Waterkeeper Alliance Coal-Ash Pilot Project: Development of Web GIS and Thermal Infrared Imagery Processing Tools for Preliminary Correlation Between Stream Temperatures and Proximity to Coal-Ash Waste
Presentation Time: Tue, 08/02/2022 - 13:00
Keywords: coal-ash, toxic waste, thermal infrared imagery, climate change, stream temperature
Coal-ash toxic waste is a “dirty energy” byproduct of coal-fired power that includes heavy metals and contaminants and is associated with climate change. Its disposal is poorly regulated in the United States; in some cases, coal-ash is discharged directly into rivers. Coal-ash contributes to elevated stream temperatures, adversely impacting living organisms and ecology. Waterkeeper Alliance, a non-profit organization advocating for clean water, initiated a coal-ash project with three aims: creating web tools that aid waterkeepers in identifying coal-ash concerns; developing a tool that processes Landsat Surface Temperature data derived from thermal infrared imagery and prepares it for analysis; and exploring the correlation between water surface temperatures and proximity to coal-ash waste using Landsat data. Coal-ash data were transformed into web layers in ArcGIS Online and formed the basis of web maps and apps for waterkeeper use. Using ArcGIS Pro ModelBuilder, a model was built to process Landsat data in a Pilot Study Area (Missouri and Kentucky). In the model, high-quality water pixels are identified, converted to Celsius, and extracted as vector points with distance to coal-ash added as an attribute. The model concludes with a bivariate spatial correlation between water temperature and coal-ash proximity. In one iteration, ninety-four percent of data showed a geostatistically significant correlation between water surface temperature and coal-ash proximity. Twenty-four percent showed a negative linear correlation and thirty-seven percent showed a complex relationship. While the analysis affirms a non-random relationship between variables, the relationship appears far more complex than two variables and Landsat imagery can explain.
