A model approach for estimating colony size, trends, and habitat associations of burrow-nesting seabirds

We present a prototype monitoring strategy for estimating the density and number of occupied burrows of burrow-nesting seabirds. We use data and management questions from Washington State as an example that can be applied to burrow-nesting seabirds at single- or multi-island scales. We also demonstrate how habitat assessments can be conducted concurrently. Specifically, we compared the density and occupancy of burrows of the Rhinoceros Auklet (Cerorhinca monocerata) at nesting colonies in the California Current and the Salish Sea and in the 1970s, 1980s, and today. We estimated 36 152, 1546, and 6494 occupied burrows on Protection and Smith islands (Salish Sea), and Destruction Island (California Current), respectively. Our estimates for the Salish Sea are 52% greater than those from the 1970s and 1980s, while that for the California Current is 60% less than that of 1975. This suggests that the Salish Sea population has increased, despite greater human effects on that ecosystem. However, some of the estimated changes between the periods could be the result of methodological and analytical differences. To address these issues we recommend an unbiased and representative sampling approach (stratified random) and an approach for optimally allocating the samples among strata within and among islands, depending on the scale of the question being addressed. Optimally allocating the sample would save a great deal of field effort; using this approach, we achieve relatively high power (\u3e0.80) to detect moderate changes (20%) sampling hundreds of fewer plots than in a sample not optimally allocated