Research On Key Technologies Of Data Assimilation Of High-Resolution Satellite Observations

With the rapid development of geostationary observation technology,the high-resolution satellite remote sensing instruments have become the most important measurements to obtain meteorological and hydrological information.These instruments provide high-resolution global observation information about atmosphere,ocean,surface and other meteorological and hydrological observation information.The high-resolution observations accounted for the largest percentange of the total data in the current meteorological and hydrological centers,such as AIRS and IASI have accounted for 60% in the European Medium Weather Forecasting Center(ECMWF).Further more,compared the contribution to the numerical weather prediction(NWP)from all kinds of observations,the high-resolution satellite observation has the largest impacts,exceeding the conventional observations such as radiodounde,aircraft and surface conventional observations and so on.In the last more a decade,the number of high-resolution satellite observations assimilated in the NWP systems has increased extensively and made continuous improvement to the initial field.This contributes very significantly to the improvement of the NWP accuracy.Although the assimilation methods and technichs have developed obviously for the high-resolution sattlelite observations in the past yeas,the assimilated data are only a small propotion of the total original high-resolution observations.There are still many challenges for assimilating the high-resolution observations.For example,the infrared high-resolution spectrame is hard to be assimilated as it is easy influenced by the cloud and the dimesion of the observation data is too large to preserve and calculate efficiently,the assimilation methods maybe not suitable for synthetic aperter radar(SAR)sea surface wind which provide observations for the wind parameters used in the forecast model.Therefore,we shoud be pioneering in bring benefits from the high-resolution satellite observations to operations by more advanced assimilation and data processing methods.In this paper,the data assimilation theories,the methods and the application status for the high-resolution stallelite observations are introduced,and the key application difficulties in the assimilation system for two typical high-resolution observations,the infrared hyper spectrum and SAR,are researched.The main works and contributions of this dissertation are as follows:As the cloud can have a grate effect on the thermal intared radiation spectrum,and the completely clear fields of view(FOV)are very few in the global scenes,the existing assimilation system is difficult to assimilate the infrared hyperspectral observations in the cloudy area.Based on the cloud detection method proposed by McNally and Watts(MW),the clear channel cloud detection algorithms MW and LMW are desighed and implemented in three-dimensional variational data assimilation for the Weather Research and Forecasts(WRFDA)system for the IASI cloud detection.All the IASI observations are assimilated after quality control,variational bias correction and cloud detection and the impact of the data assimilation on the forecasts of the super typhoon Hongxia(1506)and Molandi(1614)are assessed.The results from both the typhoon experiments are similar and indicate that the cloud detection influences the assimilation of the IASI observations very much.For the super typhoon Hongxia,the MW cloud detection scheme retained just 16.2% the number of observations by the large-threshold LMW cloud detection scheme and 9.2% of that by the MMR cloud detection scheme for the high-level channel 299,and 3.3% and 2.6% respectively for the underlying channel 921,but the analysis affected by the MW cloud detection scheme reduces the track error of typhoon Hongxia for the first 72 hours most remarkably and improves the path forecast most accurately.Generally the assimilation of IASI observations improves the skills of typhoon forecast.The high resolution observations of huge volume occupy a large storage space on-board and the ground data application center and consume a lot of transmission bandwidth and calculation overhead.Here,a workflow has been propsed that makes the dimension reduction on-board for high resolution observations based on the principal component(PC)method.The PCs are accurately accessed by the information entropy.Also the compression ratio and transmission efficiency has been tested,and the accuracy of the reconstructed radiance has been compared for the PC scores of different numbers.The noise of the radiance reconstructed from a few PC scores reduced significantly with the main information is preserved.This method is suitable for the exploitation the complete spectrum information.Systematically analyzes the transformation from the numerical model variable space,the spectral radiance space and the principal component space for the infrared hyperspectral,and a PC based optimal four — dimensional variational framework is designed.The equivalent relationship between the reconstructed radiance and the principal component under certain preconditions is demonstrated,and the error source of reconstructed radiance is analyzed.The frame for the milation of optimal reconstructed radiance is designed,and it is modified to be a sub — optimal observation operator and obersation error(H-R)frame for the assimilation the reconstructed radiance,and the sub — optimal H — R framework is implemented in WRFDA system.The observations from the area of typhoon ?Hongxia? in 2015 were reconstructed,and the reconstructed radiance of the typhoon region is assimilated in the numerical simulation experiment.The results of reconstructed radiance are accesed.High-resolution synthetic aperture radar(SAR)wind observations provide fine structural information for tropical cycles and could be assimilated into numerical weather prediction(NWP)models.However,in the conventional method assimilating the u and v components for SAR wind observations(SAR_uv),the wind direction is not a state vector and its observational error is not considered during the assimilation calculation.In this paper,an improved method for wind observation directly assimilates the SAR wind observations in the form of speed and direction(SAR_sd).This method was implemented to assimilate the sea surface wind retrieved from Sentinel-1 synthetic aperture radar(SAR)in the basic three-dimensional variational system for the Weather Research and Forecasting Model(WRF 3DVAR).Furthermore,a new quality control scheme for wind observations is also presented.Typhoon Lionrock in August 2016 is chosen as a case study to investigate and compare both assimilation methods.The experimental results show that the SAR wind observations can increase the number of the effective observations in the area of a typhoon and have a positive impact on the assimilation analysis.The numerical forecast results for this case show better results for the SAR_sd method than for the SAR_uv method.The SAR_sd method looks very promising for winds assimilation under typhoon conditions,but more cases need to be considered to draw final conclusions.This paper studies in depth the difficulties of the assimilation of high resolution satellite observations,especially the cloud detection,the data dimension reduction,the assimilation method in multi-parameter space for infrared hyperspectral data,and the assimilation method and the quality control for SAR sea surface wind.These research results can be used for reference in the application of the high resolution satellite observations in China.