Recipient Organization
UNIVERSITY OF MONTANA
COLLEGE OF FORESTRY AND CONSERVATION
MISSOULA,MT 59812
Performing Department
College of Forestry and Conservation
Non Technical Summary
In the 1995-96 and 2004-05 winters, comprehensive surveys of the Amur Tiger (Panthera tigris altaica) was conducted across their known distribution in the Russian Far East. Since 1997, the even more endangered, and less widely distributed Far Eastern Leopard (Panthera pardus orientalis) has been surveyed 5 times. Details of the survey rationale and details can be found in Miquelle et al. (2006) and Miquelle and Murzin (2000). The goal of this analysis plan is to detail the development of a quantitative habitat selection model for the Amur Tiger and Far Eastern leopard to aid in the development of habitat management and protection guidelines, and to reveal spatial aspects of Amur Tiger and leopard ecology and conservation. For example, large scale habitat models have recently been developed for huge areas for woodland caribou conservation in the boreal forests of Canada, exceeding 160,000km2 and are providing the basis for forest management and energy development to aid the recovery of this endangered species? critical habitat. A quantitative and rigorous habitat model for Amur tigers would similarly enable the development of spatially explicit conservation plans. Large-scale viability modeling of Amur tigers has provided some insight into the potential impact of future management scenarios, but do not provide sufficient detail to guide management recommendations at a sufficiently small scale, and cannot be applied easily to define potential tiger habitat in nearby China, where tigers are beginning to recolonize. Similarly, although a ?rough? analysis of leopard habitat selection has been conducted, presently there exists a need for a quantitative model that can be used to predict suitable leopard habitat for a reintroduction program.
Animal Health Component
60%
Research Effort Categories
Basic
40%
Applied
60%
Developmental
(N/A)
Goals / Objectives
The objective of this analysis plan is to describe the research hypotheses and more importantly, the analysis approach and plan for developing large-scale quantitative models of Amur tiger habitat in the Russian Far East. We will examine the following hypotheses of tiger habitat using the general framework of Resource Selection Functions described above (RSF). First, we hypothesize that modeling tiger and leopard habitat as a function of prey species (red deer, sika deer, roe deer, moose, wild boar, and musk deer) will improve our ability to predict habitat quality for tigers and leopards compared to non-prey models including only physical and vegetation measures of habitat availability. Second, we hypothesize that analyzing habitat selection at two spatial scales will reveal differences in how tigers respond to human activity similar to other recent studies of carnivores. At large spatial scales, we predict that tiger resource selection will be negatively related to human activity, whereas at finer spatial scales, tigers may show variable selection for areas with human activity because (it is assumed) at these finer scales, tiger occurrence is already constrained to lower human activity landscapes. For example, at low human activity levels, tigers may actually use human features such as forest roads and trail systems, similar to gray wolves in North America [Thurber; Hebblewhite]. Finally, multi-scale resource selection functions will be combined to test the hypothesis that identification of critical habitat for tigers will be best achieved by considering tiger habitat as the product of the two spatial scales [Johnson 2006]. The predictive capacity , or utility, of the two scales of RSF models as well as the product of the two will be compared to out-of-sample data on tiger occurrence collected during the entire winter and not just during the intensive sampling sessions with which models will be developed. We expect that habitat selection for female tigers with cubs may be more tightly correlated to areas of high prey densities than for males or possibly even non-reproducing females. Therefore, we will compare resource selection functions derived from tracks of females with cubs compared to all other tiger tracks to attempt to more precisely define parameters associated with the capacity to successfully rear young. Even though prey availability and abundance may largely dictates carnivore distribution and abundance (Hypothesis 1 above), it is imperative to develop predictive capacity for large carnivores such as tigers and leopards without these parameters, since estimates of prey abundance are often absent despite the need to make management recommendations (e.g., define habitat for recovery of tigers in Northeast China, and suitable habitat for reintroduction of leopards in southern Sikhote-Alin). Therefore, it is necessary to determine how well models without estimates of prey occurrence (presumably surrogate measures of prey) can predict tiger and leopard occurrence.
Project Methods
We will adopt the framework of resource selection functions (RSF) to address habitat selection and identify high quality habitat for Amur tigers at two spatial scales dictated by the survey design implemented by Miquelle et al. (2006) across the entire range of Amur tigers in Primorski and Khabarovsski Krai (provinces) in the Russian Far East. Resource selection functions (RSF) compare the resources available in space at locations or sample units used by a species to those available at locations where tigers could have occurred (available locations) or locations where tigers did not occur (unused locations). Given the wide geographic distribution of species like tigers and leopards, ecologically, the resources that are most critical for tigers are therefore animals which constitute their major prey. For both tigers and leopards, primary prey items in winter (when surveys are conducted) are medium to large ungulates. This definition of habitat for carnivores, namely defining carnivore habitat in terms of their major prey items, has been underdeveloped in the carnivore literature. When developing RSF or habitat models for carnivores, most carnivore ecologists have used the typical approaches of substituting surrogates for prey habitat quality, such as elevation, forest cover type, or even remotely sensed measures of primary productivity in lieu of information of the spatial distribution of ungulates. In part this is because the latter is difficult to estimate reliably across large spatial scales. However, recent work on Amur tigers shows that at least at very coarse spatial scales, tigers select for areas of good red deer and sika deer habitat, as well as wild boar, and avoid areas of moose and other prey species in the Russian Far East. While the prey distribution and relative abundance data used in this analysis was very coarse and uncertain, this and other recent approaches modeling carnivore habitat as an explicit function of prey species suggests this will be a fruitful area of research for spatial predator-prey dynamics modeling as well as understanding important habitats for predators like Tigers and leopards.