Study On The Effect Of Different Radiative Transfer Models In The Yellow Sea Fog Satellite Data Assimilation

There are three main issues that are important to the WRF numerical predictionof Yellow Sea fog: quality of initial field, boundary layer (PBL) and the choice ofmicrophysics schemes. This paper attempts to improve the quality of sea fog’s initialfield by assimilating satellite radiation data, given the severe lack of the routine seaobservation data. However, before using data assimilation to improve initial field, weneed to determine the optimal microphysics and boundary layer schemes.Ten typical widely-spread sea fog cases in the Yellow Sea were chosen as theresearch object, with the aid of WRF (Weather Research and Forecasting) model andits3DVAR (three dimensional variational) data assimilation module to carry out theresearch. First we conducted some sea fog numerical forecast tests under differentmicrophysical and boundary layer schemes, and then we analyzed their pros and consaccording to the MTSAT (Multi-functional Transport Satellite) observations. Second,based on the aforementioned tests, the YSU boundary layer and Lin microphysicsschemes were selected,12typical Yellow Sea fog processes were carried out, and theeffects of the radiative transfer models CRTM and RTTOV on the assimilation ofATOVS satellite radiation data were investigated and compared.Microphysics and boundary layer schemes’ experimental results show that:1) Nomatter it is a statistical average, or a single case, PBL scheme has much greater impacton forecasted sea fog area while microphysics scheme has less.2) The bestcombination of PBL and microphysics schemes is YSU and Lin, and the worst isMellor-Yamada and WSM5.3) The YSU and the MYNN schemes are the optimalboundary layer schemes, whereas the Mellor-Yamada and the QNSE schemes are theworst. The selection of the YSU scheme and the MYNN scheme should be determined on a case-by-case basis.4) Mellor-Yamada, QNSE boundary layerschemes have stronger turbulence near the sea surface which leads to a higherboundary layer that impedes the development and maintenance of the sea fog. Whilstthe intensity of turbulence and the height of the boundary layer delineated by theMYNN and YSU schemes are favorouble for its development and sustaining.5)Microphysical schemes’ effect to WRF model depends on the number of droplets andthe given initial value.According to the comparative analysis of radiative transfer models CRTM andRTTOV’s assimilation effects, we get the results:1) Assimilating several satellites’radiance data is much better than assimilating only one satellite’s data.2) The effectof assimilating radiance data from RTTOV is much better than that by the CRTMmodel, due to the improved initial fields of temperature, humidity and wind, and itsprocess is more stable.