Machine-learning predictions of groundwater specific conductance in the Mississippi Alluvial Plain, south-central United States, with evaluation of regional geophysical aerial electromagnetic data as explanatory variables

Courtney D. Killian; Katherine J. Knierim

doi:10.3133/sir20235099

Home/Archive/USGS

USGS · sir20235099

Machine-learning predictions of groundwater specific conductance in the Mississippi Alluvial Plain, south-central United States, with evaluation of regional geophysical aerial electromagnetic data as explanatory variables

Courtney D. Killian·Katherine J. Knierim

View original source ↗Read source document ↗Export metadata ↓

Abstract

The Mississippi Alluvial Plain, located in the south-central United States, is undergoing long-term groundwater-level declines within the surficial Mississippi River Valley alluvial aquifer (hereinafter referred to as “alluvial aquifer”), which has raised concerns about future groundwater availability. In some parts of the alluvial aquifer, groundwater availability for common uses such as irrigation, public supply, and domestic use is limited by quality (for example, high salinity) rather than quantity of water stored in the aquifer. The Mississippi Alluvial Plain region has an abundance of water-quality measurements in the alluvial aquifer and deeper aquifers; however, large areas lack direct measurements of salinity to evaluate regional groundwater availability. Statistical models can interpolate between wells to fill in spatial data gaps. In 2021, the U.S. Geological Survey trained two boosted regression tree (BRT) machine-learning models on specific conductance data available between 1942 and 2020 to predict spatially continuous surfaces of groundwater salinity at multiple depths for the alluvial aquifer and deeper aquifers. Well construction information, water levels, and surficial variables such as geomorphology and soils were included as explanatory variables in this baseline model. Additionally, subsurface electrical resistivity data from the first aquifer-wide aerial electromagnetic (AEM) survey for the region were incorporated to create a geophysical model. This work expands on prior BRT salinity predictions of the alluvial aquifer and extends predictions south to the Gulf of Mexico, where groundwater salinity is high. AEM survey data were not available for the southern extent of the alluvial aquifer at the time of modeling. A BRT model was trained without (baseline) and with (geophysical) AEM variables to test the ability of the models to predict salinity where explanatory data are missing and response data are sparse. Additionally, model sensitivity to AEM survey data was evaluated to better understand how AEM variables influence specific conductance predictions. Model performance was improved with the addition of geophysical data, which added three-dimensional information, thereby improving salinity predictions at depth. Groundwater specific conductance predictions can help inform other geophysical investigations in the southern extent of the study area, where high groundwater specific conductance can obfuscate changes in aquifer sediment resistivity and could limit groundwater resources for agricultural, public supply, and domestic uses.

Explore related subjects

Scientific Investigations Report Report

Source-reported bounding extent: 28.689695810736353° to 37.89139322749202° latitude; -92.8114869520446° to -87.62594007704502° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Courtney D. Killian, Katherine J. Knierim. 2023-11-17. Machine-learning predictions of groundwater specific conductance in the Mississippi Alluvial Plain, south-central United States, with evaluation of regional geophysical aerial electromagnetic data as explanatory variables. https://doi.org/10.3133/sir20235099

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Monitoring recreation on federally managed lands and waters—Aspects of visitor use

Federally managed public lands and waters receive around 1 billion recreational visits each year. Data on these visitors can aid in guiding policy decisions, managing resources effectively, and communicating the economic contributions of lands and waters. This report explores the methods used by agencies to collect data on aspects of recreational visitor use to Federal lands and waters (apart from visitation numbers, which are the focus of a companion publication). Aspects of recreational visitor use include visitor demographics, recreational activity participation, visitor satisfaction, visitor attitudes and experiences, trip characteristics, and economic contributions. We review practices used to understand aspects of visitor use across seven Federal agencies, revealing similarities such as the use of visitor intercept surveys and coverage of similar topic areas, and differences in how survey programs are operationalized and how specific questions on visitor surveys are worded. We also evaluate emerging technologies, such as geolocated social media and mobile device location data, for their potential to aid in understanding aspects of visitor use. This report concludes with potential opportunities to enhance data collection and coordination, ensuring cost-effective data collection and informed decision making.

Scientific Investigations Report↗

USGS2026

Evaluation of best management practices at an edge-of-field site in the Eagle Creek watershed, Ohio, 2012–20

In 2010, the U.S. Geological Survey worked in partnership with the Great Lakes Restoration Initiative and the Natural Resources Conservation Service to identify farm fields in priority watersheds—watersheds critically important to Great Lake health. Intensive best management practices were implemented in five States to test the efficacy of best management practices, also referred to as agricultural conservation practices, to reduce sediment and nutrient runoff on agricultural fields. These fields were chosen as representatives of the priority watersheds because their farming practices and geographical conditions were common among farms in those watersheds. A farm site located near Findlay, Ohio, in the Eagle Creek watershed used two conservation practices, cover crop (Natural Resources Conservation Service practice standard 340) and variable rate technology (VRT; Natural Resources Conservation Service practice standard 590), which represent common practices in the Maumee priority watershed. This study monitored surface runoff and subsurface tile runoff at the Eagle Creek watershed site. The effects of cover crop and VRT nutrient application on sediment and nutrient runoff were assessed from October 2012 to September 2020. Cover crops were applied in fall of 2016, 2017, and 2018, coinciding with VRT applications. The following parameters were analyzed as part of the study: runoff discharge; peak discharge; and concentrations and loads of total phosphorus, particulate phosphorus, dissolved reactive phosphorus, and suspended sediment. High peak runoff events disproportionally affected nutrient concentrations and loads in runoff. A threshold of the 85th percentile of the peak discharge was chosen to assess the effect of cover crop and VRT nutrient application on nutrient mitigation at each gage. Runoff events below the 85th percentile were considered small, and runoff events over the 85th percentile were considered large. Concentrations from collected water-quality samples were used for the analysis; estimated concentrations were excluded from the statistical evaluation. All surface runoff parameters during small runoff events, except surface runoff volume, were significantly lower in the period with cover crops and VRT implementation. Contrarily, subsurface runoff parameters during small events were not statistically different between the periods with and without cover crops and VRT implementation. Surface and subsurface parameters during large events were also not statistically different between the periods with and without cover crops and VRT implementation. These results indicate cover crops and VRT may improve water quality during small runoff events, but additional best management practices that mitigate large runoff events may lead to greater water-quality improvement given the contribution of large events to overall losses.

Ohio↗

USGS2026

Continuous monitoring and temporal variability of fluorescence of dissolved organic matter in karst groundwater of the Edwards aquifer, south-central Texas, 2019–24

Widespread urbanization on the Edwards aquifer recharge zone in south-central Texas has prompted the development of new approaches and tools for evaluating the current and future status of water quality in the San Antonio and Barton Springs segments of the Edwards aquifer. The U.S. Geological Survey, in cooperation with the San Antonio Water System and the City of Austin, applied continuous monitoring of fluorescence of dissolved organic matter (fDOM) to characterize the sources and transport of dissolved organic matter and associated constituents in the Edwards aquifer. Continuous fDOM time-series data were adjusted using empirically derived temperature, turbidity, and inner-filter-effect correction factors prior to analysis and interpretation. Fully corrected fDOM time-series data from the study sites were evaluated for short- and long-term temporal variability in fDOM in the context of the hydrologic and climatic conditions that occurred during the 5-year study period to understand vulnerability of the aquifer to potential contaminants in recharge. Study results indicate that fDOM is an effective proxy for dissolved organic carbon that, in turn, is indicative of the influx of recent surface water and associated contaminants, and thus, aquifer vulnerability to those contaminants. Although other continuously monitored water-quality parameters provide insights into changes in water quality in response to varying hydrologic conditions, the quantitative relation between fDOM and dissolved organic carbon directly indicates the timing and magnitude of pulses of organic constituents derived from the land surface, whereas other water-quality parameters do not.

Texas↗