Total Carbonate Hardness in Cumberland Valley Groundwater A Shippensburg University Practical Exam

Dana Heston 3/6/2015

TABLE OF CONTENTS INTRODUCTION

2

LITERATURE REVIEW

3

Hard water

3

Impacts on Health

5

Treatment Options

6

Specific Conductance

7

STUDY AREA

7

Geology

8

METHODS

10

Cumberland and Franklin County Data

11

Big Spring

12

RESULTS

13

Field Sites

13

Conductance and Hardness

14

DISCUSSION

15

Outliers

17

Specific Capacity and Sustained Yield

18

CONCLUSIONS

18

REFERENCES

19

TABLES, EQUATIONS, AND FIGURES Figure 1

8

Figure 2

9

Figure 3

15

Table 1

4

Table 2

14

Equation 1

3

Equation 2

12

1

INTRODUCTION Hard water is water that contains a high concentration of multivalent cations, commonly calcium (Ca) and magnesium (Mg) (Evens et al., 2013). These ions react with soap, preventing it from forming lather (Evens et al., 2013). Hard water is generally not considered to be harmful to human health, but makes cleaning difficult and leaves a residue called lime scale on surfaces. Hard water can be difficult to drink because of the unpleasant taste associated with the excessive mineral content. Geology plays an important role in the hardness of groundwater. Water becomes hard when mineral dissolution occurs as the flowing water comes in contact with rock and soil. Groundwater is generally harder than surface water and hardness increases as a function of the distance of the flowpath (Becher and Root, 1981). The longer the water is exposed to soluble bedrock, the harder the water will be when it is pumped for public or domestic use. Water molecules surround the disassociated Ca and Mg ions, preventing them from recombining (Meena et al., 2012). Water flowing through gypsum, limestone, and dolostone bedrock can be predicted to be hard. Conversely, water flowing through bedrock high in silica is less likely to undergo extensive mineral dissolution. The intention of this study is to answer the following questions regarding existing techniques used to calculate water hardness and the relationship between specific conductance (SpC) and hardness: 

Becher and Root showed a relationship between SpC and hardness that can aid in determining the concentration of calcium carbonate (CaCO3) in a water sample

2

by multiplying the SpC by 0.48. Is this an accurate method for calculating water hardness? 

When additional lab data are available the concentrations of Ca and Mg can be used to calculate CaCO3 concentrations. Will using this method show a relationship between SpC and hardness similar to the one identified by Becher and Root (1981)?



There are several outliers in the data, evident on the scatterplot. Why are these data points not displaying the same trend as the majority of the data? Are these water samples unique, or are data entry errors affecting the analysis? LITERATURE REVIEW

Hard Water Polyvalent, metallic ions dissolved from rock and soil cause water to become hard. Calcium (Ca) and magnesium (Mg) are the most common ions even though any positively charged divalent ions, such as iron (Fe), strontium (Sr) and manganese (Mn) can cause hard water (Meena et al., 2012). The Ca and Mg react with bicarbonate (HCO3) when water is heated to form an insoluble precipitate. The precipitate left behind is referred to as scale and leaves a residue inside pipes and on surfaces as water evaporates. The chemical reaction of hard water to scale is shown in equation 1 (Casiday and Frey, 1998). Ca 2 (aq)  2HCO3 (aq)  CaCO3 (s)  H 2 O  CO2 Equation 1: The formation of scale when hard water is heated (Casiday and Frey, 1998).

3

Water hardens as it flows through bedrock such as limestone, dolostone, marble, and gypsum. High concentrations of Ca and Mg in the bedrock quickly add these ions to the water as dissolution occurs. High concentrations of CO2 in soil and subsurface voids increase the amount of carbonic acid (H2CO3) in groundwater, which increases the rate of mineral dissolution. Carbon dioxide (CO2) is heavier than oxygen (O2) and is found in stronger concentrations at increased depths (Atkinson, 1977). Subsurface gases diffuse from high pressure to low pressure environments, such as voids. Caves and subsurface channels in karst landscapes often have elevated levels of CO2, as well as the fractures leading from these features due to the migration of CO2 into the void spaces. The stronger concentration of H2CO3 increases the rate of mineral dissolution allowing water to become hard even though conduit flow is likely in this environment. Water hardness is also increased in areas with increased calcium carbonate (CaCO3) in the soil and high concentrations of CO2 in the unsaturated zone (Atkinson, 1977). Table 1 shows the scale used by the Environmental Protection Agency (EPA) to determine the hardness of water by calculating the amount of CaCO3 (1986).

Table 1: Environmental Protection Agency (EPA) hardness scale for concentrations of calcium carbonate in water (EPA, 1986).

Classification

CaCO3 in milligrams per liter (mg/L)

Soft

0-75

Moderately Hard

75-150

Hard

150-300

Very Hard

300 and up

4

Impacts on Health The EPA has not set a maximum contaminant level (MCL) for hard water. Calcium and Mg are essential for human health in adequate quantities and hard water can contribute to these dietary needs. Research indicates that consumption of hard water can actually improve health if the ratio of Ca to Mg is within an acceptable range. The concentration of Mg is important in determining the effect hard water can have on human health (Evens et al., 2013). Total hardness of water with a concentration greater than 200 milligrams per liter (mg/L) with a Mg concentration less than 7 mg/L can cause adverse effects on human health (Evens et al., 2013). Calcium is more abundant than Mg in the environment and consuming large quantities can be harmful. Magnesium acts as a natural calcium antagonist by competing for binding sites in vascular muscle (Evens et al., 2013). The Mg content must be adequate to control the amount of Ca absorbed. A study conducted in Sweden indicated that a high Ca intake could lower s-cholesterol and s-LDL, but increased systolic blood pressure (SBP) (Nerbrand et al., 2003). Epidemiological evidence suggests that Mg plays an important role in regulating blood pressure (Anne, 2011). Drinking water with 10-100 parts per million (ppm) Mg could potentially prevent 4.5 million hearth disease and stroke deaths per year, worldwide (Rosanoff, 2013).

5

Treatment Options There are different treatment options available to soften hard water. Reducing the temperature of a water heater can help to reduce the amount of scale left behind by reducing the amount of precipitate formed. This is the most economical treatment option. Cation exchange is a popular option for softening domestic water supplies. The Ca and Mg ions in the hard water are exchanged with sodium (Na) or potassium (K). The cation exchange technique can remove nearly all of the Ca and Mg and even 5-10 ppm of Fe and Mn (Skipton and Dvorak, 2014). However, the cation exchange method increases the Na concentration in water which can have adverse effects on human health if consumed in large quantities. Cation exchange systems can be installed to soften only water going into the water heater to alleviate the concern of increased sodium intake from ingesting softened water from cooking and drinking. Reverse osmosis forces water through a filter to remove minerals and other contaminants. This technique avoids the use of salt but very hard water can clog the filter making it an expensive alternative in areas of very hard water (Cool Today, 2015). It is possible to use a water softener in conjunction with a reverse osmosis filtration system to extend the life of the filters in hard water areas. Large municipal supplies will often use a lime-soda process for ion exchange. Lime and soda ash are added to the water, combining with the Ca and Mg. The Ca precipitates as calcium carbonate (CaCO3), and the Mg as magnesium hydroxide

6

(Mg(OH)2) (Casiday and Frey, 1998). The precipitates are then removed through filtering. Specific Conductance Electrical conductivity is a measure of charged ions in the water. Pure water is a poor conductor of electricity until ions are dissolved. This parameter is temperature dependent and specific conductivity (SpC) is reported in most studies because it is normalized to 25 degrees Celsius. The Ca and Mg ions that are reported as total hardness contribute to SpC concentrations. Clean limestone waters typically have a total hardness value of close to half of the SpC (Krawczyk and Ford, 2006). This is confirmed in the formula of 0.48 * SpC given in Becher and Root (1981). Specific conductance values can also be impacted by the presence of ions such as K, Na, Cl, NO3, and SO4 that are dissolved in water from both natural and anthropogenic sources (Krawczyk and Ford, 2006). STUDY AREA The study area includes Franklin and Cumberland counties in south-central Pennsylvania, part of the Cumberland Valley portion of the Great Valley. The Cumberland Valley lies between South Mountain and Blue Mountain in the Ridge and Valley Physiographic Province (figure 1). The structural geology of the study area is complex with deformation in the rock formations common due to the Taconic, Acadian, and Alleghenian Orogenies (mountain building events) that formed the Appalachian Mountains.

7

Figure 1: The Cumberland Valley study area lies between Blue Mountain and South Mountain in south-central Pennsylvania.

Geology The stratigraphic column (figure 2) shows the geology of the study area which spans the colluvium deposited at the base of South Mountain to the Lower Cambrian Tomstown Formation. The northwestern portion of the study area is comprised of the Martinsburg Formation, which contains dark-gray shale with siltstone interbeds, and a fine-grained greywacke. The base of the Martinsburg is calcareous shale as the lithology transitions to a sequence of Ordovician carbonate formations (Becher and Root, 1981).

8

Figure 2: Stratigraphic column representing the geology of the Cumberland Valley (from Becher and Root, 1981).

The geologic formations of the study area are discussed youngest to oldest, consistent with the geologic law of superposition. The Chambersburg Formation is the youngest of the Ordovician Carbonates within the study area. The thin-bedded limestone weathers to cobbles and contains argillaceous partings and bentonite beds in 9

the upper portion (Becher and Root, 1981). These bentonite beds are the last clay deposits observed in the study area until reaching the Lower Cambrian Waynesboro Formation. The Saint Paul Group is a thick-bedded fossiliferous limestone with dolostone interbeds and chert nodules (Becher and Root, 1981). The lithology transitions to dolostone in the Pinesburg Station Formation. Quartz rosettes are present in the lower portion of the Pinesburg Station Formation. The quartz rosettes continue into the underlying Rockdale Run Formation, which is a fine-grained skeletal limestone (Becher and Root, 1981). The Stonehenge, Stoufferstown, Shady Grove, and Zullinger Formations are all predominately limestone formations. The Elbrook Formation indicates a change in the depositional environment with a change in the dominant lithology. The Elbrook Formation consists of a combination of calcareous shale and argillaceous limestone (Becher and Root, 1981). The Waynesboro Formation contains quartzite beds with worm burrows and ripple marks (Becher and Root, 1981). The oldest formation in the study area is the Tomstown Formation, a mottled dolostone with calcareous shale and limestone near the base (Becher and Root, 1981). METHODS Specific conductivity (SpC) was measured in the field using a YSI ProPlus field meter. Prior to data collection the meter was calibrated using a Ricca 250 microsiemens per centimeter (µs/cm) specific conductivity standard. Due to freezing temperatures, data collection sites were chosen from creeks fed by carbonate springs in the

10

Cumberland Valley and included: Middle Spring Creek, Dykeman Spring Creek, and Green Spring Creek. Each field measurement of SpC was multiplied by 0.48 to obtain the hardness in mg/L of CaCO3 (Becher and Root, 1981). A water sample was collected while measuring specific conductance in the field to perform a titration. The titration was performed to confirm the results of the calculation applied to the field data from Becher and Root (1981). The results of both tests were compared in table 2. The water samples were titrated using 0.16 N sulfuric acid (H2SO4), consistent with USGS protocol, to determine the amount of CaCO3. A Hach digital titrator was used to add the H2SO4 with 800 drops of acid equaling 1 mg/L. The current USGS standards require that water samples that are not filtered are titrated for acid neutralizing capacity (ANC) instead of alkalinity. The results of the titrations were entered into the U.S. Geological Survey’s online alkalinity calculator (U.S. Geological Survey, 2013) using the inflection point method to calculate ANC and determine the amount of CaCO3 present in each sample. The inflection point is set at a pH of 4.5 and the results report the amount of calcium carbonate (CaCO3), bicarbonate (HCO3), carbonate (CO3), and hydroxide (OH). Cumberland and Franklin County Data The existing water chemistry data from the Cumberland Valley was plotted in Excel using a scatterplot to show the relationship between hardness and specific conductance. Three outliers were investigated and determined to be due to data entry errors after reviewing the published data (from Becher and Root, 1981; Becher and Taylor, 1982). The errors were corrected and the correct data points added to the scatterplot. Three existing outliers showing unique trends were explored. A trendline 11

with a Y intercept of 0,0 was added to the scatterplot to determine the slope of the line. The slope value was compared to the 0.48 multiplier (from Becher and Root, 1981). Big Spring Data Big Spring Creek originates from a limestone spring in the Shadygrove Formation one mile north of Stoughstown, Pennsylvania and flows northeast to its confluence with the Conodoguinet Creek near the borough of Newville, Pennsylvania (Greene, 2002). Chemical and bacterial analyses of Big Spring Creek (Sp-22) were conducted in 1971 and 1974 and the results published in Becher and Root (1981). Pennsylvania State University (Penn State) collected Ca, Mg, and SpC data for 13 wells in 2006. Dr. Feeney provided these data for analysis in this study. Hardness (CaCO3) for the Big Spring data was calculated using Equation 2. The sum of the calcium (Ca in mg/L) divided by its atomic mass (40.08) and the magnesium (Mg in mg/L) divided by its atomic mass (24.305) was multiplied by the molecular weight of CaCO3 (100.09) to determine hardness. This calculation was performed in Excel and the corresponding data points were added to the scatter plot created to show the relationship between hardness and specific conductance in the Cumberland Valley.  Ca(mg / L) Mg (mg / L)  Hardness as CaCO3   100.09    24.305   40.08 Equation 2: An equation used to calculate total hardness of water (Casiday and Frey, 1998).

12

RESULTS Field Sites The sampling location at Green Spring Creek was on Hwy 641 in Newville, Pennsylvania. The measurement was taken upstream of the bridge at the Green Spring Brethren in Christ Church. The stream bed was predominately clay and silt. Algae and primary producers were common in the water indicating a high nutrient content at this location. The field measurement of SpC at this location was recorded as 702.3 µs/cm (table 2). The SpC value was multiplied by 0.48 to obtain a hardness of 337.10 mg/L as CaCO3. The ANC for this site reported a value of 344.2 mg/L as CaCO3. Dykeman Spring was measured near the pressure transducer previously installed by Shippensburg University, downstream from the culvert on Penn Street. The stream bed at this location was predominately a pebble to cobble grain size with some sand and silt. A SpC value of 373.3 µs/cm was measured in the field at this location. The measured SpC was multiplied by 0.48, and the value of 179.18 mg/L CaCO3 was compared to the ANC value of 179.4 mg/L as CaCO3 determined using the titration results (table 2). The Middle Spring sampling site was located on Stonewall Road, next to the Cumberland Franklin Joint Municipal Authority pumping station #10 in Shippensburg, and downstream from the University. The stream bed consisted of predominately clay and silt sediments at this location. The field SpC reading was 645.3 µs/cm, resulting in a calcium carbonate value of 309.74 after performing the calculation (645.3*0.48). The ANC value of CaCO3 for this sample was reported as 224.6 mg/L (table 2).

13

Table 2: Results of the analysis of three locations chosen within the Cumberland Valley.

Location

SpC (µs/cm)

Hardness

ANC Titration

Calculated CaCO3

Calculated CaCO3

Dykeman Spring

373.3

179.18

179.4

Green Spring

702.3

337.10

344.2

Middle Spring

645.3

309.74

224.6

Conductance and Hardness The hardness and conductivity values have been plotted using an Excel scatterplot in figure 3 to represent the overall trend of groundwater in the Cumberland Valley. The Big Spring data calculated using equation 2 is consistent with the data plotted using equation 1. All three field measurements are plotting near the trendline. Three outliers were identified as having very high SpC and hardness values. Further analysis of the data for these three outliers indicated that the Cl and Na levels are high at all three locations (Becher and Root, 1981; Becher and Taylor, 1982).

14

Figure 3: The water chemistry data from the published dataset shows the overall trend of hardness of groundwater in the Cumberland Valley (blue diamonds). The Big Spring data provided by Penn State is displayed in open red squares and the field sites are represented in triangles. The published datapoints that correspond to the Big Spring dataset and the field samples are represented in diamonds of the same color (see legend) (Becher and Root, 1981; Becher and Taylor, 1982).

DISCUSSION The water hardness values in the published data for both Cumberland and Franklin counties were calculated using the SpC value and the 0.48 multiplier discovered by Becher and Root (1981). The y=mx+b equation generated in Excel after the data were plotted was used to determine the validity of the 0.48 multiplier (from Becher and Root, 1981). The slope of the trendline (m=0.4452) is within a reasonable range to the 0.48 value; therefore, the use of Becher and Root’s (1981) 0.48 multiplier is 15

accepted as valid unless there are elevated concentrations of anions such as chloride (Cl) or sodium (Na) increasing the SpC. The outliers in the data in figure 3 are all due to elevated Cl and Na. The field sites were selected and SpC was measured. The measured value was used to calculate the hardness as CaCO3 using the 0.48 multiplier discussed in Becher and Root (1981). Titrations were performed in an attempt to confirm the validity of using the multiplier for the three field locations. The results of the calculations are compared to the results of the titrations in table 2. The ANC and calculated values for CaCO3 in Dykeman Spring and Green Spring results were within an acceptable range. However, the discrepancy in the Middle Spring results between the ANC value of CaCO3 and the calculated CaCO3 indicates that other ions are present in the sample. The location is likely impacted by road salt from the Borough of Shippensburg and Shippensburg University which are upstream of the sampling location. Further analysis would likely indicate elevated chloride (Cl) or sodium (Na) in Middle Spring, similar to the outliers in the published data. Increases of SpC do not guarantee an increase of hardness of water. The three field sites were plotted with the rest of the data from the Cumberland Valley in figure 3 and were consistent with the trend, as were the Big Spring data. Many of the data points fell on, or near, the trendline and are showing a high positive correlation between SpC and hardness. The Green Spring Creek and Middle Spring Creek points are located in the high conductance range of the scatterplot.

16

Outliers Outliers were identified after all of the data was plotted in the Excel scatterplot. After reviewing the published data, three of the identified outliers were determined to be data entry errors. These entries were corrected and the points plotted on the trendline of the Cumberland Valley data. Two of those data points were apparent because it is highly unlikely that SpC in a carbonate terrain would be less than 10 µs/cm. The corresponding points were no longer outliers, but consistent with the observed trend. Three outliers (figure 3) were investigated and identified as Cu-327, Fr-389, and Fr-499. The Cumberland County well identified as Cu-327 is located in the Rockdale Run Formation. The high conductance and hardness at this sampling location could be due to the dissolution of the pure limestone by the large volume of water flowing through this formation. Pure limestone has a high Ca content which is easily dissolved by the flowing water. The Rockdale Run Formation contains karst features and the limestone is very fine-grained with detrital and skeletal remains. Wells located in the Rockdale Run Formation report a median specific capacity of 12 gal/min/ft and sustained yields that reach 600 gal/min, with a median of 405 gal/min (Becher and Root, 1981). It is likely that subsurface CO2 levels are high due to the karst features, allowing more mineral dissolution to occur even though the residence time of the water is short. Well Fr-389 is located in Franklin County and has the largest conductance and hardness values in the Franklin County dataset. Located in the Elbrook Formation, the formation is Cambrian aged with a lithology ranging from limestone to calcareous shale sandstone. The median specific capacity of the Elbrook is 2.0 gal/min/ft; this value is 17

only surpassed by the Stonehenge Formation in Franklin County. The maximum reported well yield is 250 gal/min (Becher and Taylor, 1982). The well record for Fr389 indicates this well is 90 ft deep and the water bearing zone is in the limestone. The third outlier is Franklin County well Fr-499, located in the Martinsburg Formation. The Martinsburg Formation has a median specific capacity of 0.80 gal/min/ft and a median reported well yield of 20 gal/min (Becher and Taylor, 1982). Specific Capacity and Sustained Yield The specific capacity of a well is useful in estimating the sustained yield, which indicates the amount of water flowing through the aquifer. Specific conductance (SpC) will typically increase as the flowpath of the water increases allowing the water to dissolve more minerals along the way (Becher and Taylor, 1982). Increased residence time of the water will also increase the SpC because more mineral dissolution occurs. The outliers investigated had increased specific capacities and sustained yields, which is inconsistent with the increased SpC measurements. Further investigation indicated elevated Cl and Na concentrations elevated the SpC in those wells. CONCLUSIONS This study supports the relationship between SpC and hardness reported by Becher and Root (1981). The 0.48 multiplier applied to SpC values is consistent unless the water sample has been impacted by an outside source of dissolved ions, such as road salt. The three creeks sampled during this study maintained the relationship stated by Becher and Root (1981), also demonstrating the potential for miscalculation in a sample with high SpC, such as Middle Spring Creek. 18

Total hardness was calculated for the Big Spring dataset and added to the scatterplot. The Big Spring data, collected by Penn State (2006) and provided by Dr. Feeney, plotted with the trend of the published data and field samples, further emphasizing the relationship between SpC and hardness. Three outliers were due to data entry errors. After correcting the data using the published data, the data plotted consistently with the observed trend. Three additional outliers were explored and determined to have elevated Cl and Na concentrations. The increase of SpC due to the Cl and Na made the relationship between SpC and hardness invalid for those three wells. REFERENCES Anne, K. 2011. Magnesium and calcium in drinking water and heart diseases. Geological Survey of Finland. Accessed on Elsevier February 18, 2015. Atkinson, T.C. 1977. Carbon dioxide in the atmosphere of the unsaturated zone: an important control of groundwater hardness in limestones. Journal of Hydrology, 35(1977)111-123. Becher, A., Root, S. 1981. Groundwater and Geology of the Cumberland Valley, Cumberland County, Pennsylvania. Commonwealth of Pennsylvania Bureau of Topographic and Geologic Survey. Water Resources Report 50. Becher, A., Taylor, L. 1982. Groundwater Resources in the Cumberland and Contiguous Valleys of Franklin County, Pennsylvania. Commonwealth of Pennsylvania Bureau of Topographic and Geologic Survey. Water Resources Report 53.

19

Casiday, R., Frey, R. 1998. Water Hardness. Department of Chemistry, Washington University. http://www.chemistry.wustl.edu/~edudev/LabTutorials/Water/ FreshWater/hardness.html. accessed February 17, 2015. Cool Today Resource Library. Water softening vs. reverse osmosis. Accessed February 19, 2015. http://www.cooltoday.com/library/article/water-softening-vs.-reverseosmosis-whats-the-diference. Environmental Protection Agency (EPA). 1986. Quality criteria for water 1986. Office of water regulations and standards. EPA 440/5-86-001. Evens, E., Yanick, S., Osnick, J. 2013. Characterization of hardness in the groundwater of Port-Au-Prince. An overview on the significance of magnesium in the drinking water. Aqua-LAC. vol 5. pp 35-43. Greene, R. T. 2002. Pennsylvania Fish and Boat Commission Bureau of Fisheries, Coldwater Unit-Fisheries Management Division. Big Spring Creek (707B) Fisheries Restoration Plan. Krawczyk, W., Ford, D. 2006. Correlating Specific Conductivity with Total Hardness in Limestone and Dolomite Karst Waters. Wiley Interscience, Earth Surface Processes and Landforms, 31. pp 221-234. Meena, K.S., Gunsaria, R.K., Meena, K., Kumar, N., Meena, P.L. 2012. The problem of hardness in ground water of Deoli Tehsil (Tonk District) Rajasthan. Journal of Current Chemical and Pharmaceutical Sciences:2(1). pp 50-54.ISN 2277-2871.

20

Nerbrand, C., Agreus, L., Lenner, R.A., Nyberg, P., Svardsudd, K. 2003. The influence of calcium and magnesium in drinking water and diet on cardiovascular risk factors in individuals living in hard and soft water areas with differences in cardiovascular mortality. BMC Public Health 2003, 3:21. http://www.biomedcentral.com/147 1-2458/3/21. Rosanoff, A. 2013. The high heart health value of drinking-water magnesium. Elsevier. Medical Hypotheses 81. pp 1063-1065. Skipton, S., Dvorak, B. 2014. Drinking Water Treatment: Water Softening (ion exchange). University of Nebraska-Lincoln Extension, Institute of Agriculture and Natural Resources. G1491. Accessed February 19, 2015. http://ianrpubs.unl.edu/live/g1491/build/g1491.pdf U.S. Geological Survey online alkalinity calculator. 2013. USGS Oregon Water Science Center Alkalinity Calculator. http://or.water.usgs.gov/alk/.

21