Conference Agenda

Multiple global water cycle related satellite data products (soil moisture, vegetation optical depth, landscape freeze/thaw, snow water equivalent, ocean salinity, precipitation etc.) are available and explored by a growing community. However, very significant discrepancies among these different satellite products have been reported. The space observation of the Cryosphere needs new instruments and tools, while the global mapping of soil moisture and ocean salinity needs to be continued. In addition, the temporal-spatial resolution and accuracy of different satellite data, including the ESA Soil Moisture Ocean Salinity (SMOS, single L-band and multiple incidence angles) and the Chinese Fengyun series satellites (multi-frequency at a single incidence angle), needs to be refined for a wider global water cycle study. This project is dedicated to improving the accuracy and temporal-spatial resolution of current remote sensing products related to water cycle, including soil moisture, vegetation optical depth, landscape freeze/thaw, snow wetness and water equivalent etc., through the synergy use of multi-sources satellite observations from European and Chinese Earth observation data. It is aimed to enhance the retrieval performance through the development of radiative transfer modelling and new algorithms. Meanwhile, new satellite missions should be studied to combine the advantages of current satellite design, and continue the multi-frequency microwave observation from space.

The demand for comprehensive, continuous and reliable sea ice information from multi-source satellite data is increasing due to climate change and its impact on the environment, regional weather conditions and human activities. This paper provides an overview of our Dragon 5 project, which deals with synergistic monitoring of arctic sea ice using multi-source remote sensing data. The first topic is the implementation and development of sea ice concentration (SIC) retrieval and SIC noise reduction algorithms applied to data from the Chinese FY-3C Microwave Radiation Imager (MWRI). The SIC bias of the FY3C is 4% with dynamic tie points when compared to in-situ measurements. We also investigated whether the new AMSR2 radiometer thin ice detection algorithm could be applied to the MWRI data. The results were promising and the accuracy of MWRI thin ice detection can be improved with further work. Secondly, the ability to discriminate sea ice types at small angles of incidence provided by CFOSAT SWIM data was investigated and analysed. For sea ice thickness (SIT), we proposed a long short-term memory (LSTM) deep learning snow depth retrieval method based on the brightness temperature of AMSR-2, and then the snow depth product was used to retrieve SIT from CryoSat-2 data. Verification with OIB SIT shows that the LSTM SIT retrieval is superior to other SIT products. We also produced high reliability sea ice drift products from Chinese FY-3 and HY-2 satellite radiometer data and compared them with SSMIS and AMSR2 products. The four microwave radiometers provided relatively consistent measurements of sea ice drift. Higher resolution sea ice drift and deformation fields were retrieved from Sentinel-1 SAR imagery and used for first tests of short-term forecasts of regional sea ice conditions. Another effort was the development of robust automated methods for SAR-based iceberg detection in complex backgrounds (open ocean and sea ice). To this end, different algorithms were applied (e.g. various CFAR detectors), using both C- and L-band radar imagery. The results show good detection capability even for icebergs that are only slightly larger than the spatial resolution of the SAR system.