The environmental hazards of locating wastewater impoundments in karst terrain

Engineering Geology 65 (2002) 169 – 177 www.elsevier.com/locate/enggeo The environmental hazards of locating wastewater impoundments in karst terrain Bashir A. Memon *, M. Mumtaz Azmeh, Mary Wallace Pitts P.E. LaMoreaux & Assoc., Inc., Tuscaloosa, AL 35403, USA Abstract A wastewater storage lagoon failed due to the development of a sinkhole at a site in the Lehigh River valley in Allentown, Pennsylvania (PA). The polluted wastewater from the lagoon entered into the underlying aquifer and moved within a narrow pathway controlled by cracks, fissures, and solution channels within the karstified Allentown Formation of the Cambrian Period. The Allentown Formation serves as the principal aquifer for the public water supply of the area. To develop appropriate remedial measures, a thorough understanding of the geologic setting was required. Therefore, a geologic and hydrogeologic characterization of the area was completed, aerial photography and satellite imagery interpretations were performed, stratigraphic core holes were drilled and geophysically logged, and the data correlated to define structural control and movement of ground water and pollutants. A number of wells were drilled and constructed, and water levels were monitored on a continuous basis to correlate with climatic changes and determine the direction of flow. Water samples were collected periodically and analyzed to delineate the vertical and lateral extent of migration of pollutants. Five saturated (water-bearing) zones were identified within the bedrock based on the analysis of cores and interpretation of geophysical logs. Ground water in the lower zones is polluted; the concentration of pollution increases with depth. Monitoring stations were established in the creek, south of the site, to measure flow rate several times during different seasons, and at different reaches, to determine the losing and gaining sections of the creek. Pumping tests were conducted to determine hydraulic characteristics of the aquifer. Based on the hydrogeologic model of the karstified aquifer, flow regime and structural control, a plan of action was defined and initiated to remediate the aquifer. The ground water is being remediated using a pump and treat methodology. The cleanup effort is continuous and the pollutant level is fluctuating with an overall-declining trend. The application of this technology has also created a pressure trough, thereby controlling off the site migration of pollutants. D 2002 Published by Elsevier Science B.V. Keywords: Ground water; Wastewater; Karst aquifer 1. Introduction Ground water is a valuable water supply source. It serves as an important resource in all climatic zones throughout the world. Its use for agricultural, industrial, municipal, and domestic purposes continues to * Corresponding author. E-mail address: pela@dbtech.net (B.A. Memon). grow at increasing rates because of its good quality, convenient availability near the point of use, and relatively low cost of development. Karst aquifers are an important source of water supply for private and municipal uses. Incidents of pollution of karst aquifers due to leaking of leachate from abandoned dump sites, waste management facilities, improperly constructed and/or located waste lagoons and impoundments, uncon- 0013-7952/02/$ - see front matter D 2002 Published by Elsevier Science B.V. PII: S 0 0 1 3 - 7 9 5 2 ( 0 1 ) 0 0 1 2 5 - 9 170 B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 trolled disposal of used solvents and chemicals, accidental spills of hydrocarbs, and chemical or waste materials may reduce the availability of good quality ground water. Regional growth may be impaired due to the lack of a good quality water supply. Remediation of polluted karst aquifers requires a three-step approach. (a) Development of a hydrologic model to thoroughly understand the geologic and hydrologic conditions, and to define flow paths within the aquifer. (b) Management of the plume to prevent the movement of pollutants into usable aquifer zones. (c) Remediation and restoration of water quality to its baseline quality by removal of the source of pollution and remediating the polluted ground water. The ground water may then be returned to its beneficial use within a reasonable timeframe using cost-effective available technologies. 2. Polluting of the karst aquifer A process wastewater pond was constructed on the site of an industrial manufacturing plant in Allentown, Pennsylvania (PA). The pond was constructed with a clay liner to prevent, or at least minimize, vertical infiltration of wastewater into the underlying karst aquifer (Fig. 1). In 1979, a long dry weather spell was followed by heavy rains, which caused a sinkhole to develop under the wastewater pond. The pond liner failed and a large volume of wastewater moved quickly through the sinkhole into the underlying aquifer. The wastewater moved north of the ground water divide (see Fig. 2) because of mounding conditions caused by the introduction of a significant volume of wastewater into the underlying aquifer in a short duration. After the incident, ground water samples were collected from existing domestic shallow wells in the area and analyzed to determine if the ground water from these wells was polluted by the wastewater. No pollution was identified in the ground water from any of these local wells. Several years after the incident, two public water supply wells, and a test well, were constructed to the northeast of the Plant site, and after a few months of pumping from two public supply wells, pollutant was detected in the ground water. Pumping was suspended. A hydrogeological study was conducted to characterize the site and determine the extent of contamination. 3. Site characterization 3.1. Physiography and climate The site, at an elevation ranging from 300 to 450 ft above mean sea level (amsl), is an industrial manufacturing plant in the east-central section of Pennsylvania in the Lehigh River valley. Twelve miles to the north is Blue Mountain, a ridge with an elevation of 1000 – 1800 ft amsl. South Mountain, with an elevation of 500 –1000 ft amsl, fringes upon the southern edge of Allentown, PA. Otherwise, the topography of the area is characterized in general by rolling hills with numerous small streams. The site is in the low-lying portion of the Great Valley section of the Valley and Ridge Physiographic Province in Lehigh County, which is underlain primarily by carbonate rocks (Wood et al., 1972). Surface drainage in this broad area of gently rolling hills, bordered to the north by Blue Mountain and to the south by South Mountain, is controlled by a few irregularly spaced and shallow entrenched streams that have gentle valley slopes and transect the area. The site is bordered to the north and south by tributaries of the Lehigh River. Considerable structural deformation of the carbonate rocks has facilitated the process of weathering and solution of carbonate rocks, and therefore, the area has typical karst-terrain features including underground caverns and sinkholes. The climate in the area is characterized by a humid continental-type, where annual precipitation significantly exceeds natural losses by evapotranspiration. Rainfall records indicate a mean normal annual precipitation of 44.31 in. for the period of 1951– 1980. The wettest month of the year is August, which has a normal rainfall of 4.44 in. The driest month of the year is October, which has a normal rainfall of 3.05 in. The average monthly temperature is 51.0 jF. Temperatures as high as 105 jF and as low as 12 jF have been recorded (Wood et al., 1972). 3.2. Geology The site is underlain by the Allentown Formation of the Cambrian Period. Residuum is formed at the top of bedrock from weathering of the Formation. Data from drilling and construction of 11 wells 171 Fig. 1. Geologic map. B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 Fig. 2. Ground-water surface map for deeper zones 9-14-93. 172 B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 indicate that the residuum is typically composed of two distinct units. The upper 10 ft of the residuum is predominantly fine-grained sandy-clay to silty-clays, which are strong brown (10YR4/3) to yellowish-brown (10YR4/6). Colors at depths below 10 feet are yellow (10Y7/8), yellowish-red (5YR5/6), and reddish-yellow (7.5YR6/8). The thickness of the residuum varies significantly over relatively short distances, and ranges from 10 to 97 ft at the site. The Allentown Formation exhibits little or no intercrystalline porosity. Ground water movement in the karst aquifer occurs along solutionally enlarged fractures and bedding planes. During drilling, numerous voids or cavities were penetrated, some of which were clay-filled. Heights of the cavities range from a few inches to 5 feet. Caliper logs indicate that many smaller cavities may exist which were not identified during drilling. In the vicinity of the site, the Epler Formation has been thrust over the Allentown Formation along the Portland Thrust Fault (Berg and Dodge, 1981). Subsequently, the Epler, Allentown, and the intervening Portland Thrust Fault were folded into a series of east – west trending anticlines and synclines (Fig. 1). Erosion has locally removed the Epler Formation, creating windows in which the underlying Allentown Formation is exposed. The irregular shape of the windows and the trace of the Portland Thrust Fault are a result of the effect of erosion and the underlying geologic structure. In general, the axes of the folds are nearly horizontal, and the northern limbs of the anticlines dip more steeply than do the southern limbs. The angle of dip varies from 15j to 80j, and local beds are overturned. 3.3. Well construction Two public supply wells, and a test well, were constructed approximately 4500 ft northeast of the failed wastewater pond on the site. These wells were drilled to depths of 350, 400, and 360 ft, respectively. As pollution was identified in these wells, monitoring wells on the site were sampled in an effort to determine the extent of pollution. All wells on the site were found to be free of pollution. Other wells were located along identified lineaments, and constructed between the failed wastewater pond and the polluted water 173 supply wells. These wells were also found to be free of pollution. Then a corehole was drilled in the vicinity of the water supply wells to a depth of 400 ft below land surface (bls). Geophysical logs and core descriptions were used to identify five saturated (water-bearing) zones within the Allentown Formation. Approximate depths of these zones are as follows: zone 1 is from 175 to 190 ft bls, zone 2 is from 205 to 220 ft bls, zone 3 is from 231 to 258 ft bls, zone 4 is from 294 to 320 ft bls, and zone 5 is from 330 to 400 ft bls. To identify the polluted zone(s), the corehole was reamed and constructed into a monitoring well, cased down to a depth of 325 ft bls, thus tapping zones 4 and 5, which were polluted. A shallow well was constructed to a depth of 187.4 ft bls with 138 ft casing tapping zone 1. This shallow well was found to be clean. A third intermediate well was constructed to a depth of 265 ft bls with 223 ft of casing, tapping zones 2 and 3. This well was found to be moderately polluted. The residuum was cased off in the water supply wells and the test wells above the first saturated bedrock zone. As pollutant was found only in the lower zones of the aquifer in the vicinity of the public supply wells, it was concluded that the pollutant had migrated to and pooled in that area at the time of the pond failure. 3.4. Hydrogeology A topographic ridge extending from east to west crosses the site. Surface water from the area to the north of this ridge drains into tributaries of Coplay Creek. There are large sinkholes north of the ridge that may cause a hydraulic connection between surface water and ground water in this area. The surface water, from the area south of this ridge, drains into Jordan Creek. Both creeks eventually discharge into the Lehigh River. Ground water in the area, also, ultimately discharges into the Lehigh River. Considerable structural deformation of the carbonate rocks has facilitated the process of weathering, and the area has karstic geomorphic features, as well as underground cavities and welldeveloped fracture systems. The carbonate rocks of the Allentown Formation act as a hydrogeologic unit under artesian conditions. These rocks contain and transmit ground water 174 B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 through fracture systems and cavities. Most of which are developed with 150 – 480 ft bls. Precipitation is the main source of recharge to the ground water. The overburden (residuum) acts as a sponge and recharges ground water where it is in hydraulic connection with openings in the underlying bedrock. Ground water is recharged by runoff, which percolates downward through the joints, fissures, fractures, and cavities in the carbonate rocks. Flow measurements in the creek are effected by creek bed irregularities, flow velocities, variable channel storage, and changing stages. Therefore, the measured flow of the creek, determined by conventional stream flow techniques, which also have inaccuracies inherent to the measurement procedures, has a margin of error of 5% or more. 3.5. Water-level monitoring Aquifer tests were performed to determine the hydraulic characteristics (transmissivity, hydraulic conductivity, and storativity) of the carbonate aquifer and evaluate the hydrogeologic conditions at the site. Test wells, observation wells, stilling wells, and pumping wells were equipped with Stevens recorders and/ or electronic data loggers to collect water-level data on a continual basis prior to initiation of pumping tests, during, and after termination of pumping. The flow in the creek was measured at four stations during the pumping tests to determine if a hydraulic communication existed between surface water and the underlying carbonate aquifer. Evaluation of flow data collected during the pumping tests at selected sections in the creek demonstrates that reaches within the site limits do not lose significant amounts of water. Although there may be a small loss of stream flow within some reaches, water would not have the opportunity to travel very far from the channel, as the water is flowing as underflow, and resurfaces into the channel downstream, as indicated by flow measurements. The evaluation of ground-water level data collected during a pumping test at the monitoring wells, surface-water hydrographs, and flow data for Jordan Creek, indicated that there was no hydraulic communication between the creek and the aquifer during the pumping tests. The drawdown and recovery data collected during the tests were plotted and analyzed, using Theis and Walton methods (Lohman, 1979; Walton, 1970; Driscoll, 1987). The value of transmissivity, computed from time-drawdown and time-recovery field data graphs of the aquifer, ranges from 14,500 to 47,750 gallons per day per foot (gpd/ft). Using a saturated thickness of the aquifer as 50 ft, the hydraulic conductivity was calculated to be in the range from 190 gallons per day per square foot (gpd/ft2) to 955 gpd/ft2. All wells at the site were equipped with Stevens recorders to continuously record changes in water level. The water level in these wells along with three private wells, and three wells (shallow, intermediate, and deep wells) were measured. The depths to water level from measuring points were converted into water-level elevations using surveyed elevations of the measuring points. Using water-level data collected from all wells, except the deep monitoring well (Fig. 3), a ground water surface map was prepared (Fig. 2). The map indicates a steep gradient along the west and southwest, suggesting a poorly developed fracture system. The map also indicates the presence of a ground water divide at the site. Ground water flow north of the divide moves to the northeast, whereas ground water south of the divide flows to the south. The divide, which lies north of the failed pond, controls the ground water movement under normal flow conditions. However, failure of the pond and the resulting mounding made the ground water divide temporarily ineffective in controlling the normal flow pattern, and allowed migration of wastewater towards the public supply well. 3.6. Stream gaging Four stations were established in Jordan Creek to measure the flow in the creek. The measurements of flow at each station in the creek were taken periodically using a current Price AA meter and Pygmy meter model number 625-F (Buchanan and Somers, 1980, 1982). Flow measurements were also taken using a flowmeter (FLO-MATE Model 2000). The discharge of Jordan Creek is controlled by the complex interaction of channel characteristics, including cross-sectional area, shape, slope, and roughness. 3.7. Aquifer tests B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 Fig. 3. Pentaerythritol (mg/l) in two remediation wells and deep monitoring well (July 1993 – August 1998). 175 176 B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 The coefficient of storage ranges from 1.1  10 9.8  10 3. 2 to 3.8. Water samples for chemical analysis Water samples were collected from monitoring wells located within the site, public supply wells, monitoring wells and test wells close to public supply wells, and from the Creek. Water samples were analyzed for the following parameters: total dissolved solids (TDS), carbon oxygen demand (COD), total organic carbon (TOC), sulfate (SO4), and pentaerythritol (PE). The concentrations of these parameters, except PE, are within the permissible level. The pollutant identified to be originating from the failed wastewater pond is PE, which has been continuously monitored at selected wells since 1993. 4. Remediation To remediate ground water, public water supply wells numbers 1 and 2 were converted to remediation wells by plugging them with bentonite from the bottom of each well up to a point approximately 275 ft bls (Fig. 3). These plugs prevented entry of ground water into the well from the two deepest zones (zones 4 and 5), which have the highest concentration of pollutants. A pump was installed in each of the two remediation (public water supply) wells. The two former public supply wells (remediation wells) tapped the shallow and intermediate zones. A third pump was installed in the deep monitoring well, in the vicinity, tapping the lowest two zones. Pumping was initiated simultaneously from all three wells. The rate of pumping from the deep monitoring well is maintained equal to the combined rates from the two water supply wells to prevent migration of polluted water from the deepest zones into the upper zones. The recovered water is treated and then discharged into the creek. The remediation effort is continuous and the pollutant level is fluctuating with an overall declining trend in the two remediation wells and the deep monitoring well. Fig. 4 shows the analytical results for PE in the two remediation wells (water supply wells), and deep monitoring (pumping) wells over time. The operation of this system has also created a pressure trough, thereby controlling off site the migration of pollutants. Fig. 4. Construction details and gamma logs. B.A. Memon et al. / Engineering Geology 65 (2002) 169–177 5. Conclusions The ground water flow regime in the area is primarily controlled by secondary porosity and permeability due to karstification. . The failure of the wastewater pond liner was due to the development of a sinkhole beneath the pond. . Hydrogeological studies indicated the presence of a ground water divide. However, failure of the pond liner and the resultant mounding made the ground water divide temporarily ineffective in controlling normal flow paths. . The complexity of the karst system required a thorough understanding of the geologic and hydrogeologic setting, and geologic structural control to design and implement effective remediation measures. . To facilitate future extraction for public supply from the shallow and intermediate zones, these zones were targeted for remediation. The deeper impacted zones were plugged to control the migration of pollutants. . A pump and treat technology for remediation was implemented in the water supply wells tapping the shallow and intermediate zones. In order to prevent upconing, pumping was also initiated from a nearby well completed in the deeper zones. Pumping from the . 177 deep well is maintained at a rate equal to the combined discharge from the supply wells. . Concentrations of pentaerythritol in the remediation wells, as well as the deep monitoring well (pumping well) have declined over time. References Berg, T.M., Dodge, C.M., 1981. Atlas of Preliminary Geologic Quadrangle Maps of Pennsylvania: Pennsylvania Geologic Survey Map 61. Buchanan, T.J., Somers, W.P., 1980. Discharge Measurements at Gaging Stations (Chapter A8): U.S. Geological Survey, Techniques of Water-Resources Investigations of the United States Geological Survey, Book 3, Applications of Hydraulics, 65 pp. Buchanan, T.J., Somers, W.P., 1982. State Measurement at Gaging Stations (Chapter A7): U.S. Geological Survey, Techniques of Water-Resources Investigations of the United States Geological Survey, Book 3, Applications of Hydraulics, 28 pp. Driscoll, F.G., 1987. Groundwater and Wells, 2nd edn. Johnson Division, St. Paul, MN, 1089 pp. Lohman, S.W., 1979. Ground-water hydraulics. U.S. Geological Survey Professional Paper 708, 70 pp. Walton, C.W., 1970. Groundwater Resource Evaluation. McGrawHill, New York. Wood, C., Flippo Jr., H.N., Lescinsky, J.B., Baker, J.L., 1972. Water Resources of Lehigh County: U.S. Geological Survey Water Resources Report 31.