Freshwater salinization is an emerging water quality issue for non-tidal streams and rivers in the Chesapeake Bay watershed (CBW), USA region. A model was developed to predict specific conductance (SC; a proxy for salinity) conditions across the CBW and departures from background SC. Discrete observations of SC from 1999-2016 were acquired from a published SC data inventory and explanatory variables describing sources of SC were compiled from several sources. Random forests modeling was conducted to predict SC at four time periods (1999-2001, 2004-2006, 2009-2011, and 2014-2016) at all non-tidal National Hydrography Dataset Plus Version 2.1 (NHDPlusV2.1; 1:100K scale) stream reaches. These predictions were then compared to a national background SC dataset to determine relative departures from background SC for each NHDPlusV2.1 reach ID. This data release contains model input data, model output data, and predictions of SC. This data release contains the following three files: 1."Model_input.csv": Contains SC observations, explanatory variables, and additional columns relevant to the model application. 2. "Model_output.csv": Contains predicted SC values for the reaches contained in either the testing or training datasets, as well as the feature contributions for each explanatory variable. 3. "Model_predictions.csv": Contains predicted SC, predicted/expected (or P/E) ratios, and departure categories for all non-tidal reach IDs in the CBW for the four time periods.