Accuracy of climate-based forecasts of pathogen spread

Abstract

Species distribution models (SDMs) are a tool for predicting the eventual geographical range of an emerging pathogen. Most SDMs, however, rely on an assumption of equilibrium with the environment, which an emerging pathogen, by definition, has not reached. To determine if some SDM approaches work better than others for modelling the spread of emerging, non-equilibrium pathogens, we studied time-sensitive predictive performance of SDMs for Batrachochytrium dendrobatidis, a devastating infectious fungus of amphibians, using multiple methods trained on time-incremented subsets of the available data. We split our data into timeline-based training and testing sets, and evaluated models on each set using standard performance criteria, including AUC, kappa, false negative rate and the Boyce index. Of eight models examined, we found that boosted regression trees and random forests performed best, closely followed by MaxEnt. As expected, predictive performance generally improved with the length of time series used for model training. These results provide information on how quickly the potential extent of an emerging disease may be determined, and identify which modelling frameworks are likely to provide useful information during the early phases of pathogen expansion.

Known and estimated distribution in Mexico of Batrachochytrium dendrobatidis, a pathogenic fungus of amphibians

“…In total, we obtained 238 unpublished records from the Colección Herpetológica de El Colegio de La Frontera Sur (ECOSUR) that had been obtained via systematic sampling in sites located along three transects east of the Isthmus of Tehuantepec over an elevation range of 0 to 4,000 m asl. We obtained 136 records from publications that have reported the presence of Bd in Mexico (Cheng et al., ; Frías‐Alvarez et al., ; García‐Feria, Brousset, Vallejo‐Bravo & Cervantes‐Olivares, ; Hale, Rosen, Jarchow & Bradley, ; Luja et al., ; Murrieta‐Galindo, Parra‐Olea, González‐Romero, López‐Barrera & Vredenburg, ; Peralta‐García et al., ; Schatz et al., ; Van‐Rooij et al., ), and 116 from theses (Cortés‐García, ; Familiar‐López, ). Part of the data has been previously used to model the distribution of Bd at different scales (Hale et al., ; Frías‐Álvarez et al., ; Olson et al., ; Schatz et al., ).…”

“…We obtained 136 records from publications that have reported the presence of Bd in Mexico (Cheng et al., ; Frías‐Alvarez et al., ; García‐Feria, Brousset, Vallejo‐Bravo & Cervantes‐Olivares, ; Hale, Rosen, Jarchow & Bradley, ; Luja et al., ; Murrieta‐Galindo, Parra‐Olea, González‐Romero, López‐Barrera & Vredenburg, ; Peralta‐García et al., ; Schatz et al., ; Van‐Rooij et al., ), and 116 from theses (Cortés‐García, ; Familiar‐López, ). Part of the data has been previously used to model the distribution of Bd at different scales (Hale et al., ; Frías‐Álvarez et al., ; Olson et al., ; Schatz et al., ). For this study, presence points were included when (a) the record had been georeferenced and (b) when the record had been confirmed in triplicate using the qPCR technique of Boyle, Boyle, Olsen, Morgan and Hyatt ().…”

“…MaxEnt constructs SDMs by combining presence data with bioclimate layers. In addition to being widely used, its performance is considered one of the best for estimating the distributions of species or emergent pathogenic organisms (Baldwin, ; Elith et al., ; Schatz et al., ).…”

“…With the information they provide, it has also been established that the decline of amphibian populations and species extinction are more probable in humid regions that have stable thermal conditions (Murray et al., ). At the global (Liu, Rohr & Li, ; Olson et al., ; Rödder, Kielgast & Lötters, ; Schatz, Kramer & Drake, ) and continental scales (James et al., ; Ron, ), it is predicted that in Mexico, the pathogen can encounter ideal conditions for its development in the Sierra Madre Oriental, Sierra Madre Occidental, and Trans‐Mexican Volcanic Belt mountain systems, and in the humid and subhumid temperate regions of the country. The variables that explain the presence of Bd are temperature, precipitation, and vegetation (Liu et al., ; Olson et al., ; Rödder et al., ).…”

Associations Between Habitat Quality And Body Size In The Carpathian Land SnailVestia turgida: Species Distribution Model Selection And Assessment Of Performance

“…SDMs therefore have been widely used for predicting distributions of species in terrestrial, freshwater and marine environments, and across taxa from many biological groups (Elith, Leathwick, 2009), with increasing numbers of publications each year (Robinson et al, 2011;Brotons, 2014). SDMs have shown to be efficient in biodiversity research considering climate changes (Barbet-Massin et al, 2011;Visconti et al, 2016), conservation planning (Kremen et al, 2008), invasive species and pest risk assessments (Gallien et al, 2012;Jeger et al, 2018), pathogen spread (Schatz et al, 2017), rewilding projects (Jarvie, Svenning, 2018), and a huge number of other issues ranging from mapping snake bite risk (Yañez-Arenas et al, 2016) to Pygmy presence in Central Africa (Olivero et al, 2016). SDM tools generally correlate species' occurrence patterns with environmental variables, which are frequently selected from an array of `bioclimatic' indices (Hijmans et al, 2005;Kriticos et al, 2012;etc.…”