We developed a methodology to estimate maximum brine injection rates in subsurface formations across wide geographic areas using inverse modeling-based optimization techniques. We first defined geographic areas where groundwater was too saline to meet the standard for drinking water and where sufficient confining units existed above and below the injection layers. We then assumed concurrent brine injection into a system of wells on a consistent 25 km x 25 km spacing across the entire modeled area. Taking advantage of symmetry, we represented each 25 km x 25 km injection area as a 12.5 km-long one-dimensional radial model, divided into 100 logarithmically-sized grid blocks. A single layer of grid blocks was used because homogenous porous media were assumed. Brine injection was simulated into the leftmost (innner) grid block, and the injection rate was automatically adjusted to meet a maximum pressure buildup to 80% of the fracturing pressure, estimated as the least principal stress, at the injection location. A secondary constraint of 1 bar maximum pressure increase at the right-most (far-field boundary) grid block after 50 years of injection was applied. We demonstrated this method on three stratigraphic layers that overlie the Mt. Simon Sandstone (MSS) in the Illinois Basin, as well as in the MSS itself, because the MSS is a well-known CO2 injection target with a large estimated CO2 storage capacity. CO2 storage in the MSS could be optimized by extracting brine from that formation and injecting it elsewhere, so the brine injection rates estimated with the models contained herein could help to refine CO2 storage capacity estimates.

Two synoptic sampling campaigns were conducted near Breckenridge, Colorado, to quantify metal loading to Illinois Gulch, a tributary of the Blue River. The first campaign, conducted in August 2016, was designed to determine the degree to which Illinois Gulch loses water to the underlying mine workings, and to determine the effect of these losses on observed metal loads. The second campaign, conducted in September 2017, was designed to evaluate metal loading within Iron Springs, a subwatershed that was responsible for the majority of the metal loading observed in 2016. A continuous, instream injection of a sodium bromide (NaBr) tracer was initiated at the head of the respective study reaches several days prior to both synoptic sampling campaigns and maintained throughout the duration of each study. Bromide concentrations were subsequently used to determine streamflow in gaining stream reaches using the tracer-dilution method, and as an indicator of hydrologic connections between the Illinois Gulch and subsurface mine workings. Streamflow losses to the mine workings were quantified using a series of magnesium chloride slug additions conducted in August 2016, wherein specific conductivity readings were used as a surrogate for the tracer concentration. Study results indicate that Illinois Gulch loses water in the vicinity of the Puzzle Extension Shaft, and that water leaving the stream enters the subsurface mine workings. These losses are evidenced by the results of the slug additions and the elevated bromide concentrations observed at a collapsed mine portal in the Iron Springs subwatershed (Willard Adit 1). The primary sources of metal loading to the overall Illinois Gulch study reach include diffuse springs and groundwater near the toe of the Iron Springs mine dump and Willard Adit 1. This data release consists of 8 tables: Table 1, Locations of sampling sites for the 2016 and 2017 campaigns Table 2, Synoptic sampling results, August 18, 2016 Table 3, Synoptic sampling results, September 7, 2017 Table 4, Streamflow measurements, August 2016 Table 5, Bromide time series, August 2016 and September 2017 Table 6, Slug addition conductivity data, August 2016 Table 7, Slug addition results, August 2016 Table 8, Spatial profiles of streamflow and metal load, August 2016 and September 2017

Studies of chemical composition of natural brines from rock formations in the Illinois Basin as part of the University of Illinois deep direct-use feasibility study.

Studies of chemical composition of natural brines from rock formations in the Illinois Basin as part of the University of Illinois deep direct-use feasibility study.

This data release describes one (1) Microsoft Excel table of lumped parameter models of groundwater age for groundwater discharging to the South Loup River, Nebraska. The table (LPMAgeResults) includes final models of groundwater age and metrics by calibration of lumped parameter models to tracer concentrations using TracerLPM software (Jurgens and others, 2012). Interpreted results presented here were used to guide hydrologic interpretations of groundwater sources and flow paths of groundwater discharging to the South Loup River, NE.

This data release includes a polygon shapefile of grid cells attributed with values representing the simulated base-flow, evapotranspiration, and groundwater-storage depletions as a percentage of hypothetical well pumpage for the 2011-2060 time period. Depletions were simulated by the Phase-Three Elkhorn-Loup Model (ELM), constructed using MODFLOW-NWT (Niswonger and others, 2011). Each polygon represents one model grid cell, with pumping specified from either layer one or layer two of the model. All values are estimates and approximations. The phase three ELM simulated the High Plains aquifer in north-central Nebraska from predevelopment (pre-1895) through 2060 (Flynn and Stanton, 2018). The simulation was calibrated using an automated parameter-estimation method to optimize the fit of simulation outputs to three sets of calibration targets: estimated 1940 groundwater levels and base flows (representing pre-1940 conditions), 1940-through-2010 monthly groundwater levels, and 1940 through 2010 monthly estimated base flows. The calibrated simulation was used to estimate volumetric ratios of the reductions in base flow, evapotranspiration, and groundwater storage to the total volume of water pumped from a hypothetical well for a 50-year future time period. Ratios were then multiplied by 100 to obtain percentages. The 50-year period was selected because base-flow depletion percentages for 40- to 50-year periods are the basis of groundwater and surface-water management decisions in Nebraska.

This data release includes a polygon shapefile of grid cells attributed with values representing the simulated base-flow, evapotranspiration, and groundwater-storage depletions as a percentage of hypothetical well pumpage for the 2006-2055 time period. Depletions were simulated by the Phase-Two Elkhorn-Loup Model (ELM), constructed using MODFLOW-2005 (Harbaugh, 2005) with the Groundwater Vistas, version 5, software (Environmental Simulations, Inc., 2009). Each polygon represents one model grid cell. All values are estimates and approximations. The Phase-Two ELM simulated the High Plains aquifer in north-central Nebraska from predevelopment (pre-1895) through 2055 (Stanton and others, 2010). The simulation was calibrated using an automated parameter-estimation method to optimize the fit of simulation outputs to three sets of calibration targets: estimated 1939 groundwater levels and base flows (representing pre-1940 conditions), 1945-through-2005 decadal groundwater-level changes, and 1940-through-2005 annual base flows. The calibrated simulation was used to estimate volumetric ratios of the reductions in base flow, evapotranspiration, and groundwater storage to the total volume of water pumped from a hypothetical well for a 50-year future time period. Ratios were then multiplied by 100 to obtain percentages. The 50-year period was selected because base-flow depletion percentages for 40- to 50-year periods are the basis of groundwater and surface-water management decisions in Nebraska.

This data release includes a polygon shapefile of grid cells attributed with values representing the simulated base-flow, evapotranspiration, and groundwater-storage depletions as a percentage of hypothetical well pumpage for the 2011-2060 time period. Depletions were simulated by the Phase-Three Elkhorn-Loup Model (ELM), constructed using MODFLOW-NWT (Niswonger and others, 2011). Each polygon represents one model grid cell, with pumping specified from either layer one or layer two of the model. All values are estimates and approximations. The phase three ELM simulated the High Plains aquifer in north-central Nebraska from predevelopment (pre-1895) through 2060 (Flynn and Stanton, 2018). The simulation was calibrated using an automated parameter-estimation method to optimize the fit of simulation outputs to three sets of calibration targets: estimated 1940 groundwater levels and base flows (representing pre-1940 conditions), 1940-through-2010 monthly groundwater levels, and 1940 through 2010 monthly estimated base flows. The calibrated simulation was used to estimate volumetric ratios of the reductions in base flow, evapotranspiration, and groundwater storage to the total volume of water pumped from a hypothetical well for a 50-year future time period. Ratios were then multiplied by 100 to obtain percentages. The 50-year period was selected because base-flow depletion percentages for 40- to 50-year periods are the basis of groundwater and surface-water management decisions in Nebraska.

Groundwater residence times and flow path lengths were simulated for two major aquifers of the Mississippi embayment region using particle tracking (Pollock, 2012; Starn and Belitz, 2018) in a regional groundwater-flow model (Haugh and others, 2020). The Mississippi embayment physiographic region includes two principal aquifer systems: the surficial aquifer system, which is dominated by the Quaternary Mississippi River Valley alluvial aquifer (MRVA), and the Mississippi embayment aquifer system, which includes deeper Tertiary aquifers and confining units. The groundwater residence time simulation focused on the MRVA and two hydrogeologic units of the Claiborne Group (CLBG) from the deeper system, including the middle Claiborne aquifer (MCAQ) and lower Claiborne aquifer (LCAQ). A previously published groundwater flow model of the Mississippi embayment regional aquifer system provided the flow field for this analysis (Clark and Hart, 2009; Clark and others, 2011; and Haugh and others, 2020). Raster files were produced for seven model layers following the hydrogeologic framework for the MODFLOW groundwater-flow model of the Mississippi embayment from Clark and Hart (2009): one for the MRVA and six for the middle and lower Claiborne aquifers including four representing the MCAQ (layers 5 – 8) and two representing the LCAQ (layers 9 and 10). To determine the groundwater residence time, particles were distributed in model layers representing these aquifers using a volume-weighted algorithm then back-tracked until the particles exited the aquifer system, usually at the water-table surface. Particles were tracked under transient hydrologic conditions from March 31, 2014 backwards to January 1, 1870, then under steady-state conditions until they exited the aquifer system. The simulated residence time of each particle is the time the particle took to travel backwards from its initial location in the aquifer to its source of origin. Groundwater-residence time metrics were generated by statistically summarizing individual particles that started within each model cell. The flow-model grid resolution of one square mile was used to simulate groundwater residence times. The data were then resampled to a 1-square kilometer resolution of the National Hydrologic Grid (Clark and others, 2018). Computed metrics included the minimum, mean, maximum, standard deviation, as well as the 10th-, 20th-, 30th-, 40th-, 50th-, 60th-,70th-, 80th-, and 90th-percentiles along with the minimum, median, and maximum flow path length. Additionally, the portion of young groundwater (< 65 years old) and the mean residence time of the young portion were computed.

The hydrogeologic framework for the Lake Michigan Basin model was developed by grouping the bedrock geology of the study area into hydrogeologic units on the basis of the functioning of each unit as an aquifer or confining layer within the basin. Available data were evaluated based on the areal extent of coverage within the study area, and procedures were established to characterize areas with sparse data coverage. Top and bottom altitudes for each hydrogeologic unit were interpolated in a geographic information system for input to the model and compared with existing maps of subsurface formations. Fourteen bedrock hydrogeologic units, making up 17 bedrock model layers, were defined, and they range in age from the Jurassic Period red beds of central Michigan to the Cambrian Period Mount Simon Sandstone. Each hydrogeologic unit is referred to as its model layer number as represented in the report U.S. Geological Survey Scientific Report 2009-5060 (SIR2009-5060). They are listed below for reference as to the model layer number, and the hydrogeoloigc unit name. Dataset values represent the bottom of the layer. LSD Land surface L1_3 Quaternary unit (Bottom of Quaternary unit is Layer 3 in the model) L4 Jurassic unit L5 Upper Pennsylvanian unit L6 Lower Pennsylvanian unit L7 Michigan Formation unit L8 Marshall Formation unit L9 Devonian-Mississippian unit L10_12 Silurian-Devonian unit (Bottom of Silurian-Devonian unit is Layer 12 in the model) L13 Maquoketa Formation unit L14 Sinnipee Formation unit L15 St. Peter Formation unit L16 Prairie du Chien-Franconia unit L17 Ironton-Galesville unit L18 Eau Claire unit L19_20 Mt Simon Formation unit (Bottom of Mt Simon Formation unit is Layer 20 in the model) The Lake Michigan Basin groundwater model is discretized into a grid of 391 by 261 cells. The model has 20 layers: 3 that simulate the glacial and unconsolidated sediments and 17 that simulate the bedrock units. The model provides additional detail in the area of greatest interest, in this case, the Lake Michigan Basin, by use of smaller grid spacing in the innermost model domain compared with the grid spacing at the model boundaries. The smallest interior grid cells are 5,000 by 5,000 ft. At the model boundaries, the size of grid cells reaches approximately 68,930 ft (13 mi) from north to south by 116,490 ft (22 mi) from east to west. The grid cells each have values for the altitude to the bottom of each layer. The layer numbers are from top to bottom of the aquifer system. Three hydrogeologic units are represented by the multiple layers