POPULATION ESTIMATES AND THE TIMING OF WATERFOWL CENSUSES

The effect of the timing of waterfowl censuses on the estimates of bre eding population sizes of different species was studied in southeast F inland from 1991 to 1994. Lakes used only for breeding and a local sto p-over lake that was used for staging during spring migration were ana lysed separately. Also, the effect of the break-up of ice cover on the order of lake occupation by breeding pairs was studied. A standard wa terfowl census was made four times in May at an interval of approximat ely seven days in each year. In general, the timing of waterfowl censu ses within the three-week period considered did not cause serious bias es in population estimates, though differences between species and bet ween lake types were observed. The timing of the break-up of the ice c over was critical in the build-up of local breeding populations. Howev er, pairs of many species, especially the Mallard, Teal, and Common Go ldeneye occupied breeding lakes even when the lake still was considera bly covered by ice. A comparison of recommended census times between e arlier studies and this study suggests that in some species censuses c ould be started much earlier than previously recommended.