UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influence of the thermal environment on forest cover selection and activity of moose in summer Demarchi, Michael W. 1992

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

Item Metadata

Download

Media
831-ubc_1992_spring_demarchi_michael.pdf [ 3.44MB ]
Metadata
JSON: 831-1.0075203.json
JSON-LD: 831-1.0075203-ld.json
RDF/XML (Pretty): 831-1.0075203-rdf.xml
RDF/JSON: 831-1.0075203-rdf.json
Turtle: 831-1.0075203-turtle.txt
N-Triples: 831-1.0075203-rdf-ntriples.txt
Original Record: 831-1.0075203-source.json
Full Text
831-1.0075203-fulltext.txt
Citation
831-1.0075203.ris

Full Text

INFLUENCE OF THE THERMAL ENVIRONMENTON FOREST COVER SELECTION AND ACTIVITY OF MOOSEIN SUMMERByMICHAEL WILLIAM DEMARCHIB.Sc., The University of Victoria, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTSFOR THE DEGREE OF MASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Forest Sciences)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADecember 1991© Michael William Demarchi, 1991In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignatureDepartment of ^F-Dre..ri The University of British ColumbiaVancouver, CanadaDate  be C€44A6e—r 2-7 r rig( DE-6 (2/88)iiABSTRACTI investigated summer thermal cover and the influence of the thermalenvironment on habitat selection and activity of moose (Alces alces).Hemispherical photography was used to estimate the sky view factor (SVF) andeffective leaf area index (Le) of coniferous stands as a function of crown closureclass (CCC) values from forest cover maps. Moosehorn readings taken athemispherical photo sites indicated that the CCC scheme correctly ranked standsby canopy closure. The stand attributes of SVF and Le, together with weatherdata collected in the study area, were entered into a model that simulated theoperative temperature (Te) experienced by a moose. 'Hot' conditions existedwhen Te in the open (Teopen) exceeded the upper limit of the upper criticaltemperature (UCT) range of moose (Te >29.5 °C). 'Cool' conditions existedwhen Teopen was below the lower limit of the UCT range (Te <13 °C). Otherresearch has shown that moose are very prone to thermal stress. During thestudy, weather conditions were encountered that had the potential to thermallystress moose. Simultaneous Te values declined exponentially with increasingCCC, indicating that a gradient of thermal cover existed across CCCs. AtCCCs greater than 4, little additional thermal cover value was realized.Two hundred and fifty two radio locations were made on four adult cowmoose. Because 'hot' and 'cool' weather conditions corresponded to 'light' and'dark' conditions respectively, the effects of heat and light on habitat selectionwere indirectly assessed. Moose selected increased cover during 'hot' (thus'light') conditions (p < 0.05). The patterns of habitat selection during 'light'conditions indicated that relative use of CCC = 0 sites increased significantly (p <0.05) as Teopen decreased. Shade from willow canopies and convection bywater may have allowed or caused moose to use some CCC = 0 sites wheniiiambient conditions exceeded the UCT. When Teopen was 'hot' (thus 'light'conditions), moose located under coniferous cover tended to remain under suchcover during subsequent locations. These observations support the conclusionthat the thermal environment influenced habitat selection. No correlationsbetween the moose location attributes of distance to an edge/water/road andtime of day or Te open were found.Moose activity/inactivity was inferred from modulating/non-modulatingradio signals 326 times. Moose were more active during 'cool' (thus 'dark')conditions (p < 0.05). A negative correlation (r = -0.47) between the percent ofactive locations and mean Teopen for the hours of 11:00 until 24:00 PDTindicated that moose activity was likely thermally constrained. The mean straightline velocity (MSLV) between successive moose locations did not differ acrosshours from 11:00 until 24:00 (p > 0.05). MSLV was not correlated with Teopen ;however, a decrease in average MSLV values from 12:00 until 17:00 didcorrespond to sustained conditions of Te open values above the UCT.The susceptibility of moose to heat stress, the effectiveness of coniferstands in providing thermal cover, and the relations between habitat selectionand the thermal environment suggest that thermal cover is a required,manageable component of cow moose summer range in the study area.ivTable of ContentsABSTRACT^ iiList of Tables vList of Figures viACKNOWLEDGEMENTS^ ixCHAPTER I: GENERAL INTRODUCTION^ 1PROBLEM AND OBJECTIVES 3STUDY SITE^ 3CHAPTER II: QUANTIFYING THE THERMAL ENVIRONMENT^5INTRODUCTION 5METHODS AND MATERIALS^ 7Weather station ^ 7Hemispherical photography 7Moosehorn 9Equivalent blackbody temperature model^ 10RESULTS AND DISCUSSION^ 11Weather station ^ 11Forest canopy 15Thermal cover 23CHAPTER III: COVER SELECTION AND ACTIVITY OF MOOSE^28INTRODUCTION^ 28METHODS AND MATERIALS^ 31Moose telemetry 31Analysis of telemetry data 33RESULTS AND DISCUSSION^ 35Habitat analysis ^ 35Habitat use 35Moose activity 46CHAPTER IV: SUMMARY AND CONCLUSIONS^ 54LITERATURE CITED^ 57APPENDIX I^ 62vList of TablesTable 2.1. Crown closure class (CCC), sky view factor (SVF), mean crowncompleteness (MCC) and effective leaf area index (Le) of each site wherehemispherical photos were taken. CCC values are from the forest cover map,SVF and Le values from hemispherical photos, MCC from moosehornreadings. Parenthesized CCC values represent estimates based on therelative location of mean SVF for those four locations compared to the meanSVF of CCC = 3 and the theoretical point CCC = 0, SVF = 1 (see Methodsand Materials p. 9).  16Table 2.2. Crown closure class codes and their corresponding percent limits ofcoverage. This scheme of cover classes, used by the B.C. Ministry ofForests, does not account for gaps within individual tree crowns.  18viList of FiguresFigure 2.1. Boxplot summaries of hourly means of air temperature, wind speed andsolar radiation for the period of 90/06/17 to 90/09/23. Each plot shows themedian and the range of the quartiles. Star points are identified by SYSTAT(Wilkinson 1990) as outliers. Note: complete weather data were not obtainedfor the period of 90/07/03 to 90/07/07. N = 2247.  12Figure 2.2. Boxplot summaries of hourly means of air temperature, wind speed andsolar radiation for hours when moose activity was determined by telemetry.See Figure 2.1 for boxplot description. N = 326.   13Figure 2.3. Boxplot summaries of hourly means of air temperature, wind speed andsolar radiation for hours when moose locations were determined by telemetry.See Figure 2.1 for boxplot description. N = 252. 14Figure 2.4. Trend of the averages of mean crown completeness for crown closureclasses 3 to 8. Error bars denote mean ±1 SE. The dashed line is the line ofleast squares and is given only to illustrate the trend (r 2 = 0.87; p < 0 ^05)^ 17Figure 2.5. Relation between the mean sky view factor (SVF) and crown closureclasses 1 to 8. Error bars denote ±1 SE. The dashed line is the least squareregression from a 1/SVF data transformation ( Y 1 / [0.844 + 1.057 X], r 2 =0.85). The solid line represents the least square regression forced throughSVF = 1 from the same transformation ( Y = 1 / [1 + 1.03 • X], r 2= 0.85 ).  19Figure 2.6. Relation between effective leaf area index means and crown closureclasses 1 to 8. Error bars represent ±1 SE. The solid line represents theleast square regression. The dashed line represents the theoretical relationbetween the two variables (see eqn. 1 in text).  21viiFigure 2.7. Mean leaf area index against mean crown completeness for each ofcrown closure classes 3 to 8. The dashed line is the theoretical function givenby eqn. 1 (assuming a spherical distribution of foliar elements - see text). Thesolid line corresponds to eqn.1 when foliar elements in the vertical planepredominate (i.e., G <0.5).  22Figure 2.8. The Relation between sky view factor (SVF) and effective leaf areaindex (Le). The dashed line is the least square regression from a log(SVF)transformation ( Y = 10 0.03410 0.33x)  12 = 0.75 ).) The solid line is the leastsquare regression forced through the theoretical point Le = 0, SVF = 1 fromthe same transformation (Y = [10 0.329-x, 12 = 0 .74 ).  24Figure 2.9. Mean operative temperatures by crown closure class (CCC). Weatherdata (Fig. 2.1) were run through the simulation model (Appendix I) under twoconditions: (A) wind-speed (as recorded in the open) was held constantacross all CCCs, (B) wind-speed in CCCs >0 was held constant at 0 m•s "2 .The means for simulations A and B are represented by the lower and upperlimits of the solid rectangles respectively. Since CCC = 0 was only simulatedunder condition A, error bars denote ±1 SD. Error bars in the negative ydirection denote 1 SD for the simulations of A. Error bars in the positive Ydirection denote 1 SD for the simulations of B.  25Figure 3.1. Percentages of the entire study area (AREA, 15 970 ha) and habitatwithin 800 m of a road (HABITAT, 9855 ha) represented by each crownclosure class. For crown closure limits see Chapter II, Table 2.2.  36Figure 3.2. Percentage of the total area (ha) contained within telemetry overlaysrepresented by each crown closure class for locations when Teopen was hot(light)(n = 124) and Teopen was cool (dark)(n = 56).  38Figure 3.3. Percentage of telemetry overlays containing each crown closure classwhen hourly Teopen values were hot (n = 124) and Teopen was cool (n = 56). 39viiiFigure 3.4. Simultaneous differences between the regressions of Te versus airtemperature for CCC = 6 and CCC = 0 sites, as a function of air temperaturefrom 14 to 30 °C  41Figure 3.5. Boxplot summary of hourly Te open values when moose were located forthe hours of 11:00 until 24:00 (n = 252). Each boxplot shows the median andthe range of the quartiles. Star and circle points are identified by SYSTAT(Wilkinson 1990) as outliers. The lines corresponding to the UCT (dashed)and the thermal limit for open-mouthed panting (solid) identified by Reneckerand Hudson (1986) (see text p. 32).  43Figure 3.6. Percentage of telemetry overlay areas in CCC = 0 and CCC for theperiod from 11:00 until 24:00 (n = 243 locations). The relation between theratio of CCC = 0 to CCC ?..4 overlay areas and hour is significant (r 2 = 0.63, p< 0.05).  44Figure 3.7. Percentage of successive radio locations (i.e., <120 minutes apart)represented by cover type at time = t and cover type at time = t + 1 whenTeopen was hot (n = 83) and Teopen was cool (n = 37).  45Figure 3.8. Percentage of active and inactive radio locations under Teopen hot(light) (n = 155) and Te open was cool (dark) (n = 77).  47Figure 3.9. Average 'mean straight line velocity' between successive mooselocations from 11:00 until 24:00. Error bars denote ±1 standard deviation (n =209). The curved line represents the least square regression of operativetemperature in open areas against hour for times when moose were located(n = 242; r 2 = 0.75; p < 0.05).  50Figure 3.10. Percentage of 'active' moose locations as determined by radiotelemetry, for the hours of 11:00 until 24:00 p.m. (n = 301).  51Figure 3.11. Frequency distribution of active and inactive radio locations for 12straight-line velocity intervals. Each velocity interval corresponds to a rangeof 1 m•min -1 (e.g., velocity interval 1 = 0.00 to 0.99 m•min -1 , 2 = 1.00 to 1.99rn•min -1, etc.) 52ixACKNOWLEDGEMENTSMany thanks to my supervisor, Dr. F.L Bunnell, for granting me the meansto this achievement and for the many thought-provoking discussions. My father,Ray, whose dedication to the wildlife resources of this province continues to be agreat source of inspiration. My mother Marilyn and her husband Gerry haveprovided much support throughout my academic career. Dr. T.A. Black, Dr. M.Pitt, and Dr. T. Sullivan contributed to the study design and data analyses. Dr.J.M. Chen, Dr. M. Greig, Rob McCann, and I. Simpson gave valuable technicalsupport. J. Maedel and Dr. P. Murtha provided access to the geographicinformation system (GIS) and technical advice. K. Simpson providedaccommodations, telemetry equipment, and radio-collared animals. EnvironmentCanada donated weather equipment. The B.C. Ministry of Forests donated useof hemispherical photography equipment. The staff at the Ministry ofEnvironment office in Cranbrook provided logistical support for the final stages ofthe write-up. B. Clark and P. Bodley assisted with data collection.The project was funded by: the National Sciences and EngineeringResearch Council of Canada through a grant to Dr. F.L. Bunnell, the B.C.Ministry of Forests, Gorman Brothers Lumber, Ardew Wood Products, FletcherChallenge Canada, and Weyerhauser Canada. B. Christie and F. Harper helpedsolicit industry funding. The B.C. Habitat Conservation Fund and Mrs. J.Johnson coordinated the financial accounts. Fellowships from Canadian ForestProducts and the Van Dusen Forestry Fund provided personal support.Thanks to the 'Header House Crew' for the relaxed (usually!) yetconstructive atmosphere while I was on campus.1CHAPTER I: GENERAL INTRODUCTIONAt an elementary level, an animal's needs consist of food, water, andcover. Beyond this level, from each need there emanates a complex network ofprocesses and interactions which facilitate fitness-promoting actions.Knowledge of cover as it applies to large North American ungulates hasbeen deficient. This deficiency stems from cover's extensive and oftenoverlapping functional roles (e.g., security, escape, thermal, etc.) which aredifficult to quantify and relate to specific habitat attributes. Of the types ofungulate cover defined by North American managers and researchers, summerthermal cover has received an increasing amount of attention in the literature(e.g., Thomas et al. 1979; Bunnell et al. 1985; Smith and Long 1987; Ritcey et al.1988; Timmerman and McNicol 1988; Nyberg and Janz 1990). Ideally,managers' concepts of thermal cover are based on research into thephysiological needs of animals relative to specific environmental conditions.Having established an apparent need for thermal cover, the ability of differenthabitats to meet that need can be assessed. For example, habitat units could beranked according to a scale of thermal cover efficacy. If thermal cover isrequired, then habitat units to which we ascribe thermal cover value should beutilized predictably by wild animals. Failure to reveal a predictable pattern ofhabitat use could stem from one of several causes: thermal cover is not used, ourdefinition of thermal cover needs refining, or the metabolic cost ofthermoregulation (i.e., by being in a thermally-stressing environment) isoutweighed by some benefit (e.g., increased forage consumption).Extensive research into the physiological effects of heat on domesticungulates has been used for many years to provide optimal environments forsuch animals (e.g., Sainsbury 1967; Esmay 1969). By comparison, the amount2of comparable research on wild ungulates is meagre. Renecker and Hudson(1986) found that heart, respiratory, and metabolic rates of moose (Alves aloesL.) increased exponentially with ambient temperature. Furthermore, duringwarm summer periods feed intake was reduced and body mass subsequentlylost. Their observations suggest that moose should seek thermal cover tomitigate these negative effects. Several . authors have investigated relationsbetween moose habitat selection and the thermal environment. However, if needcannot be defined (e.g., by physiological experimentation), observations onhabitat use can only imply requirements because they fail to discriminatebetween requirement and preference (Peek et al. 1982).Most research on the effects of heat on wild ungulates has been done onfree-ranging animals in uncontrolled environments. de Vos (1958) reported thatalthough no correlation between ambient temperature and moose use of aquatichabitats was detectable, moose were more readily observed in the moming andevening than at midday. Schwab (1986) indicated that moose selected summerhabitats based on their ability to provide thermal cover; after examining his data,however, I feel his conclusions are untenable. Belovsky (1981), Ackerman(1987) and Renecker (1987) reported that moose habitat choice was correlatedwith the ambient thermal environment; however, the qualitative nature of thermalcover as these authors described it is of little use to habitat managers attemptingto provide summer thermal cover in the habitat mix.If summer thermal cover is a manageable habitat component required bymoose, there is a need to understand and quantify it well enough to manage for itwherever man is altering the landscape or managing for healthy moosepopulations. Conversely, if summer thermal cover is not a manageable habitatcomponent then time and money allocated to its management are wasted.3PROBLEM AND OBJECTIVESThe conviction that North American ungulates require summer thermalcover, and that this cover can be quantified by forest canopy closure, has beenpoorly supported by published literature.This research had two objectives:1) to determine the efficacy of forest stands of varying crown closuresin providing summer thermal cover,2) to determine if adult cow moose select habitats at the level of forestcover polygons based upon polygon-ability to provide thermalcover.Chapter II addresses the concepts of summer thermal cover and operativetemperature as applied to moose. The relations between ambient thermalenvironment and animal physiology documented by Renecker and Hudson(1986) and Parker and Robbins (1983) are used to assess the heat-stress-potential of conditions when moose were located. In Chapter III, patterns ofhabitat selection and activity of moose are examined in the context of thermalenvironment.STUDY SITEThe study site is located at approximately 50° N latitude, 120° W longitudein the Southern Thompson Upland Ecosection of the Southern InteriorEcoprovince (from British Columbia Ministry of Environment 1988). Thebiogeoclimatic zone is Montane Spruce (from British Columbia Ministry of Forests1988). Elevation ranges between 1400 and 1500 m. As a result of past fires,most forested sites in this zone are occupied by mature lodgepole pine (Pinuscontorta Dougi.) (Lloyd et al. 1990). Veteran and understory trees consist of4Englemann spruce (Picea englmannii Parry) and subalpine fir (Abies lasiocarpa(Hook.) Nutt.). The terrain is relatively level and much of the forested area isinterspersed with riparian communities. Riparian sites are dominated by sedge(Carex spp. L.), willow (Salix spp. L.) and glandular bog birch (Betula glandulosaMichx.). Forestry represents the major land use practice in this area. Fourtimber harvesting companies hold cutting rights within the study area. The areais also an important summer cattle range. Moose are present year-round(Keystone Bio-Research 1991), presumably because of a low mean annualsnowfall (270 cm; Mitchell and Greene 1981), and the abundance of deciduousbrowse.5CHAPTER II: QUANTIFYING THE THERMAL ENVIRONMENTINTRODUCTIONConiferous or deciduous overstories which shelter an animal frommeteorological processes are described as thermal cover (Black et al. 1976).Because thermal cover operates by moderating wind, precipitation and solarradiation, its quantification relies heavily on characterizing the distribution andamount of foliar elements. Norman and Campbell (1989) identify two broadmethodologies by which canopy structure is measured: direct and indirect. Directmethods encompass 'destructive sampling' techniques which require that plantorgans be clipped and measured. Indirect methods allow inference about canopystructure based on the measurement of canopy-influenced solar radiation andinclude the use of photometric sensors, hemispherical photography, andmoosehoms. Direct techniques tend to be avoided because much effort isrequired to obtain information from small areas. Conversely, indirect methodscan be used to sample large areas quite easily. Perhaps the greatest drawbackof indirect techniques is the requirement of a model which describes theinteraction between canopy attributes and radiation.The moosehorn (Robinson 1947; Bonnor 1967) is a simple tool whichquantifies tree crown cover within a conical field of view normal to the forest floor.Vales (1986) found that moosehorn estimates of forest overstory were highlypredictive of sub-canopy radiation regimes. Hemispherical photography iscommonly employed as an indirect means of determining plant cover and canopyradiation regimes (Chen et al. 1991). Having estimated the canopy values of skyview factor (SVF; Reifsnyder 1967; Kelliher 1985) and effective leaf area index(Le; Chen et al. 1991) from hemispherical photos, and given the total globalradiation under an open sky, the simultaneous radiation regime beneath a6canopy can be estimated (see Chen et al. 1991). Combining the radiation regimewith ambient air temperature, wind speed and physical properties of the bioticand abiotic environment, provides an estimate of the 'equivalent blackbodytemperature' or 'operative temperature' (Te; Campbell 1977). Te describes theair temperature of an environment with no wind and no net radiative input oroutput in which an animal would 'experience' the same thermal environment as itdid in its natural habitat. Essentially, it is the net radiative energy increment ordecrement to air temperature. Te and its similar variable 'standard operativetemperature' (Tes) (Bakken 1980) have been used by other researchers toestimate the thermal environment of wild ungulates (Schwab 1986; Parker andGillingham 1990).To assess the thermal cover value of forest cover-types, simultaneous Tevalues as a function of the crown closure class (CCC) of forest cover polygonswere estimated. A simulation model estimated the thermal environmentexperienced by a moose for a set of conditions which included: weather (airtemperature, solar radiation, and wind speed), date, time of day, and CCC. Thepurpose of these simulations was to expose the nature of the thermalenvironment under a range of canopy closures, thereby allowing hypothesesconcerning moose and thermal cover to be tested (Chapter III). CCC waschosen as the independent variable because of its importance in evaluating andmanaging wildlife habitat (e.g., Harcombe 1984). If summer thermal cover isfound to be important for moose, describing it in terms of CCC will allowmanagers to utilize a habitat attribute which is commonly described for theprovince's forests.The objective was to determine the efficacy of forest stands of varyingcrown closures in providing summer thermal cover for moose.7METHODS AND MATERIALSWeather stationA weather station was erected 90/06/17 in a meadow dominated by sedgeand glandular bog birch under 1.5 m in height. A Campbell Scientific CR21 datalogger recorded hourly averages of readings taken at 60-second intervals from alevelled Li-Cor pyranometer (cosine corrected, model LI200S), Met-Oneanemometer (model 014A) and temperature/humidity probe (Campbell Scientificmodel 207). Data were transferred to a cassette recorder and recorded on 60-minute normal bias tapes. The data logger was enclosed in a rain-tight fiberglasscase which was covered by a metal radiation shield. This assembly wasmounted at a height of 1.5 m on a steel mast. On the same mast, thepyranometer was mounted on the south side at a height of 2.5 m. Theanemometer was mounted at a height of 2.9 m. The temperature/humidity probewas placed inside a Stevenson screen on a mast at a height of 2.0 m. The entirestation was fenced to exclude cattle and moose. Weekly visits ensured that allcomponents were functioning properly. Data cassettes were read into a personalcomputer using software supplied by Campbell Scientific. The resulting ASCIIfiles were imported into SYSTAT (Wilkinson 1990).Hemispherical photographyUsing a tripod-mounted Nikon SE camera and a 180° fish-eye lense (8mm focal length), a minimum of three forest stands corresponding to each of theMinistry of Forests crown closure class codes (CCC) 3 to 8 were sampled.Kodalith Hi-Contrast, black and white, ASA 6 film was used. Photo sites were aminimum of three tree heights (i.e., >30 m in this study) from the nearest differentCCC or cover-type. Compass bearings and distances to photo sites were8determined from a 1:15 000 forest cover map. Photos were not taken directlyunder dead-falls or boughs near the ground. All photos were taken underconditions of no wind. Most photos were taken under an overcast sky. If the skywas not overcast, photos were taken when the sun was behind a cloud toeliminate direct radiation. Light was measured with a Sekonic (model L-398) lightmeter. The light meter was calibrated with a Li-Cor photometer (model Li 185).To ensure good resolution, photos were taken at a 'shutter priority' speed of 1/2second. One photo was taken at each of 3, 4, and 5 f-stops above the valueindicated by the light meter. Underexposure maximized the contrast between thetrees and sky.Film was processed by UBC Media Services in one batch to ensureconstant magnification. Using a Logitech Scanman Plus digital scanner, threesets of photos, each containing three photos were scanned. Each set wasjudged by J. M. Chen to contain a photo that was overexposed by one f-stop, onethat was correct, and one that was underexposed by one f-stop. For the purposeof this study, a correct exposure was one that maximized plant/sky contrast whileretaining foliar resolution. Using software developed by Chen et al. (1991), thedigital scans were subjected to algorithms which derived estimates of theeffective leaf area index (Le) 1 , and sky-view factor (SVF) through the integrationof gap fractions (Norman and Campbell 1989; Chen et al. 1991). Differencesbetween the SVF and Le values of the underexposed and correctly exposed andthe overexposed and correctly exposed were averaged to create two correctionfactors; one to correct SVF and one to correct Le values of incorrectly exposedphotos. From each of the remaining sets of photos, the photo nearest the correct1 'Effective leaf area index' differs from 'leaf area index'. The former is defined as one half of thetotal surface area of leaves per unit forest floor area, multiplied by a clumping index (I2) (forrandom leaf spatial distribution, LI = 1; Black et al. 1991). The latter is defined as the projectedleaf surface area per unit of ground surface area (Kimmins 1987).9exposure was scanned and subjected to the same computer proceduresdiscussed above. If the 'best' photo of a set was in my opinion underexposed oroverexposed (when compared to the 'best' photos identified by J.M. Chen), theLe and SVF correction factors (0.58 and 0.045, respectively) were added orsubtracted to obtain corrected Le and SVF values. Ten of 34 photos requiredcorrection.In addition to the photos taken under coniferous stands, seven other siteswere sampled. These sites ranged from beneath willow canopies to edgehabitats. None had a CCC value defined on the forest cover map. Theoretically,a site with no cover (CCC = 0) has SVF and Le values of 1 and 0, respectively. Ifthe SVF and Le values of these 'other' sites were closer to the theoretical valuesof a CCC = 0 site than the observed values of CCC = 3 sites, they were assignedto CCC = 1. Since no CCC = 1 sites were mapped in the study area, the purposeof this interpolation was to assess the validity of forcing the regression linethrough the theoretical point CCC = 0, SVF = 1. Regression equations werefitted to plots of the average SVF and Le values for each CCC against CCC. Anadditional regression equation was obtained by plotting SVF against Le. Whereappropriate, regression lines were forced through the theoretical coordinate pairs(i.e., CCC = 0, SVF = 1; CCC = 0, Le = 0; Le = 0, SVF = 1).MoosehornAt each site where hemispherical photos were taken, moosehorn (Bonnor1967; Bunnell and Vales 1989) estimates of mean crown completeness (MCC)were obtained. Starting directly over the photo site, readings were taken alongthe cardinal bearings. These readings were a minimum of 3 m apart (to ensureindependence of the samples), along a distance such that a minimum of threemature stems were passed within one meter of the observer along each of the1 0four bearings (T.A. Black pers. commun.). A minimum of ten readings weretaken per site. Readings from each site were averaged to determine MCC.MCC values from sites of the same CCC were averaged and correlated withCCC. The data were plotted with the least-square regression line to display thetrend of the correlation.Equivalent blackbody temperature modelTo describe the thermal environment experienced by a moose, asimulation model (Appendix I) based on the 'equivalent blackbody temperature'(Te; Campbell 1977) was developed. Using CCC, hour, day, and the weatherdata from the field station for that particular time, the thermal environments offorested and open habitats were quantified. The model was not used todetermine a moose's energy budget. Because of the assumptions in the modeland the problems inherent in extrapolating wind speed from the open to theanimal's height in the forest, the model's primary function was to rankenvironments by Te. Ranking provided an estimate of the relative differenceamong the Te values of different CCCs. When possible, the effects of altereddriving variables on Te were investigated or deduced.11RESULTS AND DISCUSSIONWeather stationRenecker and Hudson (1986) reported that in summer, the upper criticaltemperature (UCT) (i.e., the ambient temperature above which evaporative heatloss processes of a resting, thermoregulating animal are initiated; Bligh andJohnson 1973) for moose was between 14 and 20 °C. Ackerman (1987)estimated that this same value was between 15 and 17 °C. Renecker andHudson (1986) found that at air temperatures between 14 and 20 °C, thermalpanting was observed and that open-mouthed panting occurred at airtemperatures above 20 °C. 1 Many hours during the study period had anaverage air temperature in excess of 14 °C (Fig. 2.1). Further, the activity andlocation of moose were often monitored when the air temperature exceeded thealleged UCT (Figs. 2.2 and 2.3). Renecker and Hudson (1986) found that'radiant heat load' (a measure of the thermal environment incorporating airtemperature, wind speed, and solar radiation) did not predict physiologicalresponses of moose to heat better than did air temperature. R.J. Hudson (pers.commun.) expressed surprise at this result, and attributed the inadequacy ofradiant heat load as the independent variable to its poor experimentalquantification. Because solar radiation and wind speed are important indetermining an animal's operative temperature, the large radiation flux densitiesand low wind speeds recorded in this study (Figs. 2.1, 2.2 and 2.3) should haveserved to stress moose further.1 An open-mouthed, increased respiratory frequency in response to a thermoregulatory drive todissipate heat via evaporative cooling is defined by Bligh and Johnson (1973) as 'thermalpanting'. The term 'thermal panting' was used by Renecker and Hudson (1986) to describe aclosed-mouth increase in respiratory frequency. It is unclear whether 'thermal panting', asRenecker and Hudson described it, acted to dissipate heat or simply reflected an increasedmetabolic rate.12^i1 ► 1 -10^0^10^20^30AIR TEMPERATURE (°C)^Mk* *I ^ 1,—..„,1 ^ ► ^ I ^ I0 1^2^3^4^5WIND SPEED (m•s -1 )► I^ I^.^l,,^.^l^,^,^,^I-0.5^0.0 0.5 1.0SOLAR RADIATION (kWm -2)Figure 2.1. Boxplot summaries of hourly means of air temperature, wind speed,and solar radiation for the period of 90/06/17 to 90/09/23. Each plotshows the median and the range of the quartiles. Star points are identifiedby SYSTAT (Wilkinson 1990) as outliers. Note: complete weather datawere not obtained for the period of 90/07/03 to 90/07/07. N = 2247.i^13*^1^I ^I^I^1-1 0 0 10^20^30AIR TEMPERATURE (°C)^1 * *0^1^2^3^4^5WIND SPEED (m-s -l)I^I I-0.5 0.0^0.5^1 .0SOLAR RADIATION (kW.m -2)Figure 2.2. Boxplot summaries of hourly means of air temperature, wind speed,and solar radiation for hours when moose activity was determined bytelemetry. See Figure 2.1 for boxplot description. N = 326.14*^1^1^I 1^ I^I-10 0 10 20 30AIR TEMPERATURE (°C)^* *0^1^2^3^4^5WIND SPEED (m•s -1 )d^I^ I^lllllllllllllllllll^1-03 0.0 03^1.0SOLAR RADIATION (kW-rn -2)Figure 2.3. Boxplot summaries of hourly means of air temperature, wind speed,and solar radiation for hours when moose locations were determined bytelemetry. See Figure 2.1 for boxplot description. N = 252.15Forest canopyHemispherical photos and moosehom estimates were taken at 34 sites(Table 2.1). The lack of a one-to-one relation between mean crowncompleteness (MCC) and the crown closure limits represented by crown closureclass (CCC) was expected (Fig. 2.4). Crown closure, as it appears on the forestcover map, represents the proportion of ground surface covered by a verticalprojection of the crown's outermost perimeter (Harcombe 1984; Table 2.2).Because the moosehorn detects not only gaps between crowns, but those withincrowns, it is apparent that the two approaches are measuring different butcovariant features. The significant correlation (p < 0.05, r 2 = 0.87) betweenmean crown completeness (MCC) and crown closure class (CCC) indicated thatthe CCC values on the forest cover map, if not representative of the actual codelimits, correctly ranked the forest cover polygons by crown closure.The relation between sky view factor (SVF) and CCC was nonlinear (Fig.2.5). In theory a site of CCC = 0 should have a SVF of 1.0. Although theMinistry of Forests CCC = 0 corresponds to 0-5% crown closure, for the purposeof this study CCC = 0 was assumed to equal 0% crown closure. Forcing theregression line through the point CCC = 0, SVF = 1 appeared justifiable. Thenatural and forced regressions were significant (p < 0.05). Forcing the regressionthrough CCC = 0, SVF = 1 did not change the coefficient of determination (r 2 =0.85). The equation of the forced regression was used to determine CCC-specific SVFs for the simulation model.16Table 2.1. Crown closure class (CCC), sky view factor (SVF), mean crowncompleteness (MCC) and effective leaf area index (Le) of each site wherehemispherical photos were taken. CCC values are from the forest covermap, SVF and Le values from hemispherical photos, MCC frommoosehorn readings. Parenthesized CCC values represent estimatesbased on the relative location of mean SVF for those four locationscompared to the mean SVF of CCC = 3 and the theoretical point CCC = 0,SVF =1 (see Methods and Materials, p. 9).CCC SVF MCC Le0.1570.3510.3651.301.361.91(1) 0.724 0.85(1) 0.596 1.67(1) 0.698 1.15(1) 0.647 1.103 0.316 1.683 0.250 29.9 1.933 0.260 1.794 0.246 37.9 1.874 0.110 31.7 3.464 0.140 49.7 2.365 0.112 41.9 2.985 0.186 45.2 2.125 0.191 38.5 2.245 0.104 63.1 2.626 0.113 47.1 3.066 0.129 37.2 2.716 0.073 46.0 2.546 0.170 59.7 2.106 0.116 62.0 2.696 0.116 2.147 0.191 53.4 2.307 0.061 56.0 3.387 0.208 53.2 2.837 0.172 40.9 2.158 0.097 53.5 3.008 0.163 45.0 2.718 0.096 53.2 3.588 0.117 35.2 2.828 0.076 61.0 3.398 0.083 61.1 3.199 0.112 57.2 3.0760-- - -4-----170^1^1^I^I^.^I0 2^4^6^8^10CROWN CLOSURE CLASSFigure 2.4. Trend of the averages of mean crown completeness for crownclosure classes 3 to 8. Error bars denote mean ±1 SE. The dashed line isthe line of least squares and is given only to illustrate the trend (r 2 = 0.87;p < 0.05).18Table 2.2. Crown closure class codes and their corresponding percent limits ofcoverage. This scheme of cover classes, used by the B.C. Ministry ofForests, does not account for gaps within individual tree crowns.Crown ClosureClass CodeLimits(Percentage)0 0-51 6-152 16-253 26-354 36-455 46-556 56-657 66-758 76-859 86-9510 96-1000^2^4^6^8 101 .00.80.00.219CROWN CLOSURE CLASSFigure 2.5. Relation between the mean sky view factor (SVF) and crown closureclasses 1 to 8. Error bars denote ±1 std. error. The dashed line is theleast square regression from a 1/SVF data transformation ( Y = 1 / [0.844+ 1.057•X], r 2 = 0.85). The solid line represents the least squareregression forced through SVF = 1 from the same transformation ( Y = 1 /[1 + 1.03•X], r 2= 0.85 ).20In Figure 2.6, the least square regression line does not correspond to thetheoretical relation between Le and CCC adapted from Black et al. (1991):Le = -1/G • In(1-CCC/10)^(1)G represents an angle-dependent extinction coefficient per unit foliage areameasured in the direction of the solar beam. For the spherical (random)distribution of leaf inclination angles, G = 0.5 (Ross 1981). The discrepancybetween the theoretical and observed values of Le versus CCC is likely a productof two things: (1) one of the assumptions of eqn. 1 is that the values of CCCactually represent the range of 0 to 100% crown cover. From Figure 2.4 it isapparent that CCC = 8 (according to the moosehorn estimate) had a MCC of lessthan 55%. Crown cover as employed in eqn. 1 is sensitive to gaps within crowns,therefore MCC as the independent variable should yield a better fit to thetheoretical relation; (2) the value of G may not be 0.5, indicating that foliarelements were not randomly distributed, or that the radiation regime was stronglyinfluenced by vertical elements such as tree boles. The potential effects offactors 1 and 2 were supported when a value of G = 0.25 yielded a theoreticalequation which appeared to agree with the plot of CCC-specific Le and MCCmeans (Fig. 2.7).The solar zenith angle (Z) refers to the angle between the sun and a linenormal to the earth's surface (cf. solar elevation angle). The function used todetermine Le weights that part of the hemispherical photo corresponding tosmaller Z values heaviest (Black et al. 1991). Although I did not explore theexistence of Z-dependent G values, a small G value (i.e., G <0.5) at low Z (i.e., Z<0.6 radians) is characteristic of a plant structure classified as an erectophile.Erectophiles are plant structures which have a radiation regime stronglyinfluenced by components in the vertical plane (Ross 1981). This explanation isconsistent with my cursory observations that the crowns of lodgepole pine212^4^6^8^1 0CROWN CLOSURE CLASSFigure 2.6. Relation between effective leaf area index means and crown closureclasses 1 to 8. Error bars represent ±1 std. error. The solid linerepresents the least square regression. The dashed line represents thetheoretical relation between the two variables (see eqn. 1 in text).2220^40^60^80^100% MEAN CROWN COMPLETENESSFigure 2.7. Mean leaf area index against mean crown completeness for each ofcrown closure classes 3 to 8. The dashed line is the theoretical functiongiven by eqn. 1 (assuming a spherical distribution of foliar elements - seetext). The solid line corresponds to eqn.1 when foliar elements in thevertical plane predominate (i.e, G < 0.5).23tended to be sparser than those of trees such as Douglas-fir (Pseudotsugamenziesii (Mirbel.) Franco.). The result was that for the lodgepole pine, boles(thus vertical elements) tended to be more conspicuous.Owing to the nonconformity of the data to the theoretical function (Fig.2.6), T.A. Black and J.M. Chen (pers. commun.) recommended that a curve befitted through a scatterplot of SVF versus Le. In theory, when Le = 0 then SVF =1. Forcing the curve through this point changed the shape of the relations verylittle (Fig. 2.8). Using the forced regression, Le was estimated from CCC-specificSVF values.Thermal coverBecause Te is derived from many variables (Appendix I), thermal cover isvery dynamic; that is, it is a product of time, weather, and location. As defined bythe simulation model, simultaneous Te values across CCCs decreasedexponentially with increasing CCC (Fig. 2.9). From this trend it is apparent thatbeyond CCC = 4, the marginal increase in thermal cover value is relatively small.Determining the wind speed under a canopy by extrapolating a knownwind speed in the open is difficult. Sub-canopy wind speed is modified bycanopy height and structure as well as understory (Bunnell et al. 1985). Further,sub-canopy wind speed is likely influenced by the slope, aspect, and distance toan edge. When moose were located, 75% of the observations occurred whenthe mean wind speed in the open was less than 2.2 m•s- 1 (Fig. 2.3). For thisreason and the estimation difficulties noted above, no canopy-driven windattenuation function was incorporated in the model. Instead, two simulationsrepresenting extreme conditions were run: (1) a simulation which held windspeed constant across all canopies, and (2) a simulation in which beyond CCC =0, wind speed was held at 0 m-s -1 (Fig. 2.9). Because the equivalent resistance to^1240^1^2^3^4LEAF AREA INDEXFigure 2.8. The relation between sky view factor (SVF) and effective leaf areaindex (Le). The dashed line is the least square regression from alog (SVF) transformation (Y = 10 0.038.[1 0.3391 -X , 12 = 0 . 75 I .j The solid lineis the least square regression forced through the theoretical point Le = 0,SVF = 1 from the same transformation ( Y = [1 0 0.321 -X , 1 2 = 0 . 74 ).502500mo 40443044114244E.  20ra.104 1044000.^1.^2.^3.^4.^5.^6.^7.^8.CROWN CLOSURE CLASSFigure 2.9. Mean operative temperatures by crown closure class (CCC).Weather data (Fig. 2.1) were run through the simulation model (AppendixI) under two conditions: (A) wind speed (as recorded in the open) was heldconstant across all CCCs, (B) wind speed in CCCs >0 was held constantat 0 m•s- 1 . The means for simulations A and B are represented by thelower and upper limits of the solid rectangles respectively. Since CCC = 0was only simulated under condition A, error bars denote ±1 SD. Errorbars in the negative y direction denote 1 SD for the simulations of A. Errorbars in the positive Y direction denote 1 SD for the simulations of B.26heat transfer (re ) remained constant, the trend in decreasing range of mean Tewith increasing CCC resulted primarily from a reduction in the absolute amount ofradiation absorbed (Rabs) (see Figs. 2.5 and 2.8).Edgerton and McConnell (1976) showed that during summer months,mean hourly air temperatures in unlogged coniferous forests and in neighbouringclearcuts differed by less than 6 °C. Other research has shown that mean-maximum air temperatures varied little (i.e., less than 3 °C) with stand density(Jemison 1934; Spurr 1957). In their review, Bunnell et al. (1985) stated that"The simplest approach to estimating the influence of forest cover on airtemperature is to increase minimum temperatures, and decrease summermaximum temperatures by 2 °C". It appears that the effect of canopy on airtemperature is not large. My simulation model assumed that simultaneous airtemperatures beneath forest canopies did not vary with CCC, and that thesetemperatures equalled the hourly means recorded at the weather station. If thisassumption was not valid (e.g., simultaneous air temperature increasedsignificantly with decreasing CCC) the net effect would have been a greater rateof decrease in mean Te with increasing CCC on the hottest days. The trend inFigure 2.9 is a result of increased radiation attenuation by canopies of increasedCCC. Because Te is determined by adding an animal's net radiative gain to theair temperature (Appendix I), other things being equal, increasingly lower airtemperatures yield increasingly lower Te values. Therefore it can be expectedthat if both radiation and air temperature decreased with increased CCC, Tewould decrease more rapidly with increased CCC than if radiative input or airtemperature was held constant.If, as discussed above, foliar elements were not randomly distributed (i.e.,G # 0.5) but rather approximated an erectophile structure, the primary implicationwould be that at low values of Z, lodgepole pine canopies would attenuate a27smaller fraction of solar radiation) The effect of decreasing the thermal covervalue of increased CCCs would dampen the relation in Figure 2.9.If either of the assumptions discussed above (i.e., constant airtemperature and random foliar distribution) were not met, the relation between Teand CCC (Fig. 2.9) would not be expected to deviate from its present form.Therefore, although the magnitude of the differences between Te values ofdifferent CCCs may change modestly, higher CCCs should always provide betterthermal cover.CCC values for forest cover in this study area are given only for coniferousstands. Many of the sites classified as non-productive brush (NPBr) supportedclimax willow stands. A sample of the larger willow stems revealed heights up to5 m and ages over 150 years (annuli counts). In some locations, a dense canopywas formed over the ground between tree clumps. One such site (not atypical ofthe area) yielded a hemispherical photo which had a SVF (0.157) similar to thatof a CCC = 6 conifer stand (Table 2.1). Based upon the capacity to interceptsolar radiation, it is apparent that willow canopies were capable of providing alevel of thermal cover comparable to the denser conifer stands.Quantifying and comparing the thermal environments of forested sites as afunction of CCC appears to be practicable.1 Perhaps it was for this reason that Mitchell and Greene (1981) stated that lodgepole pinecanopies make poor thermal cover.28CHAPTER III: COVER SELECTION AND ACTIVITY OF MOOSEINTRODUCTIONIf summer thermal cover is required by a species, we could expect apopulation deprived of it to exhibit negative symptoms. Such symptoms mightmanifest themselves directly or indirectly to produce declines in productivity andor survivorship, or habitat abandonment via emigration. Peek et al. (1982) notedthat the predicted effects of habitat manipulations are often given in terms ofpopulation change. However, they warned that because the population responseintegrates many factors (e.g., food, predation, weather, and hunter harvest), eachmust be evaluated prior to naming cover as the causative agent of change.Therefore Peek et al. (1982) suggested that predictions of habitat use (notpopulation change) relative to habitat changes be used to anticipate the effects ofcover manipulation.Methods to evaluate resource preference of wild animals have existed formany years (e.g., Hess and Rainwater 1939). Neu et al. (1974) proposed atechnique which permits an animal's observed pattern of resource use to bedescribed as preference, avoidance or used in proportion to availability. Johnson(1980) criticized the classical approach to use-availability studies, and showedthat the arbitrary nature of researcher-defined 'availability' and researcherhandling of 'doubtful' observations can profoundly affect the conclusionspermitted by such studies. Johnson (1980) proposed a technique to evaluateresource preference, and concluded that his technique reduced researchersubjectivity and bias in estimating measures of use and availability. Alldredgeand Ratti (1986) compared four techniques for analyzing resource selection,including those of Neu et al. (1974) and Johnson (1980), and found that no one29technique consistently out-performed the others on simulated data. In theirstudy, Type I error was controlled effectively by all techniques, but theoccurrence of Type II error "depended on the number of habitats, the number ofanimals, the number of observations per animal and the magnitude of thedifferences to be detected".Assuming that preference can be demonstrated, it may be desirable toassess whether or not a preferred habitat is actually required to maintainpopulation health. Knowledge of which habitat types are required by a speciesand which are not can be very important to the process of developing anintegrated land-use strategy. Conversely, it is not unreasonable that givenadequate amounts of required resources but restricted amounts of preferredones, animals would continue seeking the latter. Catering to a species'physiological needs while ignoring learned or innate behavioural patterns mayrender any 'bare-essential' management strategy ineffective.Homeotherms employ many strategies to prevent or reduce heat-stress.Such strategies include the use of cooler microclimates (e.g., thermal cover), anddecreasing metabolic heat production via reductions in activity. Several authorshave concluded that areas providing summer thermal cover are preferred by elk(Cervus elaphus L.) (Young and Robinette 1939; Lyon 1979; Pederson et al.1980) and moose (Schwab 1986; Ackerman 1987; Renecker 1987). Peek et al.(1982) cite several reports which documented high summer densities of elk inareas with little or no thermal cover. Merrill (1991) concluded that Roosevelt elk(Cervus elaphus roosevelti Merriam) inhabiting the blast zone of Mt. St. Helensdid not require summer thermal cover. Although the animals in Merrill's studyused thermal cover when available, elk which did not use it appeared to copewith increased thermoregulatory costs. F.L. Bunnell (pers. commun.) observedelk in the blast zone of Mt. St. Helens wading into a river on hot days,30presumably to cool off. Merrill (1991) did not report any such behaviour.McCorquodale et al. (1986) believed that because heat and disturbance couldrestrict forage intake and increase metabolic costs, abundant forage andinfrequent disturbance were essential in allowing elk to summer in areas oflimited cover.Kelsall and Telfer (1974) reported that "regions where temperaturesexceed 27 °C for lengthy periods, particularly without tall trees to provide shade,or other refugia such as lakes and rivers, do not support moose". To myknowledge, no research has documented a decline in moose population size orproductivity attributable to a loss of thermal cover. Indeed, this would be difficultto do because the increases in human access and animal visibility (i.e., loss ofsecurity cover) associated with recently logged areas can rapidly increase thelegal and/or illegal hunter harvest (Eason 1985; Peek et al. 1987).Elk have a marked ability to dissipate heat via cutaneous water loss orsweating (Parker and Robbins 1983). Perhaps this is why elk appear capable oftolerating high heat loads without ill-effects (e.g., Parker and Robbins 1983;McCorquodale et al. 1986; Merrill 1991). The extent to which moose sweat todissipate heat is unclear. Sokolov and Chernova (1987) reported that moosepossess sweat glands which actively contribute to thermoregulation. Reneckerand Hudson (1986) did not address sweating in moose, but implied that pantingwas the major cooling mechanism. In areas of harsh winters, decreased summerweight gains may increase winter/spring ungulate mortality (Mautz 1978). Largeheat loads impose high thermoregulatory costs on moose which, even in thepresence of abundant forage, can reduce summer weight gains (Renecker andHudson 1986). By deduction, it could be concluded that summer thermal cover isrequired by moose. Assuming this conclusion is correct, summer thermal cover31needs to be quantified so habitat managers can ensure its existence in managedmoose habitats.The alternate hypotheses tested in this study were that if moose respondto heat load, they would: 1) be located in forest stands providing greater shelter(i.e., higher CCC) at times when heat load was highest, and 2) be least active attimes when heat load was highest.METHODS AND MATERIALSMoose telemetryFour adult cow moose collared for the Okanagan Connector FreewayUngulate Impact Assessment (Keystone Bio-Research 1991) were monitored inthis study. All cows had produced calves in years previous to 1990. No calveswere sighted with any of the cows in 1990; however, search intensity was lowerthan in previous summers (Keystone Bio-Research 1991). Because approachinga moose would likely influence its choice of habitat, all telemetry was conductedfrom roads. Because of logging, highway construction, and recreational activitieswhich occurred during the study, I felt that moose would be habituated to thesound of vehicles. The study period did not overlap with any hunting seasons inthe area. Moose were triangulated from the ground using a Lotec receiver, ahand-held Yagi-H antenna, headphones and a Silva Ranger compass. Compassbearings to each signal were taken from a minimum of three sites for eachattempted location (e.g., Springer 1979). Information recorded for each locationincluded:1) animal code,2) date, time at commencement of location,time at completion of location,3) site-specific compass bearings,4) modulating/non-modulating signal.32Moose locations were sampled at minimum intervals of one hour. If morethan one animal was being monitored on the same day, interlocation intervalsoften approached 120 minutes. A typical sampling period spanned 13 hours. Theinterval from 11:00 until 24:00 (PDT) was the most frequent sampling period,however, some hours between 5:00 and 18:00 (PDT) were also sampled. Theformer period was more frequently sampled in an attempt to discriminate theeffects of heat and light on habitat selection and activity. Moose were not fittedwith activity collars. According to Van Ballenberghe and Miquelle (1990), amodulating signal was a reliable indicator of collar, therefore moose, movement.If more than one of the location signals was modulating, that location wasrecorded as 'active'. After moose had moved to other locations within theirsummer home ranges, I investigated the general area of many telemeteredlocations to infer the potential value of each site for foraging.Renecker and Hudson (1986) reported that the upper critical [air]temperature (UCT) for moose in summer was between 14 and 20 °C. Aregression of operative temperature in open areas (Te open) against airtemperature (Tam) data for hours when moose were located in this studyindicated that air temperatures of 14 and 20 °C corresponded to Te open valuesof 13.0 and 29.5 °C respectively (Te = 0.03•Tam 2 .30 ; I 2 = 0.73, SE = 1.6, n =252). I defined hot conditions (Te open :hot) to be when Teopen was >29.5 °C, andcool conditions (Te open :cool) to be when Teopen was <13.0 °C. 1 Day or 'light'conditions existed when the mean hourly solar flux density (MHSFD) was ?_50W•m- 2 . Night or 'dark' conditions existed when MHSFD was <50 W•m- 2 .1 Renecker and Hudson (1986) conducted their observations under uncontrolled conditions withrespect to solar radiation and wind. Had their observations been made in a controlledenvironment, Te values of 14 and 20 °C would have been used as the thermal limits in thisstudy.33Analysis of telemetry dataCompass bearings were plotted on a 1:15 000 forest cover map. Theuniversal transmercator (UTM) coordinates to the nearest 25 m were recorded forthe center of the polygon formed by the intersection of at least three bearings.Habitat features near the estimated location identified by the forest cover mapwere checked against air photos. If the forest cover map lacked informationavailable from the air photo it was adjusted accordingly.Telemetry locations were rejected if they met one of two criteria:1) three bearings did not intersect,2) the polygon bounded by the bearings contained more thanone cover-type and had at least one side longer than 300 m.Other location attributes recorded directly from the forest cover map included:1) distance to road - the distance from the center of the bearingpolygon (CBP) to the nearest road (roads under thisclassification were identified as those which were relativelyfrequently travelled by vehicles),2) distance to edge - if a CBP was further than 75 m (half thediameter of telemetry overlays; see below) from a non-forestcover-type, the distance from the CBP to the nearest non-forest cover-type was recorded,3)^distance to water - the shortest distance from the CBP to anywater body that supported riparian or emergent aquaticvegetation was recorded.Using UTM coordinates, the distance between successive mooselocations was divided by the time between those locations to yield a meanstraight line velocity (MSLV). Ignoring telemetry error, interlocation distancerepresented a minimum value, and as such, likely underestimated actual meanvelocity.34The average precision of the telemetry system was determined by using a95% error arc of ±4°. A random sample of 20 moose locations was chosen.The average length of the longest side of each of the error polygons (Springer1979) was calculated. This averaged length (150 m) was used as the diameterof a circular overlay centered on each pair of UTM coordinates. Analysis of adigitized forest cover map was conducted using TERRASOFT. TERRASOFTwas used to quantify the study area by forest crown cover, and obtain informationregarding patterns of habitat use by moose from telemetry overlays. Becausetelemetry overlays often contained more than one cover-type, two types ofdependent variables were tested: 1) the area (ha) contained in telemetry overlaysby crown closure class (CCC), and 2) the frequency with which each CCCoccurred within telemetry overlays (regardless of area). Overall use-availabilityof forest cover-types was not assessed because I believed that for this type ofanalysis, the telemetered observations were autocorrelated.The small surface area to body volume ratio of large animals such asmoose results in a reduced rate of heat transfer between animal andenvironment. The potential effect on cover selection of a delayed response to ahot environment due to thermal inertia was investigated by examining theselection pattern of CCCs across a 13 hour period.The effect of the thermal environment on subsequent (<120 min. later)habitat selection was tested by examining the patterns of movement betweencover-types under Teopen :hot and Teopen :cool conditions. Habitats were labelledas open (CCC = 0) and cover (CCC >0). When telemetry overlays containedmore than one cover-type, that which represented the greater area was used.Except where noted, statistical testing was done with SYSTAT (Wilkinson1990). The probability of a Type I error was set at 0.05. If statistical tests yielded35results significant at a = 0.10 (but not a = 0.05), they are reported as such. Thefrequency distributions of crown closure class (CCC) for the study area versushabitat within 800 m of a road and Teopen :hot versus Teopen :cool, were testedwith the Kolmogorov-Smirnov (K-S) test of two independent samples. Tests ofindependent proportions were conducted according to Hicks (1982) (hereaftercalled Hicks' test). Pearson product-moment correlation coefficients (r) wereused. All variables tested for correlations were plotted to check for non-linearrelations. Regression values were compared with the paired t-test. Thelikelihood-ratio x 2 was used to test for differences in moose activity betweenTeopen :hot and Teopen :cool conditions. The K-S test was used to test for adifference between the frequency of velocity intervals for active versus inactivetelemetry readings. Tukey's HSD test was used to identify different means whensignificant ANOVA results were observed. Velocity intervals were selected by analgorithm in SYSTAT (Wilkinson 1990).RESULTS AND DISCUSSIONHabitat analysisThe results of a GIS evaluation of the proportions of the study arearepresented by each crown closure class (CCC) did not differ significantly fromthe results of a similar evaluation of habitats falling within telemetry range ofroads (Fig 3.1; K-S test, p > 0.05). The similarity between the two distributions inFigure 3.1 indicates that potential moose activity and location sampling was notbiased to a habitat mix that was atypical of the study area.Habitat useTwo hundred and fifty two telemetered moose locations were recordedacross a range of ambient light and temperature conditions (Chapter II).36504  40PLIg44AZ 3044A4.10[-• 20ZiiiU4.10„,  100AREA12 HABITATCROWN CLOSURE CLASSFigure 3.1. Percentages of the entire study area (AREA, 15 970 ha) and habitatwithin 800 m of a road (HABITAT, 9855 ha) represented by each coverclass. For crown closure limits see Chapter II, Table 2.2.37The nature of the relation between operative temperature (Te) and airtemperature (Tam) for CCC = 0 sites (i.e., Te = 0.03•Tam 2 . 30) dictated that for theobserved weather data, the upper critical temperature (UCT) for moose in thesummer was never exceeded during 'dark' hours. Conversely, only 13% of 'light'hours during the summer were below the UCT (n = 292). Because theopportunity to sample moose locations during 'cool', 'light' conditions was limited,only 3% (n = 4) of locations and 4% (n = 7) of activity samples were taken undersuch conditions. These small sample sizes precluded direct partitioning of theeffects of heat and light on the cover selection and activity of moose.Both the total area-by-CCC within telemetry overlays and the frequency ofCCCs within overlays revealed a significant difference in cover-type selectionbetween Teopen :hot and Teopen :cool conditions (Figs. 3.2 and 3.3; K-S test, p <0.05). Total use of CCC = 6 stands was greater during 'hot' (thus 'light')conditions than during 'cool' (thus 'dark') conditions, (Figs. 3.2 and 3.3; Hicks'test, p < 0.05). No difference between total use of CCC = 0 sites as a function ofthe thermal (thus light) environment was detected (Figs. 3.2 and 3.3; Hicks' test,p > 0.05). The remaining significant difference was that stands of CCC = 7 wereused more during 'cool' conditions (Fig. 3.3; Hicks' test, p < 0.05).All summer forage plants listed by Eastman and Ritcey (1987) andSingleton (1976) for moose in the vicinity of the study area are associated withriparian habitats (thus CCC = 0 in this study). With the exception of edges and afew seepage sites, no forage species were found when telemetered mooselocations in conifer stands were investigated on-foot. Increased use of CCC = 6stands during 'hot' conditions may have reflected use of those areas for thethermal cover they provided. Assuming the warmest scenario of complete windattenuation(see p. 23), the difference between the two regressions of Te versusTam for CCC = 0 (Te = 0.03•Tam 2 . 30) and CCC = 638r34 50114IX444  40•-4gIA0>4 30aI-114XPo 20•-144I-.440 10E-1ZmiUW 0aHOT12 COOLCROWN CLOSURE CLASSFigure 3.2. Percentage of the total area (ha) contained within telemetry overlaysrepresented by each crown closure class for locations when Te open washot (light) (n = 124) and cool (dark) (n = 56).3950403020100HOT12 COOLCROWN CLOSURE CLASSFigure 3.3. Percentage of telemetry overlays containing each crown closure classwhen hourly Teopen values were hot (light) (n = 124) and cool (dark) (n =56).40(Te = -4.60 + 1.36-Tam, r 2 = 0.91, SE = 2.2 °C) increased exponentiallybetween Tam = 14 and 30 °C (Fig. 3.4). Therefore, as Teopen increased abovethe UCT of moose, the relative value of CCC z6 sites as thermal cover alsoincreased. As shown in Chapter II, the marginal increase in thermal cover valuefor stands greater than CCC = 4 was slight. The relative use of CCC .4 (Figs.3.2 and 3.3) is likely a result of the availability of each cover-type (Fig. 3.1). Theregression of Te versus Tam for CCC = 4, (assuming complete wind attenuation;Te = -5.74 + 1.47•Tam, r 2 = 0.88, SE = 2.8 °C), was significantly different thanthe same regression for CCC = 6 (paired t-test, p < 0.05, n = 30). However, forthe highest recorded air temperature the largest difference between the tworegressions (2.0 °C at 27.5 °C) was within the standard error of eachregression.If CCC = 0 sites provided no thermal cover, the similar use of CCC = 0and CCC = 6 on Teopen :hot days (Figs 3.2 and 3.3) implies that thermal coverwas not being selected. As shown in Chapter II, a few sites lacking a CCCdesignation on the forest cover map (therefore taken as CCC = 0) intercepted anamount of solar radiation comparable to that intercepted by conifer stands ofCCC = 6. Cursory observations indicated that habitat polygons designated asCCC = 0 were usually associated with water. The apparent use of water bymoose as a heat-sink has been reported (Ackerman 1987; Renecker 1987).Because water's potential to act as a heat-sink and the shade properties of willowtrees were not factored into the operative temperature model (Chapter II;Appendix I), it can not be concluded that moose using areas of CCC = 0 werenecessarily heat stressed when Teopen exceed 29.5 °C. Indeed, the use ofsome CCC = 0 sites may have served to mitigate the effects of heat stress.403041201 0014. 15. It 17. 8. 13 20. 21 22. 23 24. 25 24. 27 28 23 30.AIR TEMPERATURE (deg. C)Figure 3.4. Simultaneous differences between the regressions of Te versus airtemperature for CCC = 6 and CCC = 0 sites, as a function of airtemperature from 14 to 30 °C.42For large animals such as moose, a small ratio of surface area to bodyvolume results in a reduced rate of heat gain (Peters 1983). A slow rate of heatgain could therefore mean that thermal cover is required only after a sustainedexposure to hot environments. Alternately, thermal cover might be used beforethe onset of heat stress as part of an optimization strategy. Figure 3.5 showshow Teopen values were distributed for each hour that moose locations weresampled from 11:00 until 24:00. To eliminate the effects of light, the hourly ratiosof telemetry overlay areas in CCC = 0 to areas in CCC were compared from11:00 until 21:00. A significant increase in the relative use of CCC = 0 sites (p <0.05; 11:00 until 24:00, r 2 = 0.63; 11:00 until 21:00, r 2 = 0.46) corresponded to asignificant decrease in Teopen values for the same period (p < 0.05, r 2 = 0.75)(Figs. 3.5 and 3.6). The hour when the largest Teopen values were observed(13:00) was the same hour when the ratio of CCC = 0 to CCC overlay areaswas smallest (Figs. 3.5 and 3.6). Since not all sites of CCC = 0 had willowcanopies capable of providing thermal cover, use of some forage-rich CCC = 0sites appeared to be constrained by the thermal environment of such areas.Because an increase in relative use of CCC = 0 sites corresponded to adecrease in Teopen , even though 'light' conditions prevailed until 21:00 for most ofthe summer, thermal constraints explain habitat selection better than a photo-correlated, anti-predator response. Demarchi (1990) also found that use of CCC= 0 sites by moose was least when the potential for thermal stress was greatest.When individual proportions of cover-pattern-selections were comparedbetween temperature (light) classes, moose on covered sites were most likely toremain under cover when Teopen was 'hot' (Fig. 3.7, Hicks' test, p < 0.05). Inaddition, the least common pattern of cover selection under Te open :hot conditionswas 'cover to open'; providing further evidence that habitat selection wasthermally constrained.I ii43T ITiT*1111111111111 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.HOURFigure 3.5. Boxplot summary of hourly Te o en values when moose were locatedfor the hours of 11:00 until 24:00 (n = 252). Each boxplot shows themedian and the range of the quartiles. Star and circle points are identifiedby SYSTAT (Wilkinson 1990) as outliers. The lines correspond to theUCT (dashed) and the thermal limit for open-mouthed panting (solid)identified by Renecker and Hudson (1986) (see text p. 32).8070U 40fgFAa 30201 0011 12. 13 14. 15 16. 17. 111 19 20 21 22. 2344HOUREa CCC =0• CCC >=4Figure 3.6. Percentage of telemetry overlay areas in CCC = 0 and CCC 4 forthe period from 11:00 until 24:00 (n = 243 locations). The relationbetween the ratio of CCC = 0 to CCC 4 overlay areas and hour issignificant (r 2 = 0.63, p < 0.05).4560O 50O 404.1vi0O 3020• 20O 100HOTr2 COOLN10- 3- morlcoN10'cocovs^14 cov, v•I' movv, °WIotiFigure 3.7. Percentage of successive radio locations (i.e., < 120 minutes apart)represented by cover type at time = t and cover type at time = t + 1 whenTeopen was hot (n = 83) and Te open was cool (n = 37).46No appreciable correlations (all r >-0.20 and <0.20) were found betweenmoose location distance to a road/habitat edge/water body and Teopen or time ofday. These observations are consistent with Demarchi's (1990) results whichshowed that these location variables were independent of three time intervalsbetween 15:00 and 03:00. Interpreting these results in the context of otherresearcher's findings would be difficult. The geographical variation in number,type and human-usage of roads, the degree of habitat interspersion, andphysiographic nature of habitats is certainly great across moose range (e.g.,Kelsall and Telfer 1974; Eastman and Ritcey 1987). Putman (1988) concludedthat the diel patterns of cervid activity are very plastic. If disturbance caused byhumans denied moose access to preferred or required resources near roads,moose might shift their use of such areas to times when the disturbance wasminimal (e.g., night). Because vehicular traffic in the study area did not appearexcessive to me, and roads were not associated with particular geographic andtherefore habitat features as they are in some areas (e.g., forage-rich valleybottoms), the lack of change in 'distance to road' across hours was expected.The high degree to which cover-types and riparian areas were interspersed in thestudy area (GIS analysis) may explain the findings that moose distances to anedge or water were not correlated with Teopen or hour of day.Moose activityMoose activity was sampled 326 times over a range of ambient light andtemperature conditions (Chapter II). Relative moose activity was greater underTeopen:cool (thus 'dark') conditions (x 2, p < 0.05, Fig. 3.8). Cervid activitypatterns can be temporally modified to avoid predators and other disturbances(Putman 1988). If predator activity is correlated with light conditions and mooseattempt to avoid predators (e.g., to protect calves), temporal differences in moose10008060402047COOL^HOTTEMPERATURE CLASSFigure 3.8. Percentage of active and inactive radio locations under Te open :hot(light) (n = 155) and Teopen :cool (dark) (n = 77) conditions.INACTIVEEl ACTIVE48behaviour may reflect a predator-avoidance response. The estimated mortalityrate for adult cow moose in the vicinity of the study area is extremely low (3.2%),while for calves it is quite high (49%) (Keystone Bio-Research 1991). Althoughpredation on moose calves in the study area has not been researched, potentialpredators occurring there include: black bear (Ursus americanus Pallas) andcougar (Felis concolor L.). Amstrup and Beecham (1976) reported that insummer, black bear activity peaks were crepuscular and diurnal. Van Dyke et al.(1986) reported that cougars were most active at night. Assuming the risks ofpredation by black bear and cougar were equal, the temporal segregation ofthese predators might mean that moose habitat selection or activity was notconstrained by predators.Using radio telemetry, Risenhoover (1986) concluded that variations inactivity levels of moose were attributable to linear travel. Distance travelled oractivity levels of moose in summer have been reported to be greatest at night(Phillips et al. 1973; Joyal and Scherrer 1978), at night and early in the morning(Van Ballenberghe and Miquelle 1990), and at dawn and just after and at duskand just after (Belovsky and Jordan 1978). Van Ballenberghe and Miquelle(1990) noted that shorter activity bouts 'appeared' to be associated with warmertemperatures at midday, however, their data were not tested. de Vos (1958)noted that although moose were not as readily observed at midday compared tomorning and evening, they 'seemed' to be observed more on hot compared tocool afternoons. Despite this trend, de Vos was unable to demonstrate acorrelation between air temperature and moose observability. Joyal andScherrer (1978) reported that moose 'seemed' more active on clear [thus, likelywarmer] summer days but that no significant differences between the movementson clear, cloudy or rainy days were detected. Sample sizes were not reported inany of the above papers which subjectively commented on changes in moose49activity on warmer days. Ackerman (1987) found that as black globe temperatureincreased, forage-bout duration decreased significantly and bedded durationincreased significantly. In both instances, the reported r 2 value was less than0.40. I believe that because the data presented by Ackerman in those analysesrepresented weekly averages of both independent and dependant variables, hisresults are inconclusive.Published evidence conclusively demonstrating a strong effect of heat onthe activity levels of moose is lacking. In this study, mean interlocation straight-line velocity (MSLV) was not correlated with Te open (r = -0.06, n = 205) or hour (r= 0.03, n = 205); from Figure 3.9, however, a trend of decreasing velocity existedfrom 12:00 until 17:00. The curved line in Figure 3.9 demonstrates how Te openchanged across those hours when velocity was estimated. The trend indecreasing velocity as the duration of exposure to Te open >29.5 °C increasedsupports the hypothesis that moose decrease travel distance (thereby potentiallydecreasing activity) with increasing ambient heat load. An ANOVA conducted onthe mean velocities from 12:00 until 18:00 indicated a significant differencebetween means at a = 0.10. Testing the means revealed that the significantresult was due to the difference between values at 12:00 and 16:00.The percentage of active locations (Fig. 3.10) appeared to mimic thepattern of mean hourly velocities (Fig. 3.9), but when the frequency distributionsof active and inactive locations versus velocity interval (Fig. 3.11) werecompared, no significant difference was detected (K-S test, p > 0.05). BecauseMSLV does not account for activities such as foraging in a small (e.g., <2 ha)patch, the similarity between the two activity distributions in Figure 3.11 was notunexpected.508 45406 35 C5tabea30 Fx-10425 a&.20 W15 gat4I=FAOL,10 C),..052 ►^►^►^I^►^►^►^►^►^1^►^► 011 12. 13. 14 IS 16 17. 18. 19 20. 21. 22. 23HOURFigure 3.9. Average 'mean straight line velocity' between successive mooselocations from 11:00 until 24:00. Error bars denote ±1 standard deviation(n = 209). The curved line represents the least square regression ofoperative temperature in open areas against hour for times when moosewere located (n = 242; r 2 = 0.75; p < 0.05).100011^12. 13. 14. 15. 18. 17. 18. 19. 20. 21 22. 23.51HOURFigure 3.10. Percentage of 'active' moose locations as determined by radiotelemetry, for the hours of 11:00 until 24:00. (n = 301).40301001.^2.^3.^4.^5.^6.^7.^11.^9.^10. 11. 12.52VELOCITY INTERVALINACTIVE12 ACTIVEFigure 3.11. Frequency distribution of active and inactive radio locations for 12straight-line velocity intervals. Each velocity interval corresponds to arange of 1 m•min -1 (e.g., velocity interval 1 = 0.00 to 0.99 rn•rnin -1 , 2 = 1.00to 1.99 rn-min - i , etc.).53The effect of the thermal environment on moose activity was analyzed inthe context of Teopen . Therefore, use of stands providing thermal cover mayhave allowed for increased activity when activity was thermally constrained onCCC = 0 sites. However, without knowledge of habitat-specific activities (e.g.,foraging, bedding, travel) any explanation of observed differences between theactivity levels of hot/cool conditions would be highly speculative; potentiallyarguing for or against thermal constraints on activity. A negative correlationbetween percent of locations active and mean hourly Teopen (r = -0.47, n = 13)indicated a trend of decreasing activity with increasing Te open . This observationalso favours the view that moose were responding to heat and not predators,because for most of the summer, 'light' conditions existed until 21:00.The susceptibility of moose to thermal stress in the summer (Reneckerand Hudson 1986), the influence of forested sites on the thermal environment(Chapter II), and the patterns of cover selection found in this study indicate thatsummer thermal cover for moose exists as a manageable habitat component.54CHAPTER IV: SUMMARY AND CONCLUSIONSFour adult cow moose with radio collars were monitored on the southernThompson Plateau during the summer of 1990. The opportunity to samplemoose activity and habitat selection at times when moose could have beenthermally stressed was realized. The relation between mean crowncompleteness (MCC) and crown closure class (CCC) indicated that CCC valuesfrom the forest cover map correctly ranked stands by crown cover. Theexponential decline in Te across CCCs showed that a definite gradient of thermalcover existed. The strong effect of forest canopy on solar radiation attenuationindicated that greater than CCC = 4 (corresponding to MCC >40%), little thermalcover value was gained. The effect of wind on the operative temperature (Te) ofhabitats appeared to be minimal under the observed weather conditions. Neitherthe effects of uneven air temperatures across CCCs nor a nonrandom distributionof foliar elements are believed capable of changing the thermal cover regimeacross CCCs such that hypothesis testing for the existence of thermally-correlated moose habitat selection would be hampered.With respect to forest crown closure, the relative abundance of habitatswithin telemetry range of roads was the same as that of the entire study area.This finding implied that the opportunity to sample moose locations in the variouscrown closure classes (CCC) was not determined by vehicle access. 'Hot'conditions were defined as those which exceeded the upper critical temperature(UCT) of moose in summer. 'Cool' conditions were below UCT. Because theopportunity to sample under 'cool', 'light' conditions was restricted, it was notpossible to directly partition the effects of heat and light on moose habitatselection and activity. These effects were indirectly assessed by examiningchanges in cover selection and activity patterns across a 13 hour period.55Cow moose selected different habitats between 'hot' and 'cool' (thus 'light'and 'dark') conditions. CCC = 6 was the most frequently selected coniferouscover. Use of CCC = 6 stands was greatest when Te open exceeded UCT. Whencompared to CCC = 0 sites (assuming no water or deciduous canopy cover), thethermal cover value of CCC = 6 sites increased exponentially with airtemperature. Overall use of CCC = 0 sites was not restricted when ambientconditions exceeded UCT. Because of the water and shade from willow treesassociated with some CCC = 0 sites, the ability of such sites to mitigate heat-stress could not be discounted. Prolonged periods of hot conditions in the opengeopent0 appeared to affect cover selection. As Te open declined, relative useof CCC = 0 sites during 'light' hours increased. By seeking thermal cover whenthe rate of heat build-up on CCC = 0 sites was greatest and not after theybecame thermally stressed, moose may have been able to increase total dailyuse of open (foraging) habitats. Belovsky (1981) reported that observed use of[summer thermal] cover by moose agreed with the predictions of an 'optimization'model. One of the constraints in Belovsky's model was an upper limit to bodytemperature. Putman (1988) identified predators and weather as factors whichcan influence habitat use by cervids at an intra-seasonal level. Because thepattern of habitat use from 11:00 until 21:00 changed despite continued 'light'conditions, and assuming the diel risk of predation was constant, habitatselection by moose appeared to be thermally constrained. The cover ofsuccessive moose locations was influenced by the Te and cover of previouslocations when patterns of cover selection between 'hot' and 'cool' conditionswere analyzed separately. When Teopen was 'hot', moose under coniferouscover tended to remain under such cover. The least common pattern underTeopen :hot conditions was 'cover to open'.56Likely because of habitat interspersion in the study area, the mooselocation attributes of distance to an edge and distance to water were notcorrelated with time of day or Te in the open (Teopen). The lack of a correlationbetween location distance to a road and time of day may have been a result ofmoose habituation to vehicles.Moose activity was greater under 'cool' (thus 'dark') than 'hot' (thus 'light')conditions. The mean straight line velocity (MSLV) between successive mooselocations did not differ from 11:00 until 24:00. MSLV was not correlated withTeopen . A decrease in MSLV from 12:00 until 17:00 corresponded to a sustainedperiod of exposure to Teopen >UCT. This observation implies that moose mayhave reduced movement with increased exposure to hot environments. Activitydid not increase with increased MSLV, but did increase as mean hourly Teopenvalues decreased from 11:00 until 24:00. In addition to selecting thermal cover,reduced activity at high Teopen values could eliminate the need tothermoregulate, or minimize the metabolic costs of thermoregulation.Research on moose physiology suggests that moose require summerthermal cover regardless of forage abundance (i.e., Renecker 1987). The resultsof this study indicate that moose select summer thermal cover and that this covercan be quantified. The primary management implication is that for moose in thisarea, managing summer thermal cover at the level of forest cover polygonsappears warranted. It is my opinion that if security cover needs are met andriparian sites are adequately fringed with coniferous stands of CCC thethermal cover requirement of moose on summer range will be met.57LITERATURE CITEDAckerman, T.N. 1987. Moose response to summer heat on Isle Royale. M.S.Thesis. Michigan Technological Univ., Houghton, MI. 53 pp.Alldredge, J.R. and J.T. Ratti. 1986. Comparison of some statistical techniquesfor analysis of resource selection. J. Wildl. Manage. 50:157-165.Amstrup, S.C. and J. Beecham. 1976. Activity patterns of radio-collared blackbears in Idaho. J. Wildl. Manage. 40:340-348.Bakken, G.S. 1980. The use of standard operative temperature in the study ofthe thermal energetics of birds. Physiol. Zool. 53:108-119.Belovsky, G.E. 1981. Optimal activity times and habitat choice of moose.Oecologia (48):22-30.Belovsky, G.E. and P.A. Jordan. 1978. The time-energy budget of a moose.Theor. Pop. Biol. 14:76-104.Black, H., R. Scherzinger, and J.W. Thomas. 1976. Relationships of RockyMountain elk and Rocky Mountain mule deer habitat to timbermanagement in the Blue Mountains of Oregon and Washington. p. 11-31In: S.R. Heib (ed.), Proceedings of the elk-logging-roads symposium.Idaho. 142 pp.Black, T.A., J.M. Chen, X. Lee, and R.M. Sagar. 1991. Characteristics ofshortwave and longwave irradiances under a Douglas-fir forest stand.Can. J. For. Res. 21:1020-1028.Bligh, J. and K.G. Johnson. 1973. Glossary of terms for thermal physiology. J.Appl. Physiol. 35:941-961.Bonnor, G.M. 1967. Estimation of ground canopy density from groundmeasurements. J. For. 65:544-547.British Columbia Ministry of Environment. 1988. Ecoregions of British Columbia,map.British Columbia Ministry of Forests. 1988. Biogeoclimatic Zones of BritishColumbia, map.Bunnell, F.L. and D.J. Vales. 1989. Comparison of methods for estimating forestoverstory cover: differences among techniques. Can. J. For. Res. 20:101-107.Bunnell, F.L., K.L. Parker, L.L. Kremsater and F.W. Hovey. 1985.Thermoregulation and thermal cover of deer and elk on Vancouver Island:problem analysis. Research, B.C. Ministry of Environment and Parks andB.C. Ministry of Forests and Lands. IWIFR-28. Victoria, B.C. 137 pp.Campbell, G.S. 1977. An Introduction to Environmental Biophysics. Springer,New York. 159 pp.58Chen, J.M., T.A. Black and R.S. Adams. 1991. Evaluation of hemisphericalphotography for determining plant area index and geometry of a foreststand. Agric. For. Meteorol. 56:129-143.de Vos, A. 1958. Summer observations on moose behaviour in Ontario. J.Mamm. 39:128-139.Demarchi, M.W. 1990. Daily summer movements of moose in the southerninterior of British Columbia. Unpublished. B.C. Ministry of Transportationand Highways report. 36 pp.Eason, G. 1985. Overharvest and recovery of moose in a recently logged area.Alces 21:55-75.Eastman, D.S. and R. Ritcey. 1987. Moose habitat relationships andmanagement in British Columbia. Swedish Wildl. Res. Suppl. 1:101-117.Edgerton, P.J. and B.R. McConnell. 1976. Diurnal temperature regimes of loggedand unlogged mixed conifer stands on elk summer range. USDA For. Res.Note. PNW-277. 6 pp.Erbs, D.G., S.A. Klein, and J.A. Duffie. 1982. Estimation of the diffuse radiationfraction for hourly, daily and monthly-average global radiation. SolarEnergy, 28:293-302.Esmay, M.L. 1969. Principles of Animal Environment. AVI Publishing Inc., CT.325 pp.Gates, D.M. 1980. Biophysical Ecology. Springer-Verlag, New York. 611 pp.Geiger, R. 1965. The Climate Near the Ground. Harvard Univ. Press,Massachusetts. 611 pp.Harcombe, A.P. 1984. Wildlife habitat handbooks for British Columbia: problemanalysis. British Columbia Ministry of Environment, Kamloops. 236 pp.Hess, A.D. and J.H. Rainwater. 1939. A method for measuring the foodpreference of trout. Copeia 3:154-157.Hicks C.R. 1982. Fundamental Concepts in the Design of Experiments. Holt,Rinehart and Winston. 425 pp.Jemison, G.M. 1934. The significance of the effect of stand density upon theweather beneath the canopy. J. For. 32:446-451.Johnson D.H. 1980. The comparison of usage and availability measurements forevaluating resource preference. Ecology 61:65-71.Joyal R. and B. Scherrer. 1978. Summer movements and feeding by moose inwestern Quebec. Can. Field-Nat. 92:252-258.Kelliher, F.M. 1985. Salal understory removal effects on the soil water regimeand tree transpiration rates in a Douglas-fir forest. Ph.D. thesis. Universityof British Columbia. 123 pp.59Kelsall J. and E. Telfer. 1974 Biogeography of moose with particular reference towestern North America. Nat. Can. 101:117-130.Keystone Bio-Research. 1991. Environmental assessment of the OkanaganConnector Freeway. Unpublished. British Columbia Ministry ofTransportation and Highways and British Columbia Ministry ofEnvironment report. 220 pp.Kimmins, J.P. 1987. Forest Ecology. Macmillian Publishing Co., New York, NY.529 pp.Lloyd, D., K. Angrove, G. Hope and C. Thompson. 1990. A guide to siteidentification and interpretation for the Kamloops Forest Region. B.C.Ministry of Forests, Land Management Handbook No. 23. 398 pp.Lyon, L.J. 1979. Influences of logging and weather on elk distribution in westernMontana. USDA For. Serv. Res. Paper INT-236. 11 pp.Mautz, W.W. 1978. Nutrition and carrying capacity. p. 321-348. In: J.L. Schmidtand D.L Gilbert (eds.) Big Game of North America: Ecology andManagement. Stackpole Books, PN. 494 pp.McCorquodale, S.M., K.J. Radeke, and R.D. Taber. 1986. Elk habitat usepatterns in the shrub-steppe of Washington. J. Wildl. Manage. 50:664-669.Merrill, E.H. 1991. Thermal constraints on use of cover types and activity time ofelk. Behay. Sci. 29:251-267.Mitchell, R. and R. Greene. 1981 Identification and interpretation of ecosystemsof the Western Kamloops Forest Region. First Approximation. B.C.Ministry of Forests, Land Management Handbook No. 2.Monteith, J.L. 1973. Principles of Environmental Physics. American Elsevier,New York, NY. 209 pp.Neu, C.W., C.R. Byers, J.M. Peek. 1974. A technique for analysis of utilization-availability data. J. Wildl. Manage. 38:541-545.Nilson, T. 1971. A theoretical analysis of the frequency of gaps in plant stands.Agric. Meteorol. 8:25-38.Norman, J.M. and G.S. Campbell. 1989. Canopy structure. p. 300-325 In: R.W.Pearcy, J. Ehlinger, H.A. Mooney and P.W. Rundel (eds.), PlantPhysiological Ecology: Field Methods and Instrumentation.Nyberg, J.B. and D.W. Janz (tech. eds.). 1990. Deer and elk habitats in coastalforests of Southern B.C. B.C. Ministry of Forests. 310 pp.Paldridge, G.W. and C.M.R. Platt. 1976. Radiative Processes in Meteorology andClimatology. Elsevier Scientific Publishing Co., New York. 318 pp.Parker, K.L. 1987. Body-surface measurements of mule deer and elk. J. Wildl.Manage. 51:630-633.60Parker K.L. and C.T. Robbins. 1983. Thermoregulation in mule deer and elk.Can. J. Zool. 62:1409-1422.Parker, K.L. and M.P. Gillingham. 1990. Estimates of critical environments formule deer. J. Range Manage. 43:73-81.Pederson, R.J., A.W. Adams and J.M. Skovlin. 1980. Elk habitat in an unloggedand logged forest environment. Oregon Dept. of Fish and Wildl., Wildl.Res. Rep. No. 9. 121 pp.Peek, J.M, D.J. Pierce, D.C. Graham and D.L. Davis. 1987. Moose habitat useand implications for forest management in northcentral Idaho. SwedishWildl. Res. Suppl. 1:195-199.Peek, J.M., M.D. Scott, L.J. Nelson, D.J. Pierce and L.L. Irwin. 1982. Role ofcover in habitat management for big game in northwestern United States.Trans. N. Amer. Wildl. Conf. 47:363-373.Peters, R.H. 1983. The Ecological Implications of Body Size. Cambridge Univ.Press, New York. 329 pp.Phillips, R.L., W.E. Berg and D.B. Siniff. 1973. Moose movement patterns andrange use in Northwestern Minnesota. J. Wildl. Manage. 37:266-278.Putman, R. 1988. The Natural History of Deer. Comstock Publishing Associates,New York. 191 pp.Reifsnyder, W.E. 1967. Radiation geometry in the measurement andinterpretation of the radiation balance. Agric. Meteorol. 4:255-265.Renecker, L.A. 1987. Bioenergetics and behaviour of moose (Alces alces) in theaspen-dominated boreal forest. Ph.D. Thesis, University of Alberta. 265pp.Renecker L.A. and R.J. Hudson 1986. Seasonal energy expenditures andthermoregulatory responses of moose. Can. J. Zool. 64:322-327.Risenhoover, K.L. 1986. Winter activity patterns of moose in interior Alaska. J.Wildl. Manage. 50:727-734.Ritcey, R., D. Lowe, A. Harestad, R.W. Campbell and A.P. Harcombe. 1988.Species habitat relationship models for mammals. Vol 5 In: A.P.Harcombe (tech. ed.), Wildlife habitat handbooks for the Southern InteriorEcoprovince. British Columbia Ministry of Environment and B.C. Ministryof Forests, Victoria. 252 pp.Robinson, M.W. 1947. An instrument to measure forest crown cover. For. Chron.23:222-225.Ross, J. 1981. The Radiation Regime and Architecture of Plant Stands. Dr. W.Junk Publishers, London. 391 pp.Sainsbury, D. 1967. Animal Health and Housing. Williams and Wilkins Co.Baltimore. 329 pp.61Schwab, F.E. 1986. Moose habitat selection in relation to forest cutting practicesin northcentral B.C. Ph.D. Thesis. Univ. of British Columbia, Vancouver.176 pp.Singleton, J.M. 1976. Food habits of wild ungulates in British Columbia:Bibliography and plant synopsis. B.C. Fish and Wildife Branch. 51 pp.Smith, F.W. and J.N. Long. 1987. Elk hiding and thermal cover guidelines in thecontext of lodgepole pine stand density. West. J. Appl. For. 2:6-10.Sokolov, V.E. and O.F. Chernova. 1987. Morphology of the skin of moose (Alcesalces). Swed. Wildl. Res. Suppl. 1:367-375.Springer, J.T. 1979. Some sources of bias and sampling error in radiotriangulation. J. Wildl. Manage. 43:926-935.Spurr, S.H. 1957. Local climate in the Harvard Forest. Ecology 38:37-46.Swinbank, W.C. 1963. Longwave radiation from clear skies. Quart. J. Roy. Met.Soc. 89:339-348.Thomas, J.W., J. Black, Jr., R.J Scherzinger and R.J Pedersen. 1979. Deer andelk. p.104-127 In: J.W. Thomas (ed.), Wildlife Habitats in ManagedForests: the Blue Mountains of Oregon and Washington. USDA For. Serv.,Agr. Handbook 533. WA. 512 pp.Timmerman, H.R. and J.G. McNicol. 1988. Moose habitat needs. For. Chron.64:238-245.Vales, D.J. 1986. Functional relationships between salal understory and forestoverstory. Research, British Columbia Ministry of Environment and Parksand B.C. Ministry of Forests and Lands. IWIFR-32, Victoria. 121 pp.Van Ballenberghe, V. and D.G. Miquelle. 1990. Activity of moose during springand summer in interior Alaska. J. Wildl. Manage. 54:391-396.Van Dyke, F.G., R.H. Brocke, H.G. Shaw, B.B. Ackerman, T.P. Hemcker andF.G. Lindzey. 1986. Reactions of mountain lions to logging and humanactivity. J. Wildl. Manage. 50:95-102.Wilkinson, L. 1990. SYSTAT: the system for statistics. Systat Inc. Evanston, IL.677 pp.Young, V.A. and W.L. Robinette. 1939. A study of the range habits of elk on theSelway Game Preserve. Univ. of Idaho Bull. No. 16, Moscow. 47 pp.62APPENDIX IOPERATIVE TEMPERATURE SIMULATION MODELSimulation model inputs from canopy data: Sky view factor (SVF):SVF = 1/((1.03-CCC)+1)where:CCC = crown closure class from the forest cover mapEffective leaf area index (Le):Le =1o9 10 (SVF)/-0.324Calculation of diffuse (Sd) and direct (Sb) components of total global radiation(St): Convert total global flux density (Wm - 2) to total hourly flux (MJ•m -2) (St =St•0.0036)1 0 = total hourly extraterrestrial radiation of a horizontal surface (Erbs et al. 1982)= (24/(270)-1 .360.3.6.(cos(lat)cos(5)(sin(0.26179)-sin(0))+(27c/360).15.sin(lat)sin(5))where:lat = latitude in radians5 = solar declination= 0.006918-0.399912•cos(0)+0.070257•sin(0)-0.006758•cos(20)+0.000907.sin(20)-0.002697-cos(344+ 0.00148•sin(30)where:41, = 27Ju/365where:Ju = Julian day63determine ratio (Kt) of global flux (St) to extraterrestrial flux on a horizontalsurface (10):Kt = St/I0 (Erbs et al. 1982)if Kt 0.22 then Sd = (1-0.09•Kt •Stif Kt > 0.22 and 0.8 then Sd = (0.9511-0.1604.Kt+4.388-y-16.638.Kt3+12.336.Kt4).Stif Kt > 0.8 then Sd = 0.165•Stconvert Sd back to hourly flux density (W.m - 2) (Sd = Sd/0.0036)convert St back to hourly flux density (W.m -2) (St = St/0.0036)determine direct radiation (Sb):Sb = St-Sddetermine direct radiation beneath canopy (Sbu ):Sbu = Sb•e( -G- Leicos(e)) (Nilson 1971, Black et al. 1991)where:e = base of the natural logarithmG = angle-dependant extinction coefficient per unit foliage area measuredin the direction of the solar beam (0.5 for randomly distributed foliarelements; Ross 1981)= solar incident angle with respect to the normal to the slope (= the solarzenith angle (Z given below) when slope=0)determine diffuse radiation beneath the canopy (Sdu):Sdu = Scre(-") (Black et al. 1991)Calculation of Operative Temperature (Te) (Campbell 1977):Te = Tam + (re•(Rebs - ceaTa 4))/pcpwhere:Te = operative temperature (°C)Tam = air temperature (°C)Ta = air temperature (°K)re = parallel equivalent resistance to convective and radiative heattransfer (s•m- 1 )Rabs = radiation absorbed by animal surface (W.m -2 )a = Stefan-Boltzman constant (5.67.10 "8 Wm -2 . °K -4)pCp = density of air (p) • specific heat (cp) (1200 J•rn--3 .°K -1 ; Campbell 1977)64Thermal resistance between animal's outer surface and environment (r ah1/re = 1/Rhe + 1/R rwhere:Rr = resistance to longwave radiation transfer= pcp/(4esaTa 3)where:es = emissivity of animal surface (1.0 for caribou; Monteith 1973)Rha = conditional resistance to convection:if Gr•Re 2 < 1 then Rha = 307.(d/U)" (forced convection dominant;Campbell 1977)if Gr•Re 2 1 then Rha = 840•(d/(Tsk-Tam)) 0.25 (free convection dominant;Campbell 1977)where:d = characteristic dimension (1.02 for a 350kg moose;adapted from Parker 1987)U = windspeed (m-s -1 )Re = Reynolds number= (U.d).0 -1Gr = Grashof number= agd 3. (Tsk-Tam)*1) 2where:a = coeffecient of thermal expansion of fluid (1/273 forair; Campbell 1977)g = accelleration due to gravity (9.8m•s -2)Tsk = Temperature of skin (OC)= (34.688.e (0 .0033*Tam) (Renecker and Hudson 1986)..0 = kinematic viscosity of air= 1.151.10 -5 rn•s -1 at standard temperature andpressure65Calculate longwave radiation under the canopy:sky-longwave radiation beneath canopy (L s ):Ls=SVF•Lskywhere:Ls4 = sky longwave radiation= (R.cea+(1-Rc))6Ta 4 (Swinbank 1963)where:Ca = sky emissivity (.0.674+0.007•(Ta); Gates 1980)Rc = ratio of observed 'total global solar radiation' to potential clearsky global radiation= St/Sgowhere:St = global irradiance measured by the pyranometer (W.m -2 )Sgo = clear sky global irradiance= Socos(Z)(0.271+0.706.T (llcos(Z))) (Gates 1980)where:So = solar constant (1360 Wm -2)t = atmospheric transmissivity (0.78 in this study)Z = solar zenith angle= cos -1 (sinpsin(lat)+cos(8)cos(lat) cos(TLA-13)•2•x/360)where:TLA = local apparent time= Td+LONGEQ+EQ (Paldridge andPlatt 1976)where:Td = 24 hour time of dayLONGEQ = standard longitudecorrection (0 in this study)EQ = equation of time= 0.000075+0.001868.cos(4))-0.032077.sin(4)-0.014615.cos(24))-0.040849-sin(4)66longwave radiation from plant canopy (1-SVF)•4,where:Lp = conditional function of Leif Le >2 then 4=Ecaol-ca 4where:Eca = emissivity of canopy (0.97; Black et al. 1991)Tea = temperature of canopy (assumed to equal T a : Tan et al.1978)if Le 5 2 then Lp = ((epaaTpa 4)+((1-epa). ((egraTgr 4)+((1-egr).(SVF.Lsky)))))/(1-((1-epa). (1 -egr).(1 -SVF)))where:egr = emissivity of the ground= 0.97 (Parker unpubl. cited in Parker and Gillingham1990)Tgr = temperature of the ground (approximated by Ta; Blacket al. 1991)longwave radiation from the ground (L g):Lg = egraTgr 4Calculate radiation absorbed by animal (gabs)iRabs = SW+(.5•Lg)+(.5•Ls)+(.5•(1 -SVF)-Lp)where:coefficients of 0.5 denote proportion of animal's surface area exposed toeach type of longwave radiationSW = amount of shortwave radiation absorbed= as (Ap/A-Spu/cos(0)+0.5. Sdu +0.5.SWGRwhere:coefficients of 0.5 denote proportion of animal's surface areaexposed to each type of shortwave radiationas = absorptivity to shortwave radiation (0.75; taken as the mean of theseasonal values for mule deer (0.7, 0.8) given by W.P.Porter in Parker and Gillingham 1990)Ap/A = ratio area on a surface perpendicular to the solar beam to totalsurface area for a shape simulating an ungulate (0.3;Campbell 1977)67Sbu/cos(0) = amount of direct radiation beneath the canopy on a surfaceperpendicular to the beam (Wm -2)SWGR = short wave radiation reflected from the ground= Albedo•Stu (Albedo=0.2; upper limit given for woodland byGeiger 1965)

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075203/manifest

Comment

Related Items