Monitoring reservoir water quality with Formosat-2 high spatiotemporal imagery

Abstract

Water reservoirs are the primary source of freshwater for most cities around the world. To monitor the dynamic changes in reservoir water quality, however, we need an innovative platform that is able to observe the entire reservoir with both high-spatial- and high-temporal-resolution. Formosat-2 is the first commercial satellite dedicated to site surveillance with a high-spatial-resolution sensor placed in a daily revisit orbit (2 m in panchromatic and 8 m in multispectral). In this research, we developed two empirical algorithms to map the water contents of Chlorophyll-a (Chl-a) and suspended solids (SS) from Formosat-2 multispectral imagery. These algorithms are derived from a total of 53 pairs of water-quality and surface-reflectance data collected during 14 field campaigns at Tsengwen Reservoir from 2005 to 2006. A total of 15 Formosat-2 images were selected from all available images of Tsengwen Reservoir taken in 2006 to generate water quality maps of Chl-a and SS using our new algorithms. Results from this study indicate that the Chl-a and SS concentrations can be retrieved from Formosat-2 imagery with deviations of 56% and 43%, respectively. This is the first time that the reservoir water quality can be mapped from a high-spatial-resolution satellite image at such a high-temporal-resolution. To facilitate the administration of water resources, this research encourages the application of Formosat-2 high spatiotemporal imagery in identifying areas of poor water quality and monitoring the dispersal pattern of pollutant plumes.