Sea Surface Current Retrieval Based On ASAR Data

Sea current is one of the most important phenomenon in oceanographic research,and to obtain accurate sea surface current information has significance sense on numerical ocean prediction, marine conservation,maritime military operations,navigation,marine fisheries, global climate change,etc. At present, the informatization capacity of the conventional current observation methods could not meet the growing huge needs for the current information,and the way of the remote sensing retrieval,especially the microwave remote sensing method, is well showing a tempting prospect.In the light of most of the models that microwave remote sensing methods provided are still empirical and exiting large uncertainties,this paper try to find a new model that based on Doppler centriod theory to promote a better current retrieval results which would help to study the current features and the motion rules of the sea surface current.There are four approaches for sea surface current remote sensing retrieval at present,including temporal image series invariant feature method（ MCC and CR method）, radar altimeter method, SAR Doppler centroid method and SAR along-track interferometry method（ATI）.These approaches have their own characteristics and inversionaccuracy,However, from the operational perspective of offshore current retrieval, method that based on SAR measurements are superior to other approaches.In this paper,we study the sea surface current retrieval from the Doppler centroid of the low resolution images based on SAR measurements. The Doppler centroid estimation technology is a significant technology which estimates the accurate center of the Doppler frequency spectrum from SAR raw data,and it has been used in radar imaging fields and terrestrial areas such as object tracking and recognition, while its Doppler centroid theory is relatively less researches and applications in the marine area.In this paper, Doppler shift parameters is picked up from satellite meta data,and satellite attitude error and the impact of the short Bragg waves are corrected.SAR meta data contains Doppler centroid parameters measured from the satellite and estimated by the Doppler centroid estimation technology, known as the measured Doppler frequency(f_Dc). The measured Doppler frequency partly consist of the errors caused by the satellite attitude,known as the predicted Doppler frequency（fDp）, which need to estimate from the SAR raw data in both azimuth and range directions. In addition,the Doppler shift of the Bragg waves caused by sea surface scattering would amplify the results,and thus it need to remove. To verify the retrieval results,current results computed from HYCOM and results given in the reference are compared with the sea surface velocities to confirm the method a better accuracy.After the retrieval,current retrieval methods that based on radar measurements are also summarized and reviewed in the fourth chapter,including robust empirical model based on Radarsat SAR data,error fitting model based on ASAR data, SAR along-track interferometry method（ATI） and Doppler radar imaging model（DopRIM）method.By comparison,we can find that method provided from this paper show a bright prospect on the operational sea surface current retrieval.After those in above,the impact of sea surface wind fields on sea surface current are also discussed.In this paper, we firstly conduct the sea surface wind field retrieval.Wind field is a key factor before conducting the high-resolution currents retrieval.Sea surface wind field can be obtained not only from wind products,but also by wind field retrieval from the backscattering coefficient（NRCS） images.The main idea is to build a geophysical model function（GMF） connecting the relationship between wind field and NRCS,and through NRCS images and GMF wind field can thus be calculated.In this paper, ASAR NRCS images are calculated by image segmentation and 2-D fast Fourier transform（FFT） to obtain a 2-D sea surface frequency spectrum. Since it is perpendicular with spectrum peak direction,wind direction can be easily got,and thus wind direction and scattering coefficients can be put into the CMOD5 mode function to obtain wind speed. There is a 180 degrees ambiguity problem in wind direction retrieval,which need to be combined with otherweather statistics data to eliminate.Sea surface wind field retrieval can not only help to determine the composition of the sea surface waves,but also can be used to study the impact of waves on Doppler shift and its effect on the current distribution, which show a important significance to improve the accuracy of sea surface retrieval.High-resolution current retrieval would be focused for next step in the future research since this paper has finished the current retrieval at low resolution. High-resolution current retrieval is actually a very complex issue that this article can only do some preliminary research on part aspects of it,and more work needs to be continued.