HR: 1340h
AN: SF43A-0773    [Abstracts]
TI: On the Potential of Existing Ocean Color Sensors for Monitoring Productive Turbid Waters
AU: * Dall'Olmo, G
EM: gdall@calmit.unl.edu
AF: CALMIT, School of Natural Resources, University of Nebraska-Lincoln, 113 Nebraska Hall, Lincoln, NE 68588-0517 United States
AU: Gitelson, A A
EM: gitelson@calmit.unl.edu
AF: CALMIT, School of Natural Resources, University of Nebraska-Lincoln, 113 Nebraska Hall, Lincoln, NE 68588-0517 United States
AB: Case-1 water bio-optical algorithms for remote estimation of chlorophyll-a concentration (Chl) exploit the upwelling radiation in the blue and green spectral regions. In case-2 waters other constituents, that vary independently of Chl, absorb and scatter light in these spectral regions. As a consequence, the accurate estimation of Chl in productive turbid waters has not been so far feasible from satellite sensors. In this study we calibrated and validated Chl bio-optical models based on simulated red and near-infrared (NIR) channels of three existing ocean color sensors: the Sea Wide Field-of-View Sensor (SeaWiFS), the Moderate Imaging Spectrometer (MODIS) and the Medium Resolution Imaging Spectrometer (MERIS). Reflectance spectra and relevant water constituents were collected in 270 stations over lakes and reservoirs with a wide variability in optical parameters (i.e. 4$<=$Chl$<=$240 mg m-3; 18$<=$Secchi disk depth$<=$308 cm). The relative uncertainties in Chl prediction due to the bio-optical algorithms alone were low ranging from 24% to 28% (average bias between -6% and +5%). Then, by taking into account the radiometric sensitivities of the sensors, we determined the atmospheric correction requirements for accurate Chl estimation. The radiometric sensitivity of SeaWiFS appeared to be too low for estimating Chl using the red and NIR channels. In contrast, the higher signal-to-noise ratio of MODIS allowed accurate ($<$35%) estimation of Chl in turbid productive waters for Chl$>$15 mg m-3. For a worst case scenario (solar zenith angle 60ø, Chl=15 mg m-3, uncorrelated uncertainties), the uncertainty in atmospheric correction required to obtain a $<$35% overall uncertainty in Chl was approximately three times the sensor noise. Such a requirement was less stringent at lower solar zenith angles, higher Chl and in the case of correlated atmospheric uncertainties. Thus, it appears that MODIS could be used for predicting Chl in case-2 productive turbid waters. MERIS three-band algorithms using the 705 nm channel have the additional advantages of being less affected by sensor noise and to compensate for correlated atmospheric uncertainties. This allows to accurately estimate Chl$>$10 mg m-3.
DE: 4847 Optics
DE: 4239 Limnology
DE: 4275 Remote sensing and electromagnetic processes (0689)
DE: 4552 Ocean optics
SC: Special Focus: Advances in Data Acquisition, Management, Analysis and Display [SF]
MN: 2004 AGU Fall Meeting