Where the fluke meets the snail: climate, ensemble models, and fasciolosis risk in Ukraine
Abstract
Fasciolosis, caused by the liver fluke Fasciola hepatica, is an economically important zoonotic disease that is endemic in Ukraine, yet the environmental drivers of its distribution remain insufficiently understood and no national-scale risk model has been produced. We compiled occurrence records for F. hepatica and its intermediate snail host Galba truncatula and used an ensemble species distribution modelling framework to map their potential distributions across Ukraine. Climatic predictors were drawn from the CliMond database and screened for collinearity before modelling. Several machine learning algorithms were tuned, evaluated, and combined into weighted ensembles, and the resulting suitability surfaces were compared spatially to identify areas of agreement and divergence. SHAP analysis (SHapley Additive exPlanations) was then applied to both ensembles to quantify and compare the climatic drivers of the parasite and its host. The F. hepatica model performed well, whereas the G. truncatula model was considerably weaker, indicating that climate alone does not fully explain the snail's distribution at the considered resolution. Despite this, the two suitability surfaces were strongly and positively correlated, pointing to a significant spatial concordance between parasite and snail host suitability. Temperature seasonality emerged as the dominant climatic driver for both species, and permutation tests showed that its net directional effect was similar for the two, while its overall magnitude of influence was significantly stronger for the parasite. Both species showed unimodal responses to temperature seasonality, with intermediate values most favourable; the parasite exhibited a stronger but narrower optimum, whereas the snail tolerated a broader range before the effect turned negative. Overlay analysis identified joint high-risk zones in western and northern Ukraine, parasite-only zones in the north, and snail-only zones in the south that represent potential introduction areas. No region was entirely free of climatic suitability for at least one of the two species. These results indicate that fasciolosis risk in Ukraine is concentrated in the west and north, and that the parasite's distribution is not simply a subset of its host's. Combining ensemble SDMs with explainable AI provides a transferable approach for understanding the environmental drivers of host-dependent zoonoses and for targeting surveillance and control.
Related
- Exotic catenulid flatworms (Platyhelminthes, Catenulida) and where to find them in a temperate climate – a field study in a botanic garden
- Fluke-borne viruses are a risk factor and diagnostic target for diseases caused by carcinogenic trematodes
- Vaccine antigens against fasciolosis in ruminants: a systematic review of antigen candidates, study characteristics, outcome measures, and reporting quality in cattle, sheep, and goats
- First report on chromosome number and genome size of rare endemic birch Betula klokovii Zaverucha from Ukraine
- Changes in wildlife activity patterns in response to war in Ukraine
BibTeX
@misc{tytar2026where,
title = {{Where the fluke meets the snail: climate, ensemble models, and fasciolosis risk in Ukraine}},
author = {Volodymyr Tytar},
howpublished = {EcoEvoRxiv},
year = {2026},
month = oct,
doi = {10.32942/x2111n},
url = {https://ecoevorxiv.org/repository/view/15613/},
note = {CC BY-NC 4.0},
}
RIS
TY - UNPB
TI - Where the fluke meets the snail: climate, ensemble models, and fasciolosis risk in Ukraine
AU - Volodymyr Tytar
PY - 2026
DA - 2026/10/22
PB - EcoEvoRxiv
DO - 10.32942/x2111n
UR - https://ecoevorxiv.org/repository/view/15613/
AB - Fasciolosis, caused by the liver fluke Fasciola hepatica, is an economically important zoonotic disease that is endemic in Ukraine, yet the environmental drivers of its distribution remain insufficiently understood and no national-scale risk model has been produced. We compiled occurrence records for F. hepatica and its intermediate snail host Galba truncatula and used an ensemble species distribution modelling framework to map their potential distributions across Ukraine. Climatic predictors were drawn from the CliMond database and screened for collinearity before modelling. Several machine learning algorithms were tuned, evaluated, and combined into weighted ensembles, and the resulting suitability surfaces were compared spatially to identify areas of agreement and divergence. SHAP analysis (SHapley Additive exPlanations) was then applied to both ensembles to quantify and compare the climatic drivers of the parasite and its host. The F. hepatica model performed well, whereas the G. truncatula model was considerably weaker, indicating that climate alone does not fully explain the snail's distribution at the considered resolution. Despite this, the two suitability surfaces were strongly and positively correlated, pointing to a significant spatial concordance between parasite and snail host suitability. Temperature seasonality emerged as the dominant climatic driver for both species, and permutation tests showed that its net directional effect was similar for the two, while its overall magnitude of influence was significantly stronger for the parasite. Both species showed unimodal responses to temperature seasonality, with intermediate values most favourable; the parasite exhibited a stronger but narrower optimum, whereas the snail tolerated a broader range before the effect turned negative. Overlay analysis identified joint high-risk zones in western and northern Ukraine, parasite-only zones in the north, and snail-only zones in the south that represent potential introduction areas. No region was entirely free of climatic suitability for at least one of the two species. These results indicate that fasciolosis risk in Ukraine is concentrated in the west and north, and that the parasite's distribution is not simply a subset of its host's. Combining ensemble SDMs with explainable AI provides a transferable approach for understanding the environmental drivers of host-dependent zoonoses and for targeting surveillance and control.
ER -