Evaluations of Duck Habitat and Estimation of Duck Population Sizes with a Remote-Sensing-Based System

Duck Populations

We restricted our analysis of annual change in ponds, duck population sizes, and duck production to 10 waterfowl management districts (Fig. 1) for which we had aerial video data during 1987-90. Although video was obtained of other waterfowl management districts as they were added to the remote-sensing-based system, the lack of data in some years would have confounded comparisons among years.

Regression Estimates

Estimates of breeding-population sizes were derived from a regression model. Data for constructing baseline regressions for all species except wood ducks (Aix sponsa) were gathered at the Arrowwood Waterfowl Management District (1982-84). For wood ducks, we used data gathered at the Fergus Falls (1986-87) and Detroit Lakes (1987) waterfowl management districts. Baseline regressions were constructed by regressing the number of pairs observed on each pond on the size and the square root of the size of the pond. The general form of the equations was:

Wet area in this equation is the wet area in each pond. (Estimates of the regression coefficients A and B are presented in Table 3, and the regression curves are presented on Fig. 2).

Ground Counts

The boundary of each wetland basin where breeding-pair counts were conducted was delineated on aerial photographs, and a unique identification number for the naming of the polygon of that basin was placed on the photograph.

Table 3. Baseline regression coefficients used for estimating breeding duck (Anatinae) populations on 10.4-km² plots from the area (ha) of individual ponds determined from airborne video in the prairie pothole region of the United States, 1987-90.

	Regression coefficients^a
Species	A	B
Mallard	0.0106	0.2899
Gadwall	0.0341	0.2848
American Wigeon	0.0000	0.0193
Green-winged Teal	0.0000	0.0431
Blue-winged Teal	0.0000	0.7376
Northern Shoveler	0.0136	0.1870
Northern Pintail	0.0000	0.1866
Redhead	0.0410	0.2233
Canvasback	0.0064	0.0453
Lesser Scaup	0.0173	0.0528
Ring-necked Duck	0.0022	0.0110
Ruddy Duck	0.0119	0.1226
Wood Duck	0.0064	0.5612

^aPairs = A x wet area + B x wet area.

The pair count was conducted by a person who walked around each pond. Pairs on lakes and rivers were sometimes surveyed from a boat or from a canoe. When dense emergent vegetation was present, the observer moved through the vegetation and flushed the birds. Sample ponds were more widely dispersed in the remote-sensing-based system than in surveys that are often conducted on plots or transects. We assumed that sample ponds were usually far enough apart to prevent the flushing of birds from one sample pond to another. Social groups were recorded on field forms designed for data entry into a microcomputer. Data were entered soon after completion of the counts. Error checking was done by a series of custom-designed programs that were executed at the time of data entry and immediately after data entry. The size of a breeding population was expressed as the estimated number of breeding pairs, which included observed breeding pairs and pairs that were inferred from other social groups. We followed approximately the same techniques described by Hammond (1969) and Dzubin (1969). The estimated number of breeding pairs was estimated by the computer at the time of data entry as follows: 1 observed pair = 1 estimated breeding pair; 1 lone male = 1 estimated breeding pair; 1 lone female diving duck = 1 estimated breeding pair; each flocked, male dabbling duck in flocks of five or fewer individuals = 1 estimated breeding pair except in wigeons and in northern shovelers for which the number of flocked males was not converted to estimated number of breeding pairs; each flocked female diving duck in flocks of five or fewer individuals = 1 estimated breeding pair. Flocked birds in flocks with more than five individuals were not used to estimate the number of breeding pairs.

Previous Section -- Habitat Data
Return to Contents
Next Section -- Data Analyses