Abstract
BACKGROUND: Lack of suitable analytical software and computational power constrains the comprehension of animal movement. In particular, we are aware of no tools allowing simulating spatially-explicit multistate Markovian movements constrained to linear features or conditioned by landscape heterogeneity, which hinders movement ecology research in linear/dendritic (e.g. river networks) and heterogeneous landscapes.SiMRiv is a novel, fast and intuitive R package we designed to fill such gap. It does so by allowing continuous-space mechanistic spatially-explicit simulation of multistate Markovian individual movements incorporating landscape bias on local behavior. RESULTS: We present SiMRiv and its main functionalities, illustrate its simulation capabilities and easy-of-use, and discuss its limitations and potential improvements. We further provide examples of use and a preliminary evaluation, using real and simulated data, of a parameter approximation experimental method. SiMRiv allowed us to generate increasingly complex movements of three theoretical species (aquatic, semiaquatic and terrestrial), showing the effects of input parameters and water-dependence on emerging movement patterns, and to parameterize a high-frequency simulation model from real, low-frequency movement (telemetry) data. Typical running times for conducting 1000 simulations with 10,000 steps each, of two-state movement trajectories in a river network, were of ca. 3 min in an Intel Core i7 CPU X990 @ 3.47 GHz. CONCLUSIONS: SiMRiv allows simulation of movements constrained to linear habitats or conditioned by landscape heterogeneity, therefore enhancing the application of movement ecology to linear/dendritic and heterogeneous landscapes. Importantly, the software is flexible enough to be used in linear, heterogeneous, as well as homogeneous landscapes. Using the same software, algorithm and approach, one can therefore use SiMRiv to study the movement of different organisms in a variety of landscapes, facilitating comparative research.SiMRiv balances ease and speed with high realism of the movement models obtainable, constituting a fast, powerful, yet intuitive tool, which should contribute exploring several movement-related questions. Its applications depart from the generation of mechanistic null movement models, up to population level (e.g. landscape connectivity) analyses, holding potential for all fields requiring the simulation of random trajectories.