Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Habitat selection and use in winter by moose in sub-boreal forests of north-central British Columbia,.. Eastman, Donald Sidney 1977-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


UBC_1978_A1 E28.pdf [ 32.06MB ]
JSON: 1.0094631.json
JSON-LD: 1.0094631+ld.json
RDF/XML (Pretty): 1.0094631.xml
RDF/JSON: 1.0094631+rdf.json
Turtle: 1.0094631+rdf-turtle.txt
N-Triples: 1.0094631+rdf-ntriples.txt
Original Record: 1.0094631 +original-record.json
Full Text

Full Text

HABITAT SELECTION AND USE IN WINTER BY MOOSE IN SUB-BOREAL FORESTS OF NORTH-CENTRAL BRITISH COLUMBIA, AND RELATIONSHIPS TO FORESTRY by DONALD SIDNEY EASTMAN B.Sc, University of British Columbia 1962 M.Sc, University of Aberdeen 1964 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Plant Science We accept this thesis -as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 1977 (c) Donald Sidney Eastman In presenting this thesis in partial fulfillment 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 representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Plant Science The University of British Columbia 2075 Wesbrook Place Vancouver, B.C., Canada V6T 1W5 Date Chairman: Dr. V. C. Brink ABSTRACT A study of winter habitat selection and use by moose 2 was conducted in a 11,300 km area of north-central British Columbia from May 1971 to August 19 73. The study area was located within the forested sub-boreal spruce biogeoclimatic zone, a zone that is receiving increased development, especially by forestry. Habitat selection and use was examined mainly be pellet group surveys and aerial transects. Wintering moose used partial cutovers and burns more than coniferous forests; deciduous forests and recent clearcuts were used least. Limited data suggested a similar pattern in summer. Winter use typically increased from near zero after a recent disturbance such as clearcutting, to a peak sometime between 10 and 25 years later, then declined to low levels during 25 and 9 0 years, and then apparently stabilized in the mature forest stage at slightly higher levels. On one intensively surveyed area, moose selected partial cutovers and creek bottoms even though these habitats comprised less than 6 percent of the area. Moose began concentrating on winter ranges at least by mid-November, reached a peak in November-January, and declined steadily thereafter. Food habits and diet were examined by Ill rumen analysis, trailing and post-winter browse surveys. Moose had catholic diets but ate primarily deciduous browse for most of the year. Subalpine fir becomes important in late winter. Diet varied according to season and habitat. Preferred species typically were least common. Tagged twig transects revealed that moose frequently browsed plants more than once but rarely re-browsed a twig. The time of browsing varied by species and by habitat with most use recorded in January and in April. Levels of utilization were all less than 100 percent of the previous year1s production. Utilization (weight-basis) ranged from 33 per cent on red-osier dogwood to 3 percent on subalpine fir; and from trace amounts in an upland burn habitat to more than 40 percent in deciduous forest, partial cutover and river bottom habitats. Bedding .habits were examined in an attempt to define cover for moose. Moose choose upper slopes that faced south particularly when snow depths became restrictive (> 8 0 cm). Moose tended to select larger than average trees and to bed on the southerly sides of them. Selection of bed sites varied with snow depth. As snow became deeper, moose bedded closer to larger trees in the denser canopied parts of forest stands. Moose showed greater selection for protected sites as winter conditions became more severe. Secondary serai succession was examined with respect to several attributes for mesic environments on the two commonest substrates, glacial till and lacustrine deposits. iv Floristics of serai stages from 1-200 years revealed that on lacustrine soils, vegetation was more, diverse and the deciduous phase was prolonged. Species diversity declined around year 25 on till but not on lacustrine. Several major changes occurred in the tree layer: first, a deciduous tree layer developed especially on lacustrine soils; second, after 25 years on till (45 years or more on lacustrine), lodgepole pine became most abundant; third, pine was gradually replaced by white spruce after 150-200 years or more; fourth, subalpine fir would probably become the domi nant tree species in the absence of fire. Understory phytomass, though contributing little to the mature forest mass, increased peaks early in succession and then remained low. Approximate net primary production of the understory on till was greatest at age 11 with 133 2 2 g/m /yr produced and least at age 39 with 18 g/m /yr produced. Understory production in the mature forest was an 2 estimated 27 g/m /yr. The shrubs contributed 70 percent, 26 percent, 44 percent, and 26 percent of annual production at ages 1, 11, 39, and 195 years, post-disturbance. Crude protein and lignin values were determined for 10 species (eight shrubs, one conifer, one lichen) for an annual cycle. Crude protein averaged 7 percent.and lignin, 9.8 percent. Crude protein increased abruptly from steady winter values to peaks of 10-15 percent in June-July and then returned to low levels by October. Leaf protein was higher than, and V predictable from, stem levels. Crude protein varied by-species, sometimes by substrate and rarely by habitat-type, at least for the species analyzed. The lichen, lungwort, retained a high protein value of approximately 11 percent throughout the year. Lignin levels varied seasonally, though less dramatically than crude protein. Levels were affected by species, substrate and age of serai stage. Protein levels were similar to those reported in the literature. Factors influencing crude protein were difficult to disentangle due to confounding. Winter climate was studied with respect to differences in snow features between habitats. Moose moved into winter ranges before snow depths were limiting. This indicates snow acts to trigger migration. On winter ranges, moose also moved into forested habitats in mid-winter (January) when snow depths approached 80 cm. Snow depths and densities varied between habitats. Snow cover was more variable in partially logged cutovers than in the open or forested stands. The climate of forest, ecotone and adjacent open areas were documented. Compared to adjacent open areas, the forest had higher relative humidity, less wind, more moderate temperatures and approximately 50 percent of the snow depth. The transition zone from open to forest climates appeared to be relatively narrow, less than 50 m. The relationship between carrying capacity, habitat selection and home range are discussed with reference to moose and management of their habitat. Management recommendations and suggestions for future research are provided. TABLE OF CONTENTS Page ABSTRACT ii TABLE OF CONTENTS viLIST OF TABLES xiiLIST OF FIGURES AND ILLUSTRATIONS xx LIST OF APPENDICES xxiv ACKNOWLEDGEMENTS1. INTRODUCTION 1 1.1 The Study1.2 The Approach to the Study 6 1.3 The Need for Integrated Management 8 1.4 A Land Use Perspective. 11 1.4.1 General Introduction 11.4.2 Mining 15 1.4.3 Agriculture. 18 1.4.4 Forestry 23 1.4.5 Wildfire 9 2. THE STUDY AREAS 34 2.1 Biophysical Setting 32.2 The Primary Study Areas 52 2.3 The Secondary Study Areas 68 2.4 Moose Distribution and abundance 70 vii viii 3. HABITAT USE AND SELECTION 73 3.1 Introduction 7 3 3.2 Methods • 77 3.2.1 The Synoptic Survey3.2.2 Pellet Group Counting Methods for Detailed Survey. 80 3.2.3 The Aerial Surveys 1 3.3 Results 84 3. 3.1 Habitat Use 83.3.2 Habitat Selection 9 7 3.3.3 The Timing of Migration and Occupany Periods 9 9 3.4 Discussion 102 3.4.1 The Importance of Habitat Variability. 102 4. FOOD HABITS 109 4.1 Introduction 104.2 Methods 110 4.2.1 Rumen Analysis 114.2.2 Trailing 117 4.2.3 The Browsed Stem Survey 118 4.3 Results 114.3.1 The Range of Species Taken 118 4.3.2 The Seasonal Trends 12 0 4.3.3 The Effect of Habitat-Type on Diet 125 4.4 Discussion 130 4.4.1 Methodology 13ix Page 4.4.2 Variations in the Diet 131 4.4.3 Some Management Implications of Variations in the Diet 133 4.4.4 Future Research. . . . 135 5. THE DYNAMICS OF WINTER BROWSING 137 5.1 Introduction 135.2 Methods 138 5.3 Results 143 5.3.1 The Incidence of Use 14 3 5.3.2 The Time of Use 149 5.3.3 The Level of Use 152 6. BED SITE SELECTION BY MOOSE IN WINTER 15 7 6.1 Introduction 15 7 6.2 Methods 5 9 6.3 Results 162 6.4 Discussion 179 7. SECONDARY SUCCESSION IN SUB-BOREAL FORESTS 184 7.1 Introduction 187.2 Methods 18 8 7.2.1 Stratification 188 7.2.2 Field Sampling Procedures 190 7.2.3 The Prediction of Mass and Height of Woody Plants 198 7.2.4 Date-Analysis 207 7.3 Results for Mesic Upland Sites 208 7.3.1 The Data Base 20 8 X Page 7.3.2 Floristic Changes in Serai Succession 210 7.3.3 Temporal Dynamics of the Tree Layer 237 7.3.4 Phytomass, Height and Basal Area of the Shrub Layer in Serai Plant Communities 246 For the combination of species 246 Trends in phytomass and height of food species 251 7.3.5 Phytomass of the Herb Layer in Serai Plant Communities 254 7.3.6 Net Primary Productivity of the Understory 258 7.4 Results for Riparian Sites 264 7.5 Discussion 269 7.5.1 Predicting Successional Development. 269 7.5.2 Trends in Production of Cover and Food 2 76 8. NUTRITIVE ASPECTS OF MOOSE FORAGES 2 85 8.1 Introduction 2 85 8.2 Methods 288 8.3 Crude Protein Levels 291 8.4 Lignin Levels 30 8 8.5 Discussion 314 8.5.1 Crude Protein Levels in Moose Forages 314 xi Page 8.5.2 Assessing the. Nutritive Values of Forages 318 8.5.3 Factors Affecting Nutrient Levels 322 9. EFFECTS OF FORESTS ON. WINTER CLIMATE 333 9.1 Introduction 339.2 Methods 5 9.2.1 The Estimation.of Migration and Winter Range Occupancy . 335 9.2.2 Snow Characteristics of Habitat-Types 336 9.2.3 Climate of the Forest Edge 340 9. 3 Results 341 9.3.1 Migration and Snow Accumulation of Habitat-Types 341 9.3.2 Snow Characteristics of Habitat-Types 345 9.3.3 Climate of the Forest Edge 355 9.4 Discussion 369.4.1 The Role of Snow Pack in Initiating Migration 365 9.4.2 The Role of Climate in Differential Use Between Habitats. . . . 367 10. DISCUSSION 379 10.1 Habitat Relationships in Moose Management 3710.2 The Effects of Timber Management on Moose Habitat. 39 4 10.2.1 Felling 397 10.2.2 Site Preparation 402 xii Page 10.2.3 Stand Establishment 405 10.2.4 Stand Tending . 408 10.2.5 Stand Protection 412 10.2.6 General Management Considerations. 414 10.3 Overview and Recommendations 416 11. LITERATURE CITED. . 42 7 12. APPENDICES 460 VITA LIST OF TABLES TABLE Page 1.1 Population Growth and Future Projections for Prince George and the Surrounding District 14 1.2 Major Events in the. Settlement and Growth of Prince George and the Surrounding Region 6 1.3 Number and Area of. Farms, and Numbers of Cattle for the Province and for the Prince George Region, 1881-1971 20 1.4 Trends in Logging Methods and Area Cut in the Prince George Forest District, 1950-1973 ... 27 1.5 Estimated Areas of Broad Vegetation Classes in Five Major Drainages in North-Central British Columbia (from Whitford and Craig 1918) 31 2.1 Climatic Parameters for the Study Area 40 2.2 Major Soil Associations for the Study Area and Their Relationship to Parent Materials and Moisture Regimes 44 2.3 Estimated Relative Abundance and Herd Structure for Wintering Moose on the Intensive Study Areas,, 1964-65 to 1975-76 58 2.4 Types of Analyses Conducted on the Study Area.. . . 69 3.1 Results from Trial Pellet Group Survey: Time/Plot and Number of Group/Plot 7 8 3.2 Relative Winter Use of Available Habitats on Selected Study Areas, Based on Pellet Group Surveys 8 7 3.3 Relative Winter Use of Major Habitat-Types in the Sub-Boreal Forest, Based on Pellet Group Surveys 9 xiii xiv TABLE Page 3.4 Winter Use of Ecotones Between Forests and Variously Aged Serai Stages, Based on Pellet Group Surveys 9 2 3.5 Winter Utilization of Roads and Habitats in which they were Located, Based on Accumulated Pellet Groups in 1973 at McKenzie. . . 95 3.6 Relative Summer Use of.Habitat-Types on Accumulated Summer Feces Recorded in the 1973 Synoptic Survey. . . 9 6 3.7 Distribution of Pellet Group Plots According to Habitat-Type and the Number of Groups they Contained on the Intensive Salmon Area 98 3.8 Selection of Habitat-Types by Moose in Winter as Indicated by Accumulated Pellet Groups on the Intensive Salmon Area. . 98 4.1 Components of Rumen Oligesta After Sample Preparation 112 4.2 The Effect of Analytical Method on Frequency of Occurrence of Plant Taxa Recorded in Moose Rumen Samples 114 4.3 The Effect of Analytical Method on Amounts of Plant Taxa Identified in Moose Rumen Samples . . . 116 4.4 Variety of Plant Species Eaten by Moose, by Forage Class, in Various Parts of Their North American Range . . . 119 4.5 Food Habits of Moose, in North-Central British Columbia, Based on Trailing and Rumen Analysis, 1971^-74 (%-basis) 122 4.6 Comparisons of Food Habits of Moose Between Different Habitats in Early and Late Winter. . . . 126 4.7 Winter Food Preferences of Moose in North- i Central British Columbia, by ^Habitat-Type 127 5.1 Proportions of Twigs that were Browsed Once and Twice in Major Habitats on the Eagle, Grove and Salmon Winter Ranges During the 1972-73 Winter 144 XV TABLE Page 5.2 Number of Times Plants of Subalpine Fir, Paper Birch, Red-osier Dogwood and Willow were Browsed on the Eagle, Grove and Salmon Study Areas during the 1972-73 Winter. 146 5.3 Proportion of Plants Species and Habitat, in the 1972-73 Winter 148 5.4 Time of Browsing and Level of Utilization (Weight-Basis) for All Species, Habitat Types, Study Areas and Months 15 0 6.1 Major Habitats, Snow Depth Classes and Study Areas Sampled for Bed Site Examinations 160 6.2 Example of Data Sheets-Used to Study Bed Sites . . 161 6.3 Time Spent by Moose in Beds as Indicated by Feces and Urine, According to Habitat and Month. 165 6.4 Locations of Moose.Beds with Respect to Position on Slope, and Aspect 166 6.5 Comparison of Conifer Species Available as Shelter Trees, With Those -Used by Moose 169 6.6 Orientation of Moose in Their Beds, and in Relation to the Shelter Tree 171 6.7 Location of Beds in Quamaniqs, as Affected by Habitat and Snow Depth Class 178 6.8 Comparison of Snow Depths between Moose Bedding Sites and Adjacent Areas 179 7.1 Scale Used to Assess Canopy-Coverage of Understory Vegetation (Clayer) (after Daubenmire 19 59), plus Domin Scale Equivalents 195 7.2 Summary of Features Sampled in the Synoptic Study of Succession 197 7.3 The Effect of Site on Predicting Mass for Selected Shrub Species . . • 201 xvi TABLE Page 2 7.4 Coefficients of Determination (r value) for Six Independent Variables Used to Predict Phytomass of 19 Sub-Boreal Shrubs .... 203 7.5 Regression Coefficients for Predicting Height from Diameter Measurements 2 05 7.6 Regression Coefficients for Predicing Oven-Dried, Above-Ground Phytomass of 19 Sub-Boreal Shrubs from Diameter, and from Diameter Squared by Length Measure ments. All Variables Based.on Logarithmic Transformed Data 206 7.7 Distribution of Sampling Sites for the Plant Succession Study. . 209 7.8 Plant Community Names for Successional Stages on Till and Lacustrine Substrates 211 7.9 Percent Canopy-Coverage/Frequency of Occurrence Values for Major Plant Species of the Herb (C) Layer in a Sub-Boreal Forest Sere in a Mesic Environment on the Till Substrate. .  213 7.10 Percent Canopy-Coverage/Frequency of Occurrence Values for Major Plant Species of the Herb (C) Layer in a Sub-Boreal Forest Sere in a Mesic Environment on the Lacustrine Substrate. . 215 7.11 Percent Canopy-Coverage/Frequency of Occurrence Values for Major Plant Species of the Herb (C) Layer in Partially Logged Sub-Boreal Forest Stands in a Mesic Environment on Till and Lacustrine Substrates . .217 7.12 Percent Species Composition (Stem-Basis) of the Shrub (B) Layer in a Sub-Boreal Forest Sere in a Mesic Environment Over Till Substrates 218 7.13 Percent Species Composition (Stem-Basis) of the Shrub (B) Layer in a Sub-Boreal Forest Sere in a Mesic Environment Over Lacustrine Substrates 219 xvii TABLE x Page 7.14 Percent Species Composition (Stem-Basis) of the Shrub (B) Layer in Partially Logged Stands of Sub-Boreal Forests in a Mesic Environment 220 7.15 Temporal Changes in Tree Species Composition for Mesic Sub-Boreal Forests on Till and Lacustrine Substrates 238 7.16 Temporal Trends for Coniferous Regeneration in a Mesic Environment Over Till and Lacustrine Substrates 240 7.17 Temporal Changes in Composition and Proportions of Dead Trees in Mesic Sub-Boreal Forest Stands on Till and Lacustrine Substrates 242 7.18 Temporal Changes in Basal Area, Canopy Closure, and Height of Dominant Trees in Mesic Sub-Boreal Forests on Till and Lacustrine Substrates 244 7.19 Statistics (mean = sd) for the Shrub (B) Layer in Mesic Sub-Boreal Seres Over Till and Lacustrine Substrates 247 7.2 0 Trends in Phytomass of the Herb (C) Layer for Seres in Mesic Sub-Boreal Forests on Till and Lacustrine Substrates 256 7.21 Approximate Net Primary Productivity of Understory Vegetation (Layers B and C) at Four Successional Stages on Till Substrate. . . . 260 7.22 Changes in Proportions of Plant Components of Selected Shrub Species at Four Successional Stages on Till Substrates 264 7.2 3 Major Features of Serai Stages in Forest Succession on Riparian (Alluvial) Habitats (Adapted from Sumanik (1968) and Waring (1970) ) 267 8.1 Location, Habitat, Substrate, and Species Collected for Crude Protein and Lignin Analyses 2 89 xviii TABLE Page 8.2 Estimates of Experimental Error in Protein Analyses for Selected Species.. 291 8.3 Crude Protein and Lignin Levels in Major Moose Forages Averaged Over an Annual Cycle (May 1972 - April 1973) 292 8.4 Crude Protein and Lignin Levels in the Current Year's Stems and Leaves of Selected Browse Species 295 8.5 Proportions of Stem and Leaf Tissue in Current Annual Growth of Selected Browse Species Collected in September 19 72 297 8.6 Comparison of Protein and Lignin Contents of Willow and Paper Birch Collected from Till and Lacustrine Substrates of the Burn Habitat at the Grove Study Area 299 8.7 Effect of Habitat on Crude Protein Levels in Selected Browse Species at the Eagle and Salmon Study Areas ..... 303 8.8 Year-to-Year Variations in the Content of Crude Protein and Lignin 307 8.9 Difference in Percent Lignin Content as Affected by Substrate, Habitat and Stand Age 312 8.10 Comparison of Crude Protein Values for Current Annual Growth of Common Winter Foods of Moose (November-March) 316 9.1 Location, Site Number, Elevation, and Habitat of Snow Courses. . 337 9.2 Annual Variations in Snow Depths in Open or Deciduous Forest Habitats at Eagle, Grove and Salmon Winter Ranges for March 1972, 1973, and 1974 346 9.3 Monthly Snow Depths and Densities for Three Habitats on the Eagle Winter Range 34 8 9.4 Monthly Snow Depths for Four Habitats on the Grove Winter Range 349 xix TABLE Page 9.5 Monthly Snow Densities for Four Habitats on the Grove Winter Range. 350 9.6 Monthly Snow Depths, for Five Habitats at the Salmon Winter Range. 351 9.7 Monthly Snow Densities.for Five Habitats at the Salmon Winter Range 352 9.8 Comparison of Selected Climatic Parameters Between the South-Facing Ecotone at the Grove Study Area (Station at 76 m in the Open) and Prince George. Airport. 357 9.9 Mean Monthly Temperature, Relative Humidity, Snow Pack, and Wind for Clearcut and Adjacent Forest Sites at the Bowron Study Area, 1972-73 Data 359 9.10 Mean Monthly Depth, Density, and Penetrance of Snow Across the South-Facing, Forest-Burn Ecotone at the Grove Study Area, 1972-73 Winter 363 10.1 General Features of Basic Resources Required by Moose 395 10.2 Soil Disturbance and Slash Accumulations Resulting from Different Types of Logging in Western North America (Derived from Bockheim et al. 1975) 400 LIST OF FIGURES FIGURE Page 1.1 The general relationship of sub-models that comprise a moose-forest model. Derived from Haagenrud and Hjeljord (1976) and Houston (1968) 4 1.2 Development of forestry in the Prince George Forest District as indicated by the annual cut and the number of operating sawmills, 1914-1974 25 2 1.3 Annual area (km ) burned by wildfire in the Prince George Forest District, 1910-1975 32 2.1 Locations of the study areas, and of place names mentioned in the text 35 2.2 Longterm monthly averages of some temperature and precipitation para meters for the Prince George weather station 38 2.3 Oblique aerial photographs illustrating the general terrain and vegetation of the Prince George study area 43 2.4 A schematic illustration of the major soil associations in the study area, and their topographic relationship to each other 45 2.5 Photographs of the Eagle, Grove and Salmon winter ranges 5 3 2.6 A soil association map of the Eagle study area 55 2.7 A forest cover map of the Eagle study area 6 2.8 A soil association map of the Grove study area 61 xx xx i FIGURE Page 2.9 A forest cover map of the Grove study area 6 3 2.10 A soil association map of the Salmon study area 5 2.11 A forest cover map of the Salmon study area 6 7 3.1 Photographs illustrating logged habitats in sub-boreal forests: a) selective, b) cut and leave, and c) clearcut 76-3.2 The relationship between daily snowfall and the timing of the aerial transect surveys, January 19 72 to May 19 7 3 8 3 3.3 A map showing flight lines used for the aerial transect surveys on the Grove study area 85 3.4 Relative use by moose of ecotones and adjacent habitats, based on pellet group transects 9 3 3.5 Number of moose seen/minute of flying on the Eagle, Grove and Salmon study areas during aerial transect surveys in the 1972-73 winter 101 4.1 The seasonal changes in forage classes eaten by moose in north-central British Columbia 124 5.1 Photographs illustrating the methods of tagging twigs and measuring diameter at point of browsing 141 6.1 The relationship between snow depth and the length of time moose spent in beds, as indicated by relative amounts of feces and urine 163 6.2 The relationship between snow depth and the distance between bed sites and their associated shelter trees 174 xxii FIGURE Page 6.3 The relationship between snow depth and the difference in crown closure between a bed site and the forest stand in which it was located . . . . 176 6.4 An illustration of effective snow interception by the forest canopy 177 7.1 The site and station layout used to study secondary plant succession 192 7.2 Photographs illustrating selected successional stages on mesic till and lacustrine substrates 221 7.3 Some trends in the forest stand features of basal area, dominant tree height, and canopy closure in sub-boreal forest seres on till and lacustrine substrates 245 7.4 Trends in height and mass of browse and non-browse species in sub-boreal forest succession on till and lacustrine substrates 252 7.5 Percentage composition, by forage class, of phytomass in the "C" or Herb layer at your successional stages of the sub-boreal forest on till and lacustrine substrates 255 7.6 Percentage composition, by forage class, of the net primary production of the understory vegetation (layers B and C) at four successional stages of the sub-boreal forest on the till substrate 259 7.7 Percentage composition, by species, of the net primary production of the shrubs (layers B and C) at four successional stages of the sub-boreal forest on the till substrate 262 7.8 Photographs illustrating plant successional stages on riparian (alluvial) substrates 268 8.1 Crude protein levels in major plant species eaten by moose in sub-boreal forests 293 xxiii FIGURE Page 8.2 Comparisons of crude protein levels in willow and subalpine fir growing on similar substrates but in stands of different ages 301 8.3 Comparisons of crude protein levels in different species growing at the same sites 304 8.4 Consistency in crude protein levels between aspen and willow for three different sites 306 8.5 Lignin levels in major shrub species eaten by moose in sub-boreal forests 310 8.6 Effect of site on lignin levels in selected sub-boreal shrubs 313 9.1 Photographs showing the use of the western snow sampler and the penetrometer. . . . 338 9.2 Patterns of snow accumulation and snow melt for the Eagle, Grove and Salmon study areas, and for weather stations at Prince George and Aleza Lake 342 9.3 Some temperature and relative humidity gradients across the forest-open ecotone at the Grove study area during the 1972-73 winter 358 9.4 Wind run and snow depths across the forest-open ecotone at the Grove study area during the 1972-73 winter 361 10.1 Major factors and how they inter-relate moose population levels (modified from Houston 1968) 384 LIST OF APPENDICES Page APPENDIX A. SCIENTIFIC AND COMMON NAMES OF PLANT SPECIES RECORDED IN THE STUDY AREA 4 60 APPENDIX B. SCIENTIFIC AND COMMON -NAMES OF BIRD AND MAMMAL SPECIES MENTIONED FOR THE STUDY AREA 467 APPENDIX C. STATISTICAL DATA USED FOR THE INTRODUCTION (SECTION 1) 469 , - Table C-l. Estimated Annual Economic Value of Minerals the Omineca Mining District, 1926-1974. .... 470 • . •• Table C-2. Number and Total Area of Farms, and Number of Cattle for the Province and for the Prince George Region, 1881-1971 471 Table C-3. Annual Cut of Timber (All Species) and the Sawmills Operating in the Prince George Forest District, 1909-1975 472 Table C-4. Area of Forest Land Disturbed by Wildfires and by Logging in the Prince George Forest District, 1912-1975 473 APPENDIX D. - HABITAT USE AND SELECTION DATA (SECTION 3) 474 Table D-l. Individual Plot Data on Time and Number of Accumulated Pellet Groups Counted in- the April, 19 72 Trial Used to Determine the Pellet Group Survey Method 4 75 Table D-2. Example of the Recording Format Used for the Aerial Transects of the Intensive Study Areas, and the Type of Data Recorded 477 xxiv XXV Page Table D-3. Example of the Summary Derived from Aerial Transect Data . 479 APPENDIX E. Table D-4. Data from Pellet Group Transects for Synoptic Surveys in 19 72. Table D-5. Data from Pellet Group Transects for Synoptic Surveys in 1973. CHARACTERISTICS OF SAMPLES COLLECTED FOR THE FOOD HABITS STUDY (SECTION 4) . Table E-l. Rumen Samples: Date of Kill, Sex, Age,- and Location of Kill. . APPENDIX F. SUCCESSION DATA (FOR SECTION 7) Table F-l. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, MF2-MF7. . 491 Table F-2. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, MF8-MF13 . . 494 Table F-3. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, MF14-MF19. . 497 Table F-4. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, MF20-MF22. . 500 Table F-5. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, SR1-SR6. . . 503 Table F-6. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, SR7-SR12 . . 506 xxvi Page Table F-7. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, SR13-SR18. . 509 Table F-8. Percentage Canopy-Coverage/ Frequency of Occurrence Values of Major* Plant Species Recorded in the Herb Layer at the Succession Study Sites, SR19-SR23. . 512 2 Table F-9. Phytomass (g/m , Oven-Dried Basis) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Till Substrates 515 2 Table F-10. Phytomass (g/m , Oven-Dried Basis) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Lacustrine Substrates. ... 516 2 Table F-ll. Phytomass (g/m , Oven-Dried Basis) of the Shrub Layer, by Species, in Partially Logged Sub-Boreal Forests in a Mesic Environment on Till and Lacustrine Substrates 517 2 2 Table F-12. Basal Area (cm /m ) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Till Substrates 518 2 2 Table F-13. Basal Area (cm /m ) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Lacustrine Substrates 520 2 2 Table F-14. Basal Area (cm /m ) of the Shrub Layer, by Species, in Partially Logged Sub-Boreal Forests in a Mesic Environment on Till and Lacustrine Substrates 521 Table F-15. Height (cm) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Till Substrates 522 xxvii Page Table F-16. Height (cm) of the Shrub Layer, by Species, in a Sub-Boreal Forest Sere in a Mesic Environment on Lacustrine Substrates 524 Table F-17. Height (cm) of the Shrub Layer, by Species, in Partially Logged Sub-Boreal Forests in a Mesic Environment on Till and Lacustrine Substrates 525 Table F-18. Number of Stems Sampled in the Shrub Layer, by Species, in a Mesic Sub-Boreal Forest Sere on Till Substrates 526 Table F-19. Number of Stems Sampled in the Shrub Layer, by Species, in a Mesic Sub-Boreal Forest Sere on Lacustrine Substrates 527 Table F-20. Number of Stems Sampled in the Shrub Layer, by Species, in Partially Logged, Mesic Sub-Boreal Forests on Till and Lacustrine Substrates 528 Table F-21. Oven-Dried Weights of Components of Major Shrub Species in Sub-Boreal Forests 529 APPENDIX G. DATA FOR NUTRIENT CONTENTS OF SAMPLED PLANT SPECIES (SECTION 8) . . . 534 Table G-l. Crude Protein Levels (%) in Plant Samples Collected from the Prince George Study Area, April 1972 to April 1973 ... ' 535 Table G-2. Lignin Values (%) in Plant Samples Collected from the Prince George Study Area, April 1972 to April 1973 541 APPENDIX H. CLIMATIC DATA USED FOR SECTION 9 545 Table H-l. Penetrance Values (1-11 Scale) for Snow Hardness Estimates Across the South-Facing Ecotone at the Grove Area, 1973 546 xxviii Page Table H-2. Monthly Means for Temperature and Relative Humidity Across the South-Facing Forest-Burn Ecotone at the Grove Study Area, 1972-73 Winter 547 Table H-3. Wind Run (km/day) Across the South-Facing Forest-Burn Ecotone at Grove Study Area and the Exposed Burn Site at Buckhorn 548 Table H-4. Snow Depths (cm) Across the West-Facing Ecotone at the Grove Study Area, 19,72-73 Winter 548 APPENDIX I. GLOSSARY OF TERMS AND ABBREVIATIONS USED IN THE TEXT 549 C ACKNOWLE DGEMENTS It is a pleasure to acknowledge the many persons who contributed to this project. Thanks go to my committee, Drs. P. J. Bandy, F. Bunnell, I. M. Cowan, V. C. Runeckles, J. H. G. Smith and K. Sumanik, for their advice and critical review of my work. To my supervisor, Dr. V. C. "Bert" Brink, I express my gratitude for his intellect, patience, perspective, diplomacy and honesty. It was an enriching experience to be one of his graduate students. My study was financed, encouraged and facilitated largely by the B.C. Fish and Wildlife Branch. Fish and Wildlife personnel in the Prince George region supported and assisted me in many ways. Ken Sumanik encouraged me to study the problem, and provided me with the benefit of his experience and the wit of his insight. Milt Warren was a mine of information. Conservation Officers D. Adolph, B. Clapp, L. Cox, W. Richmond, D. Turner, G. Vincent introduced me to their districts, collected rumen samples, and acted as excellent guides and sources of information. Thanks also go to P. Brade, K. Child, K. Fujino and R. Goodlad of the Prince George office, and to many in the Victoria office for their various contributions to my study. Members of the B.C. Forest Service provided much xxix XXX useful information on forestry related affairs. I wish to acknowledge C. P. Axhorn, J. Bullen, R. Clifford, D. Gilbert, E. Lemon, J. Revel, W. Young and the ranger staffs at Prince George, Hixon, Summit Lake and Aleza Lake. I received valuable help from the Resource Analysis Branch of the Environment Ministry (formerly the B.C. Land Inventory). In particular, I thank Greg Cheeseman, Al Dawson, the late Al Luckurst, Gary Runka and Jim van Barneveld. Forest companies in the area generously provided maps, information and assistance. I wish to acknowledge Rustad Brothers Lumber Company and their forester, Don Frood; Holger Thomsen; Northwood Pulp and Timber Company Ltd.; Prince George Pulp and Paper Company Ltd.; and Weldwood of Canada Company Ltd. The following people assisted me in data-collection, both in tedious laboratory analyses and in mosquito-plagued, patience-demanding field work: Rick Bonar, Dave Dunbar, Chris Easthope, Ollie Fricke, John Kelly, Ben Koop, Margaret Larkin, Mike Masson, Willa Noble, Sharon Russell, Eric Rutt, Rankin Smith, Don Stevenson and Chris Whyte. John Kelly deserves special thanks both for his invaluable help and his friendship. Other organizations who provided help: The College of New Caledonia and the Canada Agriculture Experimental Farm for laboratory space and facilities; the Canadian xxxi Wildlife Service,, for scholarships. Other individuals I wish to acknowledge are: Les Bower, master of the Cessna 185; Richard Revel, who first described the sub-boreal spruce zone to me; Ed Telfer, whose experience, common sense and ideas offered example and stimulation; John Powell and Douglas Golding of the Canada Forestry Service, for their data and advice on forest climatology; Jim Peek, for advice on snow measurements and several aspects of moose ecology; Rod Silver, for useful discussions of what makes moose "tick"; Rick Ellis, for information and discussion on plant succession; and Ralph Ritcey, for advice and comment on my ideas and writing. For making the final copy of the thesis so readable thanks go to Barbara Smith, who did an excellent job of typing; and to Laura Friis, who competently prepared the graphs and figures. As in most field projects, many individuals and families provided a human environment which complemented and enhanced my experiences while we lived in Prince George. I wish to acknowledge in particular the following families and individuals: .the Clapps, Froods, Gagnons, Jaroschs, Manns, Pagets, John Sawitsky, Spurrs and Sumaniks. Last, but certainly not least, I wish to thank my family: the support and forebearance of my wife, Elaine, and my two children, Jenny and Stuart; in many respects this thesis is as much theirs as it is mine. To my parents, Ben and Bernice, I owe an unpayable debt. My in-laws, Len and Mabel Weston, gave freely their understanding and support. 1. INTRODUCTION 1.1 The Study Moose are elusive, solitary ungulates of the sub-boreal forests in north-central British Columbia. They have been highly successful in this comparatively harsh environment. Their success is probably due to three main characteristics. Firstly, moose are adapted to winters that are long, cold and snowy (Kelsall 1969, Kelsall and Telfer 1971). Secondly, they are browsers in an area where shrubs and trees form the major food resource, especially in winter. Finally, they are a fire- or successionally-adapted species (Geist 1971) and so can capitalize effectively on the superabundance of forage produced in the early stages of forest succession (Telfer 1974). Since frequent fires leave large portions of boreal and sub-boreal forests in these early stages (Heinselman 1973 and others), this response is a distinct advantage to the species. Their response is manifested by increasing productivity when nutritious and abundant forage becomes available (Geist 1974, Markgren 1969) . Moose are an important natural resource. Historically, in the early 1900's, meat and hides were used for food and clothing by indigenous peoples and European 1 2 settlers. Since then, the recreational value of moose has increased dramatically, especially in the north-central region. During the 1970-74 period, the Prince George area (old management Areas 20-22) provided approximately an average of 16,000 man-days of recreation and a harvest of 5,400 annually (British Columbia Fish and Wildlife Branch 1970-1974). Most recently, the non-consumptive use of moose has been recognized as a growing and important value. Additionally, moose may well provide a protein source. The technical feasibility of "game ranching" moose has been demonstrated by Knorre (197 4) and others. Hence, moose will become an increasingly valuable resource. To deal adequately with this important resource, moose management must intensify. The forests that moose colonized so successfully have altered since moose first appeared. This change is an inevitable, natural consequence of forest succession. The significant difference with future changes will be the impact of man and his activities. Of these pursuits, harvesting timber and controlling forest fires will be the most important. Since these sub-boreal forests are easily accessible and highly productive, forest-related development will inevitably be widely distributed. Since moose are also widely spread in these forests, the question arises, "What will be the effects of human activities upon moose?" The purpose of this thesis is to examine moose 3 habitat with particular reference to the effects of forest-related activities. Much is known generally about moose and their habitats (Bedard et al. 1974). However, little is known about 1) moose habitat in the north-central region of British Columbia, or 2) the impact of forest practices upon this habitat. These gaps are critical links in achieving integrated management of forests and moose. Clearly, a habitat-oriented study does not deal with all those components that affect moose populations. It is therefore useful to place such a study in context. Variations in moose populations are determined by two broad types of mechanisms (Houston 1968) - environmental and population. The interface or linkage between them is the energy and nutrient supply (including water) available to moose. The available supply is determined largely by environmental mechanisms, although densities and behaviour of moose can obviously modify availability as well. The focus of the present study is on some of the major environmental mechanisms that determine the available supply of energy and nutrients to moose. This is shown schematically in Figure 1.1. The method by which I studied these mechanisms is reflected in the organization of this thesis. Four topics make up the first part of the thesis, viz., habitat use and selection, diet, level of use and bedding behaviour. In the ultimate sense, the availability of energy and nutrients to 3a Figure 1.1 The general relationship of sub-models that comprise a moose-forest model. Derived from Haagenrud and Hjeljord (1976) and Houston (1968). 4 WILDLIFE MANAGEMENT DECISIONS POPULATION MODEL CHANGES IN NATALITY, MORTALITY AND BODY WEIGHT. YEARLY HARVEST HERD STRUCTURE, DENSITY AND NUMBER OF ANIMALS BEHAVIOUR MODEL hACTIVITY ENERGY MODEL 7 SATISFAC TION COVER GRAZ ING HABIT FOOD REQUIREMENT FOREST/PATTERN/ PLANT SUCCES SION MODEL 5T • FOOD ["AVAILABLE .GRAZING QUANTITY GRAZING/ BROWSING MODEL FOREST MANAGEMENT DECISIONS TIMBER HARVEST 5 moose is set by the conversion of solar energy into the chemical energy stored in plants. However, not all of the vegetation is forage for moose, and not all the forage is available to them. Moreover, vegetation also provides escape cover and shelter from the elements. Thus the first task was an attempt to perceive the sub-boreal forest environment in terms meaningful to moose. Therefore, habitats were defined by examining the use and selection of vegetation types. Moose forage was assessed by food habit studies. Browse surveys were used to estimate the proportion of the forage that was eaten. Finally, since the availability of shelter affects the availability of food (Bunnell 1974), I also examined the selection of bed sites. These four components make up the first part of the thesis. Once these biological features were defined, I examined three major factors that modified them. These three topics comprise the second part of the thesis, viz., forest succession, forage nutritive values, and climate. Succession was studied since it determines long term trends in forage production and the provision of cover. Nutritive aspects of moose forages were quantified to determine the relations between species eaten and nutrient content, and to assess what factors influenced nutrient levels. Climate was examined, especially snow, as it influences the occupancy of winter ranges, the availability of forage, and the selection and use of habitats. These three topics that comprise the 6 second part of the thesis are also modified by forest practices. The third part of the thesis, the Discussion, attempts to combine the first two parts, to assess the effects of forestry activities on moose, and to provide an overview and recommendations for wildlife managers. 1.2 The Approach to the Study The present management of harvestable wildlife in British Columbia has emphasized population dynamics rather than habitat relationships. Until quite recently, wildlife managers regulated harvest through manipulation of hunting seasons and bag limits. Data were collected from game and hunter checks, harvest and hunter questionnaire surveys, and carry-over and post-season counts. Little attempt was made to modify the production of wildlife through habitat manipulation, although the importance of habitat was well recognized, e.g., Smith (1955). In the 1960's, the tacit recognition of habitat's role evolved to an active pursuit of habitat management. Efficient forest fire control and expanded logging were the two prime factors inducing the development of moose habitat management programs in north-central British Columbia. The increasing efficiency of protecting sub-boreal forests from wildfire vitiated the main agent for creating early successional ranges. Realizing the relationship between 7 high moose densities and early successional stages, the adverse impact of fire control on moose habitats became readily apparent. The rapidly increasing acreage of forests that were logged was also apparent. Thus, the pristine, uncontrolled agent of range creation was progressively being suppressed and replaced by an agent whose impact was controlled by human design. Although creation or modification of moose habitat was ancillary to forestry, biologists realized that habitat management for moose was readily feasible. A passive regard for vegetational changes was transformed to an active interest in modifying forests by logging for moose production. The broadened perspective of wildlife management from primarily animals-only to animals-and-their-habitat was a major development. The growing awareness of opportunities for habitat management were accompanied by the realization that moose habitat relationships in sub-boreal forests were largely unknown, except in the most general fashion. Existing information with respect to logged habitats was inadequate or poorly understood. Previous moose studies in British Columbia were neither within the sub-boreal forests nor dealt with other aspects of moose ecology (Hatter 1950, Cowan et al. 1950, Baynes 1956, Ritcey and Verbeek 1969, Finnegan 1973). Although they provided useful background information, the major problem of habitat relationships in north-central 8 forests was not addressed. Previous studies of moose habitats elsewhere in circumboreal ranges were also useful as background information (e.g., Bergerud and Manuel 1968, Houston 1968, Lykke 1964, Peek 1971, Peterson 1955,,Pimlott 1961, Stevens 1970, Telfer 1967). However, despite rather extensive studies of forested habitats, logged areas have received surprisingly little attention. Most studies have dealt with natural areas, or, where logging had occurred, it received almost incidental attention (e.g., Stevens 1970). Even in studies carried out primarily on cutovers, e.g., Bergerud and Manuel (1968), little information on relative use of different types and ages of cutovers has been provided. Studies such as those recently published by Peek et al. (1976)are exceptions to the rule. 1.3 The Need for Integrated Management Until quite recently, supplies of the common property resources exceeded demands. Low prices for land, grazing allotments, cutting rights, and hunting licences, all reflected the apparent superabundance of resources. In fact, considerable effort was expended by government to encourage settlement and development. Two concomitant pressures radically altered this situation. Human population increased at an exponential rate, increasing demands for goods and services which in turn 9 accelerated demands for natural resources. Many citizens experienced an increasing amount of leisure time resulting from wealth accrued through natural resource developments in the region. These growing demands for a variety of goods and services, confronted one basic limit: a finite land base. As demands approached the limits imposed by biophysical parameters, they inevitably led to conflict and competition. The interest in multiple and integrated use was an attempt to resolve these conflicts in a rational manner. Whether this attempt will be successful or not is still unknown. It can be assumed that an integrated system of natural resource allocation and management is a prime option available to the public. However, in the present system, important demands are usually made by the producers and users of these resources. In simplistic terms, these can be stated as: defining objectives, devising a planning process, inventorying of resources, implementation of plans followed by evaluation and re-adjustment. The objectives of identified user groups must also be defined in operational terms so that they can be evaluated, integrated and initiated within the limits of available resources, time and technology. Effective integrated forest and wildlife management must be dynamic rather than static. The need for continual re-assessment and evaluation of programs is obvious. 10 Competing demands for limited resources and their integration has important ecological implications. First, the option of "letting nature take its course" may appear less tenable because natural catastrophes such as wildfire and insect outbreaks may seriously disrupt the production of desired and needed forest products. The apparent impact of natural catastrophes on the economy may be directly related to the degree of resource commitment. That society has accepted this relationship is clearly demonstrated by commitments to forest fire suppression and insect control programs. However, it is worth noting that economic factors may be as catastrophic as natural ones (K. Sumanik, pers. comm.). A second ecological implication is that human beings impose novel means of re-distributing and using resources by altering natural patterns and processes. In forestry, wood products with their nutrient and energy content are mechanically removed and injected into eco systems often far removed from their source. In wildlife and fisheries management, vertebrate species are extracted by unnatural methods and transported to other systems. None of these implications is necessarily deleterious to the producing systems, although some logging methods may cause nutrient depletion on poor sites. Since we have great expectations from natural systems, and since we limit or at least modify their functioning, it is 11 obvious that we should understand how they work in order to manage them better. We should realize what effects our activities have upon their continued ability to produce what we require. 1.4 A Land Use Perspective  1.4.1 General introduction The purpose of this section is to relate human settlement and industrial development in the general study area, to moose and their habitat. Although the main subject of this thesis is with forestry-related impacts, this land use cannot be considered in isolation from others. A brief history of land use also clearly identifies changes in rates of land use activities: that the large scale of modifying natural systems is clearly a recent phenomenon. Morice (1905) related that from time immemorial, wildlife of north-central British Columbia "have been trapped or chased by the American representatives of the human species who call themselves Bene (men). . . ." (Morice 1905:4). From Morice's description, Carrier subsisted by hunting and fishing and depended especially upon salmon (Oneorkynehus spp. ) . They were semi-nomadic, shifting winter quarters to meet fuel needs, moving to lakes in spring for fishing, and camping at suitable salmon fishing sites in late summer and early fall (Morice 1905). Morice made no mention of the Carriers starting forest fires although he 12 stated that "black pine is fairly common all over the country. . ." (Morice 1905:2). Based on these accounts, which date back to approximately 1600, the indigenous people probably had only a minor influence upon the land and resident wildlife,.at least until the beginnings of the fur trade in the early 1800's. Furbearers were the initial reason for exploration and settlement in the study area by Europeans. The first European explorer was Alexander MacKenzie in 1793. Simon Fraser followed in 18 0 6 when he and his companions established Fort St. James, and in 18 07 when he founded Fort George (now Prince George). Fur trading continued to be the economic mainstay of the area until development of the Grand Trunk Pacific Railway in 1915. Placer mining was an early industry, but it involved primarily individual prospectors. The impact of the small number of fur-traders and placer miners on moose and moose habitat was probably inconsequential when compared with later activities. Moose were uncommon, and the need for land-clearing was minimal. The rate of tree cutting must have been slow since, for example, lumber for the Hudson Bay forts was all hand sawn. The most likely impact was an increased incidence of forest fires, although evidence on this point is scanty. The first major burst of activity resulted from the survey and construction of the Grand Trunk Pacific Railway. The era surrounding this development (1910-1916) witnessed a land boom spurred by anticipation of great expansion once the railway was completed (Kelly and Farstad 1946). By about 1915, approximately 84,000 ha of land were alienated in the general area, although more than 50 percent subsequently reverted to the Crown. This high reversion rate was attributed to economic conditions after the completion of the Grand Trunk Pacific rail line, failure of the Pacific Great Eastern to reach Prince George, and the departure of many settlers to World War I (Kelly and Farstad 1946). An interesting personal recollection of this colorful period can be found in Walker (1972). Also, it was during this time that agriculture and forestry began, primarily in response to demands associated with railway construction and by the local populace. Towns and villages were slow to develop. Most of them were sited at historical forts, or close to forestry and agricultural activities. Conflicts between settlement and moose were probably minimal at that time since both settlers and moose were uncommon and sparsely distributed. The post-World War II era saw tremendous expansion in population and industry. A useful index of this development is the growth of Prince George (Table 1.1). The expansion was accompanied by rapid development of transportation and utility corridors. These changes have had a profound impact on moose not only through loss and disturbance of habitat, but also by increasing accessibility of moose herds to 14 Table .1.1 Population Growth and Future Projections for Prince George and the Surrounding District Recorded Population Census Prince George City** Population Year District 8* did bdy new bdy Projection*** pre-1793 >12,000+ >350+ 1807 >12,000+ >350+ 1911 ca. 2,000++ c700 1915 >5,500 c3,500 1921 17,631 2,053 1931 21,534 2,479 1941 25,276 2,027 1951 40,276 4,703 1956 60,067 10,563 1961 74,240 13,877 32,268 1966 103,767 24,471 51,671 1971 128,205 33,101 49,365 64,365 1976 58,292 86,677 1981 111,279 1986 135,874 1991 160,030 1996 184,854 *Dominion Bureau of 186,440 km.2 Statistics 1 Census District 8 covers **City 1968, boundaries were extended in 1961, 1964, 1965, 1967, 1970 and most recently in 1975. ***From B.C. Research (1974). Study area boundaries included Prince George and the immediate locality. +Hudson Bay Company made these estimates in 1856. ++Area population based on Land Recording District. 15 hunters from the large population centres in southern and central British Columbia. Currently, the study area is traversed by two major provincial highways, a myriad of secondary and logging roads, two major railway lines, one electrical transmission line, and one natural gas line. Prime and critical moose habitat in the Finlay and Parsnip Rivers has been flooded and destroyed by the W. A. C. Bennett Dam and Williston Reservoir. Flooding of virtually the entire McGregor River and many of its tributaries is under active consideration by government. Dams proposed for the upper Fraser River would be devastating for moose. The Kenney Dam at the headwaters of the Nechako River has so controlled and modified the natural hydrological regime of the river that its capability as a "moose river" is likely reduced. It has also pre-empted highly productive moose habitat. Major development events are listed briefly in Table 1.2. 1.4.2 Mining As in most other regions of British Columbia, mining in the Prince George district has had an erratic history. Major developments began in 1861, when placer miners came northwards from the Cariboo placer gold fields to explore the Finlay and Parsnip Rivers. Eight years later, gold was discovered in the Manson, Germansen and Omineca Rivers and their tributaries. The following gold rush was hectic, and 16 Table 1.2 Major Events in the Settlement and Growth of Prince George and the Surrounding Region Year Event pre'^17 93 Indian village ("Lheitli") at confluence of Nechako and Fraser Rivers 1793 First recorded exploration by Europeans (A. MacKenzie) 1807 Fort George established 18 21 Merger of HBC and North West Fur companies 1861 Placer gold mining on Finlay and Parsnip Rivers 18 69 Placer gold mining on Manson, Germansen and Omineca Rivers and their tributaries 19 09 First sawmill opened in Prince George 1914 Completion of Grand Trunk Pacific Railway, W.W.I began. 1915 Prince George incorporated 1918 W.W.I, ended 1952 Kenney Dam completed. John Hart highway completed between Prince George and Dawson Creek. B.C.R. line completed from Quesnel to Prince George. 1956 Vancouver to Squamish link of B.C.R. completed 1958 Prince George to Dawson Creek, and to Ft. St. John sections of B.C.R. completed 1965 Yellowhead highway between Prince George and McBride completed 1967 B.C.R. extension to Ft. St. James 1968 Three pulp mills opened in Prince George. W. A. C. Bennett power project on Peace River completed. 1975 Agricultural land reserve established around Prince George o 1.7 it declined quickly after 1875. Mining activity was quiet during the next 50 years until in the 1930's, when the placer deposits northwest of Prince George were developed. Again, the flurry of activity was brief. Today, an old mining road can still be seen on the Salmon River study area, and until 1974, a large sluicing device was at the old E.M.K. mill site, approximately 60 km NE of Prince George. The outbreak of World War II strengthened mineral prices and stimulated exploration and development of many lode metals. After the war, activity subsided once more, but the recent, dramatic increases in the international gold prices again spurred renewed interest in placer mining on many streams and rivers in the Prince George area, especially to the southeast. In general, these activities are undertaken by small operators. These operations can adversely affect fish and wildlife through siltation and habitat destruction. Currently, the only mine operating within the Prince George area (Fraser-Ft. George Regional District) is a small limestone quarry west of Prince George. The city, however, plays an important role as a supply and service center for much of the active exploration to the north and northwest. The economic value of mines in the Omineca Mining District from 1926 to 1974 is tabulated in Appendix Table C-2. Future development possibilities appear limited, except possibly for coal (I.P.A. 1976). A non-coking coal 18 deposit, estimated at 73 x 106 t, lies on the Bowron River, approximately 60 km southeast of Prince George. Development of this deposit would see increased access, possibly a rail spur connecting to the BCR or the CNR, and direct employment of 1,000 people (I.P.A. 1976). Construction of a proposed steel mill at Prince George would also increase Prince George's population as well as increase likelihood of coal extraction from the Bowron Valley. The most important impact of mining and associated enterprises on moose relates to increased access and increased hunting. Past and potential mining activities have disturbed relatively little habitat, except some riparian habitats along placer streams. The roads leading to mines, and especially the increasing number of people at Prince George associated with supplying and servicing, have a much greater impact on moose. 1.4.3 Agriculture Agricultural development has affected moose primarily through conversion of valuable habitats in lowland areas into farmland. The first record of farming was a garden planted by D. W. Harmon in 1811 at Fort St. James (Runnalls 1946). Most initial attempts at land clearing were unsuccessful and their reversion was beneficial to moose. Kelley and Farstad (1946) noted that most early attempts to farm were small (2 to 8 ha) and scattered, with a high rate of abandonment. This pattern likely benefited moose by creating small patches of early successional vegetation in a primarily forested area. From after the railway boom until the depression, farming was in eclipse, being unable to compete with the more southern producers who had better soils and climate. At this time, many farmers turned to logging and road building (Kelley and Farstad 1946). With the reduction of these latter occupations during the depression, interest in farming and land clearing returned, and Kelley and Farstad (1946) believed that depression farming played an important role in developing agriculture. Similar to growth in settlement, agriculture expanded in the post-war period. Initiation of this phase of agriculture also represented the beginnings of incompatability between moose and farming and ranching. Large land areas were logged or burned or both, and put to the plow. Most of this land was at low elevation, e.g., along river valleys and in the glacio-lacustrine sediments below 760 m in elevation. While these areas represented the highest capability agriculture lands in the region, they also represented high capability winter moose habitats. The rate of expansion is presented in Table 1.3. By the late 1960's, agriculture was still largely in the development stage and was continuing to increase its productivity annually (Oswell 1969). 20 Table 1.3 Number and Area of Farms, Cattle for the Province Prince George Region, and Numbers of and for the 1881-1971 No. of farms/km2* Area of farms/km2* No. of ca ttie/km2* Year provincial regional provincial regional provincial regional 1881 .003 .002 .086 1891 .007 .136 1901 .007 .007 .134 1911 .018 .011 .001( 1%) ' .150 1921 .024 .012 .031( 1%) .234 (46%) 1931 .028 .011(8%) .015 .012(16%) .251 (12%) 1941 .028 .012(9%) .018 .013 .359 .228(13%) 1951 .028 .020 1956 .027 .020 1961 .021 .007(7%) .020 .014(15%) 1966 .021 .008(7%) .020 1971 .020 .012 .616 .223( 2%) *Data expressed on a per km2 basis since the sample area changed during sample time. The regional census from 1881-1931 was the Cariboo electoral district (81,417 km2); from 1941-1966, Census District 8 (186,440 km2); and in 1971, the Fraser-Fort George Regional District (51,196 km2). Area of the province is 930,528 km2. Original data in Appendix Table C-2. **Regional totals as proportion (in percent) of provincial totals in parenthesis. 21 Now, as in the past, the mainstay of regional agriculture is livestock and associated forage production. The major areas around Prince George are: a) Reid, Chief and Nukko Lakes b) the lower Salmon River Valley, east of Highway 97 c) south of Prince George to Woodpecker, primarily on the east side of the Fraser River d) scattered areas in the Willow River - Aleza Lake vicinity e) west of Prince George along Highway 97. The Pineview clay soils and alluvial materials are used most commonly for agricultural purposes. Despite optimistic expectations for agriculture, economic and biophysical characteristics foretell an uncertain future. The Pineview soils require expensive clearing and are difficult to manage: they are slow to heat, puddle when wet and pack when dry. The alluvial terraces in bottomlands are fragmented, leading to a "pocket" agriculture where access and clearing costs are high. Except for microclimatic variations along the major drainages, the growing season is cool, with frequent rainfall; crop maturity is often delayed by these factors. Crop alternatives are limited. The long and cold winters necessitate extended periods of winter feeding. Distance to markets, unstable prices for products and more productive lands elsewhere also add to the difficulties facing 22 agricultural development in the region. Other agricultural enterprises include sheep ranching, and vegetable, egg, and limited horticultural production. However, these make a minor contribution to the overall regional agriculture compared with' the beef industry. Steps by government to increase and promote agriculture in the region are many and varied. Community pastures are sponsored and partly subsidized by both provincial and federal governments. In one case at Giscome, a community pasture is developing on an important moose winter range, where land capabilities clearly favor timber and wildlife. Agricultural land surrounding Prince George was protected by an Agricultural Land Reserve (ALR) through legislation approved in 1974. These incentives are obviously encouraged by the agricultural sector and others. The need to preserve arable land in this, and elsewhere in the province, is an obvious one that meets with general public approval. A growing urban center like Prince George undoubtedly can support a local agriculture, given solution to problems such as increasing egg production quotas for local producers. Also, there is a desire to preserve ah agricultural lifestyle although the costs of this needs to be carefully evaluated. This desire assures the continued presence and development of agriculture. It is also obvious that present-day agriculture and moose and other wildlife are less than compatible. Land 23 alienation, clearing and maintenance of forage-producing areas directly remove potential moose habitat. Other items such as predator control in outlying, marginal farms pose other difficult problems. These lands often have high capabilities for moose. The disturbing aspect of these conflicting forms of land use is not only the loss of habitat, but also the apparent lack of regional objectives to give purpose and direction to proper land use, and to provide acceptable land use zoning. 1.4.4 Forestry Although furs and, to a lesser extent, minerals provided the initial impetus to development and settlement of the Prince George area, timber harvesting has been the main stay of the economy since the turn of the century. Its impact on the natural systems and human settlements of the region have been complex and extensive. Perhaps the most remarkable feature has been the rapid transformation from very simple operations in the early 1900"s to a sophisticated, integrated industrial complex in the 1970's. Prior to 1909, all wooden buildings were either of log or hand-sawn lumber. The first sawmill opened in Prince George in November, 1909 (Walker 1972). Its appearance reflected the demand for ties used in the construction of the Grand Trunk Railway, and the concomitant demand for building materials by settlers during the 1910-1916 land boom. By 1910 there were three sawmills. 24 In 1912 the Forest Branch was set up and two years later the first Crown timber auction was held in the district (Glew 1963). Talk of building pulp mills was also current during this early period (Runnalls 1946), but these proposals were not realized until more than 50 years later. Secondary wood processing began in 1912 when the first sash and door mill opened in Prince George. Whitford and Craig (1918) summarized the restricted extent of lumber operations at that time. One large mill was sited at the mouth of the Willow River, and another small one was located at Giscome. Logging was not extensive, being confined to forests adjacent to the mill sites. Virtually no logging occurred in the drainages of the Stuart, Salmon, Nation, Parsnip, Nechako and Blackwater. Lumbering activities in the Upper Fraser area were "not extensive" (Whitford and Craig 1918). By 1920, 20 sawmills were operating in the district with an annual cut of 97,000 cunits (Figure 1.2, see also Appendix Table C-3). This early phase of harvesting was typically done by horse-logging. It had only a limited impact on the forest ecosystem. Impact on moose habitat was likely beneficial. Forest management was minimal although interest in fire control was growing. Harvesting was basically an extraction activity with little regard to the future. Growth of the industry was slow and slightly erratic until the end of World War II. 24a Figure 1.2 Development of forestry in the Prince George Forest District as indicated by the annual cut and the number of operating sawmills, 1914-1974. Records for the sawmills were discontinued after 1968. Source of data: annual reports of the British Columbia Forest Service. ANNUAL CUT (CUNITS*104) o o O O o o o 00 o o NUMBER OF SAWMILLS OPFRATiMn 26 The post-war era witnessed tremendous expansion (Figure 1.2). This was made possible by a rapid transition from horse to mechanized logging. Market demands increased as well. Evidence of the rapid growth is given by the total annual cut which almost doubled from 1945 to 1972 when it was 252,000 cunits and 407,500 cunits, respectively. Silvicultural systems evolved rapidly in response to increased concern for ensuring sustained yields. Earlier logging was primarily diameter limit or commercial clear-cutting, with large amounts of the original stand being left. In 1951, single tree selection became operational (Glew 1963). Alternate strip cutting was instituted in 1954 and scarification, as a stand treatment technique, was introduced in 1956 (Glew 1963). The dominance of the partial cutting declined by the early 1960's and was replaced by clearcutting (Table 1.4). Accompanying this transition was the institution of harvesting to a close utilization standard. Although the area logged and the annual cut increased, more efficient utilization of wood meant that these increases were not parallel: the annual area cut grew more slowly than the amount of wood extracted (cf. Figure 1.1 and Table 1.4). The last twenty years also saw the end of the bush-mill where small sawmills were situated near the logging operation. Wood processing became highly integrated and 27 Table 1.4 Trends in Logging Methods and Area Cut in the Prince George Forest District, 1950 - 1973* Area logged (km2) Prop. (%) of selective methods Total area Year clearcut partial seed tree dia.limit single tree (km2) 1950 29 122(81%)** 5 78 16 151 1951 25 146(85%) 12 69 18 171 1952 21 172(89%) 14 66 20 193 1953 m*** m 22 67 11 m 1954 mm 25 63 11 m 1955 mm 40 53 7 m 1956 m m m 1957 21 225(91%) 246 1958 34 209(86%) 243 1959 103 226(76%) 329 1960 - 131 194(62%) 315 1961 118 144(55%) 262 1962 182 145(44%) 327 1963 m m m 1964 m m m 1965 m m m 1966 368+ 368+ 1967 335+ 335+ 1968 413+ 413+ 1969 485+ 485+ 1970 442+ 442+ 1971 414 5(1%) 419 1972 368 4(1%) 373 1973 386 t(l%)++ 387 *Source of data: annual reports of the B.C. Forest Service. - **Proportion of total area cut by partial cutting methods. ***Data missing. +Type of logging (clearcut or partial) not available, but likely >99% clearcut. ++t is less than 0.5%. 23 centralized, especially in Prince George. Three pulp mills began operating between 1965 - 1970. The sociological, economic and environmental impact of these developments were far-reaching. They will clearly shape the future character and nature of regional development. In future, regional forestry practices will demand a still more intensive use of timber since uncommitted wood supplies are very limited (I.P.A. 1976). The Central Report 76 (I.P.A. 1976) listed five probable developments that are paraphrased as follows: 1. A change in the proportions of various uses of forest lands. 2. A refinement of logging practices that will reduce waste and breakage in harvesting and transporting. 3. An improvement in wood processing technology. 4. An improvement in the distribution of raw materials. 5. An upgrading of products and greater specialization of wood products. ". . . the exact scale, timing and feasibility . . . will remain to be determined by actual market conditions and site (project) specific feasibility studies in the future" (I.P.A. 1976:40-41). The foregoing changes have had major impacts on the habitat and production of moose. Both the rate and nature of logging created and impaired moose habitat. Logging also produced novel types of habitat in the sense that they had no 29 natural counterparts. Murray (19 74) examined the impact of logging on harvesting moose through improved access. 1.4.5 Wildfire Wildfire is the single most significant natural factor that modified moose habitat. Even prior to the arrival of white man and moose in north-central British Columbia, circumstantial evidence indicates that fire was a major ecological force. Morice (1905) remarked that black (lodgepole) pine was fairly common in the area. His observations were made oivoa 1860, before the major influx of Europeans. The region ranks as high to very high on the Canadian Forest Fire Weather Index maps (Simard 1973). Although fire history reports, are lacking for the general study area, palynological studies by Rouse (1973, cited by Smith 1974a) in the south-central Interior revealed pollen grains of lodgepole pine and charcoal layers that predate ash layers deposited after Oregon's Mt. Mazama erupted, approximately 7,000 years ago. Detailed fire research in other boreal forest ecosystems attest to the integral and long-standing role of fire in northern forests (Heinselman 1973, and a review by Kelsall et al. 1977). As the number of Europeans increased, it is likely that the number of fires and area burned increased. The magnitude of this increase is difficult to define without detailed fire history research but again, cursory evidence 30 suggests it was large. n The arrival of moose in north-central B.C. and their remarkable southward dispersal from 1900-1950 has been attributed to increased fire activity associated with white settlement (Hatter 1950). Dawson (1879) reported extensive burns between Stuart and MacLeod Lakes. Whitford and Craig (1918) provided additional data that afford an appreciation of wildfire impact on sub-boreal forests (Table 1.5). In their classification of forest land vegetation, 53 percent of the major drainages (range of 14-70 percent) in the north-central district was "young growth!" I estimated these stands to be less than 40 years old. Thus the forests were burned, on an average, at 75-100 year intervals. Early annual reports of the Forest Branch refer to the high incidence of fires along the Grand Trunk Railway, and the need for fire patrols along the right of way. Campbell (1920) related that at least two large fires burned from the Nechako River north approximately 3 0 km almost to Summit Lake, and westward from the Fraser River. The first fire occurred 60-70 years prior to 1919, and the second in 1902 or 1903. The latter fire was likely very intense since Campbell remarked that subsoil was exposed. There was have been many small fires as well as the very large and spectacular ones. These types of reports suggest that the period from approximately 1850-1920 was one of unusually high levels of wildfire. 31 Table 1.5 Estimated Areas of Broad Vegetation Classes in Five Major Drainages in North-Central British Columbia (from Whitford and Craig 1918) Proportions of region in various vegetation classes (%) Drainage region Estimated area (km2) young growth merchantable timber incapable of timber growth above timberline upper Fraser 19,661 14 27 2 57 Willow -Bowron 8,156 27 58 4 11 Parsnip 11,634 35 38 4 23 Stuart -Salmon -Nation 27,871 59 19 15 6 Nechako -Blackwater 58,350 70 14 8 8 All regions 125,672 53 22 8 17 Since1 1920r wildfires have been more carefully and systematically monitored by the B.C. Forest Service (Figure 1.3). The data indicate that for most years relatively small areas were burned. However, when ideal fire conditions develop, large tracts burn despite fire fighting efforts. Large areas were burned in 1922, 1942, 1944, 1948, 1956, 1961 and 1971 (Figure 1.3, Table C-4). Perhaps the most important point relevant to moose habitat is the diminishing natural role of fire as an agent 31a Figure 1.3 Annual area (km2) burned by wildfire in the Prince George Forest District, 1910-1975. Source of data: records from the Protection Division, British Columbia Forest Service. of habitat creation. This change has long-term consequences on a fire-adapted species such as moose (Geist 1971, 1974) since the typical "boom-and^bust" cycle of pristine environments will become a progressively rarer event. Moose densities will continue to vary but probably with much reduced amplitude. 2. THE STUDY AREAS 2.1 Biophysical Setting The general study area is situated in central British Columbia, about 8 00 km north of Vancouver (Figure 2.1). It covers approximately 11,300 km2 roughly defined by a circle centered on Prince George, with a radius of 60 km. Most of this area was not studied in detail, but the sites studied were selected to allow generalization to the large area. Elevational range of the study area was 550 to 1220 m, with most of it 915 ± 150 m. The landscape of the study area is highly modified by glaciation and associated events such as proglacial lakes. The final glaciation was so "intense" that evidence of previous glaciations is scarce (Kelley and Farstad 1946, Tipper 1971) Physiographically, the study area can be classified as part of the Fraser Basin of Holland (1964), or the Nechako Plateau of Tipper (1971). The latter author believed the Fraser Basin distinction is an arbitrary one, and that this Basin merely represents the lowest part of the erosion surface of the Nechako Plateau (Tipper 1971:10). Regardless of this divergence of opinion, the physiography can be described as: Its flat or gently rolling surface lies for the 34 34a Figure 2.1 Locations of the study areas, and of place names mentioned in the text. 10 0 I I—I t—I I STUDY AREAS lp ^0 B PRIMARY 1 Eagle 2 Grove 3 Salmon 4 Bowron 5 Found 6 Pyfe 7 Limestone 8 McGregor Major highway KILOMETRES 5» SECONDARY 9 McKenzie 10 Pineview n Shell 12 Swamp 13 Teardrop 14 Telachick 15 Torpy 16 Whites Secondary roads 36 most part below 3,000 ft. and is covered with drift and has few exposures of bedrock. On much of the surface the drainage is poorly organized, and numerous lakes and poorly drained depressions are present. The area was occupied by ice whose movement created drumlins and drumlin-like forms in the glacial drift ... an eastward and northeastward movement of ice. (Holland 1964:67). As the Cordilleran ice sheet melted, natural drainage channels in the Nechako Plateau were blocked by ice or till. Three large lakes subsequently formed around Prince George, Vanderhoof and Fort St. James (Tipper 1971). Approximately 3035 km2 of the lower lying drumlinized till was covered by lacustrine sediments - composed typically of varved silts, clays and sands. The Prince George pro-glacial lake was caused by ice blockage of the Fraser River channel south of Prince George. Its level, and therefore the level of lacustrine deposits was approximately 790 m. This level was determined by a bedrock lip at Summit Lake, where the proglacial Fraser River flowed northward into the Peace River system via the Crooked and Parsnip Rivers. (Today it flows southward.) These lacustrine deposits have been collectively termed the Nechako Plain (Armstrong and Tipper 1948:285, quoted in Tipper 1971). Their depth varies from over 120 m to thin overlays (0.3 m) that cap still obvious drumlins: outliers of the Nechako Plateau also are found within the Nechako Plain. Macroclimate of the study area is characterized by abrupt seasonal changes from cold, snowy winters, to short 37 cool summers without a distinct dry season. As such it corresponds to the "Dfc" climate type of Koppen, or a microthermal continental sub-boreal type (Krajina 1965). Chapman (1955) distinguished a Dfb climate in the valleys of the Fraser and lower Nechako and Bowron Rivers, that is, a cool summer with at least four months above 10°C. This distinction presumably reflects the lower elevation and therefore warmer temperatures of these valley basins. Climatic data for the Prince George airport are typical of the study area (British Columbia Department of Agriculture 1976a)(Figure 2.2). Mean annual temperature is 3.3°C. Annual mean daily minimums and maximums are 2.5°C and 9°C, respectively, with extremes of -50°C recorded in January and 34.4°C in July. Precipitation averages 621 mm of which 400 mm (63 percent) falls as rain. Rain falls in all months of the year and snow has been recorded for every month except July and August. However, precipitation is not evenly distributed in all months (Figure 2.2). The null hypothesis of no difference in precipitation between months was rejected at P < 0.01 (x2 = 100.38, df = 11). The annual meteorological summary for Prince George provides a ready reference to climatic records (e.g., Anon 1973). Kelley and Farstad (1946) noted the striking variability of the local climate. Strong temperature contrasts occur as do abrupt changes in cloudiness and sequences of wet and dry weather. This variation is prob-37a Figure 2.2 Long term monthly averages of some temperature and precipitation parameters for the Prince George weather station. Source of data: British Columbia Department of Agriculture (1976a). Length of record: 30 years (1941-1970). 38 600 J JAN FEB MAR APR MAY JUNE JULY AUG SEP OCT NOV DEC 39 ably due to interaction of moist Pacific air from the west and cool, dry Polar air from the north and northeast. Major determinants of this climatic regime are latitude, elevation, continental location, and the interaction of warm, moist Pacific air and cool, dry Polar air. The relatively even topography results in noticeable horizontal climatic gradients rather than the marked vertical gradients typical for much of British Columbia. The upland areas to the east obviously are an important determinant of these.gradients, although other factors also operate. Of major importance to moose are the gradients of precipitation and cold temperature since they influence critical elements of winter climate, viz., snow deposition, characteristics and melt. Conventional climate records are not available for these parameters so the gradients must be described in terms of indicators such as snow fall and mean temperatures. A trend towards cooler temperatures to the north is evident and possibly to the west (Table 2.1). Quesnel has a mean daily temperature of 4.4°C, while Prince George, 110 km north, has 3.3°C; Vanderhoof and Fort St. James (50 km north), have temperature means of 2.7°C and 2.3°C respectively. The precipitation gradient is more pronounced than that for temperature (Table 2.1). Both rain and snow 40 increase from west to east and from south to north. Snowfall in Prince George is 127 percent that to the east in Vanderhoof while snowfall in Aleza Lake is 204 percent that of Vanderhoof. Prince George and Aleza Lake are 8 0 km and 130 km east of Vanderhoof, respectively. Snowfall in Prince George is 121 percent that in Quesnel to the South, and snowfall in Fort St. James is 101 percent that in Vanderhoof. Rainfall data show similar trends. Table 2.1 Climatic Parameters for the Study Area* Climatic parameter Climate daily daily snowfall station** mean (C) minimum (C) (mm) rain (mm) Vanderhoof (680) 2.7 -3.8 1,834 273(60%)*** Fort St. James (686) 2.3 -3.5 1,857 284(60%) Prince George (676) 3.3 -2.5 2,334 400(64%) Quesnel (545) 4.4 -1.8 1,928 361(65%) Aleza Lake (625) 3.1 -3.0 3,744 557(60%) *Source of data: British Columbia Department of Agriculture (1976a). **Elevation in m. ***Proportion of total precipitation falling as rain. Within the macroclimate there is an intricate matrix of microclimates or "climate in a small space" (Geiger 1966). (Microclimate is defined as local combinations of atmospheric factors which differ from macroclimate due to variations in olant cover, tocography, slope position, and proximity to lakes (Daubenmire 19.7:4).) Selected microclimates of the sub-boreal spruce biogioclimatic zone were studied in detail by Wali (1969) and Wali and Krajina (1973). Their site selections were based on differences in canopy density, ground cover, species composition, soil conditions and topography. The study area has been comparatively well examined for soils. Kelley and Farstad (1946) surveyed soils in a rectangular area centered on Prince George. Farstad and Laird (1954) extended the survey area westward towards Burns Lake, while Hortie et al. (1970) surveyed the upper Fraser River valley from Prince George eastward to approximately McBride. These reports are collated in a recent compendium (Keser et al. 1973). Ecological studies by Wali (1969) and a co-worker Revel (1972) related soils with ecosystematic units (sensu Krajina (1965)) in the sub-boreal spruce biogeoclimatic zone. Their study area included part of mine. The most recent and comprehensive soil survey was conducted jointly by the British Columbia Land Inventory and the Soil Survey Section, Canada Department of Agriculture (A. Dawson, pers. comm.). These surveys integrated soils, landforms and vegetation in a biophysical approach to land mapping, similar to two previous reports for more westerly regions of north-central British Columbia (Runka 1972, Cotic et al. 1974). Information from these surveys provided a basis for capability assessments regarding forestry, wildlife and agriculture. The following description of 42 soils draws upon Keser et al. (1973), and material currently in preparation (A. Dawson, pers. comm.). Soils of the study area mainly belong to the Luvisolic and Podzolic orders, as defined by the National Soil Survey Committee of Canada (1970). Smaller acreages of Brunisols and Regosols occur on mineral substrates, and Mesisols and Humisols occur on organic deposits. Soils integrate the factors of time, climate, relief, vegetation and substrate (Jenny 1941). Thus they can provide a meaningful way of stratifying a heterogeneous environment into homogeneous units. With the advice of A. Dawson and G. Runka, I stratified my study area into environmental units based on moisture regime and surficial deposit. From this subdivision, major units were derived to form the basis of much of the sampling in this project. The general distribution of soil associations and land forms is presented in Figure 2.3. Twenty-three units were delineated on mineral substrates and two on organic deposits (Table 2.2). Of these, the mesic units on basal till and lacustrine sub strates were selected for detailed study. They represented the largest map units in the study area, received most of the logging activity, and underlaid most of the intensively studied winter ranges. These selected environmental units plus seven others of importance to moose are described briefly below (taken from Dawson, pers. comm.). 42a Figure 2.3 Oblique aerial photographs illustrating the general terrain and vegetation of the Prince George study area. Photographs taken by K. Sumanik and M. Warren. 43 44 Table 2.2 Major Soil Associations for the Study Area and Their Relationship to Parent Materials and Moisture Regimes Soil association by moisture regime Surficial ; deposit xeric mesic hydric Lacustrine - clays Vanderhoof Pineview Bowron - silts Berman Bednesti Coarse outwash Alix, Mapes Saxton, Giscome Ramsey, Bear Roaring Peta Beach ridge Kluk Gunniza Basal till Barrett Deserters Dominion, Twain Ablation till Crystal (part) Crystal, Cobb Cobb (part) Shallow till/bedrock Pope Ormrod Cluculz, Averil Decker Dragon, Oona Soil association not ranked by moisture regime Recent alluvium McGregor and Stellako Glacio-fluvial Fraser and Nechako Organics Chief (sedge peats) Moxley (sphagum peats) The detailed features of all units occurring in the general study area can be found in Dawson (pers. comm.): 1. Deserters (Figure 2.4): mesic environment on gravelly and stony glacial till deposits of variable thickness. Medium to moderately coarse textures (gsl, si).* Predominantly a drumlinized and/or glacial grooved till plain with rolling and hilly topography, with inclusions of bedrock-controlled strong to very steep slopes. Elevation 745 - 1220 m. Six sub-group combinations delineated (Figure 2.4). Also included in this unit is a hydric subgroup of the drier Barrett soil association (BA 4), an orthic grey luvisol. This unit occurs as far westward as Smithers. The drier counterpart is the *Letters within parentheses are standard abbreviations used to describe soil texture. Abbreviations are: c - clay, f - fine, g - gravel, 1 - loam, s - sand, si - silt. 44a Figure 2.4 A schematic illustration showing the major soil associations in the study area, and their topographic relationship to each other. Derived from data provided by A. Dawson, British Columbia Ministry of Agriculture. DESERTERS- •-CRYSTAL-or COBB -ORMROD-or DECKER lb/ MOXLEY CHIEF J. BEDROCK 46 Barrett association; the wetter, Dominion and Twain. 2. Pineview (Figure 2.4): mesic environment on clayey glacial lake deposits which vary from up to 10 - 15 m in thickness to shallow deposits (less than 1.5 m) over glacial till. The deposits are silty at depth and gradually grade to fine (c, sic) and very fine (hvc) textures near the surface. Undulating to rolling topography, with the latter associated with the under lying drumlinized glacial till; moderately to steeply sloping adjacent to the major rivers. Elevational range is 610 - 790 m. Seven subgroups distinguished in the Prince George glacial lake basin. Drier counter part is the Vanderhoof association. 3. Bednesti (Figure 2.4): mesic environment on silty glacial lake deposits which vary from up to 10 - 15 m in thickness to shallow deposits (less than 1.5 m) over glacial till. Medium to moderately fine textured (sil, sicl). Undulating, rolling and hilly topography; strongly to very steeply sloping where the deposits are associated with the Stuart and Bednesti eskers; dissect ed adjacent to the main rivers and creeks; shallow to deep kettles occur at random. Elevational range is 610 - 790 m. Four subgroups identified. The drier counterpart is the Berman association, while the wetter one is Bowron. 4. Gunniza (Figure 2.4): mesic environment on gravelly, cobbly and sandy glacial lake beach deposits which are underlain by glacial till at variable depths from 0.3 to 3 m, with inclusions as deep as 4.5 to 6 m. Very coarse textured (g, gs, gls, s, Is). Gently to strongly rolling and moderately rolling topography which usually conforms to the underlying glacial till. Elevational range is from 730 to 790 m. This association occurs along the margins of the glacial lake centered at Prince George, and occurs on the Salmon intensive study area. 5. McGregor and Stellako (Figure 2.4): both units are recent silty and sandy laterally accreted fluvial (alluvial) terrace deposits. Medium to very coarse textured (sil, 1, fsl, si, Is, s); often inter-stratified; underlain by sands at shallow depths. Undulating topography. Subject to inundation during freshet and high water table seasonally. Elevational range is 550 - 760 m. These associations flank major rivers in the study area, e.g. Salmon, Fraser and Nechako Rivers. In combination with older glacial fluvial deposits described next, they occur on the critical valley bottom winter ranges of moose. 47 6. Fraser and Nechako (Figure 2.4): both units are older associations than the above. They are silty to sandy vertically accreted fluvial deposits; moderately coarse to medium (Nechako) or medium to moderately fine (Fraser) textured; underlain with coarse textures at variable depths (Fraser) or by sands at depth of less than 1.5 m. Undulating to nearly level topography. Elevational range is 550 to 670 m. (Fraser) and up to 7 60 m. (Nechako). Two subgroups are outlined for each association. 7. Chief and Moxley (Figure 2.4): two organic associations, the first representing deposits composed of sedge and associated hydrophytic vegetation; the second, with mainly sphagnum moss type of vegetation. Two subgroups distinguished for each association, one for generally land-locked map units (poor or non existent drainage) and one for those associated with creek drainage. Elevation- range is wide, from 610 to 137 0 m. Important summer habitats, especially Chief. These units, especially those representing mesic environments, formed the basis for most of the sampling strata for this study. Vegetation of the study area is primarily coniferous forest. Arlidge (in Hortie et al. (1970:14) provided the following appropriate description: The forests are composed of a mixture of spruces (white spruce, Engelmann Spruce and their inter-grades) and alpine fir with scattered white birch, occasional trembling aspen, and Douglas fir. Stumps and rotting trunks of Douglas fir suggest that this species had a greater representation in the recent past. Lodgepole pine occurs in pure stands and in varying mixtures with spruce and other species following fires. Aspen and birch also form pure stands or sands mixed with other species following fires. Black spruce and lodge-pole pine, together or separately, are found in bogs. Black cottonwood is found on alluvial bottom soils. Although most workers who have studied this forest type would agree with the above description, they have 48 named and classified it differently. Whitford and Craig (1918) classified this forest as an Engelmann spruce {Picea glauoa engelmanni) - alpine fir {Abies lasiooarpa) type, but also called it a lodgepole pine ( Firms oontorta) type. They considered that this species of spruce occurred from the lowest valleys up to 1,220 - 1,520 m. However, Krajina (1969) stated that white spruce ( Pioea g. glauoa) occurs at lower elevation, while white x Engelmann spruce hybrids occur at intermediate elevations, and the latter species at higher elevations (also Taylor (1959), cited in Krajina (1969)). Forests of the study area would border on Hare and Ritchie's (1972) "closed forest" zone of their boreal forest, and in Rowe's (1973) montane transition section (M.4) of the montane forest region. Inclusion in this region of Rowe's rather than the subalpine region was based on the scattered presence of interior Douglas fir (Pseudotsuga menziesii) (Rowe 1972:76). However, Douglas fir fails to reproduce itself except on the drier sites (e.g., beach deposits) and nutritionally rich Lithosols. Krajina (1959, 1965) defined a sub-boreal spruce zone of his Canadian boreal forest biogeoclimatic region that also encompassed the study area. Van Barneveld (in Cotic et al. 1974) offered the following criteria for separating the sub-boreal spruce zone from the similar northern subzone of the Cariboo-aspen (Populus tremuloides)- lodgepole pine/Douglas fir biogeoclimatic zone: 49 1. little or no potential for climax Douglas fir stands, 2. absence of pinegrass . {Calamagrostis vubescens) , 3. presence of subalpine fir regeneration below 915 m., 4. continuous and often relatively thick moss layer. These are meaningful criteria, based on my observations in the study area. Based on these attributes, and the descriptions by Krajina (1965), and Revel (1972), I believe the most applicable description is the sub-boreal spruce zone. Revel (1972) listed other authors who have studied the similar boreal forest, and stated that those by Moss (1953a, 1953b, 1955) were most relevant. Work by LaRoi (1967) and Annas (1977) can be added to this list. Classification of this forest zone into vegetation subunits has been attempted. Kujula (1945, not seen but cited in Wali and Krajina (1973)) identified forest types in the area. Subsequently, Illingworth and Arlidge (1960) proposed five site types for white spruce - subalpine fir {Abies lasiooarpa) stands based both on dominant and characteristic understory species. Their nine site types for lodgepole pine forests in south-central British Columbia have counterparts in the study area, too. Wali (1969) and Revel (1972) conducted synecological analyses in the sub-boreal spruce zone. Their classification follows the methods and concepts developed by Krajina (1965, 1969). Most recently, dissimilarity analysis was applied to Arlidge's original data plus new data from north-central 50 British Columbia (J. van Barneveld, pers. comm.). Using this hierarchical divisive approach, 41 final groups or vegetation types were provisionally identified. These were combined into 16 composite groupings that contained up to five vegetation types each. Although this recent work is not yet completed, the provisional units will probably provide the most useful subdivision of sub-boreal forests. The study area supports terrestrial fauna typical of the sub-alpine forest biotic area of Cowan and Guiget (1973). Mammalian fauna are very much like that of the boreal forest. The major ungulate is moose, with year-round 2 densities of approximately 0.4/km (K. Child, pers. comm.). Mule deer {Odoooileus hemionus) are limited in numbers, probably because of deep snow. Their critical winter habitat of steep, south-facing slopes with widely spaced mature Douglas fir is confined primarily to major river drainages. Deer are more numerous to the south, east and west of the area. Mountain caribou {Eangifev tarandus montanus) are transients. I observed tracks and pellet groups only on two occasions: at Trapping Lake and at Barney Creek, 3 7 km south and 51 km NNE of Prince George, respectively. Caribou winter to the east, southeast and northeast of Prince George. Winter habitat utilization by this species is currently under study approximately 90 km east of Prince George (Bloomfield 1976). Other vertebrate herbivores whose diets overlap that of moose are beaver {Castor canadensis), varying hare {Lepus americanus), porcupine {Hrethizon dorsatum), and ruffed grouse {Bonasa umbellus). During the period of field work, noticeable hare browsing was localized and. occurred mainly in riparian habitats. Porcupine damage is believed minimal. Ruffed grouse fed extensively on willow buds in the tops of trees and tall shrubs that were mostly beyond the reach of moose. Thus competition between these species and moose is considered minimal. Representative carnivores - wolf {Canis lupus), coyote (Canis latvans) , black bear {Ursus americanus) , grizzly bear (U. arctos) , Canada lynx {Lynx canadensis), bobcat (L. rufus) - are comparatively abundant; cougar {Felix concolor) are uncommon. Commercially valuable furbearers are also abundant. Kelly (1976) is studying habitat utilization by marten {Martes americana) in relation to logging within the study area. The area supports a varied but largely unstudied bird fauna. Studies by Munro (1947, 1949 and 1955) of birds and mammals for the Vanderhoof locality and surveys by Stanwell-Fletcher and Stanwell-Fletcher (1943) of the Driftwood River valley are probably applicable. Efforts by local birdwatchers represent the major current contributions to bird study. Many opportunities for ornithological research exist as the area is transitional for many species (e.g., Scott et al. 1976). Also, developments such as land clearing, and utility corridors have encouraged the northward extension of typically southerly species, e.g., sparrowhawk .(Faloo sparverius), western meadowlark ( Sturnella neglecta) and lazuli bunting (Passerina conoena). Agricultural land is used by many migrating waterfowl, especially Canada geese (Branta canadensis), but the capability ratings for waterfowl are generally very low, with the major limitations being reduced marsh edge, inappropriate water depth and low nutrient status. Recent surveys indicate that production levels for waterfowl are low (W. A. Munro, pers. comm.). Amphibians, reptiles and invertebrates are even less well studied than the avifauna, except for the spruce budworm (Choristoneura fumiferana) and several other insect species that damage commercial tree species. 2.2 The Primary Study Areas Three primary study areas were selected within the general study area described in section 2.1 (Figure 2.5). These were called Eagle, Grove and Salmon. All three are important moose winter ranges (K. Sumanik, pers. comm.). All three contained habitats typical of the general study area, including burns, immature forests, mature forests and cutovers. The Eagle winter range was located approximately 4 0 km ENE of Prince George. It covered approximately 2 3 90 km and measured 2 0 x 3 0 km at its greatest width and length, respectively. Western and northern boundaries 52a Figure 2.5 Photographs of the Eagle, Grove and Salmon winter ranges. 54 followed the Fraser River upstream from the Willow River to Mokus Creek, near the mouth of the McGregor River. The eastern boundary followed roughly the eastern edge of Ogilvie and Bearman Creek watersheds. Hay Creek, the north shore of Eaglet Lake and the Willow River, from its confluence with Hay Creek to the Fraser, completed the southern boundary of the Eagle range. The elevational range was from 590 to 950 m, with most of the area lying below 790 m. Soils are mostly silty and clayey lacustrine deposits, that is, Bednesti, Pineview and Bowron associations (Figure 2.6). Scattered beach deposits occur near 790 m. At higher elevations are small areas of glacial tills of the Deserters association and its wetter counter part, Dominion. Both recent' and glacial-fluvial deposits border the Fraser and Willow Rivers, and Hay Creek. Although few organic soils were mapped, many small pockets occur on the study area. Rock outcrops are uncommon. Lakes are small and scattered, though wetlands are plentiful. Vegetation is complex, varied and often lush. This is due to a combination of moisture, undulating topography, fertile soils, and a wide range and large area of logging and wildfire (Figure 2.7). The most noticeable feature is 2 the 90 km Eaglet Lake burn. The area was apparently burned twice—oa 1932 and oa 1937—.and coniferous regeneration is still scattered. Regeneration failure 54a Figure 2.6 A soil association map of the Eagle study area. Derived from maps provided by A. Dawson, British Columbia Ministry of Agriculture. 55a Figure 2.7 A forest cover map of the Eagle study area. Derived from forest cover-type maps of the Inventory Division, British Columbia Forest Service. attests to what must have been severe fires. Mature spruce-subalpine fir forests cover the northern one-third of the area. The remainder is cutover land. Among the cutting practices represented are diameter-limit and single tree selection cutting, cut and leave strips, clearcut and burn, clearcut with pre- and post-scarification and seed block logging. A small black cottonwood {Populus balsamifera tviehoeavpa) cutover occurs near the confluence of the Willow and Fraser Rivers. Some of the earliest logging in the Prince George Forest District occurred here, and it is one of the few places in central British Columbia that was rail-logged. The Eagle contained a wide variety of moose habitats. It supported both resident and migrant herds, with at least one reported migration route from the McGregor Plateau, and the watersheds of Averil, Limestone, and Olsson Creeks, southward across the Fraser River onto the study area. Typical herd structure and a relative estimate of abundance are presented in Table 2.3 (K. Child, pers. comm.). The dominant land use is timber harvesting. Winter logging is commonest as the wet soils create access problems for mechanized equipment in summer. The small forestry-based communities of Willow River and Giscome are near to the mouth of the Willow River and the western end of Eaglet Lake, respectively. A large sawmill operated in Table 2.3 58 Estimated Relative Abundance and Herd Structure for Wintering Moose on the Intensive Scudv Areas, 1964-65 to 1975-76 Study area Winter Relative abundance* Herd structure (%) Sample size bulls cows calves unclass moose time** Eagle 1964-65*** 1965-66 1.22 41 42 17 «. 93 76(D) 1966-67 2.80 29 56 16 70 25(D) 1967-68 1.15 28 48 24 - 82 71(D) 1968-69*** 1969-70*** 1970-71 2.60 23 - - 77 13 5(J) 1971-72 0.97 26 47 27 1 90 93(D) 1972-73*** 1973-74 0.69 12 56 33 — 52 rri-1974-75 m 19 61 19 - 36 1975-76 1.00 14 57 29 - 14 14 (J) Grove 1964-65 1.47 - 16 16 68 25 17 (J) 1965-66 0.53 39 43 18 - 28 53(D) 1966-67 0.02 100 1 45(M) 1967-68 0.54 31 41 27 - 51 94(D) 1968-69 1.41 32 49 18 1 82 58(J) 1969-70*** 1970-71 0.82 22 - - 78 9 H(J) 1971-72 0.98 26 48 25 - 87 89(D) 1972-73*** 1973-74 1.09 45 39 3.6 _ 140 m 1974-75*** 1975-76 0.30 21 50 .29 - 24 80(J) Salmon 1964-65 1.93 _ 10 10 81 31 16(J) 1965-66 0.85 20 46 29 6 35 41(D) 1966-67*** 1967-68 0.94 18 54 29 - 61 65(D) 1968-69*** 1969-70*** 1970-71 m 33 66 - - 9 m(J) 1971-72 1.42 38 40 18 4 50 35(D) 1972-73*** 1973-74 0.90 21 55 19 2 85 ID 1974-75 m 29 54 15 1 m 1975-76 1.00 37 40 23 - 30 30(J) Summary : E3gle 1.49(2.11)^ '24(29) 52(19) 24(17) Grove 0.80(1.39) 31(24) 41(34) 21(13) Salmon 1.17(1.08) 28(20) 47(56) 20(19) *No. of moose seen/min. helicopter survey. **Time in min. with month of survey in parenthesis (J - Jan., M - March). ***Data missing or surveys not conducted. +Missing. -H-Mean (range). Giscome until 1974 after which both it and the village closed. Logs are now hauled to mills in Prince George and Upper Fraser. Agricultural development is limited. Several small farms border the north and west shores of Eaglet Lake. The burn has been traditionally utilized by a small number of cattle. A community pasture development for the burn is currently in progress. Pasture development and maintenance could prove detrimental to moose. The burn and logged over lands are traditional, well-used hunting areas. This results from the relative abundance of animals, good access, especially after freeze-up, and proximity to Prince George. Few other recreational past-times occur except for some nature-viewing, berry-picking and limited bathing and boating along Eaglet Lake. The second intensive study area was Grove, so-named after a forest fire in August, 1961 that burned 2 approximately 320 km of mostly cutover forests. Grove lies 30 km due east of Prince George (Figure 2.1) and covers 2 approximately 450 km . The area is oval-shaped with the long axis (27 km) oriented north-south and the short axis (19 km), east-west. The western boundary roughly follows the 762 m contour (the east boundary of the Prince George Special Sale Area, and also the upper limit of the pro-glacial lake) from Buckhorn Lake northwards to Tsadestsa Creek; then easterly and southerly up the Willow River to 60 the Willow-Kale Forest Development Road; then westerly along this road to Buckhorn Lake. This southern boundary, like the others follows a topographic/physiographic break - an abandoned outwash channel that separates the Mt. George upland from the Tabor Mountain upland. Elevational range is from 640 to 1,260 m. Compared to Eagle, most of Grove is above the proglacial lake basin. Thus soils belong mostly to the Deserters association - glacial till in a moist environment (Figure 2.8). Upland areas to the west and south have shallow till over bedrock (Decker association), and rock outcrops make up to 20% of the high elevation map units. Glacial-fluvial deposits are especially common along the southern boundary but are also found at higher, more central locations. Presumably, these were meltwater deposits originating from ice wastage. Lacustrine deposits occupy the northern one-quarter of the study area, basically that land situated north of Highway 16 and below 790 m. These are mainly Bowron silts and Pineview clays. Beach deposits and deltas mark the boundary between lacustrine and till substrates. Recent alluvium, McGregor and Stellako associations, is present along the Willow River. Organics are more common than at Eagle but are confined mainly to the southwest corner of the study area. Many streams drain Grove but lakes are scattered and small, except for Buckhorn and Tabor Lakes. 60a Figure 2.8 A soil association map of the Grove study-area. Source of data: see Figure. 2.6. 61 GROVE STUDY AREA SOILS 62 Most of the plant cover is serai, due to the large fire in 1961 (Figure 2.9). Willow {Salix spp.) and paper birch {Betula papyvifera) are the major woody plants in the burn but conifers are regenerating. In the southern part of the burn, dense stands of immature lodgepole pine occur. Mature forests of mixed pine, spruce and subalpine fir are found at the north and east-central parts. Compared with Eagle, the Grove area has few cutover habitat-types as most of them were burned in 1961. For moose, the Grove Burn is primarily a winter range. Resident animals are present but not as abundant as on Eagle. Although tagging data are lacking, most of the wintering moose probably come from the upper part of the Willow River valley and the surrounding uplands southward. Possibly before the lands surrounding the Fraser River were cleared and settled, moose wintered there instead of on Grove. Typical composition and relative abundance of the winter moose herd are presented in Table 2.3. Winter distribution of moose on Grove was obviously non-random. One of the main reasons for selecting this study area was to determine why this was so. Recreation is the principal land-use on this area apart from its value as a moose winter range. As access is good throughout much of the burn, it is well used by hunters. A downhill ski facility is located along Highway 16, and cross-country skiing is commonly pursued throughout the 62a Figure 2.9 A forest cover map of the Grove study area. Source of. data: see Figure 2.7. 6 3 GROVE STUDY AREA VEGETATION r—i ^ , . .. » > Major highway EZD Conifer forest = Mature (>80yr.) Secondary road Sk, „, '™aturf (21-8°yfJ —Study area boundary KSJ Mixed forest (>21yr.) — — — Deciduous forest A Aspen B Birch Li^] Shrub type: Burn ESi] Swamp * S j Shrubby vegetation Cultivated Cutover = Clearcut Partial cut 2 1 o IHHMHHI KILOMETRES burned-over area. Trail-riders, snowmobilers, and naturalists also use the area. Difficulties in accommodating such a diverse range of outdoor activities have been encountered. An organization has been formed to try to harmonize the various recreational demands and still minimize disturbance to wintering moose. The third intensive study area was called the Salmon, taking its name from the river that flowed through this important winter range (Figure 2.1). Situated approximately 3 0 km north of Prince George, it measures approximately 3 0- x 10-km, straddling the Salmon River and Merton Creek from the Highway 97 road bridge at Salmon Valley to approximately 2 5 km up Merton Creek. This rectangle is oriented northwest to southeast. The Salmon 2 study area covers about 300 km and ranges in elevation from 610 to 780 m. Compared with Eagle and Grove, which are relatively discrete geographic units, the Salmon area is considered as a subsample of a larger area. From a soils and landform perspective, the Salmon area is more heterogeneous than Eagle and Grove (Figure 2.10). Beginning at the lowest elevation in the Salmon River valley at 610 m are the recent alluvial materials that form the river banks. Above these Stellako deposits are beach ridges intermixed with glacial till and clayey lacustrine sediments. A transitional zone occurs above this, between the Deserters and Pineview soil associations. 64a Figure 2.10 A soil association map of the Salmon study-area. Source of data: see Figure 2.6. 66 This variety is expressed in smaller map units than are found at the Eagle and Grove study areas. Vegetation on Salmon is also diverse (Figure 2.11). Stands of white spruce and black Cottonwood in various developmental stages grow on the floodplain of the Salmon River. At the southeastern end of the area is an extensive aspen and aspen-conifer stand that developed after a wildfire approximately 75 years ago. The northeastern area is primarily pine and pine-spruce mixtures with smaller areas of immature lodgepole pine. Old cutovers occur in the valley bottom, mostly at the northeast end of the area. Other commercial forest stands are currently being logged so that proportion of cutovers will increase. The Salmon River valley is one of the major moose ranges in the study area. Complexes of summer and winter habitats occur over much of it, but winter range predominates. A major migration that crosses Highway 97 south of Summit Lake indicates most wintering moose come from the McGregor uplands. The available data for wintering moose are summarized in Table 2.3. Logging occurs on the study area. Formerly, timber was rough sawn at a mill near the mouth of Merton Creek, and trucked to Prince George for drying, planing, and marketing. This small mill settlement is abandoned and logs are now hauled to Prince George for processing. A small summer ranch is situated approximately 3 km west of 66a Figure 2.11 A forest cover map of the Salmon study area. Source of data: see Figure 2.7. SALMON STUDY AREA 68 Highway 97, along the north side of the Salmon River. To develop pasture, the present trembling aspen forest is being cut down and the growth of shrubs and resprouting aspen controlled. These activities will reduce the area's suitability for moose. As with the other study areas, the Salmon range provides many recreational opportunities. It is a well known and accessible hunting locale. The fall migration of moose across Highway 97 onto the area is almost legendary with Prince George residents. The many stands of vaccinia provide fine berry-picking, while the Salmon River offers canoeing and fishing. One active trapline includes part of the study area. 2.3 The Secondary Study Areas In addition to the intensively studied winter ranges (primary study areas), other locations were sampled for various components of the project (Table 2.4). Selection of these secondary study areas was governed by the type of components, the analyses conducted, the lack of certain characteristics on the intensive areas, and by the need to gain a broad perspective for the study. Thus a wide range of soils, climatic regimes, habitat-types and cutovers was visited. Locations of both primary and secondary areas are shown in Figure 2.1. Data for food habits were collected throughout the entire study area, including the primary areas. Specific locations of rumen samples are described in Appendix E. Table 2.4 Types of Analyses Conducted on the Study Areas Topic studied* Study area Habitat Plant Browsing, Names use succession nutrients Bedding Climate Primary study areas: Eagle X X X X X Grove X X X X X Salmon X X X X X Secondary study areas: Bowron X X Found Lake X Fyfe X Limestone X McGregor X McKenzie X X X Pineview X Shell X Swamp X Teardrop X Telachick X X Torpy X Whites X *Food habits were based on samples collected over the general study area. 70 2.4 History of Moose Distribution  and Abundance The following historical account was extracted from Hatter's (1950) comprehensive treatment of the topic. Apparently, moose were absent or very low in numbers prior to and including the first half of the 19th century. MacKenzie (190 3) saw moose on the east slope of the Rockies but not in the Parsnip, Fraser or Blackwater drainages to the west. Harmon (19 03), who travelled extensively between trading posts on Stuart, Fraser and MacLeod Lakes, reported a few moose. Hatter (1950:28) concluded that these sightings referred to the northeastern part of New Caledonia adjacent to the Rocky Mountains. Brooks (1928) related that reliable reporters believed moose to be absent in the Prince George area "in their father's time"--ca. 1800 to 1850. Moose definitely inhabited the north-central region by the last half of the 1800's. In 1862, Anderson (1867) knew of moose straying in the area between Fort George and Bowron Lake, and rarely as far west as Fort George. Evidently, a moose was killed at Fraser Lake and another at Chinlac, near the junction of the Stuart and Nechako Rivers. McCabe and McCabe (1928a, b) summarized the history and status of moose in the Bowron Lake Game Reserve. They described moose being hunted in the 1870's and 1890's in this Reserve. Cast antlers and sign were also found in the 1890's around Ahbau Lake. 71 In the early 1900's, records of moose seen and shot became more plentiful than previously (Hatter 1950). Walker (1972:17) recalled that the first moose was shot at Prince George in 1914. The increased number of sightings correlates with increased settlement by Europeans. Therefore it is possible that moose were as common prior to 1900 as they appeared to be in the early 1910's. Nevertheless, their rapid range extension and population irruption from oa. 1900 to 1950 was a remarkable biological phenomenon. The general pattern of extension was from headwaters of the Fraser River west to Francois Lake and subsequently southward and westward. Both the appearance and abundance of moose in the north-central region were attributed to forest succession caused mainly by increased man-induced forest fires (Hatter 1950). Initial expansion was likely due to these vastly increased serai shrub communities. The latter spread was due to a "forced expansion": moose ranged further afield as forage became scarce in regenerating forests. Hatter (1950) identified that peak populations probably occurred in the 1920's and 1930's, based on many accounts of local abundance that occurred at this time. Similarly, his choice of fire as the main causative agent was based on a collation of historical accounts from the period and the preceding 40-50 years. In addition to natural fires were those set by 72 Indians, to attract game and provide forage for horses; by settlers, to clear land for farming; by railway and road contractors, to remove slash and debris; and by miners, to facilitate prospecting. Although all the foregoing evidence is circumstantial, the role of fire in the expansion and abundance of moose is the most plausible explanation. Moose numbers declined after the 1920-1930 peak. However, the harvest records from 1950 onwards suggest that the decline has not been to those levels of the early 1900's. Current estimate densities of moose in the general 2 study area are estimated at 0.4/km (K. Child, pers. comm.). 3. HABITAT USE AND SELECTION 3.1 Introduction Habitat is defined as the place where an animal or plant lives (Odum 1971). Occupation of the habitat is habitat use. Since moose occur throughout the sub-boreal forest, virtually all of it can be considered moose habitat. Moose do not occupy these forests uniformly. They use the various serai stages and types of forest stands to varying degrees. By sampling defined units of sub-boreal forests, relative use of these units is determinable. The relative importance of these various units, or habitat-types, can be assessed from this descriptive information. Assessments then allow wildlife biologists and foresters to identify habitat units that may require special considera tion in management. However, data on habitat use are often inadequate to identify factors responsible for differential use of habitat-types. Sometimes these data are interpretable but data on habitat selection are, in the long run, most useful in this regard. Selection differs from use. Habitat use records relative occupancy of habitat-types. It does not define selection since the availability of habitat-types is not 74 usually measured. Selection implies choice. It can therefore provide insight into what factors that moose relate to when occupying habitats. It is these factors that managers must address to manage effectively. Use is descriptive while selection is predictive. The term "habitat-type" has several meanings. In this thesis, it is used to describe units of the landscape that are characterized by obvious differences in vegetative cover or by obvious differences in position on the land scape. Common habitat-types are mature coniferous forest, upland burn, partial cutover, deciduous forest, swamp, and clear cut. Defining these types carries the implicit assumption that they are meaningful to moose. The term habitat-type is not used in the sense of Daubenmire (1959), that is, to describe a land unit that is capable of producing a certain kind of climax vegetation. Patterns of habitat use have been described for most of the circumpolar range of moose. North America and Scandinavia have received particular attention. The Quebec symposium on moose ecology (Bedard et al. 1974) presently provides the most recent and comprehensive summary on this topic since Peterson's (1955) book. Habitat use by moose in British Columbia has been virtually unstudied. Much useful, general information was thoroughly collated and synthesized by Hatter (1950). Baynes (1956) contained general remarks on habitat use. Several reports such as those by Ritcey (1967), Sumanik and Warren (1968), and Eastman (1974a) dealt with habitat use but their data have not been published. The major exception is the recent thesis by Silver (1976) on habitat use by moose in northeastern British Columbia. Therefore, the first objective of this phase of the study was to describe habitat use by moose, with particular reference to cutovers, burns and forests. Given the dearth of information, a synoptic approach was judged to be the best approach. Emphasis was on sampling a wide variety of habitat-types within the study area boundaries so that use patterns could be generalized. The primary technique used to determine habitat use was pellet group counting. The second objective of this phase of the study was to examine habitat selection. This objective required complete sampling of an area without regard to where moose occurred. This approach was undertaken through intensive study on the Salmon winter range. The site was selected as typical of moose habitat in the study area. It included a major winter habitat, the bottomlands of the Salmon River valley, with an adjacent upland that was largely unlogged. Specific boundaries of the intensive study were those of a timber sale harvesting license. Thus I tried to assess habitat selection in an area before logging. It would be useful to conduct follow-up surveys after logging. 75a Figure 3.1 Photographs illustrating logged habitats in sub-boreal forests: a) selective, cut and leave, and c) clearcut. 77 3.2 Methods  3.2.1 The Synoptic Survey The method of counting pellet groups was developed from a review of relevant articles (e.g. Neff 1968, Smith et al. 1969) and a preliminary trial in the field. The two standard plot shapes, circles and strips, were tested using three commonly used diameters and widths of belts (Table 3.1). The checking time, recorded to the nearest second, was measured from the start of one plot to the start of the next. Thus the time included both travelling and plot-reading. With the belt transects, each plot was re-checked for any missing pellet groups. The trial was located in the Grove Burn study area. A total of 1,524 m of transect or 50 circular plots was tallied for each plot shape-area combination. Based on these sources of information, I selected a 1.5- x 30.5-m belt plot for each person on the synoptic surveys. It was the quickest of all combinations tested (Table 3.1) and, of the belt plots, was the only one in which no groups were missed. Coefficient of variation (CV) compared favourably with that of other trial plot sizes and shapes (24 percent vs. a mean CV of 24.3 percent). To determine winter habitat use, pellet groups were counted after snow melt until ground vegetation obscured pellets, usually from mid-May to the end of June, in 1972 and 1973. Pellet groups were tallied in contiguous 78 Table 3.1 Results from Trial Pellet Group Survey; Time/Plot and Number of Groups/Plot.* Plot size and shape width x length of belt plots (m) radii of circular plots (m) Criteria examined 1.5x30.5 3x30.5 6.1x30.5 1.1 1.7 3.6 mean time/ plot** 44(10.6)*** 168(33.2) 205(56.6) 49(11.3) 53(11.3) 85(25.5) mean no. groups/ plot 0.65 1.47 3.06 0.10 0.21 0.66 coeffi cient of variation mean no. groups/ha 23 26 27 40 37 27 no. of groups missed 0 3 9 ndv nd nd + plot area (sq/m) 44.5 93.1 186 4.05 8.09 40.5 sample size 49 49 49 50 50 50 *Data in Appendix Table D-l. **Includes travel time. ***Mean (sd). +nd = no data. 3- x 15.2-m belt plots (46 sq m), with two people each surveying a 1.5- x 15.2-m area. Usually, two or more transects were run for a total of at least 305 m each in each homogeneous type, yielding a total sample area of 1,830 sq m per type. The site and direction of transects were selected prior to field work to minimize bias in transect location. Data from both years were combined and expressed as mean number of accumulated pellet groups per ha in each habitat. Usually pellet groups deposited during the preceding winter were easily aged as such. The autumn leaf fall usually covered pellets from previous winters. Also pellets from the preceding winter were colored differently from those deposited in previous winters. Freezing, thawing, and decomposition speeded up the rapid disintegration of pellets and likely contributed to the observed differences in color. Summer and spring feces were readily distinguisable from fall and winter pellets. Thus pellet group data were assumed to reflect habitat use during the approximately preceding eight months. Pellet group data were expressed on a per hectare basis rather than as moose-days per hectare. The conversion to moose-days assumes that daily defecation rates are known, but the published literature shows that the conventional rate of 13 groups per day is invalid. Timmermann (1974: 616-617) documented daily rate estimates that ranged from 80 10.3 (Le Resche 1970) to 32.2 (Le Resche and Davis 1971). Reasons for these differing estimates probably relate to incorrect aging of pellet groups (Le Resche and Davis 1971), diet (Smith 1964), and sex-specific rates of deposition (Des Meules 1968).. Careful study by (Franzmann et al. 1976) documented large differences in deposition rates between male and female moose. Because sex and age structure of a sampled moose herd is not always known (and this can vary within a single winter season), and since the relationship between diet and defecation rate is not established, the conversion of pellet group data to moose-days is misleading. For these reasons, pellet group data were not converted to moose-days. 3.2.2 Pellet Group Counting  Methods for the Detailed Survey clusters of three circular plots (r = 1.1 m) at stations along transects. The layout of transects and spacing of stations along transects followed methods outlined by Smith et al. (1969). The number of stations per transect was estimated using the formula: For this survey, pellet groups were counted in m m number of stations per transect, average time in hours to read a station, C 2 average travelling time in hours, 81 2 Sw = variance among stations on the same transect, 2 and S = variance among all stations. Variance estimates were obtained from preliminary field trials. Time estimates were based on Smith et al. (1969), with an upward revision of C2 to four h^due to the inaccessibility of the study area. The number of transects was determined by the following formula: n = G/(C2 + C^m), where C^, and m are as above, G = total number of available man-hours, and n = number of stations per transect. Transects were randomly selected from all possible transects, given a separation of 161 m to minimize overlaps. North and south ends of these transects were permanently marked as were station centers along each transect. Details are provided in Bonar et al. (1975) . 3.2.3 The Aerial Surveys Detailed monitoring of habitat use was accomplished through aerial surveys across the Grove study area from January.1972 to May 1973. Using a Cessna 185 aircraft, transects were flown monthly at 3.2 km intervals across the study areas at 90 - 150 m above the ground. The same flight paths were followed each month. The start and end of each transect were located to the nearest 7 5 m, and its. duration 82 was timed with a stopwatch, to the nearest second. Two observers, seated on either side of the aircraft recorded moose and their tracks on portable tape recorders. Observations were recorded to the nearest second from the start of each transect. Moose were classified as either young, adult-unknown sex, cow, large-, medium-, or small-antlered bull, or unclassified. Tracks were distinguished as old (at least one week or before the last snowfall) or fresh (made since the previous snowfall). Aging of tracks was facilitated by flying within one or two days after a snowfall (Figure 3.2). Track abundance was recorded subjectively as few, moderate or plentiful. Examples of flight summary sheets and transect data are appended in Tables D-2 and D-3. All data are on file at the Wildlife Research Section, Fish and Wildlife Branch, Parliament Buildings, Victoria, B.C. Several precautions were taken to standardize the surveys. Most flights were made in the morning during calm, stable air conditions with a high overcast cloud cover. Stable air reduced the problem of drift, while shadowless or faint-shadow light conditions minimized the problems of both glare and poor visibility. Transect tie-points were conspicuous, discrete and easily defined from the air, e.g., road junctions, bridge crossings, boundaries and corners of logged areas. Additionally, several conspicuous points were noted along the transects to check on drift. 82a Figure 3.2 The relationship between daily snowfall and the timing of the aerial transect surveys, January 1972 to May 1973. 83 A TRACE SNOWFALL ' f FLIGHT DATE 104 84 Air photo mosaics (approximately 1:50,000) proved especially helpful in keeping the aircraft on course. The transects were positioned on the study area to include as much variation as possible with respect to elevation, landforms, types and ages of logged over stands, and natural habitat-types. The flight lines used for the Grove study area are shown as an example in Figure 3.3. 3.3 Results 3.3.1 Habitat Use Synoptic survey covered 12 6 ha on a total of eight areas during May and June of 1972 and 1973. Six areas were sampled in 197 3 with 56 transects, and five areas were sampled in 1973 with 67 transects. In both years, Eagle, Grove and Salmon were surveyed with some sites checked in both winters. The remaining five areas were different. Data summaries are listed in Appendix Tables D-4 and D-5. Results of the synoptic survey are summarized briefly below (Table 3.2): 1. Eagle: Winter use was greatest in the coniferous forest cutover, with 108 pellet groups/ha (pg/ha) recorded. This was five times the level of use determined for the birch forest. Two types within the Eagle burn were used differentially. The open shrubby vegetation had almost three times the use of the birch forest. Although both types originated from the same fire, the birch forest had less browse available in winter. The cut and leave strip cutover received about the same use as the birch type. 84a Figure 3.3 A map showing flight lines used for the aerial transect surveys on the Grove, study area. 85 86 2. Salmon: The most heterogeneous habitats of aspen, mixed forest, and partial cutover, received heaviest use. The coniferous forest was used at an intermediate level, while the cleared aspen type was used the least. Compared with Eagle, differences in use of the various habitats was not great. 3. Torpy: This site was in a heavy snow belt. The forest received essentially all the winter use. 4. Whites: The two clearcuts sampled experienced similar levels of winter use although they were logged seven years apart. 5. Grove: The importance of burn habitats was emphasized on this winter range. Use of the 12 year old burn was double that of the coniferous forest at 91 pg/ha. The mature forest received heavier use than the immature forest, presumably due to its better forage supply and possibly greater ability to intercept snow. 6. McGregor: The possible effect of slash burning is illustrated at this study area. The unburned cutover received heavier use than the burned one. Apparently, the coniferous forest had little or no winter use. This may have occurred because the sampling site had little understory vegetation and because it was situated adjacent to the burned cutover. 7. McKenzie: A total of seven habitat-types were sampled, all within close proximity of each other. As at the Eagle and Salmon study areas, the heterogeneous partial cutover was very heavily used with 323 pg/ha. This was the highest density recorded during the entire study. The increasing use of older cutovers was also shown at this area with group densities increasing from 0, 16 and 49 for clearcuts aged 1, 3 and 5 years. Similar to the Grove area, immature forests were used less than mature forests. Mean levels of use varied between the study areas, although sample sizes were small for some areas (Table 3.2). Table 3.2 Relative Winter Use of Available Habitats on Selected Study Areas, Based on Pellet Group Surveys Study Area Winter use Study Area Winter use Habitat-type (pellet groups/ha) Habitat-type (pellet groups/ha) EAGLE: birch forest (2)* 22(0. 2)** coniferous forest (4) 108(0. 3) partial cutover (4) 73(0. 4) cut and leave cutover (2) 22(0. 2) burn - 35 yr (5) 57(0. 4) Mean (17) 63(1. 5) SALMON: mixed forest (3) 42(0. 3) aspen forest (3) 43(0. 3) cleared aspen forest (3) 23(0. 3) partial cutover (11) 41(1. 3) coniferous forest (14) 35(3. 3) Mean (34) 36(5.6GROVE: burn - 12 yr (13) imm. conifer, forest coniferous forest (7) (4) 91(2.2) 22(0.4) 39(0.9) Mean (24) 69(3.5) MCGREGOR: clearcut, burn - 3 yr (1) 0(0.1) clearcut - 3 yr (1) 14(0.1) cut and leave cutover (1) 22(0.2) coniferous forest (2) 0(0.05) Mean (5) 12(0.5) Table 3.2, Continued Study Area Winter use Study Area Winter use Habitat-type (pellet groups/ha) Habitat-type (pellet groups/ha) TORPY: McKENZIE: clearcut - 1 yr (2)*** 0(0.1) clearcut, burn - 1 yr (3) 0(0.3) coniferous forest (3) 36(0.1) clearcut, burn - 3 yr (2) 16(0.2) clearcut, burn - 5 yr (2) 49(0.2) Mean (5) 22(0.2) partial cutover - 5 yr (2) 65(0.2) partial cutover - 10 yr (3) 323(0.2) imm. conifer, forest (1) 11(0.1) WHITES: coniferous forest - 200 yr (3) 29(0.3) clearcut - 1 yr (1) 43(0.1) Mean (16) 67(1.4) clearcut - 8 yr (2) 34(0.2) Mean (3) 34(0.3) *No. of transects. **No. ha in sample. ***Unburned unless noted as burned. CO 00 89 Three levels were distinguishable: high at Eagle, Grove and McKenzie; moderate at Salmon and Whites; and low at McGregor and Torpy. Winter use also varied between habitat-types, based on combined information from all areas (Table 3.3). Burns were most heavily used with 8 5 pg/ha. Partial cutovers were also heavily used at 69 pg/ha. Deciduous, mixed and coniferous forests received intermediate utilization, averaging 39/ha. Least used were clearcuts less than 10 years old and immature forests, where approximately 20 pg/ha were recorded. Relative levels of habitat use ranged widely, with the highest recorded use 4.5 times the lowest. Table 3.3 Relative Winter Use of Major Habitat-types in the Sub-boreal Forest, Based on Pellet Group Surveys Habitat Winter use by moderate snowfall class* heavy Overall mean use Burn 91(2.2) 57(0.4) 85 Partial cutover 76(1.7) 51(0.6) 69 Mixed forest 42(0.3) ns 42 Mature coniferous forest 35(4.5) 68(0.7) 39 Deciduous forest 43(0.3) 22(0.2) 35 Clearcuts 26(1.0) 0 21 Immature coniferous forest 19(0.5)** ns*** 19 Means 52(10.5) 48(2.1) 51 *Use expressed as pg/ha. **Area sampled in ha. ***Not sampled. Snowfall class modified strongly the levels of use (Table 3.3). Thus in heavy snowfall areas such as McGregor and Torpy, the partial cutovers and mature coniferous forests were the most heavily used of the sampled habitats. Burns and partial cutovers received less use, while deciduous forests were used at an even lower level. Clearcuts appeared to be avoided during winter. Compared within areas of moderate snowfall, burns were clearly the most heavily utilized. Partial cutovers ranked next, followed by deciduous and mixed forests. Mature coniferous forests ranked fifth rather than first as they did in the heavy snowfall class. disturbance. This was indicated by comparing the ratio of pellet group densities in a serai stage to densities in an adjacent mature forest. Using the ratios enables comparisons to be made between the study areas. These ratios are as follows (derived from Table 3.2): Serai age Ratio of serai (yr) Habitat stage: forest Study areas 1-3 clearcut 0.3 Torpy, McGregor, The level of use changed with time since McKenzie 5 clearcut partial cut-over 1.7 2.2 McKenzie McKenzie 10-11 burn partial cut-over 2.3 Grove McKenzie 11 35 burn birch forest 0.5 0.2 Eagle Eagle 91 Serai age (yr) Habitat Ratio of serai stage: forest Study areas 50 mixed wood, aspen lodgepole pine lodgepole pine 1.2 0.6 Salmon Grove 90 0.4 McKenzie Use of serai stages varied from 0.20 to 11 times that of mature forests. Recent clearcuts and immature lodgepole pine and birch forests were used comparatively the least, with all ratios less than or equal to 0.4:1. Older clearcuts, burns and especially partial cutovers received heavier use than mature forests except at Eagle. However, the variation between study areas indicates that the fore going generalization must be applied carefully to specific study areas - other factors such as proximity to roads also influence relative use of serai stages. age and type of serai stage, type of adjacent forest and distance from the ecotone (Table 3.4, Figure 3.4). Several case studies indicated this: 1. Eagle: On this heavy snowfall study area, winter use of the open burn declined steadily from the ecotone out into the burn until approximately 300 m (Figure 3.4) from the forest edge. Beyond this, use approached the average recorded for the entire burn. In the forest, use again declined but only until about 150 m from the edge after which use increased steadily. This pattern resulted primarily from the very high pellet group counts made on one transect . (El in 1973, where 32 pellet groups were recorded in a 305 m """^^ transect (Appendix Table A-5) ) . The level of use at ecotones varied according to the Table 3.4 Winter Use of Ecotones Between Forests and Variously Aged Serai Stages, Based on Pellet Group Surveys „. . Eagle Grove Salmon McKenzie Torpy/McGregor Distance from — -ecotone (m) burn forest burn forest clearcut* aspen clearcut forest clearcut forest 0 - 30 161 161 90 215 144 0 0 0 0 0 31 - 61 108 108 197 269 0 0 0 0 0 65 62 - 91 54 0 144 0 0 36 0 0 0 0 92 - 122 54 161 215 27 36 36 0 54 0 0 123 - 152 54 161 72 108 36 0 0 0 18 0 153 - 183 0 538 90 0 0 0 0 0 184 - 213 0 431 144 0 54 0 0 54 214 - 244 0 108 43 0 0 108 0 0 245 - 274 108 1292 72 0 0 0 0 0 275 - 305 108 0 36 0 108 0 0 108 305 43 22 76 27 ns 79 ;e of stage 37 12 5 2 2-4 *In an aspen stand cleared for pasture development. 92a Figure 3.4 Relative use by moose of ecotones and adjacent habitats, based on pellet group transects. < I CO DL o DC O LU LU Q_ 200 100 100-1 200 300 200 100-1 EAGLE STUDY AREA a OVERALL MEAN USE b MEAN USE BEYOND 310 M. e ECOTONE 93 ] a —e 100 200-1 GROVE STUDY AREA -a -b -e -b -a 2001 SALMON STUDY AREA 100 H 100 E3 H-STRIP WIDTH (M) •n- -a -e -a -b 305 305 94 2. Grove: On this moderate snowfall area, use increased up to approximately 230 m from the forest edge and then declined slowly but steadily for up to at least 305 m into the burn. Extrapolating beyond this distance suggested that use would reach average levels recorded for the entire burn at about 500 m from the edge. In the adjacent forest, use declined over the 305 m sampled but still remained at greater than the average level for the forest. Again, extrapolating the trend indicated that average levels would be reached at 500 - 550 m into the forest (Figure 3.4) 3. Salmon: In the cleared aspen forest, use declined sharply away from the ecotone to reach the overall mean use by approximately 18 0 m. In the undisturbed aspen forest, use at the ecotone was virtually nil. Level of use gradually increased but had not reached mean use for this habitat even 3 05 m from the ecotone. 4. McKenzie, Torpy, and McGregor: The recently clearcut areas received little or no use: most use was recorded in the forest. No attempts were made to relate distance from the ecotone and level of use due to the low usage of open habitats. These results on use of ecotones between forested and unforested habitats can be summarized as follows: when little or no forage is available in open habitats, use of ecotones is virtually nil; when forage is sufficient in the open (see Section 7) ecotones receive heavier than average use. The decline in use away from ecotones is greatest in aspen forests, intermediate in heavy snowfall areas and least in moderate snowfall areas. Distance from ecotones at which use approaches the average for the open habitats were 180, 400, and 550 m, respectively, for these three types. Levels of use for roads and skidtrails were different from the levels in habitats in which these access routes were located (Table 3.5). This differential use occurred at least in clearcuts and partial cutovers, the only types sampled. My impression was that similar differentials occurred in burn and forest types, too, although further verification is needed. Table 3.5 Winter Utilization of Roads and Habitats in Which they were Located, Based on Accumulated Pellet Groups in 1973 at McKenzie Accumulated pellet groups/ha Sites used access habitat (type) access habitat Type of access road, skidtrail 87 road, skidtrail 221 road 832 skidtrail 475 164 (clearcut) M17-18 Ml-2 224 (partial) M19-21 M3-4 1552 (partial) M22-24 M5-6, M14 262 (partial) M25-26 M7-8 Pellet group counts on roads and skidtrails could be employed to estimate use for the habitat in which they occur. This offers a very useful field technique to assess utilization of habitats since access roads can be sampled much more easily than logging slash. The relevant prediction equation was: y = 1.8333 x -0.019, where x = no. of accumulated pellet groups/sq m on an access route, y = estimated accumulated pellet group density/sq m in the 2 habitat, n = 4, R = 0.81, s„ „ = 0.0361, and calculated F ratio = 8.31. Tabulated F ratio for 1 and 7 df = 5.59 for P = 0.05, and 12.25 for P = 0.01. Additional sampling is needed before this approach could be widely adopted. Although utilization surveys were directed primarily at winter, limited data were also collected on habitat use in summer in 1973 surveys. In the summer, moose preferred partial cutovers to forests and clearcuts (Table 3.6). Use of partial cutovers appeared to be especially heavy at the Eagle study area, where the highest density of summer feces were recorded (131/ha). It should be remembered that these comparisons were made between terrestrial habitats: aquatic habitats were not sampled. Based on other studies of summer habitat use, aquatic habitats would be used at levels equal to or greater than terrestrial ones. Table 3.6 Relative Summer Use of Habitat-types Based on Accumulated Summer Feces Recorded in the 1973 Synoptic Survey Relative summer use (accumulated feces/ha)* Habitat Eagle Grove McKenzie Salmon All areas conifer forest -(1) 33 (2) 8 (4) 11 (12) 8 partial cutover 131(1) ns 26 (5) 25 (8) 25 clearcut ns** ns 6 (8) ns 9 burn -(1) 32 (4) ns? ns 22 *Study sties were for Eagle: El-4, Grove: Gl-6, McKenzie: M1-M16, and Salmon: S1-S26 except S6. **Not sampled. 3.3.2 Habitat Selection Selection was assessed through pellet group sampling for the timber sale licence area at the Salmon winter range. A total of 2,522 plots were distributed over the 4,288 ha, for a total sample of 2.3 ha or 0.05 percent of the area. In general, few pellet groups were counted (Table 3.7). 2 Approximately 94 percent of the 9.2 m plots had no pellet groups, while 5 percent had only one group. In the remaining 1 percent of the plots, 27 had two groups, ten had three groups, two had four, and one each had five and ten groups. In total, 206 pellet groups were encountered for an average of 0.08 groups per plot. Habitats varied in their area (Table 3.8). As this area was intensively sampled, the types defined for the synoptic surveys were expanded as follows (the percentages denote the proportion of the study area occupied by that type) : 1) mature conifer forest (77 percent) 2) immature conifer forest (7 percent) 3) brush swamp (6 percent) 4) open swamp (2 percent) 5) partial cutover (5 percent) 6) clearcut (2 percent) 7) creek bottom (1 percent) The major additions were aquatic habitats. Coniferous forests clearly dominated the vegetative cover with 8 4 98 Table 3.7 Distrubution of Pellet Group Plots According to Habitat-Type and the Number of Groups they Contained on the Intensive Salmon Area Habitats 0 1 2 3 4 5 10 No. of plots per type mature coniferous forest 93 5 1 t* t t t 1857 (74%) immature coniferous forest 99 1 99 ( 4%) open swamp 99 1 99 ( 4%) brush swamp 92 8 99 ( 4%) creek bottom 92 3 5 39 ( 2%) partial cutover 88 9 2 2 135 ( 5%) clearcut 99 1 192 ( 8%) Totals 2361 120 27 10 2 1 1 2522 *t = less than 0.5%. Table 3.8 Selection of Habitats-Types by Moose in Winter as Indicated by Accumulated Pellet Groups on the Intensive Salmon Area Habitats No. of pellet groups/type (%) Mean no. of groups/plot name area (%) mature coniferous forest 77 81 0.09 immature coniferous forest 7 0.5 0.01 open swamp 2 0.5 0.01 brush swamp 6 3.9 0.08 creek bottom 1 2.4 0.13 partial cutover 5 10.6 0.16 clearcut 2 0.9 0.01 Totals (7 types) 7288 ha 206 groups 0.11 percent of the study area in these types. Brush swamps and partial cutovers were the next largest types, but each only covered approximately 5 percent. The remaining three habitat-types comprised jointly 5 percent of the timber sale area. Moose did not use habitat-types in proportion to their area (Table 3.8). The null hypothesis of relative use being proportional to areas was rejected at P = 0.002 2 (x = 20.51, 6 df). Data that most clearly reflected selection of habitat-types were average number of groups per plot in each type. Thus, partial cutovers and creek bottom types were most preferred with 0.16 and 0.13 groups per plot. Less preferred were the mature conifer forest and brush swamp types. The remaining habitats were little used with approximately 0.01 groups per plot in immature conifer forest, open swamp and clearcut types. 3.3.3 The Timing of Migration  and Occupancy Periods Migration and occupancy periods were defined primarily by ten series of aerial transects flown between January 1972 and May 1973. The number of moose seen per minute of transect flying was used as an index of the buildup and decline of moose on the three intensive study areas. The value of this approach was limited by two factors. First, the success in sighting moose depends largely upon adequate snow cover. Thus surveys made in 100 summer and spring probably underestimated moose densities. This effect was partly offset by the very open nature of much of study areas. Second, moose typically shift from open to forested habitats in mid-winter. Since moose are less observable in forested habitats, this shift would partly confound actual migrations away from winter ranges. This effect was partly offset by using tracks to record activity. These shortcomings notwithstanding, pronounced changes in the moose index were recorded. These changes were broadly similar for the three areas even though vegetative cover was different. In the 1972-73 winter, moose began concentrating on the winter ranges at least by mid-November (Figure 3.5). The pattern of concentration was not documented but probably moose arrived gradually on the winter ranges (Edwards and Ritcey 1956, Goddard 1970, Stevens 1970, Houston 1968, Coady 1974). Hunting might have delayed the timing and rate of concentration. Although I have no data on this point, general observations by wildlife biologists strongly suggest that this occurs. Peak densities occurred in mid-winter, but in different months for the three areas (Figure 3.5). The highest moose indices were recorded in November at Eagle, in December at Salmon, and in January at Grove. The indices declined steadily thereafter, reaching their lowest values in April or May. Thus moose occupied these ranges for 100a Figure 3.5 Number of moose seen/minute of flying on ' •' ." the Eagle, Grove and Salmon study areas during aerial transects in the 1972-73 winter. 102 approximately six months in 1972-73. 3.4 Discussion 3.4.1 The Importance of  Habitat Variability Moose used heterogeneous habitats more than uniform habitats. In particular, partial cutovers were preferred over all others.on most of the study areas sampled. The uniform types that were less heavily used were mainly immature single-species stands and recent clearcuts. These results were similar to many other moose habitat studies. In Minnesota, Peek et al. (1976:58) stated: "Apparently, this logging inadvertently created ideal moose habitat by removing large acreages of jackpine pulp timber and creating extensive shrub communities, interspersed with balsam fir and aspen and white birch stands." In Newfoundland, Pimlott (1953) considered that "good" moose habitat was the commercial forest, especially the mixed stands of balsam fir (Abies balsamea) and white birch (Betula papyrifera) interspersed with muskegs and alpine areas. High winter densities of moose were recorded by Bergerud et al. (1968) in logged-over Newfoundland forests. These cutovers contained scattered stands of residual conifers interspersed with regenerating balsam fir and white birch. In Nova Scotia, Telfer (1967) found that 15 - 2 0 year old pulpwood cutovers provided winter moose yards, and Prescott (1968) remarked that wintering moose 103 concentrated most frequently in partial cutovers that were approximately 15 years old. In a general review of winter habitats of .Quebec moose, Brassard et al. (1974) concluded that the best habitat occurred in the forest zones that were transitional between conifers and broadleaf forests. "Within these . . . moose select those areas that have suffered perturbation as a consequence of logging, fire or insect outbreak" (Brassard et al. 1974:79). Moose populations of moderate densities are also found in the heterogeneous transition zone that extends from north-central Alberta to north-eastern Minnesota (Berg and Phillips 1974). In western North America, the critical riparian winter habitat of the Shiras moose is quite diverse, due to flooding, erosion, changes in channels, and ice damage (Peek 1974a). In Alaska, LeResche et al. (1974) reviewed important habitats used by moose in winter; these were characterized by considerable vegetative diversity. The rapid increase in Fennoscandian moose numbers was attributed primarily to a change from the older selective methods to clearcutting (Lykke and Cowan 1968, Markgren 1974). These clearcuts were small, typically ranging only from 2 ha to 5 ha and so still represented diverse habitats. Many researchers have demonstrated the value of habitat variability as an important habitat component used during winter. Leopold (1933) expressed this concept through his 104 Law of Interspersion. Given that variability is a significant habitat requirement, several questions still remain unanswered: 1. Does use vary over time, and if so, how? 2. How much variability is needed? 3. How does variability vary over time? Results of my study indicate that winter use is minimal in very young or immature forest stages. Low utilization is due largely to an inadequate forage supply (see Section 7.3.4). Use is greatest in serai, stages ranging from 5-20 years. However, variation in this time period is attributed to such factors as site productivity and winter snow depths. The importance of the latter was shown in Table 3.3. The time span of 5 ^ 2 0 years after removal of a mature forest corresponds with findings of most other researchers, particularly logged-over forests. As noted previously, cutovers 15 - 20 years old provided winter habitat in Nova Scotia (Telfer 1967, Prescott 1968). The Siskiwit Lake burn on Isle Royale received heaviest use in the first 14 years after which, use declined (Krefting 1974) . However, the Feldtmann burn showed increasing use up to 34 years afterwards. Krefting (1974) attributed this to a better interspersion of cover in the Feldtmann Burn. Obviously, the most useful period of a sere varies, as shown by the above comparison between the Siskiwit and 105 Feldtmann burns on Isle Royale. A similar case was reported in the Prince George area (K. Sumanink, pers. comm.). The Grove and Tsus Burn both originated in 1961. The former is an important winter range and has regenerated mainly to willow and birch, while the latter is an unimportant winter range and regenerated mainly to lodgepole pine. The question of how much variability is enough varies according to factors specific to the area, and to the nature of the serai vegetation. However, the processes which initiate serai development must be sufficient to stimulate understory production. Telfer (1972) found that stimulation of browse production required reduction of the 2 stand basal area to 17 m /ha although the reduction must depend upon initial stocking. The selection system practiced in Fennoscandia was probably not intensive enough to affect browse production since moose populations did not increase until small sized clearcutting became the dominant cutting practice (Markgren 1974). In Idaho, Hungerford (19 69) found that commercial forest thinnings were insufficient to reduce stand basal area by the 20 - 30 per cent necessary to stimulate understory production. Similarly, both clearcutting and wildfires produces more forage in the early stages than moose can utilize. On the McKenzie study area, use of five year old clearcuts was less than the partial cutover. Also, 10 year old 106 partial cutovers were more heavily used than 10 year old burns. Thus the type of disturbance required to produce attractive winter range should be sufficient to stimulate browse production but need not be as drastic as complete removal of the forest cover. It appeared that partial cutting as practiced in north-central British Columbia approached the ideal. Alternatively, it appears that 2 -10 ha clearcuts achieved a similar "enhancement" effect for Scandinavian moose. The best size and shape for clearcuts is difficult to define because even very large clearcuts or burns are used by some moose, provided forage is not snow covered. Moose are very adaptable animals. A more tractable restatement would be: "what is the best size and shape for clearcuts if the management objective is to maintain current moose populations?" A specified time frame is an inherent part of this type of question. To maintain present moose densities on existing ranges the balance between cover - and food-producing areas, and how distance from cover affect the level of utilization must be determined. The temporal aspect must also be addressed. These components relate primarily to areas where the forest cover has been completely removed by clearcut logging or by intense wildfire. The relationship between edge and use has intrigued foresters and wildlife managers at least since clearcutting 107 became the dominant cutting method. Disputes regarding the best size and shape of clearcuts became acute as logging companies attempted to reduce unit costs by clearcutting larger and larger blocks. Although uncommon now, cutovers of 200 - 30 0 ha can be found in the Prince George area. However, the question of size and shape of clearcuts is unanswerable unless related to a management goal. Since moose will use open areas well removed from coniferous cover, it follows that the influence of size of clearcuts on the presence or absence of moose is essentially very limited, except in the early successional stages when snow covers forage (Stelfox et al. 1976). However, the size and shape of clearcuts will affect the level of use. If the management objective is to increase moose, then smaller clearcuts are in order. If the objective is to maximize timber production and let moose numbers adjust to this, then a different set of criteria for size and shape of clearcuts will be needed. .Results in this section demonstrated that use of open areas, primarily burns, declined with distance from cover. The rate of decline appeared to be influenced by nature of the cover, age of the serai stage and snowfall regime. Comparatively little other research has been done on this topic. Telfer (1974) summarized a few reports (Pimlott 1953:577-78, Vozeh and Cumming 1960:2-4, Bergerud and Manuel 1968:733) by saying that clearcut areas of 130 ha 108 might be utilized by moose. None of these reports addressed the functional relationship between level of use and distance from cover. Neu et al. (1974) found on the Little Sioux Burn in northeastern Minnesota that wintering moose used the 0.4 km peripheries of the burn and the adjacent forest significantly more than expected according to their respective availabilities. Irwin (1975) recorded similar results for the subsequent winter of 1971-72. Although snow depths were not given, the preferred 0.4 km periphery is remarkably similar to the situation recorded for the Eagle study area (Figure 3.4). Recent work in Ontario provides probably the most relevant data to date (Hamilton et al. 1976, Hamilton and Drysdale 1976). Studies were carried out in 1975 and 1976 on three cutovers, aged five and six years old, and that ranged from 24 ha to 525 ha in the area. In 1975, 95 per cent of browsing occurred within 80 m of boreal forest cover. However, in 1976, significant use was recorded up to 260 m from cover. These differences in browsing were independent of browse distribution. Causes of annual variations were not determined, but such factors as increased moose densities, sampling variation and age of cutover were suggested. To these reasons could be added variation in snow depth and•hardness. The data showed variation in details, but similarity in the general relationship of declining use with increasing distance from cover. 4. FOOD HABITS 4.1 Introduction Information on food habits defines what plant species moose consider as forage. It also helps to interpret habitat use and selection, and to assess relationships with forestry practices. Especially important are data that identify dietary preferences on local bases (Peek 1974b). In the boreal forests of western North America (in which my study area lies entirely), food habits of moose are poorly documented. Peek's (1974b) comprehensive review of moose feeding studies on this continent contained no references for this biome. Information from peripheral areas is available but of limited value (e.g., Hatter 1950, Ritcey 1967, Ritcey and Verbeek 1969, Silver 1976). Studies from other boreal regions have little application due to differences in vegetation, climate, and in the wintering habits and habitats of moose. Therefore, I conducted a survey to fill this gap in information. Three main questions were asked: 1) how many species do moose eat?, 2) what seasonal trends exist?, 3) do diets vary with habitat-type? These questions were answered by using three methods, viz., rumen analyses, trailing and post-winter browse transects. These 109 110 complementary techniques overcome the biases and limitations of each as mentioned by Peek (1974b!: 211-213) . 4.2 Methods 4.2.1 Rumen Analysis Rumen samples were collected within the general study area from moose killed by hunters, by poachers and by accidents. Information on these samples is presented in Appendix E. Approximately five samples per month were collected: attempts were made to obtain rumens from both sexes and a wide range of age classes. One liter samples of well-mixed digesta were collected from each rumen, labelled and frozen for later analysis. Freezing has the advantage over other storage methods of preserving helpful color-based characteristics of plants (Korschgen 1969). The method used was chosen after trial analyses of 11 initial samples (Eastman 1974b). Several approaches were tested to determine the most suitable one, and also to determine the fate of rumen material during processing. Standard procedures were followed on completely thawed samples. Oven-dried weights of each sample were determined from total wet weights (after draining to remove excess fluid) and moisture content derived from three subsamples. Samples were washed through standard sieves, 6.35 mm, 4.00 mm, and 2.00 mm, until little or no material appeared in the discharge. These sizes are commonly used in analyses of Ill ungulate rumens. Material remaining on the top two sieves was transferred to a white enamelled tray and separated by hand. This material was called the coarse fraction. Separated material was measured first volumetrically by water displacement and then weighed after being oven-dried for 4 8 hours at 5 0°C. The material remaining on the bottom sieve, the fine fraction was analyzed by point frame sampling and by manual separation of subsamples. First, the fraction was spread evenly over the bottom of an enamelled tray and the fragments sampled with a point-frame. A slanted row of ten pins, spaced equidistant on a wooden frame, was moved across the tray at 50 cm intervals. At each interval, the pins were lowered and hits recorded until a total of one hundred hits were obtained. Next, five subsamples of the fine fraction were separated manually under a binocular dissect ing microscope. Materials separated from each subsample were oven-dried and weighed. The unsampled fine material remaining on the tray was also oven-dried and weighed. During washing, approximately 7 5 percent of each rumen sample was lost (Table 4.1). This represents 154 g of the one liter samples whose mean weight was 193 g. Factors determining the proportion lost were mesh opening, the foods eaten, and the thoroughness of sieving the samples. The digesta retained on the screens were almost evenly divided between coarse (nine percent) and fine (twelve percent) fractions (Table 4.1). All of the coarse material was separated while only five percent of the fine fraction, or 0.7 percent of the total rumen sample, was used in this way. However, all the fine material was used for the point frame sampling. Table 4.1 Components of Rumen Digesta After Sample Preparation Component amount (oven-dried wt in g)* Item coarse fraction fine fraction lost in washing Type of digesta: identified taxa 5.6 ± 0.6 0.25 ± 0.004 decorticated twigs 9.6 ± 3.3 0.80 ± 0.020 other material 1.3 ± 0.3 0.05 ± 0.000 Total weight separated 16.5 ± 4.1 1.10 ± 0.020 Mean weight of component 16.5 22.4** 153.9 Proportion of mean sample weight 9% 12% 80% *Mean ± standard error of mean, where n = 11 except for lost-in-washing material where n = 8. **0nly subsamples of this were separated. Not all of the separated plant material was identifiable as to taxon (Table 4.1). Pieces of bark, shoots, petioles, etc., defy analysis unless various micro techniques are employed (e.g., Hercus 1960, Stewart 1967) -a sophistication probably unwarranted when the inherent limitations of rumen analysis are considered. The largest single item, classified as woody twigs, comprised 58 percent 113 and 73 percent of the material from coarse and fine fractions, respectively. In this and other studies, the species composition of these twigs is assumed to be similar to that of the identifiable material. To my knowledge, this assumption has not been tested. Estimates for frequency of occurrence varied according to mesh size and to procedure (Table 4.2). For separation of the coarse fraction, only nine (eleven percent) of the possible 7 6 occurrences (sum of the number of times all taxons were recorded in the eleven rumen samples) were missed. These misses consisted mostly of small pieces from uncommon taxons, such as single needles of white spruce. Similarly, the number of misses recorded in the analysis of the fine fraction was twenty (26 percent) for separation and for point frame sampling 45 (59 percent). With the latter method, most taxons missed comprised less than ten percent by weight in a sample, indicating that the point analyzer was unsuited for less abundant items. Even if results from separation of both fine and coarse fractions were combined, the proportion of misses would still be approximately three percent. The make-up of the "misses" also varied. In the coarse fraction only one taxon, the lichen Aleotovia spp., was missed; the frequency of occurrence of five others was underestimated. In separation of fine material, six taxons were overlooked and eight were under-estimated. With the 114 Table 4.2 The Effect of Analytical Method on Frequency of Occurrence of Plant Taxa Recorded -in Moose Rumen Samples Times recorded No. of times taxa were missed for 11 rumen separation point frame samples coarse fine fine TREES AND SHRUBS: Abies lasiocarpa (subalpine fir) Alnus spp. (alder) Betula papyrifeva (paper birch) Cornus stolonifera (red-osier dogwood) Pioea glauoa (white spruce) Pinus contorta (lodgepole pine) Populus tremuloides (trembling aspen) P. balsamifeva (black cottonwood) Salix spp. (willows) Sorbus spp. (mountain ash) Vaccinia spp. (vaccinia) Rosaceae Salicaceae FORBS: Cornus canadensis (bunchberry) GRAMINEAE: PTERIDOPHYTES: BRYOPHYTES: LICHENS: Alectovia spp. (old man's beard) Lob aria pulmonavia (lungwort) 10 2 4 4 6 1 1 1 11 1 4 3 6 1 4 6 1 1 9 1 1 3 2 1 1 1 1 2 4 4 6 1 5 3 3 5 1 4 5 1 1 1 Totals (19 taxa) 76 8 (11%) 20 (26%) 45 (59%) 115 point frame, eleven of the total nineteen taxons were not recorded, five were underestimated and only three agreed with the composite information. Estimates for amounts of taxons also varied according to mesh size and to procedure (Table 4.3). Using all gravimetric data as the best quantitative estimate of rumen components, the least deviation was obtained when based on separation of the coarse fraction (0.2 percent ignored), and the greatest deviation when based on point frame sampling of the fine fraction (5.7 percent, sign ignored). The discrepancies of the point frame sampling are larger than those recorded by other observers (e.g. Chamrad and Box 1964, Robel and Watt 1970). This may be due to differences in screen sizes although Chamrad and Box (1964) did not mention the size of screens they used. Samples used by these observers and myself were alike, with one hundred hits recorded in all cases. Thorough mixing of the material minimized non-randomness in the distribution of plant fragments. The most likely reason for difference was variation in size of fragments in my samples. Perhaps analysis of finer material would overcome these discrepancies. Scotter (1966:241) expressed surprise that many rumen analyses are based on volumetric procedures. However, it is more suitable for field operations because it is 116 Table 4.3 The Effect of Analytical Method Plant Taxa Identified in Moose on Amounts of Rumen Samples Proportion (%) Deviation from total wt (%) Plant taxa of total wt* separated separation coarse fine point frame fine TREES AND SHRUBS: Abies lasiocarpa (subalpine fir) 33 1 20 49 Alnus spp. (alder) 3 t** -3 -3 Betula papyrifera (paper birch) 3 t -2 -3 Cornus stolonifera (red-osier dogwood) 5 t -1 -5 Picea glauoa (white spruce) 2 t 3 -2 Populus tremuloides (trembling aspen) 4 t -4 -4 P. Balsamifera (black cottonwood) t -t -;t Salix spp. (willows) 37 2 -16 -29 Sorbus spp. (mountain ash) t t -t -t Vaccinia spp. (vaccinia) 2 t 4 -2 Rosaceae t -t Salicaceae 2 t -2 -1 FORBS: Cornus canadensis (bunchberry) t -t -t GRAMINEAE: 1 t -1 PTERIDOPHYTES: 1 t 5 BRYOPHYTES: t -t -t LICHENS: Alectoria spp. (old man's beard) t t -t Lobaria pulmonarea (lungwort) 6 1 -4 Totals 63.75 g Mean deviation, sign ignored t ' 3 6 *Oven-dried weight basis. Includes all material that was separated from coarse and fine fraction. **t = less than 1%. 11.7 faster under some circumstances and requires less equipment than gravimetric analysis. Also, the apparent disadvantage of incompatability of volumetric and gravimetric data can be overcome since they are highly correlated. (r = 0.99 , n = 101, s = 0.343). For the present data, dry weights of separated taxons were estimable from their volumes by the regression equation: y = 74.69x + 276.64, with y = 2 weight in g and ,x = volume in ml (s = 0.343, r =0.98, y • ^ n = 101) . Based on the foregoing detailed examination of the 11 initial rumen samples, the most satisfactory approach was to separate digesta remaining on a sieve with a mesh opening of 4.00 mm. Analyzing finer material than this, either by separation or point frame sampling, resulted in missing some taxons and in underestimating both frequency of occurrence and amounts of others. Although point sampling was the quickest procedure, the resulting data were incomplete. 4.2.2 Trailing Trailing provided time- and habitat-specific information on foods eaten. Trailing was done in conjunction with other field work. All evidence of browsing and grazing was recorded by species, while following freshly made trails. Each fresh stub was considered as one bite, and a minimum of one hundred bites for each trail was tallied whenever possible. Information 118 on the abundance of available food species was gained from vegetation data collected for the plant succession aspect of this study (see Section 7). 4.2.3 The Browsed Stem Survey These surveys were done in conjunction with post-winter pellet group counts. Thus they were area-specific but did not reveal at what time during the previous winter browsing occurred. They complemented the data from rumen samples. These latter data were time-specific but not necessarily area-specific since a moose may have fed in a different habitat from the one in which it was killed. The method used was to not the species present and whether or not they were browsed in the previous winter. 2 These observations were made in circular 4 m (mil-acre) plots spaced at either 20 m or 30 m intervals along the pellet group transects. At least 20 plots were sampled for each habitat visited. 4.3 Results 4.3.1 The Range of  Species Taken Moose used at least 64 plant species ranging from lichens to dicots, based on all the data (Table 4.4). One-third of the species were shrubs, one-quarter were forbs, one-quarter were lichens and mosses, with the remainder mostly conifers and ferns. While this adequately reveals 119 Table 4.4 Variety of Plant Species Eaten by'Moose, by Forage Class, in Various Parts of Their North American Range No. of species by forage class (%) Forage class Central B.C. Wyoming Minnesota Alaska Ontario Conifers 9 6 14 2 13 Shrubs 33 41 58 31 43 Forbs 23 42 14 39 20 Graminoids 2 9 11 11 9 Ferns 6 3 10 Clubmosses Liverworts 2 2 Horsetails 2 3 3 1 Mosses 13 2 Algae 3 Lichens 11 3 4 Mushrooms 3 1 Total no. species 64 69 36 62 80 Reference This study Houston Peek Le Resche Peter (1968) (1971a) and Davis (1955) (1973) the range of forage classes taken, it undoubtedly under estimates the actual number of species taken. Many aquatics and other species generally recognized as moose foods were not found (e.g., in Ritcey and Verbeek 1969), and few rumen were collected in summer when food habits were likely the most varied. Despite these shortcomings, the results show that moose eat a variety of plant species. This compares favorably with other studies except that moose in this study 120 appeared to take more mosses and lichens (Table 4.4). Whether or not all species were deliberately taken or accidentally ingested with others was not always determinable. Probably most of the mosses and lichens were eaten accidentally by foraging moose since these plants grew on or very close to staple food plants. 4.3.2 The Seasonal Trends This study supported the general pattern of seasonal trends in moose food habits as summarized by Peek (1974). It also supported Peek's observations that moose of central British Columbia have similar food habits to those in eastern North America. The spring diet consisted of 91 percent deciduous shrubs, mainly Sitka alder {Alnus orispa sinuata) , Douglas maple (Acer glabrum) , willow and paper birch. The high proportion of Sitka alder (31 percent) is unusual for moose. Its occurrence in the spring diet and that of other shade tolerant, moisture-loving species such as Devil's club (Oplopanax horridus) and red-osier dogwood {Cornus stolonifeva) , suggested that moose in the study area fed more commonly in mesic rather than in xeric forests. In summer, use of willow increased to 49 percent while use of Sitka alder dropped to 4 percent. During both spring and summer seasons, forbs and graminoids formed less than 10 percent of the diet. This figure likely under estimates the importance of these forage classes since three 121 rumens collected during these seasons contained <1 percent, 72 percent and 89 percent forbs; 50 percent, 11 percent and 2 percent graminoids; and 5 percent, 11 percent and 0 per cent horsetails (Equisetum spp.), respectively (also cf. Ritcey and V.erbeek 1969) . Rumen samples taken in early fall contained significant amounts of non-browse species (Table 4.5, Figure 4.1). Additionally, woody plants would tend to have more recognizable items in the rumen than succulents since they degrade less readily. Deciduous shrubs continued to form the largest forage class in the fall diet (63 percent), with willow and red-osier dogwood as the main species. Grasses rank as the second most important forage class (25 percent), with ferns also important (10 percent). Possibly the increase in ferns, especially the evergreen lady iAthyvium filix-femina) and grape-ferns (Botrychium multifidum), reflects the declining palatability and availability of terrestrial and aquatic forbs due to frost and lowered water levels. Early and late winter diets were examined to see if they changed when moose shifted from open habitats in early winter to coniferous forests in late winter (Table 4.5). In early winter, moose ate primarily willow, paper birch, and red-osier dogwood. In late winter, when moose used forested habitats most heavily, sub-alpine fir comprised 26 percent of the diet, with paper birch and willow making up almost comparable proportions. Thus, the ratio of conifers to Table 4.5 Food Habits of Moose in North-Central British Columbia, Based on Trailing and Rumen Analysis, 1971-74 (%-basis) Species eaten* May - June July - Aug Sept - Oct Nov - Jan Feb - Apr trailing trailing rumen trailing rumen trailing rumen CONIFERS: Abies lasiocarpa (subalpine fir) Pinus eontovta (lodgepole pine) Pseudotsuga menziesii (Douglas fir) Thuja plicata (western red cedar) 17 t 1 t 23 t 26 t DECIDUOUS TREES & SHRUBS: Acer glabrium (Douglas maple) Alnus arispa (Sitka alder) Ametanehier alnifoHa (Saskatoon) Betula glandulosa (bog birch) B. papyrifera (paper birch) Cornus stolonifera (fed-osier dogwood) Lonicera involuorata (black twinberry) Oplopanax. horridus (devil's club) Populus tremuloides (trembling aspen) P. balsamifera (black cottonwood) Rosa spp. (rose) Salix spp. (willow) Sorbus sitchensis (Sitka mountain ash) Spiraea spp. (spirea) Vaecinium spp. (vaccinia) Viburnum edule (squashberry) 13 4 1 4 t 4 31 4 1 4 1 t 1 5 6 t 6 1 1 1 1 2 9 9 4 19 27 32 12 7 4 12 23 11 3 3 4 8 4 t t 1 2 1 6 1 3 5 8 5 3 1 t 1 3 2 1 4 1 t 2 t 10 49 28 27 29 21 36 2 1 t 3 1 2 t 2 t t 1 t t 1 2 6 4 5 1 Table 4.5, Continued May - June July - Aug Sept Oct Nov - Jan Feb - Apr. Species eaten* FORBS: Aster spp. (aster) Disporwn oreganum (Oregon fairybells) Epilobium angustifolivcm (fireweed) Streptopus amplexifolius (twisted stalk) trailing trailing rumen trailing rumen trailing rumen GRAMINOIDS: Graminae (grasses) Typha latifolia (Cat-tail) 25 1 FERNS: Athyrium filix-femina (lady fern) Botrychium multifidum (leathery grape-fern) Dryopteris austriaca (spiny wood-fern) 1 1 LICHENS: Ldbaria pulmonaria (lungwort) Unidentified lichens Total number species recorded Data- base 27 15,844 bites 12,116 23 1bites 36 21 1 t 33 6 1 21 22 12(105 g)** 10,543 bites 33(321 g) 4,447 17(143 g) bites *Includes only species recorded with more than 1%. **No. samples (total oven-dried wt). 123a Figure 4.1 The seasonal changes in forage classes eaten by moose in north-central British Columbia. Derived from data in Table 4.5. I 125 deciduous shrubs changed from 1:63 in fall, to 1:18 in early winter, and to 1:3 in late winter. This shift probably reflected the effects of snow on the availability of forage species. Lungwort {Lobaria pulmonaria) was more abundant in the late winter diet than for any other season. No signs of moose pawing through snow were .seen, although this apparently occurs in the study area (K. Sumanik, pers. comm.), in northeastern and south central British Columbia (R. Silver and R. Ritcey, pers. comm.), and in Alaska (Le Resche and Davis 1973). 4.3.3 The Effect of  Habitat-Type on Diet Diet varied according to habitat. This was shown by data from rumen analyses (Table 4.6), and from post-winter browse surveys (Table 4.7). For example, yearly winter diets differed between three adjacent habitat: logged, old burn and coniferous forest (Table 4.6). Although paper birch and willow were major dietary items, the proportion of subalpine fir varied. It comprised 43 percent of the diet in the old burn, 10 percent in the logged type and was insignificant in the forest. Trembling aspen and lady fern were well represented in the samples from the forest but were uncommon or absent in moose sampled from the other two habitats. Samples from the Grove study area in the late winter period revealed dietary differences between recent burn and forest habitats. Moose in the former type ate 126 Table 4.6 Comparisons of Food Habits of Moose between Different Habitats in Early and Late Winter* Proportion of species in diet (%)** Area 1: early winter Area 2: late winter Species recorded logged old burn forest forest recent burn CONIFERS AND DECIDUOUS SHRUBS: Abies lasioaarpa (subalpine fir) 16 43 t 31 1 Acer glabrum (Douglas maple) 3 6 1 Alnus crispa (alder) 2 t 1 Be tula papyrifera (paper birch) 34 24 24 2 25 Cornus stolonifera (red-osier dogwood) 8 5 1 2 Populus tremuloides (trembling aspen) 5 1 22 7 Populus balsamifera (black cottonwood) 1 7 Salix spp. (willow) 30 25 21 16 72 Sorbus spp. (mountain ash) 2 Viburnum edule (squashberry) 5 OTHER SPECIES: Athyrium filix-femina (lady fern) 1 17 t Lobaria pulmonaria (Lungwort) t t 10 No. spp. recorded 15 11 16 13 7 No. rumen samples 10 4 4 7 4 *Early winter is September to December and late winter is January to April. Habitat-types are logged - cutovers from 1 to 20 years old, forest - coniferous forests, recent burn - 13 years old, and old burn - 60 to 70 years old. **0nly species with at least 1% in diet. 127 Table A.7 Winter Food Preferences of Noose in North-Central British Columbia, by Habitat-Type Percent occurrence/percent browsed** coniferous clear partial recent Species recorded* forest cut cutover burn CONIFERS: Abies lasiocarpa (subalpine fir) Picea glauca (white spruce) Pinus contorta (Lodgepole pine) Pseudotsuga menziesii (Douglas fir) DECIDUOUS TREES & SHRUBS: Acer gldbrum (Douglas maple) Alnus crispa (Sitka alder) Amelanahier alnifolia (Saskatoon) Betula papyrifera (paper birch) Chimaphila umbellata (prince's pine) Cornus stolonifera (red-osier dogwood) Lonicera involucrata (black twinberry) Oplopanaz horridus (devil's club) Populus (trembling aspen) Ribes spp. (gooseberry, currant) Rosa spp. (rose) Rubus idaeus (raspberry) R. parviflorus (thimbleberry) Salix spp. (willow) Sambucus racemosa (elderberry) Sorbus spp. (mountain ash) Spiraea luaida (flat-top spirea) S. douglasii (hardhack) Symphoricarpus alba (waxberry) Vaceiniwn spp. (vaccinia) Viburnum eduls (squashberry) No. of plots Prop, by type No. spp. present Prop, of spp. browsed 35/3 12/8 51/8 21/37 11 14/8 19 32/33 t 4/57 7 3/50 2 14/17 1 4/33 2 6/22 12/50 17 4 26/2 1/10C 9 10/35 41 4/13 17/15 13/37 24/44 10 2 .5/75 3/25 4/67 7/53 5/75 30 19/10 32/4 33/8 5 12 5 3 t 14/7 4/36 3 22 23 31/1 29/9 26 32/2 39/2 20 4 23 21 4 16 28 31 11 3/27 20/17 7/61 79/38 1/20 8 2 1 17 1 16/22 1 29 14 49 7 17 13 20 5 2 49/1 14/22 49/2 5 13 2 25/14 367 159 253 76 43% 19% 30% 9% 35 24 34 23 21% 58% 45% 48% *0nly species with 5% frequency of occurrence are listed. **Percent occurrence = (no. plots with spp. present) (100)/ho. plots sampled; Percent browsed = (no. plots with spp. browsed) (100)/no. plots with spp. present. 128 almost entirely willow (72 percent) and paper birch (25 per cent) , the two most abundant browse species. In the forest, moose ate paper birch and subalpine fir in nearly equal proportions, 27 percent and 31 percent, respectively. Considerably less willow was taken in the forest than in the burn (16 percent versus 72 percent). Trembling aspen, squashberry ( Viburnum edule) and lungwort were taken in the forest, but were missing from the diets of moose in the recent burn even though aspen was common in the burn. Habitat-related differences in food habits were also indicated by the results from the post-winter browse surveys. Although these surveys provided only frequency of occurrence data, they did allow limited comparisons between the availability and utilization of browse on a habitat basis. In the forest, willow, red-osier dogwood, elderberry {Sambuous racemosa) and paper birch were browsed most frequent ly but each occurred in less than five percent of the 367 plots examined. The most common taxons encountered, vaccinia (Vaooinium spp. ), regenerating subalpine fir, black twinberry '(Loniaera involuorata), flat-top spirea (Spiraea tuoida) and rose (Rosa spp.), were either very lightly browsed or not touched. In all, only 21 percent of the 35 species recorded in forest types were taken. A similar pattern occurred in the partial cutover type. The most frequently browsed species of willow, red-osier dogwood, Douglas maple, trembling aspen, Saskatoon (Amelanohier alnifolia) and paper 129 birch were among the least frequently encountered species. Similarly, regenerating subalpine fir, vaccinia, flat-top spirea and rose occurred frequently (in more than 3 9 percent of the plots) but were unbrowsed or only taken in limited amounts. In contrast with the forest-type, however, 45 percent of the 3 4 recorded species were browsed. Results for both the clearcut and recent burn types differed from the two foregoing types and between themselves (Table 4.7). In the clearcut, the commonest species were the deciduous shrubs, rose, thimbleberry (Rubus parviflorus), raspberry (Rubus idaeus), willow, black twinberry and paper birch: the most frequently browsed were red-osier dogwood, lodgepole pine, vaccinia and Douglas maple. Over 50 percent of the list of 2 4 species recorded were browsed. In the recent burn, willow, Saskatoon, black twinberry, goose berries/currants (Ribes spp. ) , and white spruce were commonest but of these, only willow was browsed frequently. Other important browse species were alder, prince's pine (Chimaphila umbellata) and red-osier dogwood. Similar to the clearcut, about 50 percent of the list of 23 recorded species were taken by moose. Several plant species showed consistent patterns, whatever the habitat-type. For example, red-osier dogwood was always uncommon but always frequently used; gooseberries/currants, rose, thimbleberry and flat-top spirea were always common and always lightly used. Other 130 species were browsed more frequently in the open habitat-types than in the forest. Examples of this include black twinberry, subalpine fir and paper birch. Also, a greater proportion of the species found growing in open sites were browsed (45-50 percent) than, in the forest (21 percent). Thus species growing in full sunlight may be more palatable to moose, possibly due to higher levels of carbohydrates and other nutrients (Laycock and Price 197 0). 4.4 Discussion 4.4.1 Methodology The strengths and weaknesses of these three, commonly-used methods to collect food habit information can be seen since all were used in the same area over the same time period. Rumen analysis provided time-specific but usually not site-specific data and so food preferences were not determinable. Rumen analysis suffers from other well-known weaknesses (e.g., Bergerud and Russell 1964). However, analysis of properly taken samples probably provides the best way of describing the range of plant species eaten. Rumen analysis is also valuable where moose feed in plant communities that are too complex or too difficult to allow direct observation. Trailing can be both time- and site-specific, especially in winter when the occurrence of snowfalls can be used to age tracks. Food preferences can also be 131 determined, although this becomes very time-consuming if trails go through many plant communities or wander along ecotones. The major weakness of trailing is in overlooking plants, especially aquatic and arboreal species. In this study, the variety of lichens taken, and the previously unrecorded use and importance of lungwort would have been overlooked if only trailing had been used. Post-winter plot studies are site-specific but not time-specific except to one winter. General information on availability of foods can be collected but this must be , carefully interpreted, considering over-winter snow conditions and depths. Similar to trailing studies, some plant species can be overlooked or their importance underrated. Thus all three methods have limitations in describ ing food habits. The best approach is to use at least two or preferably more techniques that will enable a better description of the variety of plants eaten, their relative importance, their preference in relation to availability, and the actual amounts eaten. 4.4.2 Variations in the Diet This study has shown that the diets of moose not only vary seasonally, but also spatially. Seasonal variations are documented in earlier studies, e.g., Peek (1974b), but published accounts of differences between 132 habitat for the same general area and the same winter are uncommon. These spatial variations probably are due to differences in both the palatability and availability of food species. Shrubs growing in open habitats generally have comparatively high carbohydrate levels (Oldenmeyer 1974, Laycock and Price 1970). Since carbohydrate levels are related to palatability, open-grown shrubs of the same species can be usually expected to be browsed more heavily than plants in the forest. (Of course, other factors in addition to carbohydrates determine palatability of forages.) This pattern was recorded for all species in Table 4.7 except for elderberry. Availability also influences diet. This usually refers to absolute abundance of a species (e.g., cover, phytomass, production), but also should refer to other, less obvious differences. One of these is the habitat-related differences in a species growth form. A Saskatoon plant in the forest often consists of one or two, tall spindly stalks; in the open, the same species is bushy with many stalks. If moose select shrubs on the basis of form, then these shape differences may be important. Use of species may also depend on a threshold of abundance, that is, a certain amount of a species must be present before a feeding moose will find or "choose" it. Type-related climatic effects also modify a species' availability. For example, subalpine fir regeneration in a forest may be above 133 the snow, but the tips can be covered by ice-encrusted snow caps formed by water and snow dripping from the canopy and freezing on the understory plants. This may help to explain why some individuals of a species are heavily browsed while adjacent ones are unbrowsed. 4.4.3 Some Management  Implications of  Variations in the Diet Moose eat a wide variety of foods. Thus they are able to use many stages of the successional vegetation that develop after logging. For example, lodgepole pine sites provide forage in the early successional stages after logging when deciduous species flourish. As the forest develops, the understory decreases almost to the point where no forage is produced. As the stand matures, the canopy opens, enabling shrubs and lichens to develop. These successional trends are treated more fully in Section 7. Because moose have a diverse diet does not mean that cutting virtually any stand will benefit moose. This is an oversimplification because it overlooks the concept of forage availability; the changing nutritional needs of moose, both seasonally and annually; the relative balance between quality and quantity of forage; the changes in plant nutrient levels due to soil moisture (Peek et al. 1976) and canopy closure (Cowan et al. 1950); and the use of seasonal ranges by moose. All these factors must be considered when 134 assessing the impacts of timber harvesting and post-logging treatments on forage production. Although moose eat many plant species, relatively few make up most of their diet (Table 4.5). For north-central moose, these staple species are subalpine fir, alder (in spring), paper birch, red-osier dogwood, and willow. It seems sensible to emphasize the effects of timber management practices on the production of these species. The variability in the diet raises concerns about the nature and role of forests reserved from logging. In forest development plans, immature stands are often reserved to meet the needs of moose for winter cover. While these stands may provide cover, they usually do not have commonly eaten food species. Immature stands typically lack plant species generally found in older forests such as subalpine fir, squashberry, and lungwort. How vital these winter foods are to moose is unknown—our knowledge of moose nutritional needs are too incomplete to provide decisive data. However, relying primarily on immature forests for food and cover may be limiting options for moose production. Another consideration relates to habitat management. In north-central British Columbia, complete clearcutting is the main harvesting method practiced. The same can be said for most of boreal North America and Scandinavia. It is therefore the major method of manipulating moose habitat. While this technique may be the best way of harvesting wood fibre it is not always beneficial for moose. As Peek (1974b) and this study pointed out, local variations in forage preferences are especially relevant if habitat management practices are to favour the locally preferred species. If logging is the main management practice and moose production is the major aim in an integrated forest management plan, harvesting systems and silvicultural techniques should be selected to create post-logging vegetation that will be most productive for moose. This type of predictive information is critical, yet is lacking in many places. 4.4.4 Future Research This study raises several questions for further research. The spatial variation in food habits raises the question of the relationship between diet, home range and over-winter survival of moose. If moose in winter, especially late winter, are sedentary as suggested by van Ballenberghe and Peek (1971), then the habitat mosaic of these home ranges is important. Moose restricted to an immature lodgepole pine stand would probably fare more poorly than moose in an old stand; and both would be worse off than moose in a heterogeneous home range consisting of both cover- and food-producing habitat-types. The important 136 questions then become how big are home ranges, what mix of habitat-types will produce a sustained moose population sufficient to meet public demands, and will this mix be compatible with other resource users? Pood habit studies provide part of the information needed to answer these questions, provided, of course, that .these studies are properly interpreted. Gullion (1966) offered some striking examples of how improperly interpreted food habits data can lead to inadequate and counter-productive forest management guidelines. A second area of research relates to factors influencing the availability and palatability of food species. Disentangling the complex of factors such as soil, light, plant competition, species diversity, chemical composition, browsing history, plant strategies and moose behaviour requires long-term, well-designed experiments. Many studies exist for domestic stock in temperate regions but wildlife-oriented projects, especially for boreal regions, are virtually non-existent. A third area of research is to improve ways of v assessing the nutritive values of forage in terms that are meaningful to the animal, that is, biologically interpret-able. This research need is discussed more fully in Section 8. 5. DYNAMICS OF WINTER BROWSING 5.1 Introduction The previous section described the variety of foods taken by moose, and their relative preference for specified periods of time and localities. However, it did not provide answers to questions about browsing dynamics. When were certain winter habitats used by moose for feeding? Were twigs on a particular plant browsed more than once? Were plants browsed repeatedly? Did levels and rates of browsing vary between plants and between habitats? How often was more than current annual growth removed by moose? What levels of utilization were browse species experiencing? To answer these questions, the course of individual ly marked twigs of four major food species was followed in four habitats through a winter season. Not all twigs of a plant were tagged. Instead, the tagged twigs were considered as a sample of the available browse for a parti cular species in a particular habitat. The available browse was put on a weight basis by measuring twig diameters and then converting them to oven-dried weights by appropriate regression equations (cf. Peek 1970, Telfer 1969). Thus if 50 percent of the tagged twigs of paper birch in a burn habitat were browsed, then the estimated level of browsing 137 138 on paper birch in that habitat was also 50%. The evaluation of browsing can be based on several criteria. For this section, browsing is examined with respect to incidence, time, and level of utilization. Incidence is concerned with whether or not a twig or plant was browsed, and how many times were they used (Section 5.3.1). Time of browsing deals with the temporal aspect of incidence, that is, in what months were twigs and plants browsed by moose (Section 5.3.2). Level of utilization describes how much browse was removed both by month and over the entire winter, November to April 1972-73 (Section 5.3.3) . Utilization also tends to be an ambiguous term. In the context of this section, it refers to browse removed on the basis of oven-dried weight. It does not refer to time spent in a habitat, although moose obviously spent time in a habitat while browsing. It does not reflect densities of moose, since a given level of use can be satisfied by many combinations of occupancy period, age/sex structure and moose density. 5.2 Methods Study sites were located on the intensive study areas, viz., Eagle, Grove and Salmon. The major habitat-types were coniferous forest, deciduous forest, burn, partial cutover. All were represented by one or two sites 139 at each study area. Site descriptions are given in Section 2.2. Typically, each site consisted of five stations spaced at intervals of approximately 30 m along a transect. At each station, usually five twigs on each of ten plants were marked with individually.numbered aluminum tags, for a total of 250 twigs per site (Figure 5.1). The species tagged were paper birch, willow, subalpine fir and red-osier dogwood -the major winter foods of moose in the region - with the most abundant species tagged at each site. In most cases only one or two species were sampled as other shrubs were uncommon. The initial sampling was done in October 1972, except for the river at Salmon which was set up in November. At this time, the basal diameters of current annual growth were measured to the nearest 0.1 mm, with vernier calipers, and entered on standard data sheets suitable for direct key punching. All sites were re-visited at monthly intervals. If twigs were browsed, the diameter at point of browsing, the DPB of Peek (1971), was measured (Figure 5.1) and recorded. Monthly visits concluded in early May, after snow melt was largely completed. Moose were dispersing to summer ranges and the swelling and new growth of twigs prevented further application of the method. For the twig data, diameters were converted to oven-dried weight (in grams) by using the following linear regression equations developed previously: 140a Figure 5.1 Photographs illustrating methods of tagging twigs and measuring diameters at point of browsing. subalpine fir: log (y) = 1.908 log (x) + 1.883, S = 0.118, n = 60, r2 = 0.82 y. x willow: log (y) = 3.19 log (x) - 4.01, S = 0.034, n = 120, r2 = 0.99 y .x paper birch: log (y) = 2.95 log (x) - 3.34, S = 0.040, n = 90., r2 = 0.98 y. x red-osier dogwood: log (y) = 0.92 log (x) + 0.22, S = 0.85, n = 37, r2 = 0.84 y. x where, y = oven-dried weight in grams, and x = diameter of twig or point of browsing as measured in the field in mm. The converted data were then analyzed by calculating the available browse, monthly amounts browsed, total amount browsed, and the percent utilization for the winter. The available browse was defined by the diameters measured at the initial sampling. The monthly amount browsed was calculated by subtracting the current month's dry weight from the previous month's dry weight. For example, the amount browsed in December equalled the estimated dry weight in November, minus the estimated dry weight in December. The total amount browsed was calculated as the sum of the monthly amounts browsed, and the percent utilization was calculated as the proportion of the total weight browsed to the available weight of browse. All calculations were done using a program called 143 TTT (tagged twig transect) written in Fortran 10-G for use on an IBM/3 60 computer. The program listing with explanatory notes is on file at the office of the Wildlife Research and Technical Services Section, Fish and Wildlife Branch, Parliament Buildings, Victoria, B.C. 5.3 Results  5.3.1 The Incidence of Use The following data-base was established on the three study areas. Thirteen sites were established, each with five stations. At these 65 stations, 638 plants were sampled for a total of 2,953 twigs tagged. All sites were checked eight times, except for one site on the Salmon area which was sampled seven times. A total of 25,116 observations were made during the 1972-73 winter. Incidence of use on a twig-basis varied widely between study areas (Table 5.1). Combining all species and all sites, the Eagle area had the highest rate of browsing, with 3 9 percent of the twigs browsed at least once. For Salmon the proportion was 3 4 percent, followed by 15 percent for Grove. A parallel trend was noted for the proportion of twigs browsed twice with values of 3 percent, 2 percent and 1 percent for the three study areas, respectively. Combining the data for all three areas, 7 0 percent of the twigs were unbrowsed, 3 0 percent were browsed once and 2 percent were browsed twice. 144 Table 5.1 Proportions of Twigs that were Browsed Once and Twice in Major Habitats on the Eagle, Grove and Salmon Winter Ranges During the 1972-73 Winter STUDY AREA Habitat-type No. of twigs browsed (%) Month of second use* No. of twigs once twice N D J F M A EAGLE: Conifer forest 4 2 X X 226 Burn ecotone 39 1 X X X 274 Burn centre 58 5 X X X 240 Partial cutover 54 4 X X X X 236 39 3 GROVE: Conifer forest 1 286 Burn ecotone 19 1 X 310 Burn centre 30 1 X X X X 251 Upland burn 1 150 15 1 SALMON: Conifer forest 10 231 Deciduous forest 47 2 X X X 250 Partial cutover 38 4 X X X 250 River bottom 76 1 X X X X 249 34 1 ALL AREAS 30 2 X X X X X X 2,953 *Month in which twigs browsed previously in the 1972-73 winter were browsed again. 145 Incidence of use on a plant-basis also varied widely between study areas (Table 5.2). A higher proportion of plants was utilized on the Salmon and Eagle study areas (58 percent and 49 percent) than on the Grove area (25 percent). A similar ranking occurred with proportion of plants browsed twice with 13 percent, 13 percent and 4 percent, respectively. Repeated browsing on a plant-basis was much higher than repeated browsing on twig-basis. This indicated that moose returned at least once to as many as 17-33 percent of previously browsed plants, but rarely took twigs browsed.earlier that winter. The proportion of sampled plants that were browsed varied between species (Table 5.2). Subalpine fir was browsed least frequently on all study areas, with an average .of 8 6 percent plants unbrowsed and 14 percent plants browsed once. Individual plants of this species were used least at Grove (8 percent browsed once), and approximately equally at Salmon and Eagle- (18.5 percent browsed once). Proportionately more willow plants were browsed than subalpine fir plants. Approximately 44 percent were browsed once, 11 percent were browsed twice and 3 per cent were taken three or four times. Fewer willows were browsed at Grove than at the other two winter ranges. For paper birch, approximately 56 percent of the plants were browsed at least once, 16 percent browsed at least twice and 2 percent at least three times. Paper birch plants were 146 Table 5.2 Number of Times Blants of Subalpine Fir, Eaper Birch, Red-Osier Dogwood and Willow were Browsed on the Eagle, Grove and Salmon Study Areas during the 1972-73 Winter SPECIES No. of times plants were browsed (%)* No. plants Study area 0 1 2 3 4 in sample SUBALPINE FIR: Eagle 82 18 50 Grove 92 8 60 Salmon 81 19 35 Means 86 14 145 PAPER BIRCH: Eagle 37 63 17 1 120 Grove 74 26 5 2 57 Salmon 27 73 24 5 55 Means 44 56 16 2 232 RED-OSIER DOGWOOD: Salmon (mean) 22 78 12 50 WILLOW: Eagle 54 46 17 35 Grove 65 35 7 1 85 Salmon 41 59 13 4 2 46 Means 56 44 11 2 1 166 STUDY AREA SUMMARY: Eagle 51 49 13 205 Grove 75 25 4 1 202 Salmon 42 58 13 3 196 Means 56 44 10 1 t 603 *e;g., 1 = plants browsed at least once, some browsed more than only once. **t is less than 1 percent. used most frequently at the Salmon study area, less frequently at Eagle, and least frequently at Grove. Results for red-osier dogwood showed that 7 8 percent of the plants were browsed at least once, and 12 percent at least twice. Thus incidence of browsing on plants in descending intensity was Salmon, Eagle and Grove. A comparable ranking for the browse species was red-osier dogwood, paper birch, willow, and subalpine fir. Thus variations in evidence of use was great between the species on individual study areas. The variation in incidence of use was slightly less within each species. Incidence of browsing on plants was also compared between habitats (Table 5.3). The null hypothesis tested was that among the species or habitats tested, the proportions browsed were the same. The proportion of subalpine fir plants browsed (16 percent) was essentially the same on all three study areas (x2 = 0.97, 2 df, P = 0.62). Moreover, the proportion of willow and paper birch plants browsed did not differ significantly within habitat-types nor were there significant differences between study areas (x2 = 4.09, 9 df, P = 0.91). Thus it appeared that moose on all areas treated paper birch and willow similarly with respect to proportion of plants browsed. Comparisons between all sites with paper birch and willow with red-osier dogwood were not significantly different (x2 = 5.88, df = 11, P = 0.88). Significantly fewer subalpine fir 148 Table 5.3 Proportion of Plants Browsed, by Species and by Habitat in the 1972-73 Winter Number of plants that were browsed (%) STUDY AREA+  Habitat-type Abies Betula Cornus Salix EAGLE: Conifer forest 18% (50)* Burn ecotone 60% (30) 44% (25) Burn centre 64%* (50) Partial cutover 63% (40) 50% (10) GROVE: Conifer forest** 12% (51) Burn ecotone** 35% (60) Burn centre 4.4% (25) 36% (25) Upland burn 7% (30) SALMON: Conifer forest 20% (35) 7% (15) Deciduous forest 76% (25) 68% (25) Partial cutover 64% (25) 52% (25) River bottom 78%* (50) HABITAT SUMMARY: Conifer forest 16% (137) 7% (15) Deciduous forest 76% (25) 68% (25) Burn ecotone 60% (30) 38% (85) Burn centre 57% (75) 36% (25) Upland burn 7% (30) Partial cutover 63% (65) 51% (35) River bottom 78% (50) ALL DATA MEANS 16% (137) 55% (225) 62% (65) 45% (170) *No. of plants in sample. **Sites 1 and 2 included forest and burn originally. They were re grouped as indicated above for this, presentation. +Field site numbers are 1, 2, 4, 3; 1 and 2, 1 and 2, 4, 3; 2, 1, 3, 5, respectively. 149 plants were browsed than red-osier dogwood plants (x2 = 28.29, 3 df, P < 0.001), and than paper birch and willow plants (x2 = 32.01, 13 df, P < 0.001). 5.3.2 The Time of Use Data on the time of use indicated several general points. First, browsing occurred at all study areas in all months (Table 5.4). Second, all species were browsed in all months, except for red-osier dogwood which was highly preferred yet taken only in January and February. Third, only the deciduous forest at the Salmon area was browsed in every month. All other habitats were used for two-five of the six winter months. Each study area showed some differences from the foregoing generalities (Table 5.4). At Eagle, the conifer forest was browsed primarily in the last half of the winter, but both burn sites experienced browsing almost entirely from November to February. The partial cutover was used for feeding in all months except March. At Grove, the conifer forest was used in November and in April rather than mostly in late winter as at Eagle. These differences may reflect snowfall patterns. Also different from Eagle was the occurrence of browsing at the burn sites in all winter months rather than in early winter. At Salmon, similarities existed with the two other areas. Similar to Grove, the burn or deciduous forest was used in all months. Similar to Table 5.4 Time of Browsing and Level of Utilization (weight-basis) for all Species, Habitat-types, Study Areas and Months STUDY AREA % total use removed, by month Over-winter Browse Habitat Spp.* Nov. Dec. Jan. Feb. Mar. Apr. use (%) available EAGLE: Conifer forest S 1 66 22 12 4 8,825 Burn ecotone B 32 62 1 6 23 60 W 20 30 12 38 23 89 Burn center B 42 49 10 t 29 105 Partial cutover B 1 80 5 15 47 90 W 92 3 6 1 25 31 GROVE: Conifer forest s 2 98 2 26,700 Burn ecotone w 2 8 61 6 23 14 193 Burn center B 9 71 8 2 9 17 58 w 6 94 17 46 Upland burn B 50 t 67 SALMON: Conifer forest R 100 11 212 S 62 38 10 4,486 Deciduous forest B 3 46 9 22 2 17 36 96 W 6 32 18 16 1 27 63 47 Partial cutover B 51 31 10 8 43 53 W 38 26 5 30 38 70 River bottom R NS** 96 4 36 1,059 Table 5.4, Continued STUDY AREA % total use removed, by month Over-winter Browse Habitat Nov. Dec. Jan. Feb. Mar. Apr. use (%) available AREA SUMMARY: Eagle 5 17 2 51 16 10 5 9,199 Grove 2 1 7 1 t 90 2 27,061 Salmon t 5 69 6 18 2 16 6,024 SPECIES SUMMARY: Abies lasiooarpa (subalpine fir) 1 t 23 18 20 37 3 40,010 Betula papyrifera (paper birch) 12 55 14 6 2 10 29 528 Cornus stolonifera (red-osier dogwood) NS 96 4 33 1,271 Salix spp. (willow) 5 30 32 13 t 19 25 476 HABITAT SUMMARY: Conifer forest 1 t 23 20 20 36 3 40,222 Deciduous forest •5 39 13 19 2 22 45 143 Burn ecotone 15 27 31 17 11 18 341 Burn center 26 34 36 2 t > 2 23 209 Upland burn 50 50 t 69 Partial cutover t 63 16 2 2 16 41 244 River bottom NS 96 4 36 1,059 OVERALL SUMMARY: 2 7 38 15 13 25 4 42,284 *Species abbreviated as S = subalpine fir, B **NS - site not set up this month. = paper birch, W = willow, R = red-osier dogwood. 152 Eagle, the conifer forest was used in late winter, and the partial cutover used throughout the period. The river bottom type, sited only at Salmon, was used for feeding only during January and February. Thus the time of browsing in habitats was generally similar on the three study areas. 5.3.3 The Level of Use The timing and amount of browse eaten by moose .varied considerably between study areas, shrub species and habitats (Table 5.4). The Salmon range received the heaviest level of browsing, where moose ate an estimated 16 percent.of the tagged 1972 annual growth. Moose took much less of the available browse at the Eagle and Grove ranges where only 5 percent and 2 percent, respectively, of the tagged twig phytomass was removed. These differences between study areas were not a result of comparable differences in populations of wintering moose. Rather, they probably reflected shifts in the habitats where browsing occurred. For example, only one conifer forest habitat was sampled at each study area. Other conifer stands were available and presumably used by moose. The level and timing of use of these likely varied between the study areas. In future studies, the sampling of the various habitat-types should be replicated and the number of twigs sampled, reduced. Utilization on a species basis generally reflected 153 changes in diet as described in Section 4. Subalpine fir was browsed lightly early in the winter, but by January when moose occupied coniferous forests, use reached 23 percent. This level was maintained during the remainder of the winter, ranging from 2 0 percent to 3 6 percent (Table 5.4). Overall use of subalpine fir was very light, at 3 percent of the 197 2 growth. The light usage coupled with the importance of this species in the winter diet (Section 4.3), demonstrates its abundance in the habitats used by moose in late winter. Paper birch and willow were browsed throughout the winter, though slightly more in early winter than later on. For the two species, monthly levels of use were similar. Total winter use was alike at.29 percent and 25 percent for paper birch and willow, respectively. The data for red-osier dogwood were considered for the Salmon area as no comparable sites were studied at Eagle and Grove. Apparently, red-osier dogwood was utilized very heavily for. a short period. Virtually all use in the river bottom type occurred in January, when moose were driven by deep snow to low elevations. After this month, continuing snowfalls covered most plants of this species. Compared to willow and paper birch, red-osier dogwood is low-growing in most sites. Total use of dogwood was 33 percent, the highest recorded for the four species studied. Habitats were used differentially (Table 5.4). Generally, results for the three areas were similar. Thus reference is only made to study areas to point out specifics: otherwise habitat-types are described generically. Three broad levels of use were discernible, based on winter utilization. Deciduous forest, partial cutover and river bottom types were used most heavily with 45 percent, 41 percent and 36 percent of the 1972 twig production browsed by moose, respectively. Next were the burn habitat-types, with utilization levels of approximately 20 percent. The older burn at Eagle was used slightly more than the burn at Grove, although I considered Grove to have more available browse than Eagle. Least browsed were the coniferous forest and upland burn, with over-winter utilization levels of 3 percent and less than 0.5 percent, respectively. The coniferous forest was represented at all three study areas, and all three sites were used very lightly. The previous sub-section on incidence of use (5.3.2) showed that obvious differences were apparent between habitat-types. The utilization data showed further that levels of use varied widely even between two months in which browsing was recorded. For example, the following habitat-types were used in all six winter months: conifer forest, deciduous forest, burn center and partial cutover. Yet, the levels of use differed considerably. Similar to the pattern of subalpine fir, the conifer forest type was used very 155 lightly in November and December, and then at approximately 25 percent for each of the following months. The deciduous forest was browsed heavily in December, when migrating moose likely first moved onto the winter range, and then at approximately 14 percent per month thereafter. The burn ecotone and burn center types were used most heavily in November, December and January, and then lightly for the duration of the winter, averaging about 5 percent per month. The level of use at the partial cutover type paralleled that of the burn habitat-types. Since red-osier dogwood was the only species in the river bottom type, levels of use were the same for both points of view, that is, virtually all use in January. Additional representation of this habitat-type would have been very useful. The upland burn type occurred only at Grove. It was virtually unbrowsed except for a few twigs taken in December and March. Such low use occurred despite the abundance of paper birch, a key winter browse species. This site clearly demonstrated the need to consider more than browse supply when evaluating a habitat-type's potential. The overall summary of browsing exhibited the following pattern. Levels of use were low in early winter but increased sharply in December. This upswing presumably reflected the increased density of moose on the winter ranges. Levels declined by more than one-half in February and March. Perhaps this reflected a voluntary reduction in 156 intake as has been demonstrated for blacktail deer (Wood et al. 1962) and whitetail deer (French et al. 1956). Utilization rose sharply again in April as snow depths declined and nutritive levels in forage improved. The overall level of use was only 4 percent, based on all data. 6. BED SITE SELECTION BY MOOSE IN WINTER 6.1 Introduction Minimizing energy losses is a major challenge> confronting moose in winter. Forage is often limited in amount and typically wanting in nutritive value. Movement is restricted by snow. The differentials between ambient and body temperatures reach their annual extremes. Chill factors are often high. That moose occur so widely in boreal environments demonstrates their success in adapting to these adverse winter conditions. One adaptation to winter is moose's ability to find micro-sites that minimize energy losses. This ability can be seen as an attempt to locate thermoneutral environments. Obviously, such events as wildfire and logging modify the characteristics and distribution of these micro-sites. A full appreciation of the impact of these events first requires an appreciation of what moose select for in choosing sheltered locations. From a management standpoint, this type of information is a prerequisite to prescribing logging practices that will minimize their adverse effects on potential shelter areas. The shelter sites selected by moose likely vary during the year, as do their reason for selection. In 157 158 summer, Kelsall and Telfer (1974) suggested that moose seek cool sites to avoid heat stress. However, it is probably in winter that moose face the greatest need to avoid thermal extremes, especially when lying down. Thus if moose selected particular sites for shelter in winter, it would be expected to be most obvious in their selection of bedding sites. More specifically, bedding behaviour and bed site features can be used to test the following five predictions: 1) beds will be located under denser rather than more open tree canopies. Selection will be more pronounced during periods of low temperatures or deep snow or both, than in more moderate temperatures or snow conditions. 2) beds will be located under trees affording greatest sheltering effects, i.e., under white spruce and subalpine fir rather than under lodgepole pine and Douglas fir. Deciduous species will be used only occasionally. As snow conditions become progressively adverse, large trees will be selected since they provide greater shelter than small ones. 3) in drumlinized terrain., moose will bed on upper rather than lower slopes to escape the cool air that drains into depressions. Beds will have southerly and westerly exposures to maximize exposure to sun. 4) beds will be in wind-protected locations such as behind trees or shrubbery; orientation will be » towards the wind to detect possible predators. 5) moose will select softer snow than average since it will afford greater insulation (based on Des Meules 1965). The following objectives were set for the bed site study. First, to describe generally the features of beds and their setting. Second, to quantify snow depth and 159 hardness of the beds and the surrounding area. Third, to describe the species composition, size, and crown canopy-closure of trees at the bed sites and the surrounding area. Fourth, to estimate the period and duration of occupation of beds. To ensure sampling a wide range of environments, field work was conducted in three major habitat-types and in three classes of snow depths (Table 6.1). The habitat-types were coniferous forest (undisturbed canopy), selectively logged (partially altered forest canopy), and clearcut and burns (little or no overstory canopy). Most sites were on drumlinized till materials, the dominant landform and sub strate in the region. A few were situated on lacustrine deposits, the second commonest substrate. Snow classes were defined on the basis of the generalized effects of snow depth on moose movements as reported in the literature (Coady 1974, Formosov 1946, Kelsall and Prescott 1971, Telfer 1970, Ritcey 1967). In shallow snow (0-40 cm), moose experience little or no hindrance to movement; in medium snow (41-8 0 cm), moderate impediment to movement; and in deep snow (greater than 81 cm), movement is difficult and restricted. 6.2 Methods Beds were found by following freshly made trails. Information for each bed was recorded on a standard data sheet to secure uniform and complete data collection 160 Table 6.1 Major Habitats, Snow Depth Classes, and Study Areas Sampled for Bed Site Examinations Study areas by snow depth class (cm) Habitat shallow (0-40) moderate (41-80) deep (> 81) Coniferous forest Mackenzie Mtn. Telachick Grove Limestone Creek Found Lake Partial cutover Mackenzie Mtn. Grove Found Lake Open (burn, clearcut) Mackenzie Mtn. Telachick Grove (Table 6.2). The following list outlines the type of data and methods used: a) descriptive notes on area, habitat, date, landform, vegetation and snow cover. b) crown canopy closure immediately above a bed and three m away from it, in a randomly selected direction. Closure was estimated from dot grid counts of vertical photographs taken with a 3 5 mm camera equipped with a 17 mm or fish-eye lens (Brown and Worley 1965). Dot grid counts were duplicated on each photograph to estimate percent canopy closure. c) species, number and DBH of all trees within a 28 m2 circular plot (R = 3m) centered in the middle of the bed. Distance from the edge of the bed to the closest tree stem was also recorded to the nearest cm. This tree was. defined as the "shelter" tree. d) location of the bed with respect to the shelter tree (compass bearing and whether or not in a quamaniq), position on slope (1 = crest of drumlin, 2 = upper half, 3 = lower half, 4 = swale), aspect and slope. e) snow depth in, adjacent to, and three m away from the bed, to the nearest cm. 161 Table 6.2 Example of Data Sheets Used to Study Bed Sites MOOSE-FORESTRY: Bed survey MOOSE NO. BED NO. AREA HABITAT-TYPE LOCATION DATE AREA FEATURES (general description) PHOTO DATA bedsite away a) vegetation b) physiography^ c) snow No. F-stop Speed d) recent weather Film type BED FEATURES - circle units of measurement a) physiography: slope % aspect ° bed position on slope b) snow depth (cm in): in bed ajd. away_ c) distance to closest tree (cm in) (circle closest tree below) d) in quamaniq: yes no e) excretion: urine pellets none f) age: before after g) vegetation: TREES SKETCH (N) No. Spp. DBH 1. 2. _ 3. _ 4. 5. 6. cm. in. REGEN Spp. No. stems SHRUBS Spp. No. stems 9.. 10. Notes: Distance to Forest cover, etc. 162 f) length, width and depth of bed to the nearest cm. g) orientation of moose in bed. This was determined readily by the low and wide depression produced in the bed by the moose's hindquarters. The positioning of front and hind legs in the snow provided a check on the depression features, feces and urine were usually, found only at one end of the bed. h) observed feces and urine. i) estimated time of.formation, based on snowfall records. j) a sketch of the bed, drawn to scale. Unusual features or observations on moose behaviour were also recorded on these sheets. 6.3 Results A total of 94 beds were examined during January, February, March and December, 1973. In size, they averaged 145 ± 26 (sd) cm long by ,94 ± 11 cm wide (n = 46) . The time spent in a bed could not be documented directly. Relative differences were determined, however, by assuming that the amount of excreta in a bed was directly related to time spent there. Thus it appeared that time spent in beds varied according to snow depth (Figure 6.1). The ' regression of mean numbers of pellet groups plus urinations, and snow depth was significant (calculated F ratio = 6.78, tabulated F for 1 and 17 df at 95 percent probability level = 4.45). The appropriate equation was: y = 0.03 + 0.01 x, where y = mean number of pellet groups and urinations per 2 bed, x = mean snow depth in cm, r =0.30 and SE of estimate =0.62. 162a Figure 6.1 The relationship between snow depth and length of time moose spent in beds, as indicated by relative amounts of feces and urine. 2.604 163 2.30. 2.00. 170 y -0.03+0.01 x Sy.x«0.62 r2 -0.30 n -18 F = 6.78 HABITAT TYPE • OPEN o PARTIAL A CLOSED 140 A 1.10 0.80 •o» 0.50-J • m 0.20 20 40 60 80 "ioo" —r-120 MEAN SNOW DEPTH IN CM. 164 Time spent in beds also apparently varied according to habitat-type (Table 6.3). Moose remained longest in beds in partially logged habitats, next longest in open habitats and least in coniferous forests. This result showed an inconsistent relationship with snow depth; moose might be predicted to bed longer in open habitat where snow packs were deeper than in partially logged stands. An explanation for this might be that in partial cutovers, moose were able to secure protection adjacent to food. Thus they would not be obliged to move as frequently as in the open (food but limited cover) and in the forest (cover but limited food). Also, it is likely that other factors influence the time spent in beds. Thus snow depth alone is probably an insufficient predictor. Moose stayed longer in their beds during January and February than in either December or March (Table 6.3), perhaps due to colder temperatures. Longest stays evidently occurred in February, when the mean number of pellet groups and urinations was 1.16, or 55% higher than the lowest mean recorded for March. In drumlinized terrain, moose appeared to choose particular slope positions (Table 6.4A). Generally, upper slopes were taken rather than crests, lower slopes or swales. The null hypothesis of equal numbers of beds in the four categories of slope positions was rejected (x2 = 10.68, 6 df, P < 0.05). The preference for upper slopes was most 165 Table 6.3 Time Spent by Moose in Beds as Indicated by Feces and Urine, According to Habitat and Month A. Monthly difference in time spent bedded down. Ratios/bed Month pellet groups urinations both No. of beds Dec. 0. 61 0. 18 0.79 33 Jan. 0. 87 0. 21 1.08 24 Feb. 0.79 0. 37 1.16 19 Mar. 0.50 0. 25 0.75 8 B. Habitat-•related differences on time spent bedded down. Ratios/bed Habitat pellet groups urinations both No. of beds Forest 0. 61 0. 11 0.71 28 Partial cutover 0.86 0. 45 1.32 34 Open 0.74 0.21 0. 95 22 Means/total 0.71 0. 24 0. 95 84 166 Table 6.4 Locations of Moose Beds with Respect to Position on Slope, and Aspect A. Bed location in relation to slope Snow depth Proportion of beds . (%) by slope position No. Of class crest upper half lower half swale beds shallow 33 47 20 15 moderate 35 27 31 8 26 deep 7 50 43 14 Habitat type forest 17 52 26 4 23 partial cutover 53 17 30 15 open 18 38 38 6 17 Mean/total 27 38 31 4 55 B. Aspect of bedsite Proportion of beds (%) by compass point No. of N NE E SE S SW W NW beds All samples 5 8 14 6 22 8 8 29 63 167 pronounced for deep snow conditions. Type of habitat did not appear to influence selection of slope position (Table 6.4A). The null hypothesis of equal distribution of bed sites by slope position for the three habitat-types was not rejected (x2 = 9.42, 6 df). Moose and other ungulates prefer southerly and westerly aspects in temperate and boreal latitudes, particularly in areas of pronounced relief (Stelfox and Taber 1969). Data from this study indicated that a similar preference existed but on a smaller scale where relief was not pronounced (x2 = 11.13, 7 df, p = 0.14). It should be recalled that the range of elevation in drumlinized terrain commonly does not exceed 150 m. Based on all data, beds on southerly and westerly aspects (S ->• NW) made up 57 percent of all beds (Table 6.4). Similar to slope position, the aspect preference varied with snow depth. At low depths (0-40 cm), 59 percent of the beds had northerly aspects and none had southerly aspects. At intermediate depths (41-80 cm), the proportion of beds on north-facing aspects declined to 4 8 percent and sites on south-facing aspects represented 33 percent. At restrictive depths (> 81 cm), only 17 per cent of the beds were on northerly exposure and 72 percent on south slopes. Within a forest or logged stand, moose selected certain trees as shelter trees. Choice of species varied according to species present, available tree diameters, and 168 prevailing snow conditions. These factors notwithstanding coniferous species were selected over deciduous species in all cases (n = 8). This result is not surprising since the sheltering effect of deciduous species is less than that of conifers (cf. Des Meules 1965). .Snow depths in paper birch stands on the Eagle study area, and in trembling aspen stands at Salmon were similar to.those in the adjacent open areas but deeper than those in adjacent forests (see Section 9.3.2). Moose did not select amongst the available conifer species within the sample plot (x2 = 4.30, 4 df) (Table '6.5A). This was unexpected since spruce and subalpine fir were predicted to be better shelter trees than Douglas fir and lodgepole pine on the basis of their crown characteris tics: the two former species have dense, low canopies, while the latter species have higher, more open canopies. Beds tended to be under the largest tree of the trees within the vegetation plot (Table 6.5B). This choice was most pronounced when only 2-4 trees occurred in the plot. This trend presumably reflected the more pronounced sheltering effect of individual trees in open-spaced stands than in densely stocked, closed canopy forests. Field observations support this: quamaniqs were readily identifiable in open stands but difficult to distinguish in denser stands such as even-aged lodgepole pine. The shelter trees averaged 2 9 cm in diameter. Mean 169 Table 6.5 Comparison of Conifer Species Available as Shelter Trees, With Those Used by Moose A. Tree species selection Tree species (% basis) iree • -status white spruce subalpine fir lodgepole pine Douglas fir Totals Available 46% 46% 7% 1% 132 Used 44% 49%. 5% 2% 43 Totals 80 82 11 2 175 B. Diameter selection (only samples with > two trees) No. of trees Size of shelter tree, DBH basis, with 1 = largest No. of in plot 1 2 3 4 5 6+' plots 2 3 2 5 3 5 3 1 9 4 3 1 1 5 5 1 1 1 3 6-8 2 1 2 2 7 Totals 14 8 2 3 2 29 170 diameters of the two major species were: white spruce: 36 ± 21 cm, n = 15 subalpine fir: 25 ± 16 cm, n = 24. These means were significantly different at P = 0.10 (t = 1.83, df = 37). These data were further analyzed by habitat. In the forest, white spruce trees were larger than subalpine fir (t = 2.46, df = 19). Diameters of the two species were not significantly different in partial cutovers (t = 0.48, df = 10) nor in open types (t = 1.61, df = 4). Other tree species were not commonly used as shelter trees. Two lodgepole pine shelter trees were 28 and 48 cm in diameter, and one Douglas fir was 33 cm. The direction of beds relative to the shelter trees was not random. The null hypothesis of no significant difference in location frequencies for each of the eight cardinal compass points was rejected (x2 = 22.25, 7 df. Tabulated x2 = 14.1 at 95% probability level). Beds were sited mainly in southerly directions from the trees; northerly and westerly locations occurred less frequently than expected (Table 6.6C). The orientation of moose in their beds was examined. No compass point was preferred, based on all available data (Table 6.6B). However, detailed analysis indicated that preferences in early winter (Dec.-Jan.) differed significantly from that in late winter (Feb.-Mar.). For the null hypothesis which postulated no differences between the 171 Table 6.6 Orientation of Moose in Their Beds, and in Relation to the Shelter Tree A. Orientation with respect to eight compass points Months Frequencies of occurrence (%) of orientation Total compared N NE E SE S SW W NW beds Dec. 9 21 15 12 18 12 12 33 Jan. 9 4 26 4 9 9 39 23 Feb. 21 42 11 11 16 19 Mar. 25 13 13 25 13 13 8 B. Orientation with respect to habitat-type Habitat- Frequencies of occurrence (%) of orientation Total type N NE E SE S SW W NW beds forest 7 10 14 14 14 31 10 29 open 9 15 15 18 3 15 12 15 34 partial cutover 10 15 15 25 25 10 20 Total no. 5 10 11 15 10 9 15 8 83 C. Direction with respect to shelter tree (% basis) 12 23 12 29 5 15 3 65 172 two winter periods, the calculated chi-square was 14.88. As the tabulated chi-square is 14.1 at the 95 percent probability level, the null hypothesis was rejected. Orientation also varied with habitat-type (Table 6.6B), with, partial cutovers differing from forests (x2 = 12.64, df = 7, P = 0.10) and from open types (x2 = 11.78, df = 7, P = 0.10). These differences in orientation may have reflected responses of moose to wind direction. Protection from wind afforded by bed sites was treated very briefly. Wind velocities averaged 34 percent of those away from the site (n = 4). To be conclusive, more data are needed. The average distance between a moose bed and the nearest tree was 92 ± 51 cm (n = 3 means of 14 measurements). Although mean distances varied between forest (106 cm), partial cutover (36 cm) and open (13 5 cm) habitat types, these differences were not statistically significant based on a one way analysis of variance (calculated F2 = 0.8 66, tabulated F at P Q5 = 3.98). The need for shelter from snow was predicted to increase as snow pack increased. This prediction was tested by examining four relationships: 1. stem diameter of shelter trees and snow depth, 2. distance between bed and shelter tree and snow depths, 3. difference between canopy coverage above bed and average coverage and, 173 4. snow depth and the proportion of beds in quamaniqs. The relationships should be direct for all cases except number two. The stem diameters of shelter trees was not related 2 to snow depth (r = 0.0, n =11, F ratio = 0.12). This result was unexpected. I thought that moose would have chosen larger diameter trees as snow depths increased. However, this may not be the case for one or more of the following reasons: 1. snow-shielding ability of trees may not be related to stem diameter (direct comparisons of crown canopy may be more revealing); 2. moose may have selected forest stands rather than individual trees for bedding; 3. in the areas sampled, all conifers may have been suitable as shelter trees. As snow pack increased, the distance between beds and the shelter trees decreased (Figure 6.2). Again this is expected since the shielding effect of a tree is likely greatest at its stem and least at the crown perimeter. The relationship is statistically significant (P = 0.10) and is defined by y = 142.38-1.OOx, where y = mean distance in cm between a bed's edge and the shelter tree's trunk, and x = 2 mean snow depth in cm (r = .32, S = 52.12, n = 11). The effectiveness of reduced snow pack was dramatically demonstrated at the Limestone Creek survey site. Average 173a Figure 6.2 The relationship between snow depth and the distance between bed sites and their associated shelter trees. 225 174 y = 142.4+ 100 X MEAN SNOW DEPTH (CM.) i 175 snow depths in the openings between trees was approximately 12 0 cm, but the bases of subalpine firs were exposed. Moose travelled in straight line paths between quamaniqs rather than the typical meander. The only browse available was in quamaniqs and most stems were heavily utilized. As snow depths increased, moose chose sites with crown closures that were progressively more dense than the stands' average. In other words, snow depth and the differential between crown closure of the bed sites and that of the stand, were positively related (Figure 6.3). The relevant regression equations are: forests: y = 0.12 + 0.004x, n = 21, S = 0.21 2 . x F = 5.62, r2 = 0.23, partial cutovers: y = -0.40 + O.Olx, n = 16, S = 0.17, F = 33.49, r2 = 0.71, y. x ' where y = the difference in crown canopy closure as assessed photographically and expressed as a proportion, and x = snow depth in cm. The higher r-squared value for partial cutovers probably reflects the greater opportunity for moose to select sheltered sites. From this equation, the type of "micro=stands" selected by moose from a large area of timber can be predicted. The shielding effect of conifers on snow depth is illustrated in Figure 6.4. Since increased snow pack implies increased hindrance to movements, moose would be expected to behave in 175a Figure 6.3 The relationship between snow depth and the difference in crown closure between a bed site and the.forest stand in which it was located. 176a Figure 6.4 An illustration of effective snow interception by the forest canopy. 177 178 ways that would minimize encountering adverse snow. As mentioned previously, moose at Limestone Creek kept almost exclusively to quamaniqs. The use of quamaniqs for bed sites also varied across the snow gradient, although the pattern varied between habitat-types (Table 6.7). In the forest, three of the 11 bed sites found in shallow snow were in quamaniqs. The comparable data were one of two in moderate snow pack, and 11 of 11 in deep snow packs. The corresponding proportions for partially-logged stands were one of one, two of 14, and seven of eight. The reason for this departure from the forest pattern is not known. For open types, one of the 3 0 beds in shallow and moderate snow classes were in quamaniqs, while one of two was in the deep snow. Table 6.7 Location of Beds in Quamaniqs, as Affected by Habitat and Snow Depth Class Snow depth Proportion of beds in quamaniqs (s) class (cm) forest partial cutover open Totals shallow (0-40) 3/11 1/1 1/3 5/15 moderate (41-80) 1/2 2/14 0/27 3/43 deep (> 81) 11/11 7/8 1/2 19/21 Totals 15/24 10/23 2/32 27/79 Moose bedding sites differed from ambient snow depth (Table 6.8). Differences in the compression ratio (snow depth in a bed divided by ambient snow depth (Des Meules 179 1965) varied between habitat-types. The ratio of 0.41 for partially logged types was intermediate in value, while the value for open habitats was. highest at 0.49. This trend reflected the trend in hardness from greatest in forest to least in open sites. Since most moose in forests bedded in quamaniqs, undisturbed snow adjacent to bed sites were less than average for the study site (Table 6.8). The ratio of adjacent: away depths followed a similar pattern to compression ratios, that is, least in forests and greatest in open sites (Table 6.8). Table 6.8 Comparison of Snow Depths Between Moose Bedding Sites and Adjacent Areas Habitat- Snow compression ratios "Away" snow type (n) bed/away adjacent/away depth (cm) Forest (20) 0.35 0.67 45 Partial cutover (13) 0.41 0.71 79 Open (11) 0.49 0.86 43 Means (44) 0.40 0.72 55 6.4 Discussion No attempt was made to determine number of beds made per day but this has been studied by other workers. During January and February, Quebec moose established 5.0 beds per day per cows and calves; bulls were not studied (Des Meules 1968). Although Des Meules suggested a difference in daily 180 bedding rates between cows followed by calves and unaccompanied cows, I found no significant difference (P > 0.05) in a statistical analysis of his data. In Russia, Timofeeva (1965 and 1967, quoted by Coady 1974) found that in early winter moose rested an average of five times daily (50-60 cm snow), and. in late winter (January to March) an average of eight times daily (> 7 0 cm snow). In British Columbia, Geist's (1963) results suggest that moose were less active in winter than summer, and therefore might bed less frequently but for longer periods. Most recently, Franzmann et al. (197 6) estimated that. Alaskan bull moose bed 5.5 times per day, while cow moose bed 5v4 times per day (ranges 4-7). All these above studies indicate that moose bed 5-6 times daily. Whether or not there are actual differences in bedding rates must await more detailed study. Physiographic features of moose beds are aspect, slope position, slope and elevation.. I did not record elevation since it was unimportant in the drumlinized till plain and level lacustrine basins where most of the bed surveys were conducted. The aspect of bed sites was different in Nova Scotia (Prescott 1968) and Quebec (Brassard et al. 1974). The null hypothesis of no differ ence between provinces in proportions of bed sites in each aspect class was rejected (x2 = 22.24 and 33.52, respectively, df = 7). Aspects of Nova Scotian and 181 Quebecois beds, however, were similar (x2 = 8.04, 7 df). Thus eastern Canadian moose chose predominantly southern aspects for beds (SW, S, SE), while north-central British Columbian moose chose easterly, northwesterly, and southerly aspects (Table 6.4B). These geographic differences probably reflect differences in climatic features such as prevailing winds. Shelter trees have been studied by other moose researchers. Nova Scotian moose beds were located an average of 157 cm from shelter trees whose mean DBH was 23 cm (Prescott 1968). Moose beds (67%) in Nova Scotia were near balsam fir (Abies balsamea). Des Meules (1965) presented similar results for moose in Laurentide Park, Quebec, although he defines shelter trees as those near moose tracks (and beds?). Shelter trees were usually balsam fir, the predominant species, averaging 18 cm DBH. This compares with 92 and 29 cm for my data,.respectively. Peek et al. (1976) also documented the importance of this tree species for bedding sites of moose in northeastern Minnesota. Des Meules (1964) also described changes in mean diameter of shelter trees and distance to these trees as snow depths increased (his Figures 45 and 46). At 30 cm of snow, the nearest coniferous stem diameter averaged 5 cm, while at 90 cm of snow, mean diameter rose to approximately 28 cm. Data for Prince George were 21 cm and 34 cm, respectively. It appears that B.C. moose selected larger 182 diameter trees than Quebec moose. This likely reflected differences in canopy features of the shelter tree species and possibly the generally larger size of B.C. trees. Distances to the shelter tree for the same snow depth (30 cm and 9 0 cm) were approximately 210 cm and 65 cm, respectively, in Quebec compared with 112 cm and 52 cm for my data. Again distances were less than in Quebec, suggesting that white spruce and subalpine fir did not protect moose as well as balsam fir. Both Prescott (1968) and Des Meules (1965) referred to beds located in quamaniqs. . Des Meules (1965) documented a shift in bed location from open cover-types to quamaniqs offered by large conifers once snow reached approximately 76 cm. The work by Peek et al. (1976) also implied the importance of quamaniqs in relation to a shift from open habitats to closed canopied forest stands. In other studies, differences between bed sites and the macro-environment for other climatic parameters support and extend my results. Des Meules (19 65) documented that wind speed at 19 bed sites was 25 percent of ambient velocities, a significant difference at P < 0.01. Based on a smaller sample of 4, my data suggested 34 percent. Des Meules (1965) also demonstrated a temperature differential of 2°C between bed sites and ambient (n = 19). He also pointed out that temperature in the snow profile can be considerably higher than in the atmosphere. Based on ten 183 samples, when surface snow temperature averaged -12.3°C, temperature at 51 cm below the surface was -3.2°C. In a related study on whitetail deer, Ozoga (1968) documented significant climatic differences between deer yards and the surrounding area. Thus selection of bed sites can reduce energy drains from climatic influences. Exposure to wind can be reduced significantly, thereby cutting convectional losses. The temperature differential can also be decreased to spare conductive losses. Radiational cooling can be minimized by sheltering under coniferous cover (Moen 1968). Energy can also be conserved by selecting warmer aspects and mid-slope position that escape cool air in swales and exposure on ridgetops. Moose did not exploit these tactics throughout a winter season. Rather, it appeared that the options were used as the winter environment became harsher and when moose were presumably in poorer condition. Gasaway and Coady (197 4) believe that these behavioural adaptations of moose are efficient enough to minimize thermogenesis. These authors quoted work by Markgren (1966) for moose, Moen (1968) for whitetail deer and Hart et al. (1961) for barren ground caribou in support of their statement. However, controlled studies are needed to confirm this. 7. SECONDARY SUCCESSION IN SUB-BOREAL FORESTS 7.1 Introduction For moose, change due to fire and logging is the most significant feature of sub-boreal forests. As the forest develops after disturbance, its capability to provide food and shelter for moose varies considerably. Early stages supply a superabundance of food but little cover, while latter stages supply sufficient cover but much reduced forage. Since current changes in sub-boreal forests occur on already developed ecosystems, the main focus for moose is on secondary succession. This is the non-phenological, directional change in vegetation that occurs in already established ecosystems (Mueller-Dombois and Ellenberg 1974). With minor exceptions, vegetational" changes in forests are initiated by fire, logging, land-clearing and disease and insects. As secondary succession originates from only partial disturbances of an ecosystem it involves a less complete sequence of life forms than primary succession. It proceeds relatively quickly and is considered to be free from evolutionary changes in ecological properties of the species involved (Mueller-Dombois and Ellenberg 1974). Probably only the light-intolerant species of moss disappear 184 185 completely in secondary succession (Mueller-Dombois 1965), but other, shade-intolerant species show dramatic changes. The sequence of communities developing over time is termed a sere. The sere is subdivided into component serai stages or developmental stages. Stages are usually distinguishable by changes in dominant life forms, height, biomass and vertical stratification. For secondary succession in boreal coniferous forests, the four major serai stages are commonly defined as follows (Cooper 1913): 1) Herb: vegetation dominated by annual, biannual and perennial graminoids and forbs; usually a single layer. 2) Shrub: vegetation dominated by deciduous woody vegetation; usually two layers, the upper, woody, and the lower herbaceous. 3) Early (Pioneer) forest: vegetation dominated by immature conifers; a third layer added - immature trees and tall shrubs. 4) Mature forest: vegetation dominated by mature conifers with three vegetation layers well developed;^epiphytes present. The term "dominated" is used in the sense of contributing the most to the communities' phytomass. The vertical structure of the vegetation was described in terms of the following layers or strata (adapted from Revel 1972): A) the tree layer - all trees taller than 6 m. 186 B) the shrub layer - all woody plants from 0.45 to 6 m tall. C) the herb layer - woody plants shorter than 45 cm, and all herbs, ferns, horsetails and clubmosses. Throughout this section, these layers are usually referred to by their alphabetic name to minimize confusion between the similar names used for layers and successional stages. Previous plant ecological studies of the sub-boreal spruce zone have emphasized mature forests, both climax and near-climax. Probably the earliest work, was conducted by Kujula (1945, quoted in Revel 1972). Illingworth and Arlidge (1960) defined 12 site types in lodgepole pine and white spruce-subalpine fir forests. These were character ized by common or diagnostic understory plant species and tree growth. Subsequently, Wali (19 69) explored vegetation-environment interrelationships. His co-worker Revel (19 72) pursued a synecological study in which he devised an ecological classification of this ecosystem based on methods developed by Krajina (1959, 1965). Revel (1972) referred to studies on boreal vegetation by Moss (1953a, 1953b, 1955) as particularly relevant to the sub-boreal spruce biogeo climatic zone. To this list should also be added the newly completed work by Annas (19 77) on the boreal black and white spruce biogeoclimatic zone. Extensive field work by J. van Barneveld and his associates, though largely 187 unpublished, has contributed substantially to understanding the interrelationships between environment and vegetation and the general patterns of succession. However, these foregoing studies dealt primarily with forest classifica tion and peripherally with succession. Early successional development has been virtually unreported. Given the dearth of information on secondary plant succession in sub-boreal forests, and realizing the importance of succession to moose, the following objectives were set for major serai stages: 1) To describe the plant species composition and abundance (floristics). 2) To quantify the above-ground phytomass of the understory vegetation, by forage classes. 3) To estimate the above-ground net productivity of the understory vegetation, by forage classes. 4) To describe changes in the height of the Shrub layer. 5) To compare the changes in the height and mass of browse in the Shrub layer. 6) To describe trends in the production of food and cover. 7) To assess the significance of major environmental factors on the patterns and rates of succession. These objectives were applied mainly on the Deserters environmental unit since it occupied the largest area within the study area. Other units were treated as time allowed. 188 7.2 Methods 7.2.1 Stratification Successional trends were studied by using side-by-side comparisons (Mueller-Dombois and Ellenberg 1974). Although this method is less satisfactory than studies on the same area over a time period, it is the only one feasible for short-term studies. The major factors influencing successional patterns were time-dependent and site-dependent. The time factor was of prime importance because the value of a plant community as food- or cover-producing habitat varied according to the time since logging or wildfire. For investigations of chronosequences, the main, approach was to hold all other ecological (site) factors as constant as possible, and examine stands of different ages. Of. lesser importance were site-related factors. Analogous to the need for control of site factors in studying chronosequences, is the need to hold time and non-target site factors constant while examining factors of interest. Of the many site-related factors that could be studied, I selected the broad physical environment (climate-substrate) as the principal one for study. Based on familiarity with the study area and with relevant ecological studies, this factor was judged to be of first concern. The non-target site factors were partly controlled. 189 Slope and aspect were controlled by having as far as possible nil slope and nil aspect. Cause of disturbance was not completely controlled since clearcuts older than approximately ten years were not available for study. Thus I used wildfire sites as samples for older stages of the sere. Also, the variation resulting from type and time of fire, and "quality" of logging was not controlled. These were considered important variables that should be addressed in a subsequent study. Stratification was the initial step in selecting potential study sites. The basis of stratifying was the framework of environmental units described in Section 2.1. From this framework, effort was.concentrated on mesic environments over till and lacustrine substrates. (Reasons for this choice were given in Section 2.1.) General observations were made for the other units as time permitted, especially the riparian units. Preliminary site selection followed stratification. Forest cover-type maps (1:15840) were studied to select tentative sites that represented a wide range in stand age and in composition of tree species. Subsequently, aerial photographs were examined to determine accessibility and to check that map data were still applicable. Sites were finally chosen after they were checked in the field to verify the classification of environmental unit and cover-type descriptions. Approximately 75 field checks were made 190 before selecting the final sites. 7.2.2 Field Sampling Procedures Two types of field procedures were used: detailed and synoptic. The first was to determine approximate above-ground net understory plant productivity and biomass in the four classical successional stages: herb, shrub, pioneer forest and mature forest (Odum 1971). For this detailed study, vegetation sampling began in July 1972 after maximum plant development was attained. Only the till substrate was examined. As the sites were on drumlinized till, I decided to sample only crests of drumlin to minimize confounding due to slope and aspect. Four age classes were sampled in duplicate: a) a one year old clear cut that was burned in September 1971, b) an 11 year old selectively logged site that was subsequently burned by wildlife in August 1961, c) a 40 year old wildfire, and d) a 195 year old wildfire. At each sampling site basic site features were noted, e.g., slope, aspect, microtopography, moose use. All trees greater than 2.5 cm diameter at breast height (DBH = 1.4 m) within a 20- x 20-m plot were tallied by species, diameter class and condition (alive or dead). Heights and increment borings were taken from five trees in 191 each plot. In the centre of this plot, a 5- x 5-m perquadrat was laid out, and ten 2- x 0.5-m quadrats were located within the perquadrat. All forbs, shrubs and grasses were clipped at ground level from these quadrats, and bagged separately by species. Shrubs remaining in the perquadrat were then clipped, also at ground level. In the laboratory, shrubs were separated into the following components: leaves, annual (1972) twigs, older twigs. For some species such as white spruce, bearberry (Aretostaphylos uva-ursi) and Prince's pine, complete separations were not possible. All separated shrub material, forbs and grasses were weighed after drying for at least 24 h at 50°C. The second type of field procedure, synoptic sampling, was used to extend the geographic area of coverage and to meet other objectives of this section. The above detailed methods were revised and simplified to enable sampling a wider range of stands. With this method, five stations were sampled at each site. All sites were situated from 1 to 1.5 times the adjacent stand height away from an ecotone (e.g., cut boundary). On level terrain, such as on some lacustrine substrates, stations were usually spaced 30.5 m apart. On glacial till, the drumlin was used as the basic sampling unit. Stations were placed perpendicular to the drumlin's long axis at or near where the drumlin was widest (Figure 7.1). Spacing of the 191a Figure 7.1 The site and station layout used to study secondary plant succession. A. SITE LAYOUT 192 I LONGITUDINAL AXIS WIDEST PART OF DRUMLIN OF DRUMLIN B. DETAILS OF SAMPLE STATION LAYOUT MOSS-LITTER SAMPLE 20x50 CM m M mi i I HERBS & SHORT SHRUB PLOTS 1 x 1 M. TALL SHRUB PLOT 2x2M. -rt-QUADRATS FOR FLORISTICS 20 x 50 CM., 1.5 M. APART 'STATION CENTER (CENTER OF WEDGE PRISM PLOT) 193 stations was equidistant, with actual distances dependent upon the width of the drumlin. Each transect began and finished just above the easily defined swale or receiving area between drumlins. This type.of spacing was done to maximize drumlin-based variation within sites and thus minimize its between sites effect. For each site, the following general information was recorded on a site sheet: a) location of site on 1:15,000 cover-type maps, 1:15,000 aerial photographs, and by forest management unit, b) elevation, topography and parent material, c) site photograph d) general notes on bearing and spacing of stations, type of site, evidence of moose activity both summer and winter, and specific location of site. At each station, data were recorded for understory floristics, understory phytomass and several features of the forest, viz., species composition, condition, basal area, crown closure, height and age. Slope and aspect were determined by using a compass and clinometer. Plot lay-out is illustrated in Figure 7.1. The information used to describe floral changes in secondary succession was plant species composition and abundance. The sampling method and type of data used for floristics differed for each of the three vegetation layers A, B and C. For the "A" layer, species composition was based on the species recorded in the wedge prison sampling and abundance was the percentage composition of the tree stems in these samples. Similarly for the "B" layer, composition and abundance were based on stem counts in quadrats (see second paragraph below). Floristics of the C layer were recorded using the Daubenmire (1959) method. In each of ten 20- x 50-cm quadrats, the canopy-coverage of vascular plant species (excluding inflorescence) was rated subjectively on a 1-6 scale (Table 7.1). Canopy-coverage was assessed by estimating the proportion of a quadrat occupied by the total area of vertically projected (imaginary) polygons that enclosed perimeters of plants of each species. Coverage was recorded if a plant occurred in the quadrat, even if it was rooted outside the frame. Quadrats were spaced 1.5 m apart along a transect that was at right angles to the line of stations in level terrain or that was aligned parallel to the contour in rolling terrain. This sampling provided estimates of frequency of occurrence and cover for ^ vegetation below 1.4 m. Vegetation taller than this, almost entirely shrubs and trees, was not included due to the difficulty of accurately estimating canopy-coverage. Those species missed in sampling but at the site were also noted. Understory biomass was estimated by a combination of two methods. For the tall shrub layer, woody plants taller 195 Table 7.1 Scale Used to Assess Canopy-Coverage of Understory Vegetation (C Layer) (after Daubenmire 1959), plus Domin Scale Equivalents Approximate Canopy-coverage Range of Mid-point equivalent in rating coverage of class Domin scale* 1 0.1 - 5. 0 2.5 +, 1, 2, 3 2 5.1 - 25. 0 15.0 4, 5, 6 3 25.1 - 50.0 37.5 7 4 50.1 - 75. 0 62.5 8 5 75.1 - 95.0 85. 0 9 6 95.1 - 100.0 97.5 10 *To aid comparisons with Revel1s (1972) data. than 45 cm, a 2 - x 2-m perquadrat was laid out at the samp-ling station. Diameters at 10 cm of all shrubs > exceeding 45 cm were recorded by species. Height and oven-dried weights were predicted from these diameters using regression equations (described in section 7.2.3). For the "C" or herb layer, two 1- 2-m quadrats were randomly chosen from the four possible in the perquadrat. From these two quadrats all shrubs less than 45 cm tall, and other vegetation was clipped to ground level and bagged according to the following forage classes: a) forbs b) graminoids c) ferns d) horsetails e) clubmosses 196 f) shrubs g) coniferous seedlings. These samples were weighed after drying for 24 h at 50°C. Trees were sampled using the Bitterlich variable plot method as described by Dilworth and Bell (1971). Basal area was determined using a wedge prism held vertically over the station center. A basal area factor (BAF) of 20 was selected after consultation with foresters practicing in the Prince George area. Approximately 10 trees were sampled at each station using this BAF. The species, condition (dead or alive), and DBH of each "in" tree was recorded, taking precautions mentioned by the above authors. In addition, height was determined trigonometrically for one dominant or co-dominant tree per station, for a total of five per site. Ages of these trees were determined by counting annual rings in increments extracted at 1.4 m above ground level. Estimated ages were corrected according to species and 'site class (Forest Club 1971). Also, at each station, the crown-closure was recorded photographically with a fisheye lens (17 mm) (Brown and Worley 1965). Closure was estimated by counting "hits" or canopy presence on a dot-grid (five 2 2 dots/cm ) for the center 25 cm of each print. A summary of parameters sampled, methods used and sampling intensity for the synoptic surveys is presented in Table 7.2. 197 Table 7.2 Summary of Features Sampled in the Synoptic Study of Succession Sampling intensity  Parameters sampled Method used units/plot plots/site TREES ("A" layer): spp. composition stem count (> 4.0 cm) 1 point 5 basal area wedge prism (BAF 20) 1 point 5 diameter diameter tape (DBH) 0-15 trees 5 age increment borer 1-2 trees 5 height clinometer and tape 1-2 trees 5 condition observation all trees 5 crown closure photographic (fisheye lens) 1 photograph 5 TALL SHRUBS ("B" layer): spp. composition direct count 0 o 2 one z- x z-m quadrat 5 basal area calculated from diameter as above 5 mass prediction .from diameter as above 5 height prediction from diameter as above 5 density direct count as above 5 diameter micrometer calipers as above 5 HERBS, SMALL SHRUBS ("C" layer): spp. composition observation 10 quadrats 5 occurrence presence in quadrats as above 5 canopy-coverage 1-6 scale in quadrats as above 5 phytomass clipping to ground level two 1- x 1-m quadrats 5 SITE FEATURES: substrate soil maps, pits 1 topography observation 1 aspect compass and clinometer 1 observation 5 slope compass and clinometer as above 5 elevation map, altimeter 1 history cover maps, observation, etc. 1 198 7.2.3 The Prediction of Mass  and Height of Woody Plants Mass and height were estimated from morphological measurements, using regression analysis. As Newbould (1967:10) stated, the object was "to obtain correlations between a comparatively small destructive sampling (which is both time consuming and destructive of the habitat) with a larger non-destructive sample which is representative of the stand. . ." Allometric regressions of weight on DBH have been widely and successfully applied in forestry since at least 1944 (Kittredge 1944). The 1973 IUFRO symposium on the mensuration of forest biomass provided a useful recent summary of this subject in forestry (see also Dunn 1974). In wildlife ecology, the main application of dimensional analysis has been to estimate production and utilization of browse, usually considered only as annual twig growth. These estimates have been commonly based on diameter or length measurements (Basile and Hutchings 1966, Halls and Harlow 1971, Lyon 1970, Schuster 1967, Stickney 1966, Telfer 1969), or less commonly on crown volume (Lyon 1968, Quenet 1971), canopy area and other plant-form measurements (Peek 1970). Studies on predicting total above-ground biomass for browse species are relatively uncommon. Telfer (19 69) presented data based on basal diameter measurements for 2 2 eastern North American shrub species. Some forestry and production ecology studies such as those on puckerbrush (Young 19 71), heath species 199 (Whittaker 1962), aspen (Bella 1968), birch (Gregory and Haak 1965), Douglas fir (Crossley 1967, Kurucz 1969), and red alder (Minus rubra) (Smith 1974b) are also applicable. More recently, Brown (1976a) predicted mass for 22 Rocky Mountain shrub species, and crown weights for 11 conifers (Brown 1976b). They were selected on the basis of their commonness in both successional and mature sub-boreal forest stands within a 60 km radius of Prince George. Between 15 and 4 8 samples were collected for each species following points suggested by Demaerschalk and Kozak (19 74). Each sample was classified according to the following four vegetation - parent material classes: unforested on till, forested on till, unforested on lacustrine and forested on lacustrine, to enable testing for possible site-related differences within a species as found by Peek et al. (1971). Only unbrowsed or lightly browsed plants were taken. The upper limits of diameters sampled for each species were based on field observations of the maximum size encountered. All plants were collected from July - September 1973. Each sample was measured for total height (H), living crown depth (CD), stem diameter at palm width (DHW), stem diameter at the lowest living branch (DLB), and dry weight (DW). Lengths and diameters were measured to the nearest cm and 0.01 cm, respectively. The DHW was taken at 10 cm above the base to minimize.variation due to variable' swelling at the root collar. This is analogous to the 200 measurement of DBH for trees. The DLB measurement was suggested by Newbould (1967:18). Weights were taken after drying for 4 8 h at 5 0°C. Several independent variables (and their natural logarithmic transformations) were evaluated as predictors for the dependent variables of total dry weight and height. This choice of predictors was based on an examination of the plotted data, and on suggestions in the literature. For dry weight, the following independent variables were selected: DHW, DLB, (DHW)2, (DLB)2, (DLB)2H, and (DHW)2H. For height, the predictors examined were DHW, DLB and their logarithmic transformations. In all cases, the criteria for evaluation 2 were r , the coefficient of determination, and S , the ' y .x standard error of the estimate. Calculations were performed using the statistical package, SPSS (Nie et al. 1970). Site-related differences had no significant (p > 0.05) effect on the slope of the regression equations, (b^ = b^) at least for the species, site types and independent variables that were compared (Table 7.3). This finding differs from Peek et al. (1971) who found that 87 of 100 comparisons of site-regression coefficients in Minnesota differed significantly from each other (P < 0.01). Possibly, the differences between the Prince George sites were not as great as those in Minnesota. Also, Peek et al. (1971) examined only current year's growth from plants that received varying degrees of browsing, while for the present Table 7.3 The Effect of Site on Predicting Mass for Selected Shrub Species Regression of Sample Site types log DW with sizes Species examined compared* independent var. Calculated t** (site type)* Abies lasiocarpa (subalpine fir) 1 vs. 4 log DLB 0.905 12(1) log (DLB)2H 0.584 18(4) Amelanchier alnifolia (Saskatoon) 1 vs. 4 log (DHW)2H 1.113 23(1) log (DLB)2H 0.404 17(4) Loniaera involucrata (black twinberry) 1 vs. 4 log DHW 0.981 ' 15(1) log (DHW)2H 0.655 23(4) Rosa spp. (rose) 1 vs. 4 log (DLB)2H 0.374 21(1) log (DHW)2H 0.497 8(4) Sorbus spp. (mountain ash) 2 vs. 4 log (DLB)2H 1.711 18(2) log (DHW)2H 2.455 12(4) Spiraea lueida (spirea) 1 vs. 4 same as above 0.403 14(1) - 0.693 15(1) Viburnum edule (squashberry) 1 vs. 4 same as above 1.705 16(1) 2.052 26(4) Populus tremuloides (trembling aspen) 1 vs. 3 log (DLB)2H 0.270 10(1) 1 vs. 4 0.208 11(3) 3 vs. 4 0.062 9(4) 1 vs. 3 log (DHW)2H 1.540 1 vs. 4 0.857 3 vs. 4 1.510 *Site types: 1 = unforested on till, 2 = forested on till, 3 = unforestec on lacustrine, 4 = forested on lacustrine. **A11 calculated "t"'s were not significant at the 95 percent level. 202 study entire plants were sampled that were at most, only lightly browsed. As might be expected from such a diverse grouping of plants, no one estimator was consistently the best predictor of phytomass (Table 7.4). Logarithmic transformations of 2 (DHW) H were best in 10 instances, the log transformations 2 2 of (DLB) H and the untransformed (DHW) H were best for each 2 of six species, while untransformed (DLB) H provided the best fit for one species (Table 7.4). Logarithmic 2 2 transformations of both (DLB) H and (DHW) H provided a better fit of the data than diameter-only measurements for all species except saskatoon, and better than the untransformed diameter x height combination for all species except saskatoon, birch, black twinberry, rose, squashberry and spirea. However, it should be noted that for most species there is very little difference between the r-squared values of the different regression. Log DHW was a preferable variable to log DLB, and to untransformed variables for these diameters (Table 7.4). Except for subalpine fir and saskatoon, equations based on 2 the former diameter usually had the highest r values and the lowest standard errors of the estimate. The better fit of the DHW data may result partly from the complex of factors that determine when the lowest branches die and 2 break off. For all species except thimbleberry, r values exceeded 0.90, with 14 exceeding 0.96. Although the slope Table 7.4 Coefficients of Determination (r Values) for Six Independent Variables Used to Predict Phytomass of 19 Sub-Boreal Shrubs Coefficients of determination for variables Species examined* log DHW log DLB DHW2H DLB ,2H log DHW2H log DLB2H Abies lasiocarpa (subalpine fir) 0. 91 0. 95 0. 87 0. 87 0. 85 0. 98** Acer glabrum (Douglas fir) 0. 87 0. 82 0. 36 0. 33 0. 90 0. 87 Alnus spp. (alder) 0. 97 0. 94 0. 95 0. 96 0. 97 0. ,96 Amelanohier alnifolia (Saskatoon) 0. 76 0. 79 0. 91 0. 89 0. 75 0. ,78 Betula papyrifera (paper birch) 0. 90 0. 89 0. 96 0. 90 0. 95 0. ,97 Cornus stolonifera (red-osier dogwood) 0. 92 0. 60 0. 85 0. 31 0. 97 0. ,79 Lonioera involuorata (black twinberry) 0. 90 0. 82 0. 94 0. 92 0. 93 0. ,89 Picea glauoa (white spruce) 0. 95 0. 95 0. 40 0. 40 0. 97 0. ,96 Pinus oontorta (lodgepole pine) 0. 97 0. 95 0. 48 0. 47 0. 97 0. ,97 Populus tremuloides (trembling aspen) 0. 95 0. 91 0. ,79 0. 92 0. ,97 0. ,97 P. balsamifera (black cottonwood) 0. 97 0. 94 0. ,73 0. 63 0. 97 0. ,96 Rosa spp. (rose) 0. 87 0. 76 0. ,97 0. 97 0. ,90 0. ,87 Rubus idaeus (raspberry) 0. 94 0. 95 0. ,95 0. 96 0. ,98 0. ,99 R. parviflorus (thimbleberry) 0. 27 0. 31 0. ,13 0. 18 0. ,35 0. ,42 Salix spp. (willow) 0. 95 0. 92 0. ,93 0. 43 0. ,97 0. ,95 Sorbus spp. (mountain ash) 0. 93 0. 79 0. ,98 0. 83 0. ,98 0. ,93 Spiraea luoida (spirea) 0. 98 0. 93 0. ,89 0. 83 0. ,99 0. ,97 Vaooinium spp. (vaccinia) 0. 88 0. 85 0. ,97 0. 94 0. ,92 0. .92 Vibernum edule (squashberry) 0. 76 0. 71 0. .96 0. 89 0. ,83 0. .83 *Sample sizes in Table 7.6. **Underline denotes highest r value for that species. o 204 of the regression for thimbleberry was significantly different from zero (P < 0.01), the best fit accounted for only 42 percent of the observed variation. The independent variable of DHW predicted plant height reasonably well (Table 7.5). All regression equations were highly significant (P << 0.01). Coefficients of determination ranged from 0.50 to 0.94, averaging 0.76. Standard errors of the estimate, expressed as percentages of "x", had a mean value of 11 percent with a standard deviation of 4 percent. For the purposes of this study, DHW was selected as the single most useful independent variable for predicting the mass and height of woody shrubs (Tables 7.5 and 7.6). For more detailed studies than mine, combinations involving height and diameter would be necessary. However, in a synoptic study aimed at delineating general trends, the additional time required to measure height was considered impractical and unnecessary. Thus for height, regression equations derived from untransformed data were used; for weight, equations based on logarithmically transformed data were used. During vegetation sampling, five species were encountered in addition to the 19 for which regression equations had been derived. For these five, existing regression coefficients for species of the same genus or growth form were used instead. Thus for hardhack (Spiraea Table 7.5 Regression Coefficients for Predicting Height from Diameter Measurements Regression coefficients „ r , . , — 7 ; ^ Range of hexghts Species a y.x. r (cm) Abies lasiooarpa (subalpine fir) 11. 08 51. 54 ( 4. 60) 0.79 17-166 Acer glabrum (Douglas maple) 31. 40 66. 62 ( 9. 67) 0.77 10-240 Alnus spp. (alder) 25. 42 104. 47 (10. 25) 0.81 60-291 Amelanchier alnifolia (Saskatoon) 36. 65 94. 90 (14. 67) 0.56 47-248 Betula papyrifera (paper birch) 11. 58 109. 49 ( 9. 95) 0.83 38-255 Cornus stolonifera (red-osier dogwood) -27. 81 164. 00 (14. 20) 0.90 14-189 Loniaera involucrata (black twinberry) 23. 26 77. 84 (12. 50) 0.50 9-137 Picea glauoa (white spruce) 12. 01 62. 69 ( 5. 48) 0.82 26-294 Pinus contorta (lodgepole pine) 15. 31 61. 33 (10. 95) 0.71 21-296 Populus tremuloides (trembling aspen) 36. 36 86. 82 ( 9. 07) 0.75 27-274 P. balsamifera (black cottonwood) 19. 95 62. 48 ( 3. 74) 0.94 23-294 Rosa spp. (rose) 8. 06 111. 89 (11. 57) 0.76 15-135 Rubus idaeus (raspberry) -19. 76 179. 17 (15. 60) 0.91 6-150 P. parviflorus (thimbleberry) -22. 55 133. 01 (23. 82) 0.63 12-73 Salix spp. (willow) 18. 23 123. 11 (10. 14) 0.84 9-202 Sorbus spp. (mountain ash) 24. 62 76. 14 (11. 25) 0.62 24-210 Spiraea lucida (spirea) 3. 81 147. 37 (15. 11) 0.75 14-115 Vaecinium spp. (vaccinia) -0. 50 88. 74 ( 7. 54) 0.88 16-110 Vibernum edule (squashberry) 25. 18 84. 19 ( 9.. 12) 0.65 23-224 *Sample sizes in Table 7.6. o Table 7.6 Species Regression Coefficients for Predicting Oven-Dried, Above-Ground Phytomass of 19 Sub-Boreal Shrubs from Diameter*, and from Diameter Squared by Length Measurements. All Variables Based on Logarithmic Transformed Data Regression coefficients+  diameter (diameter)2 length a b a b n Range of diameters (cm) Abies lasiocarpa (subalpine fir) 4. .00 2. ,33 0. ,63 0. 82 36 0. 24--3. 00 Acer gldbrum (Douglas maple) 3. .19 2. , 19** -0. ,26 0. 74** 16 0. 17--3. 82 Alnus spp. (alder) 3. .49 2. ,62** -0. ,93 0. 91** 26 0. 41- -2. 91 Amelanchier alnifolia (Saskatoon) 4. ,03 2, ,11 0. ,41 0. 74 35 0. 29--1. 72 Betula papyrifera (paper birch) 3. ,60 2. ,41** -0. ,14 0. 78 26 0. 30--2. 24 Cornus stolonifera (red-osier dogwood) 3. ,81 2. ,96** -0. ,39 0. 86** 16 0. 25--1. 12 Lonicera involuorata (black twinberry) 3. ,87 2. ,15** 0. ,55 0. 68** 41 0. 28--1. 60 Pioea glauoa (white spruce) 4. .10 2. ,20** 0. ,86 0. 76** 30 0. 35--4. 07 Pinus oontorta (lodgepole pine) 3. ,60 2. ,40** 0. ,13 0. 82** 15 0. 28--4. 44 Populus tremuloides (trembling aspen) 3. ,55 2. , 45** -0. ,44 0. 84** 33 0. 28--3. 23 P. balsamifera (black cottonwood) 3. ,64 2. , 28** 0. ,08 0. 80** 21 0. 30--4. 67 Rosa spp. (rose) 4. ,13 2. ,77** -0. ,25 0. 90** 31 0. 15--1. 22 Rubus idaeus (raspberry) 3. ,90 3. ,07 • -0. ,49 0. 84 15 0. 18--0. 85 R. parviflorus (thimbleberry) 2. ,55 1. ,34 0. ,68 0. 41 20 0. 21- -0. 70 Salix spp. (willow) 3. ,71 2. , 65** -0. ,58 0. 86** 30 0. 12--1. 70 Sorbus spp.. (mountain ash) 3. ,48 2. ,52** -0. ,49 0. 86** 30 0. 32--2. 58 Spiraea luoida (spirea) 3. ,58 2. ,23** -0. ,20 0. 76 34 0. 12--0. 68 Vaccinium spp. (vaccinia) 3. ,88 2. , 47** 0. ,18 0. 84** 21 0. 15--1. 14 Vibernum edule (squashberry) 3. ,75 2. ,75** -0. ,75 0. 95 48 0. 33- -2. 53 *Diameter refers to DHW in all cases. **Best equation of the two possibilities (DHW and DLB). Best of the six equations tested. K> + 2 ° General form of the regression equation is y = a+b In x, where x. s..diameter, or (diameter) length. 207 douglasii), the equations for spirea were used; for Douglas fir, white spruce was used; for currants, vaccinia was used; for elderberry, black twinberry was used; and for soapalallie (Shepherdia canadensis) , alder was used. 7.2.4 Data Analysis Floristic data were analyzed as follows. For the "C" layer, canopy-coverage ratings were converted to their mid-point percent coverage values (see Table 7.1). Then data from all sites that represented each stand were summed and divided by the total number of quadrats (50 times number of sites). Frequency of occurrence was the proportion of all quadrats for a stand in which a species was recorded, expressed as a percentage. It is important to remember that this statistic is strongly affected by the area and shape of the quadrat used in sampling (Greig-Smith 1964). However, as the 0.2 - x 0.5-m sampling frame is commonly used in western North America, many opportunities for comparison exist. For the purposes of. this study, only "major" plant species are presented. A major species was defined as one that had canopy-coverage greater than five percent in at least one of the stations representing a stand. Complete data are on file at the Wildlife Research Section office of the Fish and Wildlife Branch, Parliament Buildings, Victoria, B.C. 208 For the shrub and tree layers, species composition was based on stem counts. Phytomass estimates for the "C" layer vegetation was based on weight data for each serai stage. Shrub weight estimates for the "B" layer were based on logarithmic regression equations. As Baskerville (1972) and other foresters have pointed out, these types of predictions are biased. This bias was not corrected for in my results, but recent work by Brown (1976a)indicated that it was likely small, probably averaging five percent. Height estimates were unbiased since prediction equations utilized arithmetic rather than logarithmic data. Height and weight estimates were based on stem diameter measure ments from all 2- x 2-m quadrats for each serai stage. 7.3 Results for Mesic Upland  Plant Communities 7.3.1 The Data Base  and its Presentation Results were based on data collected from 51 sites. Thirty-four sites were in the Deserters environmental unit (mesic-till), with 15 in mesic lacustrine environments (Berman and Bednesti) and two in hydric till. Distribution by age of stand was wide, ranging from 1 year after logging to 200 year old spruce-pine stands (Table 7.7). Except for eight sites, sampled stands had originated from wildfire or from forests that were clearcut and burned. 209 Table 7.7 Distribution of Sampling Sites for the Plant Succession Study Environment -substrate Nominal age (yr) Site number (age in years in parenthesis) Mesic-till 1 MF2 (1) , MF5(1), MF6(2), Ml(l), M3(l) 5 MF8(4), MF10(5) 10 MF22 (12) , Gl (11) , G4(11) 25 MF19(?)*, MF21(23) 45 MF4(39) , G2(41) , G3 (41) , SRI(45), SR3(52) 75 SR4(69) 110 MF1(116), MF9(106), SR9(119), SR11(110) 135 SR2(130).*, SR22 (137), SR23 (134) 150 MF3(167)*, SR5(148), SR13(151)* 200 MF7(195), MF1K179)*, M2 (195) , M4(195), SR6 (174), SR10 (183) Hydric-till 110 SR16(115) 135 SR15(135) Mesic-lacustrine 1 MF12 (1) 5 MF15(4), MF17(4) 10 MF13(8), MF18(12) 25 MF20 (21) 45 SR12(54) 110 MF14 (101) , SR18(119) 150 MF16(154), SR8(153), SR14(140)*, SR17(151)*, SR20(142)*, SR21(149) Total 51 sites sampled *Partially logged stands. Others either wildfire, or clearcut and burned. 210 Results are presented in five subsections. The first one describes the serai stages floristically, that is, in terms of plant species composition and abundance. Each serai stage is described according to changes to the three vegetation layers: Tree, Shrub and Herb or A, B and C, respectively. The second subsection documents more detailed changes in the Tree layer from early to mature forest stages. Trends in species composition, height, basal area, and crown canopy closure are the major aspects treated. The third subsection deals in detail with temporal changes in the Shrub layer. Particular attention is paid to changes in phytomass and height. The fourth subsection examines the Herb layer, especially for changes in phytomass. The final subsection provides data on the net primary productivity of the vascular understory vegetation. Results are presented for all the understory, that is, for B and C layers combined. As shrubs form the major part of the winter diet of moose, most of the data are for woody plants. 7.3.2 Floristic Changes  in Serai Succession Floristic changes relates to variations in plant species composition and abundance recorded for seres on till and lacustrine substrates. Results are presented for the A, B and C layers of each serai stage. Serai plant communities were identified by three-part names (Table 7.8). Each name represents the most abundant species in the three 211 vegetation strata, beginning with the "A" layer. Only two names are provided when the tree layer was missing. Table 7.8 Plant Community Names for Successional Stages on Till and Lacustrine Substrates Nominal stand age Plant (yr) till substrate community name lacustrine substrate 1 raspberry-geranium aspen-sedge 5 spirea-fireweed saskatoon-aster 10 willow-pine willow-willow 25 pine-willow-twinflower aspen-willow-sarsparilia 45 pine-rose-spirea aspen-spirea-bunchberry 75 pine-twinberry-spruce 110 spruce-twinberry-subalpine fir pine-subalpine fir-bunchberry 135 spruce-vaccinia-subalpine fir 150 spruce-subalpine fir-cloudberry pine-spirea-bunchberry 200 spruce-thimbleberry-bunchberry Partially logged stands 135 spruce-vaccinia-bunchberry 150 spruce-maple-bunchberry subalpine fir-vaccinia-bunchberry 200 subalpine fir-squash-berry-subalpine fir 212 Plant succession proceeded through herb-, shrub- and tree-dominated stages. This pattern is typical of secondary forest succession whether it originates from old fields, wildfires, or clearcutting and slashburning. The following paragraphs highlight important changes and differences between studied serai stages (Floristic data for each site are presented in Appendix F) . Results are presented in Tables 7.9 to 7.14 and selected successional stages are illustrated in Figure 7.2. Till-year 1: Raspberry-Geranium The A and B layers were absent. The C layer was often patchy, depending upon the fire intensity, amount of residual slash and site disturbance. Also, species composition was variable, depending upon the foregoing factors and the species composition and abundance of the pre-disturbance plant community. However, raspberry and geranium (Geranium bioknellii) were the most common pioneer species. Other common shrub species in the C layer were rose and flat-top spirea; coniferous species were typically absent or very inconspicuous. Geranium was abundant particularly in year 1, but decreased sharply by year 2. Sarsaparilla (Aralia nudioaulis), f ireweed (Epilobium angustifolium) , bunchberry (Cornus canadensis), sedges (Carex spp.), and bentgrass (Agrostis spp.) were common, having average cover exceeding 5 percent or frequency of occurrence Table 7.9 Percent Canopy-Coverage/Frequency of Occurrance Values for Major* Plant Species of the Herb (C) Layer in a Sub-Boreal Forest Sere in a Mesic Environment on the Till Substrate Nominal successional age (yr) 1 5 10 25 45 75 110 135 150 200 MAJOR GROUP Species EVERGREEN TREES AND SHRUBS: Abies lasiocarpa (subalpine fir) - - 1/2 - t/1** t/4 15/35 21/54 14/32 7/19 Picea glauca (white spruce) - t/1 4/24 3/10 3/5 9/30 t/3 t/1 1/2 2/6 Pinus aontorta (lodgepole pine) - 1/9 25/46 - - 1/4 — — — — DECIDUOUS TREES AND SHRUBS: Acer glabrum (Douglas maple) - - - - - - - - - -Betula papyrifera (paper birch) - 1/10 - 2/14- - - - 1/2 — — Lonicera involucrata (black twinberry) - - - - - - - - - — Populus tremuloides (trembling aspen) - t/2 - 1/4 - - - - - -Rosa spp. (rose) 4/21 5/16 10/50 2/14 7/36 4/18 t/4 1/9 t/2 2/10 Rubus idaeus (raspberry) 7/29 7/36 t/2 - - - - t/1 t/2 2/9 R. parviflorus (thimbleberry) 3/9 7/28 - - 3/9 2/8 2/11 5/21 5/24 4/13 Salix spp. (willow) - 2/14 9/32 - t/1 - - - - -Spiraea lucida (flat-top spirea) 5/27 6/28 7/52 - 11/56 8/36 3/19 4/22 - 3/24 S. douglasii (hardhack) .- - - - - 2/24 1/5 - 1/12 -Vaccinium caespitoswn (dwarf blueberry) - - t/2 - - - - - - -V. membranaceum (mountain bilberry) t/5 - 2/20 - t/3 2/12 3/19 16/57 3/22 5/20 Viburnum edule (squashberry) t/1 1/5 t/2 - 1/4 - 2/7 t/4 1/12 t/1 FORBS AND DWARF SHRUBS: Achillea millefolium (yarrow) - - 1/14 - - 1/4 - - - -Aralia nudicaulis (sarsaparilla) 3/29 5/32 - - 2/20 - 5/24 7/38 16/66 7/30 Aster spp. (aster) t/3 1/7 7/36 - 1/5 2/8 t/1 - - 1/3 Clintonia uniflora (queen's cup) t/5 1/9 1/10 - 1/4 1/14 5/25 7/51 2/28 4/25 Cornus canadensis (bunchberry) 4/43 17/77 12/58 - 7/51 4/40 8/46 13/79 10/72 12/71 Epilobium spp. (fireweed) 8/43 21/79 2/18 - 2/8 t/2 - t/2 1/6 t/2 Table 7.9, Continued Nominal successional age (yr) MAJOR GROUP Species 1 5 10 25 45 75 110 135 150 200 Galium (bedstraw) _ 1/9 t/4 — - t/2 1/7 1/3 1/6 t/1 Geranium bioknellii (Bicknells geranium) 21/49 t/1 - - - - - - - -Hypochaeris radieata (cat's ear) - - 6/54 - - - - - - -Linnaea borealis (twinflower) t/4 3/5 14/52 18/56 9/44 3/18 1/11 6/28 2/18 4/25 Retasites frigida (coltsfoot) - - - - 1/8 3/8 t/5 t/1 - 1/6 Rubus chamaemorus (cloudberry) - - - t/2 • - 4/24 2/13 3/21 19/88 1/2 R. pedatus (trailing rubus) - • - - - - - 3/15 1/7 - 2/9 Streptopus amplexifolius (twisted stalk) - - - - 3/16 6/40 3/14 4/24 2/6 3/10 Trifolium repens (white clover) - - - - - - - - - -GRAMMINOIDS Agrostis alba (bentgrass) 6/16 - 10/34 - - - - - - t/1 Calamagrostis spp. (reedgrass) 6/2 t/1 3/10 - 10/17 - - - - -Carex spp. (sedge) 4/26 1/7 6/20 - 4/8 - - 6/1 - -OTHER TAXA: Dryopteris austriooa (spiny wood-fern) t/1 - - - t/1 3/10 3/17 7/32 16/58 4/17 Equisetum spp. (horse tail) - - - - 1/3 4/22 2/11 2/8 2/6 4/15 No. spp. recorded ..27 31 29 7 32 28 32 35 24 38 No. sites sampled (quadrats) 3(150) 2(100) 3(150) 1(50) 2(100) 1(50) 3(150) 3(150) 1(50) 3(150) *Species with at least one site where canopy-coverage s > 5% and frequency of occurrence > 20%. **t is less than 0.5% coverage. Table 7.10 Percent Canopy-Coverage/Frequency of Occurrence Values for Major* Plant Species of the Herb (C) Layer in a Sub-Boreal Forest Sere in a Mesic Environment on the Lacustrine Substrate Nominal successional age (yr) MAJOR GROUP Species 1 5 10 25 45 110 150 5/12 t/2* 1/3 12/32 7/18 - - 5/17 - 9/13 1/4 2/6 1/5 t/2 1/1 - -5/27 3/6 1/1 - -1/6 . 2/14 4/18 4/10 5/19 1/6 4/21 6/28 5/15 7/14 t/4 - - 6/1 6/36 7/52 9/43 1/10 7/33 2/17 5/29 - t/4 2/15 t/2 - - -- 1/8 1/7 1/8 1/1 t/2 1/4 t/2 1/3 19/50 7/16 3/12 - t/1 1/18 1/9 3/14 4/12 5/28 7/32 12/49 2/12 9/26 1/4 - 6/33 t/1 2/4 9/60 1/12 3/16 - - 1/6 1/2 - - - t/2 7/30 12/42 6/25 — 2/10 1/4 5/28 1/2 — 1/9 1/5 5/29 t/2 1/2 t/3 - 1/8 - 18/62 3/15 1/8 3/21 4/30 11/60 5/37 - 1/6 2/16 7/33 - 1/6 1/3 2/2 - 1/12 5/37 EVERGREEN TREES AND SHRUBS: Abies lasiocarpa (subalpine fir) Picea glauca (white spruce) Pinus contorta (lodgepole pine) DECIDUOUS TREES AND SHRUBS: Acer glabrum (Douglas maple) Betula papyrifera (paper birch) Lonicera involucrata (black twinberry) Populus tremuloiales (trembling aspen) Rosa spp. (rose) Rubus idaeus (raspberry) R. parviflorus (thimbleberry) Salix spp. (willow) Spiraea lucida (flat-top spirea) S. douglasii (hardhack) Vaccinium caespitosum (dwarf blueberry) V. membranaceum (mountain bilberry) Viburnum edule (squashberry) FORBS AND DWARF SHRUBS: Achillea millefolium (yarrow) Aralia nudicaulis (sarsaparilla) Aster spp. (aster) Clintonia uniflora (queen's cup) Table 7.10, Continued MAJOR GROUP Species Nominal successional age (yr) 1 5 10 25 45 110 150 Cornus oanadenis (bunchberry) 9/80 7/58 9/63 13/68 9/63 27/90 16/79 Epilobium spp. (fireweed) - 3/25 9/65 8/50 1/5 t/5 t/4 Galium spp. (bedstraw) - 3/22 7/21 1/10 1/8 1/9 2/17 Geranium bioknellii (Bicknells geranium) •t/2 2/17 t/1 t/2 - - -Hypochaeris radieata (cat's ear) - - 1/1 - - - -Linnaea borealis (twinflower) 1/8 t/6 t/3 t/2 7/48 7/35 3/23 Petasites frigida (coltsfoot) 3/58 8/59 9/44 2/28 4/36 9/52 5/44 Rubus chamaemorus (cloudberry) - 6/40 3/17 t/2 3/20 2/11 1/7 R. pedatus (trailing rubus) 1/8 8/28 1/4 t/2 1/5 1/4 3/13 Streptopus amplexifolius (twisted stalk) - - - - 1/7 - 3/5 Trifolium repens (white clover) - - 1/2 8/32 - - -GRAMMINOIDS: Agrostis alba (bentgrass) t/2 - - - - - -Calamagrostis spp. (reedgrass) 2/6 9/21 3/13 7/32 - 1/10 t/1 Carex spp. (sedge) 18/64 1/10 2/11 2/10 1/2 - -OTHER TAXA: Dryopteris austrioaa (spiny wood-fern) - - - - 1/14 1/7 11/20 Equisetum spp. (horse :tail) - - - - 5/26 2/14 2/8 No. spp. recorded 20 33 36 32 33 33 39 No. sites sampled (quadrats) 1(50) 2(100) 2(100) 1(50) 2(100) 2(100) 3(150) *Species with at least one site where canopy-coverage s > 5% and frequency of occurrence k 20%. **t is less than 0.5% coverage. 217 Table 7.11 Percent Canopy-Coverage/Frequency of Occurrence Values for Major* Plant Species of the Herb (C) Layer in Partially Logged Sub-Boreal Forest Stands in a Mesic Environment on Till and Lacustrine Substrates Till Lacustrine MAJOR GROUP Species 135 150 200 EVERGREEN TREES AND SHRUBS: Abies lasiocarpa (subalpine fir) Picea glauca (white spruce) Pinus contovta (lodgepole pine) DECIDUOUS TREES AND SHRUBS: Acer glabrum (Douglas maple) Betuhx papyrifera (paper birch) Lonicera involucrata (black twinberry) Populus tremuloides (trembling as Rosa spp. (rose) Rubus idaeus (raspberry) R. parviflorus (thimbleberry) Salix spp. (willow) Spiraea lucida (flat-top spirea) S. douglasii (hardhack) Vacoinium caespitosum (dwarf blueberry) V. membranaceum (mountain bilber Viburnum edule (squashberry) FORBS AND DWARF SHRUBS: Achillea millefolium (yarrow) Aralia nudicaulis (sarsaparilla) Aster spp. (aster) Clintonia uniflora (queen's cup) Cornus canadensis (bunchberry) Epilobium spp. (fireweed) Galium spp. (bedstraw) Geranium bicknellii (Bicknell's geranium) Hypochaeris radicata (cat's ear) Linnaea borealis (twinflower) Petasites frigida (coltsfoot) Rubus chamaemorus (cloudberry) R. pedatus (trailing rubus) Streptopus amplexifolius (twisted stalk) Trifolium repens (white clover) CRAMMINOIDS: Agrostis alba (bentgrass) Calamagrostis spp. (reedgrass) Carex spp. (sedge) OTHER TAXA: Dryopteris austrioca (spiny woe Equisetum spp. (horse tsil) No. spp. recorded No. sites sampled (quadrats) 16/50 10/32 16/36 1/2 1/6 1/6 150 16/34 1/3 8/22 8/15 1/1 3/5 1/4 3/8 2/9 t/21* 2/6 , 1/3 pen) - 1/3 - 1/5 2/22 2/13 3/12 1/7 E/2 - - 1/2 11/45 9/41 3/14 4/19 - t/4 -7/32 9/36 8/38 5/25 2/14 f) 26/90 9/38 -11/37 t/2 3/11 1/2 t/3 7/52 5/28 4/32 7/40 t/1 6/20 2/8 10/70 3/20 1/12 3/38 21/92 11/70 12/58 17/81 t/5 3/14 2/18 1/6 2/10 t/5 t/6 t/1 13/47 7/32 9/38 11/35 1/4 •- -3/20 t/1 - 2/25 1/6 1/5 1 1/6 2/15 1/5 . t/2 -- --fern) t/2 4/16 2/10 - 2/14 1/10 t/1 25 32 26 37 1(50) 2(50) 1(50) 3(150) where canopy -coverage s 1 5% and frequency of occurrence 2. 20%. **t is less than 0.5% coverage. Table 7.12 Percent Species Composition (stem-basis) of the Shrub (B) Layer in a Sub-Boreal Forest Sere in a Mesic Environment over Till Substrates Nominal succes ssional age (yr) Species 1 5 10 25 45 75 110 135 150 200 Abies lasiocarpa (subalpine fir) 16 15 32 15 Acer gldbrum (Douglas maple) 11 2 11 Alnus spp. (alder) 3 13 6 Amelanchier alnifolia (Saskatoon) 1 15 26 Betula papyrifera (paper birch) 4 30 Cornus stolonifera (red-osier dogwood) 2 5 Lonicera involucrata (black twinberry) 14 6 32 22 Picea glauca (white spruce) 5 15 17 2 Pinus contorta (lodgepole pine) 2 15 Populus tremuloides (trembling aspen) 5 2 1 P. balsamifera (black cottonwood) Pseudotsuga menziesii (Douglas fir) 1 2 Ribes spp. (currant) 4 10 2 2 16 5 Rosa (rose) 17 16 5 7 37 3 5 16 9 Rubus idaeus (raspberry) 73 3 R. parviflorus (thimbleberry) 3 6 9 11 31 Salix spp. (willow) 1 56 48 Sambucus racemosa (elderberry) 5 Shepherdia canadensis (soapalallie) 11 Sorbus spp. (mountain ash) 5 11 2 2 Spiraea lucida (flat-top spirea) 23 4 3 6 9 19 5 S. douglasii (hardhack) 28 10 5 16 Vaccinium spp. (vaccinia) 3 4 51 5 1 Viburnum edule (squashberry) 11 11 5 10 No. stems 41 192 55 61 64 31 174 47 19 81 Sites 3 2 1 1 3 1 4 2 1 3 Table 7.13 Percent Species Composition (stem-basis) of the Shrub (B) Layer in a Sub-Boreal Forest Sere in a Mesic Environment over Till Substrates Nominal successional age (yr) Species 1 5 10 25 45 110 150 Abies lasiooarpa (subalpine fir) 5 23 2 Acer glabrum (Douglas maple) Alnus spp (alder) Amelanchier alnifolia (Saskatoon) 23 8 13 Betula papyrifera (paper birch) 6 5 Cornus stolonifera (red-osier dogwood) 4 2 Lonicera involucrata (black twinberry) 7 13 3 16 . 21 6 Picea glauoa (white spruce) 7 Pinus contorta (lodgepole pine) 1 Populus tremuloides (trembling aspen) 93 18 4 6 P. balsamifera (black cottonwood) 2 Pseudotsuga menziesii (Douglas fir) Ribes spp. (currant) 3 1 Rosa spp. (rose) 19 1 20 21 15 Rubus idaeus (raspberry) R. parviflorus (thimbleberry) 2 4 Salix spp. (willow) 4 72 83 8 1 Sambucus racemosa (elderberry) Sheperdia canadensis (soapalallie) Sorbus spp. (mountain ash) 4 Spiraea lucida (flat-top spirea) 3 36 6 30 S. douglasii (hardhack) 15 19 18 Vaooinium spp. (vaccinia) 2 Viburnum edule (squashberry) 1 2 7 No stems (sites) 30(1) 113(2) 318(2) 64(1) 25(1) 48(2) 188(3) AO 220 Table 7.14 Percent Species Composition (stem-basis) of the Shrub (B) Layer in Partially Logged Stands of Sub-Boreal Forests in a Mesic Environment Till Lacustrine Species 135 150 200 150 Abies lasiocarpa (subalpine fir) 8 14 6 Acer glabrum (Douglas maple) 19 26 6 Alnus spp. (alder) Amelanchier alnifolia (Saskatoon) Betula papyrifera (paper birch) 1 3 3 Cornus stolonifera (red-osier dogwood) Lonioera involucrata (black twinberry) 3 17 6 Picea glauoa (white spruce) 3 Pinus contorta (lodgepole pine) Populus tremuloides (trembling aspen) 2 9 7 P. balsamifera (black cottonwood) 1 Pseudotsuga menziesii (Douglas fir) Ribes spp. (currant) 3 Rosa spp. (rose) 3 12 6 1 Rubus idaeus (raspberry) R. parviflorus (thimbleberry) 2 10 Salix spp. (willow) Sambuous raoemosa (elderberry) Shepherdia canadensis (soapalallie) Sorbus spp. (mountain ash) f 2 6 3 Spiraea lucida (flat-top spirea) 4 9 19 S. douglasii (hardhack) 14 Vacoinium spp. (vaccinia) 43 16 21 Viburnum edule (squashberry) 14 11 54 18 No. of stems (sites) 37(1) 125(2) 35(1) 72(3) 220a Figure 7.2 Photographs illustrating selected successional stages on mesic till and lacustrine substrates. Page 221 shows till sites aged approximately 1(a), 10(b), 45(c) and 200(d) years. Page 222 shows lacustrine sites aged approximately 5 and 25 years. Page 22 3 compares till and lacustrine sites at approximately 40 years (a and b) and at (c and d), respectively. 221 223 224 greater than 25 percent or both. Grasses tended to be most common in early stages south of Prince George, and uncommon or absent in northern areas. Moss cover was variable; the liverwort, marchantia (Mavohantia polymorpha) occurred on the more severely burned sites but was not abundant. Till-year 5: Spirea-fireweed The Tree (A) layer remained undeveloped, but the Shrub (B) layer showed considerable change from year 1, with respect to species diversity and canopy-coverage. Douglas maple, black twinberry, rose, squashberry and flat-top spirea were the most