Home‐range spillover in habitats with impassable boundaries: Causes, biases and corrections using autocorrelated kernel density estimation

Abstract An animal's home‐range plays a fundamental role in determining its resource use and overlap with conspecifics, competitors and predators, and is therefore a common focus of movement ecology studies. Autocorrelated kernel density estimation addresses many of the shortcomings of traditional home‐range estimators when animal tracking data are autocorrelated, but other challenges in home‐range estimation remain. One such issue is known as ‘spillover bias’, in which home‐range estimates do not respect impassable movement boundaries (e.g. shorelines and fences), and occurs in all forms of kernel density estimation. While several approaches to addressing spillover bias are used when estimating home ranges, these approaches introduce bias throughout the remaining home‐range area, depending on the amount of spillover removed, or are otherwise inaccessible to most ecologists. Here, we introduce local corrections to home‐range kernels to mitigate spillover bias in (autocorrelated) kernel density estimation in the continuous time movement model (ctmm) package, and demonstrate their performance using simulations with known home‐range extents and distributions, and a real‐world case study. Simulation results showed that local corrections minimized bias in bounded home‐range area estimates, and resulted in more accurate distributions when compared with commonly used post hoc corrections, particularly at small–intermediate sample sizes. Comparison of the impacts of local vs. post hoc corrections to bounded home‐ranges estimated from lake trout ( Salvelinus namaycush ) demonstrated that local corrections constrained the redistribution of probability mass within the remaining home‐range area, resulting in proportionally smaller home‐range areas compared with when post hoc corrections are used.