Summary

Part 1

I am examining the seasonal changes in body condition for the Beverly herd, with the objective of determining how seasonal changes in body condition are related to changes in population trajectory for this population from 1980-87.  During this time period, Thomas and Kiliaan (1998) investigated the Beverly herd's range conditions, prompted by concerns that recent wildfires had destroyed large areas of winter range.  Extensive December and March collections of caribou produced a large (n = 1258 caribou) dataset of body condition measurements (e.g., back fat depth, % femur marrow, kidney fat, reproductive status).  Over-winter trends in body conditions demonstrated that adult females maintained or gained condition over winter, implying that winter range was sufficient for the Beverly herd during the study period.  An explanation for the change in abundance was never determined.  I examined data for females only (n = 856) because of the tight relationship between female body condition and herd productivity.  I divided the data into three phases relative to the highest population size during the study period: increase (1980-83), peak (1984) and decline (1985-87) and examined over-winter (i.e., December - March) and over-summer (i.e., March - December) changes in body condition indices of the female caribou.  Body condition indices included total body mass (kg), back fat depth (mm), kidney fat mass (g), and % femur marrow fat. Caribou protein reserves also exhibit seasonal fluctuations and maternal protein reserves are important for gestation and lactation.  I thus included total body mass as a composite of protein and fat reserves, in the absence of a separate estimate of body protein.  I compared over-summer and over-winter changes in these indices among three phases using non-parametric analysis of variance.

Next steps will include examination of satellite imagery collected during the study period to quantify summer vegetation cover and timing of spring green-up, and to examine changes in winter diet throughout the study period as represented by rumen samples.  The satellite imagery will offer greater insight into the physiognomy of the Beverly range, and the rumen samples will indicate how winter diet may have changed.the three phases using non-parametric analysis of variance.

Part 2

Development of population model for caribou that includes climate effects is identified as a research priority of the Government of Nunavut, because caribou are arguably the most important species for subsistence hunting in the Canadian arctic and there is a paucity of modelling tools with which to anticipate future changes in caribou abundance (H. Robison, pers. comm.)  My objective is thus to construct and parameterize a population model for barren-ground caribou that incorporates demographic stochasticity as well as changes in survival and fecundity that result from hunting and climate effects.

This model will initially be parameterized with data from the Qamanirjuaq caribou herd.  The Qamanirjuaq herd resides in eastern Nunavut, and was estimated at 496,000 animals in 1994. Its population is believed to have been stable until 2004, and then declined (M. Campbell, pers. comm.).  Radiocollared adult females have been monitored since 1993; though their precise ages are not known, all are at least two years old.  Neither pregnancy nor parturition rates are known, though calf survival to age one may be calculated from recruitment values estimated from pre-calving censuses.  Given the absence of age-specific estimates of vital rates, the model will be a female-only, two-stage (i.e., calves and adults) Leslie matrix.  Harvest mortalities are known for collared females, but other deaths cannot be definitively attributed to predation because of the amount of time that typically elapses between caribou mortality and collar retrieval.  Harvest mortality will be incorporated by removing a quota of caribou from the adult population at each time step. Demographic stochasticity will be incorporated by considering annual survival estimates as probabilistic values.  Environmental stochasticity (i.e., climate change) will be incorporated by assuming a coefficient of variation around each vital rate.  This model is under construction, and subsequent steps include mathematical formulation in a computer modelling program and specification of harvest/climate scenarios to project.  For caribou, increased freezing rain and insect harassment as well as changes in absolute forage availability that may result from climate warming and shifts in precipitation patterns will likely increase variability of survival and reproduction.  However, given the complex metabolic pathways by which these changes occur, I will take a phenomenological approach and increase the coefficient of variation around vital rates to reflect the greater variability in these values that frequently result from climate change.  Projection of population growth under such scenarios, while not representing reality per se, integrates the hypothesized influence of environmental conditions on survival and fecundity to anticipate how a population may behave in the future, based on its current state.

Results

Preliminary results to date: 

I found that female caribou gained body mass over the summer when the herd was increasing, neither gained nor lost significant mass when the population peaked, and lost mass when the population declined. Females consistently lost kidney fat over the summer during all phases, though losses were greatest at the peak phase. Back fat depth changed very little over the summer during the increase and peak phases, but declined when the population was in decline.  In contrast, females did not lose condition over the winter, recapitulating the findings of Thomas and Kiliaan (1998).  Notably, the fetal sex ratio changed throughout the study period; it did not differ significantly from 1 female:1 male during the increase and peak phases, but was significantly biased towards females during the decrease period.  This may reflect declining maternal condition, as heavier females may be more likely to conceive male calves.  Pregnancy rate was lowest during the increase phase, similar to the summer range-regulated George River herd, where decreasing pregnancy rates preceded a population decline. These preliminary results suggest that the herd's summer range was not sufficient for females to gain body condition, though it is not known whether this stems from absolute forage availability (i.e., overgrazing) or relative forage availability (i.e., insect harassment precluding uninterrupted feeding).

Applications

These models will be valuable tools for anticipating the effects of climate change on the population dynamics of caribou in Nunavut.

Contacts

L.S. Vors
Department of Biological Sciences
University of Alberta
Edmonton, AB. T6G 2E9
Canada
Email: vors'at'ualberta.ca