Estimating Shorebird Numbers at Migration Stopover Sites

Abstract We describe a method for estimating the totalnumber of shorebirds that use a migration stopoversite during spring and fall migration. We combinedweekly shorebird counts with parameter estimatesfor detection probability, sampled proportion, andlength of stay on the Squaw Creek National WildlifeRefuge. Double sampling was used to determinedetection probability and estimated values variedamong wetland units from a low of 0.07 to a high of0.82. The sampled proportion of most wetland unitswas 100% but was lower in some of the largerunits. Length of stay (measured for Pectoral[Calidrismelanotos] and LeastSandpipers [C.minutilla] combined)averaged 10.0 days in spring and3.7 days in fall. Spring shorebird numberswere approximately five times greater than fallnumbers on the Refuge. Annual shorebird numbersvaried among years from an estimated low in 2003 of15 734 to a high in 2002 of 69 570.Peak daily counts during study years averaged only12% of estimated spring totals and 4%of fall totals. An estimate of shorebird numbersbased on summing weekly counts, not corrected fordetection probability or sampled proportion, wouldhave been only 21% (spring) to31% (fall) of the total number ofbirds. These results reveal that peak counts andnonadjusted counts can significantly underestimatethe number of shorebirds that use migrationstopover sites in the midcontinent of NorthAmerica.