A machine learning framework for estimating the probability of blacklegged tick population establishment in eastern
Hamid Ghanbari1,2, Kevin Siebels1,2, Ariane Dumas1,2
1Modelling Hub Division, Applied Public Health Sciences Directorate, Science and Policy Integration Branch, Public Health Agency of Canada, Quebec, Canada.
Plos One
|September 22, 2025
Summary
Climate change is expanding tick ranges. Machine learning models using Earth observation data accurately predict blacklegged tick (Ixodes scapularis) establishment, aiding Lyme disease risk mapping.
Area of Science:
- Ecology and Environmental Science
- Epidemiology
- Geospatial Analysis
Background:
- Ixodes scapularis ticks, vectors of Lyme disease (LD), are expanding into Canada due to climate change.
- Effective public health strategies require accurate prediction of tick population establishment.
Purpose of the Study:
- To develop and evaluate a machine learning (ML) framework for estimating blacklegged tick population establishment probability.
- To identify key environmental predictors from Earth observation (EO) data influencing tick establishment.
Main Methods:
- Utilized a comprehensive ML framework integrating active tick surveillance data.
- Derived environmental predictor variables from EO data at multiple spatial scales.
- Evaluated various ML algorithms, with XGBoost showing optimal performance.
Main Results:
- XGBoost model achieved high sensitivity (0.83) and specificity (0.71) in predicting tick establishment.
- Optimal prediction performance was achieved using environmental data within a 1 km radius of surveillance sites.
- Key predictors included temperature variables (degree-days, maximum temperature), soil properties (silty, SOC, pH), and land cover (broadleaf/mixed forests, low urban areas).
Conclusions:
- Machine learning integrated with open-access EO data provides an accurate and updatable method for LD risk mapping.
- This approach supports public health management of Lyme disease and other tick-borne diseases in a changing climate.
