology of ca r ry ing out a posteriori comparisons wi th in the non-parametric framework (Gibbons, 1971; Hol lander and Wolfe, 1973; Sokal and Rohlf, 1981). In this study, the construction of 95% confidence intervals using Tukey 's jackknife was believed to give a reasonably acceptable result in this respect (Hollander and Wolfe, 1973), since i t appeared that marked departures from normali ty existed in the data. Detailed discussions of the jackknife and its various applications m a y be found in Sokal and Roh l f (1981), Efron and Gong (1983), and Gregoire (1984). Fo r this investigation, a computer programme for calculation of the Tukey jackknife confidence l imits was wri t ten on the basis of the formulae given by Wold (1974). 121 5.3.3.2 Carbon:Nitrogen Ratios C/N ratios were calculated for each second-stage composite sample using the N and C concentration values obtained. For the organic materials, the average N value from the duplicate analyses was applied. The data for forest floor and mineral soil were analyzed statistically in the same way as outlined above for C, with the exception that no transformations were necessary for the forest floor ratios. 5.3.4 Potassium, Calcium, and Magnesium Exchangeable K, Ca, and Mg were determined by an ammonium acetate (NH„ OAc) extraction method followed by atomic absorption spectrometry (Baker and Suhr, 1982; Knudsen et al, 1982; Lanyon and Heald, 1982; Thomas, 1982). From each forest floor and mineral soil sample, five grams of material were placed in 60 ml of approximately IM NH„ OAc solution with a pH of approximately 7.0. These were then shaken mechanically for one hour, then left standing for at least one more hour. The solution was then filtered through Whatman No. 41 paper into a 100 ml volumetric flask. The material and container were then rinsed with a further 40 ml of N H a OAc in 20-ml aliquots; the leachates were filtered into the same flask. The filtrate was made up to 100 ml with fresh NH„ OAc, then placed in cold storage (1.7°C) to await analysis. Two blanks (NH 4 OAc only) were included with each set. The extracts were analyzed using a Perkin Elmer Model 4000 atomic absorption spectrophotometer. An air-acetylene flame was used for all three elements (Allen et al., 1974; Baker 122 and Suhr, 1982; Lanyon and Heald, 1982). The forest floor K , Ca, and Mg data did not fulfill the assumptions of A N O V A , and no transformation was found which could achieve this. They were analysed using the non-parametric Kruskal-Wallis and median tests, followed by construction of Tukey jackknife confidence limits—in similar fashion as described above (Section 5.3.3.1). Mineral soil data were successfully analyzed as single-classification A N O V A problems, with the individual sites as the treatments. Power transformations of the Ca and Mg data were necessary to achieve homogeneity of the variances. A posteriori comparisons were done by Tukey's method. The statistical analyses were performed using the MIDAS and G E N L I N programmes as earlier discussed. 5.3.5 Unit-Area Elemental Weights In order to obtain estimates of elemental contents on a unit-area basis, estimated mean weights of forest floor and mineral soil (to a depth of 15 cm) were multiplied by the mean concentration data obtained as described in the foregoing Sections. The forest floor weights of the sites were dealt with in Chapter Four. Mean bulk densities of the mineral soil fraction were calculated from the soil profile samples collected for that purpose (Chapter Two). These were applied to an assumed soil volume of 1500 m 3 ha" 1 for the 0-15 cm. soil layer. Since corresponding 95% confidence limits were also calculated, the method 123 used to derive the unit-area weights needs some elaboration. The weights would be derived by simple functions of the form: u = xy where: u = derived elemental mean weight per hectare; x = mean weight per hectare of soil fraction; y = mean elemental concentration. According to Black (1984), the root mean square (R.M.S . ) error is a good estimator of the result ing error in the above function, on the assumption that errors of the components are normal ly distributed. Fo r this case, the R . M . S . error of the derived quant i ty m a y be calculated as (Black, 1984): A u = / [ ( A x ) 2 + ( A y ) 2 ] where: A u = relative error of the derived mean weight; A x = relative error of the soil fraction's weight; A y = relative error of the elemental concentration. Calculated 95% confidence l imits were available for each of the quantities used in the weight derivations. The difference between the upper and lower l imits , expressed as a percentage of the corresponding mean, was used as the expression of relative error. The resultant R . M . S . error was used as the 95% interval of the variable. 124 5.4 RESULTS AND DISCUSSION In the absence of declines in forest floor weights (Chapter Four) , and also because of the method used to derive unit-area weight estimates, emphasis was placed on the concentration data for interpretive purposes. M i n e r a l soil elemental weight estimates had relat ively wide confidence intervals associated wi th them; no doubt this was due to the low number and var iabi l i ty of the bulk density samples taken. 5.4.1 pH Values The trends of forest floor and mineral soil p H values across the age sequence are presented in Figure 5.1. There were ve ry significant differences among the mean p H values for the different sites in both the forest floor and minera l soil (for both: df = 4, 45; forest floor: p < 0 .01, F = 4.1; minera l soil: p < 0 .001, F = 6.9). In the forest floor, al l cutovers except the T 8 had significantly (p < 0.05) higher mean p H values than the mature stand. The means of the cutovers themselves did not differ significantly from each other. The mean forest floor p H was approximately 3.9 for the mature stand, and ranged from approximately 4.1 to 4.3 for the harvested sites. F o r the minera l soil, al l cutovers except the T3 had significantly (p < 0.05) higher mean p H values than the mature stand. The means of the T M and T3 sites were approximately 4.0 each, while those of remaining cutovers ranged from approximately 4.3 to 4.4. W i t h i n the individual sites, ve ry significant differences between forest floor and minera l soil mean p H values were found on the T3 and 125 Vertical bars are 95% conf idence limifs 4 .8 - i 4 . 8 n 3.8 H 3 . 6 J Figure 5.1 Means sequence. and 3 4 5 6 7 8 9 10 11 Y E A R S A F T E R H A R V E S T 95% confidence limits of the pH values across the age 5 7 T 8 l 12 126 T 8 sites only. F o r the T3 site, the forest floor had a higher mean p H than the minera l soil (p < 0.01); however, the reverse was true for the T 8 cutover (p < 0.05). The above results indicate that following harvest ing there appeared to be a clear increase of about 0.3 units in p H values of both the forest floor and minera l soil materials . The increase appeared to occur wi th in three years of the disturbance in the forest floor, but lagged to between three and six years wi th in the minera l soil. The effect generally appeared to be a sustained and constant one, last ing at least over the eleven year sequence examined. The reason for the temporary decline in the forest floor values at the eight-year point is not immediately obvious. It was possibly associated wi th changes in minor vegetation and the concomitant litter input quality. This idea is somewhat supported by the T 8 p H values. The difference between T3 organic and minera l values can be viewed as par t of the lag in the mineral soil changes; however, this does not apply to the T 8 case. A very interesting observation is that at three-year point (T3) was the only one at which the mean forest floor p H value was numerical ly higher than its minera l counterpart. K i m m i n s (1974) reported forest floor layers wi th consistently higher p H values than the underlying minera l soil in spruce-fir stands near Prince George. However, data from L l o y d (in litt., A p r i l 18, 1986) show that in the Kamloops E S S F forests minera l soil surface horizons have consistently higher p H values than the overlying forest floor layers. The impact of the disturbance m a y therefore be viewed as strongest around the three-year point. 127 Comparison of the pH values observed in this study and those of the literature should be viewed within the context of methodological differences (for example, the use of water as opposed to calcium chloride). Lloyd (in litt., April 18, 1986) reported values ranging from 3.6 to 4.4 and 4.1 to 5.1 for organic and mineral layers respectively under mature E S S F stands; these were determined in calcium chloride. The values of this study fell entirely within the latter ranges. For values determined in water, Kimmins (1974) observed a remarkably constant L F H value (4.9), with the underlying Ae ranging from 4.0 to 4.6 units. Since the latter should be slightly higher than values determined in calcium chloride (McLean, 1982), the values in this study are felt to be sufficiently in agreement with those of Kimmins (1974) for comparable fractions. In terms of pH changes after harvesting in B.C. , neither Haskin (1985) nor Martin (1985) noticed differences. However, Martin (1985) cautioned that his conclusions were based on very limited sampling. Haskin's (1985) sites all had post-harvest disturbances in the form of site preparation treatments. On eastern hardwood sites, Wallace and Freedman (1986) similarly noted no differences among pH values with stand age. By contrast, following clearcutting of balsam fir in Newfoundland, Page (1974) observed an increase of approximately 0.5 pH units in surface organic and mineral layers. There was a subsequent decline past approximately five to seven years after cutting. Therefore, it may be that increased pH levels following harvesting may be a broadly site-specific phenomenon—for example, possibly limited to site and disturbance conditions similar to those of the ESSF moist subzones in this study. 128 5.4.2 Nitrogen Concentrations Patterns of N concentrations in the forest floor and mineral soil materials are illustrated in Figure 5.2. There were highly significant differences among the means of the sites for both organic and mineral fractions (both: p < 0.001, df = 4, 130; forest floor: F = 13.5; mineral soil: F = 25.3). In the forest floor, the T3 site had a significantly (p < 0.05) greater N concentration than the other sites; none of the remaining means was significantly different from any other. The mean N concentration of the T3 site was 1.53%, while those of the other sites ranged from 1.11% to 1.28%. For the mineral soil, the T6 and T8 cutovers had significantly (p < 0.01) lower mean N concentrations than the other sites, but did not differ significantly from one another. The means of the remaining sites (TM, T3, T i l ) were not significantly different from each other. The mean N concentrations of the T6 and T8 cutovers were 0.11% and 0.09% respectively; those of the other sites were between 0.16% and 0.17%. From the above result, it would appear that there was an actual increase in forest floor N concentration within the first three years after harvesting. The increase may be due to fresh inputs from slash and fine roots—especially the finer materials, and also to a subsequent change in the quality of litter with the invasion of herbaceous pioneer vegetation. The increase was temporary, disappearing by the sixth year following disturbance; moreover, between the sixth and eleventh year there may have been a tendency (non-significant) towards the development of even lower concentrations than those of the mature stand (Figure 5.2). In the absence of significant forest floor weight losses, this would imply the 129 ."o 0 J 0.25-1 0.20-0.15-0.10-MINERAL SOIL (0-15 cm) 0.05- 1 | 1 1 1 1 1 1 1 1 1 1 1 1 0 1 2 3 4 5 6 7 8 9 10 11 12 Y E A R S A F T E R H A R V E S T Figure 5.2 Means and 95% confidence limits of the N concentrations of forest floor and mineral soil materials across the age sequence. 130 possibility of very minor temporary declines i n forest floor N levels in the latter part of the time sequence. This is supported to some extent by the data from the minera l soil fraction. The latter did not appear to have the benefit of mater ia l inputs noted for the surface organic layers at the three-year mark . The pattern of concentrations after year three is very similar to but far more pronounced than for the forest floor layers . W i t h no direct major influences from inputs and changes in vegetative cover, there appeared to be definite declines in the pool of N wi thin the surface minera l soil fraction. However, by the eleventh year N levels of both the organic and minera l soil fractions appeared to re turn to pre-harvest (mature stand) levels. Under mature stands, comparable sites in the Kamloops E S S F zone had N concentrations ranging from 1.07-1.96% and 0.06-0.29% for forest floor and surface minera l soil horizons respectively (Lloyd, in litt., A p r i l 18, 1986). The values observed in this study fall squarely wi th in these ranges. Compared to values from more northerly spruce-fir stands (Kimmins , 1974), for organic materials al l but the T3 value of this study are lower; however, minera l soil values in this study were two to three times greater than those of K i m m i n s (1974). Considering changes in N concentrations after harvesting, the increase/ decrease forest floor pattern observed here is very close to that observed by M a r t i n (1985) on the B . C . coast. Covington (1981) observed no significant change in forest floor N concentration wi th age for eastern hardwood sites. However , this m a y have been because of the comparison variable employed; he expressed N as a percentage of organic matter, the latter being determined as loss on ignition. B y contrast, Page (1974) reported decreases in N concentrations wi th in 131 approximately seven years of clearcutting of eastern spruce-fir stands; concentrations increased thereafter. In addition, following wave-form diebacks of coniferous forests in the Oregon Cascades, Ma t son and Boone (1984) noted marked differences in N concentrations of forest floor materials among sites. No differences seemed evident among those of the minera l soil fractions (Matson and Boone, 1984). 5.4.3 Phosphorus Concentrations The trends of forest floor and minera l soil mean P concentrations, and also those of the resin-collected P , are presented i n Figure 5.3. The forest floor mean P values exhibited highly significant differences among sites (p < 0.001, F = 11.16, df = 4, 130). S imi l a r highly significant differences were also observed among the mean P concentrations in the minera l soil fraction (p < 0.001, F = 5.16, df = 4, 130). In the forest floor case, the mean of the T 6 site was significantly (p < 0.05) lower than a l l others except the T 8 mean. The T 8 mean was significantly different (lower) from that of the T3 site only. The T 6 forest floor mean P concentration was 0.11%, and that of the T8 0.12%. The other values fell between 0.13% and 0.14%. In the mineral soil, the means of sites T 6 and T 8 were significantly (p < 0.05) higher than those of the T i l site only. In addition, the T 8 mean was significantly higher than the T3 mean. The minera l soil mean P concentrations of the T 6 and T 8 sites were both approximately 0.07%, while those of the rest were al l approximately 0.06%. F i n a l l y , there were also highly significant differences (p < 0.001, F = 6.0, df = 4, 138) among the mean resin-P concentrations of the sites (Figure 5.3). The 132 Vertical bars are 95% confidence limits 0.16-1 0.14 H ± 0.12 A O 0.10 H 0 .08 J 0.10-1 0 . 0 8 -O ! o 0.06-1 0 .04 J 16 - i "a> 1 4 -O < ° 1 2 -| * 1 0 -^ 8 -6 -Q_ # c '(/) <D 4_ | 2 J FOREST FLOOR (F+H) MINERAL SOIL ( 0 -15 cm) ANION RESIN (at F/H) I I I I I I I I I I I I I 0 1 2 3 4 5 6 7 8 9 10 11 12 Y E A R S A F T E R H A R V E S T Figure 5.3 Means and 95% confidence l imits of the forest floor and minera l soil P concentrations, including resin-P concentrations. 133 T i l mean was significantly (p < 0.01) lower than those of the T 6 and T 8 , but not significantly different from the T M and T3 means. The T 6 and T 8 means were not significantly different from each other or those of the T M and T 3 sites. A s s u m i n g that only the H 2 PO4 ionic species was involved, the mean concentration on the T i l cutover was 4.3 m g bag" 1 . Those of others ranged from 6.6 m g bag" 1 to 10.7 m g bag" 1 . Whi le significant differences among the site P variables have been evident, it should be noted that the absolute magnitudes and ranges of these differences were relat ively smal l . It is debatable whether such smal l changes in P concentrations could have had any practical significance for growth of advance regeneration. This point notwithstanding, the results can be interpreted in terms of an assart pattern. The forest floor pattern implies that there was a slight, temporary impoverishment of the pool of P in the organic layers after the three years. The lowest point of this occurred around the six-year point, following which the pool was restored to its pre-harvest levels. Al though not clearly demonstrated, the trend in the minera l soil P fraction was towards a marg ina l enrichment by year eight, then a return to pre-harvest (or even lower) levels. Taken together, the forest floor and minera l soil patterns imply a temporary and margina l downward movement of P on the sites, spread over two to three years . This movement seemed initiated some three years following the harvest. The peak of this movement occurred between the s ixth and eighth years. The trend of the resin-P data gives some support to this conclusion. According to Olsen and Sommers (1982), the resin method is useful in approximating root uptake mechanisms for P , as we l l as in measuring phosphate avai labi l i ty . While the 134 earlier portion of the sequence exhibited no significant differences, the trend % of resin-P implies a rise and fall in plant-available P coinciding extremely wel l wi th the forest floor and minera l soil patterns. Thus, it can be concluded that levels of plant-available P were increased slightly between three and eight years after logging. This increase occurred principal ly through mobilization of P from its forest floor pool. There were relat ively few available literature sources dealing wi th soil "total" P concentrations under E S S F forests. D a t a from L l o y d (in litt., A p r i l 18, 1986) included minera l soils only; moreover, methodological differences precluded direct comparisons. For ion exchange resin-P, although the resins have been frequently used in laboratory extractions, there have been almost no reported in situ applications. H a r t et al. (1986) was one such study; unfortunately, actual resin-P concentrations were not published. The forest floor P concentrations of this study are very close to but more variable than those of K i m m i n s (1974) for more northerly spruce-fir forests. However , the minera l soil P concentrations are two to three times those of the latter study. The forest floor concentrations are also similar to those of Wooldridge (1970) for several coniferous sites, but slightly higher than those of Freedman and M o r a s h (1985) under softwood cover in Eas te rn Canada. M i n e r a l soil concentrations were approximately two times those of Freedman and M o r a s h (1985). Concerning levels of P following disturbance, there was little published comparable data. L a n g et al. (1981) did not observe any major P concentration differences wi th age among their stands. However, they noted that their forest floor values were generally ve ry high. Page (1974) observed decreases in "available" P of 20-50%, followed by recovery, 135 in surface organic layers; minera l soil levels varied little. 5.4.4 Exchangeable Potassium, Calcium, and Magnesium Figure 5.4 il lustrates the patterns of exchangeable K , C a , and M g for the forest floor and minera l soil components. The non-parametric tests detected highly significant (p < 0.001) differences among sites for forest floor K and C a concentrations, and very significant (p < 0.01) differences for the forest floor M g values. In the minera l fraction, no significant (p > 0.05) differences were observed among the mean K concentrations. However, there were highly significant differences among the mean C a and M g concentrations (both: df = 4, 130, p < 0.001; F = 18.3 and 10.1 for C a and M g respectively). The jackknife 95% confidence intervals around the forest floor mean K concentrations indicated that the T 8 concentration was significantly lower than those of the other sites. The latter exhibited no significant differences among themselves. A s stated above, the minera l fraction showed no differences. The T 8 forest floor mean K concentration was 0.8 me (100g)" 1 ; the other forest floor means ranged from 1.1 me (100g)" 1 to 1.4 me (100g)" 1 . M i n e r a l soil mean K concentrations were a l l approximately 0.2 me (100g)" 1 . Fo r C a , according to the 95% intervals , a l l sites showed significantly higher forest floor mean concentrations than the mature stand. In addition, the T 6 mean appeared significantly higher than the T M , T 3 , and T8 values. The mean forest floor C a concentrations were 11.3 me (100g)" 1 and 20.0 me (100g)" 1 for the T M and T 6 sites respectively. The others fell between these extremes (Figure 5.4). The 136 Ver t ica l b a r s are 95% c o n f i d e n c e l imits FOREST FLOOR MINERAL SOIL 0 1 2 3 4 5 6 7 8 9 10 1112 0 1 2 3 4 5 6 7 8 9 10 11 12 Y E A R S A F T E R H A R V E S T Figure 5.4 Means and 95% confidence limits of the forest floor and mineral soil exchangeable K, Ca, and Mg concentrations. 137 minera l soil C a concentrations paralleled this pattern. The T 6 mean was significantly (p < 0.01) greater than those of the T M , T 3 , and T 8 values. However , in this case, the T i l mean was also significantly higher than those of the T M and T 3 sites. M e a n mineral soil C a concentrations were 1.4 (100g)" 1 and 3.6 me (100g)" 1 for the control and T 6 sites, w i th the other site values ly ing between these two (Figure 5.4). F ina l ly , the jackknife 95% limits suggested that the forest floor mean M g concentration of the T3 site was significantly higher than those of the T 8 and T i l sites. No other significant differences were apparent among the means. The T 3 , T8 , and T i l mean forest floor M g concentrations were 2.7, 2.1, and 2.2 me (100g)" 1 respectively. The mean minera l soil M g concentration of the T6 site was significantly (p < 0.05) greater than those of a l l the other sites. B y contrast, the T3 site had a significantly lower mean M g concentration than all but the T i l site. The means of the T3 and T 6 sites were approximately 0.3 me (100g)" 1 and 0.5 me (100g)~ 1 respectively, wi th the means of the other sites fal l ing between these values. A s wi th the phosphorus concentrations earlier, apart from the C a trends, the absolute magnitudes and ranges of differences encountered for the exchangeable bases wi th in the fractions are quite smal l . Added to this, the t ransi tory nature of these concentrations over even the short term detracts further from their interpretive value. The comments which follow should be viewed wi th in these l imitations. The data imply a slight and gradual decline in exchangeable K levels culminat ing around the eighth year and recovering by the eleventh. Un l i ke the P case earlier, this trend is suggested in both forest floor and minera l soil components—though the differences are non-significant in the 138 latter instance. The trend in C a concentrations might be considered the clearest of the three bases examined. Fol lowing harvest ing, both forest floor and mineral soil fractions generally exhibited increased levels of exchangeable C a ; this lasted beyond the eleventh year. The M g data were somewhat inconclusive. Forest floor exchangeable levels of this element appeared to decline slightly in the latter portion of the time sequence. The minera l soil fraction's M g declined wi thin the first three years, then appeared to increase up to a m a x i m u m at year six before re turning to pre-harvest levels. A clear explanation of this trend is not readily apparent. In any case, it is open to question whether the changes in al l except exchangeable C a concentrations in the forest floor had any marked impact on tree growth. The C a result is itself extremely interesting in light of the results of K l i n k a et al. (1980). In the latter study, exchangeable C a was the cation best correlated (albeit not very well) wi th forest productivity. The result here therefore suggests a post-harvest increase in site qual i ty (for forest growth) which lasted for the period covered by the sequence—and possibly beyond. Comparisons of these data wi th those of other studies have very limited usefulness. The values compare favourably wi th those of L l o y d (in litt., A p r i l 18, 1986) for comparable E S S F soils under mature forest cover. The K values of this study are sl ightly higher in general, but wi th the forest floor concentration ranges overlapping those of L l o y d . W i t h the exception of minera l soil M g , the concentrations of other exchangeable cation fractions fall in the middle of Lloyd 's ranges. The minera l soil M g concentrations overlap wi th those of L l o y d , but are generally lower (Lloyd, in litt., A p r i l 18, 1986). Compared to K i m m i n s (1974) exchangeable base concentrations, al l the forest floor values as well as the 139 mineral soil Mg concentrations of this study are much lower. Mineral soil K and Ca values are similar to those of the latter study. The mineral soil cation concentrations of this study fall well within the ranges observed by Van Ryswyk (1969) in subalpine and alpine soils further south in B.C. In terms of post-disturbance patterns, although there were analytical differences, these results agree very well with those of Covington (1981) for the Hubbard Brook sites. He found no major trends with stand age in K or Mg concentrations; however, there appeared to be a significant increase in Ca concentrations in the earlier portions of his age sequence (Covington, 1981). In the main Hubbard Brook study (Bormann and Likens, 1979), stream-water concentrations of K, Ca, and Mg were dramatically increased (as much as 11-fold) within the first three years. These losses commenced within weeks of the disturbance. In this study, revegetation was not suppressed as at Hubbard Brook; however, the possibility exists that the cation fractions with increased levels after harvesting may have experienced losses in stream-water. Where Ca is concerned, the results of this study are similar to those of Page (1974); however, subsequent declines were noted in the latter study. The trend of Mg values in the mineral soil are also similar to those of Page (1974). The pattern of K values and forest floor Mg concentrations do not follow those of Page (1974). 5.4.5 Carbon Concentrations and Carbon:Nitrogen Ratios Patterns of C concentrations and C/N ratios in the organic and mineral fractions are presented in Figure 5.5. Forest floor mean C concentrations exhibited no significant differences among themselves (p >0.05, F = 0.9, df = 140 Ver t ica l b a r s are 95% c o n f i d e n c e limits 45 -i 4 0 -(_> 3 5 -q "5 30 35 3 0 -2 5 -20 15 J FOREST FLOOR 6- i 5 -4 -MINERAL SOIL I—I—i—I—l—I—l—I—l—l—I—l I 0 1 2 3 4 5 6 7 8 9 10 1112 H 3 5 - i 30-25-1 20 15 r ~ l — I — r — r ~ l — I — i — I — i — i — I — I 0 1 2 3 4 5 6 7 8 9 10 1112 Y E A R S A F T E R H A R V E S T Figure 5.5 Means and 95% confidence limits of the forest floor and mineral soil C concentrations and C/N ratios. 141 4, 130). B y contrast, the non-parametric tests indicated highly significant (p < 0.001) differences among the groups for the minera l soil fraction values. The jackknife 95% confidence limits indicated that the minera l soil mean C concentration of the T 3 , T4 , and T5 cutovers were a l l significantly lower than those of the T M and T3 sites. The T M and T3 means were not significantly different; however, the T 6 and T8 means were significantly lower than that of the T i l site. Forest floor mean C concentrations ranged from 37.2% to 39.9%. For the minera l fraction, mean C concentrations were highest in the T M (4.8%) and lowest in the T 8 materials (2.3%). H i g h l y significant differences were detected among the C / N ratio site values for both forest floor and minera l soil fractions (both: p < 0.001; forest floor: F = 20.5, df = 4, 130). In the forest floor case, the T3 mean ratio was significantly (p < 0.05) lower than a l l others; in addition, the T 6 and T8 means were significantly higher than the T M but not significantly different from the T i l . N o significant differences were detected between the T M and T i l forest floor mean ratios. The jackknife 95% limits indicated that al l the cutover minera l soil mean C / N ratios were significantly lower than that under the mature stand. A m o n g the cutover sites, the T i l mean ratio was significantly lower than the other cutovers; the latter exhibited no significant differences among themselves. Forest floor mean C / N ratios ranged from 19.8 on the T 3 site to 29.6 on the T 8 site. Fo r the mineral fraction, the T M site had the highest mean ratio (31.1) and the T i l the lowest (24.7). The C concentration data indicate that there were no major changes in concentrations of forest floor organic matter after disturbance. However , in the minera l soil there was a noticeable decline; this reached its m a x i m u m at about 142 eight years after logging, then recovery was initiated. A t the l l ^ y e a r point, organic matter levels in the minera l soil had not yet returned to pre-harvest values. Given the s imi la r i ty in weights of F / H layers across the sequence (Chapter Four) , the lack of major change in organic matter levels in the forest floor materials is not surpr is ing. The pattern of decline and recovery in the minera l pool implies that assart impacts affect this more than the forest floor materials . Decreased organic matter concentrations in the mineral soil could have arisen from greatly increased microbial act ivi ty. This explanation would be more acceptable i f s imilar changes had occurred in the overlying forest floor materials; i f true, the N avai labi l i ty indices (Chapter Six) should reflect this increased act ivi ty . Perhaps more plausibly, the minera l soil declines could have come from a downward movement of organic matter outside of the sampling zone. This would imply that the organic matter outputs from the 0-15 cm fraction were not balanced by inputs from the forest floor pool above in the first hal f of the time sequence. Fur ther , the recovery would have required a slowing in the rate of organic matter output to the extent that inputs from above first balanced and then outstripped losses. The C / N ratio patterns present an interesting picture. For both the organic and minera l soil components, there was a definite decrease in ratio values wi th in three years after the harvest. Forest floor ratios returned relat ively quickly to pre-harvest levels and even sl ightly higher. B y contrast, minera l soil ratios continued a gradual decline over time. A s noted earlier, previously developed concepts of the control of C / N ratios over N availabil i ty in forest soils have been questioned in recent times. However , insofar as the C / N ratios 143 influenced decomposition and N mineral izat ion processes, the sharp decrease at the three-year mark suggests that N avai labi l i ty (and perhaps decomposition processes) might have been greatest wi th in the first three years . Th i s trend would have been quickly arrested in the forest floor materials , but would have remained in effect in the minera l fraction. The patterns of both fractions i n the first three years are entirely in consonance wi th what might have been expected wi th disturbance and additions of organic mater ia l . The rapid return of the forest floor to in i t ia l ratio values is probably a result of the post-harvest dynamics of minor vegetation and litter inputs. B y contrast, it is l ikely that the trend i n the minera l soil—buffered at least in the in i t ia l stages from the effects of vegetative changes—is an indication of what might occur without revegetation after logging. One unexpected result of this phase of the study was the low C concentrations and C / N ratio values encountered—especially under the mature stand. The question of how the above results compare wi th those of other studies must now be considered. It should be noted that methodological differences would affect the validity of comparisons. L l o y d (in litt., A p r i l 18, 1986) provided data from comparable E S S F mature stands in the Kamloops Region. Fores t floor C concentrations were between 28.7% and 46.7%, while its C / N ratios ranged from 24 to 41. For the minera l soil , C concentration and C / N ratio ranges were 1.6% to 7.1% and 23 to 27 respectively. A s s u m i n g that it was one of the dichromate oxidation procedures (Nelson and Sommers, 1982), the recovery rates in L l o y d (op. cit.) would have been somewhat lower than for the method used in m y study. In light of this, it can be seen that the C concentrations of this study are wi th in Lloyd ' s ranges, but lie towards their lower ends. The C / N 144 ratios observed are somewhat lower, but overlap with L loyd ' s ranges for comparable fractions. K i m m i n s (1974) reported ranges from spruce-fir forests near Prince George; the analyt ica l methods used in my study were v i r tua l ly the same as his. The forest floor C concentrations of this study are close to but s l ight ly higher than those of K i m m i n s (1974); however, the mineral soil values were approximately one order of magnitude higher than in the latter study. Concomitantly, C / N ratios of the forest floors were comparable between the two studies, but for the minera l soil those of this study are approximately two times those of the other. The ranges of minera l soil C concentrations and C / N ratios are very close to those observed by V a n R y s w y k (1969) in high-elevation forest soils in B . C . F i n a l l y , both organic and minera l fraction C / N ratios were wi th in the same range as those observed in E S S F forests of the coast/interior t ransi t ion (Kl inka et al, 1982). It can therefore be safely concluded that the values observed in this study are entirely acceptable. One extremely interesting implication here is that the generally low C / N ratio values of the E S S F forests in this and other studies m a y be indicative of a much greater level of microbial act ivi ty and N avai labi l i ty than might have been expected from accepted concepts (see Chapter One). In terms of trends in C concentrations and C / N ratios over time after disturbance, reported results have been predictably variable, depending {inter alia) on the nature and extent of the disturbance and general site and cover conditions. The lack of differences in the forest floor C concentrations of this study are in general agreement w i th data presented by Piene (1978), L a n g et al. (1981), and Mat son and Boone (1984). The existence of differences in the 145 mineral soil concentrations wi th age are s imilar to the result of M a t s o n and Boone (1984); however, the observed direction of change is the opposite of that reported in the latter study. L a n g et al. (1981) did not observe any trends wi th age in mineral fractions. Trends in C / N ratios were not reported directly by the above mentioned studies. However , inspection of the data of M a t s o n and Boone (1984) implied that there m a y have been significantly higher C / N ratios i n the forest floors of stands wi th regrowth compared to those of old-growth stands. The mineral soil data did not support any differences in the ratios w i th age. The increased forest floor C / N ratios of this study at eight years after disturbance may be analogous to Ma t son and Boone's (1984) increases; however, the differing circumstances of disturbance and ages involved preclude concrete statements in this regard. The ratio patterns after harvest ing are very close to those of Page (1974) for organic mater ia l ; those of the mineral fraction are not in agreement wi th those of the latter study. 5.4.6 Unit-Area Elemental Weights The estimated weights per unit area of the various elements in the organic and mineral fractions are presented in Tables 5.1a and b respectively. Fo r completeness, the data on bulk densities and weights of the minera l fraction have been included. The range of bulk density values lies squarely wi th in that of K l i n k a et al. (1982) for comparable depths in soils under mature E S S F cover. They are also wi th in the range for the same depths in podzols at high elevations in the dry interior (Hask in , 1985). The values are much lower than those observed by K i m m i n s (1974) for comparable horizons in spruce-fir forests to 146 Table 5.1a Means and 95% confidence l imits of the elemental unit-area weights for the forest floor fraction. F R A C T I O N / S I T E / V A L U E S V A R I A B L E T M T3 T 6 T 8 T i l (values rounded; figures in brackets are 95% limits) FOREST FLOOR Total N t 1045 (1197,893) 1116 (1285,947) 927 (1018,836) 796 (882,710) 895 (993,797) Total P 109 (124,94) 101 (116,86) 87 (95,79) 84 (92,76) 92 (102,82) E x c h . K 42 (48,36) 38 (44,32) 35 (38,32) 23 (26,20) 33 (38,28) E x c h . C a 189 (222,156) 221 (248,194) 315 (353,277) 217 (238,196) 243 (275,211) E x c h . M g 25 (29,21) 24 (29,19) 22 (24,20) 19 (21,17) 18 (20,16) Tota l C , M g / h a 31.4 (35.7,27.0) 27.1 (31.4,22.9) 30.5 (33.1,27.8) 27.1 (29.8,24.5) 27.8 (30.5,25.1) tUnless otherwise stated, units are kg ha" 147 Table 5.1b Means and 95% confidence limits of the elemental unit-area weights for the minera l soil fraction. F R A C T I O N / S I T E / V A L U E S V A R I A B L E T M T3 T 6 T 8 T i l (values rounded; figures i n brackets are 95% limits) UNERAL SOIL Bulk density, M g / m 3 0.77 (0.99,0.54) 0.71 (0.72,0.70) 0.84 (1.16,0.53) 0.98 (1.34,0.64) 0.68 (1.05,0.32) Weight, M g / h a 1152 (1491,813) 1070 (1084,1056) 1266 (1732,800) 1478 (2003,953) 1022 (1570,474) Total N t 1843 (2412,1274) 1819 (2035,1603) 1393 (1921,865) 1330 (1851,809) 1737 (2692,782) Total P 703 (917,489) 621 (675,567) 823 (1135,511) 1049 (1432,666) 562 (868,256) Exch . K 89 (117,61) 85 (94,76) 97 (134,60) 97 (133,61) 73 (113,33) E x c h C a 323 (454,192) 287 (374,200) 919 (1326,512) 590 (857,323) 602 (970,234) E x c h . M g 50 (66,34) 35 (39,31) 74 (103,45) 63 (87,39) 40 (62,18) Total C, M g / h a 55.7 - (72.5,38.9) 46.6 (50.4,42.9) 35.8 (49.7,30.0) 34.3 (46.7,21.9) 37.3 (57.4,17.3) tUnless otherwise stated, units are kg ha 148 the north; the latter range was 1.22 to 1.83 M g m " 3 . However , they are much higher than those of V a n R y s w y k (1969) for Alp ine B r o w n forest soils (0.3 to 0.5 M g m " 3 ) . The minera l soil mean bulk density estimates m a y thus be considered acceptable. Nevertheless, the wide confidence intervals associated with them severely constrained attempts to compare derived estimates of minera l soil elemental contents. Here , the only definite semblance of change was wi th C a values. Calculated forest floor elemental contents are l ike ly to be overestimates; they are based on the mean weights of the entire F / H fractions, while concentrations are from the fine fractions (< 2 mm) only. It has been shown that in organic materials the finer fractions can have much higher N levels than others (Wil l iams, 1983a). Moreover, as noted in Chapter Four , no attempt was made to quantify the actual coverage of the sites by forest floor materials (that is, as opposed to rock outcrops, exposed mineral soil, etc.). A s stated earlier, interpretative priority was accorded the elemental concentration data considered in the foregoing. Nevertheless, i t was considered at least instructive to examine where and to what extent possible differences in absolute elemental contents might have occurred. The derived 95% confidence intervals were taken as an indicative basis in this regard. A s might be expected, the trends in mean unit-area elemental contents paralleled those of the mean concentrations. Nevertheless, some minor differences were apparent. 149 5.4.6.1 Nitrogen Content M e a n N contents for forest floor and minera l fractions ranged between 796 and 1116 k g h a " 1 and 1330 to 1843 kg h a " 1 respectively (Tables 5.1a and b). In the forest floor, unlike the trend in concentrations, there was no suggested increase at the three-year mark; moreover, the T 8 value suggested an absolute decline in N content at eight years of almost 250 kg N ha" 1 less than mature stand values. In light of the width of the derived 95% limits , it is tempting to accept this as indicating an actual decline in forest floor N contents over t ime, as in other studies. However , given the lack of change in this direction in both forest floor weights and concentrations, it is believed best to reject this anomalous result. Based on the review by K i m m i n s et al. (1985) the forest floor N contents are at the higher end of the range of values reported for Abies forests. They are also greater than the estimate of 842 k g N ha" 1 for spruce-fir forest floor layers near Prince George (Kimmins , 1974). This supports the belief that the values are indeed overestimates. However , higher values have been reported for forest floors of Abies forests; either equivalent or higher values have been observed in spruce and spruce-fir types i n eastern Canada , the U . S. A . , and Europe (Cole and Rapp, 1981; K i m m i n s et al., 1985). The minera l soil values are wi th in the ranges of values for s imi lar depths under spruce and/or fir types (Kimmins et al., 1985). 150 5.4.6.2 Phosphorus Content The ranges of mean P contents were from 84 to 109 kg ha" 1 for forest floor layers and from 562 to 1049 kg ha" 1 for the minera l soil fraction (Tables 5.1a and b). The forest floor trends are s imi lar to the concentration data in that a significant decline and recovery in the organic layer P are suggested. However , while the concentration data implied a m a x i m u m decline at about year six, the content data suggests that year eight is the point of m i n i m u m P levels. Both measures point to some decline; the suggested magnitude of the latter is approximately 25 kg ha" 1 from mature stand levels. Concerning comparisons wi th other reported P values, what was said earlier for N contents also applies here. The values are somewhat higher than the general reported ranges; nevertheless, s imi la r or higher values have been encountered by other workers. For reasons given earlier, the forest floor P content values are believed to be overestimates. However , this should not seriously detract from the estimates of decline given earlier. F r o m the review of K i m m i n s et al. (1985), the mineral soil P values compare favourably wi th s imi lar data from spruce and/or spruce-fir forests in various locations. 5.4.6.3 Exchangeable Potassium, Calcium, and Magnesium Contents In the forest floor, mean exchangeable K , C a , and M g content ranges were 23 to 42 kg ha" 1 , 189 to 315 kg ha" 1 , and 18 to 25 kg ha" 1 respectively. In the same order, mineral soil mean values ranged between 73 and 151 97 kg ha" 1 , 287 and 919 kg ha" 1 , and 35 and 74 kg ha" 1 respectively. The forest floor trends a l l supported those of the concentration data earlier, though wi th lessened sensitivities owing to the relat ively wide derived 95% limits . Declines in forest floor exchangeable levels of approximately 19 kg ha" 1 for K and seven kg ha" 1 for M g were suggested; the M g decline was rejected for reasons s imi lar to those given for forest floor N contents earlier. Fo r exchangeable C a , the content data suggested an increase in forest floor levels by year six of 126 k g ha" 1 . M i n e r a l soil data also suggested an increase, but of 596 k g ha" 1 in that fraction—the only case in which significant differences in minera l fraction contents were suggested. A s stated earlier, comparisons of exchangeable cation data with results of other studies are of extremely l imited value. A s wi th N and P , it is believed that the values obtained above for the organic fraction were overestimates; s imi la r ly , minera l soil contents are probably acceptable wi th in the very wide range possible from point-in-time samples. It is interesting that when compared to the results of K i m m i n s (1974), the exchangeable K and M g contents of both forest floor and minera l soil fractions in this study are much lower. The values from K i m m i n s (1974) for exchangeable C a contents of both fractions lie easily wi th in the range observed in this study. These observations support somewhat the acceptability of the results. 152 5.4.6.4 Carbon Contents A s given in Tables 5.1a and b, mean C contents ranged between 27.1 M g ha" 1 i n the forest floor layers, and between 34.3 M g ha" 1 and 55.7 M g ha" 1 in the minera l soil fraction. A s wi th the related concentration data, there appeared to be no indication of any differences over the age sequence in forest floor C content. Unl ike its concentration counterpart, the pattern of minera l fraction C contents showed no evidence of marked changes among the sites. However , since the overall pattern in this case was s imi lar (though without statist ical differences) to that of the concentration data, the apparent lack of marked content differences may have been due main ly to the wide confidence l imi ts . F o r reasons discussed earlier, the estimates of C contents of the forest floor horizons are l ikely to be overestimates. A survey of the literature yielded few studies wi th directly comparable data. O n the assumption that organic matter is 58% C , comparisons can be made wi th values from various spruce and/or spruce-fir forests. The range of F / H C contents of this study is higher than the 20 M g ha" 1 calculated from data of K i m m i n s (1974) for the total forest floor of more northerly mature spruce-fir stands. However , the range is wi th in that of F / H horizons under coniferous stands i n N o v a Scotia (Freedman and M o r a s h , 1985), and very much lower than those for entire forest floors of spruce cover i n A l a s k a and Europe (Cole and Rapp, 1981). The minera l soil data are more difficult to compare wi th those of other studies since sampling depths differed markedly among studies. Their range overlaps wi th that of the C contents 153 observed in the 0-15 cm fraction under different-aged Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco var . menziesii] stands in Oregon (Kimmins et al, 1985). 5.5 CONCLUSIONS F r o m the foregoing results, it was concluded that both subsidiary hypotheses H 0 and H 0 should be rejected. There was clear evidence that major changes had indeed occurred in the general chemical characteristics of both the forest floor and minera l soil fractions during the post-harvest period. In both, increases in p H levels and exchangeable C a concentrations appeared to last beyond the 11-year span of the age sequence. Temporary increases in forest floor total N and available P were noted; temporary decreases occurred in the total P , exchangeable K , and C / N ratios of this fraction. In the mineral fraction, C / N ratios declined consistently to beyond the 11-year mark. For . the remainder, temporary declines were noted i n total N , exchangeable K and M g , and also in C concentrations. Total P concentrations in the minera l soil appeared temporari ly increased. The absolute values and magnitudes of changes in total P and exchangeable K and M g concentrations m a y not be of major practical significance to plant growth. Moreover, for the exchangeable bases, the ephemeral nature of values over a growing season places l imits on the importance which can be attached to the same. Taken as a whole, the P data suggest that there was a downward movement of P through the profile; this started approximately three years after logging, and was spread over some three or more years. Plant-available P may have been derived pr incipal ly from the forest floor fraction. The C data also suggest the occurrence of a slight downward movement of organic 154 matter. Unl ike P, this appears to have occurred more within the minera l profile, and was not ini t ial ly balanced by inputs from the overlying organic horizons. F i n a l l y , the relatively low C / N values observed imply that in general there m a y have been a higher level of microbial act ivi ty and N availabi l i ty than might have been expected. Insofar as the results m a y be interpreted in terms of an assart pattern, it can be seen from the foregoing discussion that trends varied according to the elements and fractions under consideration. However , it m a y be concluded that in general the first eight years after harvest ing emerged as the pr incipal period of importance in this context. W i t h i n this, the period from three to eight years—and possibly main ly around the six-year mark—appeared to be that in which m a x i m u m change occurred. The pr incipal manifestations of chemical change seemed to have vir tual ly disappeared by the eleventh year after disturbance. The unit-area elemental contents were believed to be overestimates; nevertheless, they can be used as indications of magnitudes of changes. The wide confidence l imits associated with mineral soil elemental contents prevented the estimation of changes for al l except the exchangeable C a values of that fraction. The latter was increased by nearly 600 kg ha" 1 , or approximately 185% over the mature stand value, by year six of the sequence. No changes occurred in forest floor total N and C contents. Forest floor P decreased by 25 kg ha" 1 , approximately 23% of the T M value, by year eight of the sequence. Exchangeable K contents declined by 19 kg ha" 1 , approximately 45% of the T M value, by the same point in the age sequence. Exchangeable" M g contents declined by 7 k g ha" 1 (31% of T M value) by year 11. B y contrast, forest floor exchangeable C a was 155 increased by 126 kg ha" 1 , approximately 67% of the mature stand value, by year six of the sequence. CHAPTER 6 POST-HARVEST NITROGEN AVAILABILITY PATTERNS In Chapter F i v e , consideration was given to post-harvest patterns of total N in the forest floor and minera l soil fractions. However, it is generally accepted that total measures are often poor indicators of the avai labi l i ty of soil N to plants. Thus, a more explicit examinat ion of post-harvest N avai labi l i ty trends is now presented in this chapter. 6.1 INTRODUCTION There have been numerous studies and publications dealing wi th N cycl ing and availabil i ty in forested ecosystems. Recent reviews concerning various facets of the problem of N avai labi l i ty to forest trees include those of Keeney (1980), H e a l et al. (1982), Robertson (1982), T a m m (1982), Gosz (1984), and Powers (1984a). A s noted earlier, the key to the avai labi l i ty of inorganic forms of N lies in the net mineral izat ion level of that element. For such reasons, there has been long-standing interest in both agriculture- and forestry-oriented circles i n the development of a reliable index of plant-available N . Bremner (1965) summarized the many approaches which have been used, classifying them as either chemical or biological methods. A s imi lar but much more recent review of approaches was carried out by Keeney (1982). The general consensus was that the most reliable procedure for the indication of plant-available N in an unknown soil is by the determination of mineralizable N (MN) in a laboratory incubation (Bremner, 1965; Hesse, 1971; Keeney, 1980 and 1982). 156 157 The basic principles of mineral izat ion studies by incubation have been wel l documented elsewhere. The general methods and problems—especially where forest soils are involved—were examined elsewhere by this writer, t The issues involved are much too lengthy and complex to be repeated here in detail. Nevertheless, some aspects w i l l be highlighted very briefly. Fo r many applications, the basic anaerobic procedure pioneered by W a r i n g and Bremner (1964) has become the most highly recommended biological index of N availabil i ty (Hesse, 1971; Keeney, 1980 and 1982). Specifically, some form of the one-week anaerobic incubation at 4 0 ° C has emerged as the best available index (Keeney, 1982). W i t h this, the results can be calculated in two ways . The first is the original method of War ing and Bremner (1964), in which M N is calculated as the difference between the pre- and post-incubation concentrations of N H J - N ; this w i l l be referred to as "net M N " . The second way is that proposed by Powers (1980 and 1984b), in which only the post-incubation N H J - N concentration is considered—the "gross M N " measure. Regardless of which incubation approach or method is employed, there are several difficulties which have served to make the index an imperfect one. Such problems include the possible effects of sample collection, handling, and storage; of note here is the wel l -known effect of d ry ing and re-wetting (Birch, 1959 and 1960; Bremner , 1965; Keeney and Bremner , 1966; Hesse, 1971; Salonius, 1978). W h e n forest soils are under consideration, additional factors often render problems more acute or br ing further complications. These include stocking levels, t This was done in an unpublished report prepared for Forestry 512 (Dr. T . M . Ballard—Professor), Facu l ty of Fores t ry , U . B . C . Copies are available from the author. 158 management practices, and internal cycling patterns (Lamb, 1975; Keeney, 1978 and 1980). One of the most important factors is the presence of a forest floor—the principal characteristic differentiating forested soils from agricul tural ones (Pritchett, 1979). Opinions have been divided as to whether forest floor layers should be excluded (in part or as a whole) from M N studies—even in the mind of a single investigator at different points in time (Keeney, 1978 and 1980; Powers, 1984a, b). There is an overwhelming amount of evidence indicating that for many reasons, consideration of forest floor layers is crucial to M N and tree nutrition studies, especially i n northern forest ecosystems (Viro, 1963; T a m m and Pettersson, 1969; Bhure , 1970; L a m b , 1975; K i m m i n s and Hawkes , 1978; Youngberg, 1978; T a m m , 1979; V a n Cleve et al., 1981; Weber and V a n Cleve, 1981; Mahendrappa and Salonius, 1982). Powers (1984b) presented a succinct summary of the major sources of variat ion encountered in N mineral izat ion studies. He identified three ma in types of variat ion—spatial , seasonal, and analytical—and made several proposals aimed at standardization and avoidance of their effects. A m o n g these was the use of a standard reference depth, composite samples taken in a t r iangular pattern, and the gross M N measure. This last was reported to be extremely useful in overcoming intractable problems deriving from sample pre-treatments and storage (Powers, 1984b). Va r i ab i l i t y problems and effects were also discussed in s imi lar fashion by M c N a b b (1984). A s for incubation methods in general, the anaerobic procedure has been regarded with skepticism by some investigators. One of the m a i n cri t icisms seems to revolve around a difficulty on the part of some researchers to accept it as adequately simulating or reflecting actual field conditions [Hase and T immer , 1982; Regional Forest Nutr i t ion Research Project ( R . F . N . R . P . ) , 1984]. It has been suggested that 159 at least in coniferous forest soils the method measures microbial N pr imar i ly (Myrold , 1985). The above notwithstanding, the anaerobic procedure has been strongly recommended for standard use in forest soil applications (Keeney, 1980; Powers, 1980 and 1984b; Smi th , 1984; R . F . N . R . P . , 1984). Th i s study employed such a procedure. A n attempt was made to refine it on the basis of temperature differences among the sites (see Chapter Three). G i v e n the flaws of the incubation approach, continued investigation of this as wel l as other indices—mainly chemical extractions—has been urged. Examples of promising laboratory methods include a calcium chloride-autoclave method (Keeney, 1980 and 1982; Mahendrappa, 1980), hot-water-extractable N (Keeney, 1980 and 1982; R . F . N . R . P . , 1984), the use of hot acid chromate (Keeney, 1980), acid or alkaline KMnO<, (Keeney, 1982; Carsk i and Sparks , 1985), and alkaline steam disti l lat ion (Keeney, 1982; Gianello and Bremner , 1985). Powers (1980), while preferr ing the anaerobic incubation as an overall best choice, observed that in western forest soils KCl-extractable N was actually the best choice from a statist ical standpoint; however, its - results were considered too variable in general for routine applications (Powers, 1980). A KCl-ext rac t ion method was employed in this study. Several researchers have indicated preferences for a re turn to some form of in situ approach to mineralization studies —either separately or in conjunction wi th laboratory incubations. There have been noteworthy innovations in the use of buried soil-filled containers (Powers et al, 1978; Bonneau, 1980; Keeney, 1980). One recent and exciting alternative to the direct involvement of soil 160 materials has been the in situ use of ion exchange resins. A s noted earlier (see Chapter Five) , such resins have long been employed in laboratory studies of P availabil i ty to plants. However , the in situ use of resins for the examination of N (and P) avai labi l i ty has been largely pioneered recently by D . Bink ley and co-workers (Binkley and Packee, 1982; Bink ley and Matson , 1983; Bink ley , 1984; Cromack and Bink ley , 1984; H a r t et al., 1986). Intui t ively, such an approach has the advantage of circumventing many of the difficulties inherent in the use of soil sample materials (e.g. pre-treatment effects, chemical and microbiological changes). It has been noted that the resins have the advantage of indicating the integrated effects of soil conditions ( R . F . N . R . P . , 1984). B ink l ey (1984) has shown that their ion collection is very sensitive to ion mobil i ty and transport; this makes them useful for assessing N avai labi l i ty to tree roots. Moreover, differences in microbial competition for N , mineral izat ion rates, and plant uptake are also reflected by the method (Binkley, 1984). There remains the disadvantage that such an index can only be used for comparat ive purposes currently (Binkley, 1984). However , this is a relatively minor setback given the potential advantages. On the analyt ica l side, the degree of ion recovery is method-dependent. Nevertheless, this is of much less importance in comparative studies among sites under a common method (Hart and Bink ley , 1984). B i n k l e y (1984) noted that increases in water flow to resin bags m a y increase N H J - N capture more than that of N O i - N . Moreover , he felt that the greater mobil i ty of N O ^ - N would emphasize the role of nitrification in determining resin bag estimates of N availabi l i ty (Binkley, 1984). This last point is of great importance to this study; anion resins were used to reflect relative differences in N O 3 - N levels cumulative over a period, thereby indicating points of m a x i m u m (or otherwise) N availabil i ty 161 across the age sequence. In B . C . , Binkley and Packee (1982) and M a r t i n (1985) have studied N avai labi l i ty patterns after harvest ing using the resin approach. Elsewhere , Vitousek and Matson (unpubl.) incorporated the method in a recently-initiated (in 1982) long-term study of the effects of disturbance on N dynamics in the southeastern Uni t ed States. 6.2 O B J E C T I V E S A N D H Y P O T H E S E S The m a i n objective of this phase of the study was to examine whether any changes in N avai labi l i ty had occurred on the sites since harvesting, and to attempt to relate the findings to the overall investigation of an assart effect. This phase concerns m a i n hypothesis H 0 directly; m a i n hypothesis H 0 is also X o indirectly and to a lesser extent involved. The subsidiary hypotheses under test and their alternates may be stated as follows: H 0 y N o major changes occurred in N avai labi l i ty from the forest floor materials dur ing the post-harvest period. H , : Major changes indeed occurred in N avai labi l i ty from the forest floor materials dur ing the post-harvest period. H 0 No major changes occurred in N avai labi l i ty from the mineral soil fraction dur ing the post-harvest period. 162 H , : Major changes indeed occurred in N avai labi l i ty from the mineral soil fraction dur ing the post-harvest period. If H 0 ^ and H 0 ^ are true, there should be no marked differences among the N mineral izat ion and inorganic N levels observed in the relevant fractions from the sites. Stat is t ical tests for such differences among the sites were accepted as the falsification cr i ter ia of the nul l hypotheses. 6.3 METHODS The basic field layout, sampling scheme, and system of creating composite samples for analysis were presented i n Chapter Two. For both forest floor and minera l soil fractions, all analyses and tests in this phase involved the second-stage composites only. Three categories of N avai labi l i ty indices were employed—incubation studies, chemical extractions, and ion exchange resins. Wi th the exception of the resins, data were obtained for both organic and mineral fractions in a var ie ty of ways. These are discussed below. 6.3.1 Incubation Studies Incubations were carried out on the forest floor and minera l soil samples of each site. In a l l cases, the seven-day, 40 °C anaerobic technique was employed. F o r the organic samples, three grams of sieved, air-dry mater ial were placed wi th 24 m l distilled water i n a 20 X 125 m m borosilicate glass culture 163 tube. The sample weights and water volumes used were chosen (by pre l iminary t r ia l and error) such that the water v i r tua l ly filled the tubes after sample absorption, but yet allowed a constant water volume to be used for a l l samples. E a c h tube was t ightly sealed by a combination of Teflon thread tape (approximately 13 m m wide) and a phenolic screw cap. M i n e r a l soil samples were treated s imi lar ly , except that five grams of air-dry soil were used in the same volume of water. Complete saturation of sample materials was ensured by use of a vortex mixer. Samples were then placed i n batches in a Precis ion Scientific Model 4 E G gravi ty convection incubator; two blanks (distilled water only) were included wi th each batch. Incubation was initiated immediately after the preparation of a batch of samples. Temperature was monitored using a mercury-in-glass thermometer; this was checked twice dai ly during the incubation period. After incubation, the contents of each tube were gravity-filtered into storage bottles through Wha tman No. 41 filter paper. The tube and sample mater ia l were rinsed wi th 24 m l of 2 M KC1 in 12 m l aliquots. The leachate was also filtered into the same containers, which were then placed in cold storage (1 .7°C) unt i l analysis. Concentrations of N H J - N in the sample extracts were determined by the indophenol blue procedure on a Technicon Au toAna lyze r (Technicon Instruments Corporation, 1973). In i t ia l concentrations of N H J - N and N O ^ - N were taken to be the values obtained by 2 M KC1 extractions (see next Section). Fo r both organic and mineral fractions, these data allowed the use of two mineral izat ion indices—the traditional net M N and the newer gross M N . In addition, in light of the temperature differences among the sites and their 164 implications (see Chapter Three), these indices were adjusted according to the mean forest floor temperatures observed. This was done on a per-sample basis assuming Q , 0 = 2 for N mineral izat ion (Stanford et al, 1973; Powers, 1980). Adjusted M N levels were obtained by incorporating the latter in the general discount formula (Chapman and M e y e r , 1947) to yield the following calculation formula: A M N T = M N /{2**[(40-T)/10]} where: A M N T = M N level adjusted to temperature " T " ; T = relevant mean forest floor temperature (Chapter Three); M N 4 Q = unadjusted M N level at T = 40. M i n e r a l soil M N values were adjusted using the simple assumption that that fraction's mean temperatures would be 2 ° C lower than the corresponding forest floor means ( M . Novak , Professor, U . B . C ; pers. comm. M a r c h 18, 1986). Thus, for each soil fraction there were four indices of potential N mineralization [sensu Powers (1984a) and Vitousek and Mat son (1985)] —unadjusted and adjusted net and gross M N levels i n each case. There appear to have been very few attempts to adjust anaerobic incubation results by temperature values; that by Powers (1984a) is noteworthy. Burger and Pritchett (1984) applied both temperature and moisture adjustments to the aerobic N 0 procedure. For the four indices, the data in each instance were analyzed as a single-165 classification Model I A N O V A , wi th the sites as the treatments. A posteriori comparisons were performed using Tukey ' s test. These analyses were carr ied out us ing the G E N L I N package (Greig and Bjerr ing , 1980). Conformity of a l l data to the assumptions of a valid A N O V A was checked using both the M I D A S (Fox and Gui re , 1976; Anonymous, 1976) and G E N L I N programmes. Transformations were necessary in al l cases to achieve homogeneity of variances. Power transformations sufficed for a l l net M N measures in both fractions; the same was true of natural-log transformations for a l l gross M N measures. Some estimate—however crude—of the result ing potentially mineralizable N on a unit-area basis was believed desirable. The anaerobic technique does not readily lend itself to direct derivations of such quantities. F r o m Powers (1984a), a regression equation between field and laboratory incubation values showed that under a var iety of forest conditions field M N values over a 12-month period could be much higher than the concentrations observed from laboratory anaerobic incubations, depending on the soil temperature level chosen. Therefore, in this study, it was assumed that the concentrations observed during the incubations would be at least equal to those which might be observable in the field for a nominal base of one year. However , these concentrations were expressed as a percentage of the total N measurement (see Chapter Five) of each sample; a new data set was thereby generated for each fraction and N mineral izat ion index. F o r minera l soil samples, these data were analyzed statist ically in s imi lar fashion to the original data above. Arcs ine transformations were necessary in these cases to achieve homogeneity of the variances. No transformations were found which could have allowed the forest floor data to conform to the 166 assumptions for A N O V A . Therefore, these were a l l analyzed by the non-parametric K r u s k a l - W a l l i s and median tests (Gibbons, 1971; Hollander and Wolfe, 1973; Sokal and Rohlf, 1981) w i th in the M I D A S programme. A s in- earlier cases of this nature (Chapter Five) , the construction of 95% confidence l imits us ing Tukey ' s jackknife was used as a basis for a posteriori comparisons. Per-hectare estimates of M N and their associated 95% confidence limits were derived using the per-cent-of-total concentration data, the total N data, and the approach presented earlier (Chapter Five) involv ing R . M . S . error calculations. 6.3.2 Chemical Extractions In each sample, concentrations of N H J - N and N O 3 - N were determined by equi l ibr ium extraction wi th 2 M KC1 (Keeney and Nelson, 1982). F o r both forest floor and mineral mater ials , five grams of sieved, air-dry material were shaken mechanically in 50 m l 2 M KC1 for approximately (but not less than) one hour. The samples were then left to settle for an additional hour. The extracts were gravity-fil tered into storage bottles through W h a t m a n No. 41 filter paper, then place in cold storage (1 .7°C) unt i l analysis . Concentrations of N H J - N in the extracts were determined s imi la r ly to those i n the incubations. Concentrations of N O 3 - N were also determined colourimetrical ly on a Technicon A u t o A n a l y z e r (Technicon Instruments Corporat ion, 1971b). None of the data conformed to the assumptions for a va l id A N O V A , and no transformations were found which could achieve this. Thus, a l l such "absolute" KC1 extraction data were analyzed using non-parametric methods and Tukey jackknife 95% intervals as described earlier. Separate analyses were done for N H J - N , N O ^ - N , and their sum (IN). 167 Weights of KCl-extractable N H J - N and NO3 - N per unit-area were estimated using the same approach as described earlier for M N unit-area values. Observed sample concentrations were expressed as a percentage of their respective total N levels, thereby generating new data sets for analysis . W i t h the exception of forest floor NO5 - N , a l l analyses were performed using the non-parametric tests and Tukey jackknife 95% confidence l imits as earlier outlined; these data did not fulfill the requirements of a valid A N O V A . The organic fraction's NO3" - N data were analyzed as a single-classification Model I A N O V A , wi th the sites as treatments. A posteriori comparisons were performed using Tukey ' s test. Statist ical data checking and analysis were carr ied out as described earlier using the G E N L I N and M I D A S programmes. A power transformation was necessary to achieve homogeneity of the variances. Per-hectare estimates were derived using the R . M . S . error calculations as presented earlier. 6.3.3 Ion Exchange Resins Ion exchange resins were used to monitor relative changes in N H J - N and N O I - N levels over an extended period. Cations and anions were collected by separate but paired resin lots. In the fall of 1983, 30 lots of cation and anion resins were placed at the F / H interface, 30-50 cm from the bases of selected sample trees on each site. E a c h lot consisted of separate nylon mesh bags containing 17 g and 11 g oven-dry equivalent respectively of F isher R e x y n 101 (H) strong acid organic cation and R e x y n 201 (OH) strong base organic anion exchange beads. The mesh size was between 16 and 50. Both res in types employed a polystyrene divinylbenzene matr ix ; the cation resins were sulphonated 168 and in the hydrogen form, while the anion exchangers were aminated wi th a l k y l quaternary amines and in the hydroxyl form. A l l bags were treated wi th a mercuric chloride solution to prevent microbial interference wi th collected ions; approximately two to five per cent of the exchange sites were so saturated. A l l "plots" of the T M and T 3 sites had resins; selections for resin placement on other sites was by random selection among sample "plots". The resin lots were retrieved after approximately 12 months in situ. Resins were extracted in their bags wi th KC1 solutions. For the cation resins, I M KC1 was used after H a r t and Binkley (1984), but the volume of extractant was 200 m l for each bag. In the anion case 100 m l of 2 M KC1 solution was used. The bags were shaken in the solution for approximately 30 minutes, then left to equilibrate for approximately 24 hours. The solutions were then gravi ty-filtered through W h a t m a n N o . 41 paper. Extracts were kept i n cold storage (1 .7°C) unt i l analysis . Al iquots were used to determine N H J - N and N O 3 - N concentrations colourimetrical ly on a Technicon Au toAna lyze r (Technicon Instruments Corporat ion, 1971b and 1973). In both the cation and anion cases, the data were analyzed as a single-classification Model I A N O V A , wi th the sites as treatments. A posteriori comparisons were performed using Tukey ' s method. The G E N L I N and M I D A S statist ical programmes referred to earlier were used for analyses—including checks for conformity to the A N O V A assumptions. N o transformation was necessary for the N H J - N data; however, the NO5 - N data required a natural- logar i thm transformation to achieve homogeneity of the variances. 169 6.4 RESULTS AND DISCUSSION 6.4.1 Incubation Studies The trends of forest floor and mineral soil N mineral iza t ion concentration levels according to the various indices are presented in F igure 6.1. 6.4.1.1 Forest Floor Materials There were highly significant differences (p < 0 .001, F = 9.7, df = 4, 130) among the means of the unadjusted data from the net M N procedure ( N U M N ) . The means of the T 6 and T 8 sites were significantly (p < 0.05) lower than those of the T M , T 3 , and T i l sites; wi th in each of the two different groups thus formed, there were no differences among the means of the member sites. The T6 and T 8 means were 482 and 503 m g N (kg soil)" 1 respectively; those of the other sites ranged between 636 and 746 m g N (kg soil)" 1 . This pattern was repeated in the unadjusted gross M N data ( G U M N ) , but at predictably higher values. There were highly significant differences among the G U M N means (p < 0 .001, F = 34.8, df = 4, 130). The a posteriori comparisons yielded exactly the same groupings of means as did the N U M N data, at the same significance level. The T 6 and T 8 means were 543 and 569 m g N (kg soil)" 1 respectively, while the others ranged from 963 to 993 mg N (kg soil)" 1 . The unadjusted net and gross M N procedures both indicated the same basic pattern of N avai labi l i ty levels across the age sequence. 170 Vertical bars are 95% confidence limits FOREST FLOOR MINERAL SOIL 9 0 0 "1 ^ 800-|> 700-^ 600-^ 500-Z 400-_ 120 - i E 100 80H < Z 60-1 ~v 12001 O) 1100-\ 1000-D> 900-E, 800-z 700-600-500-400-1 140-1 |» 120H ^ 100 H o < 80-60-20-i 15-10-5-0-2.5 2-1.5 1-0.5-0 25-20-15-10-5-3 2.5-2 1.5-1-l I I I I l l I l l I I I 0.5-0 1 2 3 4 5 6 7 8 9 10 1112 I I I I I—I I I ! 1 — I — I — I 0 1 2 3 4 5 6 7 8 9 10 1112 YEARS AFTER HARVEST Figure 6.1 M e a n s and 95% confidence l imits of the N minera l iza t ion concentrations given by the four measures. (See text for abbreviations.) The adjusted data presented somewhat different patterns to the above; moreover, the adjusted gross M N ( G A M N ) values exhibited a different trend to those of the adjusted net M N ( N A M N ) (Figure 6.1). There were very significant differences (p < 0.01, F = 4.0, df = 4, 130) among the N A M N means. The T i l mean was significantly (p < 0.05) greater than those of the T M and T6 sites; no other differences among the means were evident. The T i l mean N A M N was 95 m g N (kg so i l ) - 1 . The others ranged between 71 (T6) and 89 (T3) mg N (kg soil)" 1 . H i g h l y significant differences (p < 0 .001, F = 16.6, df = 4, 130) were observed among the means of the G A M N data. The T 6 mean was significantly (p < 0.05) lower than those of a l l except the T 8 site, from which it was not significantly different. In addition, the T 8 mean was s imi lar ly significantly lower than those of the T3 and T i l sites. The T 6 and T 8 G A M N means were 82 and 97 m g N (kg soil)" 1 respectively; those of the other sites ranged between 115 (TM) and 141 (T3) mg N (kg soil)" 1 . The trends from both the N U M N and G U M N data suggest that no change in potential N mineralization rates took place unt i l after the third year following disturbance. A t this point, the rates declined for the next five years, then recovered wi th in the three subsequent years . The N A M N and G A M N trends were assumed to be more reflective of field conditions than their unadjusted counterparts. They present conflicting pictures. The N A M N data pattern implied no real increase or decrease in potential N mineral izat ion occurred during the first eight years. However , a significant increase was indicated by year 11. This conflicts wi th the trends of both the unadjusted indices and the G A M N values. The G A M N trend agreed wi th the unadjusted indices in that a definite decline 172 and recovery pattern was indicated between years three and eleven. One extremely interesting point is that both adjusted indices indicated a non-significant but noticeable "increase" in potential N mineral izat ion by the third year. The margins defining this point as not being significantly different from the T M mean were extremely narrow. A n increase in potential N mineral izat ion by year three would have been entirely possible in light of the increases in N concentrations and drop in C / N ratios observed in the organic materials at that point (see Chapter Five) . Considering this possibility, and the results presented above, it appears that the G A M N trend may be a more accurate representation of the post-harvest potential N availabil i ty situation in the forest floor than the other M N indices considered for that fraction. V e r y few published studies have applied the anaerobic incubation technique to forest floor materials . K l i n k a et al. (1982), Ma t son and Boone (1984), and H a s k i n (1985) were three such studies; the first two used the net M N measure, while , the th i rd employed the gross M N index. K l i n k a et al. (1982) were examin ing mature E S S F stands; the other two studies were investigating disturbance effects in other high-elevation western forests. Both the N U M N and G U M N levels observed in organic materials of this study are much higher (two to three times) than those reported in the latter three studies. K l i n k a et al. (1982) reported a net M N level of 239 mg N (kg soil)" 1 ; Ma t son and Boone's (1984) observed range was between 40 and 225 m g N (kg soil)" 1 approximately. H a s k i n (1985) observed a range of gross M N concentrations of between 219 and 540 mg N (kg soil)" 1 . The reasons why the forest floor values of this study were so much higher than those of the others are not 173 immediately obvious. Given that the minera l soil M N values (see next Section) were ent irely comparable wi th those of the others for a given technique, no analyt ical bias is suspected. However , it is difficult to see why the potential N mineral izat ion levels of the organic materials of the study area should have been so markedly higher than elsewhere. Comparisons of trends of M N levels after disturbance can also be made. H a s k i n (1985) reported that forest floor gross M N trends were inconclusive in her study. However , the G U M N trends of this study are s t r ikingly s imilar to that of Ma t son and Boone (1984) for the O l horizon—a mixture of L and some F materials (Pritchett, 1979). Tha t is, a decrease in potential N mineralization after disturbance wi th subsequent recovery were indicated. The opposite trend (increase followed by decrease) was noted i n the 0 2 horizon (Matson and Boone, 1984). Ma t son and Boone (1984) theorized that woody materials and high C / N ratios m a y have operated to decrease the potential for N mineralization in the 0 1 horizon. These were unlikely to have been the reasons for decreased levels in this study. A direct decrease in N mineral izat ion (and availabil i ty) would run contrary to expectations based on the concept of an assart effect. This might therefore imply that the unadjusted indices of forest floor N mineralization were not sufficiently sensitive to the on-site conditions in this study. Al ternat ively , such trends m a y be viewed as an indirect consequence of the so-called "Gadgi l effect" (Gadgil and Gadgi l , 1975 and 1978; B e r g and Lindberg , 1980). W i t h mycorrhizal suppression of decomposition—presumably highest on the mature stand—there would be a build-up in the pool of potentially mineralizable N . A sample extracted from such materials would undergo a release from the suppression, and 174 thereby exhibit a relatively high M N value—indicating a spurious N avai labi l i ty level. O n harvested sites, mycorrhizal suppression should be greatly decreased ini t ia l ly ; this would allow an ini t ial ly high rate of mineralization—as in the case of the samples above. A s time progressed, the potential for N mineral izat ion would be decreased as the more recalcitrant fractions are encountered; samples taken from such mater ia l would correctly reflect this decline in their M N values. W i t h increasing canopy (tree) development, the suppression phenomenon would gradual ly reassert its control. Therefore, laboratory incubation techniques m a y not adequately reflect the "true" situation in horizons where mycorrh iza l suppression is a major factor—especially in comparative studies such as this one. B y extension, the minera l soil incubations of this study m a y wel l be more indicative of the " true". trends —assuming lower numbers of mycor rh iza l roots. The results of Ma t son and Boone (1984) for the 0 2 horizon do not fit the foregoing conjectures. Unfortunately, no forest floor depth data were given in the latter study; this might have allowed some speculation as to the importance of mycor rh iza l suppression to the materials involved. There were no available data wi th which the magnitudes of the adjusted values ( N A M N and G A M N ) could be directly compared. A s noted earlier, their indicated trends are more in line wi th theoretical expectations than their unadjusted counterparts. Nevertheless, the cri t icisms outlined above should also be applicable here. Thus , the mineral soil values m a y again be more reflective of the "true" trends of N availabil i ty across the age sequence. 175 6.4.1.2 Mineral Soil Materials There were highly significant differences (p < 0.001, F = 10.8, df = 4, 130) among the means of the unadjusted data for the net M N procedure ( N U M N ) (Figure 6.1). The means of the T 6 and T i l cutovers were significantly (p < 0.05) higher than those of the T M , T 3 , and T 8 sites; wi th in each of the two different groups thus formed, there were no differences among the means of the member sites. The T 6 and T i l means were 12 and 14 mg N (kg soil)" 1 respectively; those of the other sites fell between five and six mg N (kg soil)" 1 . Un l i ke the forest floor case, this pattern was not fully repeated in the unadjusted gross M N data ( G U M N ) . H i g h l y significant differences were indeed observed among the G U M N means (p < 0 .001, F = 9.2, df = 4, 136). The T 6 and T i l means were significantly (p < 0.005) greater than those of the T M and T 8 sites; in addition, the T 3 mean was significantly greater than that of the T 8 . There appeared to be no other significant differences among the G U M N means. The T 6 and T i l means were 16 and 20 m g N (kg soil)" 1~ respectively; the others ranged from 10 (T8) to 14 (T3) m g N (kg soil)" 1 . Thus, the unadjusted net and gross M N procedures both indicated the same basic trend, but w i t h diss imilar individual differences among the sites. The trend was that of increased potential N mineral izat ion at the six-year mark after harvesting, followed by a sharp decline to mature stand levels by year eight. Somewhat curiously, potential mineralization levels apparent ly increased again by 11 years after the harvest. Ac tua l N availabil i t ies are unl ikely to have increased with increasing vegetative cover and stand closure. It is more plausible that the secondary increase was a manifestation of the "potential-measuring" nature of 176 incubation methods. Total N concentrations in the minera l fraction declined to year eight then recovered (Chapter Five); the secondary increase is believed to be a reflection of this. In sum, then, it was concluded that the N supplying power of the minera l fraction was increased to a peak at six years after logging; a sharp decline to pre-disturbance levels occurred by year eight. A secondary increase in N availabil i ty is not believed to have occurred. The adjusted data reinforced the basic patterns and interpretations noted above, but at much lower magnitudes. The adjusted gross M N ( G A M N ) pattern exhibited a different trend to that of the adjusted net M N ( N A M N ) (Figure 6.1). There were highly significant differences {p < 0 .001, F = 11.2, df = 4, 125) among the N A M N means. The a posteriori comparisons yielded exactly the same result as in the unadjusted net ( N U M N ) instance—the T 6 and T i l site means were significantly (p < 0.05) greater than the rest, wi th no other apparent differences. The T 6 and T i l means were approximately 1.6 and 1.7 m g N (kg soil)" 1 respectively; the others fell between 0.5 (TM) and 0.8 (T8) m g N (kg soi l ) ' 1 . H i g h l y significant differences {p < 0 .001, F = 10.8, df = 4, 130) were also observed among the means of the G A M N data. However , individual differences were dissimilar to those for their unadjusted counterparts. W i t h the exception of the T8 mean, al l cutover means were significantly (p < 0.05) greater than that of the mature stand. In addition, the T i l mean was once again greater than that of the T 8 . There were no differences between the T 8 and mature stand values. The T M and T 8 means were 1.1 and 1.5 m g N (kg soil)" 1 respectively; the others ranged from 1.8 • (T3) to 2.2 ( T i l ) m g N (kg soil)" 1 . Thus , the main feature of G A M N data here was that increased 177 mineralization was indicated earlier—at the three-year point—than for the others. The trends of both the NUMN and GUMN data were discussed earlier. The NAMN and GAMN trends were assumed to be more reflective of field conditions; the reasoning used in the NUMN and GUMN cases is also applicable here. The NAMN and GAMN trends strongly supported those of the unadjusted indices. Moreover, the GAMN pattern suggested that potential N mineralization levels were increased as early as three years after harvesting. The unanimity found in the trends, as well as their concordance with the theoretical expectations, supports the argument that the mineral soil MN trends may be more indicative (than those of the forest floor) of the "true" situation which obtained on the sites. The mean NUMN and GUMN values observed in this study compare well with others found in the literature for comparable techniques and conditions. The range of the NUMN means is below that noted by Klinka et al. (1982) under mature ESSF stands, but above that of Matson and Boone (1984) in the Oregon Cascades. The range of the GUMN means is very similar to those noted by Haskin (1985), and well within the ranges reported by Powers (1980 and 1984a). However, it was below that observed by McNabb et al. (1986) for a similar technique and soil fraction in the Oregon Cascades. These favourable comparisons lend support to the idea that the mineral fraction MN values may present a more accurate picture of the post-harvest situation than their forest floor counterparts. The indicated patterns of potential N availability after disturbance also compare reasonably well with the findings of other studies which 178 employed anaerobic incubation techniques. The pattern of the N U M N values for the first eight years after harvesting is similar to that of Matson and Boone (1984) for a sequence of disturbed sites; however, the seeming secondary increase (after the eight-year point) was not apparent in the latter study. The G U M N trends of this study are strikingly similar to those of Haskin (1985)—including the apparent secondary increase in mineralization levels. For her sites, the first peak and decline occurred at eight and ten years after logging respectively, while the secondary increase was at year 12. Interestingly enough, Haskin (1985) also noted a later (secondary) decline by year 13—beyond the range of cutover ages used in this study. This raises the question of whether the dual peaks implied in this study and noted in hers were indeed accurate representations of their respective "true" situations, rather than artifacts of the method and total N changes as discussed earlier. Unfortunately, Haskin (1985) did not include a consideration of total N values. As with their forest floor counterparts, there were no available data with which the magnitudes of the adjusted values ( N A M N and G A M N ) could be directly compared. As noted earlier, they reinforce (and, in the G A M N case, perhaps refine) the trends indicated by the unadjusted values. The four unanimously pointed toward a trend of increased post-harvest N availability after the three- and before the eight-year points, with a peak occurring at six years after harvesting. Moreover, the G A M N data suggested that the increase could have occurred at the three-year point or even a little earlier. This is in accordance with the forest floor G A M N trends, as well as with the seeming increase in forest floor total N at three years after logging (Chapter Five). 179 Therefore, this index m a y wel l be the most "accurate" one. The results also indicated a secondary increase in N avai labi l i ty between eight and 11 years . While this was noted by an earlier investigation (along wi th a secondary decline), the rather conservative position was taken that the phenomenon was an artifact in this study. 6.4.1.3 Unit-Area Weights O n a per-cent-of-total-N basis, highly significantly differences (p < 0.001) were again found among the site means in each of the four data sets. These differences wi l l not be discussed, since interest was more in obtaining mean and interval data for the unit-area weight estimates. Under the assumption stated earlier, the result ing estimates of M N levels on a per-hectare basis are presented in Table 6.1. The patterns maintained a close reflection of the paral le l concentration data presented earlier, but wi th less indicated differences across the site sequence. Because of the phenomena involved and methodological differences, comparisons of these data wi th values from the literature on an absolute basis are of extremely l imited value. The adjusted forest floor data ( N A M N and G A M N ) are reasonably close to the annual N mineralization estimate for humus given by Wi l l i ams (1983b) —approximately 17 kg ha" 1 . However , the adjusted mineral soil data ranges are far below those calculated by M c N a b b et al. (1986). The unadjusted ( N U M N and G U M N ) minera l soil values better fit the latter range. The unadjusted forest floor values are closer to the values observable from L- layer materials (Wil l iams, 1983b). Since that layer was specifically excluded from sampl ing, this might support the view that the adjusted indices 180 Table 6.1 Means and 95% confidence limits of the estimated unit-area weights of mineralized N given by the four measures. (See text for abbreviations.) F R A C T I O N / S I T E / V A L U E S V A R I A B L E T M T 3 T 6 T 8 T i l (values rounded; figures in brackets are 95% limits) FOREST FLOOR N U M N 56 t (65,47) . 58 (71,45) 40 (45,35) 38 (43,33) 56 (65,47) G U M N 84 (98,70) 79 (95,63) 44 (49,39) 42 (48,36) . 70 (80,60) N A M N 6 (7,5) 8 (10,6) 6 (7,5) 6 (7,5) 7 (8,6) G A M N 10 (12,8) 11 (13,9) 7 (8,6) 7 (8,6) 9 (10,8) (INERAL SOIL N U M N 6 (9,3) 7 (9,5) 15 (21,9) 8 (12,4) 15 (24,6) G U M N 13 (18,8) 17 (20,14) 21 (29,13) 15 (21,9) 21 (33,9) N A M N 0.6 (0.9,0.3) 0.3 (0.4,0.2) 0.3 (0.4,0.2) 1.1 (1.6,0.6) 1.7 (2.7,0.7) G A M N 1.4 (1.9,0.9) 2.1 (2.5,1.7) 2.8 (3.9,1.7) 2.2 (3.1,1.3) 2.4 (3.8,1.0) f A l l units are k g ha" 1 181 give a more representative picture. Insofar as this is true, it may be stated that absolute magnitudes of M N levels did not change drastically across the site sequence. The main change from mature stand levels appeared to be a maximum increase of approximately 1.4 kg ha" 1 from the mineral fractions approximately six years after harvesting. 6.4.2 Chemical Extractions The trends of concentrations of KCl-extractable NHJ -N, N O i -N, and their sums (referred to as TIN—total inorganic N) are given in Figure 6.2 for the organic and mineral fractions. 6.4.2.1 Forest Floor Materials There were highly significant differences (p < 0.001) among each of the NHJ -N, NO5 -N, and TIN concentration values of the sites. The jackknife 95% limits indicated that the mean N H J - N concentrations of the T6, T8, and T i l sites were significantly lower than that of the mature stand. In addition, the T6 and T8 means were significantly lower than the others, but not different from each other. The mean NHJ -N concentrations ranged from approximately 346 mg N (kg soil)" 1 on the T M site down to approximately 51 mg N (kg soil)" 1 on the T6 site. For the NO3" -N data, with the exception of the T6 site, the mean concentration of the T3 site appeared to be significantly higher than the others. In addition, the three remaining cutover means were not significantly different from each other; however, of these only the T8 and T i l means were 182 Vertical bars are 95% confidence limits FOREST FLOOR MINERAL SOIL YEARS AFTER HARVEST Figure 6.2 Means and 95% confidence limits of the inorganic N concentrations given by KC1 extraction. (TIN = total of N H J -N and NO5-N.) 183 significantly different from that of the mature stand. The margins for significance or otherwise in the last three compared to the T M were very narrow. The T3 mean N O 3 - N concentration was approximately 13 mg N (kg so i l ) - 1 , while the others fell between 1.2 and 3.3 mg N (kg soi l ) - 1 . The T I N means exhibited exactly the same pattern of differences as the N H J - N concentrations. The above results indicate that N H J - N was by far the dominant form of inorganic N on all sites. Its trend across the site sequence was surpris ingly s imi lar to that indicated by the unadjusted incubation data ( N U M N and G U M N earlier). The indicated decline can be considered in s imi lar terms to those outlined earlier; that is, the extractions reflect the fact that as time proceeded the more resistant fractions were encountered. A s an index of N avai labi l i ty , the N H J - N values are probably seriously constrained by the lack of incorporation of environmental and microbiological conditions and processes. However , they do provide l imited support for the incubation results wi th in these constraints. The extractable N O 3 - N trends are of extreme interest in that they reflect some measure of nitrification which was highest at the three-year point, and possibly elevated sl ight ly above mature stand levels for the rest of the period considered. This result is even more exciting in view of the fact that chemical extraction dur ing the p re l iminary survey in the fall of 1982 had indicated negligible [less than 1.0 m g N (kg s o i l ) - 1 ] N O 3 - N levels in forest floor materials on a l l sites. It is possible that the summer sampling highlighted a N O 3 - N presence that might have been missed by fall sampling because of intervening precipitation events. M a r k e d l y increased N O ^ - N levels at year three would imply that that was at least one point of m a x i m u m N availabil i ty across the age sequence. The 184 resin data should therefore shed further light in this respect. Taken together, the N H J - N and N O ] - N extraction data substantiate the patterns hinted at by first the total N data (Chapter Five) and subsequently the adjusted incubation data. That is, there appeared to be increased N avai labi l i ty at three years following the harvest; Th is avai labi l i ty declined between years three and eight, but recovered its mature stand level by year 11. The m a x i m u m KCl-extractable N H J - N concentrations observed in the organic materials of this study are one order of magnitude higher than those reported in the l i terature. The reasons for this are not immediately apparent; this result accords wi th the high M N values noted earlier. However , both the lower N H J - N and the general range of N O 3 - N values are comparable to those observed in forest floor materials in Eas t e rn Canada (Freedman and Morash , 1985; Wal lace and Freedman, 1986). N o directly comparable data were available for higher elevation forest floor materials . In terms of trends following harvesting, Wallace and Freedman (1986) noted higher N H J - N in the forest floors of clearcuts as opposed to those of mature stands; no differences in N O ] - N were observed among sites. It should be noted that i n the latter study the frequency of sampl ing for N O i - N was such that accumulations might have been missed i f precipitation events occurred between sampl ing dates. A s noted in Chapter One, increased NO5 - N levels following harvest ing have been documented by several studies. 185 6.4.2.2 Mineral Soil Materials A s wi th the organic materials, highly significant differences {p < 0.001) were apparent among each of the N H J - N , N O 3 - N , and T I N concentration values of the sites. Fo r the N H J - N data, the jackknife 95% limits indicated that the T 3 mean was significantly higher than all others; i n addition, that of the T 6 site was significantly lower than the T M mean concentration. The mean N H J - N concentration of the T3 site was 10.1 m g N (kg soil)" 1 ; the others ranged between 3.9 (T6) and 6.4 (TM) mg N (kg s o i l ) " 1 . Fo r the N O 3 - N data, the means of the cutover sites were al l significantly greater than that of the mature stand. The only other significant difference was that the T 3 mean was greater than the T 6 mean. The mean N O 3 - N concentration of the T M materials was 0.5 m g N (kg soil)" 1 ; the others fell between 1.8 (T6) and 6.7 (T3) mg N (kg soil)" 1 . F o r the T I N values, the only significant difference was that of the T 3 mean being greater than those of the T M , T 6 , and T 8 sites. The means ranged from 5.7 (T6) to 16.7 (T3) mg N (kg soil) ' 1 . The above results both parallel and complement the results for the forest floor materials . Unl ike the latter, the trends of a l l three variables are s imi lar across the age sequence. Moreover, N H J - N and N O 5 - N mean values were much closer together than in the forest floor case. The observed patterns indicated a significant increase i n al l inorganic N levels by year three; levels had declined by year six. Fo r N H J - N , the decline went temporar i ly below mature stand levels, but recovery was evident subsequently. Considered wi th the forest floor data, there was the implication of a downward movement of N H J - N from the 186 organic horizons to the upper minera l soil. The N O 3 - N data supported the pattern observed in the forest floor fraction; this is extremely interesting for reasons given earlier. The observed patterns generally supported the trends suggested by the incubation data. The data obtained compare favourably wi th those reported in the li terature. T a k i n g differences in sampl ing depth into account, the mean T I N values are s imi lar to those of Powers (1980) from western high-elevation coniferous (true fir and other) cover stands. The range of N H J - N mean values was very s imi lar to and narrower than those of Freedman and Morash (1985). Whi le there is some evidence that nitr if ication m a y not be as rare in acid forest soil conditions as once believed (Lee and Stewart , 1978; Robertson, 1982; Lee et al., 1983) the low mean N O 3 - N value of the mature stand accords well with accepted ideas. The elevated NO5 - N levels in the cutovers agree with previously documented observations. The range of NO5 - N values were somewhat lower than those of Freedman and M o r a s h (1985). Fo r sites involving disturbance, the ranges of values for inorganic N in this study are higher than those of Vitousek and Mat son (1985) in the southeastern Uni ted States, but lower than those of M a r t i n (1985) in coastal B . C . M a r t i n (1985) noted no differences in N H J - N concentrations across his age sequence; however, increases in NO5 - N concentrations were observed by year four—a result supported by the findings of this study. 187 6.4.2.3 Unit-Area Weights H i g h l y significant differences (p < 0.001) were observed among the site values of each of the three data categories on a per-cent-of-total N-basis . A s stated earlier, these differences wi l l not be discussed. The resulting estimates of KCl-ext rac table inorganic N levels on a per-hectare basis are presented in Table 6.2. A s might be expected, the trends reflected those of the concentration data earlier. The low overall contents of NO3" - N i n the forest floor materials became even more apparent, while the same is true for the closeness of N H J - N and NO5-N contents of the mineral soil. Fo r the reasons stated in Chapter F i v e , the forest floor unit-area weights are l ikely to be overestimates. This is supported by the fact that the values of this study are generally much higher than those of eastern softwood forest floors of s imi lar depths (Freedman and Morash , 1985). The trends of change are thus of more importance than absolute magnitudes. M i n e r a l soil values are not directly comparable to those of other studies because of methodological differences. Nevertheless, t ak ing differences in sampl ing depths into account, the mineral soil values are comparable to those noted by M a r t i n (1985) on the B . C . coast. The trends of change wi th time after harvest ing of the inorganic N contents are also very s imi lar to those of M a r t i n (1985). 6.4.3 Ion Exchange Resins A total of six cation and seven anion resin bags were excluded because of damage—presumably by soil fauna. The patterns of change in the N H J - N and N O 3 - N concentrations adsorbed onto the remain ing resins after approximately one 188 Table 6.2 Means and 95% confidence limits of the estimated unit-area weights of inorganic N given by KC1 extraction. (TIN = total of N H ; -N and NO5 -N.) F R A C T I O N / SITE / V A L U E S V A R I A B L E T M T3 T6 T8 T i l (values rounded; figures in brackets are 95% limits) FOREST FLOOR N H J - N 29t 20 4 4 14 (34,24) (26,14) . (5,3) (5,3) (17,11) N O i - N 0.1 0.6 0.1 0.2 0.1 (0.1,0.1) (0.8,0.4) (0.2,0.1) (0.2,0.1) (0.2,0.1) TIN 29 21 4 4 14 (34,24) (27,15) (5,3) (5,3) (17,11) UNERAL SOIL N H J -N 8 11 6 7 6 (11,5) (13,9) (9,3) (10,4) (9,3) NO5-N 0.6 7 3 4 4 (0.9,0.2) (10,4) (4,1) (7,2) (7,1) TIN 8 18 8 11 10 (11,6) (22,13) (12,5) (16,6) (16,3) tAll units are kg h a - 1 189 year in situ are presented in F igure 6.3. N o significant differences were observed in the mean N H J - N concentrations among the sites (p > 0.05, F = 0.6, df = 4, 139). The means ranged from 3.0 to 3.9 m g b a g ' 1 . B y contrast, there were highly significant differences among the mean N O ^ - N values (p < 0 .001, F = 10.5, df = 4, 138). The T3 mean was significantly (p < 0.05) greater than a l l the others; in addition, the T i l mean was significantly lower than that of the mature stand. No other significant differences were apparent among the sites. The T3 mean N O 3 - N concentration was 2.1 m g b a g - 1 , and that of the T i l 1.2 mg b a g - 1 . The other means ranged between 1.3 and 1.6 m g bag" 1 . The resin N H J - N results do not support those of the incubations and KC1 extractions considered earlier. Four m a i n possibilities exist: F i r s t , it is possible that the resin data accurately reflect the "true" situation; that is, there were no post-harvest differences in N H J - N avai labi l i ty . To some extent, such a result would be similar to that of Wallace and Freedman (1986), who noted no differences in ammonification after harvest ing. However, it would run contrary to the evidence of the other N avai labi l i ty indices of this study, as wel l as to the documented assart phenomena. The second possibility is that the result is more related to the factors governing adsorption onto the cation resins than to any reflection of N availabil i ty. The findings of B ink ley (1984), presented earlier, support the latter view; in addition, M a r t i n (1985) noted a s imi lar lack of resin N H J - N differences where other measures suggested that differences in N availabi l i ty did indeed exist. The th i rd possibility is that high within-site var iabi l i ty may have masked "true" differences; this possibility m a y also be related to the second one. The fourth possibility is that the resin bags a l l 190 0 J 0 1 2 3 4 5 6 7 8 Y E A R S AFTER H A R V E S T 7 i 9 n i 1 10 11 12 Figure 6.3 Means and 95% confidence limits of the inorganic N concentrations adsorbed by the ion exchange resins. 191 became saturated over the period in situ, thus g iv ing the same reading for N H J - N ; this is believed most unl ikely, given the high rated exchange capacity of the cation resins (3.3 meq g" 1 dry weight) and relat ively low concentrations recovered by extraction. The other results of this s tudy and the above considerations led to the conclusion that the resin N H 4 - N data did not accurately reflect the post-harvest situation; the second and third possibilities listed above are believed to have been effective in this regard. In contrast to the above, the NO5 - N concentrations adsorbed onto the anion resins presented an excit ing picture. They strongly supported (and were supported by) many of the earlier indications, but were more concrete in that they represented an integration and expression of the conditions over t ime. The data strongly indicated that a peak in N O 3 - N formation—hence N availability—occurred during the fourth year after harvest ing; this had subsided by the seventh year, and apparently had declined to sl ightly below pre-harvest levels by year 12. This result is comparable to that of M a r t i n (1985), who noted a peak and subsequent decline at five and eight years after harvest ing respectively. While significant differences were observed in this study, their magnitudes relative to that of the mature stand's resin N O 3 - N levels were low when compared to other s imi la r studies. The peak increase here was only 1.3 times the level of the mature stand; both Bink ley and Packee (1982) and M a r t i n (1985) reported very much larger increases —up to 30 times control levels in the case of the latter study. Considering the KC1 extraction and resin results together, it would appear 192 that neither the KC1 N H J - N extraction nor the cation resin N H 4 - N adsorption methods could yield conclusive results when applied to or in the organic materials under consideration. The mineral soil KC1 data all gave consistent and instructive results. The anion resin NO5 - N procedure also yielded a consistent and informative result, but went further in that it represented an integration of conditions over time. In general, all data obtained from N O 3 - N assessments pointed toward the conclusion that levels of that anion were generally increased after harvesting. 6.5 CONCLUSIONS The results discussed in the foregoing suggest that the subsidiary hypotheses H 0 and H 0 should be rejected. Major changes in N availability from both forest floor and mineral soil materials appeared to have occurred during the post-harvest period. The several measures used differed in their apparent efficacy and indications given. The "potential" nature of the incubation measures needed to be borne in mind in examining the results. Moreover, in the case of the organic materials it is believed that influences of the "Gadgil effect" needed consideration. For the mineral fraction, changes in concentrations of total N appeared to influence the results in the latter portion of the age sequence. For reasons given earlier, the mineral soil incubation results are believed to be more reflective of the "true" post-harvest situation; similarly, the temperature-adjusted incubation measures are believed to be more indicative than their unadjusted counterparts. The value and interpretation of KC1 extraction results are limited by their point-in-time character; in addition, they do not take any 193 direct account of microbial ly mediated processes. These were considered more as secondary (support) measures for interpretive purposes. Where the resins were concerned, the efficacy of the cation resin method appeared to be heavily constrained by factors governing ion adsorption; its results were therefore considered inconclusive. B y contrast, it is believed that the anion resin results yielded important quali tat ive information concerning post-harvest N avai labi l i ty trends. The latter method m a y thus be far more useful than the cation resins for s imilar studies which involve relat ively long in situ placements. Under the assumptions and limitations stated above, some general statements can now be made concerning the overal l indications. The incubation procedures were unanimous in suggesting a decline i n N avai labi l i ty from forest floor materials between years three and six of the sequence. W h a t was less clear from the incubations was whether there was any increase in the first three years; there was a strong (though non-significant) trend in this direction in the adjusted M N data. B o t h the KC1 and resin NO5 - N data supported the idea of a peak in N avai labi l i ty at years three to four; it was concluded that this was indeed the case. Therefore, in the forest floor N avai labi l i ty appeared to increase to a peak during the first three or four years, then to decline by year six. Though the "better" adjusted incubation ( G A M N ) result suggested that the decline went below pre-harvest levels, the NO5 - N (KC1 and resin) refuted this. It was thus concluded that the decline at year six went down to pre-harvest levels only, and remained there. Insofar as the adjusted incubations were indicative of actual changes in N mineral izat ion, the increase from the organic materials would have been fair ly small—less than two kg ha" 1 . The minera l soil results support the 194 view of increased post-harvest N mineralization by year three; however, the duration appeared to be unti l year eight in this case. The magnitude of any increase would have been the same as for the forest floors. The KC1 extractions also suggested a downward movement of N H , - N from the organic to the mineral fraction (and perhaps beyond). The forest floor N H J - N content declined by approximately 25 kg ha" 1 ; the 0-15 cm mineral fraction had an increase of approximately 3 kg ha" 1 at the year six. Both forest floor and mineral fractions exhibited apparent increases in NO3 - N contents; these were less than one kg ha" 1 and approximately six kg ha" 1 respectively. V i e w e d i n terms of an assart pattern, it can be concluded that the first eight years of the sequence witnessed increased levels of N avai labi l i ty . These were at a peak around the three-to-four-year mark for the forest floor. The minera l soil peak also occurred at year three, but m a y have been more sustained (to year six) than that of the forest floor. The possible magnitudes of changes were stated above. These conclusions and those of Chapter F ive are mutua l ly supportive. CHAPTER 7 PRE- AND POST-HARVEST TREE GROWTH In the foregoing Chapters , consideration was given to soil physical and chemical characteristics following the harvest. W e now turn our attention towards a s imilar examinat ion of the subalpine fir advance regeneration. In this Chapter, their height and diameter growth before and after logging are considered. 7.1 INTRODUCTION The results of the phases considered previously provided evidence of an assart effect occurring principal ly wi th in the first eight years after harvesting, wi th a peak between three and six years. A s discussed earlier (Chapter One), the next step was to examine whether there were benefits accruing to the advance regeneration present; the context is therefore that of m a i n hypothesis H 0 2 stated previously. It was believed that an examinat ion of pre- and post-harvest growth patterns over the site sequence would provide insights in this respect. Considering two extremes, i f regeneration exhibited no increase in annual growth rates after release by logging, it might be assumed that they were not benefiting from the assart flush. Conversely, i f a full and immediate response was noted, major benefits could be assumed to have accrued to the regeneration. The period of response delay is of great importance to second-growth yields—especially where subalpine fir is concerned (see Chapter One). Harr ington and M u r r a y (1982) noted that while true firs can recover from early suppression to give excellent growth, they remain shorter than trees of the same age that 195 196 emerged more rapidly from the juvenile phase. Severa l studies have examined the post-harvest growth of subalpine fir, both for logging residuals and advance regeneration. Those in B . C . include Stettler (1958), H e r r i n g (197.7), Her r ing and M c M i n n (1980), Monchak (1982), and Bergs t rom (1983). Others include Crossley (1976) and Johnstone (1978) in Alber ta , and McCaughey and Schmidt (1982) in the Uni ted States. Stettler (1958) recognized three periods in the release response of subalpine fir residuals. The first five years was the period of adaptation; a period of m a x i m u m response occurred between ten and fifteen years after harvest ing, while declines in growth were noted after 20 years. For advance regeneration, the response delay in height growth is usual ly two to five years after logging (Herring, 1977; Johnstone, 1978; McCaughey and Schmidt, 1982); rad ia l growth response appears to be more rapid (Herr ing, 1977). Crossley (1976) noted that radial growth gave a more consistent and pronounced response than height growth. Other studies suggest that the response delay period for height growth noted above is applicable to other western true firs as well (Gordon, 1973; Seidel, 1980; Helms and Standiford, 1985). The physiological origins and other aspects of the response delay phenomenon were discussed by Ferguson and A d a m s (1980) and Helms and Standiford (1985). The question of what annual growth rate indicates that a release response has occurred is of importance. There are at least two ways of approaching this. The first would be by comparison of observed current growth rates to those which occurred historically both before and after the harvest—the latter 197 presumably relat ively constant for advance regeneration. The second approach is more in tune wi th practical forest management; it would involve a simple assessment of the time taken to achieve a pre-set rate defined as "acceptable growth". Where current height growth of subalpine fir advance regeneration in the southern interior of B . C . is concerned, the m i n i m u m standard can be taken as 15 cm y r " 1 (Ivanco, 1985; B . C . M . O . F . , 1986b). Severa l studies have focussed on the issue of predicting growth response of true fir (and other) advance regeneration; these employed models of vary ing complexity wi th va ry ing degrees of success (Johnstone, 1978; Ferguson and A d a m s , 1980; McCaughey and Schmidt, 1982; He lms and Standiford, 1985; Seidel, 1980 and 1985; Ferguson et al, 1986). Detai led consideration of these efforts is outside of the scope of this study. Nevertheless, some of their results—along wi th those of non-predictive studies mentioned earlier—highlight three points which are germane here. F i r s t l y , i t appears the general consensus that total height at time of logging affects the release response growth. What remains an area of argument is precisely wha t the effects are. Some results suggested that higher ini t ia l response rates occur in taller trees compared to shorter ones (Stettler, 1958; Gordon, 1973; Johnstone, 1978); Seidel (1980) noted no such effect; yet other studies impl ied that shorter trees respond better than taller ones (Ferguson and Adams , 1980; M c C a u g h e y and Schmidt, 1982). Both H e r r i n g (1977) and Monchak (1982) m a y be included in the last category, though the effects of release height were indirect in their cases. The fact that ini t ial height m a y exert some influence on absolute growth suggests that a relative measure of release growth m a y be more suitable. The second issue highlighted is 198 that growth in the five-year period preceding the harvest has been found a useful predictor of response growth (Ferguson and A d a m s , 1980; McCaughey and Schmidt , 1982; Helms and Standiford, 1985; Seidel, 1985). Thus, it might be useful to include this growth period in that used to provide comparative historical growth patterns. The final point of interest is that of the effect of age at release on subsequent growth. Several investigators have either concluded or implied that age at logging does not directly affect growth response of subalpine fir (Crossley, 1976; Herr ing , 1977; Johnstone, 1978; Monchak, 1982). This appears to be the dominant view in practice in the southern interior of B . C . at present. However , Stettler (1958) was of the opposite view for the same species. Moreover , age has been found to affect growth response inversely in other western true firs (Ferguson and Adams , 1980; Seidel, 1985). It was therefore believed useful to include a simple assessment of possible "effects of age on release growth. 7.2 O B J E C T I V E S A N D H Y P O T H E S E S The p r imary objectives of this phase were to provide information concerning pre- and post-harvest height and diameter growth patterns, and to assess par t ia l ly to what extent growth of advance regeneration might have been enhanced by the assart pattern noted earlier. A s stated earlier, ma in hypothesis H 0 g is involved here. A s noted in Chapter One, i f this hypothesis is true, there should have been no detectable differences i n growth of advance regeneration over time on the sites. Unl ike the soil considerations, no formal statement of subsidiary hypotheses wi l l be made in this case. Rather, conclusions concerning 199 the above main hypothesis w i l l be d rawn from a simple inspection of the data patterns. Addi t ional investigations of the veraci ty of ma in hypothesis H 0 are also presented in Chapter Eight . 7.3 METHODS The basic layout and sampl ing scheme were discussed in Chapter Two . This included the procedure used in the measurement of annual height growth in 1983. Fur the r details are provided below. 7.3.1 Height growth A l l data were processed wi th in the M I D A S programme. U s i n g the terminat ion of 1982 growth as a s tar t ing point, the height at release of cutover trees was estimated by subtracting each successive year 's measured growth back to the year of logging. Only trees yielding a release height greater than zero centimetres were retained in the data set. A relative measure of each year 's height growth (RHG) was calculated by dividing each year 's absolute annual height increment by the height at release. The same units were used in the numerator and denominator for this calculation; the R H G figures were expressed as cm c m " 1 , but could also be considered in other terms. No R H G calculations were applied to the mature stand's advance regeneration. Fo r each cutover, the period commencing at five years before the year of logging and ending in 1982 was used as a basis for establishing absolute annual 200 height growth trends. These trends were also compared graphical ly to those exhibited by the mature stand's advance regeneration for the same period. The R H G values were plotted in s imi lar fashion, but without any comparisons to mature stand values. For assart comparisons, the period of response delay was rather arbi t rar i ly considered to be the time taken to achieve and sustain for two consecutive years a mean growth rate of not less than two times the mean periodic increment of the five years prior to the year of harvest. This was applied to both the absolute annual growth and R H G values. In addition, i n the case of absolute growth, the time taken to achieve the 15 cm y r - 1 s tandard was assessed. 7.3.2 Diameter Growth and Age Diameter growth patterns and age were determined from the stump-height (0.30 m) discs taken in the fall of 1984. S tar t ing at the terminat ion of 1984 growth, radial growth and age measurements were performed on each disc using a Parker Instruments Model Three annual growth r ing measur ing machine coupled to an Apple He microcomputer. M a n y of the discs were asymmetr ica l , and wi th extremely narrow rings; a measurement path giving the m a x i m u m number of rings perpendicular to it was chosen on the wider portion of the disc. The data obtained above were processed wi th in the M I D A S programme. Absolute diameter growth in each year was calculated as two times the rad ia l growth measurement; in this sense, the diameter growth data are somewhat idealized. The stump height diameters (Dsh) and ages at release were determined 201 by the same approach as used wi th the height data. No corrections were applied to the age estimates for time taken to attain stump height. This could have been as low as four and as h igh as seventeen years (Watts, 1983). In addition, a relative measure of each year 's diameter growth (RDG) was also calculated in s imi lar fashion. The data were examined using the same approach as wi th the height data. In order to examine s imply the idea of whether there were any associations between age at release and release growth, simple correlation analysis was carried out on the grouped data of the three oldest cutovers. This analysis examined the degree of association between age at release and both the total height growth and the total diameter growth for the five-year period following the year of harvest. The rationale here was that any influence of age should be most pronounced dur ing the period of adaptation to changed conditions. 7.4 R E S U L T S A N D D I S C U S S I O N The patterns of absolute annual height and diameter growth are presented in Figures 7.1 and 7.2 respectively. Figures 7.3 and 7.4 illustrate the trends of the R H G and R D G measures respectively. The vertical scales of the latter two Figures have been kept the same so as to facilitate comparisons between them. A summary of the indicated response delay estimates is presented in Table 7.1. It was immediately obvious that the T 3 result was anomalous; that is , a noticeable increase in growth occurred even before the year of harvest. Th is is explained by the fact that understory growth on the relatively smal l T3 area was influenced by the harvest ing of the large neighbouring T i l site. It appears that the T3 understory responded to the increased availabil i ty of light. It is 202 = Advonce regeneration in mature stand (or same period Cut 1980 V T3 • — • -24-* - 21-I t— 18-o cc 15-t— 12-I o UJ 9 -X Z 6-< 2 i-Cut 1977 T6 2 41 21-X 18-o cc 15-12-X o U J 9 -X z 5 -< U J 2 3-Cut 1975 / T8 • — • T V 24-o 21-X 18-o Q. 15-O t— 12-X o U J 9 -X z 6 -< 2 J -Cut 1972 T11 -6 - 5 - 4 - 3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 YEARS BEFORE/AFTER HARVEST Figure 7.1 Pat terns of absolute annual height growth of advance regeneration. 203 : Advance regeneration in mature stand for some period Cut 1930 T3 y \ Cut 1977 T6 Cut 1975 18 • — • — • — « Cut 1972 Til / ° i i i i i i i i i — i i — i — i — ~ i — i — i — i — i — i - 6 - 5 - 4 - 3 - 2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 YEARS BEFORE/AFTER HARVEST Figure 7.2 Pat terns of absolute annual diameter growth of advance regeneration. E 1-] 0.8-E 0.6-o RH 0.4-< 1 • 1 0.2-L_LJ 0-1-. 0.8-E 0.6-o RH 0.4-< 1 i I 0.2-1 * 1 0-1-i 0.8-£ 0.6-o a: 0.4-< l . I 0.2-l - L J 0-1-i £ ^ 0.8-j £ 0.6H o £ 0.4 H z < 0.2 UJ 0 Cut 1980 Cut 1977 \ Cut 1975 Cut 1972 1 ' ' i i r - — i 1 1 1 i 1 2 3 4 5 6 7 8 9 10 11 -6 - 5 - 4 - 3 -2 -1 0 YEARS BEFORE/AFTER HARVEST Figure 7.3 Patterns of relative height growth of advance regeneration I-i IT 0 . 8 -E 0 . 6 -o RD 0 . 4 -< 1 . 1 0 . 2 -l_l_J 2 Cut 1972 / \ ~i i i i r 2 3 4 5 6 1 1 1 1 9 10 11 12 - 6 - 5 - 4 - 3 - 2 - 1 0 1      7 8 YEARS BEFORE/AFTER HARVEST Figure 7.4 Patterns of relative diameter growth of advance regeneration. 206 Table 7.1 S u m m a r y of response delay estimates. V A R I A B L E S I T E / V A L U E S T 6 T 8 T i l DIAMETER GROWTH absolute 2 t 1 3 relative 1 1 1 HEIGHT GROWTH absolute 4 5 5 relative 2 4 3 to 15 cm y r _ 1 5 6 10 f A l l units are years . unl ike ly that this effect extended characteristics in the unharvested T3 patterns prompted their exclusion from below. 207 to any soil or nutrient avai labi l i ty site. This anomaly in the T 3 growth the bulk of the discussion which follows 7.4.1 Height Growth Considering the three oldest cutovers, pre-harvest growth appeared to be in consonance wi th that which obtained in the mature stand's advance regeneration (Figure 7.1). O n l y the T 8 site appeared to have exhibited margina l ly better growth. This reinforced the assumption that the cutovers had s imi lar pre-harvest characteristics. F o r the oldest cutovers, five-year pre-harvest growth averages ranged from 4.3 cm y r " 1 (T6) to 6.1 cm y r " 1 (T8). Pre-harvest R H G trends were s imi lar but somewhat less uniform (Figure 7.3). The five-year averages were between 0.06 cm c m ' 1 (T6) and 0.09 cm cm" 1 (T8). In the post-harvest stage, the response delay period indicated by absolute height growth was consistently greater than that from the relative measure. Absolute growth rates indicated a response delay of four to five years, while that indicated by relative rates was two to four years (Table 7.1). The absolute delay indication agrees wel l wi th values from other studies as noted earlier. However , the relative measure may be a more sensitive reflection of the "true" reaction of the trees. If so, the results suggest that for subalpine fir there may be less of a lag between diameter and height growth responses than previously envisioned. More important ly in this study, the shorter delay of the R H G 208 measure indicates that the trees were able to respond when the assart effect was at its peak; absolute growth suggests that the peak might have been missed by about one year. Considered in terms of the 15 cm y r " 1 standard, there was a delay of five to ten years before this was achieved. This represents a serious lag from the viewpoint of practical forest management. Regardless of the length of the response delay, the post-harvest height growth rates fell wel l below those of which the species is capable. The highest post-harvest rate observed was 22 cm y r " 1 , seven years after harvesting on the T8 site; the other cutovers al l had m a x i m a below 20 cm y r" 1 . B y contrast, Monchak (1982) observed average release growth of 21 cm y r " 1 , wi th a range up to 25 cm y r " 1 , for comparable sites. H e r r i n g (1977) noted periodic post-harvest height increments of 16 to 24 cm y r " 1 , but regarded these as serious underestimates; he considered 44 cm y r " 1 a better estimate of post-harvest growth performance. Her r ing and M c M i n n (1980) reported mean current increments of 34 cm yr" 1 in advance growth on scarified sites. It is recognized that, according to Stettler 's (1958) phases, the period of m a x i m u m growth may not yet have been achieved. E v e n so, it seems somewhat unlikely that the higher rates above w i l l be achieved. This is supported by the diameter data (Figures 7.2 and 7.4), which indicated at least a subsequent levelling off of growth; whether this was temporary or applied only to diameter growth is unknown. A s s u m i n g these considerations to be va l id , i t would appear that while the regeneration was potentially able to benefit from the assart release, such growth m a y nevertheless have been constrained throughout the post-harvest period considered by this study. 209 7.4.2 D i a m e t e r G r o w t h Once again considering mainly the three oldest cutovers, pre-harvest diameter growth appeared very s imilar to but margina l ly below that which obtained in the advance regeneration of the mature stand (Figure 7.2). A s wi th the height growth, this reinforced the assumption of pre-harvest uniformity. F ive -year pre-harvest growth averages ranged from 0.03 to 0.06 cm y r" 1 (T8 and T 6 sites respectively). It is somewhat interesting that the T 8 site, which exhibited the highest height growth average, possessed the lowest diameter growth average. However , since statistical tests were not performed, the significance of any differences cannot be assessed here. Pre-harvest R D G trends were s imi lar and fair ly uniform (Figure 7.4). Their five-year averages fell between 0.04 and 0.05 cm c m " 1 . In the post-harvest phase, as wi th height growth, absolute growth indicated consistently greater response delay periods than relative diameter growth. Absolute growth indicated a delay of one to three years, while the relative measure suggested a period of one year throughout (Table 7.1). The delays accord wel l w i th those noted earlier from the l i terature. It would appear that the use of a relative measure did indeed improve delay estimates for diameter growth response. It is perhaps noteworthy that the R D G patterns are somewhat steeper than those of the R H G measure (Figures 7.3 and 7.4); this possibly suggests that growth efforts were concentrated radia l ly rather than vertically. The relat ively quick diameter growth response levels further support the argument that the trees were in a position to benefit in whole or in part from the peak 210 of the assart flush. There appears to be little published data on post-harvest diameter growth of subalpine fir advance regeneration. Fo r the T 6 site, mean absolute diameter growth ranged between 0.10 and 0.39 cm y r " 1 ; that of the T 8 fell between 0.07 and 0.50 cm y r" 1 ; for the T i l site the range was 0.07 to 0.35 cm y r " 1 . D a t a from Johnstone (1978) for five-centimetre (Dbh inside bark) subalpine fir logging residuals suggested post-harvest diameter growth rates ranging from 0.08 to 0.25 cm y r" 1 . The ranges observed in this study compare favourably with, these values. They also are very close to the ranges observed by Seidel (1985) for other released western true firs. Thus , it appears that diameter growth generally suffered none of the constraints seemingly operating on release height growth. The reasons for this apparent ly conflicting result are not immediate ly obvious; they m a y be related to possible differences in the physiological cost to the trees of generating p r ima ry as opposed to secondary growth. 7.4.3 Tree Age Considerations The mean ages at release and corresponding 95% confidence l imits for the advance regeneration on the cutover sites are presented in Table 7.2. Their values are generally s imilar . B y comparison, the subalpine fir advance growth in the mature stand had a mean age in 1984 of 92 years (n = 30; upper 95% l imi t = 103 y r , lower 95% l imi t = 81 yr) . The T M m a x i m u m and min imum values observed were 148 and 32 y r respectively in 1984. The ages of mature 211 Table 7.2 M e a n release ages for advance regeneration on the cutovers. S I T E n t A G E ? 95% L I M I T S (yr) Upper Lower T 3 28 35 49 21 T 6 100 31 35 27 T 8 90 42 48 36 T i l 70 48 56 40 tDenotes sample size. t V a l u e s are rounded means. 212 stand advance regeneration when the other sites were harvested can be estimated as 88, 85, 83, and 80 y r for the T 3 , T 6 , T8 , and T i l sites respectively. The corresponding cutover values appear to have been definitely lower than these. This implies that the older stems on the harvested sites were somehow eliminated. This could have been an unintentional result of the harvest ing process, or otherwise effected by the post-harvest slashing treatments applied (see Chapter Two). He r r ing (1977) noted that the larger advance growth were very generally speaking the older representatives; moreover, such larger trees are more susceptible to logging damage than smal ler ones. The ages at release observed in this study are well wi th in the general range reported by Her r ing (1977) for subalpine fir; they are ve ry close to those observed under s imilar conditions at nearby Sock Lake (Herr ing, 1977). W i t h respect to the possible effects of age on growth response, there was a very significant but weak negative correlation between age at release and total five-year height growth (p < 0 .01, r = -0.36, df = 231). No significant correlation was observed . w i t h the corresponding diameter growth variable (p > 0.05). It is fully acknowledged that a significant correlation between variables does not necessarily point to a causal relationship. Nevertheless, this result might indeed be reflective of a negative influence on post-harvest height growth which was related (directly or indirectly) to age values. This idea runs counter to currently accepted concepts as discussed earlier; however, it is extremely interesting to note that such an association appeared very significant for the seemingly constrained height growth, but non-significant for the seemingly unconstrained diameter growth. This supports the view that age might have 213 played a direct or indirect role in influencing release height growth. Whi le Monchak (1982) remained convinced that there were no direct age effects on response growth, he conceded that growth performance was usual ly lower where the average (release) age of the advance regeneration at breast height was greater than 80 yr . A s s u m i n g a time to grow to breast height of 30 y r (Achuff and L a Roi , 1977; Har r ing ton and M u r r a y , 1982; Bergstrom, 1983) and that to stump height as ten years (Watts, 1983), the equivalent "threshold age" at stump height would be 60 yr . The mean ages at release for the trees in this study were al l well below this value. The foregoing considerations tempt an affirmation that age played a role (direct or indirect) in constraining release height response. Such a conclusion can at best be accepted only tentatively; nevertheless, the results supporting it at least indicate that further investigations of this aspect might be desirable. 7.5 CONCLUSIONS Some general statements can now be made wi thin the context of stated objectives of this phase of the study. The details of patterns and values pertaining to pre- and post-harvest height and diameter growth were presented. Absolute height growth response was four to five years, but the relative measures for both height and diameter growth indicated an earlier response to release which would have coincided wi th the assart peak noted earl ier . It m a y be that in situations such as these there is less of a lag between diameter and height growth responses than previously indicated by the l i terature. The application of relative measures appeared to improve interpretations of both height 214 and diameter data over the absolute measures. For diameter growth, the relative measure indicated consistently that response to the changed conditions occurred wi th in one year of harvest ing. W i t h respect to the achievement of the 15 cm y r " 1 level, there was a five- to ten-year lag associated wi th this. Th is could represent a serious problem from the viewpoint of the forest manager. It was concluded that the regeneration was at least potentially in a position to benefit from the assart effect. There is thus par t ia l (but not. conclusive) support for a rejection of ma in hypothesis H 0 . This wi l l be referred to later in the course of a more concrete examination of the val idi ty of this hypothesis (Chapter Eight) . However , release height growth appeared to be constrained by influences which seemed not to l imit diameter growth response. There was evidence suggesting that the age of the advance regeneration might have played a role here, even at the relatively low release ages observed. CHAPTER 8 POST-HARVEST FOLIAR CHEMISTRY The results discussed i n the previous Chapter suggested that there m a y have been some post-harvest l imitations on growth, and gave pre l iminary consideration to one such possibility. In this Chapter , the post-harvest foliar chemistry of the advance regeneration is considered—both in terms of possible consequences of the assart patterns noted in the soil data, and also i n l ight of the possibility of nutrient l imitations to growth. Because of its established importance to tree growth, the m a i n emphasis was placed on patterns of foliar N . 8.1 INTRODUCTION M a n y of the reviews noted earlier dealing wi th N avai labi l i ty in forest soils (see Chapter Six) logically included some consideration of foliar analysis as a method of assessing nutrient avai labi l i ty to forest trees. There is a substantial body of literature on this subject. Noteworthy efforts not previously mentioned include those of L e a f (1973), Mor r i son (1974a), v a n den Driessche (1974), Lamber t (1984), T i m m e r and M o r r o w (1984), and B a l l a r d and Car ter (1986). Stone (1968) also considered microelement nutri t ion of forest trees in some detail. Conceptually, two general purposes of foliar analysis m a y be stated; these are the diagnosis of nutrient deficiency and the prediction of tree response to some stimulus (Morrison, 1974a; v a n den Driessche, 1974; Ba l l a rd and Carter , 1986). Other analytical approaches do exist; however, the consensus appears to be that 215 216 foliar analysis is by far the best currently available diagnostic tool for assessing tree nutrient status (Morr ison, 1974a; van den Driessche, 1974; Powers , 1984a; Ba l l a rd and Car ter , 1986). Moreover, when coupled wi th fertilizer treatments, its predictive value has been clearly established (Morrison, 1974a; T i m m e r and Mor row, 1984). Studies involving foliar analysis have usually examined nutrient concentrations; nutr ient contents and unit needle weights have also been included in many instances. The concept of cri t ical levels has become the one most widely used in data interpretations (Morrison, 1974a; Powers, 1984a; Ba l l a rd and Carter , 1986). However , ratios of different nutrients have also been employed, in part because of possible antagonisms between them within plants (van den Driessche, 1974; B a l l a r d and Car ter , 1986). St i l l other researchers have preferred the use of opt imum and/or balanced nutrient levels (Morrison, 1974a; Powers , 1984a; Ba l l a rd and Car te r , 1986). Severa l ve ry interesting methods have been developed which combine the above to v a r y i n g degrees. D u r i n g the 1960s, H . Krauss , and later D . Heinsdorf, presented a graphical diagnostic technique for examining fertilizer responses (Morrison, 1974a; T i m m e r and Stone, 1978). This was later modified by V . T i m m e r and co-workers to give a rapid method of evaluating nutrient status and predicting fertilizer response (Timmer and Stone, 1978; T i m m e r and Mor row, 1984). In B . C . , this method has been coupled with a mini-plot technique to yield a rapid, low-cost method of screening forest stands for their nutr ient status and potential for response to fertilizer treatments (Weetman and Fournier , 1982 and 217 1984). T. Ba l l a rd and co-workers have developed a comprehensive system for diagnosis, interpretation, and prescription of fertilizer treatments for major western species. The sys tem is based on cri t ical and op t imum ratios, and can accommodate both macro- and micronutrient evaluations (Bal lard, 1984; Ba l l a rd and Carter , 1986). A very interesting diagnostic technique was developed by P . Zinke and co-workers; it entails the comparison of data against a non-linear (Weibull) fit of cumulat ive probabilities developed from benchmark data (Zinke and Stangenberger, 1979; Zinke, 1986). F ina l ly , perhaps the most comprehensive system yet to be applied to such forestry problems was developed in agricultural circles—the Diagnosis and Recommendation Integrated Sys tem (DRIS) of Beaufils (1973). The system uses nutrient ratios, and has as its objective the calculation of directly comparable indices for the factors affecting yield; non-nutrit ional factors may be included. There have been few forestry applications; T r u m a n and Lamber t (1980), Leech and K i m (1981), H o c k m a n et al. (1985), and Zasoski and Porada (1986) are examples. According to Powers (1984a), D R I S presents a solution to most foliar analysis problems in forestry. Its m a i n difficulty—and a serious one—was perceived to be in the establishment of norms; this cr i t icism may also apply to the probability method above, but perhaps to a lesser extent. In this study, cr i t ical levels and ratios were the pr incipal variables used. In addition, the screening technique was applied on selected cutover sites. 8.2 OBJECTIVES AND HYPOTHESES The objective of this phase of the study was to ascertain the post-harvest nutrient status of the advance regeneration. The m a i n aspects of interest were 218 the post-harvest patterns of change, and to what extent these paralleled the assart patterns noted earlier; also, whether any nutrient l imitations to post-harvest growth were apparent. This fell wi th in the ambit of tests of main hypotheses H 0 , H 0 f . , and H 0 . Limitat ions include the effects of toxic levels as wel l as those of deficiencies. If H 0 ^ is true, there should have been little change in the foliar levels of nutrients the supply of which depended main ly on mineral izat ion processes in the soil. Also , for such elements, neither should any paral le l ism have been evident between post-harvest trends of change in foliar levels and those noted in the soil. A valid H 0 implies that for the macro-o nutrients no levels or ratios indicating either deficient or toxic conditions should have been evident in the foliar macronutrient values. A n argument s imilar to the latter applies to H 0 and micronutrients. The question of which specific nutrients could be taken as diagnostic arises. Fo r m a i n hypothesis H 0 , N levels were chosen; those of N , P , K , C a , M g , and S were used i n testing H 0 g—including a screening approach involving N only. F ive micronutrients were considered appropriate; these were a lumin ium (Al), boron (B), copper (Cu), iron (Fe), and manganese (Mn). The general essentiality of A l beyond certain plant groups has not been confirmed; while it is a common constituent of a l l plants, its physiological functions have not been clearly identified (Kabata-Pendias and Pendias, 1984). However, A l toxicity effects have been frequently noted for plants growing on acid soil—including forest trees (Pritchett, 1979; Kabata-Pendias and Pendias, 1984). Moreover , there are interactions between A l and some macronutrients which influence uptake of the latter; excess A l can reduce cation uptake, and also hamper metabolism of P (Kabata-Pendias 219 and Pendias, 1984; Ba l l a rd and Carter , 1986). B , C u , and Fe were included on the basis of recent observations in interior B . C . by T. B a l l a r d . t It was noted that low B levels m a y have been among the m a i n problems hampering old-growth spruce-stands. L o w C u and Fe levels were observed i n spruce plantations, especially after slashburning. F ina l ly , the inclusion of M n was from the viewpoint of toxicity. The idea of M n toxicity appears to have stemmed from research with true firs in the Vosges region of France; it was associated wi th seedlings growing on acid mulls (Rousseau, 1960; Val lee , 1967). The concept has fallen into disfavour, since many forest tree species exhibit very high M n levels without any apparent i l l effects (Stone, 1968; T. B a l l a r d , pers. comm. M a r c h 3, 1986). More recent work wi th true firs in the Vosges region has also tended to refute this idea (Drapier , 1983). Nevertheless, Powers (1979) believed the concept to be val id for some western true firs at high elevations. H e reported that in the absence of other nutrient l imitations, poor fir growth can often be traced to inhibit ing M n effects; the latter occurred at foliar concentrations above 300 mg kg" 1 . However , M n levels of up to 1000 m g k g " 1 have been observed in foliage of subalpine fir i n B . C . without any apparent adverse effect on growth (T. Ba l l a rd , pers. comm. M a r c h 3, 1986). The question of M n toxicity may be unresolved; nevertheless, in the absence of other apparent nutrient limitations, very high foliar levels (e.g. 2000 m g k g " 1 or greater) in slow-growing trees might war ran t consideration in this respect. W i t h i n the context of the above considerations, the subsidiary hypotheses under test and their alternates can now be stated: t Professor, Facu l ty of Fores t ry and Dept. of Soi l Science, U . B . C . ; pers. comm. M a r c h 3, 1986. H 0 : No major changes occurred i n the foliar N levels of advance regeneration dur ing the post-harvest period. H , y There were indeed major changes i n the foliar N levels of advance regeneration dur ing the post-harvest period. H 0 : There was no marked similar i t ies in pattern between post-harvest foliar N levels of advance regeneration and that previously described for post-harvest N availabil i ty. Hi 2: There were indeed marked similar i t ies in pattern between post-harvest foliar N levels of advance regeneration and that for post-harvest N avai labi l i ty . H 0 : N o detectable N or other macronutrient deficiencies occurred in the advance regeneration after harvesting. g: Detectable macronutrient deficiencies occurred in the advance regeneration after harvest ing. H 0 • Neither toxicities of A l or M n nor deficiencies in any of B , C u , and Fe occurred i n the advance regeneration after harvesting. 221 H , : Toxici ty effects of A l or M n and/or deficiencies in one or more of B , C u , and Fe occurred in the advance regeneration after harvest ing. There should have been no differences in foliar N levels among the sites i f H 0 ^ is true. Stat is t ical tests for such differences were accepted as the falsification cr i ter ia . The cri teria for judging H 0 consisted of a combination of visual examinat ion of plotted data and stat ist ical tests for differences among sites. If H 0 g and H 0 ^ are true, cr i t ical l eve l s . or ratios indicative of deficiencies (or toxicities as the case may be) should not have been evident in the foliar macro-and micronutrient data from the sites. Such comparisons thus formed the cri teria for falsification; as before, statist ical tests identifying differences were included where necessary. 8.3 METHODS The basic field layout, sampl ing scheme, and system of creating composite samples were presented in Chapter T w o . Fur the r details specific to this phase of the study are given below. F ie ld sampl ing followed the guidelines of Ba l la rd and Car te r (1986) in principle. Al though they are considered together below, analysis for micronutrients and S was done after (and on the basis of) the main macro-nutrient analyses. 222 8.3.1 Main Foliage Analysis Sampl ing was done in the fall of 1983 on al l sites. Ten twigs were taken from the upper third of the live crown of each tree; the uppermost branches were excluded. Each twig consisted of the current terminal and la teral shoots on that branch; these were placed in plastic bags. The twigs were placed i n cold storage (1 .7°C) until d ry ing could be done. In the laboratory, samples were dried at 6 0 ° C for 24 hours; needles were then carefully removed from the twigs and stored temporarily in paper bags at room temperature. Random grouping to provide second-stage composites was done at this point. Es t imates of needle weights were obtained from 1000 needles (five lots of 200 each) from each second-stage composite; The numbers used were based on application of the H u s c h et al. (1972) formula (see Chapter Two) to needle weights of samples taken during the prel iminary survey. Samples were re-dried (six hours at 6 0 ° C ) before counting, and afterwards placed in paper envelopes into cold storage as before. Because the amount of foliage per sample was fairly low, the samples were not ground before analysis; this is unl ike ly to have affected substantial ly the results obtained (Salonius et al, 1978). Tota l N and P concentrations were determined as described in Chapter F ive for soil samples; in this instance, duplicate analyses were performed on al l samples. Concentrations of K , C a , and M g were obtained by atomic absorption spectrometry after dry ashing and HC1 digestion (Allen et al, 1974; Baker and Suhr , 1982; Knudsen et al, 1982; L a n y o n and Heald , 1982). Analyses for micronutrients and S were carr ied out on a subset of the 223 samples, after results had been obtained from the above analyses. Sub-sampling was restricted to those foliage samples which exhibited needle weights of five or less g (1000 needles)" 1 and N concentrations of 1.35% or greater. Thus , micro-nutrient and S effects were examined for "worst case" development which was not apparently l imited by N . The subset was comprised of four randomly chosen sub-samples from each site except the T 8 . In the latter case, no samples were found which fulfilled the stated cri ter ia . Tota l B was analyzed by dry ashing and colourimetric determinations, pr incipal ly following the azomethine-H method as applied in the Pedology laboratory at U . B . C . (Anonymous, 1984). Because of the smal l amounts of mater ia l available, total A l , C u , Fe , and M n concentrations were determined employing a sulphuric acid-hydrogen peroxide digest procedure followed by atomic absorption spectrometry (A. Gammel l , t in litt., June 16, 1986). Fo r S, samples were ground in a Wi l ey m i l l , re-dried (three hours at 6 0 ° C ) , analyzed for total S using a Fisher Model 475 analyzer (Guthrie and Lowe , 1984). D a t a were processed and analyzed using the M I D A S , A N O V A R , and G E N L I N programmes as outlined earlier. Fo r the macronutrients except S, nutrient contents per 1000 needles were determined using the needle weight and concentration data; for N and P , the average of the duplicate measurements was applied here. No content calculations were done for the micronutrients or S. The cr i t ical levels and ratios used for interpretation were those given by Monchak (1982) for needle weights and N levels of subalpine fir regeneration, as wel l as those of Ba l l a rd and Car ter (1986). Total N concentration and content data were t Labora tory Supervisor, Woodlands Services Div i s ion , M a c M i l l a n Bloedel, L t d . 224 analyzed as a two-level nested A N O V A wi th unequal sample sizes (Sokal and Rohlf, 1981); the individual sites formed the first level, while the samples formed the second. Tukey 's test (Sokal and Rohlf, 1981; Dowdy and Wearden, 1983) was employed for the a posteriori comparisons. The following were separately analyzed as single-classification Model I A N O V A problems, wi th the sites as treatments: Needle weights, K concentrations and contents, C a concentrations and contents, S concentrations, N / S and K / C a ratios, and al l micronutrient concentrations. A n arcsine transformation was necessary for the C a concentration data, while logarithmic transformations were applied to the A l , B , N / S , and K / C a data. In the A l and B cases, a very slight heterogeneity of variances st i l l existed in the transformed data. Nevertheless, since no significant differences were detected by the A N O V A (see Results section), the analyses were accepted in spite of this condition (Lindeman, 1974). No transformations were found which allowed concentration and content data for P and M g to be va l id ly analyzed by an A N O V A ; the same was true for the C a / M g ratio data. These were therefore analyzed using the non-parametric median and Kruska l -Wal l i s tests in conjunction wi th Tukey jackknife 95% confidence intervals as described earlier. 8.3.2 Screening Trial Approach Dur ing the 1982 pre l iminary survey, urea was applied to the freshly cut, five-, and ten-year-old cutovers of the in i t ia l sequence investigated (see Chapter Two). The latter two sites were wha t became the T 6 and T i l sites respectively of the final sequence. Twelve 0.01 ha circular plots were located along an open traverse on each site. The distances between plots were variable; plot centres 225 were chosen in such a way as to cover as much of the "representative" portion of the cutover as possible. O n each site, four replicates of each of three treatments were randomly allocated among the 12 plots. The treatments corresponded to 0, 150, and 300 kg N ha" 1 , designated as NO, N I , and N 2 respectively. A pair of sample trees was selected from the dominants (excluding residuals) as the plot-centre and off-centre trees (Weetman and Fournier , 1982). The off-centre tree was the largest dominant nearest to the centre tree. E a c h centre tree was marked wi th flagging and an a luminium identification tag. The fertilizer was applied throughout each plot using a cyclone seeder. Unfor tunate ly , only 26 of the original 36 plots could be re-located in 1983. A s noted earlier, in the T i l case the fertilized plots were in an entirely separate area to that on which other sampl ing activities took place. In the T 6 case, care was taken to ensure that identifiable fertilized areas of the cutover were excluded from any other sampling activities. Owing to the number of plots lost, neither the forest floor weight change nor the N mineral izat ion studies were attempted (see Chapter One). In the fall of 1983, samples of current growth were taken from the two selected trees on each remaining plot us ing the same scheme as discussed in the previous Section. The samples were dried at 6 0 ° C for 24 hours; needles were then carefully removed from the twigs, counted, and weighed as described previously. In this case, the samples were ground in a Wi ley m i l l for analysis. Concentrations of N , P , K , C a , and M g were determined using the procedures outlined in the previous Section; nutrient contents per 1000 needles were also calculated i n the same manner. Graphica l representations of the data were constructed and interpreted in 226 terms of observable shifts (Timmer and Stone, 1978; Wee tman and Fournier , 1982). The needle weight, nutrient concentration and content data were analyzed separately for each site as a single-classification, Model I A N O V A . In addition, the needle weight and N data from a l l sites were pooled and analyzed as a two-way, Model I A N O V A , wi th the individual sites and N applications as the treatments. A s before, data processing and analysis was done us ing the M I D A S and G E N L I N programmes. 8.4 RESULTS AND DISCUSSION 8.4.1 Main Foliage Analysis The post-harvest trends of needle weight, N concentration, and N content values across the age sequence are presented in F igure 8.1. F igure 8.2 il lustrates those for concentrations and contents of the other macronutrients (S excepted), while N / P , K / C a , and C a / M g ratio trends after harvest ing are presented in Figure 8.3. 8.4.1.1 Needle weights and Nitrogen There were highly significant differences (p < 0.001) among the mean needle weights, mean N concentrations, and mean N contents of the sites (for al l : df = 4, 130; F = 20.4, 24.2, and 22.3 for needle weights, N concentration, and N content respectively). The cutover mean needle weights were all significantly greater {p < 0.01) than that of the mature stand; in addition, 227 Vertical bars are 95% confidence limits 3 0 i 1 1 1 1 1 1 1 1 1 1 1 0 1 2 3 4 5 6 7 8 9 10 11 YEARS AFTER HARVEST Figure 8.1 Means and 95% confidence limits of the foliar needle weights, N concentrations, and N contents. 228 Vertical bars are 95% confidence limits CONCENTRATION (%) CONTENT (mg/1000 needles) C L D O cn 0.27-i 0.24 0.21-0.18-1.3 —| 1.2-1.1 1-0.9-0 .8-0.50 0.45 H 0.40 0.35-0.30-0.25-0.18-j 0.16-0.14-0.12-0.10-0.08-18-, n—I—I—l—I—I—I—l—l—I—I 1 2 3 4 5 6 7 8 9 10 11 6 70 6 0 -5 0 -40-30-25-20-15-10 10 8-6 4 2- i—I—I—I—l—l— I—\—l—I—l—l 0 1 2 3 4 5 6 7 8 9 10 11 YEARS AFTER HARVEST Figure 8.2 Means and 95% confidence limits of the foliar P, K, Ca, and Mg concentrations and contents. 229 1 J | 1 1 1 1 1 1 1 1 1 1 I 0 1 2 3 4 5 6 7 8 9 10 11 Y E A R S AFTER H A R V E S T Figure 8.3 Means and 95% confidence limits of the foliar N/P, K/Ca, and Ca/Mg ratios. 230 the T 6 mean was significantly greater than that of the T 3 . No other significant differences among the needle weight means were detected. The mean needle weight in the mature stand was 3.89 g (1000 needles) " 1 ; those of the cutovers were between 5.07 (T3) and 5.98 (T6) g (1000 needles)" 1 . For N concentrations, the T 3 and T i l means were significantly (p < 0.05) greater than those of the other sites; wi th in these two sub-groups no further differences among the means were apparent. The T3 and T i l mean N concentrations were 1.64% and 1.56% respectively; the other means ranged from 1.21% (TM) to 1.32% (T6). F ina l ly , the mean N contents of the cutovers were a l l significantly (p < 0.05) greater than that of the mature stand; moreover, the T 8 mean was significantly lower than those of the T 3 and T i l sites. No other differences were apparent. The T3 and T i l means were 84.0 and 86.9 m g (1000 needles)" 1 respectively; the others fell between 47.1 (TM) and 78.8 (T6) m g (1000 need le s ) " 1 . The mean needle weights, N concentrations, and N contents observed in this study are well wi th in the ranges reported or implied by Beaton et al. (1965) and Monchak (1982) for young or released subalpine fir. The concentrations are slightly above those observed in young balsam fir (Morr ison, 1974b). The N contents are well wi th in the range of those from Monchak (1982); however, they are much higher than those reported by Husted (1982) for "wel l g rown" Pacific silver fir trees after logging. M a r t i n (1985) also noted that foliar N N concentrations (and contents) were highest in the early portion of his post-harvest age sequence. The post-harvest needle weight trends (Figure 8.1) suggest that the advance regeneration increased their individual needle weights by approximately 54% in the first six years after harvest ing. Sizes remained constant thereafter 231 imp ly ing that physiological adjustment had been completed. However , even the m a x i m u m size attained is well below the 7.5 m g (1000 needles)" 1 given by Monchak (1982) as the level below which N m a y be l imi t ing growth of released subalpine fir. The N concentration and content trends (Figure 8.1) present conflicting pictures. Both indicate that changes in foliar N levels did occur in the post-harvest period. Moreover, the three-year-mark appeared to represent at least one peak of such change; values declined thereafter up to year eight. The trend of mean N contents suggests that N uptake generally increased to a marked degree following harvesting. The two variables differ i n their indications of N deficiency according to Monchak 's (1982) standards. The concentration data indicate that the advance regeneration was not N-deficient on any of the sites—not even in the understory of the mature stand. B y contrast, the N content trends suggest that the advance regeneration m a y have experienced N deficiencies between six and eight years (and possibly beyond) after harvesting. If this is true, it would mean that the trees were N- l imi ted immediately after the height growth response delay period (Chapter Seven). W h i c h variable presents the closest representation of the "true" post-harvest picture? O f the two, concentration data has tradit ionally found more widespread use. However , T immer and Stone (1978) noted that the implicit assumption on which this is based (i.e. that concentration data by itself is an adequate reflection of nutrient status) has not been tested to any substantial degree. Other factors complicate such interpretations. Fo r example, the decline in N concentration wi th a simultaneous increase in needle weight between years three and six (Figure 8.1) could be taken as a manifestation of the "Steenbjerg effect" or growth dilution (Tamm, 1964; Jones and Eck , 1973; Morr ison, 1974a). Thus , it m a y be that in this 232 case nutr ient content (an expression of both needle weight and concentration) may-be the more legitimate indicator of nutrient status. There is yet another question which affects the above arguments: To what extent are the standards given by Monchak (1982) val id indicators of N status of subalpine fir advance regeneration? B y his cr i ter ia , released subalpine fir advance growth exhibit ing N concentrations of 1.2% (or less) and/or needle weights of 7.5 g (1000 needles)" 1 (or less) m a y be considered N-deficient; this corresponds to a foliar N content of 90 m g (1000 needles)" 1 . Examina t ion of Monchak ' s (1982) data indicated that the latter needle weight and concentration levels were equalled or exceeded by approximately 40% and 50% respectively of his data points. The concentration level can be assailed as being too low. Other than Monchak ' s (1982) study, there are apparently no other published standards for subalpine fir trees—definitely none for released advance growth. However, subalpine fir is reputed to have higher nutrient levels (foliar and other components) as wel l as to be more nutrient-demanding than its companion spruces (K immins , 1974; K i m m i n s and H a w k e s , 1978). D a t a from a study by Beaton et al. (1965) support this view. Under this assumption, comparisons can be made against the standards noted for white (or Engelmann) spruce; in the latter, N deficiency m a y be expected at foliar levels below 1.4 to 1.5% (Morr ison, 1974a; Ba l l a rd and Car ter , 1986). This is also approximately the same level as might be applied to western hemlock [Tsuga heterophylla (Raf.) Sarg.] (Bal lard and Car ter , 1986), another shade-tolerant species which produces advance regeneration as part of its reproductive strategy. Therefore, to the extent that concentration is a val id indicator wi th in this context, it m a y be more logical to 233 increase the assumed critical level. The range 1.35% to 1.40% might be a point of departure; 25% of Monchak's (1982) points were above the 1.35% value. Regarding the implici t N content standard, other studies have reported levels in current foliage of unfertilized true firs ranging from 13.9 to 42.0 m g (1000 needles)" 1 ; 75 mg (1000 needles)" ' was that observed in fertilized trees (Husted, 1982). Husted (1982) reported a value of 26.0 mg (1000 needles)" 1 in current foliage of "wel l grown" Pacific si lver fir trees after logging. M a r t i n (1985) observed values ranging from 34.8 to 92.2 m g (1000 needles)" 1 in the same species after logging. While Pacific si lver fir generally appears to exhibit lower N levels than subalpine fir (Beaton et al., 1965; Husted, 1982; Monchak, 1982), the content standard implicit in Monchak (1982) might be excessive. W i t h the increased concentration standard, this implies that a lowering of the needle weight standard may be desirable. I f the level observed in fertilized Pacific si lver fir is used wi th a 1.4% N crit ical level, the needle weight standard becomes approximately 5.4 g (1000 needles)" 1 . It is extremely interesting to note that this is very close to the value at which the needle weights of this study appeared to have stabilized. Appl icat ion of the above "new" standards to the results given earlier would change the interpretation obtained substantial ly; an element of conflict would s t i l l exist. The N concentrations would then indicate that N was non-l imi t ing at the three- and eleven-year marks only. The needle weights would indicate no such l imitat ion from the six-year point onwards; further, the value at year three could also, be taken as being more reflective of adaptation than nutrient l imitat ion. Assuming the content data to be more reliable (as before), the 234 indication would be that N deficiencies existed around year eight only of the sequence. These "new" results would imply that the advance regeneration wi th in the mature stand was N deficient—an entirely believable notion. The boost in values at year three to wi th in the range of N adequacy would then accord very wel l wi th the soil N O 3 - N peaks and their implications as discussed earlier (Chapter Six) . Once the adaptation (response delay) period was over, a "Steenbjerg effect" became evident going from year three to year six (and possibly also year eight). It is not certain whether this could be interpreted as a "true" deficiency or l imitat ion to growth; the "new" content standards would suggest that this was indeed true for the eight-year point. The screening t r ia l data should shed more light on this. N-deficient trees at year eight would be a tremendously interesting result, since this would imply that any assart flush for N was t ru ly over by that point; such an idea would be in basic agreement wi th the results of Chapter S ix . W h a t would then remain unclear would be the subsequent rise to levels of N adequacy by year 11; one possible explanation would be an increase in the efficiency of internal cycling of N by the advance growth. A t best, these conjectures support the concept of an assart effect last ing less than eight years; beyond that point trees became temporarily N deficient, but this condition had disappeared by year 11. A t the very least, the above considerations demonstrate that further work is necessary to facilitate accurate evaluations of the post-harvest N status of subalpine fir advance growth. 235 8.4.1.2 Phosphorus, Potassium, Calcium, and Magnesium High ly significant differences (p < 0.001) were apparent among the site groups for P , K , C a , and M g concentrations (for K : F = 44.0; for C a : F = 10.1; df = 4, 130 for both). The jackknife 95% limits indicated that the mean P concentration of the T i l site was significantly lower than a l l except that of the mature stand; no other significant differences were apparent. The lack of further significant differences m a y have been due to the greater var iab i l i ty of the T M samples; it is noteworthy that the trend (Figure 8.2) shows a s t r ik ing paral lel to that of the resin measure of "available" P in the soil (Figure 5.3). The mean P concentration of the T i l site was approximately 0.22%; the others ranged between 0.22 and 0.24% approximately. A s with the soil data, it is questionable whether such smal l changes were of practical significance to growth. F o r K , the mean foliar concentrations of the cutovers were a l l significantly (p < 0.05) lower than that of the mature stand. In addition, the T 8 mean was significantly lower than those of the other cutovers; the T i l mean was greater than that of the T6 (in addition to the T 8 mean). The trend was thus one of lowered K concentrations to a m i n i m u m at year eight of the sequence, w i th a recovery thereafter. There is no apparent simple parallel between this trend and that noted in the soil K data. The mean foliar K concentration of the mature stand was 1.24%; the others fell between 0.94% (T8) and 1.09% ( T i l ) . The mean C a concentrations of the T 6 , T 8 , and T i l sites were significantly {p < 0.05) greater than those of the T M and T 3 sites, wi th no further differences wi th in these two groupings. The trend thus formed is in very close agreement wi th that obtained from the soil C a data earlier. The mean concentrations of the 236 T M and T3 sites were 0.33 and 0.31% respectively; the others ranged between 0.39 and 0.40%. For foliar M g concentrations, the jackknife 95% limits indicated that the T 8 mean was significantly lower than all others; in addition, the T i l mean was significantly greater than those of the T3 and T6 sites. Here , again, the relatively higher var iabi l i ty of samples from the mature stand m a y have masked further differences. The trend of mean M g concentrations does not paral lel those observed earlier in the soil fractions. The T8 site had a mean foliar M g concentration of 0.11%, while, the others ranged between 0.13 and 0.15%. The P , K , C a , and M g concentration ranges agree very wel l wi th those reported by Beaton et al. (1965) and Monchak (1982) for subalpine fir. The only exceptions to this statement were in the case of K concentrations compared to Monchak 's (1982) and M g concentrations compared to those of Beaton et al. (1965); in both cases the values of this study were slightly higher. Compared to concentratons for young ba lsam fir, this study's P and K values are higher, its C a values are lower, and its M g values in basic agreement wi th those of Mor r i son (1974b). Where nutrient status interpretations are concerned, the constraints to the use of concentrations only were discussed previously. Moreover , standards for these elements have not been developed specifically for subalpine fir. W i t h i n these l imits , the cr i t ical levels of Ba l l a rd and Carter (1986) for white spruce and western hemlock can be tentatively applied. F r o m this, it would appear that a l l of the elements P , K , C a , and M g can be said to have been in adequate supply. Confirmat ion (or otherwise) of this statement can be expected from the consideration of cr i t ical ratios. 237 Turn ing to nutrient content data, there were highly significant differences (p < 0.001) among the site groups for P , K , C a , and M g contents (for K : F = 7.5; for Ca : F = 27.1; for both: df = 4, 130). The jackknife 95% confidence l imits suggested that the mean P contents of the cutovers were a l l significantly higher than that of the mature stand. Moreover, the T 6 and T 8 means were significantly greater than the T3 mean; the T 6 mean (but not the T8) was also significantly greater than that of the T i l . Here, again, the somewhat parabolic trend is one which shows a very close paral lel to that of the "avai lable" P measure. It also suggests that P uptake levels generally increased after harvesting. The mean P content of the mature stand was approximately 8.7 m g (1000 needles)" 1 , w i th the others ranging from 11.9 to 14.7 mg (1000 needles)" 1 . For K , the mean foliar contents of the T 6 and T i l sites were significantly (p < 0.05) greater than those of the T M and T 3 ; neither of these two sub-groups differed significantly from the T 8 mean, nor were their constituent means significantly different from each other. The trend suggests that K uptake levels were generally elevated during the post-harvest period; this implies that the lowered concentrations earlier might have been a consequence of a growth dilution process which tended to disappear towards the end of the sequence. The mean K contents ranged from 47.5 mg (1000 needles)" 1 in the mature stand to 61.2 m g (1000 needles)" 1 on the T i l site. The C a content pattern was the same as that noted earlier for foliar C a concentrations. The means of the T6 , T 8 , and T i l sites were significantly (p < 0.05) greater than those of the T M and T 3 , wi th no other significant differences. The suggestion is one of elevated C a uptake levels in the post-harvest period—a result which reflects and supports the soil C a post-harvest patterns as discussed earlier 238 (Chapter Five) . M e a n foliar C a contents ranged from 12.9 mg (1000 needles)" 1 on the T M site to 23.7 m g (1000 needles)" 1 on the T6 . F i n a l l y , the jackknife 95% confidence l imits indicated that, w i th the exception of the T 8 site, mean foliar M g contents of the cutovers were significantly greater than that of the mature stand. In addition, the T 6 and T i l means were significantly greater than the T 8 ; the T i l mean was also greater than that of the T3 site. Here , again, the trend suggests that M g uptake levels were generally higher after harvest ing. A s wi th M g concentrations, this does not parallel the observed soil M g patterns. M e a n foliar M g contents fell between 5.3 mg (1000 needles)" 1 in the mature stand and 8.4 m g (1000 needles)" 1 on the T i l site. A s noted earl ier , there is apparently little published information wi th which the nutrient content data can be compared. W i t h the exception of C a , the content values of this study are very close to those of Monchak (1982) for released subalpine fir advance growth. The C a contents were lower than those of the latter study. The P content values of this study are higher than those observed in Pacific si lver fir by Husted (1982). However, the content levels noted for the other three elements are one to two orders of magnitude lower than those reported in the latter study. One common facet of interest in the nutrient content trends across the age sequence is that they all implied increased nutrient uptake by the trees dur ing the post-harvest period. For P and C a , there were also s t r ik ing parallels w i th the trends noted previously in the soil data. These findings (along w i t h a s imi lar one for N earlier) lend support to the argument (Chapter Seven) that the trees were able to benefit from the assart effect. 239 Some general statements can now be made concerning both the concentration and content data for the four elements under consideration. Both measures suggested that there had been changes i n nutrient levels following harvesting. Tree nutrient uptake apparently increased; however, growth dilution effects m a y have been operational—notably in the case of K . A need for standards by which the status of these elements could be judged was apparent. Tentatively, i t is unl ikely that deficiencies of any of these elements l imited post-harvest growth; the nutrient ratio data should shed further light on this. The magnitudes of changes were relatively low; it is debatable whether they were of any practical significance to tree growth. Nevertheless, the P and C a trends suggested a close para l le l i sm between levels of these nutrients i n foliage and levels of their "avai lable" forms in the soil across the age sequence. This supports the view that trees were able to benefit from the assart release described earlier. Since neither P nor C a seemed to be in short supply, it might be argued that no "true" or direct benefit accrued to the trees—despite greater uptake. Par t icu la r ly in the case of C a , the benefits were probably indirect, derived from such factors as improved conditions for microbial act ivi ty. Such indirect benefits would therefore form the basis of the suggested increase in site quali ty noted earlier (Chapter Five) . 8.4.1.3 NIP, KICa, and CalMg Ratios There were highly significant differences (p < 0.001) among the sites in their N / P , K / C a , and C a / M g ratios (for K / C a : F = 20.5; df = 4, 130). The jackknife 95% confidence l imits indicated that the mean N / P ratios of the T 3 and 240 T i l sites were significantly higher than those of the others, but not from one another. A l so , the T 8 mean was significantly lower than those of the T M and T 6 sites (Figure 8.3). The mean values ranged between 5.0 for the T 8 site and 7.3 for the T i l site. Since none of the values rose above 12.5, P appeared to be in adequate supply on a l l sites (Ballard and Carter , 1986). The mean K / C a ratios of the T 6 , T 8 , and T i l sites were significantly (p < 0.05) lower than those of the T M and T3 sites; no other significant differences were apparent. The mean ratios of the T M and T3 sites were 3.8 and 3.3 respectively; the others fell between 2.4 and 2.8 (Figure 8.3). None of the values fell below 0.5; there was thus no evidence of a K deficiency on any of the sites (Bal lard and Car ter , 1986). The T M mean K / C a ratio fell marginal ly wi th in a range which is suggestive of a possible Fe deficiency (Ballard and Carter , 1986). This result by itself is not believed to be important within the context of this study; the possibili ty of Fe deficiencies could be explored more directly using the micro-nutrient data. F i n a l l y , the jackknife 95% limits indicated that for C a / M g ratios, the T 6 and T 8 means were significantly greater than those of the T 3 and T i l ; no other significant differences were apparent. The relat ively high var iabi l i ty of the samples from the mature stand (Figure 8.3) m a y have served to mask further differences. The mean values fell between 2.4 (T3) and 3.5 (T8); thus, growth should not have been limited on any of the sites (Bal lard and Carter , 1986). The above results suggest that changes in nutrient ratios occurred in the post-harvest period. However , deficiencies of P , K , and C a did not seem to occur at any time wi th in the sequence. These results confirm the tentative conclusion 241 derived earlier using the concentration (and content) data: The elements P, K, Ca, and Mg can be said to have been in adequate supply. 8.4.2 Micronutrients and Sulphur The means and 95% confidence limits of the micronutrient concentrations, S concentrations, and N/S ratios of the foliage sub-samples are presented in Table 8.1. No significant differences were detected among the mean foliar concentrations of A l , B, Cu, and Mn (for all: p > 0.05; df = 3, 12; F = 2.6, 2.1, 1.2, and 1.3 for Al , B, Cu, and Mn respectively). There were significant differences among the mean Fe concentrations (p < 0.05, F = 4.7, df = 3, 12), very significant differences among the mean N/S ratios (p < 0.01, F = 6.3, df = 3, 12), and highly significant differences among the mean S concentrations (p < 0.001, F = 15.8, df = 3, 12). In the case of Fe, the T3 mean concentration was significantly (p < 0.05) greater than those of the T6 and T i l sites, but not significantly different from the mean of the mature stand. The T3 and T i l mean N/S ratios were significantly (p < 0.05) higher than those of the T6 site, but not different from one another; the T i l mean was also significantly greater than the mean of the mature stand. Finally, the mean S concentration of the T i l site was significantly lower (p < 0.05) than that of the other sites; no further differences were apparent. Interpretations of the above must be considered in light of the deliberate "worst case" sub-sample selections. Kabata-Pendias and Pendias (1984) quoted a wide range of A l concentrations in higher plants, from a usual value of the 242 Table 8.1 Means and 95% confidence limits of the micronutr ient concentrations, S concentrations, and N / S ratios in the foliage sub-samples. (See text for details.) E L E M E N T S I T E / V A L U E S T M T3 T 6 T 8 T i l (values rounded; figures i n brackets are 95% limits) A l 162a t? 201a 140a 223a (216,121) (268,151) (186,105) (297,167) B 18a 13a 18a 18a (23,14) (17,10) (23,14) (23,14) C u 6a 11a 5a 8a (12,1) (17,6) (10,0) (14,3) Fe 42ab 71a 38b 39b (58,26) (87,55) (54,21) (55,23) M n 728a 600a 615a 473a (928,528) (801,399) (816,414) (673,272) S, % 0.102a 0.107a 0.102a 0.081b (.108..096) (.114,.101) (.108,.096) (.088,.075) N / S § 14 . l ab 17.6bc 13.8a 18.0c (16.0,12.5) (19.9,15.6) (15.6,12.2) (20.3,15.9) tUn le s s otherwise stated, units are m g nutrient (kg foliage)" 1 . $ Values followed by the same letter are not significantly different (p > 0.05) from each other. §A ratio, hence dimensionless. 243 order of 200 m g kg" 1 to more than 1000 m g k g " 1 in A l accumulators. The mean A l concentrations of the sites were a l l around the lower end of this. Since the mean P concentrations of the m a i n set of foliage samples were a l l greater than (almost double) 0.13%, the idea of P uptake being hindered by A l interactions can be dispensed wi th (Bal lard and Car ter , 1986). The possibility that cation uptake might have been reduced by excess A l (Kabata-Pendias and Pendias, 1984) can also be ruled out in light of the results discussed earlier. It would thus seem safe to conclude that A l levels did not in any major way l imi t growth of the advance regeneration on the sites. The B concentrations are sl ightly below those noted for other field-growing true firs wi th no apparent B deficiencies (Stone, 1968). The T3 mean B concentration does fall slightly below the 15 m g kg" 1 cri t ical value, suggested by B a l l a r d and Carter (1986). However , in light of the "worst case" characteristic, it is not believed that B deficiencies were generally in evidence on the T3 site. Therefore, it may be said that B was not a principal l imi t ing factor to the growth of subalpine fir regeneration on the sites. A s imilar conclusion can be reached concerning possible l imitations imposed by C u and Fe . The mean C u and Fe concentrations in this study were comparable to those observed in field-growing, non-limited true firs and Picea spp. respectively; the Fe values were also comparable to those in western hemlock (Stone, 1968). The mean C u concentrations were a l l above B a l l a r d and Car ter ' s (1986) suggested cr i t ical level. Interestingly, the Fe concentrations of this study were approximately one order of magnitude lower than those in the released subalpine fir trees examined by Monchak (1982). While the T 6 and T i l means fell below the tentative criterion (50 mg kg" 1 ) of B a l l a r d and Car ter (1986), this is not believed to be indicative of a general Fe 244 deficiency on those sites. The rationale is the same as in the B case above; moreover, the mean K / C a ratios of the m a i n set of samples from these sites did not indicate such a condition (Bal lard and Car ter , 1986). The idea of a possible M n toxicity must also be rejected on the basis of the data. The mean concentrations noted fell at the lower end of the range reported by Monchak (1982) for released subalpine fir. However , they were below the concentration ranges observed in other field-grown, non-limited true firs (Stone, 1968). Whi le the mean values of this study are above the value noted by Powers (1979), they are nowhere near a level that might be considered a growth-limiting one. The final consideration here is that of S, and it is here that possibly the most interesting result of the sub-sample analysis occurred. The ranges of mean S concentrations and N / S ratios of this study are in close agreement wi th those observed by Monchak (1982) in released subalpine fir advance regeneration. The S concentrations fell below and the N / S ratios above the relevant ranges noted in foliage of young subalpine fir trees (Beaton et al, 1965). The concentrations were also below the mean value reported by Guthrie and Lowe (1984) for ba l sam fir foliage. W h a t is of extreme interest is that both the mean S concentration and the mean N / S ratio values suggest that S deficiencies existed i n the trees the foliage of which was sub-sampled. According to Ba l l a rd and Car te r (1986), total S concentrations of less than 0.14% and/or N / S ratios of greater than 13.6 suggest possible deficiencies of S; where the N / S ratio is greater than 14.6, S deficiencies can be assumed. Such a result agrees with that of Monchak (1982) for his sub-sample of released subalpine fir trees; the possibili ty that S deficiencies might be induced by N fertilization was inferred in old-growth subalpine fir stands in the Prince George Region of B . C . (T. Ba l l a rd , r 245 pers. comm., March 21, 1986). It would therefore be tempting indeed to conclude that S was a primary growth-limiting factor on the sites. However, this must be tempered by the "worst case" characteristic of the data. The most that can be said is that the possibility exists that the advance regeneration was S-deficient during the post-harvest period. The veracity of this needs to be tested in a more rigorous fashion, especially in light of such indications elsewhere. 8.4.3 Screening Trial Analyses N was the only nutrient element applied in the fertilizer treatments. Also, the foregoing results suggested that there were no P, K, Ca, or Mg deficiencies apparent across the sequence. For these reasons, only the results pertaining to N values are considered here. 8.4.3.1 Directional Relationships Graphical representations of the effects of the fertilizer treatments on needle weights, N concentrations, and N contents of foliage on the three sites are presented in Figure 8.4. With the exception of N concentrations on the 1972 site, no significant differences were detected among the treatment means for any of the three measurement variables [for all: p > 0.05; df = (2,4), (2,7), and (2,6) for the 1982, 1977, and 1972 sites respectively]. In the 1972 N concentration case, very significant differences among the treatment means were apparent (p < 0.01; F = 12.4; df = 2, 6). The N I and N2 mean concentrations were significantly (p < 0.05) greater than that of the control, but 246 Figure 8.4 Directional relationships among needle weights, N concentrations, and N contents. 247 not significantly different from each other. The mean of the control was 1.39%, while those of the N I and N 2 treatments were 1.82% and 2.06% respectively. The confidence interval around the control value ( ± 0 . 2 4 % ) suggested that i t was not significantly different from the mean N concentration observed in the m a i n analyses for the 1972 ( T i l ) site. G i v e n the lack of significant differences, it must be stated that there was no evidence of N deficiencies on any of the sites. However , it is believed that this could have been due to the smal l number of replicates recovered and/or high within-site variabi l i ty; the wide 95% in terva l of the 1972 control N concentration given above illustrates this point. In light of this, it was sti l l considered useful—and perhaps instructive—to discuss directional shifts in detail. The terminology used and interpretations given follow pr incipal ly those of T immer and Stone (1978) and T i m m e r and Mor row (1984). O n the freshly-cut 1982 site, both the N I and N 2 (but par t icular ly the N I ) treatments tended towards an " E " shift. This is believed to be more an indication of non-nutritional growth constraints than one of toxic N levels (Timmer and Stone, 1978). W i t h i n the l imited context of these interpretations, it m a y be that this is a reflection of the physiological inabil i ty of the recently released trees to respond to changed nutrient and other conditions. Such a directional trend was also very weakly observed for both treatments on the 1972 cutover. On the 1977 site, the N I treatment tended towards a " C " shift—an indication of N deficiency. W i t h the N 2 , a " D " shift ( luxury consumption) seemed to be in evidence. These results could be construed as hinting at the existence of slightly N-l imi ted conditions around the six-year mark of the sequence. Th is would accord well wi th the idea of a temporary N deficiency around this point, 248 as discussed earlier. 8.4.3.2 Overall Analyses The fertilizer treatment means and 95% confidence limits of the needle weights, foliar N concentrations, and foliar N contents from the overall analysis are presented in Table 8.2. There were no significant differences (p > 0.05) among the mean needle weights of the three treatment levels. Highly significant differences were observed among the mean N concentrations (p < 0.001; F = 11.6; df = 2, 17). The N I and N2 means were significantly (p < 0.05) greater than that of the control, but not different from one another. Similarly, significant differences were apparent among the mean N contents (p < 0.05; F = 5.2; df = 2, 17). The N2 mean was significantly greater than that of the control; neither the control nor the N2 means were significantly different from the N I mean. Taken together, these results constitute a "D" shift (luxury consumption) for both N I and N2, thus reinforcing the idea that N was not the principal limiting factor on these sites. However, inspection of the confidence limits suggests that high within-treatment variabilities may have served to prevent detection of "true" differences—especially for needle weights. 8.5 CONCLUSIONS From the results discussed in the foregoing, it was concluded that subsidiary hypothesis H 0 should be rejected. Clearly, there were major changes in the foliar N levels of advance regeneration during the post-harvest period. 249 Table 8.2 Means and 95% confidence l imits of the needle weights, foliar N concentrations, and foliar N contents for the grouped fertilizer plots. T R E A T M E N T N E E D L E W E I G H T (g/1000) N C O N C ' N (%) N C O N T E N T (mg/1000) (Values rounded; figures in brackets are 95% limits) NO: control 5.5at 1.50a 81a (6.2,4.7) (1.68,1.32) (97,65) N I : 150 kg N / h a 5.6a 1.84b 10 l a b (6.4,4.7) (2.03,1.65) (118,84) N 2 : 300 kg N / h a 5.6a 2.05b 114b (6.3,4.9) (2.20,1.89) (127,100) tColumn values followed by the same letter do not differ significantly from each other (p > 0.05). 250 Moreover , while the interpretations of the data differed according to the assumptions made, it can be said that some degree of paral le l ism was apparent between foliar N levels and the soil trends earlier. Thus, subsidiary hypothesis H 0 ^ was also rejected. The question of whether macronutrient deficiencies occurred after harvest ing was not completely resolved. The rather cautious position is taken that H 0 ^ has not been conclusively rejected; the conclusion must therefore be that no detectable macronutrient deficiencies occurred in the advance regeneration. Nevertheless, the possibility exists that the foliar N and needle weight cr i ter ia used were inadequate. W i t h revised standards, there was a strong suggestion of a temporary N-deficient condition at or around year eight of the sequence.. S was also inferred to have been deficient in some trees on a l l sites. Because of the "worst case" aspect of this sub-sample, this cannot be said to be the general case for the advance growth. No evidence of A l or M n toxicities was uncovered, nor was there any which might suggest deficiencies of the other micronutrients considered. Therefore, subsidiary hypothesis H 0 has not been rejected. It appears that the advance regeneration was indeed able to benefit from the assart effect in the post-harvest period. Uptake of the macronutrients appeared to have increased generally. Considering the nutrient content data of the first eight years only, this increase was as much as 78% for N , 69% for P , 28% for K , 84% for C a , and 43% for M g . Post-harvest trends of P and C a foliar levels were s t r ik ingly paral lel to those observed earlier i n the soil. However , i n terms of foliar nutrient concentrations, the magnitudes of changes in P , K , C a , and M g were fai r ly smal l . Fo r N , it is believed that Monchak ' s 251 (1982) deficiency cr i ter ia should be revised. Accordingly , it is proposed that 1.40% be employed as the cr i t ical foliar concentration below which N deficiencies can be assumed. In addition, the value of 5.5g (1000 needles)" 1 is tentatively suggested as the corresponding needle weight standard; however, it should be noted that non-N factors can also influence this measurement variable. A more rigorous investigation of the possibilities of both temporary N deficiencies and S deficiencies in the post-harvest period would be highly desirable. One alternative would be the application of the screening tr ia l approach for both; wi th this, tests of the val idi ty of the proposed N standards could also be incorporated. The results of this phase can also be discussed in terms of the relevant main hypotheses stated i n Chapter One. The supported conclusion above that advance growth did benefit from the assart release constitutes a rejection of ma in hypothesis H 0 ^. The failure to reject conclusively subsidiary hypotheses H 0 and H 0 . suggests that m a i n hypotheses H 0 „ and H 0 . should be also accepted. CHAPTER 9 GROWTH IN RELATION TO SOIL FACTORS AND FOLIAR CHEMISTRY 9.1 INTRODUCTION In previous Chapters, aspects of the soil, foliar chemistry of the advance regeneration, and growth were examined separately. The results thus far have provided some hints as to some of the factors which might have played important roles in moderating post-harvest growth—directly or indirectly. These included moisture and temperature regimes (Chapter Three), the release age of advance growth (Chapter Seven), and possible N and/or S deficiencies (Chapter Eight). However, none of these were unconditionally or conclusively implicated as growth constraints. It was therefore believed useful to explore the data further for clues regarding other potentially important influences. In addition, this would also serve to highlight which of the variables examined previously might be the most important. An integrated examination of these facets must now be attempted in an effort to relate post-harvest growth to soil characteristics and foliar chemistry. The soil variables which were examined earlier for post-harvest changes were principally measurement variables sensu Sokal and Rohlf (1981). However, it is possible that soil attributes such as the presence or absence of rotting wood and the humus forms present could have exerted detectable influences on growth. Such attributes were therefore included in the analysis which follows. It was established earlier (Chapter Seven) that the release height growth of the 252 253 subalpine fir advance regeneration appeared to have been constrained in some way . Ideally, height growth should have been used for this examination. Unfortunately , no height data were collected for 1983—the year in which soil and foliar samples were taken; diameter growth was therefore substituted. 9.2 O B J E C T I V E S A N D H Y P O T H E S E S The m a i n objective of this phase of the study was to identify the soil and foliar variables which were potentially the most important to post-harvest growth of the advance regeneration. Included i n the latter was an assessment of the influences of rotting wood in the forest floor, and also of the different humus forms present. This phase was exploratory, and not l inked to any of the ma in hypotheses stated in Chapter One; because of this, no formal statements and tests of hypotheses were attempted. Nevertheless, it is acknowledged that there are indeed hypotheses and/or assumptions under lying such an approach. 9.3 M E T H O D S It should be clear that this phase consisted entirely of manipulation and examinat ion of data already collected or derived as outlined in earlier Chapters. Absolute diameter growth in 1983 was chosen as the indicative growth variable. It should be recalled that the chemical analyses and needle weights were al l based on second-stage composites (see Chapter Two). Thus , where these were involved, diameter growth and soil attributes [presence/absence of rotting wood, humus forms by both the Bernier (1968) and K l i n k a et al. (1981) systems] had 254 to be brought to a compatible form. This was done using exactly the same random sample groupings which formed the second-stage composites. On this basis, mean diameter growth values were calculated to represent each second-stage composite. Similarly, "average" humus form classifications were assigned to the latter on the basis of the classification previously given to a majority of its first-stage member samples. This led to the elimination of Bernier's (1968) fibrimor category. The same approach was also applied to determination of the rotting wood categories. Therefore, this phase consisted of two sets of analyses: Rotting wood and humus form influences involved first-stage sample data only, while the effects of other variables required analysis of the second-stage data only. For examination of the effects of the three soil attributes, the diameter data were analyzed in each case as a two-way Model I A N O V A , with the individual sites as one set of treatments. This allowed elimination of variation due to among-site differences. The other treatment groups consisted in turn of the rotting wood categories, Bernier's (1968) humus form classifications, and the Klinka et al. (1981) humus form classifications. The MIDAS and G E N L I N programmes were used as before for data checks and analysis. Power transformations were necessary in each case to achieve homogeneity of the variances. Multiple linear regression analysis was used to attempt identification and quantification of the possible influences on absolute diameter growth of 38 of the variables considered earlier. The soil variables were forest floor and mineral soil 255 (variables considered separately) KCl-ext rac tab le N H J - N , KCl-extractable NO5 - N , total N , total P , "available" K , "avai lable" C a , "available" M g , C / N ratio, and the four N mineralization measures ( G U M N , N U M N , G A M N , N A M N ) . Fol ia r variables were needle weight, concentrations and contents of N , P , K , C a , and M g , along wi th N / P , K / C a , and C a / M g ratios. W i t h diameter growth as the dependent variable, backward el iminat ion and forward step-wise procedures were applied to these using the M I D A S programme; the significance level for inclusion/ exclusion was p = 0.05. The analysis was performed on the grouped T6 , T 8 , and T i l data. Fo r the variables selected, the val idi ty of the suggested regression models was checked by an examinat ion of the residuals compared to the predicted values (Draper and Smi th , 1981). In adopting the linear regression approach, two limitations were fully recognized. F i r s t l y , the existence of a val id regression relationship does not necessarily imply a causal relationship. Secondly, for many—perhaps all—of the independent variables under consideration, their actual relationship wi th growth would more l ikely be a curvil inear one. That is, i t is possible that the assumption of the existence of a val id linear relationship was not fully met. Nevertheless, in spite of these l imitations, it was believed that application of the techniques presented a simple and rapid way of identifying potential influences. The use of the grouped data from the older cutovers ensured that a sufficiently large and general release growth base was employed. 256 9.4 RESULTS AND DISCUSSION No significant differences (p > 0.05) were detected among the mean absolute diameter growth values for any of the soil attributes under consideration (for rotting wood: F = 0.01, df = 1, 291; for Bernier 's humus forms: F = 0.69, df = 1, 288; for K l i n k a ' s humus forms: F = 0.92, df = 2, 282)!" The observed means and their 95% confidence l imits are presented in Table 9.1. The K l i n k a humus form results suggest that the low sample numbers involved in some of the cases might have contributed to the above lack of differences by increasing estimates of variation. A more adequate design might have detected significant differences for this attribute. The backward elimination regression procedure yielded a highly significant relationship (p < 0.001), in which seven variables from the minera l soil and foliar categories accounted for approximately 48% of the variat ion in diameter growth [standard error of estimate (SE) = 0.11 cm, df = 7, 73]. These were minera l soil total N , KCl-extractable N H J - N , G U M N , and N A M N , along wi th foliar N contents, P contents, and N / P ratios. O f these, the foliar N and P contents together accounted for approximately 38% of the variat ion in diameter growth (SE = 0.12 cm, df = 2, 78); P content alone "explained" approximately 29% (SE = 0.13 cm, df = 1, 79). F r o m the practical or interpretative viewpoint, some degree of correlation (or redundancy) was anticipated between the selected G U M N and N A M N variables, and also between the foliar N and P contents and the N / P ratio. However , this was not observed to be serious enough to affect the strength of the model derived. Rather, somewhat surpr is ingly , Table 9.1 Means and 95% confidence l imits of absolute diameter growth for each soil attribute considered. A T T R I B U T E D I A M E T E R G R O W T H t 95% L I M I T S Upper Lower ROTTING WOOD present absent BERNIER HUMUS FORMX humi-fibrimor fibri-humimor KLINKA HUMUS FORM§ hemimor hemihumimor humimor 0.41 0.41 0.40 0.43 0.34 0.39 0.39 0.45 0.44 0.43 0.49 0.41 0.42 0.49 0.37 0.38 0.37 0.37 0.28 0.37 0.29 t V a l u e s are rounded means in cm. ?See Bernier (1968). §See K l i n k a et al. (1981). 258 removal of any one of the N mineralization measures rendered the contribution of the other non-significant; simultaneously, the proportion of "explained" variation decreased noticeably. Interpretations of the above results are within the context of the methodological limitations stated earlier. The results suggest that neither the presence (or absence) of rotting wood in the profile nor the type of humus form present affected diameter growth. Of the soil variables investigated, it is extremely interesting that all of those selected were measures of mineral soil N levels. This strongly suggests two points: Firstly, tree growth on the older cutovers was influenced more by soil N than any other soil nutrient; secondly, the advance regeneration was more dependent on the mineral soil for that N than on the forest floor. The apparent influence of foliar N content on growth may be taken as confirmation of the concept of N as a controlling factor. These findings thus provide some level of support for earlier conjectures concerning temporary N deficiencies (Chapter Eight). Interpretation of the apparent importance of foliar P content to growth is less straightforward. From the viewpoint of the strong parallelism noted earlier between cutover foliar and soil "available" P measures (Chapters Five and Eight), the result was somewhat expected. What is surprising is the apparent marked influence of P on growth when no deficiency of P was indicated (Chapter Eight). A tentative explanation of this is that while P levels were adequate, they had not yet reached the optimum for growth of the advance regeneration. Verification of this would necessitate a more direct examination of this possibility. 259 The proportion of var ia t ion in diameter growth accounted for by the selected variables is relat ively low. Husted (1982) observed much higher levels (up to 75% by four variables) in relat ing height growth to fall foliar chemistry of Pacific silver fir; a single variable often "explained" more than 40% of the var ia t ion. Given the range and number of variables examined, the re la t ively low level of "explained" var ia t ion suggests that factors not considered here m a y also have played important roles in regulating growth. The methodological constraints under which this phase operated would have contributed here. Fo r example, the derivation of an average diameter growth for use with second-stage samples would have itself contributed a degree of unexplainable variat ion. Nevertheless, it is believed that the results provide indirect support for the idea that growth was influenced by other factors. It m a y be of benefit to consider a few l ikely candidates. F r o m wi th in this study, moisture and/or temperature constraints are suggested (Chapter Three). Nut r i t iona l ly , the possibility of S deficiencies (Chapter Eight) would meri t further investigation. Other studies also provide ideas in this respect. In general, competition from minor vegetation was not believed to be a problem on any of the sites. Ericaceous vegetation was present, but believed insufficient to present the allelopathic effects described by Malco lm (1975) and L e Tacon et al. (1984). However , evidence of auto-toxicity in true firs has been found in France (Drapier, 1983; Becker and Drapier , 1984). F ina l ly , incipient root rot diseases could have affected growth i f they were present. Such effects have been noted to be important in ba lsam fir saplings and older trees (Whitney and M y r e n , 1978); several such pathogens have been observed to infect subalpine fir (James and Goheen, 1981). 260 9.5 CONCLUSIONS There was no evidence that any of the soil attributes considered affected growth on the older cutovers. Soil N variables appeared to exert more influence on growth than any other soil variables considered. In addition, the advance regeneration seemed to be more dependent on the mineral soil fraction for N than on the forest floor layers. Some support was thus provided for the ideas concerning temporary N deficiencies presented earlier. There were strong indications that P played an important role in affecting growth. It is tentatively suggested that while P levels were not deficient, they may not have been near those which would have been optimum for growth. It would appear that factors which were not under consideration in this phase may have exerted major influences on growth on the older cutovers. Identification and/or verification of these would require further study. CHAPTER 10 SYNTHESIS: HARVESTING EFFECTS, GROWTH, AND SILVICULTURE In the foregoing Chapters, several aspects of the post-harvest condition of the sites and advance regeneration were examined separately. These efforts must now be brought together in an effort to synthesize an integrated conception of the changes which appeared to have been effected over the age sequence. In this way, some insights may be gained concerning the silvicultural effects of such harvesting on similar E S S F sites. A recapitulation of the previous results is first presented; construction of an overview is then attempted. Finally, some general silvicultural consequences and implications are discussed. 10.1 RECAPITULATION OF PRINCIPAL RESULTS Within the context of the objectives stated in Chapter One, the investigation proceeded by examining post-harvest changes in the soil, the advance regeneration, and the microclimatic regimes operating on both of these. Only summaries of results are presented here; the earlier Chapters should be consulted for complete details. 261 262 10.1.1 The Soil 10.1.1.1 General Chemical Changes There was no evidence of any major physical forest floor changes after harvesting. The humus forms encountered suggested very minor changes within the first six years only. By contrast, major post-harvest changes appeared evident in the soil chemical characteristics. In both the forest floor and mineral soil fractions pH levels and exchangeable Ca concentrations were increased up to the 11-year point of the sequence, and possibly beyond. For the forest floor these increases were noticeable from the three-year mark, while in the mineral fraction they were evident from the six-year point. The exchangeable K concentrations of both fractions also declined temporarily. The two fractions differed in the trends of other elemental concentrations examined. Forest floor total N and "available" P concentrations were increased temporarily, while in the same fraction temporary declines were noted in total P concentration and C/N ratios. Concentrations of Mg and C in the forest floor showed no change over the age sequence. Mineral soil total P concentrations increased temporarily; total N , exchangeable Mg, and C concentrations of this fraction exhibited temporary declines. Moreover, mineral soil C/N ratios declined consistently up to the 11-year point, and possibly beyond. For both fractions, the range of C/N ratio values was lower than expected, but nevertheless very similar to those of other E S S F and/or high-elevation sites. Taken together, the soil P concentrations and "available" P measure suggested a mobilization and downward movement of forest floor P, starting at the three-year mark and lasting for three or more years. There was also a hint of a 263 temporary downward movement of organic matter wi th in the minera l profile. V iewed i n terms of the detection of an assart pattern, the chemical changes pointed ma in ly towards the first eight years of the age sequence as the most important period. The period from year three to year eight appeared to be that of m a x i m u m change, and perhaps especially around the six-year point. However , i t was clear that the t iming and direction of perceived change could differ according to the variable and fraction under consideration. Relative to mature stand values, some m a x i m u m degrees of change were calculated. In the forest floor, exchangeable C a content increased by approximately 67%, while total N and C contents showed no change. P, K , and M g content declined by 23%, 45%, and 31% respectively. Fo r the mineral soil, only an increase of 185% in C a content could be estimated. 10.1.1.2 Nitrogen Availability Changes A s stated i n Chapter One, strong emphasis was placed on N variable in this study. Fores t floor N avai labi l i ty appeared to increase to a peak at three to four years after harvest ing, then decline by year six—ostensibly to pre-harvest levels. N avai labi l i ty in the minera l soil attained its peak at the same time as i n the organic layers ; however, the peak was more extended, last ing up to year eight. NO3 - N was present, and used to identify points of peak N availabil i ty; however, the dominant form of N present across the sequence was N H J - N . There was a suggestion of a downward movement of the latter ionic species from the forest floor to the minera l soil. 264 The N availability trends therefore supported the concept of an assart effect lasting eight years, with a peak between years three and six. Within the methodological limits imposed, the increase in the magnitude of N supplying power of the soil fractions was approximately one kg ha" 1 in each case. 10.1.2 The Advance Regeneration 10.1.2.1 Post-Harvest Growth Post-harvest growth trends indicated different response delay periods for absolute as opposed to relative growth measures. Absolute height growth indicated a delay of four to five years, while that given by relative height growth was two to four years. Similarly, absolute diameter growth indicated a response delay of one to three years, while a one-year delay was given by the relative measures. The relative measures were believed to have given better estimates than absolute growth. Therefore, the advance regeneration was in a position to derive some benefit from any increased nutrient availability. From the viewpoint of the forest manager, there was a lag of five to ten years before an acceptable current rate of height growth was achieved. Post-harvest height growth appeared to be constrained in some way; this did not seem to hold true for diameter growth. There was evidence that the age at release may have played a negative role here —directly or indirectly. 265 10.1.2.2 Foliar Chemistry M a r k e d changes were observed in foliar levels of N and other macro-nutrients in the post-harvest period. No unconditional evidence was found of any macronutr ient deficiencies. However, it is believed l ikely that the criteria by which this judgement was made are in need of revision. If this is correct, the subalpine fir advance growth could have experienced a temporary N deficiency at year eight of the sequence. Revised cri t ical levels for foliar N were proposed; these were a concentration of 1.40%, and—tentatively—a needle weight of 5.5 g (1000 needles)" 1 . There was evidence also that at least some of the advance regeneration on most of the sites was deficient in S; the extent to which this might have been the general case could not be confirmed. There was no evidence of any l imita t ion to growth associated wi th the micronutrients considered. It appeared that nutrient uptake was generally increased after harvesting. Considering m a x i m u m increases relative to regeneration under the mature stand, and for the first eight years only, increases were 78%, 69%, 28%, 84%, and 43% for N , P , K , Ca , - and M g respectively. There were strong parallels between foliar N , P , and C a levels and their soil "avai lable" measures for those elements—notably for P and C a . This seemed to confirm that benefits (direct and/or indirect) from an assart effect had indeed accrued to the advance regeneration. B y the same token, it also provided support for the view that an assar t effect had indeed been operational. 266 10.1.2.3 Influences on Tree Growth Assessments here were based on the use of diameter growth in 1983. Growth did not appear to be affected by either the presence (or otherwise) of rotting wood or the type of humus form present. Of the soil variables considered, the ones representing soil N levels appeared the most influential. In addition, the trees appeared to be far more dependent on the mineral soil for N than on the forest floor. Surprisingly, P seemed to have played a major role in affecting growth, though there was no indication of a P deficiency. The above notwithstanding, it was concluded that growth appeared influenced by factors which had not been included in the statistical analyses. Suggestions were made as to likely candidates; concrete verification of these was beyond the scope of this study. 10.1.3 Microclimatic Regimes Harvesting appeared to have increased growing season air and forest floor temperatures. Increases of between 3 ° C and 6 ° C were evident for at least eight years. However, it is likely that growing season temperatures were generally sub-optimal for growth. Unlike temperature, neither growing season moisture contents nor the retention characteristics of the forest floor seemed affected by harvesting. However, insofar as forest floor formed a rooting medium, there was clear evidence of severe and sustained moisture stresses on the advance regeneration during the growing season. Decomposition and mineralization processes should not 267 have been thus affected. 10.2 OVERVIEW OF THE POST-HARVEST SITUATION Construction of a l ikely post-harvest scenario can now be attempted. It can be assumed that one of the first impacts of harvest ing was to effect an increase in growing season mean temperature levels on the cutovers—both in the soil and in the air immediately above it. I f differences in sample numbers are taken into account, the range of temperature var ia t ion did not seem to be great ly changed. In spite of the noticeable increases, growing season temperatures on the cutovers remained below that which might be considered optimal for growth of trees. Un l ike the temperature regime, the growing season moisture regime of the forest floor was not changed markedly from that which obtained under the mature stand. However, the forest floor (and possibly the minera l soil immediately below it) presented moisture avai labi l i ty conditions which were continually stressful for trees which were dependent on it. This was not an impact of harvesting, but rather a phenomenon which might be considered a site characteristic. No major physical changes occurred i n the forest floor; i t can be assumed that the same was true for the under lying minera l soil. Apparen t ly , harvest ing did not disturb the forest floor to any great degree. W i t h i n the context of the physical variables considered, it m a y be said that harvest ing had little impact at the physical level. Quite a different picture presents i tself when the chemical and/or microbially-mediated levels are examined. Whi le there may have been little 268 churning and mixing of organic and minera l horizons, many other changes had occurred. Fo r example, temperatures had been increased, soil p H levels increased sl ight ly, inputs of fresh organic mater ia l had occurred, and vegetative uptake of moisture and nutrients had been decreased. Microb ia l activity flourished wi th the changing conditions; an assar t effect was under way . Nutr ients wi th large proportions held in organic fractions—notably N and P—underwent increased mineral izat ion, and were thus more available to plants. Wi th in three years after harvest ing, peak levels of N avai labi l i ty were attained, wi th increased nitrification becoming evident. Ini t ial ly low C / N ratios i n both the forest floor and minera l soil probably contributed to this. Insofar as "available" C a concentrations are indicative of potential forest productivi ty, it can be said that the site class was at least temporari ly increased s tar t ing wi th in the first three years after harvest ing. Wha t was the reaction of the trees and other to this point? The advance regeneration was able to respond to the changed conditions wi th in one year after harvest ing. However, the response involved an emphasis on radial growth, while height growth was constrained; for the latter, younger trees were able to respond better than older ones. B y the time peak N avai labi l i ty levels were attained, the trees, while benefiting, were not physiologically able to utilize fully the pools of available N and P in the forest floor. The growing season moisture deficits in that fraction also helped to prevent such use. A portion of N and P available in the forest floor was transported downward to the minera l horizons; this included some of N O ^ - N which was i n evidence at this point. Thus, while mineral soil N mineral izat ion rates did not reach a m a x i m u m unt i l approximately six years after 269 harvesting, the mineral fraction nevertheless had relatively high levels of N H J - N and N O 3 - N even at the three-year point. The prominence of fireweed and other herbaceous vegetation from the three-year point onward was perhaps an indicator of both the increased N avai labi l i ty and the fact that some benefits were accruing to non-crop vegetation also. This is . viewed as a beneficial effect, since the possibility also exists that some of the mobilized N was lost through leaching during the peak. Since the sites were not exceptionally fertile, any such losses should have been relatively low. The post-harvest situation described thus far conforms well to that which H e a l et al. (1982) viewed as a fourth nutr i t ional stage to be added to Mil le r ' s (1981) three. The assart effect lasted for at least eight years; peak avai labi l i ty was st i l l in evidence at six years after logging. However , by the s ix th year, the situation had changed subtly. The advance regeneration were now more adapted to the post-logging conditions, and had increased their nutrient uptake levels. Forest floor N availabil i ty levels had declined. The mineral soil enrichment process had reached its m a x i m u m ; moreover, the mineral soil also reached its peak N supplying power at that t ime. The increased demand for N was satisfied from the mineral soil pool, rather than from that of the forest floor. The supply of P also became more important to tree growth, and it is l ikely that this need too was met from the minera l soil pool. W i t h the moisture stress conditions l imi t ing their exploitation of forest floor resources, the mineral component became the soil fraction which most directly influenced growth of the trees. F r o m the eight-year point, the v iew of the post-harvest situation becomes 270 obscured. The original period of increased soil N availability appeared to be over by year eight. What was unclear was whether this was a truly temporary decline in N availability, with a secondary peak to follow; whether it was a falsely indicated decline, with levels sustained in reality beyond the eleven-year mark; or whether it was a true termination of the effect, with the eleven-year point being a falsely indicated increase. The limitations of the methods employed do not allow any firm resolution of this problem. Tentatively, the third alternative above was accepted as the one believed to be valid. There is the additional problem of whether the regeneration became even temporarily N-deficient around the eight-year point. Acceptance of the third alternative above would support, and be supported by, the idea of such a deficiency; yet, it would not allow a decision as to whether this was temporary or more prolonged. It is felt that no firm statements can be made in this respect. One final consideration was that growth on the cutovers appeared to be influenced by factors which were not considered in depth by this investigation. Regardless of its position in the time sequence, evidence of S deficiencies was apparent in at least some trees on all but one of the sites. Tentatively, it is suggested that S availability was at least among the more important influences on the post-harvest nutrient status and growth of the advance regeneration. 10.3 SILVICULTURAL CONSEQUENCES AND IMPLICATIONS Silvicultural systems and treatments have to be considered within the context of specific objectives and goals—managerial, biological, and otherwise—for a 271 given forest. Therefore, the presentation which follows deals w i th broad s i lvicul tural issues and principles only. Emphasis has been placed on the consequences and implications involved in the use of subalpine fir advance regeneration in the manner encountered on the sites of this study. While other resource values were obvious, only timber production was discussed. N o consideration was given to economic realities or corporate objectives. 10.3.1 Background Highlights It is desirable to place the cutt ing system employed on the study sites wi th in the context of the general s i lv icul tural systems which have been used in or recommended for the E S S F type. Interest in the management of these forests has existed at least since the beginning of this century (Hodson and Foster , 1910), wi th a marked increase in the same in B . C . since W o r l d W a r II. B r i e f accounts of major steps in the development of concepts were given by Smi th (1955) and Bergstrom (1983). M u c h of the effort in developing a body of knowledge concerning this type was carried out in the Rocky Mounta ins in the Uni ted States, notably by R. Alexander and co-workers [for example, Alexander (1974 and 1977), Alexander and Edminster (1980), and m a n y others]. However, the insights which have been so gained should be applied wi th great caution in the B . C . si tuation, since there are crucial differences between the E S S F forests of the latter and those of the south and central Rockies in the Uni ted States (Alexander, 1986). For example, much higher elevations are involved in the U . S. A . ; 3,000 m was considered a 272 mid-elevation position by R. Alexander (1977). A t such elevations, h igh light intensities were believed to inhibit photosynthesis, leading to tissue damage and subsequent death of seedlings—a phenomenon known as solarization (Ronco, 1970; Alexander , 1974). According to Alexander (1986), major establishment and growth problems in the Rocky Mounta ins stemmed pr imar i ly from drought and moisture avai labi l i ty as wel l as heat. B y contrast, it was felt that low soil temperature might pose a major problem on many E S S F sites in B . C . W i t h i n the context of this report, perhaps the most important difference between the two E S S F forests is a notable lack of Enge lmann spruce advance regeneration in the understory of mature stands in B . C . ; moreover, understories in B . C . appear much denser (Alexander, 1986). A n example can be seen by comparing the mensurat ional data from Chapter Two wi th the findings of Alexander (1985). Alexander (1986) surmised that the lack of spruce in the understory is a possible indication of a more serai role of the species in the B . C . E S S F forests; the "true" ' c l imax species might indeed be subalpine fir only. Regeneration objectives in E S S F forests have t radi t ional ly emphasized spruce as the crop of interest. In this, it has long been recognized that one of the major difficulties has been that most harvest ing disturbances favour regeneration of or by subalpine fir (Hodson and Foster, 1910; Smi th , 1955; Alexander , 1974; Alexander and Engelby, 1983). This has also been found to be true of spruce beetle {Dendroctonus rufipennis K i r b y ) attacks (Schmid and Hinds , 1974), and even fire suppression policies (Cole, 1981). In B . C . , the lack of spruce advance growth no doubt exacerbates this problem. Several investigators have proposed cutting and s i lvicul tural systems for E S S F forests. The most detailed presentation so far has been that of Alexander 273 (1974). He contended that, under the right conditions, a l l but one of the high-forest systems (Smith, 1986)—even- or uneven-aged—could be successfully applied. The exception was the seed-tree system, rejected because of the risk of windthrow. The major determinants of which system would be successful included such aspects as stand conditions, associated vegetation, the r isk of windthrow, and susceptibility to the spruce beetle. In old-growth stands, differences in vert ical stand structure were considered part icularly important in determining the system or variant employed (R. Alexander , 1974 and 1977; Alexander and Engelby, 1983). Current practice in the central and southern Rocky M o u n t a i n areas of the United States involves even-aged systems only. Above 3000 m on north and east aspects, smal l clearcuts (approximately 1.5 ha) and a complete dependence on natural regeneration have been found successful. Below 3000 m , the shelterwood system was the easiest in achieving regeneration on these aspects. High-elevation south and west aspects were the most difficult to regenerate, notably wi th clearcutting. O n these, the shelterwood system is applied, but takes a long time to achieve regeneration objectives (Alexander, 1985 and 1986). Investigators in B . C . have also put forward suggestions concerning appropriate s i lvicul tural systems for the E S S F type wi th in the province. Recognizing that more than one si lvicultural system could be applied, Smi th (1955) suggested cr i ter ia for choosing between clear or par t ia l cutting. These were based pr incipal ly on stand structure, age and volume, and the presence of adequate advance growth. H e noted that clearcuts might be difficult to regenerate natural ly , and that seed supplies should be ensured. The latter study was part of a comprehensive long-term investigation of the E S S F . A s par t of this, Smi th and Cla rk (1974) presented detailed results of the effects of different cutt ing and site treatment methods on regeneration 274 success. They observed that in a l l cases natural regeneration was barely adequate. Clearcut t ing methods, especially wi th s lashburning and plant ing, produced the best results. Under the then existing circumstances, no advantages were believed gained by the use of par t ia l cutting. However , w i th changed uti l ization standards and acceptance of subalpine fir, the potential value of advance growth of the latter was recognized. It was believed that w i th adequate stocking and proper protection, spacing of such reproduction combined w i t h enrichment planting of spruce or lodgepole pine was the best approach in this respect (Smith and Cla rk , 1974). The recommendations of Smi th and C la rk (1974) probably had the best research base of its k ind so far in Br i t i sh Columbia . Indeed, recent ecosystem interpretations by L l o y d (1983) supported both the application of clearcuts and light burns, and the use of subalpine fir advance regeneration i n several ecosystem associations wi th in the E S S F m l var iant . Nevertheless, technological and other changes over the last 30 years l imi t the value of the Smi th and C la rk (1974) recommendations today. For example, their sites were a l l horse-logged; wheeled skidders are ubiquitous today in the E S S F . Logs were bucked on site at that time; today, tree length extraction is common. W i t h decreased diameter l imits of cut and different classification methods, what was considered advance growth by Smi th and Clark (1974) included trees which would current ly be termed "residuals". In expressing concern for the stability of Canada 's spruce forests in general, Weetman (1980b) believed that light part ia l cuts (shelterwood or selection methods) or smal l clearcuts which allowed use of advance regeneration could achieve successful stand renewal in E S S F forests. He felt that such practices would mimic the natural dynamics of such forests to a degree which would enhance their stability. The B .C .M.O .F . (1979) developed 275 detailed guidelines for silvicultural activities in the E S S F zone; these appeared to incorporate many of the earlier suggestions. The methods recommended were complete clearcuts (with further treatments), clearcutting "with protection of residuals", and selection cutting. For the • second method above, it was stressed that much planning and care were needed; post-harvest evaluations were also believed necessary. Such points were also emphasized by Alexander (1974) for such a choice. Smith (1962 and 1986) included them as cornerstone principles in the application of silvicultural systems which employ the use of advance reproduction. 10.3.2 The Present System From the considerations thus far, it seemed the consensus that more than one silvicultural system could be applied in E S S F forests; specific site and stand conditions would be among the main determinants of the choice made. The method of cutting applied to the sites of this study was clearcutting with protection of residuals; the only subsequent stand treatment was "quality slashing" (B.C.M.O.F. , 1979) (see Chapter Two). Where does this method fit within the context of the recommended silvicultural systems? What has been its actual impact from the silvicultural viewpoint? These issues will now be considered briefly. 276 10.3.2.1 Theory Versus Reality S m i t h (1986) pointed out that the te rm "clearcutting" has been used—perhaps misused—to refer to a variety of cutting methods. The method under consideration is perhaps a case in point. Alexander (1974) indeed considered this to be a type of clearcut regeneration system; however, it was subsequently termed a "simulated shelterwood" method (R. Alexander , 1977; Alexander and Edminster , 1980). In terms of s i lv icul tural systems described by S m i t h (1986), the method is closest to that termed the "one-cut shelterwood". S m i t h (1986) preferred this term for any system employing advance regeneration i n which the overstory was removed in a single cut. In principle, then, the method applied should be considered part of a shelterwood system rather than of a clearcut system. This wri ter believes that the distinction is at least a psychologically important one. The shelterwood view implic i t ly recognizes the need for such components as pre-harvest evaluations, careful planning and stand entries, and post-harvest evaluations and treatments. Th is is less obvious i f the clearcut term is applied—especially where relat ively large cut blocks are envisaged. The issues and techniques involved in the proper application of a shelterwood system have been well documented by Smi th (1962 and 1986). R. Alexander (1974 and 1977) and Alexander and Edmins ter (1980) dealt wi th such considerations specifically for E S S F forest situations. S m i t h (1986) discussed the one-cut shelterwood method at some length. M u c h was found to commend the system. In N o r t h Amer i ca , it has been applied wi th considerable effectiveness at 277 the nascent stages of forest management. Its main use has been in bringing degenerate old-growth or selectively logged stands under management; the existence of markets for small or defective trees is desirable. The principal disadvantages of the system were perceived to lie in the complete lack of control over regeneration, and the substantial losses of volume during the establishment of the regeneration. The system was seen as a blunt approach which is of value in the initial stages only of the development of silvicultural practices (Smith, 1962 and 1986). One concept has been perceived as a vital prerequisite to successful application of a shelterwood system: The overall adequacy of the advance regeneration to the task of forming the next crop must be verified. This has been expressed in various ways; for example, advance growth should be "adequate" (Smith, 1955), form a "manageable stand" before overstory removal (R. Alexander, 1977), or be "satisfactory" (Smith, 1986). The concept is simple enough; however, its practical application in the E S S F forests in B.C. involves a plethora of issues and problems. The lack of spruce representatives and the contempt for subalpine fir (Chapter One) were discussed earlier; of necessity, commercial and corporate considerations enter the picture. Even with acceptance of subalpine fir, a much more basic issue is precisely what comprises advance regeneration as opposed to residuals. Size at the time of logging has been the main criterion. For Alexander (1974), advance regeneration meant trees less than approximately ten centimetres Dbh; the B . C . M . O . F . (1979) implied that trees greater than eight centimetres Dbh should be considered residuals. Other investigators have considered trees greater than three metres in height to be 278 residuals (Herr ing, 1977; Monchak, 1982; Bergs t rom, 1983). The two views are believed to be incompatible. Such definitions are of part icular importance in B . C . where subalpine fir and the method of "clearcutting wi th protection of residuals" are concerned. Irrespective of the way(s) in which size and age at harvest ing affect release growth (see Chapter Seven), the definitions seriously influence the management and projected yields of residual stands. Fo r example, guidelines for post-harvest treatments have included removal of non-crop and/or unresponsive trees and spacing ( B . C . M . O . F . , 1979; Monchak, 1982). Residuals greater than five metres in height have been considered to be non-crop trees ( B . C . M . O . F . , 1986b). Such decisions would also affect the age characteristics of the stand, w i th concomitant questions concerning of growth response, vigour, susceptibility to disease, and length and yield of the second rotation. Bergs t rom (1983) argued that it is the residuals (trees taller than 3 m at the harvest) that could maximize volume increments and decrease the time to the second harvest . F i n a l l y , there is the consideration that shelterwood (and clearcut) methods are aimed at creating essentially even-aged stands (Smith, 1986). The variety of ages and sizes that can be included in the residual stand under the present sys tem mili tates against such a management approach. Indeed, Bergstrom (1983) demonstrated convincingly that uneven-aged management was the logical and feasible approach to use on such residual stands. The situation becomes even more complicated wi th the inclusion of other cr i ter ia for deciding which subalpine fir representatives should be selected as crop trees—in both the pre- and post-harvest stages. This problem has been a general one associated with the use of true fir advance regeneration both in B . C . and the northwestern Uni ted States (Herr ing, 1977; Monchak, 1982; Ferguson, 1984). In B . C . , much attention has 279 been paid to these aspects, resul t ing in m a n y interesting suggestions (Herring, 1977; B . C . M . O . F . , 1979 and 1986b; V y s e et al, 1979; Monchak, 1982; Ivanco, 1985). Cur ren t thinking appears to reflect much of Monchak 's (1982) th inking (Ivanco, 1985; B . C . M . O . F . , 1986b). Thus far, we have established that the harvest ing methods applied on the sites of this study properly belong to the shelterwood family of s i lvicul tural systems. However , the question of wha t comprises a second crop is in need of better resolution. While currently planned and future operations should benefit greatly from recent developments, it is reasonably certain that many older cutovers were not created wi th in such an enlightened context. Smi th (1986) noted that the one-cut shelterwood system has often been applied quite unintentionally throughout N o r t h Amer i ca . This is probably true of many of the older E S S F cutovers in Br i t i sh Columbia . The statement m a y also be largely true of the sites of this study. Fo r the Clearwater Dis t r ic t (and, presumably, the Kamloops Region) i n general, the attempts to use subalpine fir advance regeneration arose largely because of funding constraints which prohibited other actions ( M . E . Monte i th , pers. comm., J u l y 12, 1984; Fern ie , B . C . ) . Moreover, at that t ime, much of the logging wi th in that portion of T F L 18 (though not on the study sites) was governed at least i n part by fears concerning spruce beetle attacks (E. R. Swanson, pers. comm., J u l y , 1983). G i v e n the above circumstances, it is felt safe to assume that little of the careful planning and other considerations necessary to a successful application of a shelterwood system was applied in the case of the study sites. Fur ther , it is highly l ikely that this was also true for m a n y E S S F cutovers of s imi lar or older ages in B . C . To some extent, this 280 assumption was supported by such features of the cutovers as skid-road layout and density, and the general state of the residual stands. O n older cutovers in the E S S F , the question also arises of the t iming and implementation of post-harvest evaluations and treatments—if any. According to E . R. Swanson (in litt., M a r c h 15, 1984), evaluations and follow-up treatments would normal ly be carr ied out wi th in five years of logging. Ivanco (1985) indicated that at least the post-harvest evaluations should be completed wi th in that period. Such was not the case on the study sites; in addition, subsequent examinations of other cutovers wi th in the E S S F zone tended to confirm that such practices had not a lways been followed. The current backlog problems wi th in the E S S F zone in the southern interior are perhaps a fair indication of the absence of the same. It is noteworthy that under s imilar circumstances in the U . S. A . , a forest manager would have been legally required to regenerate the stands art if icial ly after five years (Alexander and Edminster , 1980). A summary of the s i lvicul tural reali ty of the cutting method used can now be attempted. A one-cut shelterwood system was applied. However , implementation appeared to have been largely unintentional, or otherwise executed wi th inadequate amounts of planning, care i n stand entries, and subsequent treatments. Questions such as the composition and quality of the second crop were apparently not fully resolved; the crucial concept of verifying the overal l adequacy of the advance regeneration appears to have been largely ignored. The financial and biological constraints operating at that time were among the pr incipal causal factors; in large part, they rendered the situation beyond the control of the forest manager. It was noted earlier that harvest ing disturbances 281 i n these forests heavily favour regeneration of and by subalpine fir. It is possible that an optimistic view of the second-growth potential of such regeneration also aided acceptance of such a poorly defined system of operation. G iven the foregoing considerations, the logical outcome of such harvesting methods on the sites would be to promote realization of the disadvantages of the one-cut shelterwood system of regeneration. 10.3.2.2 Silvicultural Consequences A n overview of the s i lv icul tura l effects of the harvesting efforts on the sites can now be presented. The residual stands were comprised overwhelmingly of subalpine fir trees, w i th few spruce representatives. The spatial dis tr ibution was very irregular. Fo r "acceptable" well-spaced stems, various post-harvest m i n i m u m stocking levels have been recommended. Examples are 500 stems ha" 1 ( B . C . M . O . F . , 1986b), 625 stems ha" 1 (Monchak, 1982), and between 900 and 1200 stems ha" 1 (Lloyd, 1983). In the central Rocky Mountains, 1500 stems ha" 1 was considered the m i n i m u m acceptable level after five years; successfully regenerated stands seldom exhibited densities greater than 5000 stems ha" 1 at age ten years (Alexander and Edmins ter , 1980). Considering only advance regeneration less than three metres ta l l at the time of logging, the residual densities encountered in this study were greater than the highest recommended level above. Densities more than two times the highest levels recommended for B . C . were encountered (see Table 2.5). Thus, it is highly l ikely that the residual stands were overstocked. This notwithstanding, the regeneration did not fully occupy any of the sites. There was a proliferation of herbaceous vegetation; 282 shrubs were also present, but did not appear to have increased their cover to any marked degree. With the residuals included, the trees on the cutovers exhibited a wide range of sizes and ages. In spite of the "quality slashing" done after logging, there were many trees on all the cutovers which did not meet some of the suggested requirements of acceptability. The lack of control over regeneration which is so characteristic of the one-cut shelterwood system appeared to be very evident. An assart effect was triggered by harvesting; this lasted for at least eight years. The residual stands achieved some benefits from this; nutrient status was generally improved for the duration of the effect. The noticeable increase in numbers of herbaceous representatives suggests that the minor vegetation component also enjoyed such benefits. Soil and air temperature regimes were improved by harvesting, but remained at sub-optimal levels. There was no marked effect on the soil moisture regime. Harvesting was not sufficient to disturb the organic layers seriously; coupled with the high, irregularly distributed stem densities and a lack of follow-up treatments, this could have had some interesting consequences. In the first place, a heavier removal of vegetation and disturbance or mixing of the organic layer could have improved the soil temperature regime even more. At the same time, any moisture stress effects derived from the organic layers would have been alleviated. Further, more receptive seedbeds would have been created for ingress of spruce (and fir) regeneration. The possibility exists that the constrained growth of the advance regeneration was in part related to the moisture conditions of the forest floor and the sub-optimal temperatures. The regeneration were mainly dependent on the 283 mineral soil layers for nutrients, while the forest floor acted ma in ly as a reservoir for their replenishment. The undisturbed forest floor can therefore be viewed as hav ing been more of a l iabil i ty than an asset to the establishment and growth of the second crop. It should be stressed that this v iew must be tempered by a consideration of the detrimental aspects of forest floor removal mentioned earlier (Chapter One). Therefore, the implication here is that successful application of the one-cut shelterwood method to such sites wi l l also involve some degree of forest floor manipulat ion or humus form management; delicacy w i l l be required here. W i t h increased disturbance levels, there would also be the problem of heightened nutr ient avai labi l i ty when the tree crop is unlikely to utilize it completely. The minor vegetation component might be increased, to the later detriment of the tree crop; al ternatively, leaching losses—or even denitrification losses (Mar t in , 1985) —might become marked. The origins of the constrained growth of the residual stands raise other si lvicultural issues. The phenomenon did not appear to be associated wi th N deficiencies on any widespread or long-term basis for at least several years after harvesting. However , if, as hinted, S deficiencies were the pr incipal cause, the question of ferti l ization of the residual stands arises. The economics of such an option would need to be examined carefully. The results of this study have suggested that such constraints could "have been associated wi th higher release ages —even at the lower values encountered on the sites. If age was indeed a main influence—directly or indirectly—the application of the harvest ing method and selection procedures for the residual crop would have to be tightened considerably. N e w cri ter ia incorporating age estimations would have to be devised. Bo th 284 Her r ing (1977) and the data of this study have indicated that size and age of subalpine fir advance regeneration are only loosely related; developing rapidly-applied selection cri ter ia in this respect could prove to be a difficult problem. The question of the influence of soil moisture regime is a ve ry important one. This study has demonstrated that even wi thin an area considered to be part of the "Wet Belt" , there are sites on which serious moisture stresses can develop. The sensitivity of subalpine fir to such stresses was established earlier (Chapter Three). Taken together, these two aspects indicate strongly that the variety of sites on which this species should be considered a viable commercial option for regeneration . m a y be much more l imited than that suggested by its range. F ina l ly , Alexander (1986) raised an issue which has far-reaching implications: A r e the expectations concerning regeneration, growth, and yield of second-growth stands within the E S S F zone of B . C . unrealist ically high? A n examinat ion of this question is beyond the scope of this presentation. A n answer in the negative would imply that much more effort is necessary to uncover biologically and commercially sound keys to successful regeneration of E S S F forests. A n affirmative answer would require hard decisions concerning renewabil i ty and further use of the resource; technical, economic, and socio-political factors would al l be involved. 10.3.3 Prognosis Hav ing examined the context, application, and consequences of the harvesting of the study sites, we can now conclude by focussing attention briefly on two final s i lv icul tura l issues. The first is the l ike ly fate of the cutovers 285 investigated—and, by extension, s imi lar E S S F sites harvested in this fashion. The second issue is what broad lessons might these considerations have for future harvests and stand removal in the E S S F zone. 10.3.3.1 The Untreated Cutovers Left as they are, wi th no further stand treatments, the fate of the cutovers can be projected on the basis of existing stand conditions and current knowledge of such situations. A s noted earlier, post-harvest regeneration by natural means wi th in the E S S F zone is extremely difficult. This is par t icular ly true in the Kamloops Region above 1500 m in elevation ( B . C . M . O . F . , 1979). According to Alexander and Edminster (1980), ingress of Enge lmann spruce m a y be slow and poorly distributed wi th any cutting method, and despite the best efforts of the manager. In the case under consideration, circumstances prevented the manager from displaying his or her best efforts. Moreover , the presence of an undisturbed forest floor would prohibit germination and su rv iva l of both spruce and fir seedlings (Eis, 1965; Danie l and Glatzel , 1966; Dan ie l and Schmidt, 1972). In any case, one crucial prerequisite was missing over much of the area—a nearby and adequate seed source for spruce. The paramount importance and value of this was stressed by Smi th (1955), and again by Alexander (1986). Even more to the point, it has been f i rmly established that the effective seeding distance for successful spruce regeneration should not be more than approximately 120 m (Alexander, 1974; Noble and Ronco, 1978; Alexander and Edminster , 1980). Fur the r ingress by either spruce or fir should therefore be min ima l ; the residual stands i n their current state should be ' the only source of any future 286 reality. It was established that the post-harvest state of these stands is far from encouraging. G r o w t h is below average, and a regeneration lag was apparent; this situation is unl ikely to change. O n the basis of B . C . M . O . F . variable density yield curves, the expected rotation and yield of Enge lmann spruce in old-growth stands of that area are approximately 100 y r and 250 m 3 ha" 1 respectively. Fo r subalpine fir, the figures were 70 y r and 200 m 3 ha" 1 respectively (A. Kokoshke , t pers. comm., J u l y 29, 1986). The rotations and yields associated wi th the existing residual stands should be noticeably longer and lower respectively. The example given by the Barkervi l le E S S F cutovers in the Cariboo Region support this view, w i th estimated second-growth rotations of at least 180 years (Vankka , 1983). The pathological rotation of subalpine fir is some 150 to 200 years (Henderson, 1982; Alexander et al, 1984). A . Kokoshke (pers. comm., J u l y 29, 1986) was convinced that the untreated residual stands of this study could not be depended upon for a second harvest in any realistic time frame. The writer agrees w i t h this view. The extension of the foregoing arguments is that the one-cut shelterwood system as encountered in this study should not be considered a viable s i lv i -cultural option for such E S S F sites. Major revisions and stricter adherence to s i lvicul tural principles and recommendations w i l l be necessary for success in future applications. In this, f inancial and other constraints could seriously frustrate the best intentions and efforts of forest managers. t P lanning Forester , C T P . 287 10.3.3.2 Future Harvests The point was stressed at the inception that specific objectives and goals form the context of any consideration of s i lvicul tural activities. A reminder of this was considered useful here, since the statement is true for any future harvest ing activities in the E S S F type. What follows can only consist of broad generalizations made on the basis of insights gained earlier. The rather bleak outlook for the cutovers does not in any way refute the idea that several s i lvicul tural systems can be applied successfully in B . C . ' s E S S F forests. W h a t it does stress is that the environmental and biological characteristics of these forests make high levels of p lanning and care a pre-requisite for success. This point was developed from a poorly applied one-cut shelterwood example; however, earlier considerations show it to be the general rule for most attempts at harvesting and regeneration i n the E S S F type. Fo r success i n the latter, future cutting methods w i l l need to be considered increasingly as methods of manipulat ing the forest environment to achieve the desired regeneration result (Smith, 1962 and 1986). European approaches have demonstrated the feasibility of obtaining any desired species mixture using canopy manipula t ion alone (Matthews, 1983; Smi th , 1986). According to Mat thews (1983), the clearcut, shelterwood, and selection systems can be viewed as an evolutionary progression in s i lv icul tural practices. The selection sys tem was regarded as the highest form, providing a continuous supply of goods and services. S i lv i -cul tural—and thus harvesting—approaches in B . C . w i l l probably evolve in the same direction. 288 Whether the progression above w i l l extend to large-scale application of shelterwood or selection methods wi th in the E S S F forests of B . C . remains to be seen. This would involve a level of sophistication and care in stand manipulation that is apparent ly not yet in evidence in forest management in the E S S F zone. According to R. Alexander (1977), either a s tandard or modified shelterwood cut would require two or three stand entries over a short period (e.g. 20 years); each entry would have fair ly precise objectives. Fo r selection cuts, between three and six entries might be needed over the rotation (R. Alexander , 1977). Mat thews (1983) pointed out that the attitudes which promoted the evolution of high-level practices in Europe arose dur ing the 17th and 18th centuries. This occurred when the pressures on the forest were perceived by the majority of the population to pose a serious threat to their well-being. Such perceptions are only l ikely to arise where the population as a whole is heavi ly dependent—and in close contact with—the forest resource. This is not the case where the extensive E S S F forests in B . C . are concerned. Thus , it is felt that economic concerns wi l l be the major determinants of the systems applied. G i v e n the difficulties and complex issues involved in obtaining natural regeneration in E S S F types, there is considerable interest in art if icial regeneration options. F o r m s of this include simple replacements of harvested cl imax species in single-species plantations, species mixtures through combinations wi th fill-in planting, and outright species conversions. It is in circumstances such as these that the considerations of this study m a y assume increased importance. The assart and related effects triggered by such large-scale operations would need consideration. Issues would include the degree of disturbance and the micro-289 environments created, effects on competing vegetation, t iming of plant ing operations, and control of possible nutrient losses. Moreover , species selections for renewal w i l l be based increasingly on site-specific information and ecological cr i ter ia (Nuszdorfer and K l i n k a , 1982; L l o y d , 1983). Such directions should force greater care and attention to principles, wi th some bearing on the features of the cutt ing methods used. The record of plantat ion successes in the E S S F zone of B . C . is not a good one. Even with such improvements, in the near future it w i l l r emain open to question whether this approach can fully meet future reforestation needs. In this respect, the use of species mixtures or conversion of suitable E S S F sites to lodgepole pine for a second-rotation crop might meri t greater consideration than the conventional approach of replanting wi th spruce. Lodgepole pine usual ly forms a long-lasting par t of the natural succession to spruce-fir cover after fire in the E S S F forests (Alexander, 1974 and 1986; B . C . M . O . F . , 1979). A s such, it would appear to be a logical candidate i n any species selection procedure for a second-rotation crop. Monchak (1982) appeared to favour such conversions where feasible, as did A . Kokoshke (pers. comm., J u l y 29, 1986). O n the basis of the successional argument and inspection of lodgepole pine stands at high elevations, this writer finds much in this option to commend it. Harves t ing and s i lvicul tural practices associated wi th lodgepole pine are much different from those associated with spruce-fir cover. Thei r impacts i n assart terms m a y be very different to those observed in this study. The effects of such conversions would thus form an interesting area of future research. 290 10.4 CONCLUSIONS In this Chapter, an attempt was made to develop an overall conception of the changes which were effected over the age sequence. The objective was to dis t i l l from this some insights which might be of value to si lvicultural practices in the E S S F type. It was established that the harvest ing method applied approximated a one-cut shelterwood system. Unfortunately, largely because of financial and other constraints, the application of the sys tem left much to be desired. The residual stands were almost entirely comprised of subalpine fir. Their characteristics i l lustrated the lack of control of regeneration typica l of the cutting method. Considerations of possible growth constraints and the relatively low degree of disturbance led to the suggestion that successful application of the one-cut shelterwood method to such sites would also involve some measure of forest floor manipulat ion or humus form management. The origins of the seemingly constrained growth could have important consequences for s i lvicul tural practices in the E S S F type. These could include decisions concerning fertilization, more efficient and rigorous application of proper harvest ing and second-crop tree selection procedures, and perhaps the incorporation of additional age criteria. The question of whether expectations of growth and yield of second-crop stands in the E S S F zone are unrealist ically high needs to be addressed urgently. Lef t untreated, the prognosis for the cutovers studied is a bleak one. O n this basis, it was concluded that the one-cut shelterwood system as encountered 291 in this study should not be considered a viable s i lvicul tural option for such E S S F sites. This in no way negates the idea that several s i lvicultural systems might be applied successfully in B . C . ' s E S S F forests. It is l ikely that in the future economic concerns wi l l be the major determinants of the systems applied. Several art if icial regeneration alternatives exist for these forests. One of the more excit ing ones might be conversion of traditional spruce sites to culture of lodgepole pine. The long-term impact of such conversions and their concomitant practices would form an interesting area of future research. C H A P T E R 11 CONCLUSIONS The principal aim of this study was to examine whether N supply problems existed in the early stages of the second-rotation tree crop of an E S S F forest. The main intention was to provide insights concerning post-harvest N dynamics in relation to growth and silviculture in such forests. The primary objectives included investigation of the assart effect, N status and growth of advance regeneration, and the implications the findings might hold for silvicultural practices. Some general conclusions can now be stated in terms of the objectives and the main hypotheses developed to aid their achievement (see Chapter One). In presenting this study, results were interpreted and reported in a cumulative fashion. This was done in an effort to stress implicitly the gradual, cumulative, and continuous nature of the process by which ideas were formulated. This was particularly true of the latter stages of the thesis, which culminated in a synthesis of the separate phases. Because of this, there is some degree of overlap between earlier statements and those given below. For the sake of brevity, a synoptic view only is presented here. It was felt that little purpose would be served by a reiteration of the detailed interpretations and comments given earlier. Hypothesis H 0 was clearly rejected; an assart effect was indeed operational on the cutovers. The phenomenon manifested itself in various ways, with the magnitudes and directions of change dependent on the characteristic or variable examined. The effect lasted for at least eight years after harvesting, 292 293 wi th a peak of change between years three and six. L o w C / N ratios were encountered; forest floor mean values ranged from 19.8 to 29.6, while those of the mineral fraction fell between 24.7 and 31.1. The low ratios probably contributed to the increased N avai labi l i ty . The existence of low C / N values—especially under mature stand conditions—could be indicative of a generally higher level of microbial act ivi ty and N availabi l i ty in E S S F forests than previously supposed. For N , peak avai labi l i ty from the forest floor was near the three-year point; that of the minera l soil occurred around year six. There appeared to be some downward movement of N , from the forest floor to the mineral soil. The period of increased N availabi l i ty ended by year eight. The increase in the N supplying power of the two soil fractions was tentatively estimated as one kg ha" 1 for each fraction. Hypothesis H 0 ^ was also clear ly rejected; the advance regeneration benefited from the assart effect. The trees responded to the changed conditions wi th in one year of harvesting; rad ia l growth was emphasized, while height growth was constrained. Nut r ien t uptake appeared to increase generally. D u r i n g the first eight years, m a x i m u m increases (relative to regeneration wi th in the mature stand) were estimated as 78%, 69%, 28%, 84%, and 43% for N , P , K , C a , and M g respectively. However , in the early stages of the N avai labi l i ty peak, the trees were unable to utilize fully the available N pools. A portion of the mobilized N may have benefited minor vegetation, thereby perhaps contributing to its increased representation—particularly of the herbaceous component. In the latter portion of the N avai labi l i ty peak, the trees were more adapted to the post-logging conditions. The resulting increased demand for N was 294 satisfied from the minera l soil pool, rather than from that of the organic layers. Coupled wi th low forest floor moisture, this made the mineral soil fraction a more direct influence on the growth of trees than the organic forest floor layers. There was no clear or unconditional rejection of H 0 . Nevertheless, there was evidence that at least some of the advance regeneration on most of the sites was deficient in S; the extent to which this might have been the general case could not be confirmed. Moreover , the possibility exists that the cr i ter ia used to judge the N status of the trees are in need of revision. If this is true, growth of the subalpine fir advance reproduction could have been N- l imi t ed at year eight of the sequence. In keeping wi th the approach explained earlier (Chapter One), the cautious conclusion was accepted; that is, there was no macronutrient l imitat ion on the abil i ty of the advance regeneration to benefit from the assart effect. However , it is proposed that a revised cri t ical level of 1.40% for foliar N concentrations be used in evaluating the N status of subalpine fir advance regeneration. In addition, i t is tentatively suggested that a needle weight of 5.5 g (1000 needles) - 1 be considered as possibly indicating an N deficiency. It should be emphasized here that non-N factors can also influence this measurement variable. F i n a l l y , it is believed that a more rigorous investigation of the possibilities of both N and S deficiencies in the post-harvest period is needed. One alternative would be the application of the screening tr ia l approach for both; wi th this, the val idi ty of the proposed N standards could be assessed. There was no equivocality where hypothesis H 0 was concerned. The latter was not rejected; there was no evidence of any l imitat ion to growth associated wi th the micro-nutrients considered. 295 Both hypotheses H 0 and H 0 were rejected. Harves t ing increased o o growing season air and forest floor temperatures by 3°C to 6°C. Nevertheless, temperatures were generally sub-optimal for growth. The