論文
Ocean Color Algorithms for Estimating Water Clarity (Secchi Depth) from SeaWiFS
1998 年 4 巻 2 号 p. 301-306
詳細
1998 年 4 巻 2 号 p. 301-306
Relationships between water-leaving radiances (Lwn) in the blue:Green region of electromag netic spectrum, Secchi disk depth and the diffuse attenuation coefficient at 490 nm (K490) were examined using data from several CalCOFI cruises (1993-1997). We were able to explain ap proximately 90% of the variance in Secchi depth and K490 using band ratios of Lwn490/Lwn555. We present results that support the view that a ratio of Lwns at 490:555 perform better than the 443:550 in explaining the variance of K490 and Secchi depth. We also performed a neural net (NN) analyses to predict Secchi depth using a combination of Lwn ratios and information from a series of channels corresponding to SeaWiFS (Lwns at 412,443,490, 512,555 and Lwn ratios at 443/510, 443/555,490/555). The best training set from the NN analyses resulted in an r of 0.91. A sensitivity analysis using the neural network methodology allowed us to specify the variables that are most useful in predicting Secchi depth. We present results of predicted Secchi depth and K490 using SeaWiFS data from Southern California and other locations where water turbidity is of great interest to recreational and commercial fishing and diving.
