Empirical algorithms for the total absorption coefficient and absorpti
on coefficient by pigments for surface waters at 440 nm were developed
by applying a quadratic formula that combines two spectral ratios of
remote-sensing reflectance. For total absorption coefficients ranging
from 0.02 to 2.0 m(-1), a goodness of fit was achieved between the mea
sured and modeled data with a root-mean-square difference between the
measured and modeled values for log10 scale (RMSDlog10) of 0.062 (15.3
% for linear scale, number of samples N = 63), while RMSDlog10 is 0.11
1 (29.1% for linear scale, N = 126) for pigment absorption (ranging fr
om 0.01 to 1.0 m(-1)). As alternatives to pigment concentration algori
thms, the absorption algorithms developed can be applied to the coasta
l zone color scanner and sea-viewing wide-field-of-view sensor data to
derive inherent optical properties of the ocean, For the same data se
ts, we also directly related the chlorophyll a concentrations to the s
pectral ratios and obtained an RMSDlog10 value of 0.218 (65.2% for lin
ear scale, N = 120) for concentrations ranging from 0.06 to 50.0 mg m(
-3). These results indicate that it is more accurate to estimate the a
bsorption coefficients than the pigment concentrations from remotely s
ensed data. This is likely due to the fact that for. the broad range o
f waters studied the pigment-specific absorption coefficient at 440 nm
ranged from 0.03 to 0.2 m(2) (mg chl)(-1). As an indirect test of the
algorithms developed, the chlorophyll a concentration algorithm is ap
plied to an independent global data set and an RMSDlog10 of 0.191 (55.
2% for linear scale, N = 919) is obtained. There is no independent glo
bal absorption data set available as yet to test the absorption algori
thms.