• Functional habitat heterogeneity and large herbivore seasonal habitat selection in Northern Botswana

      Fynn, Richard W.S.; Chase, Michael J.; Röder, Achim (2014)
      This study aimed to determine the functional seasonal attributes for herbivores of the major habitats and landscapes of the Savuti-Mababe-Linyanti ecosystem (SMLE) of northern Botswana and how various herbivore species responded to this heterogeneity. Floodplain grasslands and dambo grasslands provided the only significant green forage and biomass during the late dry season, whereas short grasslands of the Mababe Depression provided the highest forage quality of all habitats during the wet season....
    • Reconsidering habitat associations in the Anthropocene

      Hennessy, Sarah McCullough; Marczak, Susanne A.; Nordstrom, Lisa A.; Swaisgood, Ronald R. (2018)
      The California ground squirrel (Otospermophilus beecheyi) is generally undervalued despite serving as an ecosystem engineer in grassland ecosystems. Evidence of significant engineering effects by squirrels indicates that population reductions have cascading effects on other species, including several conservation-dependent species. While the theory and practices behind habitat association studies are already well established, our application of this approach helped identify priority management options in degraded grasslands expected to change further under shifts in climate. In this study we conducted surveys for California ground squirrels throughout San Diego County grasslands and examined habitat covariates to determine the ecological variables currently associated with occurrence. The primary objectives were to 1) improve our understanding of the habitat variables associated with squirrel presence, and 2) develop a predictive model for squirrel habitat suitability at a local scale. The most predictive models included significant main effects for percent sand (as a component of soil texture) and vegetation cover. A 10% increase in vegetation cover was associated with 1.3 fold lower odds of squirrel presence, whereas a 10% increase in percent sand was associated with 2.0 times higher odds of squirrel presence. Comparison of the predictive accuracy of soil texture data at two scales (fine-scale field vs. landscape scale GIS layers) showed fine-scale field sampling has greater predictive strength. Because soil type is a logistically non-malleable factor for wildlife managers, it is important to categorize management sites by soil type to identify the potential for promoting fossorial species on the landscape. With the prospect of shifting landscape ecotones due to climate change, it is as important to understand the basic habitat requirements of keystone species as for rare species.