WP2 led by the University of Reading, focuses on translating weather-related risk such as rainfall deficit or heat stress into agricultural risks such as drought and crop yield loss.
This is accomplished by developing a series of algorithms that link meteorological and climate variables with weather-driven agricultural losses. The work is based on novel modelling with the integration of remotely-sensed data products and data assimilation methods.
Combining on-the-ground field observations such as farmer management practices enhances the accuracy of crop loss predictions to better inform risk mitigation scenarios.
Focusing on maize crops in pilot sites in Tanzania, the team has been investigating the impact of factors such as rainfall and temperature changes in relation to region- and crop-specific conditions such as soil conditions and cyclical or seasonal phenomena using a multi-model approach with WRSI, DSSAT, and APSIM/EPIC crop models. Crop model simulations consider drought tolerant and non-drought tolerant maize varieties under short, medium, and long growing periods. Combined with on-the-ground field observations such as farmer management practices, fertilizer use or seed variety selection, the team can enhance the accuracy of crop loss predictions and better inform risk mitigation scenarios.
To further ensure the results are as accurate as possible, WP2 will assess the sensitivity of various models to rainfall inputs. For example, while ARC2 is typically used for index-based insurance to determine when a pay-out is triggered, WINnERS will compare outputs to the recently developed quasi-global CHIRPS dataset. Earth observation datasets that provide additional information on temperature, vegetation, and evapotranspiration will also be examined to better understand the relative importance of rainfall and temperature at different seasonal stages to crop growth. Finally, the team will test new dynamic methods for defining the start of season and length of growing period and will develop new strategies for incorporating this information in weather index-based instruments.