The Locust service offers timely information on the locust breeding areas using Machine Learning modelling trained with historical data and operating with weather and earth observation data. The following methodology takes into account particularities of locust life cycle with respect to the occurring environmental conditions during the locust breeding. Forecasting of locust potential breeding areas is of particular importance especially since locust breeding could take place in a very large area.

Locust Potential Breeding Areas Forecast

The Service provides output operational in a monthly frequency covering the area of Africa where desert locust can potentially breed.

The ML model is built around a species distribution approach and the produced output has the form of density maps depicting locust breeding areas for the present time and for six-month forecast. Environmental conditions expressed by various variables are used as input.