A retrospective approach for evaluating ecological niche modeling transferability over time: the case of Mexican endemic rodents DOI Creative Commons
Claudia N. Moreno-Arzate, Enrique Martínez‐Meyer

PeerJ, Journal Year: 2024, Volume and Issue: 12, P. e18414 - e18414

Published: Nov. 29, 2024

Ecological niche modeling (ENM) is a valuable tool for inferring suitable environmental conditions and estimating species’ geographic distributions. ENM widely used to assess the potential effects of climate change on species distributions; however, choice algorithm introduces substantial uncertainty, especially since future projections cannot be properly validated. In this study, we evaluated performance seven popular algorithms—Bioclim, generalized additive models (GAM), linear (GLM), boosted regression trees (BRT), Maxent, random forest (RF), support vector machine (SVM)—in transferring across time, using Mexican endemic rodents as model system. We retrospective approach, from near past (1950–1979) more recent (1980–2009) vice versa, evaluate their in both forecasting hindcasting. Consistent with previous studies, our results highlight that input data quality significantly impact accuracy, but most importantly, found varied between While no single outperformed others temporal directions, RF generally showed better forecasting, while Maxent performed hindcasting, though it was sensitive small sample sizes. Bioclim consistently lowest performance. These findings underscore not all or algorithms are suited projections. Therefore, strongly recommend conducting thorough evaluation quality—in terms quantity biases—of interest. Based assessment, appropriate algorithm(s) should carefully selected rigorously tested before proceeding transfers.

Language: Английский

Optimising Species Distribution Models: Sample size, positional error, and sampling bias matter DOI Creative Commons
Vítězslav Moudrý, Manuele Bazzichetto, Ruben Remelgado

et al.

Published: Dec. 4, 2023

Species distribution models (SDMs) have proven valuable in filling gaps our knowledge of species occurrences. However, despite their broad applicability, SDMs exhibit critical shortcomings due to limitations occurrence data. These include, particular, issues related sample size, positional error, and sampling bias. In addition, it is widely recognized that the quality as well approaches used mitigate impact aforementioned data are dependent on ecology. While numerous studies experimentally evaluated effects these SDM performance, a synthesis results lacking. without comprehensive understanding individual combined effects, ability predict influence modelled species-environment associations remains largely uncertain, limiting value model outputs. this paper, we review bias, ecology We integrate findings into step-by-step guide for assessment spatial intended use SDMs.

Language: Английский

Citations

2
Shifts in ecological patterns and processes under global changes DOI Creative Commons
Mingzhen Lu, Lifei Wang, Lixin Wang

et al.

Landscape Ecology, Journal Year: 2024, Volume and Issue: 39(4)

Published: April 5, 2024

Language: Английский

Citations

0
A retrospective approach for evaluating ecological niche modeling transferability over time: the case of Mexican endemic rodents DOI Creative Commons
Claudia N. Moreno-Arzate, Enrique Martínez‐Meyer

PeerJ, Journal Year: 2024, Volume and Issue: 12, P. e18414 - e18414

Published: Nov. 29, 2024

Ecological niche modeling (ENM) is a valuable tool for inferring suitable environmental conditions and estimating species’ geographic distributions. ENM widely used to assess the potential effects of climate change on species distributions; however, choice algorithm introduces substantial uncertainty, especially since future projections cannot be properly validated. In this study, we evaluated performance seven popular algorithms—Bioclim, generalized additive models (GAM), linear (GLM), boosted regression trees (BRT), Maxent, random forest (RF), support vector machine (SVM)—in transferring across time, using Mexican endemic rodents as model system. We retrospective approach, from near past (1950–1979) more recent (1980–2009) vice versa, evaluate their in both forecasting hindcasting. Consistent with previous studies, our results highlight that input data quality significantly impact accuracy, but most importantly, found varied between While no single outperformed others temporal directions, RF generally showed better forecasting, while Maxent performed hindcasting, though it was sensitive small sample sizes. Bioclim consistently lowest performance. These findings underscore not all or algorithms are suited projections. Therefore, strongly recommend conducting thorough evaluation quality—in terms quantity biases—of interest. Based assessment, appropriate algorithm(s) should carefully selected rigorously tested before proceeding transfers.

Language: Английский

Citations

0