Influence of Particle Composition on Remote Sensing Reflectance and MERIS Maximum Chlorophyll Index Algorithm: Examples From Taihu Lake and Chaohu Lake

Using data collected from two eutrophic lakes located in eastern China (Taihu Lake, 2330 km² and Chaohu Lake, 760 km²), the influence of variable particle composition on remote sensing reflectance (Rrs, in sr^-1) properties and on the Medium Resolution Imaging Spectrometer (MERIS) maximum chlorophyll index (MCI) algorithm for estimating near-surface chlorophyll-a concentrations (Chla, in μg · L^-1) is demonstrated. Although separated by a distance of only ~200 km, the two lakes showed dramatic differences in particle composition, with Taihu Lake dominated by inorganic particles and Chaohu Lake dominated by organic particles. Such differences led to variable Rrs spectral slopes in the red and near-IR bands and perturbations to the MCI algorithm. A modified MCI algorithm (MCIT) was then developed to reduce the impact of turbidity caused by inorganic particles. Root-mean-square errors in Chla retrievals decreased from 129.5% to 43.5% when using this new approach compared with the MCI algorithm in Taihu Lake for Chla ranging between ~5 and 100 μg · L^-1. Application of this approach to other turbid water bodies, on the other hand, requires validation and possibly further tuning.

I. Introduction

BOTH inherent and apparent optical properties (IOPs and AOPs) in aquatic environments are affected by variable concentrations and compositions of particulate materials (e.g., [2], [23], [25], [27], and [30]), which in turn determine the design of remote sensing inversion algorithms. One such example is the maximum chlorophyll index (MCI) algorithm, which was designed for the Medium Resolution Imaging Spectrometer (MERIS) to estimate chlorophyll-a concentrations (Chla, in ) in highly productive waters [11].