Modeling population variation in Peromyscus leucopus: An exploratory analysis

Time-series analysis was applied in an exploratory analysis of density estimates for Peromyscus leucopus obtained over 9 years by weekly nest-box checks supplemented by live trapping. We developed a model that quantified the relationship between population density in a given month and density 1 month prior (1st-order autocorrelation), 12 months prior (seasonal effects), and 36 months prior (multiannual effects). The remaining variation was examined to determine the influence of stochastic variations in parameters of weather. In order of importance, predictors of density in a given month were density in the previous month (60.4% of variance explained), seasonality (6.1%), and multiannual effects (7.0%). Of the remaining 26.5% of variance, 3.4% was due to two weather factors; deviation from mean precipitation with a 5-month lag and high temperature with a 2-month lag. The population of P. leucopus was relatively resilient to abiotic effects over the 9 years. Quantitative models of this type are rare in the literature because gathering and adequately analyzing long-term ecological data is a demanding task. However, quantitative descriptions of variations in population density are essential to assessment of the relative importance of various mechanisms contributing to population regulation.