This data release contains: (1) ASCII grids of predicted probability of elevated arsenic in groundwater for the Northwest and Central Minnesota regions, (2) input arsenic and predictive variable data used in model development and calculation of predictions, and (3) ASCII files used to predict the probability of elevated arsenic across the two study regions. The probability of elevated arsenic was predicted using Boosted Regression Tree (BRT) modeling methods using the gbm package in R Studio version 3.4.2. The response variable was the presence or absence of arsenic >10 µg/L, the U.S. Environmental Protection Agency’s maximum contaminant level for arsenic, in 3,283 wells located throughout both study regions (1,363 in the Northwest region and 1,920 in the Central). The original database used to develop the BRT model consisted of 127 predictor variables which included well characteristics, land use, soil properties, aquifer properties, depth to water table, and predicted nitrate. After optimization steps, a final database of 33 predictor variables was used to predict the occurrence of elevated arsenic across the two study regions.

This data release provides total and aqueous arsenic (As) determinations and associated field readings collected from groundwater sampled from 254 newly constructed private residential wells from 2014-2016. The study focuses on three regions of Minnesota that differ geologically: south-central (herein called central), northwest, and northeast. These study regions were chosen due to their prevalent elevated As concentrations in drinking water. Each of the 254 wells were sampled in three rounds by the Minnesota Department of Health (MDH). The timing of the three sampling rounds was (1) immediately or shortly after well construction (round 1); (2) 3-6 months after initial sample collection (round 2); and (3) 12 months after initial sample collection (round 3). During each round, samples were collected for both total and aqueous As. Physicochemical characteristics, including specific conductance, pH, dissolved oxygen, oxidation reduction potential, and temperature, were also measured to gage the well water stability prior to sample collection. Round 1 sampling was timed to co-occur and mimic well driller regulatory sampling. Drillers collected samples after well development from the drill rig groundwater pump or from the residential plumbing and the MDH sampler replicated the sample location and timing used by the driller. Sampling from the drill rig's groundwater pump occurred after the well was drilled and developed, when the water was visibly clear, with little visible sediment particles. Samples from plumbing were collected after the plumbing was flushed out and physicochemical characteristic readings stabilized. Round 2 and round 3 by MDH staff were collected only from plumbing. Samples collected from plumbing were taken from faucets, hydrants, or pressure tanks prior to filters or treatment systems.

This data release provides total and aqueous arsenic (As) determinations and associated field readings collected from groundwater sampled from 254 newly constructed private residential wells from 2014-2016. The study focuses on three regions of Minnesota that differ geologically: south-central (herein called central), northwest, and northeast. These study regions were chosen due to their prevalent elevated As concentrations in drinking water. Each of the 254 wells were sampled in three rounds by the Minnesota Department of Health (MDH). The timing of the three sampling rounds was (1) immediately or shortly after well construction (round 1); (2) 3-6 months after initial sample collection (round 2); and (3) 12 months after initial sample collection (round 3). During each round, samples were collected for both total and aqueous As. Physicochemical characteristics, including specific conductance, pH, dissolved oxygen, oxidation reduction potential, and temperature, were also measured to gage the well water stability prior to sample collection. Round 1 sampling was timed to co-occur and mimic well driller regulatory sampling. Drillers collected samples after well development from the drill rig groundwater pump or from the residential plumbing and the MDH sampler replicated the sample location and timing used by the driller. Sampling from the drill rig's groundwater pump occurred after the well was drilled and developed, when the water was visibly clear, with little visible sediment particles. Samples from plumbing were collected after the plumbing was flushed out and physicochemical characteristic readings stabilized. Round 2 and round 3 by MDH staff were collected only from plumbing. Samples collected from plumbing were taken from faucets, hydrants, or pressure tanks prior to filters or treatment systems.

This dataset provides well construction information provided by driller well logs. Well logs can be accessed on the Minnesota Well Index online (http://www.health.state.mn.us/ divs/eh/cwi/).

This dataset provides aqueous nitrate+nitrite, aqueous manganese, aqueous iron, and total sulfate measurements in groundwater samples from 254 newly constructed private residential wells between 2014 and 2016. The study focuses on three geologically distinct regions of Minnesota: central, northwest, and northeast. These study regions were chosen due to their prevalent elevated As concentrations in drinking water. Each of the 254 wells were sampled in three rounds by the Minnesota Department of Health (MDH). The timing of the three sampling rounds was (1) immediately or shortly after well construction (round 1); (2) 3-6 months after initial sample collection (round 2); and (3) 12 months after initial sample collection (round 3). During each round, samples were collected for both total and aqueous As, aqueous nitrate+nitrite, aqueous manganese, aqueous iron, and total sulfate. Physiochemical characteristics, including specific conductance, pH, dissolved oxygen, oxidation reduction potential, and temperature, were also measured to gage the well water stability prior to sample collection. Round 1 sampling was timed to co-occur and mimic well driller regulatory sampling. Drillers collected samples after well development from the drill rig groundwater pump or from the residential plumbing, and the MDH sampler replicated the sample location and timing used by the driller. Sampling from the drill rig’s groundwater pump occurred after the well was drilled and developed, when the water was visibly clear, with little visible sediment particles. Samples from plumbing were collected after the plumbing was flushed out and physiochemical characteristic readings stabilized. Round 2 and round 3 by MDH staff were collected only from plumbing. Samples collected from plumbing were taken from faucets, hydrants, or pressure tanks prior to filters or treatment systems.

This dataset provides aqueous nitrate+nitrite, aqueous manganese, aqueous iron, and total sulfate measurements in groundwater samples from 254 newly constructed private residential wells between 2014 and 2016. The study focuses on three geologically distinct regions of Minnesota: central, northwest, and northeast. These study regions were chosen due to their prevalent elevated As concentrations in drinking water. Each of the 254 wells were sampled in three rounds by the Minnesota Department of Health (MDH). The timing of the three sampling rounds was (1) immediately or shortly after well construction (round 1); (2) 3-6 months after initial sample collection (round 2); and (3) 12 months after initial sample collection (round 3). During each round, samples were collected for both total and aqueous As, aqueous nitrate+nitrite, aqueous manganese, aqueous iron, and total sulfate. Physiochemical characteristics, including specific conductance, pH, dissolved oxygen, oxidation reduction potential, and temperature, were also measured to gage the well water stability prior to sample collection. Round 1 sampling was timed to co-occur and mimic well driller regulatory sampling. Drillers collected samples after well development from the drill rig groundwater pump or from the residential plumbing, and the MDH sampler replicated the sample location and timing used by the driller. Sampling from the drill rig’s groundwater pump occurred after the well was drilled and developed, when the water was visibly clear, with little visible sediment particles. Samples from plumbing were collected after the plumbing was flushed out and physiochemical characteristic readings stabilized. Round 2 and round 3 by MDH staff were collected only from plumbing. Samples collected from plumbing were taken from faucets, hydrants, or pressure tanks prior to filters or treatment systems.

This data release contains data used to develop models and maps that estimate probabilities of exceeding various thresholds of arsenic concentrations in private domestic wells throughout the conterminous United States. Three boosted regression tree (BRT) models were developed separately to estimate the probability of private well arsenic concentrations exceeding 1, 5, and 10 micrograms per liter (µg/L). A random forest (RF) model was developed to estimate the most probable arsenic concentration category (≤5, >5 to ≤10, or >10 µg/L). The models use arsenic concentration data from private domestic wells located throughout the conterminous United States and independent variables that are available as geospatial data. The models were used to produce maps that are included in this data release. The model input data (predictor variables) that were used to make the maps are within a zipped folder (Map_Input_Data.zip) that contains 85 tif-raster files, one for each model predictor variable. The map probability estimates that are outputs from the model are in a zipped folder (Map_Output_Data.zip) that contains 13 tif-raster files, one model estimate map for each of the BRT models and four for the RF model, as well as 2 confidence interval maps for each BRT model.

The U.S. Geological Survey (USGS), in cooperation with the U.S. Centers for Disease Control and Prevention and the Maine Center for Disease Control and Prevention, assessed the physical and chemical characteristics and the occurrence, distribution, and oxidation state of inorganic arsenic in drinking water from selected domestic well-water supplies in Maine in 2001–2 and 2006–7. The data collected provide support for evaluating arsenic-removal efficiencies of household water-purification systems and provide information to State and local officials that can be used in determining a water-treatment approach for the removal of arsenic from drinking water.

Documented in this data release are data used to model and map the probability of arsenic being greater than 10 micrograms per liter in private domestic wells throughout the conterminous United States during drought conditions (Lombard and others, 2020). The model used to predict the probability of arsenic exceeding 10 micrograms per liter in private domestic wells was previously developed and documented by Ayotte and others (2017). Independent variables in the model include groundwater recharge and annual precipitation. In order to assess the impact of drought these variables were altered to simulate drought by reducing the 30-year average annual values by 25 and 50 percent. The impact of drought was also assessed by using groundwater recharge and precipitation values from the year 2012 when approximately 66 percent of the contiguous United States experienced drought. Data sources for groundwater recharge and precipitation for the year 2012 differ from those used in the original model and the drought simulations, therefore a 30-year average climate model was also produced using these new data sources (Thornton and others, 2018; Hay, 2019). Data are documented from the original model, the drought simulations with reduced values of groundwater recharge and precipitation, the year 2012 and the average annual precipitation and groundwater recharge from 1981 - 2010 from the new data sources. The model input data that were used to make the prediction maps are within a zipped folder (Prediction_Input_Data.zip) that contains 50 files, one for each model predictor variable. These include the predictor variables from the original model as well as the updated precipitation and groundwater recharge variables for the year 2012 and the average annual values based on the years1981 - 2010, and groundwater recharge and precipitation variables that were systematically decreased for drought simulations. The model prediction outputs are within a zipped folder (Prediction_Output_Data.zip) that contains 10 tif-format raster files, one for each of the eight drought simulations, one for the year 2012, and one for the updated average annual precipitation and groundwater recharge variables for 1981 - 2010. A third zipped folder (Change_Prob_Maps.zip) contains 10 tif-raster files that show the change in probability of arsenic exceeding 10 micrograms per liter in private domestic wells based on the drought simulations and the data used for the year 2012.

This data release contains a table of measured arsenic concentrations and associated model input variables used to test existing multivariate logistic regression models that predict the probabilities of arsenic concentrations exceeding threshold values of 1, 5, and 10 micrograms per liter in bedrock aquifers of New Hampshire. Location data are censored to the county level.