Spatiotemporal heterogeneity in predator activity can influence the availability of refugia to prey. Prey are more able to exploit refugia through passive and active mechanisms if refugia persist over time. White-footed mice (Peromyscus leucopus) and eastern chipmunks (Tamias striatus) are abundant generalist rodents and large-scale removal experiments have confirmed them as important predators of gypsy moth (Lymantria dispar) pupae and songbird eggs. However, models indicate that small-scale (10s of m) variations in rodent activity may be important to persistence of their prey, and removal experiments at such small scales are rare. Our objectives were to quantify spatiotemporal structure and persistence of heterogeneity in mouse and chipmunk activity, and evaluate how small-scale (30 x 30 m “spots”) rodent removal impacts these characteristics. Track plates were used to measure rodent activity (a strong predictor of predation risk to pupae) on and around three control/experimental pairs of oak-dominated plots in upstate New York for 3, two-week sampling sessions each summer, 2008-2009. To manipulate persistence of refugia on experimental plots, rodents were removed from 8 spots each year, although only 4 spots received removal trapping in both years (persistent removal). Mean and variation of track activity in each plot and session were quantified by fitting a beta-binomial distribution to track data via maximum likelihood, and temporal persistence was quantified by disattenuated correlations.

Results/Conclusions

Over 36,000 plate checks were conducted each year. Overall track activity (2008 means: 0.036 – 0.12, 2009 means: 0.074 – 0.12) and coefficients of variation (2008 CVs: 1.38 – 2.85, 2009 CVs: 1.29 – 1.51) were similar between years. Between sampling periods, disattenuated correlation values ranged from -0.014 to 0.50 in 2008 and -0.12 to 0.34 in 2009, indicating low to moderate persistence of track activity. Geostatistical analysis of mouse and chipmunk activity indicated little spatial structure at the scales sampled (15-250 m), possibly due to low rodent densities in both summers. Preliminary analysis indicates that rodent removal may have counterintuitively increased local track activity of mice and chipmunks, but correlations between years were stronger for spots with persistent treatments (control or removal both years) than for spots with removal in only 1 year. These results suggest that sufficient spatial heterogeneity in small-mammal activity can provide persistent prey refugia, but the spatial scale may be quite small. The counterintuitive effect of rodent removal on track activity warrants further analysis and may indicate a disconnect between our measure of activity and local abundance.