Modelling and mapping habitat suitability of species in a given study area has become a popular method to assess biogeographical questions on the habitat requirements of species, and to study the effects of land use or climate change on species distributions. Beside purely academic aims, model projections are increasingly used for devising conservation and management strategies and they are an established tool in conservation planning. Species distribution models (SDMs) link geographic presence records of species with environmental conditions found at these locations, and extrapolate the occurrence probabilities of species throughout the study area. While the modelling framework is similar in the terrestrial, marine and freshwater realms, each realm comprises specific challenges for combining the spatial scale, the environmental data and the species records for building reliable models. In this paper, we focus on these key issues from a stream ecology perspective and highlight three critical challenges to be kept in mind when building SDMs in stream ecosystems - (1) the spatial configuration in terms of the hierarchical structure of catchments and dendritic stream networks; (2) obtaining relevant and spatially continuous environmental predictors along the stream network, sourcing from three interfaces (stream vs. atmosphere, catchment area, groundwater) and (3) species detectability and thus the challenge of obtaining freshwater species occurrence data along the stream network. To depict freshwater species distributions at the stream level in the best possible way, multi-scale models are recommended, which account for scale dependence, a stronger integration of various environmental data types and sources, and heterogeneous species data. While model projections have the potential to serve as reliable tools for conservation and management purposes, this is only true if these relevant factors are considered in the modelling framework.
+++ title = “Application of species distribution models in stream ecosystems: the challenges of spatial and temporal scale, environmental predictors and species occurrence data” date = “2015-02-01” authors = ["Domisch, S.", “Jähnig, S.C.”, “Simaika, J.P.”, “Kuemmerlen, M.”, “Stoll, S.”] publication_types = [“2”] publication = “Fundamental and Applied Limnology, (186), 1-2, pp. 45–61, https://doi.org/10.1127/fal/2015/0627" publication_short = “Fundamental and Applied Limnology (186), 45–61” abstract = “Modelling and mapping habitat suitability of species in a given study area has become a popular method to assess biogeographical questions on the habitat requirements of species, and to study the effects of land use or climate change on species distributions. Beside purely academic aims, model projections are increasingly used for devising conservation and management strategies and they are an established tool in conservation planning. Species distribution models (SDMs) link geographic presence records of species with environmental conditions found at these locations, and extrapolate the occurrence probabilities of species throughout the study area. While the modelling framework is similar in the terrestrial, marine and freshwater realms, each realm comprises specific challenges for combining the spatial scale, the environmental data and the species records for building reliable models. In this paper, we focus on these key issues from a stream ecology perspective and highlight three critical challenges to be kept in mind when building SDMs in stream ecosystems: (1) the spatial configuration in terms of the hierarchical structure of catchments and dendritic stream networks; (2) obtaining relevant and spatially continuous environmental predictors along the stream network, sourcing from three interfaces (stream vs. atmosphere, catchment area, groundwater) and (3) species detectability and thus the challenge of obtaining freshwater species occurrence data along the stream network. To depict freshwater species distributions at the stream level in the best possible way, multi-scale models are recommended, which account for scale dependence, a stronger integration of various environmental data types and sources, and heterogeneous species data. While model projections have the potential to serve as reliable tools for conservation and management purposes, this is only true if these relevant factors are considered in the modelling framework. " abstract_short = "” image_preview = "” selected = false projects = [] tags = [] url_pdf = “https://doi.org/10.1127/fal/2015/0627" url_preprint = "” url_code = "" url_dataset = "" url_project = "" url_slides = "" url_video = "" url_poster = "" url_source = "" math = true highlight = true [header] image = "" caption = "" +++