HR: 14:10h
AN: H12F-03    [PDF]
TI: Towards a unified approach for remote estimation of chlorophyll-a in both terrestrial vegetation and turbid productive waters
AU: Dall Olmo, G
EM: gdall@calmit.unl.edu
AF: University of Nebraska-Lincoln, 113 Nebraska Hall, Lincon, NE 68588-0517
AU: * Gitelson, A A
EM: gitelson@calmit.unl.edu
AF: University of Nebraska-Lincoln, 113 Nebraska Hall, Lincon, NE 68588-0517
AU: Rundquist, D C
EM: drundquist1@unl.edu
AF: University of Nebraska-Lincoln, 113 Nebraska Hall, Lincon, NE 68588-0517
AB: The photosynthetic pigment chlorophyll-a is an indicator of biomass and productivity of both terrestrial and aquatic ecosystems. Numerous medium specific, independent techniques to extract information on chlorophyll-a concentration from reflectance have been developedŸn. However, the differences and similarities of methods for extracting chlorophyll-a amounts from reflectance spectral data collected over different types of media (e.g. water bodies and plant leaves) have not been compared and generalized. Recently a conceptual model, relating remotely sensed reflectance and pigment content in higher plant leaves has been developed. The model was devised to isolate the absorption coefficient of the pigment of interest from reflectance spectra using three spectral regions. The model allowed accurate estimation of chlorophyll content in leaves and leaf area index in crops. In this study we tested the applicability of the model to retrieve chlorophyll-a concentrations from reflectance spectra of turbid productive waters. We tuned the conceptual model according to the optical characteristics of the aquatic medium, and accurately predicted chlorophyll-a concentrations in water bodies over a wide range of optical conditions (chlorophyll from 7 to 194 mg m$^{-3}$; total suspended matter from 0.1 to 214 mg L$^{-1}$; absorption coefficient of dissolved organic matter from 0.7 to 2.3 m$^{-1}$). Three spectral bands (the red, red edge and near infra-red) were used in the model, which accounted for 94% of the variance (p$<$0.0001) of chlorophyll-a concentrations measured analytically. In the range of chlorophyll variation from 7 to 194 mg m$^{-3}$, the root mean square error (RMSE) of chlorophyll-a estimation was less than 11 mg m$^{-3}$. The model was validated by independent data set yielding a RMSE of chlorophyll-a prediction lower than 13 mg m$^{-3}$. Our results provide evidence that this technique may be considered as a general solution, independent of the type of medium, for assessing chlorophyll concentration in optically deep media using remotely sensed data.
DE: 1845 Limnology
DE: 4275 Remote sensing and electromagnetic processes (0689)
DE: 4552 Ocean optics
DE: 4847 Optics
SC: Hydrology [H]
MN: 2003 Fall Meeting