An intuitive method to calculate the utilization distribution of an animal from step‐selection analysis

Abstract Step‐selection analysis (SSA) is a popular tool for estimating resource/habitat selection conditional on the local availability of conditions as determined by an animal's movement capacity. These models can be subsequently used to parameterize a movement model; however, most SSAs focus instead merely on interpreting the direction and magnitude of model coefficients. This is at least partly due to perceived complexities of generating such models through simulation or Partial Differential Equation (PDE)‐based approaches, suggesting a need for alternative methods that are intuitive, approachable, and ideally, already familiar to ecological application. We advocate here for the full predictive functionality of SSA and provide methods to use a fitted step selection function to predict an animal's space use. Specifically, we use fundamental mathematical principles to construct a transition matrix from the step selection movement process. We demonstrate that transition matrices can efficiently produce the emergent spatial pattern from SSA using two different methods (eigen decomposition and iterations). This alleviates the need for thousands of simulations of the underlying model while leveraging a well‐established tool in applied ecology. We show how such matrices may be parameterized under competing definitions of the underlying movement kernel and link these back to realistic ecological processes. To emphasize the flexibility of this approach, we apply these transition matrices to field data from two species with markedly different movement capacities and strategies: wild pigs ( Sus scrofa ) and long‐tailed tits ( Aegithalos caudatus ). We discuss the caveats in transition matrix parameterization and resulting predictions of space use with respect to the underlying movement process and provide suggestions for application of this procedure to SSA. Our framework represents a powerful tool for space‐use prediction from SSA while also highlighting critical gaps in our current knowledge about and tools to describe animal space‐use processes. In particular, we discuss higher order Markovian processes and spatiotemporally dynamic variables as particularly important topics for further development.