The International Waterbird Census (IWC) is one of the longest-running bird monitoring schemes in the Western Palearctic. Most of its data are collected with land-based counts, the reliability of which is largely unknown. This study compared estimates of land-based vs. aerial counts, and the relative conservation values of coastal sites obtained with the two methods. The data were collected in the West-Estonian archipelago of the Baltic Sea in 1993, and analyzed at two spatial scales (mean area of plots 9 and 36 km²). Among nine waterfowl species, land-based and aerial census provided closely correlated local population estimates for the Mute Swan (Cygnus olor), Steller’s Eider (Polysticta stelleri), Long-tailed Duck (Clangula hyemalis) and, at the larger scale, for the Mallard (Anas platyrhynchos). However, small numbers of all species, except the swan, remained undetected with the land-based census, and numbers of Mallard and Steller’s Eider were also systematically underestimated. The areas having the highest conservation value were reliably identified with the land-based census, particularly at the larger scale. Hence, land-based studies are in general accordance with the aims of the IWC, but the absolute population estimates should be interpreted with care.