Merging of rain rate estimates from opportunistic sensors and geostationary satellites

• Applicants: Dr. Christian Chwala; Dr.-Ing. Uwe Siart
• Subject Area: Atmospheric Science ; Hydrogeology, Hydrology, Limnology, Urban Water Management, Water Chemistry, Integrated Water Resources Management ; Communication Technology and Networks, High-Frequency Technology and Photonic Systems, Signal Processing and Machine Learning for Information Technology
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 541482432

The implementation plan of COP27 made a very clear statement. “One third of the world, including sixty per cent of Africa, does not have access to early warning and climate information services”. This is in particular true with regard to rainfall related warnings and climate information. The reason for this is the almost complete absence of weather radars on the African continent and the lacking density of rainfall measurement stations. In contrast, Geostationary satellites (GEOsat) and potentially also commercial microwave links (CMLs) and satellite microwave links (SMLs) are available in near real-time and can be used to derive rainfall estimates. However, quantitative precipitation estimation (QPE) from GEOsat data is challenging due to the very indirect relation between rain rate and the actual measurements which are carried out in the visible and infrared spectrum. For QPE from SML and CML data, in particular based on large-scale CML studies in Europe, it has been shown that it can be on-par with QPE from radar and rain gauges. In the absence of reference data, as it is often the case in developing countries, the existing tailored semi-empirical processing methods lack applicability, though. GEOsat data has the potential to support CML/SML processing in these regions, and vice-versa, CML/SML QPE could be used to adjust GEOsat QPE. The overarching goal of the project MERGOSAT is therefore to develop novel methods for the generation of improved near-real-time rainfall maps for data-scarce regions via a combination of data from GEOsat and opportunistic ground-based sensors, namely CMLs and SMLs. To reach this goal, we will focus on three different objectives: 1) Create a basis for more generic CML/SML QPE models by improving the understanding of the processes that affect CML and SML EM propagation focusing on wet antenna (WAA) and drop size distribution (DSD), including the distinction between liquid and melting hydrometeors along slanted paths. 2) Develop adequate CML/SML QPE models applicable in data-scarce regions, building on the new insights on WAA and DSD and utilizing GEOsat data in an innovative manner to constrain the models. 3) Improve GEOsat QPE with DeepLearning methods and develop a novel way to allow merging with CML/SML data at sub-hourly resolution. We will base our research on our large archive of CML data, also from Africa, and the increasing availability of SML data. Additional data from new field experiment setups will be combined with cutting edge simulations of EM propagation. In addition we will leverage the newest advancements of DeepLearning in hydrometeorology and our high-performance compute infrastructure. Combined with the extended capabilities of the recently launched Meteosat Third Generation GEOsat, that will allow us to successfully tackle our challenging objectives and produce the methodological basis needed for supplying data-scarce regions with improved near-real-time rainfall information.

• DFG Programme: Research Grants
• International Connection: Czech Republic
• Partner Organisation: Czech Science Foundation
• Co-Investigator: Professor Dr. Harald Kunstmann
• Cooperation Partner: Dr. Vojtech Bares