Retrieval of ice water path from the FY-3B MWHS polarimetric measurements based on deep neural network

• DOI: 10.5194/amt-2022-2
• Publication Date: 2022-01-28T08:34:03Z
• Authors: Wenyu Wang; Zhenzhan Wang; Qiurui He; Lanjie Zhang
• Keywords: 13. Climate action; 01 natural sciences; 0105 earth and related environmental sciences

Abstract. Ice water path (IWP) is an important cloud parameter in atmospheric radiation, and there are still great difficulties in retrieval. The artificial neural network is a popular method in atmospheric remote sensing in recent years. This study presents a global IWP retrieval based on deep neural networks using the measurements from Microwave Humidity Sounder (MWHS) onboard the FengYun-3B (FY-3B) satellite. Since FY-3B/MWHS has quasi-polarization channels at 150 GHz, the effect of polarimetric radiance difference (PD) is also investigated. A retrieval database is established using collocations between MWHS and CloudSat 2C-ICE. Then two types of networks are trained for cloud scene filtering and IWP retrieval, respectively. For the cloud filtering network, using IWP of 10 g/m2 and 100 g/m2 as the threshold show the filtering accuracy of 86.48 % and 94.22 % respectively. For the IWP retrieval network, different training input combinations of auxiliary information and channels are compared. The results show that the MWHS IWP retrieval performs well at IWP > 100 g/m2. The mean and median relative errors are 72.02 % and 46.29 % compared to the 2C-ICE IWP. PD shows an important impact when IWP is larger than 1000 g/m2. At last, two tropical cyclone cases are chosen to test the performance of the networks, the results show a good agreement with the characteristics of the brightness temperature observed by the satellite. The monthly MWHS IWP shows a good consistency compared to the ERA5 and 2C-ICE while it is lower than MODIS IWP.