Abstract
A general approach is presented to quantify retrieval errors in in-water constituent concentra
tions induced by uncertainty in inherent optical property (IOP) submodel parameterization. Chlo
rophyll concentrations are retrieved by inverting a radiance model with nonlinear IOP submodels.
We demonstrate quantitatively how uncertainty in the IOP submodel parameterization influences
the accuracy of the chlorophyll concentration retrieval at different chlorophyll concentration lev
els. Two complete sets of simulations were designed and conducted, representing two extreme
cases between which ”real” cases are expected to occur. The simulations show that precise knowl
edge of spectral shapes of IOP submodels is important in chlorophyll retrieval.
