Evaluation of survey methods for colonial waterbirds at Chase Lake National Wildlife Refuge, North Dakota

Abstract

Estimating the number of breeding pairs in a mixed-species waterbird colony is difficult because colonial waterbirds are vulnerable to human intrusion and their colonies are often in remote areas with limited access. We investigated methods to estimate the number of nests of waterbirds at a large, mixed-species colony at Chase Lake National Wildlife Refuge in south-central North Dakota. The primary goals of this study were to evaluate survey methods for shrub- and ground-nesting colonial waterbirds at Chase Lake National Wildlife Refuge and to develop protocols for estimating abundance of the different species. The specific objectives were (1) to assess visible-nest counts for ciconiiform species from the perimeter of nesting areas (hereafter, perimeter counts) and observational surveys from fixed points outside the colony to count flights of adult ciconiiforms in and out of the colony (hereafter, flightline surveys) as alternatives to within-colony counts of ciconiiform nests, and (2) to assess semiautomated, pixel-based image-analysis techniques to estimate abundance of American White Pelicans ( Pelecanus erythrorhynchos ) as an alternative to traditional manual counts from aerial photographs. For shrub-nesting ciconiiform species, observers counted 2,259 and 1,759 active ciconiiform nests in 2012 and 2013, respectively, during within-colony counts of ciconiiform nests. Results from within-colony counts of ciconiiform nests indicated a positive relation between the number of nests and the area of the shrub subcolony for the three most common ciconiiform species and all ciconiiform species combined. The perimeter nest counts of ciconiiform nests at Chase Lake represented only 18.8 percent of the total active ciconiiform nests counted in 11 subcolonies in 2012, which was well below the recommended target of 50 percent. Although we found a positive relationship between the number of nests counted during perimeter counts and the number of nests counted during within-colony counts for the three most common ciconiiform species and all ciconiiform species combined, perimeter counts at Chase Lake were hampered by disturbance to nesting birds. Thus, we discontinued the perimeter counts before they were completed. We did not develop predictive models from these perimeter counts in 2012 because these models could be misleading due to inconsistent application of the survey methods, which likely would have provided inaccurate perimeter counts. The extent of this issue is unknown. Flightline surveys at Chase Lake documented patterns of ciconiiform activity that were unknown for this region. For the common ciconiiform species, the number of flights to and from the South Island at Chase Lake were greatest in the morning (7:00−12:00 central daylight time [CDT]) and least in the afternoon (12:00−17:00), and least early in the breeding season (May 29–June 20, 2013) and greatest later in the breeding season (June 24–August 1, 2013). Flightline surveys are an index but lacked comparability with within-colony nest counts because the two methods provide measures of different things (that is, adult activity away from the colony as compared to the number of nests within the colony). The overall proportions of flights generally reflected the proportions of the within-colony nest counts for the four most common species: Black-crowned Night-Heron ( Nycticorax nycticorax ), Cattle Egret ( Bubulcus ibis ), Great Egret ( Ardea alba ), and Snowy Egret ( Egretta thula ). Flightline surveys at Chase Lake indicated apparent variation related to the time of day and season, as well as a variation in detection of inbound and outbound adult ciconiiforms. For ciconiiforms at Chase Lake, the most appropriate combination of survey approaches will depend on the need for annual estimates of nest abundance of ciconiiform species, balanced with the financial, personnel, and logistical constraints associated with the survey methods. For ground-nesting American White Pelicans, the results from this study indicated that digital-image processing using remote-sensing software provides an accurate estimate of the number of American White Pelican nests. Estimates of the number of pelican nests from digital-image processing, using two commercially available remote-sensing software packages, produced nest estimates that were comparable to those of traditional manual counts from aerial photographs.