POSTER SESSION 3
15:30–17:30, Wednesday, September 16
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1338 | POSTER 15
DERIVING WATER QUALITY IN COASTAL AND INLAND WATERS: BEST PRACTICES FOR ALGORITHM DEVELOPMENT AND VALIDATION
Coastal and inland waters offer essential resources and services, but effective monitoring is hindered by limited in situ sampling and uncertainties in satellite-derived data. Higher-resolution sensors and modern empirical algorithms have advanced capabilities in water quality monitoring, but there is limited consensus on best practices and minimum criteria required to ensure robust model approaches. Moreover, algorithms tuned for one waterbody or sensor often fail to generalize across regions or platforms, resulting in a patchwork of regionally tuned algorithms with no direct transition mechanism. In this work, we present findings from algorithm development activities and a global round robin of existing algorithms spanning inland, estuarine, and coastal waters, highlighting novel approaches and shared features of best-performing models. Importantly, new models are benchmarked against existing global / regional approaches, while models of increasing complexity are justified against performance of simpler approaches. Particularly in data limited regions (< 100-200 matchups), conventional algorithms may thus prove more robust derivations than those based on machine learning. Even in such traditional algorithm frameworks, approaches based on Rayleigh corrected reflectance can outperform those based on remote sensing reflectance, potentially due to residual atmospheric correction uncertainties. Other design criteria for the most successful algorithms include dynamic determination of optically shallow pixels and optical water types, as well as product-level bridging to maximize cross-sensor consistency. Together, these efforts reinforce the need for well-justified algorithm design processes, while resultant algorithms have strengthened the reliability of aquatic data products and enabled characterization of water quality patterns and trends in these critical waterbodies.
Brian Barnes, University of South Florida, [email protected], https://orcid.org/0000-0003-0056-3500
Cheng Xue, University of South Florida, [email protected]
Madjid Hadjal, University of South Florida, [email protected]
Chuanmin Hu, University of South Florida, [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
