<title>Remote sensing of natural water quality parameters: retrieval algorithm development and optimization analysis</title>

Abstract

Development of water quality retrieval algorithms is discussed in terms of causal dependence of the upwelling spectral radiance upon the water composition. Unlike clean marine/oceanic waters for which linear regression retrieval relationships are valid, inland and coastal zone water masses with high degree of optical complexity necessitate the development of nonparametric retrieval approaches. At the basis of these techniques are models considering the optical competitiveness of several coexisting aquatic components. Such models for Lakes Ladoga and Ontario are described and compared. Monte Carlo simulations have been performed to analyze the spectral and angular variations of the upwelling radiance scattered by the water column out into the atmosphere. Analysis of optical conditions for running remote sounding natural waters of various optical complexity is carried out. Relevant recommendations are formulated.