REMOTE SENSING OF BIOGEOPHYSICAL FACTORS INFLUENCING LOCAL WATER QUALITY AND CORAL REEF HEALTH, U.S. VIRGIN ISLANDS

Coral reefs offer an important source of food, biological diversity, and protection against flooding in coastal areas. Island economies rely on healthy reefs as major attractions in the lucrative industry of tourism. Increases in development and changes in land use have created an influx of sediment and nutrients entering the coastal waters of the U.S. Virgin Islands (USVI), causing detrimental effects on water quality (Ennis et al. 2016). Consequently, coral reefs have started to degrade, particularly those located within nearshore waters (Ennis et al. 2016). Following Kerrigan (2015) and Hollister (2016), we employed remote sensing as a method of water quality monitoring, which offers a spatial advantage and cost effective alternative over traditional water quality monitoring.

This study integrates NASA satellite data with field spectroscopy in order to determine bio-optical properties and to quantify water quality parameters that affect coral reef health in the coastal waters of the USVI. Landsat-8 (L8) surface reflectance imagery from August 2016 through August 2017 were analyzed by calculating the derivative of the measured visible/near infrared spectra and then using Varimax-rotated Principal Component Analysis (VPCA) decomposition to identify color producing agents in the water column (Ortiz et al. 2013). VPCA loadings were standardized and results were matched to libraries of reflectance derivative spectra for known pigment and mineral standards. To verify these results, surface reflectance measurements were made (400-900nm) using an ASD Fieldspec HH2 spectroradiometer (ASD) in December 2016 and May 2017 around the coastal waters of St. Thomas Island, USVI (Ortiz et al., KSU NASA EPSCoR Annual Report 2016). ASD data were averaged to 10nm resolution for comparison with the spectral library and to L8 resolution for comparison to the VPCA decomposed L8 imagery.

VPCA-decomposition of the L8 images revealed four matching spectral signatures across dates, which are graphically similar to the spectral signatures of the in situ hyperspectral measurements. The detection and analysis of water quality parameters is a necessity in current and future remediation efforts and VPCA-decomposition will likely prove beneficial as an inexpensive method of near real time water quality monitoring.