Modeling the spectral absorption by CDOM in Meiliang Bay, Taihu Lake

Qinghua Fu; Shixin Wang; Yi Zhou; Fuli Yan; Jingjing Wang

doi:10.1117/12.760499

Translator Disclaimer

26 July 2007 Modeling the spectral absorption by CDOM in Meiliang Bay, Taihu Lake

Qinghua Fu, Shixin Wang, Yi Zhou, Fuli Yan, Jingjing Wang

Author Affiliations +

Proceedings Volume 6752, Geoinformatics 2007: Remotely Sensed Data and Information; 67521L (2007) https://doi.org/10.1117/12.760499
Event: Geoinformatics 2007, 2007, Nanjing, China

Abstract

The interpretation of remotely sensed images of turbid coastal waters or inland lake waters is more difficult than case 1 water, because their optical properties are complex, and their optical constituents are independent of phytoplankton concentrations. In recent years significant efforts have been made to develop ocean color satellite missions with improved spectral and radiometric performance, and in the same time, techniques for constituent retrieval have evolved from empirical towards analytical algorithms. Analytical models can be developed and inverted to yield concentrations (Carder et al. 1999) of substances in the water from reflectance measurements, which require a suitable parameterization of the optically active constituents and their optical properties. This paper focused on absorption by chromophoric dissolved organic matter (CDOM; also Gelbstoff or yellow substances), which was the pool of absorbing substances in water and one of the main optically active constituents in Case 2 waters. The absorption of CDOM is generally considered as the exponential form model, which have three important main parameters, S, a(λ₀), λ₀. The S results got from the exponential form model fit using CDOM normalization absorptions by 350nm, or 400nm, or 440nm absorption were the same, and the final value of S for CDOM in Meiliang Bay, Taihu Lake was 0.0106, namely the mean of S for all samples; Normally, a(λ₀) is simply taken to be the mean of the absorption coefficient of CDOM of field samples in the reference wavelength, however, this study found that a(λ₀) a varied greatly between samples ( λ₀ = 400, 1.93 <a(_λ0)<4.09; λ₀=440, 1.12<a(λ₀)<2.56; λ₀ =350, 4.07 <a(λ₀)<7.85), and the correlation coefficient between a(_λ) and TN(total nitrogen) concentration increased with the decrease of wavelength, in 350nm, up to 0.83, so a(_λ0); a submodel was constructed by lineal regression of a(λ₀) and TN(total nitrogen) concentration; The errors of reversion were compared and analyzed in yellow matter absorption model using different parameter, and three reference wavelengths _λ0, namely 350nm, 400nm and 440nm, were considered in the paper; and the results showed that parameter a(_λ0) was most important parameter in absorption model of CDOM than other parameters, and sub-model of a(λ₀) was more reasonable parameter to CDOM absorption model than average a(λ₀). Because absolute relative error using a(_λ0) submodel in 350-700nm was greatly reduced, its average absolute relative error was 15.1 percent, and correlation coefficients between measured absorption coefficient of CDOM and estimated absorption coefficient of CDOM in 350- 700nm were remarkably improved, its average value in 350-700nm was 0.73, whereas average absolute relative error using average a(λ₀) a was approximately 24 percent and its correlation coefficients between measured absorption coefficient of CDOM and estimated absorption coefficient of CDOM were 0. Furthermore, choice of reference wavelength had little effect on CDOM absorption coefficient reversion.

Citation Download Citation

Qinghua Fu, Shixin Wang, Yi Zhou, Fuli Yan, and Jingjing Wang "Modeling the spectral absorption by CDOM in Meiliang Bay, Taihu Lake", Proc. SPIE 6752, Geoinformatics 2007: Remotely Sensed Data and Information, 67521L (26 July 2007); https://doi.org/10.1117/12.760499

ACCESS THE FULL ARTICLE