Understanding how the distribution and occurrence of organisms respond to habitat changes and disturbances are recurring themes in ecological research and conservation planning. In this presentation we discuss the application of a community model that estimates community summaries as a function of individual species detection and occupancy probabilities. We use this model to make inference about the impact of habitat disturbance caused by timber harvesting on the structure of breeding forest bird communities. We prove the utility of this technique to estimate community summaries (e.g., species richness) and its application in management and conservation planning by using bird count data collected under a robust design in managed forests and national parks in La Sierra Tarahumara region (STR) of northern Mexico. Specifically, we (1) estimate species occupancy and richness of the bird community, and (2) evaluate the influence of covariates such as residual basal area, average height, and stand structural diversity on our estimates.
Results/Conclusions
Our analysis yielded realistic estimates of site-specific and total species richness, and these were positively related to high stand structural diversity and residual basal area values. These patterns support our hypothesis that degradation caused by recent timber extraction reduces the size of bird communities, species composition and occupancy. However, we expected to find higher estimate uncertainty in sites with higher richness due to the high number of rare (i.e., infrequently detected) species found in such sites, and lower uncertainty in sites with low richness and few rare species. We note that in this particular situation, additional sampling will not reduce uncertainty in sites with a high number of species and sites with fewer species will require additional sampling efforts. Although our results can be readily incorporated into forest management plans to increase species occupancy and richness in the STR, this modeling framework can also be applied to other taxa and test hypotheses that address questions of community ecology and natural resource management. Finally, we also describe the importance of estimate uncertainty for sampling design and inference regarding community structure and dynamics.