Data Observation Error Estimation Based On WRF/GSI Assimilation System And Its Impact On Typhoon Prediction

Under the background of global change,typhoon disasters are more and more serious.Therefore,it is particularly urgent to improve the prediction of typhoon.The rapid development of data assimilation technology improves the data application in typhoon forecast,but it still lacks the quantitative assessment of the estimation of data observation error and its influence on typhoon forecast.This paper used the mesoscale WRF(Weather Research and Forecasting)model as the forecasting model and interconnected GSI(Gridpoint Statistical Interpolation)as the assimilation system.This paper took Super Typhoon "Soudelor" in 2015 as an example,and the experiments of the influence of the different meteorological elements and observation errors estimation in the conventional observation data are carried out.At the same time,the observation error of the satellite retrieved temperature and humidity profiles is estimated by using the radiosonde data.The following are the conclusions of this paper:Based on the interconnection of WRF model and GSI system,the 3DVAR method is used to carry out cyclic assimilation.The analysis field which changes after GSI assimilation is put into the WRF model for 72-hour prediction experiments.The assimilation experiment of conventional observation data observation error estimation shows that the change of observation error of temperature and humidity data in the assimilation experiment makes the deviation less than before.The assimilation result has positive influence on the path and strength of the simulated typhoon,and the accurate description of the observation error will affect the assimilation results.The observation error estimation of the satellite retrieved temperature and humidity data reveals that the AIRS retrieved temperature profiles are in good agreement with the radiosonde data,whilst the AIRS retrieved relative humidity profiles show the phenomenon of wetter in higher layers and drier in lower layers.The RMSE difference of temperature and relative humidity range from 1.02? to 2.49? and from 12.91%to 23.43%,respectively.At the same time,the retrieved error estimation under different cloud coverage can provide a guarantee for the assimilation of temperature and humidity data and the application of cloud affected satellite data in the future.