Estimating occupancy dynamics for large-scale monitoring networks: amphibian breeding occupancy across protected areas in the northeast United States.

摘要: Regional monitoring strategies frequently employ a nested sampling design where finite set of study areas from throughout region are selected and intensive occurs within subset sites the individual areas. This protocol naturally lends itself to hierarchical analysis account for dependence among subsamples. Implementing such an using classic likelihood framework is computationally challenging when accounting detection errors in species occurrence models. Bayesian methods offer alternative approach fitting models that readily allows spatial structure be incorporated. We demonstrate general estimating occupancy data come design. analyzed regional program wood frogs (Lithobates sylvaticus) spotted salamanders (Ambystoma maculatum) vernal pools static dynamic observations 2004 2013 were collected 14 protected located northeast United States. use estimate trends at both area levels. show level was relatively stable species. However, substantial variation occurred areas, with some populations declining increasing In addition, When not accounted for, one would conclude stronger support latitudinal gradient than our accounts contrast model does nesting, did include effect latitude 95% credible interval. These results shed light on range-level population status these pond-breeding amphibians, provides can used examine drivers local dynamics.