The impact of modelling choices in the predictive performance of richness maps derived from species‐distribution models: guidelines to build better diversity models

Abstract

1. The stacking of species-distribution models (S-SDMs) is receiving attention by conservation researchers because this approach is capable of simultaneously predicting species richness and composition. However, the steps required to build S-SDMs implies at least two choices that influence its predictive performance which have not been extensively assessed: the selection of the modelling algorithm and the application of a threshold to transform the species-distribution models into binary maps to be added together to build the final S-SDM. Our goal was to provide guidelines concerning the best combinations of modelling algorithms and thresholds with which to build more accurate S-SDMs. 2. We generated 380 S-SDMs of 1224 tree species in Mesoamerica by combining 19 distribution modelling methods with 20 different thresholds using presence-only data from the Global Biodiversity Information Facility. We compared the predicted richness and composition with inventory data obtained from the BIOTREE-NET forest plot database. We designed two indicators of predictive performance that were based on the diversity factors used to measure species turnover: a (shared species between the observed and predicted compositions), b and c (the exclusive species of the predicted and observed compositions respectively) and compared them with the Sorensen and Beta-Simpson turnover measures. 3. Our proposed indexes and the Sorensen index proved suitable as indicators of predictive performance for S-SDMs, whereas the Beta-Simpson turnover measure presented issues that would prevent its application to evaluate S-SDMs. 4. Some modelling methods-especially machine learning and ensemble model forecasting methods performed significantly better than others in minimizing the error in predicted richness and composition. Our results also points out that restrictive thresholds (with high omission errors) lead to more accurate S-SDMs in terms of species richness and composition. Here, we demonstrate that particular combinations of modelling methods and thresholds provide results with higher predictive performance. 5. These results provide clear modelling guidelines that will help S-SDM modellers to select the appropriate combination of modelling methods and thresholds to build more accurate S-SDMs, and therefore will have a positive impact on the quality of the diversity models used to assist conservation planning.

ssdm: An r package to predict distribution of species richness and composition based on stacked species distribution models

“…The final maps of local species richness and composition can be computed using six different methods: (1) by summing discrete presence/absence maps (bSSDM) derived from one of the six metrics available to compute binary maps detailed in the next section (e.g. Benito et al., ; Brown et al., ; Fitzpatrick et al., ; Midgley et al., ; Moraes et al., ; Ogawa‐Onishi et al., ; Raes et al., ); (2) by summing discrete presence/absence maps obtained by drawing repeatedly from a Bernoulli distribution (see Dubuis et al., ; Calabrese et al., for further details); (3) by summing continuous habitat suitability maps (pSSDM) (e.g. Mateo et al., ; Murray‐Smith et al., ; Pouteau, Bayle, et al., ; Schmidt‐Lebuhn et al., ); (4) by applying the PRR of the SESAM framework (a number of species equal to the prediction of species richness is selected on the basis of decreasing probability of presence calculated by the SDMs) with species richness as estimated by a pSSDM (referred to as “PRR.pSSDM”) (D'Amen, Dubuis, et al., ); (5) by applying the PRR with species richness as estimated by a MEM (“PRR.MEM”) (D'Amen, Dubuis, et al., ; D'Amen, Pradervand, et al., ; Guisan & Rahbek, ); and (6) using the maximum‐likelihood adjustment approach proposed by Calabrese et al.…”

“…Macroecological models and pSSDMs both tend to perform similarly and to overestimate at sites with low species richness and underestimate at sites with high species richness (Calabrese et al., ). In contrast, bSSDMs tend to overpredict species richness, which is associated with generally higher and asymmetric prediction errors than MEMs, and may be affected by the choice of threshold for making binary predictions (Benito et al., ; Calabrese et al., ; Cord, Klein, Gernandt, de la Rosa, & Dech, ; D'Amen, Pradervand, et al., ; Dubuis et al., ).…”

Downscaling patterns of complementarity to a finer resolution and its implications for conservation prioritization

“…(), sunshine and topographic diversity (Benito et al. ) from USGS (n.d.). We calculated the mean or range of each environmental variable across each 50‐km and 10‐km grid cell.…”

Spatially explicit cost-effective actions for biodiversity threat abatement

“…Finer-resolution models can provide managers a better indication of priority management areas for species of interest (Bombi et al 2011, Gillingham et al 2012, and may be modelled at an appropriate spatial scale to match intended conservation applications . It is also increasingly feasible to better address uncertainty in SDM modelling (Beale and Lennon 2012) and guidelines to better mapping are evolving (Benito et al 2013). On the other hand, expert-derived geographic range maps may have more perceived credibility, and are more generally available than specific project-derived SDMs created from dedicated resources.…”