Spatial confounding in Bayesian species distribution modeling
1) Species distribution models (SDMs) are currently the main tools to derive species niche estimates and spatially explicit predictions for species geographical distribution. However, unobserved environmental conditions and ecological processes may confound the model estimates if they have direct im...
Published in: | Ecography |
---|---|
Main Authors: | , , , , |
Other Authors: | , , , , , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Wiley
2022
|
Subjects: | |
Online Access: | http://hdl.handle.net/10138/350497 |
id |
ftunivhelsihelda:oai:helda.helsinki.fi:10138/350497 |
---|---|
record_format |
openpolar |
spelling |
ftunivhelsihelda:oai:helda.helsinki.fi:10138/350497 2024-01-07T09:45:30+01:00 Spatial confounding in Bayesian species distribution modeling Mäkinen, Jussi Numminen, Elina Niittynen, Pekka Luoto, Miska Vanhatalo, Jarno Organismal and Evolutionary Biology Research Programme Research Centre for Ecological Change Faculty of Biological and Environmental Sciences Environmental and Ecological Statistics Group Department of Mathematics and Statistics Faculty of Science Department of Geosciences and Geography BioGeoClimate Modelling Lab 2022-11-07T11:10:03Z 12 application/pdf http://hdl.handle.net/10138/350497 eng eng Wiley 10.1111/ecog.06183 Suomen Akatemia Projektilaskutus This work was funded by the Academy of Finland (grant no. 317255) and Jane and Aatos Erkko foundation (Jussi Makinen, Elina Numminen and Jarno Vanhatalo); Kone foundation, Societas pro Fauna et Flora Fennica (Pekka Niittynen). Mäkinen , J , Numminen , E , Niittynen , P , Luoto , M & Vanhatalo , J 2022 , ' Spatial confounding in Bayesian species distribution modeling ' , Ecography , vol. 2022 , no. 11 , e06183 . https://doi.org/10.1111/ecog.06183 ORCID: /0000-0001-6203-5143/work/122636483 ORCID: /0000-0001-6599-8279/work/122643838 ORCID: /0000-0002-7290-029X/work/122643869 d3f3f279-4081-4e1d-bc29-51e90f8f64d7 http://hdl.handle.net/10138/350497 000847425300001 cc_by openAccess info:eu-repo/semantics/openAccess estimation bias Gaussian process spatial confounding spatial random effect species distribution model BIOTIC INTERACTIONS RED HERRINGS AUTOCORRELATION PREDICTION REGRESSION INFERENCE PRIORS 1181 Ecology evolutionary biology Article publishedVersion 2022 ftunivhelsihelda 2023-12-14T00:02:34Z 1) Species distribution models (SDMs) are currently the main tools to derive species niche estimates and spatially explicit predictions for species geographical distribution. However, unobserved environmental conditions and ecological processes may confound the model estimates if they have direct impact on the species and, at the same time, they are correlated with the observed environmental covariates. This, so-called spatial confounding, is a general property of spatial models and it has not been studied in the context of SDMs before. 2) We examine how the estimation accuracy of SDMs depends on the type of spatial confounding. We construct two simulation studies where we alter spatial structures of the observed and unobserved covariates and the level of dependence between them. We fit generalized linear models with and without spatial random effects applying Bayesian inference and recording the bias induced to model estimates by spatial confounding. After this we examine spatial confounding also with real vegetation data from northern Norway. 3) Our results show that model estimates for coarse scale covariates, such as climate covariates, are likely to be biased if a species distribution depends also on an unobserved covariate operating on a finer spatial scale. Pushing higher probability for a relatively weak and smoothly varying spatial random effect compared to the observed covariates improved the model's estimation accuracy. The improvement was independent of the actual spatial structure of the unobserved covariate. 4) Our study addresses the major factors of spatial confounding in SDMs and provides a list of recommendations for pre-inference assessment of spatial confounding and for inference-based methods to decrease the chance of biased model estimates. Peer reviewed Article in Journal/Newspaper Northern Norway HELDA – University of Helsinki Open Repository Norway Ecography 2022 11 |
institution |
Open Polar |
collection |
HELDA – University of Helsinki Open Repository |
op_collection_id |
ftunivhelsihelda |
language |
English |
topic |
estimation bias Gaussian process spatial confounding spatial random effect species distribution model BIOTIC INTERACTIONS RED HERRINGS AUTOCORRELATION PREDICTION REGRESSION INFERENCE PRIORS 1181 Ecology evolutionary biology |
spellingShingle |
estimation bias Gaussian process spatial confounding spatial random effect species distribution model BIOTIC INTERACTIONS RED HERRINGS AUTOCORRELATION PREDICTION REGRESSION INFERENCE PRIORS 1181 Ecology evolutionary biology Mäkinen, Jussi Numminen, Elina Niittynen, Pekka Luoto, Miska Vanhatalo, Jarno Spatial confounding in Bayesian species distribution modeling |
topic_facet |
estimation bias Gaussian process spatial confounding spatial random effect species distribution model BIOTIC INTERACTIONS RED HERRINGS AUTOCORRELATION PREDICTION REGRESSION INFERENCE PRIORS 1181 Ecology evolutionary biology |
description |
1) Species distribution models (SDMs) are currently the main tools to derive species niche estimates and spatially explicit predictions for species geographical distribution. However, unobserved environmental conditions and ecological processes may confound the model estimates if they have direct impact on the species and, at the same time, they are correlated with the observed environmental covariates. This, so-called spatial confounding, is a general property of spatial models and it has not been studied in the context of SDMs before. 2) We examine how the estimation accuracy of SDMs depends on the type of spatial confounding. We construct two simulation studies where we alter spatial structures of the observed and unobserved covariates and the level of dependence between them. We fit generalized linear models with and without spatial random effects applying Bayesian inference and recording the bias induced to model estimates by spatial confounding. After this we examine spatial confounding also with real vegetation data from northern Norway. 3) Our results show that model estimates for coarse scale covariates, such as climate covariates, are likely to be biased if a species distribution depends also on an unobserved covariate operating on a finer spatial scale. Pushing higher probability for a relatively weak and smoothly varying spatial random effect compared to the observed covariates improved the model's estimation accuracy. The improvement was independent of the actual spatial structure of the unobserved covariate. 4) Our study addresses the major factors of spatial confounding in SDMs and provides a list of recommendations for pre-inference assessment of spatial confounding and for inference-based methods to decrease the chance of biased model estimates. Peer reviewed |
author2 |
Organismal and Evolutionary Biology Research Programme Research Centre for Ecological Change Faculty of Biological and Environmental Sciences Environmental and Ecological Statistics Group Department of Mathematics and Statistics Faculty of Science Department of Geosciences and Geography BioGeoClimate Modelling Lab |
format |
Article in Journal/Newspaper |
author |
Mäkinen, Jussi Numminen, Elina Niittynen, Pekka Luoto, Miska Vanhatalo, Jarno |
author_facet |
Mäkinen, Jussi Numminen, Elina Niittynen, Pekka Luoto, Miska Vanhatalo, Jarno |
author_sort |
Mäkinen, Jussi |
title |
Spatial confounding in Bayesian species distribution modeling |
title_short |
Spatial confounding in Bayesian species distribution modeling |
title_full |
Spatial confounding in Bayesian species distribution modeling |
title_fullStr |
Spatial confounding in Bayesian species distribution modeling |
title_full_unstemmed |
Spatial confounding in Bayesian species distribution modeling |
title_sort |
spatial confounding in bayesian species distribution modeling |
publisher |
Wiley |
publishDate |
2022 |
url |
http://hdl.handle.net/10138/350497 |
geographic |
Norway |
geographic_facet |
Norway |
genre |
Northern Norway |
genre_facet |
Northern Norway |
op_relation |
10.1111/ecog.06183 Suomen Akatemia Projektilaskutus This work was funded by the Academy of Finland (grant no. 317255) and Jane and Aatos Erkko foundation (Jussi Makinen, Elina Numminen and Jarno Vanhatalo); Kone foundation, Societas pro Fauna et Flora Fennica (Pekka Niittynen). Mäkinen , J , Numminen , E , Niittynen , P , Luoto , M & Vanhatalo , J 2022 , ' Spatial confounding in Bayesian species distribution modeling ' , Ecography , vol. 2022 , no. 11 , e06183 . https://doi.org/10.1111/ecog.06183 ORCID: /0000-0001-6203-5143/work/122636483 ORCID: /0000-0001-6599-8279/work/122643838 ORCID: /0000-0002-7290-029X/work/122643869 d3f3f279-4081-4e1d-bc29-51e90f8f64d7 http://hdl.handle.net/10138/350497 000847425300001 |
op_rights |
cc_by openAccess info:eu-repo/semantics/openAccess |
container_title |
Ecography |
container_volume |
2022 |
container_issue |
11 |
_version_ |
1787427056490905600 |