Study On Inversion Model Of Chlorophyll Concentration Based On GOCI Image Data In The North Yellow Sea

Chlorophyll is an important source of energy in marine ecosystem.As one of the three elements of water color,chlorophyll can reflect the quality of the water,the biomass of phytoplankton,the eutrophication degree of the water body and the primary productivity of the sea area.The early warning of harmful algal bloom can also be realized,so it is of great value and significance to measure,invert and analyze the distribution characteristics of chlorophyll concentration.GOCI is a geosynchronous water color sensor.The application of GOCI data has great value to research.The North Yellow Sea is a semi-enclosed sea with a long renewal cycle and limited self-purification capacity.The coastal waters are seriously affected by human activities.The waterbody of the North Yellow Sea belongs to case 2 waters,and it is difficult to invert the water color elements.The main work and conclusions of this paper are as follows:(1)Based on the MODIS reflectivity data,the correlations between the GOCI data and the MODIS data after using FLAASH,QUAC and KOSC atmospheric correction methods were analyzed in this paper,and the three methods were evaluated the correlation between the GOCI data and the MODIS data..The results showed that the KOSC algorithm is stricter in discriminating cloud and fog,the data utilization ratio is lower,the accuracy of QUAC is a little lower than that of FLAASH algorithm,and the performance of FLAASH algorithm is more stable and the precision is higher.And the trend of spectral changes is closer to MODIS data than QUAC.Therefore,the FLAASH was used for atmospheric correction in this study.(2)The inversion results of three classical water color inversion algorithms,OC2,OC3and YOC,are compared in this paper,R~2 of OC2 is the highest,which is 0.7360.The coefficients of OC2,OC3 and YOC algorithms were optimized and the factors were improved.The OC2 model with B412*B490/B555^2 as a factor has the highest comprehensive accuracy,and its R~2,MAPE and RMSE are 0.7374,27.63%,and0.2091?g/L respectively.A new double-factor model was proposed.It is found that when the two factors are(B443+B490)/B555 and B443*B490/B555^2,the comprehensive performance of the model is the best.Its R~2,MAPE and RMSE were0.7414,27.32%and 0.2057?g/L respectively.Compared with the OC2,OC3 and YOC models,the accuracy of the double-factor model was improved.(3)The inversion ability of the double-factor model was tested.The spatio-temporal variation of chlorophyll concentration in the North Yellow Sea in 2015 were analyzed.The distribution characteristics of chlorophyll concentration are as follows:the chlorophyll concentration in the coastal waters is high and the area of high concentration is large,and the chlorophyll concentration in the center of the sea is lower.The high concentration areas are mainly distributed in north,northeast and east of the North Yellow Sea,and the waters near Weihai in Shandong province.There are also high chlorophyll concentration and large area of high chlorophyll concentration in the waters near river estuary and islands.