Modeling the surface water and groundwater budgets of the US using MODFLOW-OWHM

https://doi.org/10.1016/j.advwatres.2020.103682Get rights and content

Highlights

  • A generalized, calibrated set-up of the USGS MODFLOW-One World Hydrologic Model across CONUS is presented that includes human impacts.

  • The model reasonably simulates the spatial patterns of hydrologic fluxes across the model domain, with the largest errors occurring in regions with heterogeneous topography.

  • Water budget analysis highlights the variability in precipitation, recharge, and streamflow with evapotranspiration demonstrating the least interannual variability.

Abstract

Assessments of groundwater and surface water budgets at a large scale, such as the contiguous United States, often separately analyze the complex dynamics linking the surface and subsurface categories of water resources. These dynamics include recharge and groundwater contributions to streamflow. The time-varying simulation of these complex hydrologic dynamics, across large spatial and temporal scales, remains a scientific challenge due to the complexity of the processes and data availability. In this study, groundwater fluxes and surface hydrologic processes are simulated across the contiguous US for 1950-2010. The simulation estimates the monthly water budget components, such as groundwater recharge, surface runoff, and evapotranspiration; streamflow in major rivers is routed while accounting for groundwater exchange. Human impacts are included through groundwater pumping, and climate variability is included, including variability in precipitation, temperature and potential evapotranspiration. The simulated groundwater level and river discharge have strong correlation with USGS observation wells and streamflow gages, with R2 values of 0.992 and 0.946, respectively. The simulated evapotranspiration is compared with three other published estimation methods, showing that it is able to capture the magnitude and seasonality of evapotranspiration over the Mississippi River basin. As such, the model is able to reasonably simulate the surface and groundwater budgets over the US, allowing for questions of the relative importance of climate and human impacts to be explored in the future.

Introduction

Surface water and groundwater interact with one another, with surface hydrologic processes impacting groundwater recharge and groundwater levels determining baseflow in streams. The water budgets are dynamically affected by climate conditions, such as precipitation and temperature; human activities, such as water withdrawals from rivers, groundwater pumping, and land use change; and terrestrial processes such as plant water use. Estimating the surface and groundwater budgets is critical for quantifying water resources across large spatial and temporal scales, yet it remains a challenging task due to a lack of in situ observations of critical hydrologic processes and poor characterization of subsurface hydrogeologic properties. This study presents a contiguous US (CONUS) set-up of the MODFLOW-One-Water Hydrologic Model Version 2 (MF-OWHM2) (Hanson et al., 2014a; Boyce et al., 2020), which is a surface and groundwater model capable of simulating the hydrologic fluxes and storages for water budget assessment.

Climate, subsurface conditions, and human activities all impact groundwater levels. For example, during periods of low rainfall, groundwater levels decline due to a reduction in groundwater storage from low recharge rates (Wada et al., 2014) if the lateral groundwater flow is limited (de Graaf et al 2017, Condon and Maxwell, 2017). Furthermore, excessive groundwater pumping can threaten groundwater sources with high groundwater depletion (Döll et al., 2014b; Scanlon et al., 2012). In some regions, such as the Great Plains, groundwater levels increase slowly due to low recharge rate and relatively low hydraulic conductivity (Peterson et al., 2016). The sensitivity analysis of groundwater level is analyzed on global scale with respect to hydraulic conductivity, groundwater recharge, and surface water body elevation by Reinecke et al., (2019). To maintain the groundwater availability and baseflow to rivers, some global studies (Döll et al., 2014a; Pokhrel et al., 2012; Wada et al., 2010) analyzed the groundwater recharge rate and groundwater storage from water budget components; and other studies (Peterson et al., 2016; Faunt et al., 2009) analyzed the streamflow in regional aquifer systems.

Although it is widely accepted that understanding surface and groundwater fluxes and storages is critical for water resources planning, collecting extensive in situ observations, such as groundwater levels and subsurface geological properties, is prohibitively expensive over large regions. Remote sensing data can provide estimates of hydrologic fluxes over large spatial scales, but remote sensing has relatively short time series and can typically only observe surface processes. The exception is GRACE (Tapley et al., 2004), which can estimate groundwater storage change through gravity field changes, but only at coarse spatial and temporal resolutions. To overcome the remote sensing limitations numerical models can be utilized to estimate and reconstruct the hydrologic fluxes and stores, thereby overcoming the observational data limitations. These models can be useful tools to address critical questions: for instance, what is the impact of climate variability and groundwater pumping on groundwater availability.

Regional, hydrological and groundwater, models have been used to simulate numerically the surface water and groundwater hydrologic processes. For example, three regional models have been set up in detail over three major aquifers in the United States in Central Valley, High Plains, and Rio Grande aquifer systems. The first groundwater flow model simulates the Northern High Plains aquifer (Peterson et al., 2016) and uses MODFLOW with the Newton-Rhapson solver (Niswonger et al., 2011). The second is the Central Valley Hydrologic Model CVHM (Faunt et al., 2009) for California's Central Valley aquifer using MODFLOW-FMP2 (Schmid and Hanson, 2009). The forthcoming new version of CVHM (CVHM2) is being updated to use MF-OWHM2. A third model is the Rio Grande transboundary integrated hydrologic model (RGTIHM) that simulates portions of New Mexico, Texas, and Mexico (Hanson et al., 2020) using MF-OWHM2. These regional models provide a better understanding of the surface and subsurface regional water budgets and their dynamics. However, regional models typically cover a single aquifer, neglecting processes at the model boundaries. This means the streamflow, surface runoff, and lateral groundwater flow may be neglected at the boundaries of these regional models. These assumptions can misrepresent the modeled water budgets (Schaller and Fan, 2009; Krakauer et al., 2014), depending on the regional conditions such as topography, geology, and climate.

Continental and global scale groundwater models were developed (de Graaf et al., 2015; Fan et al., 2007; Maxwell et al., 2015) to not only solve the problem of lateral flow at model boundaries, but because they can provide a better understanding of the hydrology of an entire large-scale system. For example, large scale models can analyze the spatial variation of climate, geology, and topography and its effect on surface water and groundwater availability, and they can analyze the surface-groundwater interaction in different aquifers.

One of the first studies to simulate the groundwater depth across the contiguous United States was conducted by Fan et al. (2007), where the water table was estimated as the equilibrium of long-term climatic forcing on groundwater level under a steady state model, followed by a study of global observations of water table depth (Fan et al., 2013). These studies led to the ability to connect groundwater with other hydrologic fluxes, which has been investigated in several later studies over the United States. For example, Maxwell et al. (2015) used an integrated groundwater model (ParFlow, Maxwell and Miller, 2005) to analyze the surface and subsurface flow system over the majority of CONUS. To do a high spatial resolution run, the model was run under steady state conditions but can be run transiently as in Kollett (2009) and Maxwell et al. (2016). Other studies simulate the groundwater on the global scale using a model run at steady state (de Graaf et al., 2015) and transient (de Graaf et al., 2017), then analyzing the groundwater depletion (de Graaf et al., 2019). These models simulate recharge and river discharge from the global hydrological model PCR-GLOBWB (Wada et al., 2011), then simulates the groundwater lateral flow from MODFLOW. The MODFLOW One-Water Hydrologic Model (MF-OWHM2) used in this paper simulates both surface hydrologic processes and groundwater flow simultaneously in one model, overcoming the limitation of coupling two models.

Some of these large scale models assume the magnitude and direction of the fluxes are constant over time (Fan et al., 2007; Maxwell et al., 2015; de Graaf et al., 2015), thereby neglecting the temporal dynamics of seasonal and interannual climate variability. In addition, human activity, such as groundwater pumping, is often neglected. However, in recent studies (de Graaf et al., 2017; Condon et al., 2019; de Graaf et al., 2019) the groundwater pumping is simulated. In addition, incorporating streamflow routing is still a challenging component to model with a coupled groundwater model because of modeling limitations or the scarcity of needed data such as river hydraulic conductance over such a vast area such as the United States. These challenges have led some past studies to simplify the modeling of such a complex system by neglecting human activity and simplifying or neglecting the streamflow routing and its interaction with groundwater.

This study's objectives are 1) to evaluate the feasibility of using MF-OWHM2 at continental scales for water budget estimation and 2) to develop and validate a CONUS-wide MF-OWHM2 model set-up that explicitly simulates hydrologic processes that link the surface and groundwater hydrologic processes and human impacts. The focus is on capturing large-scale hydrologic patterns and human impacts to provide a baseline model for the hydrologic community, which can then expand the model to include more detailed processes relevant for specific research questions. This model is complementary to other US-focused modeling efforts, such as that of ParFlow, as it is more focused on including human impacts in a parsimonious framework that would not require high performance computing resources. In this paper, we first present the model set-up, including simplifications, the observation-based input data, and the assumptions made in some input variables. Then, the model validation and water budget results are presented, including both groundwater and surface water. The conclusions are presented in the final section.

Section snippets

Model and data

To simulate the surface and subsurface hydrology of the contiguous US (CONUS), this study uses MODFLOW-One-Water Hydrologic Model Version 2 (MF-OWHM2) (Boyce et al., 2020; Boyce, 2020), which is a modular modeling software developed by the U.S. Geological Survey and U.S. Bureau of Reclamation. MF-OWHM2 builds on the widely-used MODFLOW model (Harbaugh et al., 2000, Harbaugh, 2005), which simulates groundwater fluxes at a range of spatial and temporal scales. In addition to simulating

Model spin-up and calibration

For calibration purposes, a simplified 12-month model—using the seasonal averaged climate variables from 1950 to 2010—was created for the model spin-up and calibrated to estimate the groundwater level in an equilibrium state and to calibrate the extinction depth of groundwater-sourced evapotranspiration using the Farm Process version 4 (FMP) of MODFLOW-OWHM2. The simplified, 12-month model was run with the seasonal climatology, allowing for errors in initial conditions to dissipate. Evaluation

Groundwater validation

The full monthly model simulation from 1950 to 2010 with groundwater pumping is validated against the observed groundwater levels from the selected USGS wells for unconfined aquifers. The validation was implemented between USGS well measurements and the simulated groundwater level in the grid cell that contains that well for the months when a measurement was made. The criteria to select the USGS observation wells for the groundwater level validation is similar to the criteria used for the model

Conclusions

In this paper, the groundwater and surface water budgets were simulated on a monthly time step across the United States, using the transient MODFLOW-One-Water Hydrologic Model Version 2 (MF-OWHM2, Boyce et al. 2020, Boyce 2020). The model includes the impact of climate variability and human activities, specifically groundwater pumping, on the groundwater levels, including their seasonality. In addition, the lateral groundwater flow is simulated and interacts with surface water. This model

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors gratefully acknowledge the support of NSF Water Sustainability and Climate Grant 1360446, “WSC-Category 3 Collaborative: America's Water - The Changing Landscape of Risk, Competing Demands and Climate” and the support of NSF Division of Mathematical Sciences 1720114. They thank Nathaniel Chaney for sharing the POLARIS dataset. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

References (59)

  • S. Channan et al.

    Global Mosaics of the Standard MODIS Land Cover Type Data

    (2014)
  • V.T. Chow

    Open-Channel Hydraulics: New York

    (1959)
  • I.E.M. de Graaf et al.

    Environmental flow limits to global groundwater pumping

    Nature

    (2019)
  • I.E.M. de Graaf et al.

    A high-resolution global-scale groundwater model

    Hydrol. Earth Syst. Sci.

    (2015)
  • P. Döll et al.

    Seasonal water storage variations as impacted by water abstractions: comparing the output of a global hydrological model with GRACE and GPS observations

    Surv. Geophys.

    (2014)
  • P. Döll et al.

    Global‐scale assessment of groundwater depletion and related groundwater abstractions: combining hydrological modeling with information from well observations and GRACE satellites

    Water Resour. Res.

    (2014)
  • Y. Fan et al.

    Global patterns of groundwater table depth

    Science

    (2013)
  • Y. Fan et al.

    Incorporating water table dynamics in climate modeling: 1. Water table observations and equilibrium water table simulations

    J. Geophys. Res.

    (2007)
  • C.C. Faunt et al.

    Groundwater availability of the Central Valley Aquifer, California

    U.S. Geol. Surv. Prof. Pap.

    (2009)
  • C.R. Ferguson et al.

    Evapotranspiration estimates from satellite remote sensing: are present-day products of sufficient accuracy?

    AGU

    (2008)
  • L.E. Flint et al.

    Regional analysis of ground-water recharge

    (2007)
  • R.T. Hanson et al.
  • R.T. Hanson et al.

    One-water hydrologic flow model (MODFLOW-OWHM)

    US Geol. Surv.

    (2014)
  • R.T. Hanson et al.

    Rio Grande Transboundary Integrated Hydrologic Model and Water-Availability Analysis, New Mexico and Texas, United States, and northern Chihuahua, Mexico

    (2020)
  • A.W. Harbaugh et al.
    (2000)
  • A.W. Harbaugh

    MODFLOW-2005, the U.S. Geological Survey Modular Ground-Water Model: The Ground-Water Flow Process: U.S. Geological Survey Techniques and Methods 6-A16

    (2005)
  • R.C. Heath
    (1983)
  • S.J. Kollet

    Influence of soil heterogeneity on evapotranspiration under shallow water table conditions: transient, stochastic simulations

    Environ. Res. Lett.

    (2009)
  • N.Y. Krakauer et al.

    Groundwater flow across spatial scales: importance for climate modeling

    Environ. Res. Lett.

    (2014)
  • Cited by (11)

    • Augmenting the National agroecosystem model with physically based spatially distributed groundwater modeling

      2023, Environmental Modelling and Software
      Citation Excerpt :

      Modeling is performed at the river basin (Sun et al., 2016; Mohammed et al., 2018; Aliyari et al., 2019; Tran et al., 2022a; Tran et al., 2022b), regional (Schuol et al., 2008; Peterson et al., 2016; Hanson et al., 2020; Čerkasova et al., 2021), national (Condon and Maxwell, 2019; Alattar et al., 2020), and even continental (Abbaspour et al., 2015; Naz et al., 2022) scale. Many of these models have been used to inform policy regarding water management and climate change effects (e.g., Abbaspour et al., 2015; Čerkasova et al., 2021), whereas others are used to improve understanding of hydrologic systems and feedbacks (Maxwell et al., 2016) or are prepared for potential application in scenario analysis (Alattar et al., 2020). The degree of model complexity within these studies is wide-ranging, with several approaches using the semi-distributed model SWAT (Arnold et al., 1998), which lumps hydrologic fluxes at the level of hydrologic response units (HRUs; unique geographic areas of topographic slope, soil type, and land management) and simulates groundwater storage and flow using steady-state assumptions and linear reservoir equations [Sun et al. (2016) in the Yangtze River Basin (China), Yan and Roy (2016) in the Yellow River Basin (China), Schuol et al. (2008) in west Africa, Abbaspour et al. (2015) for the European continent, and Čerkasova et al. (2021) for the Nemunas River Basin (Europe].

    • Approximate solutions for flows in unconfined double porosity aquifers

      2022, Journal of Hydrology
      Citation Excerpt :

      Understanding groundwater flow in unconfined aquifers (UAs) is of first importance in several environmental and geotechnical applications. These aquifers are of ecological and economical significance as they play a major role in controlling surface-subsurface water exchange (Alattar et al. 2020), linking groundwater resources and surface water, as well as in the recharge of confined aquifers and the contaminant attenuation processes. UAs represent an important source of freshwater and water supply (Şen, 2015).

    • Developing a pan-European high-resolution groundwater recharge map – Combining satellite data and national survey data using machine learning

      2022, Science of the Total Environment
      Citation Excerpt :

      For the Continental USA, large modelling efforts using coupled and integrated groundwater surface water models, have been conducted (Alattar et al., 2020; Maxwell et al., 2015). These enable calibration and validation against both stream flow records and groundwater level time series, in an attempt to reduce the uncertainty on groundwater recharge simulation (Alattar et al., 2020). Similarly, the PCR-GLOBWB model has been run both globally and at a European scale, producing groundwater recharge estimates (Sutanudjaja et al., 2018).

    View all citing articles on Scopus
    View full text