Submitted Abstract
Rain has a significant influence on the global economy. It influences agriculture and the success of harvest, creating inflationary/ deflationary effect on agricultural produce and societal consumption of the same. It creates river and flash floods, which may vanish homes and businesses causing loss of life and property. It can interrupt traffic connections, supplies for electricity, gas, water and medical care bringing economic activity to a standstill. Apart from this indirect economic impact, rain strongly influences the insurance business. Accurate estimation of rainfall would allow insurance companies to calculate the premiums precisely. The measurement of rainfall might appear to be straightfor¬ward at first; however, it is highly variable spatially and temporally, making it very difficult to measure satisfactorily. Therefore, denser measurement networks are required in order to capture this variability.However, due to high operational cost of such infrastructure, only 1,675 rain gauges can be found across 10 million square kilometres of Europe and around 8000 worldwide with satisfactory temporal resolution. There are other solutions for rainfall measurement: rain radars and Infrared Satellite Imagery. Although they have advantage of higher coverage area, but they suffer from high cost and poor temporal accuracy. This motivates the need for a cheap, accurate and innovative technology for rainfall measurement.RAFAEL (Rainfall estimation using signalling data of satellite communication network) develops an innovative technology for this purpose. The main idea of RAFAEL is to extract rainfall information from the signalling data between satellites and satellite ground terminals by employing advanced machine learning techniques. There are more than 300,000 satellite ground terminals across Europe and 2 million worldwide; these can be transformed into reliable and real time rainfall measurement sensors. The signalling data is already aggregated into a single database, eliminating the need to create an elaborate network to feedback such sensor data. Further, replenishment of satellite ground terminals does not add to the database generation cost. Thus, the proposed idea of using signalling data is an efficient alternative to other database generation methods in terms of accuracy and CAPEX/ OPEX.This Proof of Concept project builds on the results of the cooperation between SES and the University of Luxembourg/SnT. The aim is to valorise these results into minimum viable products that deliver reliable and cost effective rain estimation with high resolution targeting among others, the insurance and agriculture sector.