Abstract:
A series of three mathematical procedures is derived to discriminate the light absorption by phytoplank-ton,
colored dissolved organic matter, and nonpigmented particulates in waters in which absorption
is dominated by factors other than phytoplankton (i.e., case 2 waters). The procedures utilize
normalized absorption cross-sectional spectra of the absorption components and matrix inversion to
solve for the coefficients that scale the normalized spectra. The procedures differ in the amount of
ancillary measurements incorporated to reduce the variability of the estimates. The procedure that
incorporates no ancillary information is expected to be unbiased only over long time periods. Ap-plication
of the procedures to a 15-day time series of continuously monitored data from the Rhode
River, Maryland, revealed the presence of large (approximately twofold) changes in absorption at 440
nm over periods of a few hours. Hourly sampling over a 24-h period confirmed that the changes in
measured optical coefficients corresponded to changes in water quality. Errors in estimates of
absorption components were of a magnitude consistent with those observed in development of the
procedures and confirmed the progressive improvement achieved by incorporation of additional
information. Over the time period observed, changes in optical properties appeared to be driven by
advective processes.
