ABSTRACT. Aboriginal harvests of barren-ground caribou (Rangifer tarandus) potentially offer a large sample for assessing body condition. The purpose of this study was to determine the probability that a certain amount of fat would be at designated anatomical sites when Dënesóline hunters qualitatively report the condition of an animal. Hunters' impressions were used to evaluate the condition of adult female barren-ground caribou (R. t. groenlandicus) in late winter. A semi-objective body condition index (BCI) was developed using fat indices described by biologists as useful for measuring caribou body condition. Fat deposits from 217 adult female caribou harvested in late winter (41 in 2000 and 176 in 2001) were used to calibrate hunters' impressions with BCI estimates. Variation in hunters' impressions and BCI estimates indicated that adult female caribou were fatter in 2000 than in 2001. Multinomial log-linear models indicated that hunters' impressions were related to each of the variables that make up BCI. The probability of pregnancy was significantly related to both BCI and hunters' impressions in 2000 and in 2001. Both models indicated that fat adult female caribou had a greater probability of being pregnant than thin cows. Monitoring of barren-ground caribou body condition provides common ground for northern aboriginal communities and government biologists to collaboratively manage a wildlife resource.

Key words: Dënesóline, caribou, body condition, hunters' impressions, pregnancy, herd well-being

RÉSUMÉ. Les prélèvements du caribou des toundras (Rangifer tarandus) par les Autochtones pourraient offrir un vaste échantillon permettant d'évaluer l'état corporel. Cette étude avait pour but de déterminer la probabilité qu'une certaine quantité de gras soit présente dans des zones anatomiques désignées, au moment où les chasseurs Dënesóline font un rapport qualitatif sur la condition physique d'un animal. On a eu recours aux opinions des chasseurs pour évaluer l'état physique du caribou femelle des toundras à l'âge adulte (R. t. groenlandicus) à la fin de l'hiver. On a créé un index semi-objectif de l'état corporel (IEC) en utilisant les indices de gras que les biologistes ont trouvés utiles pour mesurer l'état physique du caribou. Les dépôts de gras provenant de 217 caribous femelles adultes prélevés à la fin de l'hiver (41 en 2000 et 176 en 2001) ont servi à étalonner l'opinion des chasseurs par rapport à 1'IEC estimé. Les variations entre l'opinion des chasseurs et 1'IEC estimé ont montré que le caribou femelle adulte était plus gras en 2000 qu'en 2001. Des modèles log-linéaires multinomiaux ont révélé que l'opinion des chasseurs était reliée à chacune des variables composant 1'IEC. En 2000 et 2001, la probabilité que la femelle soit en gestation était fortement corrélée à la fois à 1'IEC et à l'opinion des chasseurs. Les deux modèles ont révélé que la probabilité que la femelle caribou adulte soi t gravide était beaucoup plus grande pour les femelles grasses que pour les maigres. Le suivi sur la condition physique du caribou des toundras constitue une plate-forme commune propice à une gestion de la faune menée en collaboration par les communautés autochtones du Nord et les biologistes du gouvernement.

Mots clés: Dënesóline, caribou, condition physique, opinion des chasseurs, état de gestation, niveau de santé de la harde

Traduit pour la revue Arctic par Nésida Loyer.

INTRODUCTION

Community-based monitoring programs are becoming increasingly common in wildlife and environmental management. Biologists and managers representing government and industry are recognizing the value and long-term viability of involving indigenous user groups. A major advantage of community-based monitoring programs is their ability to combine user traditional knowledge with Western science-based information. This potential contribution to management is well recognized in northern aboriginal communities like Lútsël K'é that harvest Bathurst barren-ground caribou, Rangifer tarandus groenlandicus (Parlee, 1998).

Barren-ground caribou are one of the most valuable renewable resources in Canada's Northwest Territories (NT). They have immense cultural, nutritional, and biological value, and consequently the caribou harvest is of great importance to aboriginal hunter-gatherer groups in this area (Smith, 1978; Berkes, 1998). Effective caribou management should be guided by ecological monitoring that includes aboriginal knowledge.

Both aboriginal people and scientists recognize caribou body condition as an important indicator of general herd condition (Kofinas et al., 2003). Body condition directly affects mortality, pregnancy, calf survival, age to first breeding, and breeding pauses in adult female caribou (Thomas, 1982;Cameron, 1994; Gerhart et al., 1996,1997; Thomas and Kiliaan, 1998b; Allaye Chan-McLeod et al., 1999). Changes in caribou body condition may reflect environmental limitations to herd growth.

A community-based caribou monitoring system required body condition indices that would not be an imposition on hunters and would use their existing techniques to assess caribou condition. We did not want hunters to simply report on the status of fat at one or more anatomical sites because that was not customary. Rather, we wanted to ask hunters to use the method that they would normally use.

Our research objective was to calibrate hunters' qualitative impressions of harvested adult (two or more years old) female caribou condition with a semi-objective body condition index (BCI), and in turn determine the relationship of each to pregnancy rates. This required developing a BCI from fat indices similar to those used in other caribou monitoring programs (Kofinas et al., 2001 ). It was thought that observer variation would be reduced if BCI parameters were used, especially if in the future a large number of observers from different communities were involved. Aboriginal hunters use a number of characteristics (primarily the fat indices measured in this study) to form an overall impression of caribou body condition. Therefore, the probability that a certain amount of fat would be at designated anatomical sites when hunters refer to an adult female caribou as being in a particular condition was determined. Variation between the two study years in body condition predictions and pregnancy probabilities was also assessed.

METHODS

Assessment of Caribou Body Condition

The study was conducted from the Dënesóline (Chipewyan) community of Lútsël K'é (62°24' N, 110°48' W), located in the east arm of Great Slave Lake, NT, Canada (Fig. 1). Three observers accompanied 36 Lútsël K'é hunters on hunting forays for caribou between mid February and late April in 2000 and 2001. An Inuit hunter from Baker Lake, two Yellowknife hunters from Dettah, and two Dogrib hunters from Fort Rae were also surveyed during this period in 2001, as they were hunting in the Lútsël K'é area.

The selection of each technique to measure body condition was based on ease of data collection, reliability of measure, ease of analysis, acceptability of the technique to local people, and cost of sampling. Two indices, hunters' impressions and a body condition index (BCI), were used to evaluate the body condition of each caribou. After preliminary discussions with hunters, a qualitative index was developed using their impressions of caribou body condition (Table 1). Hunters provided observers with an evaluation of body condition after dressing out each caribou in the field. The second index used was semi-objective. Scores were given to five body-condition categories and then summed (Table 1). BCI ratings for each caribou could range from 5 to 17. Observers measured these fat indices as the hunter dressed out each animal. Brisket and back fat depths were recorded in inches because this provides easy-to-judge measurements and conforms to the protocol used in other hunter body-condition studies conducted around North America (Kofinas et al., 2003; Table 1). Back fat depth was measured along the dorsal line about 5 cm forward from the base of the tail (Gerhart et al., 1996). Kidney and stomach fat were the amount of fat judged to be covering each of these organs (Table 1 ). The pregnancy status of each adult female caribou harvested was also recorded.

Analysis of Body Condition Indices

Multinomial log-linear models (using the "multinom" function in the neural network S-Plus library, Venables and Ripley, 1997) were used to evaluate the relationship between hunters' impressions and individual body condition indices. For each body condition index (stomach, kidney, back, brisket, and marrow fat), four models were fitted and compared:

1. Body condition = impression + year + impression-year (i.e., the hunter's impression of an animal with the same body condition rating differs between years);