Evaluation of Environmental Drivers and Land-Use Exposure of the Invasive Pest Halyomorpha halys Using Interpretable Ensemble Machine Learning Across Europe

Halyomorpha halys, the brown marmorated stink bug, is a highly polyphagous invasive pest that has spread across Europe since 2004. However, studies of continental-scale climatic suitability modelling and land cover analysis for H. halys remain limited, as the few existing models covering Europe are static and do not relate predicted suitability to land cover. Here, a four-member ensemble species distribution model was fitted for H. halys occurrence data from the Global Biodiversity Information Facility (GBIF) database across Europe for 2020 to 2025, using an initial set of 117 environmental covariates. Prediction accuracy was consistently high for all four algorithms (ROC AUC 0.966 to 0.991) and the combined ensemble (true skill statistic 0.846 to 0.912), and predictions transferred well to later years (Boyce index 0.942 to 0.997), but less so to records more than 50 km from the training records (0.676 to 0.968). Mean annual temperature was the most important covariate in four of six years, with its anomaly in 2023 and human population density in 2022. The suitable area ranged from 10.0% to 34.3% of Europe (11.7% to 23.1% at a common threshold) and was highest in 2020, the year with the fewest and most spatially restricted records. Among 15 land cover classes, mean suitability was highest for urban and built-up land in every year (0.52 to 0.74), whereas croplands (25.7% to 26.1% of land area) reached only moderate suitability (0.16 to 0.37) but formed the largest share of the suitable area (31.6% to 42.2%, against 4.5% to 10.3% for urban land). The predicted niche of H. halys therefore overlaps most strongly per unit area with urban and built-up land, but in absolute area mainly with croplands.

Authors

Institutions

Publication Details

Journal
AI
Published
2026-11-18
DOI
https://doi.org/10.3390/ai7110493
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
NOV
article

Evaluation of Environmental Drivers and Land-Use Exposure of the Invasive Pest Halyomorpha halys Using Interpretable Ensemble Machine Learning Across Europe

Mladen Jurišić, Helena Ereš, Dorijan Radočaj, Ivana Majić
AI
Species Distribution and Climate Change
article

Evaluation of Environmental Drivers and Land-Use Exposure of the Invasive Pest Halyomorpha halys Using Interpretable Ensemble Machine Learning Across Europe

Mladen Jurišić, Helena Ereš, Dorijan Radočaj, Ivana Majić
article en

Abstract

Halyomorpha halys, the brown marmorated stink bug, is a highly polyphagous invasive pest that has spread across Europe since 2004. However, studies of continental-scale climatic suitability modelling and land cover analysis for H. halys remain limited, as the few existing models covering Europe are static and do not relate predicted suitability to land cover. Here, a four-member ensemble species distribution model was fitted for H. halys occurrence data from the Global Biodiversity Information Facility (GBIF) database across Europe for 2020 to 2025, using an initial set of 117 environmental covariates. Prediction accuracy was consistently high for all four algorithms (ROC AUC 0.966 to 0.991) and the combined ensemble (true skill statistic 0.846 to 0.912), and predictions transferred well to later years (Boyce index 0.942 to 0.997), but less so to records more than 50 km from the training records (0.676 to 0.968). Mean annual temperature was the most important covariate in four of six years, with its anomaly in 2023 and human population density in 2022. The suitable area ranged from 10.0% to 34.3% of Europe (11.7% to 23.1% at a common threshold) and was highest in 2020, the year with the fewest and most spatially restricted records. Among 15 land cover classes, mean suitability was highest for urban and built-up land in every year (0.52 to 0.74), whereas croplands (25.7% to 26.1% of land area) reached only moderate suitability (0.16 to 0.37) but formed the largest share of the suitable area (31.6% to 42.2%, against 4.5% to 10.3% for urban land). The predicted niche of H. halys therefore overlaps most strongly per unit area with urban and built-up land, but in absolute area mainly with croplands.

AIVol. 7(11)
University of Osijek (HR)
Sveučilište Josipa Jurja Strossmayera u Osijeku
Life on land, Zero hunger
Species Distribution and Climate Change
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.

Laika access coming soon