Comparison of two inversion techniques of a semi-analytical model for the determination of lake water constituents using imaging spectrometry data

In this study, two different inversion techniques for the determination of chlorophyll-a in water were compared by a sensitivity analysis: (i) a matrix inversion method, and (ii) a curve-fitting routine. Adding white noise to the reflectance spectrum led to clearly better results for the curve-fitting routine. If, however, the atmospheric parameter visibility was not exactly known, both methods behaved similarly well. The analyses implied that the performance depended on the quality of the input spectra, the knowledge of model parameters, and also on the inversion methods, even if they were based on the same semi-analytical model. Of course, not only the uncertainties of model parameters had to be considered for the testing of the performance, but also other factors, such as processing time, implementation of the inversion algorithm, number of relevant parameters, and the application of the method to different times and different lakes.