MaxHop

MaxHop Run

Inspiration

We were inspired by Sustainable Development Goals #2 (Zero Hunger) and #13 (Climate Action). Climate change has increased the likelihood that locust swarm outbreaks will occur more frequently and threaten the global food supply. MaxHop will help address this problem by predicting the probability of locust presence, which can then be used to determine early intervention and deterrence. MaxHop is based on a paper titled, “Prediction of breeding regions for the desert locust Schistocerca gregaria in East Africa”.

What it does

The Jataware locust model predicts the probability of locust presence in Africa using a maximum entropy model over environmental data such as climate and soil texture. Our model provides a command line interface which allows users to manipulate environmental input conditions against a pre-trained model to predict future locust distributions. We attempted to adapt MaxHop to the ASDI dataset, CMIP6 GCMs. Training MaxHop on the CMIP6 GCMs would produce forecasts for locust presence from 1980-2100 based on the climate simulations from CMIP.