Estimating seabird abundance: a case study in Kenai Fjords National Park, Alaska

Abstract

Estimation of breeding seabird population size and trends are integral components of studies or programs seeking to understand how seabird populations respond to changes in marine or coastal environments, to identify threatened or declining species, and to inform management actions and decisions. In Chapter 1, I conduct a review of the challenges, considerations, tools, and methods involved in efforts to estimate and monitor breeding seabird abundance. I discuss challenges in terms of two broad categories: 1) seabird life history, behavior, and breeding environments, and 2) challenges inherent to survey methods and logistics. I introduce methods and tools used to access seabird colonies, detect birds, and design methods to collect and analyze count or abundance data. The focus of Chapter 2 is to find effective methods to estimate the breeding abundance of glaucous-winged gulls (Larus glaucescens) in Kenai Fjords National Park (KEFJ), Alaska, which has been designated as an Important Bird Area (IBA) for this species. There are numerous inherent challenges in this effort, as L. glaucescens breeds in widespread colonies on vertical cliff faces of the fjords and associated islands, and their nests are not visually detectable from boat-based surveys. I conducted and compared field counts to replicated photographic counts, and found enough variability between replicates for both count methods to preclude calculation of precise abundance estimates using counts alone. I then developed a more intensive method of analyzing images using a modified mark-resight (MR) approach to identify all potential nest locations, and I took advantage of both attendance and behavioral data collected from repeat photographs to estimate what proportion of them have a high probability of containing nests. I quantified two potential survey error rates and their effects on the results of our modified MR approach. Finally, I considered temporal and environmental factors likely to affect both repeated counts and the results of my modified MR approach. I found that: 1) the modified MR approach provided a better approximation of breeding abundance than simple field counts and addressed variability between replicate surveys; 2) low misidentification survey error rates had a negligible effect on the results; and 3) general patterns of attendance of birds at colonies were influenced by different factors than the attendance patterns at locations that were likely nests. I recommend similar methods for other colonial or cliff-nesting bird species, species that have variable attendance, or species that make nests that are hard to see. These methods may also be helpful in areas that are remote or infrequently visited or where time in the field is a limiting factor in how much data can be collected.