ERADACS is a novel, multi-disciplinary project that brings together: - a light-weight but powerful forecast system, state-of-the-art land surface data assimilation using NASA soil moisture data - new methods for visualizing and communicating forecasts, co-developed with farmers in Africa - socioeconomic studies on building resilience through early warning and incentivizing action on forecasts The project will be led by the University of Reading and carried out in partnership the Kwame Nkrumah University of Science and Technology (KNUST), a leading Ghanaian university; and Evidence for Development, a well-established NGO. The impact of agricultural drought on the world's poorest communities can be devastating. Changes in weather patterns due to changing climate can mean that decisions about when to plant and how to manage crops based on experience of weather in recent years are not reliable. Timely forecasts of the developing likelihood of agricultural drought have the potential to have a significant positive impact on the lives of small-scale farmers and their surrounding communities during such events. The ERADACS project will develop a forecast system for agricultural drought using multiple streams of satellite data, the Met Office land surface model (JULES) and state-of-the-art mathematical techniques (Data Assimilation) to combine these. The resulting forecast will use climatological information from the 30 year TAMSAT data record to predict likely trajectories of rainfall. These seasonal forecasts will be produced across Ghana and made available openly via the TAMSAT website. Beyond the lifetime of the ERADACS project we will sustain the forecast system via the operational TAMSAT platform. A key aspect of the ERADACS project will be a pilot of how this information might be useful to specific communities in Ghana that are reliant on subsistence farming practices. We will visit and collect socioeconomic information from a number of communities to establish their vulnerability to agricultural drought as well as using serious game play to elicit likely responses. Our forecasts, and their likely uncertainties, will be discussed with the farmers and we will trial different methods of presenting this data. We will use feedback from these groups to refine the means of communication and to tailor the information produced from the forecast system. The work in Ghana will be carried out by KNUST and Evidence for Development, both of whom have many years experience of capacity building in Africa.