Waterfowl were counted in a sample of wetlands located along the Yukon road system. These wetlands are distributed along the Alaska Highway between Watson Lake and Beaver Creek, the south Klondike Highway between Whitehorse and Carcross, the North Klondike Highway between Whitehorse and Carmacks, and the Aishihik Road (Table 1, Figure 1). Each wetland was surveyed five times at weekly intervals beginning in early May to cover the range of breeding dates for the various species of waterfowl. Each wetland was censused from one to four vantage points with the aid of binoculars and tripod-mounted telescope. In some cases it was impossible to census the entire wetland, but the same portion of the wetland was surveyed each time. At each wetland, the observer recorded the species composition, age, and sex of all waterbirds in as detailed a fashion as possible, as indicated on the data sheet (Appendix 1). Ice cover, if any, was noted as well.

Data were keypunched into the SAS system (SAS Institute 1985) and all computer analyses were done using SAS. Four measures of waterfowl abundance were used for each species:

Estimates of the above parameters were generated by calculating the mean and standard error over the five replicate surveys. From the standard error, 95% percent confidence limits were calculated by multiplying by t_(.05,n-1) where n=5, and t=2.776. Paired t-tests (SAS Institute 1988) were employed to test for differences between the numbers of birds seen on the 5 surveys in 1997 vs. 1998.

To determine trends from 1991 to 1998, we calculated the percentage change from 1991 to 1992 using the waterbodies surveyed in common between both years, then did the same for 1992-93, etc., using the waterbodies common to those pairs of years. The 1991 population index, being the starting point, was assigned an arbitrary value of 100 percent.