This groundwater-flow model archive/data release contains the model input and output files for 1) edited versions of four of the five NAWQA steady- state, inset MODFLOW-NWT models of regional model of Lake Michigan Basin (https://doi.org/10.3133/sir20185038) and 2) general models simulating the same four basins as the four inset models. Two HUC8 basins in the lower peninsula of Michigan (Kalamazoo (KALA) and Boardman-Charlevoix (BOARD) basins) and two HUC8 basins in Wisconsin (Upper Fox (UFOX) and Manitowoc-Sheboygan (MANI) basins) are represented in the inset and genera-simulation models. The inset models are designed to serve as a training area for metamodels to estimate groundwater age in glacial wells. The construction and details of the original four inset models are outlined in the U.S. Geological Survey Scientific Investigations Report 2018-5038 (https://doi.org/10.3133/sir20185038), and the construction and details of the general models are outlined in the Water Resources Research journal article (https://doi.org/10.1029/2017WR021531). The original four inset models are archived in the data release at https://doi.org/10.5066/F76D5R5V. Groundwater withdrawals from wells in the original four inset models were removed in the inset models in this archive because the general models did not have wells and to be able to compare the results from the two types of models in the archive. The boundary conditions of these “pre-development” versions of the inset models were changed from constant-head boundaries (reflecting 2005 conditions) to no-flow boundaries. The general-simulation models apply an innovative modeling approach that allows for rapid,automated construction and calibration of models at a scale appropriate to the problem at hand (https://doi.org/10.3133/sir20215142). Results from the four general models in this archive were compared to results from the edited versions of the four inset models to evaluate the degree to which the general models reproduce behavior simulated by the inset models that use conventional flow modeling techniques. The underlying directories contain all the input and output files for the MODFLOW-NWT simulations and MODPATH particle tracking analysis for the edited versions of four inset models, and the four general models simulating the same four basins as the inset models described in the USGS Scientific Investigations Report (https://doi.org/10.3133sir20215142). The MODFLOW-NWT (v 1.0.9) and MODPATH 6 (version 6.0.1) executables and source codes, various ancillary python scripts written for this project, and model geospatial data are also included in the archive. Descriptions of the data in each subdirectory are provided to facilitate understanding of this this model archive. File descriptions are provided for select input and output files to provide additional information that may be of use for understanding this this model archive.

A previously developed three-dimensional groundwater flow model (https://doi.org/10.3133/sir20125183) of the three primary aquifers in the region north of Aberdeen, South Dakota, was revised to assist the City of Aberdeen with water-resource planning. The principal aquifers are the Elm, Middle James, and Deep James. The numerical model is intended to be used to (1) simulate hydrologic scenarios of interest to groundwater managers and to advance the understanding of groundwater budgets and components including recharge, discharge, and aquifer storage, (2) compute historical and projected system response to natural and anthropogenic stresses, and (3) evaluate potential withdrawal scenarios to assist with water-management decisions. The three-dimensional groundwater-flow model was developed using the USGS's numerical modeling software, MODFLOW-NWT. The model included both steady-state (mean) and transient (temporally varying) conditions. Steady-state data were from October 1, 1974 to September 30, 2009, and the transient time period included data from October 1, 1974 to September 30, 2015. The model was calibrated by attempting to match simulated and measured or estimated hydraulic heads, differences in hydraulic heads between aquifers, and stream base flow. Water budgets for the model's layers area were produced with ZONEBUDGET. This USGS data release contains all of the input and output files for the model described in the associated model documentation report (http://doi.org/10.3133/sir20185137). The data release also includes (1) MODFLOW-NWT (version 1.1.4) source code, and (2) ZONEBUDGET (3.01) source code.

A previously developed three-dimensional groundwater flow model (https://doi.org/10.3133/sir20125183) of the three primary aquifers in the region north of Aberdeen, South Dakota, was revised to assist the City of Aberdeen with water-resource planning. The principal aquifers are the Elm, Middle James, and Deep James. The numerical model is intended to be used to (1) simulate hydrologic scenarios of interest to groundwater managers and to advance the understanding of groundwater budgets and components including recharge, discharge, and aquifer storage, (2) compute historical and projected system response to natural and anthropogenic stresses, and (3) evaluate potential withdrawal scenarios to assist with water-management decisions. The three-dimensional groundwater-flow model was developed using the USGS's numerical modeling software, MODFLOW-NWT. The model included both steady-state (mean) and transient (temporally varying) conditions. Steady-state data were from October 1, 1974 to September 30, 2009, and the transient time period included data from October 1, 1974 to September 30, 2015. The model was calibrated by attempting to match simulated and measured or estimated hydraulic heads, differences in hydraulic heads between aquifers, and stream base flow. Water budgets for the model's layers area were produced with ZONEBUDGET. This USGS data release contains all of the input and output files for the model described in the associated model documentation report (http://doi.org/10.3133/sir20185137). The data release also includes (1) MODFLOW-NWT (version 1.1.4) source code, and (2) ZONEBUDGET (3.01) source code.

A MODFLOW-NWT model was used to simulate the water budget for Haskell Lake and Tower Creek in WI using the Lake, Streamflow Routing, and Unsaturated Zone Flow packages. Particle tracking was performed with the MODFLOW solution (using MODPATH 6). This USGS data release contains all of the input and output files for the simulations described in the associated model documentation report (https://doi.org/10.3133/sir20205024).

A MODFLOW-NWT model was used to simulate the water budget for Haskell Lake and Tower Creek in WI using the Lake, Streamflow Routing, and Unsaturated Zone Flow packages. Particle tracking was performed with the MODFLOW solution (using MODPATH 6). This USGS data release contains all of the input and output files for the simulations described in the associated model documentation report (https://doi.org/10.3133/sir20205024).

Groundwater flow models have the potential to predict spatial groundwater discharge dynamics within river networks, but models are often not evaluated against discharge dynamics. The objective of this study was to understand the variation in simulated discharge dynamics (discharge location, flowpath depth, and subsurface travel time) for models with common, but varying frameworks and assumptions. The University of Connecticut in collaboration with the United States Geological Survey developed a groundwater flow model (MODFLOW-NWT) for the Farmington River Watershed (1,570 km2) in the northeastern United States and systematically varied the type of typical calibration data (well head and stream elevation); calibration parameters; parameters related to permeability of the surficial materials, bedrock, and riverbed sediments; control of river-aquifer exchange directionality; and model resolution. Each model variation has an associated particle tracking (MODPATH) model. Subsequent work, not described in this model archive, compared with simulated spatial patterns of groundwater discharge with patterns observed with hand-held thermal infrared imagery. This dataset contains model inputs and outputs, post-processing python scripts, and pest calibration input files for 12 model variations described in the associated journal article (https://doi.org/10.1029/2020WR028027)

Groundwater flow models have the potential to predict spatial groundwater discharge dynamics within river networks, but models are often not evaluated against discharge dynamics. The objective of this study was to understand the variation in simulated discharge dynamics (discharge location, flowpath depth, and subsurface travel time) for models with common, but varying frameworks and assumptions. The University of Connecticut in collaboration with the United States Geological Survey developed a groundwater flow model (MODFLOW-NWT) for the Farmington River Watershed (1,570 km2) in the northeastern United States and systematically varied the type of typical calibration data (well head and stream elevation); calibration parameters; parameters related to permeability of the surficial materials, bedrock, and riverbed sediments; control of river-aquifer exchange directionality; and model resolution. Each model variation has an associated particle tracking (MODPATH) model. Subsequent work, not described in this model archive, compared with simulated spatial patterns of groundwater discharge with patterns observed with hand-held thermal infrared imagery. This dataset contains model inputs and outputs, post-processing python scripts, and pest calibration input files for 12 model variations described in the associated journal article (https://doi.org/10.1029/2020WR028027)

A three-dimensional, steady-state groundwater-flow model representing 2003-13 mean hydrologic conditions was developed and calibrated to assess groundwater and lake-water exchanges and the effects of groundwater withdrawals and precipitation on water levels in lakes in the northeast Twin Cities Metropolitan Area, Minnesota. The USGS groundwater-flow model program MODFLOW-NWT version 1.0.8 was used to simulate groundwater flow in the approximately 1,000-square-mile area of the northeast Twin Cities Metropolitan Area and western Wisconsin. Water levels were below normal for several lakes in the northeast Twin Cities Metropolitan Area during 2003 through 2013. Previous periods of low lake-water levels generally correlate with periods of below-normal precipitation but increases in groundwater withdrawals and land-use changes have put into question whether recent lake-water-level declines are due solely to declines in precipitation. The groundwater-flow model was developed to provide a thorough understanding of regional groundwater and surface-water exchanges under different groundwater withdrawal and precipitation scenarios. This USGS data release contains all of the input and output files for the simulations described in the associated U.S. Geological Survey Scientific Investigations Report 2016-5139B (https://doi.org/10.3133/sir20165139B). This data release also contains input and output data for ancillary soil-water balance (SWB) models used to simulate runoff to a subset of lakes and areal groundwater recharge.

In the Chenango River Basin, New York, glacial valley-fill unconfined and confined aquifers are an important source of drinking-water supplies. The U.S. Geological Survey developed three-dimensional groundwater-flow models (MODFLOW-NWT code) of three study areas in the basin to improve understanding of groundwater flow and delineate areas contributing recharge to 16 production wells as part of an effort to protect the source of water to these wells. The modeled areas were the Cortland study area in Cortland County (CCSA), Greene study area (GSA), and the Cincinnatus study area (CSA). Production wells in CCSA tap unconfined aquifers, whereas the production wells in GSA and CSA are screened in confined aquifers. The CCSA, which has an extensive calibration dataset, was calibrated by nonlinear regression and included a prediction uncertainty analysis. The GSA and CSA were manually calibrated because of the sparse model design and calibration information. The groundwater-flow models along with particle tracking (MODPATH code) were used to simulate water levels, base flows, water budgets, and flow paths. The USGS data release contains all the input and output files and source codes for the simulations described in the associated report: https//doi.org/10.3133/sir20215083.

Three groundwater flow models (KMS model, Pumping Test model, and Modified LMB model) were developed for the Kettle Moraine Springs State Fish Hatchery using the U.S. Geological Survey codes MODLOW-NWT, GWM-2005, MODFLOW-2005, and SEAWAT-2000. The KMS inset model was derived from a published USGS regional Lake Michigan Basin model, and was constructed to simulate groundwater pumping from the semi-confined Silurian bedrock aquifer. The LMB modified model is a version of the published Lake Michigan Basin model that was modified with aquifer parameters refined in an area around the hatchery. The Pumping Test model, was constructed to evaluate a pumping test conducted in the Cambrian-Ordovician aquifer system and to simulate groundwater pumping from this deep bedrock aquifer at the hatchery.