Cloud Detection For Microwave And Hyperspectral Infrared Radiances And Its Application In Typhoon Monitoring

The Fengyun(FY)-3A/B/C/D microwave temperature sounder-2(MWTS-2)is similar to Advanced Microwave Sounding Unit-A(AMSU-A)except for lacking two window channels located at 23.8 GHz and 31.4 GHz.This makes a clear-sky data determination challenging for MWTS-2 due to an unavailability of cloud liquid water path(LWP)retrievable from the two window channels.In this study,a clear-sky data selection algorithm is developed for FY-3C/D Microwave Temperature Sounder-2 based on a high correlation between the differences between observations and background simulations with both the latitudinal bias(?(?))and scan bias(?(?))subtracted,O-B-?(?)-?(?),of the MWTS-2 50.3-GHz channel 1(or equivalently AMSU-A channel 3)with the liquid water path.A data is defined as a clear-sky point if(i)all points within its 60-km radial distance satisfying the criteria of O-B-?(?)-?(?)< 2K;or(ii)it is within a 100-km radial distance from a “cloudy” point defined by O-B-?(?)-?(?)>2K that the mean value of O-B-?(?)-?(?)in the 100-km circle is less than2 K.Compared with LWP,more than 70%(95%)of the clear-sky(cloudy)data points are successfully identified for both AMSU-A and MWTS-2 observations.This success rate of clear-sky data selection is 20% higher than a previous scheme that removing all data points within a 60-km radial distance from a “cloudy” point.We also show that the MWTS-2channel 1 O-B-?(?)-?(?)well reveals the the eye,eyewall and spiral rainband structure of super typhoon Halong(2014).Cloud detection remains a challenging task when assimilating hyperspectral Cross-track Infrared Sounder(Cr IS)observations,for which clear channels above and not contaminated by cloud must be identified.By pairing Cr IS longwave and shortwave channels with weighting function(WF)peaks and cloud insensitivity levels at similar altitudes,a set of cloud emission and scattering indices(CESIs)is generated using the European Centre for Medium-Range Weather Forecasting analyses as the training dataset.Using three CESIs located at around 200(CESI-5),400(CESI-9),and 1,085 h Pa(CESI-19)after removing latitudinal and scan-dependent biases,we show that CESI-5 is strongly(weakly)affected by those ice clouds with cloud-top pressures(CTPs)above(below)?200 h Pa.CESI-9 is strongly(weakly)affected by those ice clouds with CTPs above(below)?400 h Pa.CESI-19 is highly correlated with the ice cloud optical depth(ICOD)obtained from the Atmospheric Infrared Sounder(AIRS)retrieval product and is not affected by sun glints.CESI-19 is greater than 5K(6.5 K)in the presence of ice clouds with the probability of correct typing of 88.3%,the false alarm rate is 9.5%,and the leakage rate of 4.5%,during the daytime.