GIS coverages were created for ten mourning dove call-count routes in Mississippi for three time periods (1960’s, 1980’s, and 1990’s). Habitats were classified within 1.64 km. of routes. FRAGSTATS was used to spatially analyze landscapes. Metrics were then condensed using Principle Component Analysis. PC scores were regressed against call-count data to examine correlations. Akaike’s Information Criterion selected the model containing PC 1 (adjusted R2 = 0.148). PC 1 was summarized as simplicity/complexity measures. Patch density, edge density, Shannon’s diversity index, and % forest were negatively correlated with relative dove abundance. Total core area and % agriculture were positively correlated with relative dove abundance.
Dove variability was examined over a two-year period. Doves heard were variable in both summer and winter, while doves seen were variable in winter. The variability detected in summer could cause biased results from call-count surveys in some instances.