The origin and global distribution of second order variability in satellite ocean color and its potential applications to algorithm development

Empirical algorithms based on first order relationships between ocean color and the chlorophyll concentration ([ChI]; mg m(-3)) are widely used, but cannot explain the statistical dispersion or "anomalies" around the mean trends. We use an empirical approach that removes the first order effects of [ChI] from satellite ocean color, thus allowing us to quantify the impact on the ocean color signal of optical anomalies that vary independently of the global mean trends with remotely sensed [ChI]. We then present statistical and modeling analyses to interpret the observed anomalies in terms of their optical sources (i.e. absorption and backscattering coefficients). We identify two main sources of second order variability for a given [ChI]: 1) the amount of non-algal absorption, especially due to colored dissolved organic matter; and 2) the amplitude of the backscattering coefficient of particles. The global distribution of the anomalies displays significant regional and seasonal trends, providing important information for characterizing the marine optical environment and for inferring biogeochemical influences. We subsequently use our empirically determined anomalies to estimate the backscattering coefficient of particles and the combined absorption coefficient for colored detrital and dissolved materials. This purely empirical approach provides an independent assessment of second order optical variability for comparison with existing methods that are generally based on semi-analytical models. (C) 2008 Elsevier Inc. All rights reserved.