Spatial patterns and drivers of invasive rodent dynamics in New Zealand forests

Abstract

Populations of invasive ship rat (Rattus rattus) and house mouse (Mus musculus) vary greatly over both time and space in New Zealand’s indigenous forests. Both species threaten endemic fauna, and their spatiotemporal variation over short time scales and steep environmental gradients present substantial challenges for conservation management. We fitted models to 18 years of 3-monthly records from a large-scale tracking-tunnel network to (1) predict, classify and describe forest-wide spatial patterns of temporal dynamics of unmanaged rat and mouse populations, and (2) understand the population-limiting roles of environment and biotic interactions. We distinguish six classes of forest rodent dynamics by classifying deciles of predicted rat-tracking rates. Classes form sequences across broad latitudinal and elevation gradients, and grade from ‘irruptive’, with low median but synchronous high maximum rat and mouse tracking rates in colder forests, to ‘continuously ratty’, with high median rat and low unsynchronised mouse tracking rates in warmer forests. Mice irrupt alone more frequently in colder forests, and their tracking rates are spatially reciprocal with rat tracking rate minima but not maxima. We conclude that predictable spatial patterns of rodent population dynamics arise from stronger low-temperature limitations on rats than mice, biotic limitation of mice by rats, and spatiotemporal food resource patterns affecting both species. Practical implications are that native species conservation in New Zealand forests requires spatially-differentiated predator management regimes; ship rats are likely to become increasingly prevalent at higher elevations as climate warms; and suppression of ship rats alone will release house mouse populations, especially in warmer forests.