气候变化领域集成服务门户

Retrievals of water quality parameters from satellite measurements over optically shallow waters have been problematic due to bottom contamination of the signals. As a result, large errors are associated with derived water column properties. These deficiencies greatly reduce the ability to use satellites to assess the shallow water environments around coral reefs and seagrass beds. Here, a modified version of an existing algorithm is used to derive multispectral diffuse attenuation coefficient (K-d) from MODIS/Aqua measurements over optically shallow waters in the Florida Keys. Results were validated against concurrent in situ data (K-d(488) from 0.02 to 0.20 m(-1), N = 22, R-2 = 0.68, Mean Ratio = 0.93, unbiased RMS = 31%), and showed significant improvement over current products when compared to the same in situ data (N = 13, R-2 = 037, Mean Ratio = 1.61, unbiased RMS = 50%). The modified algorithm was then applied to time series of MODIS/Aqua data over the Florida Keys (in particular, the Florida Keys Reef Tract), whereby spatial and temporal patterns of water clarity between 2002 and 2011 were elucidated. Climatologies, time series, anomaly images, and empirical orthogonal function analysis showed primarily nearshore-offshore gradients in water clarity and its variability, with peaks in both at the major channels draining Florida Bay. ANOVA revealed significant differences in K-d(488) according to distance from shore and geographic region. Excluding the Dry Tortugas, which had the lowest climatological K-d(488), water was clearest at the northern extent of the Reef Tract, and K-d(488) significantly decreased sequentially for every region along the tract. Tests over other shallow-water tropical waters such as the Belize Barrier Reef also suggested general applicability of the algorithm. As water clarity and light availability on the ocean bottom are key environmental parameters in determining the health of shallow-water plants and animals, the validated new products provide unprecedented information for assessing and monitoring of coral reef and seagrass health, and could further assist ongoing regional zoning efforts. (C) 2013 Elsevier Inc. All rights reserved.