The Doñana Digital Twin: integrating heterogeneous data into iterative, mechanistically grounded forecasts of biodiversity change
Abstract
Natural ecosystems are increasingly impacted by multiple global-change drivers, and these impacts propagate non-linearly from individuals to populations and to communities. To anticipate ecological change under such complexities and mitigate negative impacts, we require integrated, mechanistically-grounded approaches to prediction. For mitigation strategies to be adaptive, predictions must also be updated efficiently as new evidence becomes available. In many protected areas, decades of heterogeneous monitoring data are available, but remain siloed within research groups or projects and are rarely turned into integrated forecasts. Here we present an ecological digital twin of Doñana Natural Area, a Mediterranean biodiversity hotspot under increasing climate and land-use pressures. The digital twin combines long-term biodiversity data with weather, hydrological and remote-sensing products. This data is then used to link demographic, abundance and occurrence models to interspecific interactions, which results in community and ecosystem dynamics. Annual forecasts are generated under global-change scenarios, and uncertainty is explicitly carried through every stage of this chain, from driver to outcome. The digital twin is hosted open-access on the LifeWatch ERIC e-infrastructure, thus ensuring transferability and continuity. Annual evaluation of forecast skill against benchmark models, open forecast submissions by different user groups, and a web interface support continuous updating and decision-making. Importantly, the digital twin is also built on standardized workflows, allowing new models, species, and data streams to be incorporated rapidly and consistently. This design makes the approach directly scalable: the same infrastructure can be extended to other protected areas that already hold numerous ecological datasets but currently lack the integrated products needed to translate that data into actionable ecological knowledge.
Related
- Long-term biodiversity responses to anthropogenic change are spatially and taxonomically heterogeneous across coastal wetlands
- Weak dispersal and landscape size inevitably promote local biodiversity in heterogeneous metacommunities of competing species
- Integrating images, sounds, citizen science and AI to assess the biodiversity of the Atlantic Forest of Brazil
- Rarefaction profiles for repeated biodiversity surveys: integrating temporal variability and trend along sample-based accumulation curves
- A heterogeneous biomedical knowledge network framework for rare disease drug candidate prioritization: integrating Orphadata and DisGeNET via gene-bridge harmonization
BibTeX
@misc{evers2026doñana,
title = {{The Doñana Digital Twin: integrating heterogeneous data into iterative, mechanistically grounded forecasts of biodiversity change}},
author = {Sanne M Evers and Julio Paneque and Aimara Planillo and Emma Underwood and Guillermo Gómez Peña and Hanna Serediuk and Teresa Sánchez-Mejía and Francisco Lloret and Eloy Revilla and Maria Paniw},
howpublished = {EcoEvoRxiv},
year = {2026},
month = oct,
doi = {10.32942/x27q4j},
url = {https://ecoevorxiv.org/repository/view/15591/},
note = {CC BY-NC 4.0},
}
RIS
TY - UNPB
TI - The Doñana Digital Twin: integrating heterogeneous data into iterative, mechanistically grounded forecasts of biodiversity change
AU - Sanne M Evers
AU - Julio Paneque
AU - Aimara Planillo
AU - Emma Underwood
AU - Guillermo Gómez Peña
AU - Hanna Serediuk
AU - Teresa Sánchez-Mejía
AU - Francisco Lloret
AU - Eloy Revilla
AU - Maria Paniw
PY - 2026
DA - 2026/10/20
PB - EcoEvoRxiv
DO - 10.32942/x27q4j
UR - https://ecoevorxiv.org/repository/view/15591/
AB - Natural ecosystems are increasingly impacted by multiple global-change drivers, and these impacts propagate non-linearly from individuals to populations and to communities. To anticipate ecological change under such complexities and mitigate negative impacts, we require integrated, mechanistically-grounded approaches to prediction. For mitigation strategies to be adaptive, predictions must also be updated efficiently as new evidence becomes available. In many protected areas, decades of heterogeneous monitoring data are available, but remain siloed within research groups or projects and are rarely turned into integrated forecasts. Here we present an ecological digital twin of Doñana Natural Area, a Mediterranean biodiversity hotspot under increasing climate and land-use pressures. The digital twin combines long-term biodiversity data with weather, hydrological and remote-sensing products. This data is then used to link demographic, abundance and occurrence models to interspecific interactions, which results in community and ecosystem dynamics. Annual forecasts are generated under global-change scenarios, and uncertainty is explicitly carried through every stage of this chain, from driver to outcome. The digital twin is hosted open-access on the LifeWatch ERIC e-infrastructure, thus ensuring transferability and continuity. Annual evaluation of forecast skill against benchmark models, open forecast submissions by different user groups, and a web interface support continuous updating and decision-making. Importantly, the digital twin is also built on standardized workflows, allowing new models, species, and data streams to be incorporated rapidly and consistently. This design makes the approach directly scalable: the same infrastructure can be extended to other protected areas that already hold numerous ecological datasets but currently lack the integrated products needed to translate that data into actionable ecological knowledge.
ER -