Real time rainfall estimates through neural networks
At present, quantitative data on precipitation is obtained by rainfall measurements and weather radar networks. However, these networks do not cover the total surface of the planet, nor do they provide instantaneous information. The use of geostationary satellites, in orbits that keep them in a fixed position relative to a specific point on the earth, can provide real time data for monitoring rainfall on a much wider scale. The EurainSat project has used neural networks, to develop a new method for simulating complex rain estimation models. Neural networks are particularly effective for predicting events when the networks have a large database of prior examples to draw on. The project has developed UOB net 1.0 combining data obtained via infrared-based techniques and visible observations with data gathered via microwave sensor measurements. These satellite rainfall data help to obtain precise and real time rainfall estimations covering the whole of the earth's surface. This kind of information is highly valuable as it helps to make more precise global scale weather forecasts and enables better monitoring of severe weather phenomena. Estimations are useful for efficient hydrology and water management, and support decision-making in civil protection as they have a direct impact on the monitoring of emergencies and other serious events (flood forecasts and management, search and rescue operations). Forecasts and assessments are also useful for global climate change studies, monitoring of remote areas, agriculture and famine reduction. The potential users of this innovation include satellite exploitation organisations, national and regional weather services and TV weather broadcasters, the international organisations that deal with weather analysis, as well as forecasting and civil protection agencies. Further, the Food and Agriculture Organisation (FAO) of the UN will benefit from the estimations as they will facilitate food production monitoring, especially in developing countries.
