Glassboro Study Area, NJ

Mahantango Creek Watershed, PA

Upper Mississippi Basin, MN

Glassboro Study Area, New Jersey

Working with Art Baehr of the USGS New Jersey District and with the NAWQA Program,we are evaluating the utility of several recharge estimation techniquess for use in humid environments. There appear to be 3 basic recharge regimes at this site: (1)continuous steady state, (2)seasonal steady state, and (3)transient. The methods we are using include the Darcian steady-state centrifuge method, water table fluctuation method, and numerical simulation of flow and energy transport.

Because unsaturated hydraulic properties, which are used to estimate steady-state recharge, are difficult and time consuming to measure accurately, models that estimate these properties indirectly are often used. Using data from six locations in southern New Jersey that appear to have steady-state flow conditions, five hydraulic property prediction and parameterization techniques were evaluated for recharge estimation. The unsaturated zone at this site, as in many coastal plain regions, is mainly sand to sandy loam in texture, which is considered a highly favorable case for soil hydraulic property estimation. Annual recharge has been estimated for several southern New Jersey watersheds by water budget methods and ranges from 33 to 49 cm/yr. Using these estimates as a gauge of reasonable values for steady flow, several indirect methods were compared to determine which are appropriate for recharge estimation in the coastal plain environment.

The methods used for estimation and representation of unsaturated hydraulic properties in this study were: (1) simple power-law and hand-drawn curve fits to measured unsaturated hydraulic conductivity, (2) measured water retention and measured unsaturated hydraulic conductivity fit using the van Genuchten-Mualem model with fixed and optimized parameter values, (3) unsaturated hydraulic conductivity estimated from measured water retention with fixed and optimized parameter values, (4) estimation of water retention and hydraulic conductivity from bulk density and minimal textural information using the Rosetta pedotransfer function model, and (5) estimation of water retention and hydraulic conductivity using high resolution particle size distributions with the Arya-Paris and van Genuchten-Mualem models.

Preliminary results show that while reasonable estimates can come from directly measured unsaturated hydraulic conductivity, the curve fitting method is critical because of the extreme nonlinearity of the relationship between hydraulic conductivity and water content. Even a relatively good visual fit can lead to unreasonable recharge estimation values. A hand-drawn interpolation technique, and in some cases a simple power-law fit, yielded better results for these data than the van Genuchten-Mualem curve fit. In all cases, the Rosetta model lead to unreasonably high recharge estimates, primarily due to the over-prediction of saturated hydraulic conductivity. Water retention was predicted relatively well by the Rosetta Model. A combination of the Arya-Paris and van Genuchten-Mualem models also lead to over-prediction of recharge rates, and did not do as well as Rosetta in predicting water retention for this sandy material.

Mahantango Creek Watershed, Pennsylvania

In collaboration with the Dennis Risser and others in the Pennsylvania District, Chris Heppner and John Nimmo began research in 2003 on an investigation titled Comparison of Methods for Estimating Ground-Water Recharge and Conceptual Modeling of Ground-Water Accretion Behavior at the USDA-ARS East Mahantango Creek Watershed in East-Central Pennsylvania.

In this research we make use of available hydrologic data to compare various methods for estimating ground-water recharge at the U.S. Department of Agriculture–Agricultural Research Service (ARS) East Mahantango Creek watershed in east-central Pennsylvania. Because ARS has collected detailed ground-water, surface-water, and meteorological data for the past 30 years, the site affords a unique opportunity to compare different approaches for estimating long-term ground-water recharge rates and to evaluate conceptual processes controlling movement of ground water in the unsaturated-zone. Hydrologists at ARS have shared their historical data and have invited USGS to cooperatively investigate ground-water resources at the experimental watershed. Recently, the USGS National Water Quality Assessment, Chesapeake Bay Ecosystem, and Bedrock Regional Systematics (BRASS) programs have conducted cooperative research with USDA-ARS on water quality, ground-water age dating, and geologic controls of ground-water movement in the East Mahantango Creek watershed.

The objectives of this study are to: (1) compare four methods for estimating the timing and magnitude of ground-water recharge for a single watershed, (2) evaluate hydrograph-separation methods at three watershed scales, and (3) use conceptual modeling to test hypotheses concerning the processes controlling ground-water accretion behavior. In our collaboration, he Pennsylvania District is pursuing the first two of these objectives and the Unsaturated-Zone Flow Project is pursuing the third.

Recharge can be estimated on a monthly and annual basis from the volume of percolate collected in zero-tension pan lysimeters at a location known as the Masser Recharge Site. This upland site was selected by ARS to minimize lateral movement of subsurface flow so that processes characterizing vertical ground-water recharge could be studied. Twenty-eight zero-tension pan lysimeters (16 monitored continuously) have been installed at the Masser Recharge Site to monitor the amount and timing of percolate since 1993. Five observation wells, 100-150 feet deep, are located near the lysimeters. This is the method that most closely represents an actual direct measurement of recharge; however, previous results so far contain significant discrepancies and surprises, including:

• The ground-water accretion method and unsaturated-zone lysimeter method do not consistently agree. For example, in summer, the water-table response to storms sometimes indicates significant recharge while lysimeter percolate is negligible.

• Estimates of recharge by the lysimeter method combined with water-table hydrographs gives an unusually low estimate of specific yield (<1%). This has yet to be explained in terms of dynamics of flow in the unsaturated zone.

The conceptual modeling is proposed to: (1) test hypotheses concerning processes responsible for the observed responses in the gravity lysimeters and nearby wells, (2) enhance understanding of ground-water movement in the larger-scale fractures versus the soil and rock matrix, and (3) establish the basis for further field or numerical investigations of the subsurface critical to recharge. Some of the research activities in this are to:

q       Formulate a basic conceptual model of recharging fluxes based on observable features of the subsurface.

q       Assess in what ways the model can account for the qualitative nature of observed hydrologic behavior.

q       Modify the model according to hypotheses related to basic fracture types; fracture orientations, porosity, roughness, connectedness; film flow; matrix flow; air-trapping; episodic nature of recharge.

q       Evaluate these and ascertain quantitative ranges of fracture scale, aperture, density, and connectedness that might lead to the observed hydrologic behavior

Upper Mississippi Basin, Minnesota

In collaboration with Geoff Delin and others in the Minnesota District, and Rick Healy of BRR, Central Region, Chris Heppner and John Nimmo began research in 2003 on an investigation titled Temporal and Spatial Variability in Ground-Water Recharge in Upper Mississippi Basin, Minnesota.

The area of study is the Upper Mississippi River Basin that includes about 47,000 mi² in Minnesota and Wisconsin and includes the Mississippi, the Minnesota and the St. Croix River Basins. Average annual precipitation ranges from 22 inches in the west to 32 inches in the east. Land use and cover in the area include forests, lakes, agriculture, and urban areas. Within this are are the Bemidji and IRI sites, extensively investigated in other USGS studies

The primary objective of this study is to quantify recharge to unconfined sand and gravel aquifers in the Upper Mississippi River Basin using several methods applied at a variety of scales. Intensive (local scale) data-collection networks and existing data sets from the Crude Oil research site, near Bemidji, and from the Interdisciplinary Research Initiative (IRI) research site, near Akeley, will provide detailed foundational information that will compliment USGS and State long-term, regional data on ground-water levels and streamflow.

The approach the research team is following includes the investigations listed below. The Unsaturated-Zone Flow Project is contributing mainly to the specific yield estimations and hydraulic property measurements, items 1 and 3. There is much conceptual overlap between this study and our recharge investigation at Mahantango Creek Watershed, Pennsylvania. We gain considerable insight from the fact that these two sites strongly complement each other. Unsaturated flow at the Minnesota site occurs within a granular, mainly sandy, medium, whereas at the Pennsylvania site it occurs predominantly within fractured rock.

1. Ground-water hydrograph analysis is used to estimate recharge based on existing water-level information from more than 150 wells throughout the Basin. Application of the water-table fluctuation method provides information on both temporal and spatial variability of recharge across the basin and at local scales.

A critical element in estimating ground-water recharge from fluctuations of water levels in wells is the estimation of specific yield (a property of the unsaturated zone and the saturated zone near the water table). Measurement of unsaturated hydraulic properties will be made from core samples collected for this study using the steady-state centrifuge (SSC) method, and related methods, at representative depths in the unsaturated zone. These estimates will be used to determine values of specific yield. Specific yield will also be computed using other measurements and techniques including those developed in the investigation at the Mahantango Creek site. This method will improve upon the uncertainty in estimating recharge using approximate handbook values of specific yield.

2. Continuous soil-moisture data from the unsaturated zone from the Bemidji site (1997 to present) and from the IRI site (1999 to present) will be used to estimate temporal variability in recharge based on the zero-flux plane method (ZFP).

3. Unsaturated hydraulic property measurements, described above in connection with water-table fluctuation estimates, will be used with basic variably saturated flow modeling (VS2DT) to refine our knowledge of the position and behavior of the ZFP. This will reduce a significant source of uncertainty in the ZFP method.

4. Ground-water samples from approximately 20 locations across the Basin will be obtained and analyzed for sulfur hexaflouride (SF6) to estimate ground-water ages. Ground-water age dates will be used to estimate recharge. This information will be combined with ground-water age-dating information from the Bemidji site and from data from the Upper Mississippi River NAWQA study (tritium/helium and CFC techniques). These site-specific estimates will be used with precipitation information (up-scale estimate) to obtain estimated values of ground-water recharge for the region.

5. Sub-basin scale estimates of recharge will be obtained by applying hydrograph separation to records of streamflow and by compiling steady-state estimates of recharge from ground-water-flow models developed for surficial aquifers in the Basin. These numbers will represent the largest of different spatial scales of recharge estimates for the Basin.

6. A detailed basin characteristic GIS database exists for the upper Mississippi River basin. Information from this database and from the STATSGO data base (soils) will be used in a multiple-regression analysis to identify parameters (hydrologic, geologic, meteorological, soils. and vegetation) that relate to recharge rate variability from the different techniques described above.