Research On Remote Sensing Retrieval Algorithms Of Secchi Disk Depth

The Secchi disk depth(SDD)is an important parameter describing the optical properties of seawater.Monitoring the temporal and spatial changes of seawater transparency is of great significance for studying the physical and chemical properties of seawater,fishery and marine military activities.The traditional method for SDD measurement has a large workload,and it is impossible to obtain a large temporal and spatial distribution characteristics of the SDD.However,with the development of ocean color remote sensing technology,real-time and large-scale oceanic SDD monitoring has become possible.In order to retrieve high-precision SDD values from remote sensing data,it is necessary to focus on the SDD inversion algorithms.Among all the SDD inversion algorithms,the semi-analytical algorithms are widely used by scholars and have been developed so far.But there are still some disadvantages when these semi-analytical algorithms applied to the Case Ⅱ water bodies,such as low inversion accuracy and incompatibility of inversion algorithms.Therefore,this paper summarized and analyzed the previous research results of SDD inversion and further studied the shortcomings of the semi-analytical algorithms.And our research had improved the related algorithms and use the in-situ SDD data in the Yellow Sea of China to verify and analyze the inversion precision.Its main work and results are as follows:(1)In the process of inverting SDD using the semi-analytical algorithm,it is necessary to use the QAA(Quasi Analytical Algorithm)to solve the absorption coefficients and backscattering coefficients at wavelengths of narrow band,and these parameters are key intermediate parameters.Therefore,three versions of QAA models including QAA＿V4,QAA＿V5 and QAA＿V6 were compared.It is found that the QAA model has higher accuracy when inverting the absorption coefficient at wavelengths of narrow band,and the QAA＿V4 model performs better than the other two versions in inverting the absorption coefficient at wavelengths of narrow band,whose root mean square error and the average relative error reached 0.1995 and 11.82%,respectively.(2)Two SDD inversion modes including Poole-Atkins mode and DuntleyPreisendorfer mode were analyzed for the sensitivity of their undetermined parameters.Then the accuracies of the two models were verified with the in-situ SDD data in the Yellow Sea.Although both the inversion accuracies of the two SDD inversion modes in the Yellow Sea are relatively low,the Poole-Atkins model is generally more stable.(3)In order to further improve the inversion accuracy of Poole-Atkins model,this research introduces the Forel-Ule Index to improve the Poole-Atkins model.Through the analysis of the IOCCG(International Ocean Color Coordinating Group)data set,it was found that there is strong correlation between the FUI and parameter A.Comparing the inversion results with the in situ measured data in the Yellow Sea showed that the RMSE(Root Mean Square Error)and MRE(Mean Relative Error)of the quadratic formula are 1.83 and 43.74%,respectively,which showed a better performance than the traditional Poole-Atkins model.(4)Finally,the improved Poole-Atkins model in this study was used to map the SDD inversion results for the Yellow Sea and analyzed the spatial and temporal changes.And the factors affecting the inversion and change of transparency are analyzed.