Supporting information for publication titled "Method to Rapidly Track Salinization in Surface Waters"

In streams located throughout temperate regions of the United States, salinity and chloride concentrations have been increasing over time, which can threaten stream health and drinking water quality. During snow melt events, chloride concentrations can rapidly increase and exceed ecological thresholds, making it important to frequently determine chloride concentrations. However, traditional sampling poses difficulties to rapidly determining chloride concentrations due to timing constraints. To supplement traditional sampling, specific conductance, which is monitored remotely and at 15-minute intervals in many streams, has been used as a primary predictor of chloride concentrations in surrogate regression models. Prior to developing surrogate models, geochemical modeling can be used to calculate the percent contribution of chloride to specific conductance to determine if site-specific regression models are appropriate. In Chester County, Pennsylvania, chloride surrogate models have been previously published for three stream sites using streamflow and specific conductance as explanatory variables. The previously published models use data collected through late 2022/early 2023. However, these models were recalculated to include data through 2024. This data release contains supporting information for the publication titled "Method to Rapidly Track Salinization in Surface Waters", and included in the data release is code to complete the geochemical analysis, compare the published models both visually and statistically to newly calculated models that incorporate data collected through 2024, and to create figures evaluating the data.
Modeling data and code are included for 3 stream sites:

Valley Creek at PA Turnpike Bridge near Valley Forge, PA (USGS station 01473169) with discrete water-quality data from nearby downstream site Valley Creek at Wilson Road near Valley Forge, PA (USGS station 01473170);
White Clay Creek near Strickersville, PA (USGS station 01478245); and
Brandywine Creek at Chadds Ford, PA (USGS station 01481000).

 
This data release contains 11 files:
1. "READ_ME_SuppInfo_MethodTrackSalinization.txt": This file contains important information about the data release files and should be read prior to using the data and code.
2. "GEOREF_SuppInfo_MethodTrackSalinization.txt": This file contains information about the sites used in the analysis.
3. "PHREEQC_salinitycalcs.R": This is the code to complete the geochemical analysis using PHREEQC.
4-6. "phreeq_xxxxxxxx.csv": These are the input files for the "PHREEQC_salinitycalcs.R" script. These input files contain discrete water-quality data. xxxxxxxx refers to the three stream sites: USGS-01473170, USGS-01478245, USGS-01481000.
7. "models_ancova.R": This is the code to develop the published and newly calculated chloride surrogate regression models. There is also code to complete an Analysis of Covariance to statistically compare the surrogate models.
8. "Prediction_Intervals_Plots_Flow_Duration_Curves.R": This is the code to visually compare the published and newly calculated surrogate regression models, as well as to create flow duration curves.
9-11. "combined_data_xxxxxxxx.csv": Three input files for the "models_ancova.R" script and the "Prediction_Intervals_Plots_Flow_Duration_Curves.R" script. These files contain chloride, specific conductance, and streamflow data. xxxxxxxx refers to the three stream sites: USGS-01473169, USGS-01478245, USGS-01481000.