Environ Geol (2007) 51: 731–735 DOI 10.1007/s00254-006-0392-0

Richard R. Parizek

Received: 9 October 2005 Accepted: 23 May 2006 Published online: 27 October 2006  Springer-Verlag 2006

R. R. Parizek Department of Geosciences, The Pennsylvania State University, University Park, PA, USA E-mail: [email protected]

ORIGINAL ARTICLE

Opportunities to enhance management of karstic aquifers

Abstract Methods exist to obtain ‘‘new sources of water.’’ Examples include: (1) capturing and enhancing stormwater recharge and retention within diffuse-flow portions of karst and other aquifers; (2) recycling and reuse of waste water; (3) reducing evapotranspiration and rejected recharge; and (4) ameliorating atmospheric acid deposition through use of alkaline groundwater. These little used management methods have immense potential to sustain future water demands. Full utilization of ‘‘new’’ and traditional water resources requires an understanding of

Opportunities to enhance management of karstic aquifers This paper provides examples of two of the four such methods to develop ‘‘new water sources,’’ (i.e., Capture of Stormwater, and Recycling Municipal Waste Water) Capture of stormwater Stormwater runoff can be highly variable depending upon basin characteristics and changes in land use. In general, there is an inverse relationship between stormwater and baseflow. As the proportion of drainage area underlain by carbonate rock increases, stormwater decreases. ‘‘New sources’’ of water result from reductions in diffuse infiltration and evapotranspiration as plant cover is reduced. Routing stormwater to surface streams via detention basins and drains is a widespread practice to reduce flood peaks and the potential for groundwater

the hydrogeologic framework of karstic aquifers. Reliable conceptual, numerical flow and transport models are needed to help evaluate, select, and design viable water management options. Three such simulation examples are provided together with a discussion of Penn State’s Wastewater reuse project where recharge approaches 3.785 · 109l/year Keywords Karst Æ Artificial recharge Æ Stormwater Æ Wastewater reuse Æ Numerical modeling

pollution, as well as to protect down stream properties and to avoid sinkhole formation. In the U.S., there is a shift in emphasis to sustain or enhance groundwater recharge and reduce stormwater pollutant loads. If the intent is to optimize groundwater recharge, then aquifer storage potential and water residence time should be considered. Two such modeling projects are described below. In the first example, carbonate rocks within the Ridge and Valley Province of the Appalachian Mountains offer numerous opportunities to divert stormwater and enhance groundwater recharge. Figure 1a and b illustrates common occurrences of carbonate aquifers flanked by mountain ridges underlain by clastic rocks. The hydrologic budget of the Spring Creek basin in Centre County, PA reveals sources and magnitude of recharge within this (and other basins) in this province. Its surface water and groundwater drainage areas are 375 and 453 km2 respectively. Nearly 60% of its recharge is derived from mountain ridges. Ridges account for 20% of

732

Fig 1 a Relationship of stratigraphy, bedrock structure, topography, conduits and mountain runoff (From Schmotzer et al. 1974); b Use of gravity injected wells to capture excess stormwater (Modified from Parizek and Mundi 1990)

the total basin area, while carbonate rocks represent nearly 80%. Porosity and permeability are not uniformly distributed (Siddiqui and Parizek 1971; Chin and Parizek 1997). Rauch and White (1970) showed that of the 2,220 m of carbonate rocks; only about 280 m contain explorable caves. These cavernous host strata lie just below mountain slopes. Although some diffuse recharge enters these conduits between water gaps (Konikow 1969), most recharge occurs via swallow holes and channel seepage opposite water gaps. Conduits rapidly divert this water parallel to stratigraphic strike away from central portions of carbonate uplands to springs (Fig. 1a). This is evident from tracer experiments (Schmotzer et al. 1974; Underwood 1994), seasonal chemical and flow responses of springs (Shuster and White 1971) and groundwater and cave exploration. Where strata are nearly vertical, some mountain tributaries channel large quantities of runoff to master streams without significant recharge. Portions of this water could be captured for recharge (Parizek

and Mundi 1990). Water can be piped across conduits and areas of rejected recharge for injection into diffuseflow portions of aquifers (Fig. 1b) Methods exist to reduce turbidity and to select optimum high capacity injection sites (Parizek 1987); however, some water may have to be released to sustain ecosystems and aquatic habitats. Numerical groundwater models help in the evaluation of artificial recharge schemes including appropriate well counts, spacing and injection rates, as well as expected changes in the groundwater flow field as it is stressed by recharge and pumpage. Yeh and Chang’s (1993) FEWA 2-D finite element code was used in the first analyses. Some model assumptions included: groundwater flow through an equivalent porous media; no-flow boundary aquifer base; homogeneous hydraulic properties within each grid cell; horizontal flow; grid alignment with principal hydraulic conductivity; maximum principal hydraulic conductivity set at 5x’s minimum value; and steady state (Chin and Parizek 1997). A hydrogeologic conceptual model was developed that incorporated data obtained through years of study. Hydraulic conductivity was vertically averaged over the region. Spatial distribution of final hydraulic conductivity values were derived from calibrated results. The 80 km2 model domain, 578 nodes and 816 elements, was oriented in the direction of principal hydraulic conductivity. All, except the northwestern boundary, were simulated as constant head boundaries. The northwestern boundary along an influent stream was simulated as a constant flux boundary. Groundwater is known to flow from a mountain ridge toward and below this stream where it enters a diffuse flow-dominated aquifer in the vicinity of a thrust fault where it is constrained by a regional groundwater trough (Fig. 2a) and stratigraphic sequences. Bottom elevations of the aquifer were estimated based on a number of observations. Included among these are: depth; size; frequency of water-yielding openings observed within test and water supply wells; logs of drill cores and cuttings; borehole geophysical logs; inspection of fractures and solution openings within mines and quarries (up to 290 m deep); chemical and flow responses of springs; and water level responses. The average recharge rate (254 mm/year) was constrained by a hydrologic budget analysis of the Spring Creek basin. Local recharge rates varied according to land uses (e.g., irrigated turf, crop land, areas of concentrated runoff and infiltration, and Penn State University¢s waste water renovation site). Flow nets and cones of depression also were analyzed to estimate recharge (Smith 1986). Porosity values were obtained from cores available from carbonate strata. Average values ranged from 1.5 to 5% (Brown 1977). Specific yield was

733

determined from a 38-day groundwater recession analysis of three tributary mountain basins (Konikow 1969) and the entire Spring Creek basin (Giddings 1974). The average value was 1.5% from mountain tributaries and 3.0 to 4.0% for carbonates. Coefficients of storage obtained from multiple well pumping tests varied from 0.001 to 0.000001 (Siddiqui 1969). Model calibration was based on best-fit analysis of simulated and measured hydraulic heads and similarities in observed and predicted water table configuration. Calculation of root-mean-square error (RMSE), and other methods were generally used. Steady state was assumed. The spatial distribution of recharge rate and boundary conditions fixed during the calibration process were used in the final model. About 180 model replications and parameter adjustments were required to match the final simulated to the observed potentiometric surface (Fig. 2a, b), (Chin and Parizek 1997). The model is most sensitive to a decrease in hydraulic conductivity and increase in anisotropy, but not to the assumed bottom elevation of the aquifer. Figure 3 shows the steady state cones of impression that might be achieved by recharging six existing production wells at their present production rates: UN-14, 27.72 l/s; UN-16, 17.35; UN-17 28.4; UN-24, 28.4; UN-26, 47.3; and UN28, 31.6 l/s. It is unlikely that this quantity of water could be diverted from the nearest mountain tributaries without adversely affecting aquatic ecosystems even though recharge and long term storage were assured. Following grid refinement, ten injection wells were located perpendicular to the Gatesburg groundwater trough dominated by diffuse flow. In this second analysis four injection rates were considered (i.e., 1.6, 3.2, 4.7, and 6.3 l/s/well). One simulation is shown in Fig. 4. These results illustrate the opportunity of using stormwater to enhance aquifer recharge and storage. This unused method of water management has immense potential to sustain future water demands within some carbonate aquifers. A third modeling analysis examined a ribbon-like aquifer within the Valley and Ridge due to folding and faulting of relatively thin carbonate strata. Despite favorable well yields, their sustained-yield potential is limited by available recharge. Two dimensional finite difference models were developed for two areas that feature this hydrogeologic setting (Mundi 1972; Meiser 1995; Parizek and Mundi 1990). The model included dipping clastic confining beds, poorly permeable soil-colluvial overburden, and mapped water table. Recharge rates, aquifer storage, and transmission properties were varied to estimate the yield potential of a 25 km length of this aquifer. Fourteen production wells were assumed, each yielding 63.1 l/s, three wells at 31.6 l/ s together with five recharge wells each injecting 31.6 l/s. Natural recharge was assumed to vary from 127- to 254 mm/year for this 153 m thick aquifer and 2.44 km outcrop width. Storativity was assumed to be 0.08 and

Fig 2 a Water-table configuration [ft] within the model domain showing regional and local groundwater troughs (From Parizek et al. 1967); b calibrated water table map (Modified from Chin and Parizek 1997)

aquifer transmissivity, 124.2 and 1,242 m2/day. Where streams crossed the dipping aquifer, streambed seepage was based on the observed rate and assigned a value of 1.02 m2/day. A transmissivity value 17.69 m2/day was assigned to shale aquitards that confine this aquifer. Figure 5 illustrates drawdown for two assumed transmissivity values. Depending on aquifer hydraulic properties that exist for portions of this aquifer and injection rates that might be adopted, its sustained yield could be

734

14 Wells at 63 l/s 3 Wells at 31.6 l/s 5 Recharge Wells at 31.6 l/s T = 92.9 m2/day M = 76.2 m Recharge 12.7 cm/yr Time 3.75 yrs

14 Wells at 63 l/s 3 Wells at 31.6 l/s 5 Recharge Wells at 31.6 l/s T = 1242 m2/day M = 76.2 m Recharge 12.7 cm/yr Time 3.75 yrs

Fig 3 Simulated cones of impression [ft.] under steady state conditions. The combined injection rate is assumed to be 180.8 l/s

increased by a factor of two to five times over that of the natural rate (Parizek and Mundi 1990). No communities are known to have adopted this significant water-management option despite the presence of thousands of kilometers of such narrowly exposed aquifers in close proximity to towns and cities within the Valley and Ridge. Recycling municipal waste water Today, a high level of treatment is required before municipal and industrial effluents are released to surface. Some beneficial uses of waste water include irrigation of

Fig 4 Simulated cones of impression [ft.] resulting from ten injection wells each receiving 3.16 l/s

Fig 5 Cumulative drawdown [ft.] in a thin carbonate aquifer with two transmissivity values tapped by 14 production wells each with a 63l/s yield, 3 with a 31.6 l/s yield together with five recharge wells each injecting 31.6 l/s. (Modified from Parizek and Mundi 1990)

cropland and turf, cooling, secondary oil recovery, and control of salt water encroachment. The need to achieve a higher level of treatment and enhanced recharge has led to development of land application systems. Overland flow, spray, and trickling irrigation methods are in use, as well as natural and engineered wetlands. Unwanted contaminants must be compatible with renovation processes operating either in uplands or wetlands. Suitable sites must be selected and projects managed to ensure long term success of land-treatment systems. Sites can be selected to optimize recharge, groundwater storage and reuse (Parizek et al. 1967). The Pennsylvania State University relies on this ‘‘Living Filter’’ concept to achieve tertiary treatment of effluent, reuse of nutrients and to enhance recharge. Following eleven years of research and experimentation with application rates (2.54, 5.8, 10.16 and 15.24 cm/ week), a 5.8 cm/week rate was selected for year round use under Pennsylvania’s northern climate. Both cropland and forests are irrigated to treat all of Penn State’s secondary effluent, up to 1.514 · 107 l/day. Two spray fields were selected underlain by thick residual soils (3.1–47.5 m) derived from the weathering of carbonate rock together with a deep water table (30– 91 m). Full operation began in 1983 using a solid set spray system. Irregular surface topography resulting from the differential weathering of carbonate bedrock ensures the retention, infiltration and recharge of effluent and precipitation. Annually, the sites receive 3.62 m of water resulting in nearly 2.49 m average recharge. Area wide recharge approaches 3.785 · 109 l/year. This

735

exceeds the annual groundwater withdrawal rate required to sustain a campus population > 55,000. The FEWA finite element flow and transport model used to simulate cones of impression discussed in the first two examples was developed to address the transport of nitrate within the diffuse fracture-flow dominated carbonate aquifer that underlies these sprayfields (Chin and Parizek 1997). The calibrated steady-state flow model (Fig. 2b) includes enhanced recharge derived from Penn State’s two spray fields.

Conclusions Worldwide competition for available water supplies will increase as a result of growing population and economic

activity. Innovative water development and management plans must be devised and adopted to meet these water requirements while protecting aquatic resources and ecosystems. Enhancing groundwater recharge through recycling of wastewater and capture of excess stormwater offer but two examples. Hydrogeologic frameworks must be understood to select and optimize surface and groundwater management schemes that consider all available sources of water. Numerical flow and transport models provide helpful tools to assist in this process, but require new insights to address conduitdominated karstic aquifer characterization and modeling requirements. Acknowledgments The author thanks E. Baumrucker and B.R. Parizek for their editorial comments.

References Brown CE (1977) Multivariate analysis of petrographic and chemical factors influencing aquifer properties in selected carbonate aquifers on Central Pennsylvania. Unpublished PhD Dissertation, The Pennsylvania State University, University Park, PA, 206 pp Chin CV and Parizek RR (1997) Hydrogeologic and water quality investigations. In: 1996 Forecast of loading report for the Pennsylvania State University, University Park Campus, Waste Water Treatment Plant, State College, PA, Office of Physical Plant, Operations Division, 64 pp Gidding Jr MT (1974) Hydrologic Budget of Spring Creek Drainage Basin, Pennsylvania. Unpublished PhD Dissertation, The Pennsylvania State University, University Park, PA, 76 pp Konikow LF (1969) Mountain runoff and its relation to precipitation, ground water and recharge to the carbonate aquifers of Nittany Valley, Pennsylvania. Unpublished Master’s Thesis, The Pennsylvania State University, University Park, PA 128 pp Meiser Jr EW (1995) Ground water geology and digital simulation of sustained yield potential of the Altoona area, Pennsylvania, The Pennsylvania State University, University Park, 183 pp

Mundi EK (1972) The physical characteristics of some fractured aquifers in Central Pennsylvania and a digital simulation of their sustained yields. Unpublished PhD Dissertation, Department of Goesciences, The Pennsylvania State University, University Park, PA, 250 pp Parizek RR, Kardos LT, Sopper WE, Meyers EA, Davis DD, Farrel MA, Nesbitt JS (1967) Waste water renovation and conservation, Penn State Studies Series 23, The Pennsylvania State University, 74 pp Parizek RR (1987) Identification of ten potential high capacity well fields in the greater State College, PA Region, Phase III Study for the Center Regional Planning Commission, State College, PA Parizek RR, Mundi EK (1990) Artificial ground-water recharge: an under used water resources management option, Chpt 22:367–283. In: Majumdar SK, Miller EW, Parizek RR (eds) Water resources in Pennsylvania: availability, quality and management. The Pennsylvania Academy of Sciences, Typehouse of Easton, Philipsburg, 580 pp Raunch HA, White WB (1970) Lithologic controls on the development of porosity in carbonate aquifers. Water Resour Res 6(4)1175–1192 Schmotzer JK, Jester WA, Parizek RR (1974) Groundwater tracing using post activation analysis techniques. J Hydrol 20(3) 217–236

Shuster ET, White WB (1971) Seasonal fluctuations in the chemistry of limestone springs: a possible means for characterizing carbonate aquifers, J Hydrol 14:93–128 Siddiqui SH (1969) Hydrogeologic factors influencing well yields and aquifer hydraulic properties of folded and faulted carbonate rocks in Central Pennsylvania, Unpublished PhD Dissertation, The Pennsylvania State University, University Park, PA, 502 pp Siddiqui SH, Parizek RR (1971) Hydrogeologic factors influencing well yields in folded and faulted carbonate rocks in Central Pennsylvania. Water Resour Res 7(5)1295–1312 Smith M (1986) Evaluating ground-water recharge potential in the valley and ridge region of the Central Appalachians using a digital-overlay technique. Unpublished MS Thesis, The Pennsylvania State University, University Park, PA, 289 pp Underwood KL (1994) Evaluation of pesticides and nitrate mobility in a conduitflow dominated basin. Unpublished M.S. Thesis, The Pennsylvania State University, University Park, PA, 205 pp Yeh GT, Chang RJ (1993) A finite element model of water flow through aquifers: version 2.0, Class notes issued May 1993, Department of Engineering, Pennsylvania State University