A carregar...
Publicação
SiMRiv: an R package for mechanistic simulation of individual, spatially-explicit multistate movements in rivers, heterogenous and homogenous spaces incorporating landscape bias
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.32 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
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
Descrição
Software article
Palavras-chave
connectivity dendritic ecological networks landscape heterogeneity linear habitats hidden Markov models mechanistic movement models individual-based movement simulation movement ecology river networks resistance
Contexto Educativo
Citação
Quaglietta and Porto Movement Ecology (2019) 7:11
Editora
BMC