Error Analysis Of Sea Surface Salinity Data From Satellite Microwave Remote Sensing And Correction Of Land Contamination

Ocean salinity is the key component of ocean dynamics and has influence on water cycle,atmospheric circulation and global climate.Although Sea surface salinity(SSS)using microwave remote sensing has achieved the application requirement in open ocean with the launch of Soil Moisture and Ocean Salinity(SMOS)and Aquarius/SAC-D,significant errors still exist in some areas near the coast and the two poles.Based on the sea surface microwave radiation model and the SSS retrieval algorithm,this thesis analyzes the accuracy of SSS retrieval,the measurement method and data products of SSS from satellites,and the influence and correction of land contamination.The main contents include the following aspects.The influence of error sources on SSS accuracy including sea surface temperature,sea surface wind,foam,and TB measurement is analyzed through the sea surface microwave radiation model and the SSS retrieval algorithm.Then the simulation of SMOS accuracy verify that the SSS measurement can achieve an accuracy of 0.2 psu or 0.1 psu when retrievals are averaged over 10 days with spatial resolution of ??11 ? or ??22 ?.In the next,the global SSS is retrieved using SMOS-L1 c TB data and the corresponding auxiliary data and compared with SMOS-L2 SSS data.The comparison shows that about 58% data differences are less than 0.1psu and the cause of the difference is also presented.Both SMOS and Aquarius are missions for observing SSS using L-band radiometer.SMOS carries a two-dimensional Microwave Imaging Radiometer Aperture Synthesis(MIRAS)while Aquarius uses active and passive combined method with three real aperture radiometers and a scatterometer.The differences of their payloads lead to the different data processing method and data products.The comparisons are applied to L3 data products between SMOS and Aquarius and then with the in-situ salinity data from Argo float.It is found that,the data correlation between two satellite products is improved through temporal-spatial averaging.The comparisons with Argo data show that the accuracy of Aquarius is higher than SMOS in global scale.With respect to SSS accuracy in different regions,SMOS outperforms Aquarius in open ocean while achieves much lower accuracy in coastal areas.One of the main researches of this thesis is the impact of land contamination on TB measurement.For SMOS data,significant errors exist in coastal areas because of the contamination by the nearby land.It is found in this thesis,after the reducing of the influence from sea surface temperature,wind speed,and polarization,the SMOS brightness temperature(TB)data in coastal areas reveals a decreasing trend with the increase of the distance to coast(DC).A SMOS TB measurement model according to parameters of geophysics and SMOS/MIRAS antenna array is established to simulate how land contamination affects TB.Within 40 km of DC,TB error due to land contamination is large since land is warmer than ocean and is observed by the mainlobe of the antenna array.TB error decreases rapidly with the increase of DC because the mainlobe is departing from land to ocean.When DC exceeds 40 km,TB error is still higher than zero since the near sidelobes are still observing land and still makes considerable contribution to TB.TB error decreases smoothly from 40 km to 400 km since the near sidelobes are departing from land.When DC exceeds 400 km,TB error approaches zero smoothly since the far sidelobes are low and have less influence on TB.The land contamination analysis in thesis letter is used for the correction of SMOS TB and SSS data in coastal areas of South China Sea(SCS).First,the variations of SMOS TB data with respect to DC and incidence angle in coastal area of SCS are analyzed.Since SCS belongs to area which is contaminated by radio frequency interference(RFI),TB outliers within this area are counted.TB outliers within this area which exist in DC exceeds 40 km are considered as contamination by RFI are counted and then abandoned.Then,TB contribution from land nearby SCS is calculated according to the geophysical parameters of land near SCS.Finally,the correction values caused by land contamination for SMOS TB in coastal area of SCS are calculated and the corrected TB data are applied for SSS retrieval.After the correction of land contamination,the accuracy of SSS in coastal area of SCS improves from 1.31 psu to 0.67 psu.