| Summary: | Small rodents are some of the most important elements of boreal/arctic food webs, where they play essential functional roles. Their population dynamics are characterized by large amplitude multi-annual cycles regulated by direct and delayed density-dependence. These drastic variations in abundance have deep cascading effects into the whole ecosystem. Hence, the study boreal rodent population processes and drivers is important to understand/predict future states of northern ecosystems. To monitor animal populations, it is important to obtain reliable of estimates of abundance, which involves accounting for observation process errors. For small rodents, it is common to use the capture-recapture methodology, which collects information on both the number of observed animals and on their detectability, allowing to infer the number of non-observed individuals. Time series of abundance corrected for the observation process can then be used to model population processes of interest. Capture-recapture, although being optimal, is resource-intensive and limited to favorable field conditions, restricting the spatial and temporal resolution of the abundance data. This can be particularly limiting when studying populations of multivoltine rodents, with fast-changing population dynamics subject to strong effects of seasonality. New methods based on camera traps allow to increase spatio-temporal community-based data resolution. However, they require species-specific calibration studies. This thesis focuses on three specific research goals. (1) Develop a statistical framework to account for different sources of sampling error (i.e., capture heterogeneity) when estimating direct and delayed density-dependence in rodent population processes. In addition, assess estimation biases for different process parameters through a comprehensive simulation study. (2) Assess the adequacy of tunnel-based camera trap activity data as an index for abundance, calibrated against estimates obtained from capture-recapture in two different small ...
|
|---|