Species distribution model transferability and model grain size – finer may not always be better

Abstract

Species distribution models have been used to predict the distribution of invasive species for conservation planning. Understanding spatial transferability of niche predictions is critical to promote species-habitat conservation and forecasting areas vulnerable to invasion. Grain size of predictor variables is an important factor affecting the accuracy and transferability of species distribution models. Choice of grain size is often dependent on the type of predictor variables used and the selection of predictors sometimes rely on data availability. This study employed the MAXENT species distribution model to investigate the effect of the grain size on model transferability for an invasive plant species. We modelled the distribution of Rhododendron ponticum in Wales, U.K. and tested model performance and transferability by varying grain size (50 m, 300 m, and 1 km). MAXENT-based models are sensitive to grain size and selection of variables. We found that over-reliance on the commonly used bioclimatic variables may lead to less accurate models as it often compromises the finer grain size of biophysical variables which may be more important determinants of species distribution at small spatial scales. Model accuracy is likely to increase with decreasing grain size. However, successful model transferability may require optimization of model grain size.

Climate change and the future restructuring of Neotropical anuran biodiversity

“…To evaluate the effect of these two common sources of uncertainty (Pearson and Dawson 2003, Braunisch et al 2013, Manzoor et al 2018), we reestimated baseline and future conditions using climatic variables with a spatial resolution of 10 arc-min (i.e. To evaluate the effect of these two common sources of uncertainty (Pearson and Dawson 2003, Braunisch et al 2013, Manzoor et al 2018), we reestimated baseline and future conditions using climatic variables with a spatial resolution of 10 arc-min (i.e.…”

How does spatial resolution affect model performance? A case for ensemble approaches for marine benthic mesophotic communities

“…A coarser grid size, and thus a smoothing of variable values, may result in simpler relationships between the response and predictors. This can prevent overfitting of models (Merow et al., ) and is supported by observations that model transferability is improved at coarser grid sizes (Manzoor et al., ). This may explain the greater performance of the macroalgae and sponge models at coarser grid sizes when evaluated with independent data.…”

“…Modelling studies attempt to draw conclusions about the important ecological processes that are influencing the distribution of organisms. Ecological processes occur across a range of spatial scales which will influence an organisms’ distribution (Kendrick, Holmes, & Van Niel, ; Manzoor et al., ). Sessile taxa are more affected than mobile taxa as their distributions may be more reliant on fine‐scale patterns (Guisan et al., ).…”

“…The use of coarse variables may result in large overgeneralizations and more homogenous landscapes (Gottschalk et al., ; Rahbek, ; Scales et al., ). This results in a reduction in the number of “distinct” grid cells which separate out different classes within the data (Gottschalk et al., ; Manzoor et al., ). For example, grid cells comprising of steep slopes are more prevalent, with a greater range of values for slope (0.015–26.34) at 10 m grid sizes.…”

“…However at 250 m grid sizes, slope values ranged from 0.11 to 4.03 resulting in less classes at the extremes as values are averaged across a wider area for each grid cell. Fine‐scale patterns in topography can be associated with changes in other environmental factors such as local terrestrial climate (Manzoor et al., ), or in the marine environment light and wave and current exposure. Corals may be particularly receptive to these factors while they also require areas of high relief topography to attach and avoid smothering by sedimentation.…”