At-sea Density Monitoring of Marbled Murrelets in Central California: Methodological Considerations

We conducted at-sea line transect surveys for Marbled Murrelets (Brachyramphus marmoratus) to determine density off the coast of central California and to explore the utility of various survey protocols. Surveys were designed to compare line versus strip transect methods, and reveal the effects of distance from shore, viewing conditions and seasonal trends on density estimates. On consecutive days, we conducted 12 paired (24 total) at-sea line and strip transect surveys that were 20 km long at 400 m and 800 m from shore. We also performed nine surveys that were 10 km long and at distances of 400 m, 900 m, 1,400 m, 2,400 m, 3,400 m and 4,400 m from shore. Density estimates calculated using line transects were significantly greater than estimates based on strip transects of 100 m and 200 m widths. Marbled Murrelet density ranged from 2.4-39.4 birds km-* at 400 m from shore, and from 0.0-16.5 birds km-* at 800 m from shore. Density was higher on the 400 m than on the 800 m survey on 22 of 24 survey days. Densities measured on consecutive days were highly correlated on the 400 m transect but not on the 800 m transect. Line transect densities on the 400 m transect were higher when conducted under better viewing conditions. Line transects had higher statistical power to detect trends than strip transects. Statistical power analyses indicated only a 24% chance of detecting a population declining by 5% per year over 5 years when surveying line transects 20 km in length five times a year. Power to detect a 5% annual change increased to 57% when surveying line transects five times per year over 10 years. Survey design should strive to minimize variability in bird density in order to maximize likelihood of detecting population trends. An increase in the number of surveys per breeding season, length of a transect, or duration of monitoring effort should increase power to detect trends in murrelet density. We suggest that at-sea surveys should focus on detecting trends in density rather than population size.