Retrieval Method Of Path-integrated LWC For Airborne Upward-Looking Microwave Radiometer Using Cloud Model

A new retrieval method for deducing path-integrated cloud liquid water content(L) by airborne upward-looking microwave radiometer is developed in this paper.The coefficients of retrieval equation are obtained through statistic regression methodlike former but the statistical samples are produced by the one-dimensional stratiformcloud mode 1 (called method 1). This method is compared with the former one(called method 0) in which historic radiosonde data is employed to determine thevertical distribution of cloud liquid water content by testing for relative humidityabove some threshold value. The case of 8 July 2001 in Jilin shows that as thedynamic and microphysical processes of stratiform cloud are considered, thestatistical samples produced by the cloud model are consistent with the synopticobservations much better than method 0 and the retrieval errors caused by theuncertainty of the background atmospheric condition (including the verticaldistribution of vapor and oxygen) and the vertical distribution of cloud liquid watercontent decrease to some extent. As a result, the relative accuracy of retrieved valueof L is improved at all altitudes by 0.15%～7.28%.The sensitivity test of the one dimensional stratiform cloud model shows thatthe statistical samples produced by the cloud model approximately cover all thepossible profiles of the cloud liquid water content. But the actual profiles during thefield observation are only part of them. Besides, the factor analysis of the case of 8July 2001 in Jilin indicates that the errors caused by the uncertainty of the height ofthe cloud base and the vertical distribution of cloud liquid water content can not beignored when compared with the errors caused by the background and the instrumentdrift errors. Under most of the circumstances the former ones are much greater thanthe latter ones. Hence, In order to decrease the retrieval errors caused by the above-mentioned factors, we use selected cloud samples which are supplied by cloudmodel to regress retrieval coefficients according to the synoptic observation (calledmethod 2). The numerical simulation results of the case of 8 July 2001 in Jilin showthat method 2 makes a further improvement of relative accuracy of retrieved value ofL at all altitude by 0.8％～5.9％when compared with method1. This indicates thatusing independently observed information of cloud to restrain sample 1 can furtherimprove the retrieval accuracy.The cases of 8 July and 9 July in 2001 and 22 April in 2002 in Jilin show thatthe statistical samples of method 2 are consistent with the synoptic observations quitewell and the retrieval accuracy increases by 1.9％～37.2％when compared withmethod 0.This improvement is meaningful when compared with the instrument drifterrors. Besides, the horizontal distributions of L are correlated with the distributionsof PPI echo intensity of surface weather radar very well.