Does the interpolation accuracy of species distribution models come at the expense of transferability

By Risto Heikkinen, Mathieu Marmion, Miska Luoto. 2012

279 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Nature and Landscape Conservation

View Paper: https://www.semanticscholar.org/paper/385b37ecd50ee1a145fcf0c37bf09167451a3cdb

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Ecological Modeling

In Ecology

In Nature and Landscape Conservation

Citations in the corpus

Statistical Forecasting of Current and Future Circum‐Arctic Ground Temperatures and Active Layer Thickness (2018, background): “Similarly, the parametric GLM reduces the risk of overfitting (and thus the most spurious forecasts) providing relatively low 95th percentile uncertainty, but with a cost of modeling accuracy (Heikkinen et al., 2012).”
Climatic niche shifts between species' native and naturalized ranges raise concern for ecological forecasts during invasions and climate change (2014, result): “This approach has yielded equivocal conclusions (Leathwick, 1998; Pearson et al., 2006; Randin et al., 2006; Heikkinen et al., 2012; Sánchez-Fernández et al., 2012) and is fundamentally unsuited to assessing niche conservatism, for three principal reasons.”
Outstanding Challenges in the Transferability of Ecological Models. (2018, background): “How Should Uncertainty Be Quantified, Propagated, and Communicated When Transferring a Model? Uncertainty arises from many sources [64], including: sampling methodology, species vagrancy, data quality, environmental stochasticity, initial conditions, species identification, model specification, pred”
Integrating ensemble species distribution modelling and statistical phylogeography to inform projections of climate change impacts on species distributions (2013, methodology): “For many reasons, an SDM that performs well on current data cannot be presumed to perform well on projected climate data (Heikkinen et al. 2011), though this assumption is routinely made.”
A 40‐year, continent‐wide, multispecies assessment of relevant climate predictors for species distribution modelling (2014, background): “This probably results from a tendency of RF to overfit the data or to perform not as well when extrapolation is necessary (Heikkinen et al., 2012; Dormann et al., 2013).”
Climatic Risk and Distribution Atlas of European Bumblebees (2015, background): “However, GLMs had the overall best performance and their clear and simple mathematical formulation allows highly accurate extrapolations into new environmental space (Elith et al., 2006; Heikkinen et al., 2012).”