Abstract Submission: The USEPA issued the Coal Combustion Residual (CCR) Rule in 2015 to regulate CCR storage units. This rule requires evaluation of CCR units for a release to groundwater. If results indicate a potential release, an alternative source demonstration (ASD) must provide an explanation and a source other than a release from the regulated unit. Otherwise, the unit begins more comprehensive monitoring which can lead to corrective measures. ASD development under dynamic groundwater conditions can be challenging, with a variety of factors complicating identification of alternative sources. An advanced statistical method helps differentiate between natural variability and changes from operations, and identify causes for observed changes. The Principal Component Analysis (PCA) approach can benefit ASD preparation, allowing simultaneous analysis of multiple variables and accounting for their interactions. The PCA model is one type of latent variable model in which the latent variables are oriented in the direction of maximum variance of the input data. The latent variables help identify patterns and relationships between actual variables. Further, the residual errors and score distances of a PCA model help understanding the data distribution in the reduced-dimensional space, model diagnoses, and validation. These statistics can be plotted and statistically tested documenting whether sampling results were regular or extreme, and identify the constituents making the event more extreme. A case study is presented using a real-life dataset showing the use of the PCA modeling for ASD development. The PCA supported and clarified hydrogeologic, geochemical, and univariate statistical analyses, for a stronger and more complete evaluation.
.jpg "Aditya Tyagi, PhD, PE (he/him/his) photo")
