IMPROVED RRS AND TURBIDITY RETRIEVAL FROM OLI IMAGES IN COMPLEX INLAND WATERS: A CASE STUDY FOR LAGUNA DE BAY
Keywords: Laguna de Bay, Landsat-8 Operational Land Imager (OLI), remote sensing, turbidity, atmospheric correction
Abstract. Spatiotemporal monitoring of water quality parameters such as turbidity in inland waters is desirable to better understand productivity and mitigate the negative impacts of pollution induced by increasing anthropogenic activities. However, precise retrieval of water quality parameters in complex turbid waters from the remote sensing reflectance Rrs, remains a challenging task due to the varying optical complexity of the water body. In this study, a modified version of the Atmospheric Correction for OLI-lite (ModACO) scheme for turbid inland waters, which implements a linear extrapolation of NIR aerosol reflectance in the visible bands instead of a nonlinear function, is presented. The performance of the proposed method and other existing algorithms such as the Atmospheric Correction for OLI-lite (ACOLITE), Management Unit of the Noth Seas Mathematical Models (MUMM) scheme, SeaDAS standard processing, and SWIRE were evaluated. The Rrs retrievals from these models were then used as input for turbidity estimation and mapping of Laguna de Bay. Results show lowest Rrs error in all five spectral bands (443, 482, 561, 655, and 865 nm) on ModACO-based retrievals. Relative to the ACOLITE and other atmospheric correction schemes, the proposed method reduced the Rrs retrieval errors in terms of RMSE and MAPE by more than 50%. Similarly, significant improvements in turbidity retrievals were achieved from ModACO-based Rrs values, wherein comparable accuracy was observed from red/green ratio and the single NIR band turbidity models. Turbidity maps of Laguna de Bay show elevated values from the mid of dry season, which may be associated with point source discharge and wind-induced resuspension of bottom sediments. The lake turbidity then drops by the end of dry season, which is linked to the absence of prevailing strong winds that may increase in-water mixing. Using the aforementioned method, accurate monitoring of turbidity can be done to determine and mitigate possible degradation on the water quality of Laguna de Bay and other productive turbid inland waters.