Wildlife Survival Estimation: Recent Advances and Future Directions

Abstract

Survival is a major determinant of animal population size and trend, and survival estimation has received considerable attention from population and conservation biologists seeking to understand patterns of abundance and numerical change in ecological systems. Although researchers commonly consider survival to be a population parameter, it is an attribute of individual animals. Because genetic, demographic, ecological, and environmental factors act upon individuals to influence survival, the objective of modern survival analysis is to quantify the role of such factors on mortality risk. In recent years, survival has been a prevalent topic of study in the Journal of Wildlife Management compared to other demographic attributes like productivity and dispersal (Fig. 1). However, survival tends to be more difficult to estimate properly than other demographic parameters (McCallum 2000, Fox 2001), especially among free-ranging animals where cryptic behavior patterns and difficulties in marking and monitoring often lead to small sample sizes and a high proportion of subjects succumbing to an unknown fate. These problems can limit the amount and quality of survival data obtained from ecological studies, and the challenge for researchers is to apply robust methodology that will make effective use of survival information obtained under sometimes constraining field conditions. Survival estimation has its early roots in human medicine and epidemiology (Oakes 2001), but in recent decades, considerable effort has focused on the development of reliable survival methods for application with free-ranging animals (Brownie et al. 1985, Burnham et al. 1987, Pollock et al. 1990, Lereton et al. 1992). In the last 20 years, substantial improvements have been made in this arena, both in terms of capture, marking, and monitoring techniques facilitating comprehensive survival assessment for free-ranging animals (Krebs 1999, Kenward 2001, Millspaugh and Marzluff 2001), and in the development of software programs allowing for the analysis of complex survival functions (Williams et al. 2002). Together, these changes have elicited a flurry of innovative survival research in animal ecology (Lebreton et al. 1992, Cam et al. 2002, DelGiudice et al. 2002, Dinsmore et al. 2002). These recent advances notwithstanding, there remains significant interest in the current state of survival estimation in wildlife research, particularly in terms of basic assumptions that may be difficult to meet in many research studies, or in the application of specialized analytical techniques requiring particular specification (Krebs 1999, Williams et al. 2002, Zens and Peart 2003, Rotella et al. 2004). Thus, a general review of current methodology and new developments in wildlife survival estimation should be timely and beneficial.