Data Analysis

Waterbird counts were examined by species for each area, as lake totals for each year of the survey, and a combined five-year summary. Five-year species means were calculated by first averaging counts from all years for each survey period. Next, an overall mean for each species was computed by averaging the 17 survey period means. The same process was applied to specific survey periods of interest for each species to arrive at a more accurate estimate of population size during periods of peak occurrence. For example, counts of Wilson’s phalaropes from all survey sites were totaled for each survey period for each of the five years. Yearly totals for survey period 1 were averaged. This was repeated for the remaining 16 survey periods resulting in average numbers of phalaropes by survey period through the season. To calculate an overall average of Wilson’s phalaropes, the 17 survey period means were averaged together. Also, selected periods of phalarope presence were averaged to get an estimate of the species’ peak occurrence at GSL. Species distribution maps illustrate mean counts over survey periods when the species are present at Great Salt Lake. Means for suites of species were also calculated.  Suites included unidentified groups that were not assigned to any species totals. For example, the DUCKSX suite includes all duck species and the “unidentified duck” (DUCK) category that cannot be assigned to any one species. Unidentified numbers are considerable in many cases and should not be overlooked. Peak numbers reported are the largest 5-year period mean for a particular species or suite.

An important consideration during the five-year survey was the fluctuation of lake elevation and its affect on habitat. For analytical purposes we determined to evaluate habitat changes and their subsequent species use for the years of lowest and highest lake elevation during the years of the study. The highest lake elevation year was 1999. Two years, 1997 and 2001, were both years of low lake elevation. The 2001 survey season was chosen to be the representative low lake year because the data set was more complete than that of 1997. To provide an assessment of the length of time individual bird populations occur within the ecosystem, bird use days were estimated from the data set. A bird day is defined as one bird spending 24 hours within the study area during the study period. These figures were computed by multiplying the mean number of birds by the number of survey days.  For 1998-2001, the study period each year was 170 days, April through September. The field season was considerably shorter in 1997 and so the mean bird numbers was potentially inflated by as much as 25%. For a more accurate comparison of bird use days between years, data from 1997 were omitted.