USGS Home
Contact USGS
Search USGS

Some investigators (e.g., Stewart and Kantrud 1973) used a combination of the rules set forth in those papers. Sauder et al. (1971) used only lone males and pairs to calculate estimated breeding pairs. Like Dzubin, they concluded that several counts were necessary to increase reliability of breeding pair estimates. The guidelines set forth by Hammond were used originally in operational surveys of refuge lands in the United States. Recently, local managers have tended to modify the procedures because of unique local problems or insufficient funds. The outcome of these minor changes has been that results among studies, areas, or years are not strictly comparable.

A number of other biologists used Dzubin's method of estimating breeding populations of ducks, especially diving ducks, from a count of nests. In a study of reproduction of Black Ducks and Eiders on the St. Lawrence River (Milne and Reed 1974, Reed 1975), the number of nests found in a complete search of the study area was used to estimate the breeding population. Isakov (1963) used nest counts to estimate the number of breeding ducks on islands in the Black Sea. Nest searches are generally too time-consuming for operational surveys. Sugden and Butler (1980) considered nest counts for diving ducks impractical and recommended using pairs plus lone males to estimate breeding Canvasbacks and total females to estimate breeding pairs of Redheads.

The assumptions underlying use of nests as an estimator of breeding population present additional problems. It is necessary to assume that all nests are found, and some hens have more than one nest. In addition, if only a portion or even none of the population actually nests, the breeding population as we have defined it would be severely underestimated. The method would require that breeding population be defined as hens that have at least one nest, and a correction for renesting would be required.

Use of lone males to indicate the presence of a breeding female (assumed to be on a nest) can introduce a serious bias in populations with a preponderance of males, as is typically the case with diving ducks (Dzubin 1969a, Erskine 1971). Dzubin recommended correcting population counts for unbalanced sex ratios, but estimates of sex ratio require detailed field studies that may not be available for operational surveys.

Definition of breeding population for a local area requires that transient and breeding birds be distinguished. Smith and Hawkins (1948) arbitrarily divided total birds counted by 2.5 to account for the presence of nonresident birds on transects. Ducks also move among ponds during the day and may be found on one class of pond during the middle of the day and on other classes during the morning and evening (Klett and Kirsch 1976, Longcore and Ringelman 1980). This behavior can cause biases because visibility varies among pond classes. Individual hens change social status as the season progresses. Breeding birds can be separated from migrants to some extent by timing of the census. Counts made too early or too late will be seriously biased because they may contain a high proportion of migrant birds, or birds that have already attempted to breed may have left the local area. In addition, the proper time for conducting pair surveys varies among species, years, and locations. Dzubin (1969a:219) treated this subject in detail and made recommendations based on extensive field data. Kauppinen (1983) investigated many of the same factors studied by Dzubin and made recommendations for the best timing of census periods for ducks in Finland. Dzubin set two counts as a minimum but recommended an average of four or five counts. Where only trends are required, Hammond (1969) suggested that a single count would suffice. Single counts of breeding ducks are frequently used, however, to estimate the breeding population, a procedure that can lead to highly biased estimates.

Timing of census presents less of a problem for geese because most species nest in colonies in the northernmost part of their range, and breeding birds are not confused with migrants. Malecki et al. (1981) planned their survey of Canada Geese in northern Manitoba for late in the nesting cycle to minimize inclusion of subadults. King (1982) recommended that swans in Alaska be counted between mid-May and June when most of the females are incubating and can be seen from the air, or in August when young are large enough to be counted from the air.