Validation Of The Remote Sensing Products Retrieved By Geostationary Ocean Color Imager

The Geostationary Ocean Color Imager(GOCI)is the first ocean color sensor on a geostationary satellite,which have high temporal resolution and be conductive to study the dynamic of sea water.On account of the radiometric calibration of imager,the response and set of bands,the atmosphere rectification algorithm and so on,its ocean color products have uncertainty.Validation is an important mean to evaluate the consistency,stability and uncertainty of remote sensing products objectively,and be of great significance to improve the quantitative level,algorithm,and the application reliability of products.By using in situ data and MODIS data,the satellite-ground validation and cross check analysis are carried out respectively for GOCI ocean color products.Taking Hangzhou Bay and Zhejiang coastal as research area,the satellite-ground data matching criterion is established,and the Rrs,CHL and TSS retrieved by GDPS built-in algorithm are validated via in situ data.Taking MODIS products as reference,the star to star match-up procedure is made,the GOCI Rrs and CHL are cross checked in the clean East China Sea and turbid Zhejiang coastal area respectively,and the error sources and its relationship between error are analyzed.The innovative achievements of the study are mainly reflected in the following aspects:(1)The KOSC algorithm has good applicability in low turbid water area,but its Rrs is underestimated,and fail retrieved in high turbid water.The uncertainty of Rrs in low turbid water decreases with the increase of wave band,while opposite in high turbid water.The general uncertainty of Rrs increases with the increase of TSS.The relative errors of Rrs from visible to near infrared bands are: 25.95%,23.06%,24.21%,26.48%,31.24%,28.39%,52.26%,66.50%.(2)The CHL retrieval algorithm OC3 G is best used in the study area.CHL values retrieved by OC2,OC3 G,YOC are underestimated,and the relative errors are30.68%,14.27%,31.30% respectively.The TSS retrieval algorithm Case2 is well applied in study area,and YOC is not suitable for high turbid water.TSS valuesretrieved by Case2 and YOC are underestimated,errors increase with TSS increasing,and the relative error is 20.91% and 44.43% respectively.(3)The GOCI products value in the clean East China Sea is greater than MODIS(except Rrs(490 nm)),and the ratio of its each band Rrs and CHL to MODIS is 1.78,1.10,0.94,1.40,1.64 and 1.73 respectively.The products value of GOCI and MODIS in turbid Zhejiang coastal area is basically consistent and the ratios are 1.05,0.90,0.88,0.97,1.10 and 1.09 respectively,with errors greater than clean water.(4)The imaging condition of the sensor is an important source of products' uncertainty,and the sensor's zenith angle is a vital factor.The deviation between two sensors' products increases with the increase of the observed zenith angle.