A database of geochemical compositions of aqueous species in produced water reported to the PA DEP. Samples were collected between mid-2012 to early-2020. Data from publicly-available PA DEP 26r reports were scraped from pdf files and cumulated into tabular spreadsheet format for >1000 produced water streams from Marcellus wells in Pennsylvania. In addition to providing the original values, the NETL NEWTS team has reformatted the dataset to allow sample streams to be easily copied into OLI Studio and Geochemist WorkBench (GWB) software for modeling the geochemistry and the recovery of critical minerals, such as lithium, from these produced water streams.

In addition, a version of the dataset has been included with predictions for some missing values in the original dataset using machine learning techniques within CoDaRT software, a public ML software developed by the Nation Energy Technology Laboratory. We have made the Input into CoDaRT and one example output from CoDaRT available in this dataset.