101 DEMOGRAPHY AND SUSTAINABLE HARVEST RATES OF LOW-DENSITY MOOSE IN NORTHERN BRITISH COLUMBIA Ian W. Hatter Nature Wise Consulting, #49-640 Upper Lakeview Road, Invermere, British Columbia V0A 1K3, Canada ABSTRACT: Numerous moose (Alces alces) populations throughout Alaska, Yukon, and the Northwest Territories occur at low-density, a condition that often persists for decades and is referred to as a low-density equilibrium (LDE). The demographic conditions for these populations include low-density (≤ 0.4 moose/km2), low annual recruitment of calves (~ 0.25 calves/cow), and static pop- ulation growth (λ ~ 1.00). I used data from aerial surveys and hunter harvest surveys to assess if these conditions applied to 4 moose populations in northern British Columbia, Canada over a 20-year period from 1996/97–2015/16. All populations exhibited low-density, low recruitment, and static growth suggesting that moose in this part of the province exist within a LDE state. Harvest and moose densi- ties were positively related. Harvest rates from survey data ranged from 2.4–3.2% of the total popula- tion and 6.1–10.5% of the bull population. A stochastic model was used to estimate sustainable harvest rates defined as rates where the harvest risk was ≤ 10% probability that the post-hunt bull:cow ratio dropped below a given adult sex ratio threshold after 50 years of harvest. Sustainable harvest rates averaged ≤ 2.4% of the total population or ≤ 8.4% of the bull population with 0.50 bulls/cow as the threshold, ≤ 3.2% of the population or ≤ 13.0% of bulls with 0.40 bulls/cow as the threshold, and ≤ 4.1% of the population or ≤ 20.4% of bulls with 0.30 bulls/cow as the threshold. Modelling indicated that even small changes in harvest rates could greatly affect the probability of bull:cow ratios dropping below adult sex ratio thresholds. Research focussed on specific factors contributing to low moose density and increased population survey effort should improve estimates of sustainable harvest rates and management of moose in northern British Columbia. ALCES VOL. 58: 101–112 (2022) Key words: adult sex ratio threshold, Alces alces, bull:cow ratio, calf recruitment, demography, low-density equilibrium, moose, predation, sustainable harvest rate Numerous moose (Alces alces) populations throughout Alaska, Yukon, and the Northwest Territories occur at low-density (≤ 0.4 moose/ km2) in a state often referred to as a low-den- sity equilibrium (LDE) or low-density dynamic equilibrium (LDDE), that may persist for decades (Bergerud 1992, Gasaway et al. 1992, Stenhouse et al. 1995, Lake et al. 2013, Joly 2017). In this state, body condition, productiv- ity, and survival of adult moose are relatively high and sufficient to allow population growth. However, the population does not increase because wolf (Canis lupus) and bear (Ursus spp.) predation limit calf recruitment. Harvest is typically restricted to bull-only to prevent further decline in these predator-prey systems (Gasaway et al. 1992). Bergerud (1992) hypothesized that for each low-density popula- tion there is also a stabilizing recruitment of calves (Rs) where the growth rate is static. In unhunted or lightly hunted moose populations Rs ~ 0.25 calves/cow (Bergerud 1992, Bergerud and Elliott 1998). Messier (1994) modelled moose-wolf interactions over a broad spectrum of moose densities in North America and predicted that moose would DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 102 stabilize at ~ 2.0 moose/km2 in the absence of predation and at ~ 1.3 moose/km2 in the pres- ence of wolves. However, if moose productiv- ity was diminished through deteriorating habitat quality or bear predation on calves, then a LDE (0.2–0.4 moose/km2) was pre- dicted. These demographic conditions for LDE populations have received general support from additional reviews on the effects of pre- dation on moose numbers (Van Ballenberghe and Ballard 1994, Ballard and Van Ballenberghe 1997, Van Ballenberghe and Ballard 1997). Hatter (1999) reviewed moose harvest management in central and northern British Columbia and suggested that the demog- raphy of 4 northern moose populations adjacent to Yukon and the Northwest Territories may exist in a LDE state, and that sustainable harvest rates may be sub- stantially lower than in higher density populations in central British Columbia. The objectives of this analysis were to: 1) assess if the demographic conditions for LDE populations (i.e., low-density [≤ 0.4 moose/km2], low calf recruitment [R ~ 0.25 calves/cow], and static popula- tion growth [λ ~ 1.0)]) applied to these 4 populations from 1996/97–2015/16; 2) estimate average harvest rates for each population based on survey and harvest densities during this 20-year period; and 3) determine sustainable harvest rates for moose existing in a LDE state based on a stochastic demographic model. STUDY AREA The study area included Game Management Zones (GMZ) 6f, 6e, 7Pd, and 7Pe in northern British Columbia, Canada (Fig. 1). GMZs are amalgamations of adjacent Wildlife Management Units which share similar eco- logical characteristics and hunter harvest pat- terns, and thus provide a suitable framework for assessing moose demography and harvest rates over large geographical areas (Hatter 1999, Kuzyk et al. 2018). GMZs 6f, 6e, 7Pd, and 7Pe comprised 29,256 km2, 44,587 km2, 38,330 km2, and 61,204 km2 of habitable moose range, respectively. Six moose eco- types have been identified in British Columbia based on ecological, climatic, and physio- graphic differences in their habitats (Eastman and Ritcey 1987). The Boreal Upland ecotype occupies GMZs 6f, 6e, and 7Pd while the Fig. 1. Location of Game Management Zones 6f, 6e, 7Pd, and 7Pe in northern British Columbia. ALCES VOL. 58, 2022 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER. 103 Boreal Lowland ecotype occupies GMZ 7Pe. The Boreal Upland ecotype primarily occurs within the Boreal Black and White Spruce, the Subalpine Spruce-Willow-Birch, and the Alpine Tundra biogeoclimatic zones. GMZ 7Pe is comparatively homogeneous and the Boreal Lowland ecotype almost exclusively occupies the Boreal White and Black Spruce biogeoclimatic zone. In general, moose habi- tat quality is higher in the Spruce-Willow- Birch zone of GMZs 6f, 6e, and 7Pd and lowest in the Boreal Black and White Spruce zone in GMZ 7Pe. The climate for all 4 GMZs is characterized by long, cold winters and cool, short summers. Moose co-exist with wolves, grizzly bears (U. arctos), and black bears (U. americanus) in all 4 GMZs. Wolf density ranged between 5–15 wolves/1000 km2 with lower densities in GMZ 7Pe (BC FLNRO 2014, Serrouya et al. 2015). Grizzly bear density ranged from 10–30 bears/1000 km2 in GMZs 6f, 6e, and 7Pd with ≤ 10 bears/1000 km2 in GMZ 7Pe (BC FLNRO 2020). Black bears occur at unknown density in all 4 GMZs. Caribou (Rangifer tarandus) occur throughout much of the study area, with mountain goats (Oreamnos americanus) and thinhorn sheep (Ovis dalli) in mountainous regions. Elk (Cervus canaden- sis), mule deer (Odocoileus hemionus), and white-tailed deer (O. virginianus) are moder- ately abundant in the extreme southern portion of GMZ 7Pe, and either absent or occur at very low to low density elsewhere. Small numbers of reintroduced bison (Bison bison) occupy portions of GMZs 7d and 7e. Licensed hunting seasons were restricted to bull-only (bag limit of 1 moose/hunter) in all GMZs with season dates generally between 15 August and 30 November. Bull hunting was primarily regulated by general open seasons although limited-entry seasons with no antler restrictions, or a combination of general open and limited entry seasons occurred in some areas (Kuzyk et al. 2018). Only bulls with spike or fork antlers and bulls with tri-palm antlers could be harvested during September – October from 1996– 2002 in GMZs 7Pd and 7Pe. After 2002, bulls with at least 10 points on one antler could also be harvested. Indigenous people also harvest moose at unknown levels throughout the 4 GMZs. Harvest is generally tied to access from roads and waterways, and although considerable road access exists within GMZ 7Pe, it is limited elsewhere. METHODS Moose and Hunter Harvest Surveys Moose were counted in helicopter surveys in each GMZ, generally in mid-December through late February in suitable weather and snow conditions. Stratified random block or distance sampling survey methods were used to enumerate bulls (1+ year-old males), cows (1+ year-old females), and calves (Kuzyk et al. 2018); unclassified ani- mals were typically ≤5% of total count. The survey areas were assumed representative of moose density and composition within a GMZ. Although herd composition counts were also conducted periodically, they were deemed less reliable and not used in this analysis. A total of 20 surveys were used to estimate density, and bull:cow and calf:cow ratios from 1996/97–2015/16 in the 4 GMZs. Annual hunter surveys were used to esti- mate the licensed resident harvest, number of hunters, and hunter days from 1996–2015. Annual mail questionnaires were sent to 15,477 (average) provincial hunters chosen randomly, with an average response rate of 68% (Kuzyk et al. 2018). Reporting of non-resident licenced harvest was manda- tory and obtained from guide declarations. Moose Demography Moose within each GMZ were assumed to comprise a discrete population (Hatter 1999, DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 104 Kuzyk et al. 2018). I used the resident licensed kills/100 hunter days of effort (KPUE) as an annual population index for each GMZ, and estimated the growth rate (λ) by regressing ln(KPUE) against year from 1996–2015 where λ = er and r is the expo- nential rate of growth. The 95% CIs were estimated by bootstrapping with 2000 sam- ples (Haddon 2001). I considered the popu- lation to be stable if the 95% CIs for λ included 1.00. I determined the average win- ter moose density, calf:cow ratio and bull:cow ratio for each GMZ from 1996/97– 2015/16 by weighting each estimate by the size of the survey area. This assumed larger survey areas provided a more accurate esti- mate of the population parameters at the GMZ level, that the population was stable, and that each survey area was randomly placed within the GMZ. I used 2000 boot- strap samples to estimate 95% CIs on the weighted (by area) estimates of the popula- tion ratios and moose density. Harvest den- sity was calculated as the bull harvest/1000 km2 of moose range within each GMZ and the 95% CIs with 2000 bootstrap samples. Harvest rates were calculated for each GMZ as (bull harvest/1000 km2)/(moose/km2 ×1000) for the population harvest rate and (bull harvest/1000 km2)/(bulls/km2 ×1000) for the bull harvest rate. Sustainable Harvest Rates Sustainable harvests were characterized as those where the post-hunt bull:cow ratio remained above a specified adult sex ratio threshold. I considered 3 ratio thresholds: 0.3, 0.4, and 0.5 bull:cow. I used these thresholds as both 0.3 and 0.5 bull:cow ratios are used as management targets in British Columbia (BC MOE 2010), and 0.4 is used in Yukon (Czetwertynski 2015). I assessed the licensed harvest only (bulls), and did (could) not consider the Indigenous harvest as it was unknown. I estimated harvest sustainability with a demographic model parameterized on the average density and population composition from moose winter surveys of the 4 GMZs; density dependence was not modelled. Nutritional density dependence affecting pop- ulation performance was unlikely due to the low moose density, and inadequate informa- tion existed to model density dependence of mortality due to predation. The modelled post- hunt population (N) was partitioned into bulls (B), cows (C), and calves (Ca) as follows: Nt+1 = Bt+1 + Ct+1 Cɑt+1 (1) where: Bt+1 = BtSmt + 0.5(CatSjt) – (MhtNt) (2) or Bt+1 = BtSmt + 0.5(CatSjt) – (MhtBt) (3) Ct+1 = CtSft + 0.5(CɑtSjt) (4) Cɑt+1 = Ct+1 × Rpst,t+1 (5) where t was year, Sm was the annual bull survival rate, Sf was the annual cow sur- vival rate, Sj was the annual calf survival rate (i.e., from post-hunt or 0.5 years-of- age to the next post-hunt period or 1.5 years-of-age), Rpst was the post-hunt recruitment rate (i.e., the average calf:cow ratio from moose density surveys of the 4 GMZs), and Mh was the harvest rate. The post-hunt calf sex ratio was assumed to be 1:1 (Ballard et al. 1991, Boer 1992). Sf was set to 0.89 which was the average survival rate from 5 cow mortality studies within low-density moose populations (Sf = 0.91, Larsen et al. 1989; Sf = 0.91, Gasaway et al. 1992; Sf = 0.88, Stenhouse et al. 1995; Sf = 0.88, Bertram and Vivian 2002; Sf = 0.90, Joly et al. 2017). Previous studies of natu- rally fluctuating moose populations reported adult sex ratios near parity, ALCES VOL. 58, 2022 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER. 105 suggesting that Sm and Sf are equal or 1:1 (Timmerman 1992). Conversely, the annual survival rate of bulls was lower than cows in the unhunted moose population on Isle Royale (Peterson 1977), with the average adult sex ratio of 0.8 bull:cow on the island from 1950–1981 (Page 1992). I therefore modelled 3 bull survival rates correspond- ing to unhunted adult sex ratios: 1.0, 0.9, and 0.8 bull:cow. The annual calf survival rate was determined by iteratively adjusting Sj until the modelled population achieved a stable age distribution with λ = 1.0. I used this value of Sj in the model projections as I was interested in sustainable harvest rates for stable, low-density moose populations. Bergerud (1992) considered moose calves at 6–9 months of age to be fully recruited, or that calves and cows have similar winter sur- vival rates. However, several studies found that winter survival rates of calves are lower than cows (Ballard et al. 1991, Joly et al. 2017, Kuzyk et al. 2019), suggesting that recruitment should be measured when calves are 1 year-of- age (Hatter 2020). I determined the annual recruitment rate (R) from Rpst and the ratio of the winter calf (Sjw) and cow (Sfw) survival rates where: R = Rpst × Sjw ∕ Sfw (6) (Hatter 2020). Mortality of cows was assumed negligible from calving to the post- hunt period (i.e., Sfw = Sf ). Similarly, mor- tality from 1 year-of-age to the post-hunt period was assumed negligible (i.e., Sjw = Sj). Rs was equal to R when λ = 1.0. Harvest rates were applied to either the total population (bull harvest/post-hunt population, Eq. 2) or the bull population (bull harvest/post- hunt bulls, Eq. 3) because both metrics are com- monly used by moose managers in British Columbia (Kuzyk et al. 2018). I considered har- vest rates up to 6% of the total population and up to 25% of the bull population. I made the demographic model stochastic to assess harvest risk (i.e., the probability that the harvest rate was not sustainable) by includ- ing the SE for each parameter based on esti- mates of the coefficient of variation (CV). I set CV(Sj) = 0.15 (Ballard et al. 1991) and CV(Sf) = 0.017 from the 5 cow mortality studies. The CV for the winter calf:cow ratio was calculated from survey estimates in the 4 GMZs. The CV for implementation uncertainty (i.e., annual variation in harvest rates) was set to 19% to match annual changes in reported harvests for each of the 4 GMZs during 1996–2015. The CV for survey uncertainty of the population size and bull:cow ratios was set to the survey standard by Gasaway et al. (1986) where a CV of 15.2% equals + 25% of the true value 90% of the time. For the survival rates, I portioned the total vari- ance for Sj, Sf, and Sm into SEenvironmental and SEsampling by assuming 50% of the variance was due to environmental uncertainty and 50% was sampling error. Finally, I generated correlated random winter survival rates (Burgman et al. 1993) with r = 0.75 for bulls and cows, r = 0.5 for cows and calves, and r = 0.5 for bulls and calves. All random variables were drawn from a normal distribution. I conducted 2000 Monte Carlo simulations of the model for each bull survival rate and har- vest rate during a 50-year period. I chose 50 years because mean annual bull:cow ratios changed very little after that. Survey uncertainty was incorporated into the initial simulation year. The harvest rate was considered sustainable if the percentage of the simulations that resulted in final (year 50) bull:cow ratios below the adult sex ratio threshold was ≤ 10%. Simulations were performed with the Microsoft Excel add-in program PopTools 3.2 (Hood 2011). RESULTS Moose Demography Six moose surveys ranging from 1943–7766 km2, 7 surveys ranging from 1533–7319 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 106 km2, 3 surveys ranging from 5675–8917 km2, and 4 surveys ranging from 11,022– 34,260 km2 were conducted in GMZ 6f, GMZ 6e, GMZ 7Pd, and GMZ 7Pe, respec- tively, in 1996/97–2015/16 (Fig. 2). Moose population growth rate based on KPUE was approximately stable in the 4 GMZs over the 20-year period (Table 1). Although GMZ 7Pe had a slightly declining growth rate (λ = 0.98), the 95% CIs included λ =1.00. Further, no decline occurred in the density estimates in GMZ 7Pe from the 4 moose surveys con- ducted in 2004–2016, and calf:cow ratios were among the highest in the 4 GMZs (Table 1). The average moose density was 0.27/ km2, 0.36/km2, 0.16/ km2, and 0.09/km2 in GMZ 6f, GMZ 6e, GMZ 7Pd, and GMZ 7Pe, respectively. The winter calf:cow ratio ranged from 0.24 (GMZ 7Pd) to 0.32 (GMZ 7Pe), averaging 0.28 (Table 1); the CV was 0.27 from the 20 moose surveys in the 4 GMZs. The bull:cow ratio ranged from 0.58 (GMZ 7Pe) to 0.97 (GMZ 7Pd), averaging 0.75 across the 4 GMZs (Table 1). Assuming a stable population (λ = 1.0), 0.28 calf:cow ratio in early winter, a den- sity of 0.22 moose/km2, and a cow winter survival rate of 89%, the modelled winter calf survival rate was 79% and stabilizing recruitment (Rs) was 0.25 calves/cow. R for each moose population, after adjust- ment for winter survival rates of cows and calves, was 0.25 (GMZ 6f), 0.26 (GMZ 6e), 0.21 (GMZ 7Pd), and 0.28 (GMZ 7Pe). Harvest density based on hunter survey data was lowest in GMZ 7Pe (3.0/1000 km2) and highest in GMZ 6e (8.5/1000 km2) (Table 1). Population harvest rates from moose and hunter survey data were 2.5% (GMZ 6f), 2.4% (GMZ 6e), 2.9% (GMZ 7Pd), Fig. 2. Survey and harvest parameters for Game Management Zones 6f, 6e, 7Pd, and 7Pe from 1996/97– 2015/16. N/km2 = moose/km2, B:C = bull:cow ratio, Ca:C = calf:cow ratio, and KPUE = resident kill per 100 resident hunter days. ALCES VOL. 58, 2022 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER. 107 and 3.2% (7Pe); bull harvest rates were 7.3% (GMZ 6f), 6.1% (GMZ 6e), 6.7% (GMZ 7Pd), and 10.5% (GMZ 7Pe). The average harvest and moose density were highly cor- related among GMZs (r = 1.00, P = 0.001, Fig. 3), but harvest rate and bull:cow ratio were not (population harvest rate: r = −0.18, P = 0.82; bull harvest rate: r = −0.75, P = 0.25). Sustainable Harvest Rates Sustainable harvest rates were affected by different bull survival rates, with lower sur- vival rates associated with reduced harvests (Table 2). Population harvest rate up to 2.0% of the modelled post-hunt moose population posed little risk of reducing the bull:cow ratio below threshold values (Table 2a). Sustainable population harvest rates, aver- aged across the 3 estimates of Sm, were ≤2.4% with the 0.5 bull:cow ratio threshold, ≤3.2% with the 0.40 threshold, and ≤4.1% with the 0.30 threshold. The average popula- tion harvest rate was 2.8% based on survey data which was the upper limit of sustain- ability for the 0.5 bulls/cow threshold. However, the bull:cow ratios from winter moose surveys were ≥0.5, suggesting that the modelling results are conservative. Simulated bull harvest rates up to 7% of the modelled post-hunt bull population presented little risk of lowering adult sex ratios below bull:cow thresholds (Table 2b). Sustainable bull harvest rates averaged ≤8.4% with the 0.5 bull:cow threshold, ≤13.0% with the 0.40 threshold, and ≤20.4% with the 0.30 threshold. The Table 1. Demographic status of low-density moose populations within 4 Game Management Zones (GMZs) located in northern British Columbia, 1996/97–2015/16. Moose/km2 and sex/age composition are from moose density surveys during winter. Growth rate and harvest are from hunter harvest surveys. Numbers in brackets are 95% CIs. GMZ Moose/km2 Bulls/Cow Calves/Cow Growth rate (λ) Harvest/1000 km2 6f 0.27 0.65 0.28 1.01 6.5 (0.20–0.35) (0.57–0.74) (0.24–0.34) (0.99–1.02) (6.1–7.0) 6e 0.36 0.81 0.29 1.00 8.5 (0.29–0.43) (0.69–0.90) (0.24–0.35) (0.98–1.01) (7.9–9.1) 7Pd 0.16 0.96 0.24 0.99 4.5 (0.07–0.34) (0.73–1.16) (0.21–0.29) (0.97–1.02) (4.2–4.7) 7Pe 0.09 0.58 0.32 0.98 3.0 (0.06–0.12) (0.51–0.69) (0.25–0.36) (0.97–1.00) (2.6–3.4) x̅ 0.22 0.75 0.28 1.00 5.6 SD 0.12 0.17 0.03 0.01 2.4 Fig. 3. Relationship between average surveyed hunter kill density (harvest/1000 km2) and average surveyed moose density (moose/ km2) in four Game Management Zones in northern British Columbia, 1996/97– 2015/16. Error bars display minimum and maximum values. DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 108 average bull harvest rate based on survey data was 7.7% which was within the range of modelled sustainable rates for the 0.5 bull:cow threshold. Harvest risk rose sharply as the mod- elled harvest rates were increased. For exam- ple, a population harvest rate of 2.5% had a 0% probability of reducing the bull:cow ratio below 0.5 bulls/cow, whereas a harvest rate of 3.5% had a 95% probability (Fig. 4). Similarly, a bull harvest rate of 8% had a 0% chance of reducing the adult sex ratio below 0.5 bull:cow, but a 12% harvest had an 81% chance (Fig. 5). DISCUSSION Moose from all 4 GMZs in northern British Columbia met the demographic conditions for LDE populations. All populations were at low density (≤0.4 moose/km2) and their long-term (20 year) population growth rate was static (λ ~ 1.0). Across the GMZs, sur- veys indicated that the winter calf:cow ratio averaged 0.28, the modelled winter calf Table 2. Sustainable harvest rates (Mh) where bull survival rates (Sm) are 0.890 (unhunted sex ratio = 1.0 bull/cow), 0.878 (unhunted sex ratio = 0.9 bulls/cow), and 0.862 (unhunted sex ratio = 0.8 bulls/cow). a. Harvest rate applied to total population (Mh = bull harvest/post-hunt population) Threshold value Sm = 0.890 Sm = 0.878 Sm = 0.862 0.3 bulls/cow Mh ≤ 4.4% Mh ≤ 4.1% Mh ≤ 3.8% 0.4 bulls/cow Mh ≤ 3.6% Mh ≤ 3.1% Mh ≤ 2.8% 0.5 bulls/cow Mh ≤ 2.8% Mh ≤ 2.4% Mh ≤ 2.0% b. Harvest rate applied to bull population (Mh = bull harvest/post-hunt bulls) Threshold value Sm = 0.890 Sm = 0.878 Sm = 0.862 0.3 bulls/cow Mh ≤ 21.6% Mh ≤ 20.6% Mh ≤ 19.1% 0.4 bulls/cow Mh ≤ 14.2% Mh ≤ 13.2% Mh ≤ 11.5% 0.5 bulls/cow Mh ≤ 9.6% Mh ≤ 8.4% Mh ≤ 7.0% Fig. 5. Relationship between harvest risk and bull population harvest rate for post-hunt adult sex ratio thresholds of 0.30, 0.40, and 0.50 bulls/ cow. The harvest risk is the estimated probability that a specified harvest rate will result in a bull:cow ratio (B/C) below the adult sex ratio threshold. The adult bull survival rate was 0.890. Fig. 4. Relationship between harvest risk and total population harvest rate for post-hunt adult sex ratio thresholds of 0.30, 0.40, and 0.50 bulls/ cow. The harvest risk is the estimated probability that a specified harvest rate will result in a bull:cow ratio (B/C) below the adult sex ratio threshold. The adult bull survival rate was 0.890. ALCES VOL. 58, 2022 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER. 109 survival was 78%, and annual calf recruit- ment averaged 0.25 calves/cow. Bergerud (1992) hypothesized that the mechanism for a LDE in a moose population is density-dependent wolf predation on calves, augmented by density-independent bear predation on neonates; food was not considered to be a limiting factor. Joly et al. (2017), however, studied a low-density (0.06–0.12 moose/km2) population in north-central Alaska and found that produc- tivity was significantly lower in areas with less high-quality habitat, indicating that for- age resources/nutrition could be a contribut- ing factor in a LDE population. They suggested that the relative roles of predation on young calves, winter weather, and nutri- tional constraints likely interact to maintain low moose density. Much of the moose range in GMZ 7Pe consists of low to moderate hab- itat capability (Thiessen 2010), suggesting that habitat quality may be a local limiting factor in this population. Alternatively, the lower moose density in GMZ 7Pe may reflect a lack of suitable habitat for moose to space- out and avoid predation (Bergerud 1992). Two important considerations in manag- ing low-density moose populations are to maintain an appropriate bull:cow ratio that provides hunting opportunities for large-ant- lered bulls and to ensure adequate numbers of bulls are available to locate and breed all receptive cows (Timmermann 1992). Given these considerations, adult sex ratios for low-density populations (≤0.2 moose/km2) in British Columbia are managed with an objective to maintain a bull:cow ratio ≥0.5, with higher density populations managed with a threshold bull:cow ratio ≥0.3 (BC MOE 2010). In Yukon, low-density moose are managed to maintain an adult bull:cow ratio of at least 0.4 (Czetwertynski 2015), while in parts of Alaska the management objective is a ratio of 0.3 (Young and Boertje 2008). I used a stochastic model to assess harvest risk, or the probability that a speci- fied harvest rate will reduce the bull:cow ratio below the adult sex ratio threshold. I considered the harvest rate as sustainable if the harvest risk was ≤10%. Sustainable har- vest rates averaged ≤2.4% of the total popu- lation or ≤8.4% of the bull population at the 0.50 bull:cow ratio threshold, ≤3.2% of the population or ≤13.0% of bulls at the 0.40 threshold, and ≤4.1% of the population or ≤20.4% of bulls at the 0.30 threshold. Similarly, a harvest rate of 10% of adult bulls or 2.2–3.3% of the total population was sustainable in low-density Yukon moose populations with a bull:cow ratio objective (threshold) of 0.4 (Czetwertynski 2015). While the surveyed harvest and moose densities were highly correlated, harvest rates and bull:cow ratios were not. This may indicate that the surveyed bull:cow ratios were not representative of the GMZ, or some compensatory mortality existed in the licensed harvest. In addition, unmeasured variation in the mortality rates of cows could have affected the bull:cow ratio. I used KPUE to ascertain long-term (20 year) stability of the moose population within each of the 4 GMZs. While several studies have found that KPUE is correlated with moose abundance (Crête et al. 1981, Fryxell et al. 1988), there is potential for error or bias in growth rates measured with KPUE (Fryxell et al. 1988, Bowyer et al. 1999, Hatter 2001, DeCesare et al. 2016). I assumed that the survey areas provided a representative sample of the moose density and bull:cow and calf:cow ratios within each GMZ. Another data concern was that the Indigenous harvest was unknown and not an annual harvest metric (Kuzyk et al. 2018). I considered 10% as an acceptable, although subjective level of harvest risk. While higher risk levels such as 15% or 20% could have been chosen with greater risk tol- erance, sustainable harvest rates would have DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 110 only increased slightly. For example, select- ing a 20% harvest risk versus 10% risk only increased the sustainable population harvest rate from 2.8% to 2.9% for the 0.5 bull:cow ratio threshold with Sm = 0.870. Computation of stabilizing recruitment required an estimate of the winter cow sur- vival rate. I assumed all cow mortality occurred during winter such that the winter survival rate was equal to the annual survival rate. However, studies of low-density moose populations have documented that some cow mortality occurs during summer (Larsen et al. 1989). Calculation of Rs would have been improved with a more reliable estimate of the winter cow survival rate. I did not include density-dependent nutritional effects in the demographic model as moose density in the 4 GMZs ranged from 0.1–0.4 km2, and thus were well below the habitat carrying capacity of 1.5–2.0 moose/ km2 for North American populations (Bergerud 1992, Messier 1994). Despite this, nutritional influences may have been present within GMZ 7Pe where habitat qual- ity was lower. Inclusion of density depen- dence in the model would have increased sustainable harvest rates (Caughley 1977). The stochastic modelling indicated that minor changes in harvest rates could greatly affect the probability of bull:cow ratios dropping below adult sex ratio thresholds, revealing that small changes in harvest man- agement could affect the resulting bull:cow ratios. Sustainable harvest rates were also affected by different bull survival rates with lower survival rates sustaining reduced har- vests. I did not model non-stationarity in environmental variation although global warming is expected to be highly influential to ungulate population dynamics in northern latitudes where variation in demographic rates, and thus sustainable harvest rates, are closely aligned to annual cycles in climate and primary productivity (Brown 2011). For these reasons, wildlife managers should be cautious about applying estimates of sustain- able harvest rates from this study to other low-density moose populations. Despite these limitations, I was able to offer support for the LDE hypothesis for all 4 GMZs in northern British Columbia and provide additional insight on sustainable harvest rates for moose within these north- ern ecosystems. More study is recommended to identify the relative roles of hunting, pred- ators (wolves and bears), habitat quality, and winter weather in maintaining low densities of moose across northern British Columbia. Further understanding of sustainable harvest rates would benefit from additional moose surveys, Indigenous harvest surveys, and studies of moose survival rates. ACKNOWLEDGEMENTS M. Bridger, D. Heard, and H. Schindler kindly reviewed an early draft of the manu- script and made numerous helpful and insightful suggestions. The manuscript was also greatly improved by constructive reviews from N. DeCesare and 2 reviewers. REFERENCES Ballard, W. B., J. S. Whitman, and D. J. reed. 1991. Population dynamics of moose in south-central Alaska. Wildlife Monographs 114: 3–49. _____, and V. Van BallenBerghe. 1997. Predator/prey relationships. Pages 247– 273 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and Management of the North America Moose. Wildlife Management Institute, Washington, DC, USA. Bergerud, A. T. 1992. Rareness as an anti- predator strategy to reduce predation risk for moose and caribou. Pages 1008–1021 in D. R. McCullough and R. H. Barrett, editors. Wildlife 2001: Populations. Elsevier Applied Science Publishers, London, United Kingdom. ALCES VOL. 58, 2022 DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER. 111 _____, and J. P. elliott. 1998. Wolf preda- tion in a multiple-ungulate system in northern British Columbia. Canadian Journal of Zoology 76: 1551–1569. doi: 10.1139/z98-083 Bertram, M. R., and M. T. ViVion. 2002. Moose mortality in eastern interior Alaska. Journal of Wildlife Management 66: 747–756. doi: 10.2307/3803140 Boer, A. H. 1992. Fecundity of North American moose (Alces alces): a review. Alces Supplement 1 (1992): 1–10. BoWyer, R. T., M. C. nicholson, E. M. molVar, and J. B. Faro. 1999. Moose on Kalgin Island: are density-dependent pro- cesses related to harvest? Alces 35: 73–89. British columBia ministry oF enVironment (BC MOE). 2010. Moose harvest manage- ment procedure. Fish and Wildlife Branch, Victoria, British Columbia, Canada. British columBia ministry oF Forests, lands and natural resource operations (BC FLNRO). 2014. Management plan for the Grey Wolf (Canis lupus) in British Columbia. Fish and Wildlife Branch, Victoria, British Columbia, Canada. _____. 2020. British Columbia Grizzly Bear Population Estimate for 2018. Fish and Wildlife Branch, Victoria, British Columbia, Canada. BroWn, G. S. 2011. Patterns and causes of demographic variation in a harvested moose population: evidence for the effects of climate and density-dependent drivers. Journal of Animal Ecology 80: 1288–1298. doi: 10.1111/j.1365-2656.2011.01875.x Burgman, M. A., S. Ferson, and R. akçakaya. 1993. Risk assessment in conservation biology. Chapman and Hall, New York, New York, USA. caughley, G. 1977. Analysis of vertebrate populations. John Wiley and Sons, New York, New York, USA. crête, M., R. J. taylor, and P. A. Jordon. 1981. Optimization of moose harvest in southwestern Quebec. Journal of Wildlife Management 45: 598–611. doi: 10.2307/3808693 czetWertynski, S. 2015. Sustainable harvest rates for moose in Yukon, draft report. Fish and Wildlife Branch, Department of Environment, Government of Yukon, Whitehorse, Yukon, Canada. decesare, N. J., J. R. neWBy, V. J. Boccadori, T. chilton-radandt, T. thier, D. Waltee, K. podruzny, and J. A. gude. 2016. Calibrating minimum counts and catch-per-unit effort as indi- ces of moose population trend. Wildlife Society Bulletin 40: 537–547. doi: 10.1002/wsb.678 eastman, D. S., and R. ritcey. 1987. Moose habitat relationships and management in British Columbia. Swedish Wildlife Research Supplement 1: 101–117. Fryxell, J. M., W. E. mercer, and R. gellately. 1988. Population dynamics of Newfoundland moose using cohort analysis. Journal of Wildlife Management 52: 14–21. doi: 10.2307/ 3801050 gasaWay, W. C., R. D. BoertJe, D. grangaard, D. G. kellyhouse, R. O. stephenson, and D. G. larsen. 1992. The role of predation in limiting moose at low densities in Alaska and Yukon and implications for conservation. Wildlife Monographs 120: 3–59. _____, S. D. duBois, D. J. reed, and S. J. harBo. 1986. Estimating moose popula- tion parameters from aerial surveys. Biological Papers of the University of Alaska, Number 22. Institute of Arctic Biology, University of Alaska, Fairbanks, Alaska, USA. haddon, M. 2001. Modelling and quantita- tive methods in fisheries. Chapman and Hall/CRC, London, United Kingdom. hatter, I. W. 1999. An evaluation of moose harvest management in central and northern British Columbia. Alces 35: 91–103. _____. 2001. An assessment of catch per unit effort to estimate rate of change in deer and moose populations. Alces 37: 71–77. DEMOGRAPHY AND HARVEST OF LOW-DENSITY MOOSE – HATTER ALCES VOL. 58, 2022 112 _____. 2020. Revisiting the recruit- ment-mortality equation to assess moose growth rates. Alces 56: 39–47. hood, G. M. 2011. PopTools version 3.2.h. (accessed September 2022). Joly, K., T. craig, M. D. cameron, A. E. gall, and M. S. sorum. 2017. Lying in wait: limiting factors on a low-density ungulate population and the latent traits that can facilitate escape from them. Acta Oecologica 85: 174–183. doi: 10.1016/j.actao.2017.11.004 kuzyk, G., I. hatter, S. marshall, C. procter, B. cadsand, D. lirette, H. schindler, M. Bridger, P. stent, A. Walker, and M. klaczek. 2018. Moose population dynamics during 20 years of declining harvest in British Columbia. Alces 54: 101–119. _____, C. procter, S. marshall, H. schindler, H. schWantJe, M. scheideman, and D. hodder. 2019. Determining factors affecting moose population change in British Columbia: testing the landscape change hypothesis. Wildlife Working Report. No. WR-127. Progress Report: February 2012–April 2019. British Columbia Ministry of Forests, Lands and Natural Research Operations and Rural Development, Victoria, British Columbia, Canada. lake, B. C., M. R. Bertram, N. guldager, J. R. caikoski, and R. O. stephenson. 2013. Wolf kill rates across winter in a low-density moose system in Alaska. Journal of Wildlife Management 77: 1512–1522. doi: 10.1002/jwmg.603 larsen, D. G., D. A. gauthier, and R. L. markel. 1989. Causes and rate of moose mortality in southwest Yukon. Journal of Wildlife Management 53: 548–557. doi: 10.2307/3809175 messier, F. 1994. Ungulate population mod- els with predation: A case study with the North American moose. Ecology 75: 478–488. doi: 10.2307/1939551 page, R. E. 1992. Can we expect to achieve natural moose sex ratios? Pages 26–34 in D. F. Hatler, editor. Proceedings of the 1991 Moose Harvest Management Workshop. Wildlife Branch, British Columbia Ministry of Environment, Victoria, British Columbia, Canada. peterson, R. O. 1977. Wolf ecology and prey relationships on Isle Royale. United States National Park Service Scientific Monograph No. 11. Superintendent of Documents, Washington, DC, USA. serrouya, R., H. Van oort, and C. demars. 2015. Wolf census in Three Boreal Caribou Ranges in British Columbia; Results from 2015. Report prepared by the Alberta Biodiversity Monitoring Institute, Edmonton, Alberta, Canada. stenhouse, G. B., P. B. latour, L. kutny, N. maclean, and G. gloVer. 1995. Productivity, survival, and movements of female moose in a low-density popu- lation, Northwest Territories, Canada. Arctic 48: 57–62. thiessen, C. 2010. Horn River Basin Moose Inventory, January/February 2010. Peace region Technical Report. British Columbia Ministry of Environment, Fort St. John, British Columbia, Canada. timmermann, H. R. 1992. Moose sociobiol- ogy and implications for harvest. Alces 28: 59–77. young, D. D., and R. D. Boertje. 2008. Recovery of low bull:cow ratios of moose in interior Alaska. Alces 44: 65–71. Van BallenBerghe, V., and W. B. Ballard. 1994. Limitation and regulation of moose populations: the role of preda- tion. Canadian Journal of Zoology 72: 2071–2077. doi: 10.1139/z94-277 _____, and _____. 1997. Population dynam- ics. Pages 223–245 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and management of the North American moose. Wildlife Management Institute, Washington, DC, USA. https://poptools.software.informer.com/3.2/ https://poptools.software.informer.com/3.2/