Development of a MODIS Data–Based Algorithm for Retrieving Gage Height in Nearshore Waters along the Louisiana Gulf Coast

• Published in: Journal of coastal research Vol. 34; no. 1; pp. 220 - 228
• Main Authors: Wang, Jiao, Deng, Zhiqiang
• Format: Journal Article
• Language: English
• Published: Fort Lauderdale The Coastal Education and Research Foundation 01.01.2018; COASTAL EDUCATION & RESEARCH FOUNDATION, INC; Allen Press Inc
• Subjects: Algorithms; Artificial neural networks; Artificial satellites; Bathymetry; Coastal management; Coastal water; Coastal waters; Coastal zone; Coastal zone management; Coasts; Correlation coefficient; Correlation coefficients; Data; Earth science; Fisheries; Geological surveys; Ground truth; Height; Imaging techniques; Limnology; Mathematical models; MODIS; MODIS Aqua data; Nearshore water; Neural networks; Physical properties; Reflectance; Remote monitoring; Remote sensing; remote sensing algorithm; Resource development; Resource management; Satellites; Sea level; Shallow water; Shellfish; Spectroradiometers; Studies; Surveying; TECHNICAL COMMUNICATIONS; Training; Water depth; Louisiana; United States > US
• ISSN: 0749-0208; 1551-5036
• DOI: 10.2112/JCOASTRES-D-16-00161.1

Wang, J. and Deng, Z., 2018. Development of a MODIS data–based algorithm for retrieving gage height in nearshore waters along the Louisiana Gulf Coast. Gage height is one of most important physical parameters commonly used for the description of daily sea levels and generally monitored at sparsely scattered tidal stations. This paper presents a novel remote sensing algorithm for the retrieval of spatially distributed gage height data in coastal waters with emphasis on nearshore waters using the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data. The algorithm was trained using the Artificial Neural Networks toolbox in the MATLAB program and over 4 years (2007–11) of cloud-free MODIS Aqua data for water-leaving reflectance as well as ground truth measurements collected daily from U.S. Geological Survey stations along the Louisiana Gulf Coast. The algorithm was validated using 3 additional years of independent data sets, which were not used in the algorithm training and collected from 2012 to 2014. Cross-validation results indicated that the gage heights derived from the new algorithm were in good agreement with observed height, as evidenced by the high linear correlation coefficient of 0.8465 and low root mean square error of 0.2238 m. The new algorithm makes it possible to produce daily, spatially distributed gage height data for coastal management and resources development in shallow coastal areas where there are no gages.