Related Experiment Video
Updated: Dec 24, 2025
05:16
Flying Insect Detection and Classification with Inexpensive Sensors
Published on: October 15, 2014
25.6K
Modelling the monthly abundance of Culicoides biting midges in nine European countries using Random Forests machine
Ana Carolina Cuéllar1, Lene Jung Kjær2, Andreas Baum3
1Division for Diagnostics and Scientific Advice, National Veterinary Institute, Technical University of Denmark (DTU), Lyngby, Denmark. anacarocuellar@gmail.com.
Parasites & Vectors
|April 17, 2020
Summary
This study models Culicoides midge abundance across Europe using machine learning. The models predict vector distribution, aiding in understanding and managing the spread of economically significant livestock diseases.
Area of Science:
- Veterinary Entomology
- Epidemiology
- Geospatial Analysis
Background:
- Culicoides biting midges transmit arboviruses causing significant economic losses in European livestock.
- Predicting vector abundance is crucial for managing the spread of diseases like bluetongue and African horse sickness.
Purpose of the Study:
- To model and map the monthly abundance of Culicoides species across Europe.
- To provide a foundation for predicting vector-borne disease transmission risk.
Main Methods:
- Utilized entomological data from 904 European farms (2007-2013).
- Employed Random Forests machine learning with satellite-derived environmental and climatic predictors.
- Predicted monthly average Culicoides abundance at 1 km² resolution.
Main Results:
- Model performance varied by month and Culicoides species, with lower accuracy in winter.
- The Obsoletus and Pulicaris ensembles showed higher predictive power than Culicoides imicola.
- Models captured broad distribution patterns but lacked farm-level prediction accuracy.
Conclusions:
- Developed models and maps represent a novel approach to understanding large-scale Culicoides abundance variations.
- These models are a crucial first step for R₀ modeling of Culicoides-borne infections continent-wide.
