Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment
Convolutional Neural Networks (CNNs) are statistical models suited for learning complex visual patterns. In the context of Species Distribution Models (SDM) and in line with predictions of landscape ecology and island biogeography, CNN could grasp how local landscape structure affects prediction of species occurrence in SDMs. The prediction can thus reflect the signatures of entangled ecological processes. Although previous machine-learning based SDMs can learn complex influences of environmental predictors, they cannot acknowledge the influence of environmental structure in local landscapes (hence denoted “punctual models”). In this study, we applied CNNs to a large dataset of plant occurrences in France (GBIF), on a large taxonomical scale, to predict ranked relative probability of species (by joint learning) to any geographical position. We examined the way local environmental landscapes improve prediction by performing alternative CNN models deprived of information on landscape heterogeneity and structure (“ablation experiments”). We found that the landscape structure around location crucially contributed to improve predictive performance of CNN-SDMs. CNN models can classify the predicted distributions of many species, as other joint modelling approaches, but they further prove efficient in identifying the influence of local environmental landscapes. CNN can then represent signatures of spatially structured environmental drivers. The prediction gain is noticeable for rare species, which open promising perspectives for biodiversity monitoring and conservation strategies. Therefore, the approach is of both theoretical and practical interest. We discuss the way to test hypotheses on the patterns learnt by CNN, which should be essential for further interpretation of the ecological processes at play.
Main Authors: | , , , , , |
---|---|
Format: | article biblioteca |
Language: | eng |
Subjects: | F70 - Taxonomie végétale et phytogéographie, U10 - Informatique, mathématiques et statistiques, F40 - Écologie végétale, modélisation environnementale, distribution des populations, distribution géographique, réseau de neurones, biogéographie, écologie, paysage, http://aims.fao.org/aos/agrovoc/c_9000056, http://aims.fao.org/aos/agrovoc/c_6113, http://aims.fao.org/aos/agrovoc/c_5083, http://aims.fao.org/aos/agrovoc/c_37467, http://aims.fao.org/aos/agrovoc/c_915, http://aims.fao.org/aos/agrovoc/c_2467, http://aims.fao.org/aos/agrovoc/c_4185, http://aims.fao.org/aos/agrovoc/c_3081, |
Online Access: | http://agritrop.cirad.fr/598364/ http://agritrop.cirad.fr/598364/1/journal.pcbi.1008856.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
dig-cirad-fr-598364 |
---|---|
record_format |
koha |
spelling |
dig-cirad-fr-5983642024-04-24T11:44:25Z http://agritrop.cirad.fr/598364/ http://agritrop.cirad.fr/598364/ Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment. Deneu Benjamin, Servajean Maximilien, Bonnet Pierre, Botella Christophe, Munoz François, Joly Alexis. 2021. PLoS Computational Biology, 17 (4):e1008856, 21 p.https://doi.org/10.1371/journal.pcbi.1008856 <https://doi.org/10.1371/journal.pcbi.1008856> Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment Deneu, Benjamin Servajean, Maximilien Bonnet, Pierre Botella, Christophe Munoz, François Joly, Alexis eng 2021 PLoS Computational Biology F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 France http://aims.fao.org/aos/agrovoc/c_3081 Convolutional Neural Networks (CNNs) are statistical models suited for learning complex visual patterns. In the context of Species Distribution Models (SDM) and in line with predictions of landscape ecology and island biogeography, CNN could grasp how local landscape structure affects prediction of species occurrence in SDMs. The prediction can thus reflect the signatures of entangled ecological processes. Although previous machine-learning based SDMs can learn complex influences of environmental predictors, they cannot acknowledge the influence of environmental structure in local landscapes (hence denoted “punctual models”). In this study, we applied CNNs to a large dataset of plant occurrences in France (GBIF), on a large taxonomical scale, to predict ranked relative probability of species (by joint learning) to any geographical position. We examined the way local environmental landscapes improve prediction by performing alternative CNN models deprived of information on landscape heterogeneity and structure (“ablation experiments”). We found that the landscape structure around location crucially contributed to improve predictive performance of CNN-SDMs. CNN models can classify the predicted distributions of many species, as other joint modelling approaches, but they further prove efficient in identifying the influence of local environmental landscapes. CNN can then represent signatures of spatially structured environmental drivers. The prediction gain is noticeable for rare species, which open promising perspectives for biodiversity monitoring and conservation strategies. Therefore, the approach is of both theoretical and practical interest. We discuss the way to test hypotheses on the patterns learnt by CNN, which should be essential for further interpretation of the ecological processes at play. article info:eu-repo/semantics/article Journal Article info:eu-repo/semantics/publishedVersion http://agritrop.cirad.fr/598364/1/journal.pcbi.1008856.pdf text cc_by info:eu-repo/semantics/openAccess https://creativecommons.org/licenses/by/4.0/ https://doi.org/10.1371/journal.pcbi.1008856 10.1371/journal.pcbi.1008856 info:eu-repo/semantics/altIdentifier/doi/10.1371/journal.pcbi.1008856 info:eu-repo/semantics/altIdentifier/purl/https://doi.org/10.1371/journal.pcbi.1008856 info:eu-repo/semantics/reference/purl/https://gitlab.inria.fr/bdeneu/cnn-sdm |
institution |
CIRAD FR |
collection |
DSpace |
country |
Francia |
countrycode |
FR |
component |
Bibliográfico |
access |
En linea |
databasecode |
dig-cirad-fr |
tag |
biblioteca |
region |
Europa del Oeste |
libraryname |
Biblioteca del CIRAD Francia |
language |
eng |
topic |
F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 http://aims.fao.org/aos/agrovoc/c_3081 F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 http://aims.fao.org/aos/agrovoc/c_3081 |
spellingShingle |
F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 http://aims.fao.org/aos/agrovoc/c_3081 F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 http://aims.fao.org/aos/agrovoc/c_3081 Deneu, Benjamin Servajean, Maximilien Bonnet, Pierre Botella, Christophe Munoz, François Joly, Alexis Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
description |
Convolutional Neural Networks (CNNs) are statistical models suited for learning complex visual patterns. In the context of Species Distribution Models (SDM) and in line with predictions of landscape ecology and island biogeography, CNN could grasp how local landscape structure affects prediction of species occurrence in SDMs. The prediction can thus reflect the signatures of entangled ecological processes. Although previous machine-learning based SDMs can learn complex influences of environmental predictors, they cannot acknowledge the influence of environmental structure in local landscapes (hence denoted “punctual models”). In this study, we applied CNNs to a large dataset of plant occurrences in France (GBIF), on a large taxonomical scale, to predict ranked relative probability of species (by joint learning) to any geographical position. We examined the way local environmental landscapes improve prediction by performing alternative CNN models deprived of information on landscape heterogeneity and structure (“ablation experiments”). We found that the landscape structure around location crucially contributed to improve predictive performance of CNN-SDMs. CNN models can classify the predicted distributions of many species, as other joint modelling approaches, but they further prove efficient in identifying the influence of local environmental landscapes. CNN can then represent signatures of spatially structured environmental drivers. The prediction gain is noticeable for rare species, which open promising perspectives for biodiversity monitoring and conservation strategies. Therefore, the approach is of both theoretical and practical interest. We discuss the way to test hypotheses on the patterns learnt by CNN, which should be essential for further interpretation of the ecological processes at play. |
format |
article |
topic_facet |
F70 - Taxonomie végétale et phytogéographie U10 - Informatique, mathématiques et statistiques F40 - Écologie végétale modélisation environnementale distribution des populations distribution géographique réseau de neurones biogéographie écologie paysage http://aims.fao.org/aos/agrovoc/c_9000056 http://aims.fao.org/aos/agrovoc/c_6113 http://aims.fao.org/aos/agrovoc/c_5083 http://aims.fao.org/aos/agrovoc/c_37467 http://aims.fao.org/aos/agrovoc/c_915 http://aims.fao.org/aos/agrovoc/c_2467 http://aims.fao.org/aos/agrovoc/c_4185 http://aims.fao.org/aos/agrovoc/c_3081 |
author |
Deneu, Benjamin Servajean, Maximilien Bonnet, Pierre Botella, Christophe Munoz, François Joly, Alexis |
author_facet |
Deneu, Benjamin Servajean, Maximilien Bonnet, Pierre Botella, Christophe Munoz, François Joly, Alexis |
author_sort |
Deneu, Benjamin |
title |
Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
title_short |
Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
title_full |
Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
title_fullStr |
Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
title_full_unstemmed |
Convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
title_sort |
convolutional neural networks improve species distribution modelling by capturing the spatial structure of the environment |
url |
http://agritrop.cirad.fr/598364/ http://agritrop.cirad.fr/598364/1/journal.pcbi.1008856.pdf |
work_keys_str_mv |
AT deneubenjamin convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment AT servajeanmaximilien convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment AT bonnetpierre convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment AT botellachristophe convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment AT munozfrancois convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment AT jolyalexis convolutionalneuralnetworksimprovespeciesdistributionmodellingbycapturingthespatialstructureoftheenvironment |
_version_ |
1798165063619575808 |