Ground-Water Recharge in Minnesota “Ground-Water Recharge” Broadly Describes the Addition of Water to the Ground-Water System

Home , Aquifer

Photograph by Geoffrey Delin, U.S. Geological Survey Ground-Water Recharge in Minnesota “Ground-water recharge” broadly describes the addition of water to the ground-water system. Most water recharging the ground- water system moves relatively rapidly to surface-water bodies and sustains streamflow, lake levels, and wetlands. Over the long term, recharge is generally balanced by discharge to surface waters, to plants, and to deeper parts of the ground-water system. However, this balance can be altered locally as a result of pumping, impervious surfaces, land use, or climate changes that could result in increased or decreased recharge. • Recharge rates to unconfined aquifers in Minnesota typically are about 20–25 percent of precipitation. • Ground-water recharge is least (0–2 inches per year) in the western and northwestern parts of the State and increases to greater than 6 inches per year in the central and eastern parts of the State. • Water-level measurement frequency is important in estimating recharge. Measurements made less frequently than about once per week resulted in as much as a 48 percent underestimation of recharge compared with estimates based on an hourly measurement frequency. • High-quality, long-term, continuous hydrologic and climatic data are important in estimating recharge rates.

of the hydrologic cycle, is the process Where Does Recharge Occur? The first section of this fact sheet by which water moves to the water table introduces the process of ground-water Recharge to the water table occurs and then away from that area through recharge, including definitions of in most areas of the landscape but com- saturated materials. Figure 1 shows how related terminology and clarification of monly at varying rates. Given the same water recharges an unconfined aquifer common misconceptions. The second precipitation, recharge rates in areas and moves toward a river where it dis- section describes how ground-water where sediments are sandy are greater charges. recharge rates vary in Minnesota. The than in areas where sediments are finer The process of ground-water third and final section is a more techni- grained. In addition, surface-water runoff recharge is like refilling a leaky tank. cal overview of several methods used to to low lying depressions in the land-sur- Over the long term, recharge is generally estimate recharge rates in Minnesota. face generally promotes recharge (fig. 1). balanced by discharge to surface waters Surface-water bodies can be a source of or deeper parts of the ground-water recharge in areas where the water level system or by uptake by plants. However, What is Ground-Water Recharge? (hydraulic head) in the underlying aquifer this balance can be altered locally by “Ground-water recharge” broadly is below the lake or river water level. pumping, impervious surfaces, land use, describes the replenishment of water to Water in confined glacial and bed- or climate changes resulting in increased a ground-water flow system (Winter and rock aquifers is typically replenished by or decreased recharge. others, 1998). Recharge, an integral part leakage of water through overlying

Precipitation Transpiration Pumped Evaporation well Surface runoff Unsaturated Wetland sediment or rock Lowland Recharge to unconfinedWater aquifer table River Unconfined aquifer Direction of ground-water flow Leakage to and from confined aquiferConfining layer

Confined aquifer

Figure 1. Ground-water recharge is an important part of the hydrologic cycle in Minnesota.

U.S. Department of the Interior Printed on recycled paper Fact Sheet 2007–3002 U.S. Geological Survey January 2007 confining layers. Leakage is a process Figure 3a illustrates the temporary 56 similar to but not the same as recharge effect recharge water has on the water RRecharge = specific yield x e water-level rise (fig. 1). Rates of leakage to confined table. The water level in a well installed c 57 h = 0.23 x 2.1 feet aquifers are generally less than rates of in an unconfined aquifer generally rises a = 0.483 feet per year recharge to unconfined aquifers. Upward rapidly in response to recharge from r leakage also occurs from a confined aqui- precipitation. In the absence of recharge, 58 fer in areas where hydraulic head in the the water table declines in response to confined aquifer is greater than that in the movement of water away from the area. 59 overlying formation; water moves from 2.1 feet areas of greater to lower hydraulic head. When Does Recharge Occur? 60 Projected water-level decline Confined aquifers also are replenished in The amount of ground-water areas where the confining layer is perme- Depth to water, feet below land surface recharge varies seasonally in response to 61 able or absent. Jan Mar May July Sept Nov precipitation; spring and fall are the times Feb Apr June Aug Oct Dec of greatest ground-water replenishment in Minnesota (fig. 3b). Evidence of smaller Figure 3a. Example recharge estimation recharge events also can be seen in the based on the water-table fluctuation figure 3b hydrograph. During winter, method using hypothetical data with only water levels generally decline because one recharge event in a year (modified from recharge rates are negligible due to snow Delin and others, 2007). cover and frozen soil. During the growing season, plant transpiration rates generally exceed precipitation rates. A large 15.0 0 proportion of the water that infiltrates 0.4 Photograph by Lee Grim, National Park Service at land surface, therefore, is returned to 15.4 0.8 the atmosphere before it can reach the water table. Occasionally, as shown in 1.2

15.8 Precipitation, inches Figure 2. Much ground water recharge in figure 3b, summer rainfall in excess of 1.6 Minnesota ends up in streams, lakes, and ET rates results in net recharge. wetlands. 16.2 Ground-water recharge is not… 16.6 Why Do Recharge Rates Vary? …equivalent to “infiltration” of ET effects

Numerous factors, including physi- water at the land surface. Most water Depth to water, feet below measuring point 17.0 cal characteristics of the soil, vegetation that infiltrates at the land surface Jan Mar May July Sept Nov cover, land use, topography, water con- is returned to the atmosphere by Feb Apr June Aug Oct Dec tent of surface materials, and the presence plant transpiration and evaporation Figure 3b. Hydrograph for calendar year and depth of the confining layers, influ- from soil and water surfaces (fig. 1). 2000 from a well near Windom, Minn., ence the spatial variability of recharge Recharge is typically only a small completed in an unconfined aquifer, rates. Weather patterns, including the fraction of infiltration. showing the effects of evapotranspiration timing and intensity of spring snowmelt, …equated to the process of “perco- (ET) and of multiple recharge events spring rains, evapotranspiration (ET) lation;” instead, percolation refers (modified from Cowdery, 2005). during the summer growing season, and to the movement of water through fall rains, play a major role in control- unsaturated sediments. The percolat- ling spatial and temporal variability in ing water can be viewed as potential Why is Knowledge of Ground-Water recharge rates. recharge, however. Recharge Rates Important? …to be confused with the term Knowledge of ground-water What Happens to the Water After it is “aquifer yield.” This term refers to recharge rates is important to studies of Recharged? the amount of water that an aquifer can yield to pumping. water availability, sustainability, well- In semihumid regions like Minne- …the same as “sustainable yield.” head protection, contaminant transport, sota, most water recharging the ground- Recharge is less than sustainable ground-water and surface-water interac- water flow system moves relatively yield. If all recharge water was tions, effects of urbanization, and aquifer rapidly to surface-water bodies (fig. 1) utilized, stream, lake, and wetland vulnerability to contamination (Scanlon and sustains streamflow, lake levels, and levels could be adversely affected. and others, 2002). Estimates of recharge wetlands (fig. 2). Smaller portions of the Furthermore, it cannot be assumed rates are necessary to quantify the amount recharged water move to deeper confined that pumping at less than the of water moving through near-surface aquifers, are extracted by plants, or are recharge rate will not cause water- ground-water systems and are important withdrawn from aquifers by wells for level declines and ground-water to understanding the water balance of an irrigation, public supply, or industrial and storage depletions. area and the ways that human activities, other uses. such as landscape practices, affect that balance. For example, by estimat- rate long-term measurements of stream- porated in the RRR method. The results ing the seasonal and spatial distribution flow, ground-water levels, and climate are illustrate the spatial variability of annual of recharge rates, one can estimate the needed to estimate recharge rates. The recharge rates to surficial materials across total volume of water entering a system. final section of this fact sheet provides a the State. An abbreviated description of Recharge is a sensitive component of general description of several methods the RRR method is in the Regional-Scale ground-water flow models and is the one that can be used to estimate recharge in Methods section of this fact sheet. The that is least understood (Delin and others, Minnesota. recharge estimates shown in figure 4 are 2007). representative of average recharge rates How do Recharge Rates Vary for 1971–2000 because they are based on How are Recharge Rates Estimated? Across Minnesota? data from that time period. There is no “recharge gage” analo- The statewide basin-scale estimates gous to a rain gage, but many methods of ground-water recharge rates in Min- Recharge rates to unconfined have been developed for estimating nesota (fig. 4) are based on the regional aquifers in Minnesota typically are recharge rates. Selection of the appropri- regression recharge (RRR) method of about 20–25 percent of precipita- ate method for a given study is important Lorenz and Delin (2007). Stream base- tion. A crude, preliminary estimate and can be challenging. Recharge rates flow estimates in selected basins and of recharge rate is sometimes are best estimated by use of multiple statewide climate and soil data are incor- methods and the results compared. Accu- made on the basis of this assump- tion. Recharge rates where glacial clays or till are present, however, generally are less than 10 percent 96º of precipitation. By comparison, leakage rates to confined aquifers 94º CANADA generally are less than 1 percent of precipitation. 92º 90º 48º Spatial trends in recharge in Min- nesota (fig. 4) reflect general trends in precipitation. In the western and north-

Lake Superior western parts of the State where precipitation is least (20–25 in/yr (inches per year)), recharge also is least (0–2 in/yr). In contrast, recharge increases in the east-

EXPLANATION ern part of the State to greater than 6 in/yr NORTH DAKOTA DAKOTA NORTH Mississippi Water as precipitation increases to greater than Recharge in about 30 in/yr. inches per year In addition to being affected by 46º Greater than 12 precipitation, recharge rates are locally 10.01 to 12 affected by soil properties. Sandy areas, 8.01 to 10 such as the Anoka sand plain and others 6.01 to 8 in the central and east-central parts of 4.01 to 6 the State, have coarse-textured soils that 2.01 to 4 correspond well with RRR rates in the WISCONSIN 0 to 2 6–10 in/yr range (fig. 4). Clay-rich soils, Anoka Unclassifiable Sand such as those in the western parts of the Plain 0 20 40 60 80 KILOMETERS State, correspond well with RRR rates in Minnesota 0 20 40 60 80 M ILES the 0–4 in/yr range. The “Unclassifiable”

River SOUTH DAKOTA areas in figure 4 represent primarily peat- 44º River lands, where the organic content is too great to estimate RRR rates accurately. Fine-scale variability in weather MINNESOTA patterns, soil properties, land use, and Base from U.S. Geological Survey digital data IOWA topography were not included in the RRR 1:2,000,000, 1972, Universal Transverse Mercator Projection model of the State. Low permeability units at or near land surface, ET in areas Figure 4. Average annual recharge rate to surficial materials in Minnesota (1971–2000) of shallow ground water, and impervious estimated on the basis on the regional regression recharge method (modified from Lorenz surfaces could effectively reduce the and Delin, 2007). estimated recharge rates locally. Focused use but is limited by the accuracy with 31.6 recharge to depressional areas by runoff which specific yield can be determined; (a) from upland areas also may increase it is also affected by water-level measure- 31.8 recharge locally. In addition, 30-year ment frequency and accuracy. averages of precipitation and other data were used to construct the RRR model 32.0 (Lorenz and Delin, 2007). Thus, actual recharge rates may vary during a given Delin and others (2007) 32.2 showed that water-level measure- year as well as from year to year, depend- Recharge / precipitation = 7.8 percent ing on weather patterns. Nonetheless, ment frequency is important in 32.4 the RRR recharge estimates (fig. 4) can applying the water-table fluctua- 31.6 be a useful source of input for regional tion method. Measurements made (b) ground-water flow models and should be less frequently than about once 31.8 helpful to resource managers in develop- per week resulted in as much as ing water-management plans at State and a 48 percent underestimation of 32.0 regional scales. recharge based on an hourly measurement frequency (fig. 5). Methods Used to Estimate 32.2 Recharge Rates in Minnesota Recharge / precipitation = 5.3 percent 32.4 Ground-water recharge rates cannot The age dating of ground water 31.6 easily be measured, thus recharge rates method uses estimated ground-water (c) must be estimated using indirect methods. Depth to water, feet below measuring point ages (the time elapsed since the water 31.8 Recharge rates in Minnesota have been entered the aquifer as recharge) and well- estimated by means of many methods. A depth information to obtain an estimate 32.0 detailed description of all recharge-esti- of vertical ground-water velocity. The mation methods and their limitations is velocity is then multiplied by aquifer Data point beyond the scope of this fact sheet. The porosity to obtain an estimate of recharge 32.2 reader is referred to Scanlon and others rate. This method is limited primarily Recharge / precipitation = 3.6 percent (2002) for a thorough review of recharge by the accuracy with which the ground- 32.4 estimation methods and their limitations. Jan Mar May July Sept Nov water age is determined and by the accu- Feb Apr June Aug Oct Dec The methods commonly used in Min- racy with which aquifer porosity can be nesota are briefly described below and determined. Environmental tracers such Figure 5. Hydrographs and recharge grouped on the basis of spatial scale: as chlorofluorocarbons and sulfur-hexa- estimates for a well near Bemidji, Minn., for local, representing areas up to thousands fluoride (as described by Busenberg and calendar year 2002 based on the water- of square feet; basin, representing tens to Plummer, 1992 and 2000, among others) table fluctuation method and a water-level hundreds of square miles, and regional, are used to estimate ground-water age to measurement frequency of once every representing hundreds to thousands of within about 1 to 2 years. (a) day, (b), month, and (c) two months. square miles. The chloride tracer method is used to estimate recharge rates as the product Local-Scale Methods of precipitation and the ratio of chloride The water-table fluctuation (WTF) concentration in wet and dry deposition The unsaturated-zone water method may be the most widely used on land surface to chloride concentra- balance (UZWB) method (Delin and method for estimating recharge rates in tion in the unsaturated or saturated zones others, 2000) is based on the premise that humid regions (Healy and Cook, 2002). (Scanlon and others, 2002). The unsatu- soil water moves upward in response to The WTF method relates changes in mea- rated zone is that part of the subsurface ET above the zero-flux plane boundary in sured water-level elevation to changes in from land surface down to the water the unsaturated zone and that water below the amount of water stored in the aquifer. table. Chloride is conservative (not prone that boundary eventually percolates Recharge is assumed to be equal to the to change in concentration as a result of downward to the water table. Recharge product of water-table rise and specific chemical reactions), and its mass inflow is estimated by measuring the change in yield (fig. 3a). Specific yield is the must be balanced by mass outflow or by water stored below the boundary over amount of water a unit volume of satu- a change in storage. In areas where total time. This method, sometimes called rated permeable material will yield when infiltration is less than total precipitation, the zero-flux plane method, is limited drained by gravity. Several approaches because of surface runoff, the chloride primarily by the requirement of intensive have been used in Minnesota for estimat- tracer method will overestimate actual collection of soil-moisture data from the ing the peak water-table rise attributed recharge. unsaturated zone. to the recharge event (Delin and others, 2007). The method is relatively easy to The unsaturated-zone drainage rates (Delin and others, 2007). RRR method can be used to estimate recharge High-quality, long-term, con- rate estimates could be a useful source rates by directly measuring the vertical tinuous hydrologic and climatic data of input for regional ground-water flow flow of water from gravity lysimeters are important in estimating recharge models. installed in the unsaturated zone. This rates. High-quality streamflow The water-balance equation is drainage from the lysimeters represents (fig. 6), ground-water level (fig. 7), a common approach for estimating water that has not yet reached the water and climate data are required for some recharge rates at regional scales. A gen- table and has been termed “potential recharge estimation methods. Miss- eral form of a water budget is: ing data add an additional degree of recharge” by Scanlon and others (2002). P + Q = ET + Q +∆S Lysimeters are not routinely used to esti- uncertainty to the results. on off mate recharge because they are expensive where P is precipitation (including irriga- and difficult to construct, and they also tion), Qon and Qoff are water flow onto and require substantial maintenance. In the streamflow-hydrograph off the site respectively, ET is evapo- The Darcian flux method is used separation methods, base flow is used transpiration, and ∆S is change in water to estimate recharge rates on the basis as a proxy for recharge. Base flow is storage (Scanlon and others, 2002). Many of Darcy’s Law, by using estimated defined as ground water that discharges other variations of this equation can be hydraulic head gradient and estimated to a stream and sustains the flow during formulated by incorporating subcompo- hydraulic conductivity. Several variants dry periods. In using these methods, one nents of each variable. All ground-water of this method can be used with data assumes that ground-water underflow, and surface-water flow models fit under from saturated and unsaturated materi- ET, and other losses of ground water this method because they are based on a als. In the unit gradient variation of this are minimal. PART (Rutledge, 1993), water-balance equation. Although listed method, one assumes the matric potential HYSEP (Sloto and Crouse, 1996), and here as a regional-scale method, a water in the unsaturated zone is constant with BFI (Wahl and Wahl, 1988) are computer budget estimate of recharge can also be depth and gravity is the only driving force programs that separate base flow from a made on the basis of local- or basin-scale (Nimmo and others, 2003). streamflow hydrograph on the basis of data. Seepage meters can be used to different criteria. estimate seepage to or from surface- water bodies (Scanlon and others, 2002). Regional-Scale Methods A seepage meter typically consists of a The regional regression recharge cylinder that is pushed into the bottom of (RRR) method of Lorenz and Delin a stream or lake; an attached reservoir, (2007) yields an estimate of spatial vari- commonly a plastic bag, collects water. ability of annual recharge rates within a A ground-water recharge rate can be esti- region. Many other methods are docu- mated from the rate at which water in the mented in the literature, and the reader is cylinder infiltrates, as determined from referred to Lorenz and Delin (2007) for changes in the reservoir volume. This a review of these references. The RRR method is inexpensive and easy to apply; method is based on a regression of RORA however, numerous measurements may recharge rate estimates with climate and be necessary because of the point nature soil data for the region. The accuracy of of the estimate. the RRR estimates are representative of the soils data, which are collected over Photograph by Charlie Smith, U.S. Geological Survey Basin-Scale Methods areas ranging from about 2 to 150 square RORA (Rutledge, 1998) is an auto- miles. Recharge rates estimated on the mated method for estimating the average basis of the RRR method (fig. 4) com- recharge rate in a basin from analysis of pared favorably with results from local- streamflow records, and is based on the and basin-scale methods in Minnesota. recession-curve-displacement method of The RRR rates on average were about Rorabaugh (1960). RORA accounts for 41 percent less than UZWB rates, ranged the effects of ET, underflow (the flow from 35 percent greater to 12 percent less Figure 6. High-quality, continuous of ground water beneath and bypassing than WTF rates, were about 12 percent streamflow data are essential for accurate a stream), and other losses or gains of less than age-dating-method rates, and recharge estimation using basin-scale ground water after a precipitation event. were about 5 percent greater than RORA methods. The USGS maintains a nationwide network of stream-gaging stations, with stream water level and flow data served in real-time to the Internet. Delin, G.N., Healy, R.W., Landon, M.K., Rutledge, A.T., 1998, Computer pro- and Böhlke, J.K., 2000, Effects of grams for describing the recession of topography and soil properties on ground-water discharge and for esti- recharge at two sites in an agricultural mating mean ground-water recharge field: Journal of the American Water and discharge from streamflow Resources Association, v. 36, no. 6, records—Update: U.S. Geological p. 1401–1416. Survey Water-Resources Investigations Report 98–4148, 43 p. Delin, G.N., Healy, R.W., Lorenz, D.L., and Nimmo, J.R., 2007, Comparison Scanlon, B.R., Healy, R.W., and Cook, of local- to regional-scale estimates of P.G., 2002, Choosing appropriate tech- ground-water recharge in Minnesota, niques for quantifying groundwater SOIL USA: Journal of Hydrology, v. 334, recharge: Hydrogeology Journal, v. 10, SAND AND GRAVEL no. 1-2, p. 231–249. p. 18–39.

U.S. Geological Survey Photograph by Jeffrey Stoner, CLAY Healy, R.W., and Cook, P.G., 2002, Sloto, R.A., and Crouse, M.Y., 1996,

WATER TABLE Using groundwater levels to estimate HYSEP—A computer program for recharge: Hydrogeology Journal, v. 10, streamflow hydrograph separation AQUIFER p. 91–109. and analysis: U.S. Geological Survey Water-Resources Investigations Report Lorenz, D.L., and Delin, G.N., 2007, A 96–4040, 46 p. Figure 7. High quality, continuous ground- regression model to estimate regional water level data are essential for accurate ground-water recharge in Minnesota: Wahl, K.L., and Wahl, T.L., 1988, Deter- recharge estimation with the water-table Ground Water, v. 45, no. 2, 10.1111/ mining the flow of Comal Springs at fluctuation method. In this picture a USGS j.1745-6584.2006.00273.x. Braunfels, Texas, in Texas Water ’95, hydrologist is programming dataloggers to San Antonio, Texas, August 16-17, measure ground-water levels in Otter Tail Nimmo, J.R., Stonestrom, D., and Healy, 1995, Proceedings: American Society County. The schematic diagram beneath R.W., 2003, Aquifer recharge in Ency- of Civil Engineers, p. 77–86. the picture illustrates an observation well clopedia of Water Science: Stewart, completed in a glacial aquifer. B.A., and Howell, T.A., eds., New Winter, T.C., Harvey, J.W., Franke, O.L., York, Marcel Dekker, p. 1– 4. and Alley, W.M., 1998, Ground water and surface water—a single resource: References Rorabaugh, M., 1960, Use of water levels U.S. Geological Survey Circular 1139, in estimating aquifer constants in a 79 p. Busenberg Eurybiades, and Plummer finite aquifer: International Association L.N., 1992, Use of chlorofluorocar- of Scientific Hydrology Commission of Subterranean Waters, publication 1 2 bons (CCl3F and CCl2F2) as hydrologic By G.N. Delin and J.D. Falteisek tracers and age-dating tools—The 52, p. 314–323. alluvium and terrace system of Central Oklahoma: Water Resources Research, Rutledge, A.T., 1993, Computer pro- v. 28, no. 9, p. 2257–2283. grams for describing the recession of ground-water discharge and Busenberg, Eurybiades, and Plummer, for estimating mean ground-water L.N., 2000, Dating young groundwater recharge and discharge from stream- with sulfur hexafluoride—Natural and flow records: U.S. Geological Survey anthropogenic sources of sulfur hexa- Water-Resources Investigations Report 93–4121, 45 p. fluoride: Water Resources Research, Information on the 1U.S. Geologi- v. 36, no. 10, p. 3011–3030. cal Survey Minnesota Water Science Cowdery, T.K., 2005, Hydrogeology and Center can be obtained at following ground-water/surface-water interac- web site http://mn.water.usgs.gov/ 2 tions in the Des Moines River Valley, and information on the Minnesota southwestern Minnesota, 1997–2001: Department of Natural Resources U.S. Geological Survey Scientific Division of Waters can be obtained at Investigations Report 2005–5219, 51 p. http://www.dnr.state.mn.us/waters/