A Methodology to Estimate Water Demand for Thermal Power Plants in Data-Scarce Regions Using Satellite Imagesby , and -
Thermal power generation is thirsty, but disclosure of power plant water usage is not required in many countries. WRI has created a new methodology to calculate water usage for thermal power plants, even those in data-scarce regions. Plant cooling and fuel type, identified through satellite data, can be compared to empirical water-use factors to estimate water usage. The data gleaned from this method can help companies, governments, and investors better understand water risks to thermal power plants, and the power sector as a whole.
Thermal power plants depend on and consume water for cooling. Detailed power sector water-use data can help aid in evaluating the security of water access for power plants as well the impact of their water use on downstream water users. While power generation data are widely available globally, no readily available global data exist on power plant water withdrawal and consumption. This paper proposes an innovative methodology to help close the power sector water-use data gap for data-scarce regions. The methodology uses free, easily accessible satellite images to visually identify cooling and fuel type, and empirical water-use factors to estimate water demand. The method was tested against 200 U.S. power plants and achieved 90 percent precision in cooling technology identification, 89 percent precision in fuel identification, and 69 percent precision in water demand estimation. Information gleaned from this methodology can be used by companies, investors, and the general public to better understand water risks to energy and to devise a better path forward.