You are here:
Estimation of diffuse attenuation of ultraviolet light in optically shallow Florida Keys waters from MODIS measurements
Citation:
Barnes, B., C. Hu, J. Cannizzaro, S. Craig, P. Hallock, D. Jones, J. Lehrter, N. Melo, B. Schaeffer, AND R. Zepp. Estimation of diffuse attenuation of ultraviolet light in optically shallow Florida Keys waters from MODIS measurements. REMOTE SENSING OF ENVIRONMENT. Elsevier Science Ltd, New York, NY, 140:519-532, (2014).
Impact/Purpose:
the objective of this study was to develop and validate a method to estimate Kd and ag in UV wavelengths from MODIS/Aqua measurements over optically shallow waters. The method is based on a new approach to implement the EOF method, as well as extensive field data collected from the Florida Keys region for algorithm tuning and validation.
Description:
Diffuse attenuation of solar light (Kd, m−1) determines the percentage of light penetrating the water column and available for benthic organisms. Therefore, Kd can be used as an index of water quality for coastal ecosystems that are dependent on photosynthesis, such as the coral reef environments of the Florida Reef Tract. Ultraviolet (UV) light reaching corals can lead to reductions in photosynthetic capacity as well as DNA damage. Unfortunately, field measurements of Kd(UV) lack sufficient spatial and temporal coverage to derive statistically meaningful patterns, and it has been notoriously difficult to derive Kd in optically shallow waters from remote sensing due to bottom contamination. Here we describe an approach to derive Kd(UV) in optically shallow waters of the Florida Keys using variations in the spectral shape of MODIS-derived surface reflectance. The approach used a principal component analysis and stepwise multiple regression to parsimoniously select modes of variance in MODIS derived reflectance data that best explained variance in concurrent in situ Kd(UV) measurements. The resulting models for Kd(UV) retrievals in waters 1–30m deep showed strong positive relationships between derived and measured parameters [e.g., for Kd(305) ranging from 0.28 to 3.27m−1; N=29; R2=0.94]. The predictive capabilities of these models were further tested, also showing acceptable performance [for Kd(305), R2= 0.92; bias=−0.02m−1; URMS=23%]. The same approach worked reasonably well in deriving the absorption coefficient of colored dissolved organic matter (CDOM) in UV wavelengths [ag(UV), m−1], as Kd(UV) is dominated by ag(UV). Application of the approach to MODIS data showed different spatial and temporal Kd(305) patterns than the Kd(488) patterns derived from a recently validated semi-analytical approach, suggesting that different mechanisms are controlling Kd in the UV and in the visible. Given the importance of water clarity and light availability to shallow-water flora and fauna, the new Kd(UV) and ag(UV) data products provide unprecedented information for assessing and monitoring of coral reef health, and could further assist ongoing regional protection efforts.
