Modeling biotic and abiotic influences on population size in small mammals

Most previous work in population ecology has modeled density-dependent effects in isolation. In this paper, we concurrently modeled the effect of density-dependent and density-independent factors on the rate of population change (R _t ) in Peromyscus leucopus (white-footed mouse), using a Ricker equation expanded to include weather and seasonality. From 1973 to 1996, we live-trapped P. leucopus monthly in a 2-ha Ohio woodlot. Population peaks (July to August) varied from 27 to 181 individuals, while troughs (December to March) varied from 4 to 46 individuals. We used time-delayed densities to act as surrogates for unobserved density-dependent factors, and principal components to represent 12 highly collinear weather variables. We identified time-delayed correlations by season between R _t and the independent variables (i.e., previous densities and weather principal components) using transfer function analysis. In summer, when P. leucopus densities were above the seasonal mean for the month, R _t was lower in the following 2 months; however, in winter, R _t was greater in the first but lower in the second succeeding month. R _t also correlated positively in autumn with contemporaneous precipitation, and was negatively correlated with `extreme' weather in summer with 2- and 3-month lags and in winter with a 3-month lag. We hypothesize that precipitation affected juveniles by influencing food resources and that `extreme' weather affected breeding. Our model explained 65% of the variability in R _t , and density-dependent and density-independent factors explained an equal percentage of that variability. This model created good forecasts of population density up to 12 months in the future.