Font Size: a A A

The Feasibility Of Deriving The Vertical Distribution Of Chlorophyll Concentration From Sea Surface Reflectance Measurements In Case 1 Waters

Posted on:2010-02-01Degree:MasterType:Thesis
Country:ChinaCandidate:B LiuFull Text:PDF
GTID:2178360275986007Subject:Physical oceanography
Abstract/Summary:PDF Full Text Request
The feasibility study of the vertical structure of the chlorophyll in Case I waters with technique of ocean color is proposed in this study. Optical remote sensing of ocean color is a well-established technique that is used to produce maps of marine constituents on a routine basis. The traditional retrieval algorithms used to derive chlorophyll concentrations from measurements of ocean color were usually based the assumption that the upper ocean column was vertically homogeneous. However, the vertical distribution of chlorophyll in majority of ocean is stratified. In this paper, the vertically distribution of chlorophyll in Case I waters is assumed to consist of three homogeneous layers with different concentrations of chlorophyll. Four parameters used to describe the vertical structure of the model is the background value of chlorophyll concentration, the thickness of mixed layer of the upper ocean, the peak value of chlorophyll concentration, the thickness of the layer with the peak value of chlorophyll concentration. Based on the two-flow radiative transfer theory, the relationship of the vertical structure of chlorophyll and sea surface reflectance is modeled, and the effects of the four parameters on the sea surface reflectance is examined. And on this basis of above, a feasibility study based on one three-layer artificial neural to derive the various parameters of chlorophyll concentrations vertical structure is carried out. The sixty input variables of the three-layer structure of BP neural network correspond to 60-band sea surface reflectance. The output is a parameter of the vertical distribution of chlorophyll. 60% of simulated data is used to train the network, 40% is used to test network. The results show that the derivation of the background concentration of chlorophyll, the thickness of mixed layer of the upper ocean, and the peak concentration of chlorophyll is possible in theory, but the thickness of the layer with the peak value of chlorophyll concentration is impossible.
Keywords/Search Tags:Ocean color, Vertical distribution of chlorophyll, Artificial neural network
PDF Full Text Request
Related items