Harnessing the power of machine and deep learning for transferring joint species distribution models considering the structure of biotic interactions

The transferability of single or joint species distribution models ((j)SDMs) depends on their ability to predict beyond the observed environmental range and to remain consistent despite shifts in biotic interactions. Transfer accuracy may be improved by recent advances in the application of deep learning that provide greater flexibility and potentially superior predictive accuracy than traditional approaches. We implemented jSDMs with deep and machine learning algorithms and measured the transfer accuracy from continental to regional areas in communities with different species composition. We ran jSDMs with deep neural networks (DNN), elastic net (EN), and stacked SDMs (sSDM) with random forests (RF). We used 134 689 occurrence records representing 1776 species of six taxonomic groups (beetles, birds, bryophytes, fungi, lichens and plants) from 2387 forest plots in Europe. We employed an agnostic modelling approach that covered most of the environmental conditions by including more than 100 satellite‐derived variables and 98 climatic variables. The predictive power of the models within the training continental area was evaluated using AUC, whereas the transfer accuracy in the regional area was evaluated with the Boyce index calculated with independent presence records. We found that the DNN–jSDMs outperformed other models at continental scale, but model transfer from continental to regional extent was less accurate. We found that the accuracy of regional predictions was higher for taxonomic groups with better representation in the continental data, such as birds, bryophytes and plants. Depending on the algorithm and the taxonomic group, we achieved acceptable (Boyce > 0) to accurate (Boyce > 0.5) transferability for 32–78% of the species. Our findings underscored the need of considering trade‐offs among hyperparameter tuning, spatial scales and model complexity. Our findings also suggest that the varying biotic interaction structures and, particularly, the different species compositions of the transfer areas, may affect model transferability more than previously considered.

Authors

Institutions

Publication Details

Journal
Ecography
Published
2026-04-15
DOI
https://doi.org/10.1002/ecog.08269
Primary Topic
Species Distribution and Climate Change
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Harnessing the power of machine and deep learning for transferring joint species distribution models considering the structure of biotic interactions

Luc De Keersmaeker, Sabina Burrascano, Andrin Gross, Kastytis Šimkevičius et al.
Ecography
Species Distribution and Climate Change
article

Harnessing the power of machine and deep learning for transferring joint species distribution models considering the structure of biotic interactions

Luc De Keersmaeker, Sabina Burrascano, Andrin Gross, Kastytis Šimkevičius, Gediminas Brazaitis, Frédéric Gosselin, Ettore D’Andrea, Fotios Xystrakis, Francesco Rota, Elena Haeler, Frédéric Archaux, Francesco Chianucci, Flóra Tinya, Péter Ódor, Kris Vandekerkhove, Thomas Campagnaro, Maximilian Pichler, Anders Mårell, Nathalie Korboulewsky, Mariana Ujházyová, Carlos Miguel Landivar Albis, Anja Leyman, Nicolas Strebel, Daniel Kozák, Francesco Valerio, Tommaso Sitzia, Zoltán Elek, Lorenzo Balducci, Marion Gosselin, Marco Basile, Pallieter De Smedt, Martin Mikoláš, Sérgio Godinho, Xiang Liu, Sebastian Kepfer-Rojas, Christophe Bouget, Thibault Lachat, Kris Verheyen, Itziar García‐ Mijangos, Yoan Paillet, Miroslav Svoboda, Radim Matula, Michael Wohlwend, Silvia Stofer, Wouter Dekoninck
article en

Abstract

The transferability of single or joint species distribution models ((j)SDMs) depends on their ability to predict beyond the observed environmental range and to remain consistent despite shifts in biotic interactions. Transfer accuracy may be improved by recent advances in the application of deep learning that provide greater flexibility and potentially superior predictive accuracy than traditional approaches. We implemented jSDMs with deep and machine learning algorithms and measured the transfer accuracy from continental to regional areas in communities with different species composition. We ran jSDMs with deep neural networks (DNN), elastic net (EN), and stacked SDMs (sSDM) with random forests (RF). We used 134 689 occurrence records representing 1776 species of six taxonomic groups (beetles, birds, bryophytes, fungi, lichens and plants) from 2387 forest plots in Europe. We employed an agnostic modelling approach that covered most of the environmental conditions by including more than 100 satellite‐derived variables and 98 climatic variables. The predictive power of the models within the training continental area was evaluated using AUC, whereas the transfer accuracy in the regional area was evaluated with the Boyce index calculated with independent presence records. We found that the DNN–jSDMs outperformed other models at continental scale, but model transfer from continental to regional extent was less accurate. We found that the accuracy of regional predictions was higher for taxonomic groups with better representation in the continental data, such as birds, bryophytes and plants. Depending on the algorithm and the taxonomic group, we achieved acceptable (Boyce > 0) to accurate (Boyce > 0.5) transferability for 32–78% of the species. Our findings underscored the need of considering trade‐offs among hyperparameter tuning, spatial scales and model complexity. Our findings also suggest that the varying biotic interaction structures and, particularly, the different species compositions of the transfer areas, may affect model transferability more than previously considered.

Ecography
University of Copenhagen (DK), Bern University of Applied Sciences (CH), University of Debrecen (HU), University of Padua (IT), University of Freiburg (DE), University of the Basque Country (ES), Copenhagen Business School (DK), Universidade do Porto (PT), Vytautas Magnus University (LT), Research Institute for Nature and Forest (BE), Czech University of Life Sciences Prague (CZ), University of Évora (PT), Swiss Ornithological Institute (CH), Swiss Federal Institute for Forest, Snow and Landscape Research (CH), Institute of Natural Sciences (BE), Humboldt-Universität zu Berlin (DE), Institut National de Recherche pour l'Agriculture, l'Alimentation et l'Environnement (FR), Institute of Ecology and Botany (HU), Swiss Federal University for Vocational Education and Training SFUVET (CH), Technická univerzita vo Zvolene (SK), Laboratoire Interdisciplinaire de Physique (FR), Forestry Research Centre (IT), Leibniz Institute of Freshwater Ecology and Inland Fisheries (DE), HUN-REN Centre for Ecological Research (HU), Forest Research Institute (BG), University of Veterinary Medicine (HU), Research Institute on Sustainable Economic Growth (IT), Austrian Research Centre for Forests (AT), Budapest Institute (HU), University of Regensburg (DE), University of Sopron (HU), Sapienza University of Rome (IT), Université Grenoble Alpes (FR), University of Palermo (IT)
European Commission, Grantová Agentura České Republiky, Università degli Studi di Padova, Magyar Tudományos Akadémia, Novo Nordisk, Vedecká Grantová Agentúra MŠVVaŠ SR a SAV, Nemzeti Kutatási, Fejlesztési és Innovaciós Alap
Life in Land
Openalex Percentile: Top 14%
Species Distribution and Climate Change
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.