Species Distribution Models (SDM)
 for freshwaters

Member of the

Freshwater Information Platform

How to build Species Distribution Models for freshwaters

Ana Filipa Filipe and Núria Bonada (University of Barcelona) and Teresa Connolly (University of Oxford)

What is this model

Species distribution models (SDMs) predict the potential distribution of a species in space and/or time. They are also known as environmental niche models, habitat suitability models or bioclimatic envelope models. Correlative SDMs are built based on the relationship between species’ occurrence and selected environmental variables. By using the field observations of species’ occurrences and environmental variables used as predictors, a number of algorithms can be applied to build SDMs that predict the habitat suitability of species in a wider landscape. These predictions can also be extended through time and thus provide future scenarios of species distributions. Common outputs of SDMs are maps of the probability of occurrence of the species through landscapes and/or through time (e.g. future, seasonal).

From the initial developments, SDMs have been broadened to be applicable to distinct taxa, non-terrestrial habitats, and using distinct environmental predictors beyond climate-only datasets. For the particular case of SDMs applied to freshwaters, habitat constraints as stream networks, lake patches, and costal buffers are also included. Also, different species use freshwater habitats exclusively, or they use multiple habitats (e.g. amphibians use both freshwater and terrestrial habitats).

Why use it

SDMs have been widely developed during recent years as they meet the needs of environmental research, resource management, and conservation planning. These developments result from the growing need of information on the geographical distribution of biodiversity and of the growing availability of data and techniques.

In the particular case of freshwater environments, SDMs are being used as tools for conservation and management purposes, from predicting species invasions to forecasting climate change impacts (e.g. Filipe et al. 2013; Domisch et al. 2013).

Building SDMs

Building SDMs applied to freshwaters is a complex task, and decisions made during the procedure depend on the aims of the particular study. Here we exemplify the particular case of forecasting stream fish species in Europe using the BIOMOD package of the R software and QGIS. The following text outlines the main steps for obtaining SDMs:

  • Data preparation: Merge datasets of fish species occurrence and GIS environmental layers at comparable scales, regarding resolution (unit of analyses) and extent of data (area studied). Examples of environmental datasets:
  • Modelling: Calibration of SDMs using individual or joint algorithms (i.e., ensemble approach) in the BIOMOD package (R software), including selection of environmental predictors and of threshold of binary prediction. Example of algorithms are GLM (Generalized Linear Model), GAM (Generalized Additive Model), and MARS (Multivariate Adaptive Regression Splines).
  • Modelling evaluation: Assessment of the predictive performance of the SDMs (e.g. AUC Area Under the Receiver Operator Curve statistics), including the significance of models for data not used in model
    calibration (model validation).
  • Transferability. Transfer SDMs to future climates, if that is the aim for the study.
  • Outputs from BIOMOD can be visualised using QGIS software and exported as a layout map file.

BIOMOD is a computer platform for ensemble forecasting of species distributions, which includes the treatment of a range of methodological uncertainties in models and the examination of species-environment relationships (see website for further reading).
QGIS is a user friendly open source Geographic Information System (GIS). For download and further information see QGIS website.

Problem solving

Detailed instructions and guidance on how to run the suite of models available in BIOMOD can be found in the Presentation Manual for BIOMOD. The version BIOMOD2 does not have a manual but rather a list of vignettes, accessible from within the software explaining the different steps to follow. In addition you can engage in user forums on the BIOMOD R-Forge page.

For QGIS support see QGIS website.

Contribute your findings to the BioFresh platform

Both BIOMOD and QGIS are open source softwares so you are encouraged to share your data with an open licence too. We invite you to contribute your data and/or results to the BioFresh platform. You can also see examples of SDMs built with BIOMOD in the online Global Freshwater Biodiversity Atlas:

Forecasts of Salmo trutta distribution in European basins

Climatic Suitability of European Stream Macroinvertebrates under Climate Change


For an overview of SDMs procedures:

Franklin, J. (2009) Mapping Species Distributions: Spatial Inference and Prediction, Cambridge University Press, Cambridge, UK.

Peterson, A.T., Soberon, J.M., Pearson, R.G., Anderson, R.P., Martinez-Meyer, E., Nakamura, M. & Araujo, M.B. (2011) Ecological niches and geographic distributions, (ed. by S.A. Levin and H.S. Horn) Princeton University Press.

For a scientific presentation of BIOMOD:

Thuiller W., Lafourcade B., Engler R. & Araujo M.B. (2009). BIOMOD – A platform for ensemble forecasting of species distributions. Ecography, 32, 369-373.

Some publications that exemplify the development of freshwater SDMs under the BioFresh project:

Filipe, A.F.; Markovic, D.; Pletterbauer, F.; Tisseuil, C.; De Wever, A.; Schmutz, S.; Bonada, N.; Freyhof, J. (2013). Forecasting fish distribution along stream networks: brown trout (Salmo trutta) in Europe. Diversity and Distributions, 19:1059-1071.

Domisch, S.; Araujo, M.B.; Bonada, N.; Pauls, S.U.; Jahing, S.C. & Hasse, P. (2013). Modelling distribution in European stream macroinvertebrates under future climates. Global Change Biology, 19:752-762.