Improving the Accuracy of the SeaUV Algorithms in Dark Marine Waters
Abstract
Our long-term objective is to develop a robust set of algorithms for the global ocean to provide accurate surface UV attenuation and CDOM retrieval from remotely sensed ocean color for use in optical, photochemical, and photobiological investigation. The central objective of this project is to generate new, high quality optical data sets for a variety of darker coastal systems to be used in evaluating SeaUV algorithms and retraining them for accurate use in the highly variable optical conditions typical of nearshore waters. Previous ONR funding in our lab produced two improved and ready-to-use algorithms (SeaUV and SeaUVC) detailed in Fichot (2004) and Fichot et al. (2008). These algorithms are used for estimating Kd(320-490) and ag(320) from measurements of spectrally resolved remote sensing reflectance, Rrs(lambda). Our general approach for this project is to collect new in situ optical data sets for inshore and dark waters, apply the SeaUV algorithms to this new data set for evaluation of current predictive capability, and incorporate these new data into the training data set for evaluation of improved predictive capability using new "dark trained" algorithms. The final product will be a single model that will predict Kd(UV) and ag(lambda) from ocean color in optical domains ranging from the clear open ocean to the dark waters found in close proximity to the coast. We will then apply these trained algorithms to independent data sets where possible for validation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2008
- Accession Number
- ADA517437
Entities
People
- William L. Miller
Organizations
- University of Georgia