Testing the predictive performance of comparative extinction risk models to support the global amphibian assessment

Abstract

Assessing the extinction risk of species through the IUCN Red List is key to guiding conservation policies and reducing biodiversity loss. This process is resource-demanding, however, and requires a continuous update which becomes increasingly difficult as new species are added to the IUCN Red List. The use of automatic methods, such as comparative analyses to predict species extinction risk, can be an efficient alternative to maintaining up to date assessments. Using amphibians as a study group, we predict which species were more likely to change status, in order to suggest species that should be prioritized for reassessment. We used species traits, environmental variables, and proxies of climate and land-use change as predictors of the IUCN Red List category of species. We produced an ensemble prediction of IUCN Red List categories by combining four different model algorithms: Cumulative Link Models (CLM), phylogenetic Generalized Least Squares (PGLS), Random Forests (RF), Neural Networks (NN). By comparing IUCN Red List categories with the ensemble prediction, and accounting for uncertainty among model algorithms, we identified species that should be prioritized for future reassessments due to high prediction versus observation mismatch. We found that CLM and RF performed better than PGLS and NN, but there was not a clear best algorithm. The most important predicting variables across models were species range size, climate change, and land-use change. We propose ensemble modelling of extinction risk as a promising tool for prioritizing species for reassessment while accounting for inherent models' uncertainty.

Modelling the probability of meeting IUCN Red List criteria to support reassessments

“…Comparative extinction risk models have often been promoted as useful tools to provide a first prediction of extinction risk for species not yet included in the Red List (Darrah et al., 2017; Zizka, Andermann, et al., 2022; Zizka, Barratt, et al., 2022), for data deficient species (Bland & Böhm, 2016; Borgelt et al., 2022; Cazalis et al., 2023; He et al., 2021; Luiz et al., 2016) or to prioritise reassessments (Di Marco et al., 2014; Lucas et al., 2023). But so far, these have remained largely unmet promises, with hardly any uptake of these modelling approaches in the Red List (Cazalis et al., 2022).…”

“…To investigate the relationship between species traits, extrinsic factors and extinction risk, we used cumulative link models (CLM, also known as ordinal regression models) from the R package ‘ordinal’ (Christensen, 2019), which allow preservation of the ordinal structure of the Red List categories (Lucas et al., 2019, 2023; Luiz et al., 2016). Moreover, CLMs have been demonstrated to be the best algorithms to deal with the ordinal structure of Red List categories when compared to other algorithms traditionally used in comparative models of extinction risk, such as Random Forest, Neural networks or Phylogenetic Generalised Linear Models (PGLS) (Lucas et al., 2023).…”

“…Finally, our models were employed to predict the most probable Red List category for all species, which was then transformed into a binary prediction (threatened/not threatened) and compared with the actual Red List category. The mismatches thus identified serve as flags for prioritising reassessments of a specific list of species, as classically done in comparative extinction risk analyses (Di Marco et al., 2014; Lucas et al., 2023).…”

“…change in forest canopy cover, distance to cities). We modelled the relationship between extinction risk predictors and each species' Red List category met under each specific criterion, using ordinal regression models that best match the ordinal structure of the Red List categories (Lucas et al, 2019(Lucas et al, , 2023Luiz et al, 2016), and combined these models into a single final prediction. We then compared the performance of these criterion-specific models with a criterion-blind approach, as well as new approach to increase the uptake of extinction risk models in Red List assessments, bridging a long-standing research-implementation gap.…”

“…human density, land‐use change, climate change; Chichorro et al., 2019, 2022; Purvis et al., 2000). The models are then used to predict the Red List categories of species that have not yet been assessed for the Red List (Darrah et al., 2017; Zizka et al., 2021), that are Data Deficient (Bland & Böhm, 2016; Luiz et al., 2016) or that need updating (Di Marco et al., 2014; Lucas et al., 2023), with the objective of accelerating and helping to prioritise the work of Red List assessors by providing additional information on species' extinction risk.…”