METHODS OF ASSESSING BREEDING POPULATION SIZE 



A first requirement of most population studies is to obtain a precise 

 count or an estimate of the number of organisms in the study area. This 

 presents little difficulty when the organisms are large, sedentary, conspic- 

 uous, diurnal, and live in homogeneous, open habitats. However, such 

 conditions seldom hold for a given species. Waterbirds, in general, are 

 relatively large and conspicuous, and are primarily diurnal except for night 

 herons, petrels, and perhaps black skimmers. Their mobility while feeding and 

 their tendency to nest in yery dense, patchy vegetation (e.g. herons, egrets) 

 or in holes or crevices (petrels, guillemots), however, presents problems for 

 censusing. In addition, the large size of many waterbird breeding colonies 

 makes accurate estimating extremely difficult. 



The problems of locating and censusing waterbird colonies have been the 

 subject of a number of publications (for example, Belopol'skii 1957,Kartaschew 

 1963, Kadlec and Drury 1968a, b, Drury 1973, Nisbet 1973, Drury 1974, Nettle- 

 ship 1976, Buckley and Buckley 1976). In general, the most appropriate census 

 methods will be dictated by the size of the geographic region to be surveyed, 

 the species involved, the number of qualified personnel available, and funds 

 available. At one extreme, censusing only a few colonies of cryptic-nesting 

 species under conditions of limited support demands a ground count of nests; 

 while, at the other extreme, a regional census with substantial funding would 

 probably best be conducted using a combination of aerial censusing and photo- 

 graphy with nest counts. A point which receives recurrent emphasis from 

 experienced census workers is the need for replicability and standardization 

 of procedures and reporting. For best results, censusing should be conducted 

 (1) by the same individual (s) each year, (2) using the same, pre-tested tech- 

 niques each time, (3) at comparable times (of day and of nesting phase). This 

 should minimize the variance in the numbers estimated and allow more rigorous 

 statistical evaluation of the "true" variance, i.e., the variability due to 

 biological processes independent of sampling biases. 



Natural variation of adult numbers in a colony results from several 

 sources. Feeding conditions vary during the day depending upon tide cycle, 

 wind conditions, weather, etc., all of which influence the number of foraging 

 birds (Dunn 1973, Nisbet 1973, Custer 1977). The stage of nesting also 

 influences colony attendance. Incubation requires at least one adult at the 

 nest at all times but, after hatching, both members of the pair may need to 

 forage simultaneously to meet the demands of the brood. Other factors affect- 

 ing numbers at a colony are distances between nesting and feeding grounds, the 

 number of "surplus" (nonbreeding) adults in attendance (Kadlec and Drury 

 1968a), and disturbances by predators (Nisbet 1975, Robert and Ralph 1975). 



Because this study covered the entire northeastern U.S. coast, regional 

 (large-scale) inventorying using rotary- and fixed-wing aircraft was employed 

 in both 1976 and 1977. At selected colonies in each State, ground counts were 

 made to provide a basis for assessing the accuracy of aerial data. Because 

 conditions under which nest counts of the different species are taken vary so 

 greatly, the accuracy of small-scale, "within-colony" sampling was tested in 

 several colonies. These two levels of assessing census methods are treated 

 separately below. 



34 



