Drinking Water Quality in Rural Pennsylvania and the

Drinking Water Quality in Rural Pennsylvaniaand the Effect of Management Practices

This research was sponsored by a grant from the Center for Rural Pennsylvania, a legislative agency of thePennsylvania General Assembly.

The Center for Rural Pennsylvania is a bipartisan, bicameral legislative agency that serves as a resource forrural policy within the Pennsylvania General Assembly. It was created in 1987 under Act 16, the RuralRevitalization Act, to promote and sustain the vitality of Pennsylvania’s rural and small communities.

Information contained in this report does not necessarily reflect the views of individual board members orthe Center for Rural Pennsylvania. For more information, contact the Center for Rural Pennsylvania, 200North Third St., Suite 600, Harrisburg, PA 17101, telephone (717) 787-9555, fax (717) 772-3587, email:[email protected].

Drinking Water Quality in Rural Pennsylvania andthe Effect of Management Practices

ByBryan R. Swistock, M.S., Stephanie Clemens, M.S. and William E. Sharpe, Ph.D.,

School of Forest Resources and Institutes of Energy and the Environment,Pennsylvania State University

January 2009

EXECUTIVE SUMMARY

4 The Center for Rural Pennsylvania

Millions of rural and suburban Pennsylvania residents rely on private wells for drinkingwater, and, each year, 20,000 new wells are drilled. While research has shown that manyprivate wells in the state have failed at least one drinking water standard, Pennsylvania remains

one of the few states without any private well regulations.This study was conducted to better understand the prevalence and

causes of private well contamination and to evaluate the role ofregulatory versus voluntary management of private wells. It set out todetermine if specific indicators, including natural factors, wellconstruction features, and nearby land uses, could be correlated withwater quality parameters in private wells.

To complete the study, the researchers enlisted the help of morethan 170 Master Well Owner Network (MWON) volunteers to collectsamples from 701 private wells statewide.

The findings indicated that bedrock geology was statisticallysignificant in explaining variations in all of the water quality param-eters, with the exception of arsenic. Soil moisture, at the time ofsampling, was the most important factor in explaining the occurrenceof bacteria in wells. Individual well construction features were not

statistically important in explaining well water quality but combinations of well constructioncomponents were statistically significant in explaining both coliform and E. coli bacterialcontamination. No specific land activities were correlated with bacterial contamination, butDNA fingerprinting of E. coli bacteria from wells found that the majority were from animalsources. Nitrate concentrations in wells were statistically correlated with the distance to the nearestcornfield and other crop fields. Lead contamination was found to be largely from metal plumbingcomponents that were exposed to acidic and soft raw groundwater.

About half of the homeowner participants in this study had never had their water testedproperly, which resulted in low awareness of water quality problems. MWON volunteers weregenerally two to three times more likely to know about a health-related pollutant in their well,suggesting that education can greatly improve awareness of problems. Overall, up to 80 percentof the well owners that were shown to have unhealthy drinking water took steps to successfullyavoid the problem within one year after having their water tested.

Results from this study suggest a combination of educational programs for homeowners andnew regulations to overcome the largest barriers to safe drinking water. Regulations are war-ranted to increase mandatory testing of private water wells at the completion of new wellconstruction and before finalization of any real estate transaction. For existing well owners, thisstudy demonstrated the effect education can have to increase the frequency of water testing, theuse of certified labs and awareness of water quality problems.

While this study showed that education increased the use of sanitary well caps on existingwells, most well construction features need to be included at the time the well is drilled.Homeowners having new wells drilled are difficult to reach with educational programs and, asa result, the voluntary approach to encourage proper well construction has largely failed. Giventhe benefits of well construction and the difficulty in reaching the target audience for newwells, statewide regulations requiring well construction components appear to be warranted.

The results of this study do not make a strong case for the need for mandatory wellheadprotection areas around private wells. In most cases, voluntary wellhead protection areasalready existed around private wells in this study. As a result, the data seem to confirm theimportance and success of de facto wellhead protection areas of 50 to 100 feet that alreadyexist around most wells.

Overall, 63 to 78 percent of well owners were supportive of potential regulations targetingwell construction, well location and well driller certification.

Introduction ............................... 5

Goals and Objectives ................ 6

Methodology .............................. 6

Results ........................................ 8

Conclusions .............................. 21

Policy Considerations ............. 21

References ................................ 23

TABLE OF CONTENTS

Drinking Water Quality in Rural Pennsylvania 5

INTRODUCTIONMore than 3 million rural and suburban residents in

Pennsylvania rely on a private well for drinking water,and about 20,000 new wells are drilled each year in thestate1 (U.S. Census Bureau, 1970, 1990). Only Michi-gan has a larger population served by private watersupplies. Unlike residents that use community watersystems, homeowners with private wells are notprotected by any statewide regulations.

Despite many past attempts, (the most recent wereHouse Bill 1591 and Senate Bill 870 in 2001), Pennsyl-vania remains one of the few states where well loca-tion, construction, testing and treatment are the volun-tary responsibility of the homeowner. Some countiesand townships have passed ordinances, and consider-able educational efforts have been made to meet thedemands of private well owners interested in properlymanaging their water supply (Mancl et al., 1989;Swistock et al., 2001).

The voluntary management of private wells is aproblem because most health-related pollutants inwater are symptomless. As a result, homeowners withprivate water supplies may be exposed unknowingly tohealth related pollutants unless they voluntarily havetheir water tested for the correct water quality param-eters.

Several studies have documented the occurrence ofvarious water contaminants in private water systems(Rowe et al., 2007; New Jersey DEP, 2004; Iowa DNR,2004; Liu et al., 2005; Zimmerman et al., 2001). Large-scale national or statewide studies typically report thatabout 15 to 50 percent of private water systems fail atleast one safe drinking water standard. Smaller, re-gional studies often report much lower or highercontamination rates. Only a few of these studies havemade any attempt to determine the causes, such asnatural geology, land use and well construction, ofcontamination of private water systems.

A significant portion of the rural population may beexposed to unhealthy drinking water unless it properlytreats the water or uses bottled water. However, docu-menting the impact of polluted drinking water on thehealth of residents using private water supplies isdifficult because most pollutants require long-termexposure and mimic the effects from other air- or food-borne pollutants. Those that create acute effects, suchas bacteria, have symptoms similar to common viral or

bacterial illnesses. A comprehensive study of 228waterborne illnesses in the 1970’s by Craun (1986)determined that residents with private or small, semi-public wells were most vulnerable to waterborneillnesses. Recent research in Pennsylvania has furtherdocumented the occurrence of disease-causing bacteriain private wells (Lindsey et al., 2002; Swistock et al.,2004).

Contamination of groundwater wells can occur fromboth above and below the surface. Pollution of entiregroundwater aquifers may occur from failing septicsystems, manure and fertilizer applications, mining, orother land uses. Individual water supplies may also becontaminated around the exposed well casing (well-head) from surface water flowing along the well casingand/or from a loose fitting or absent well cap thatallows insects, animals or surface water to directlyenter the well. Thus, the wellhead area, especially inpoorly constructed wells, represents a very sensitivearea that can serve as an open conduit to undergroundaquifers that threaten nearby private and public watersupplies.

Surface contamination can be prevented by extendingthe well casing above the ground surface, installing acement-like grout seal around the casing, and fitting avermin-proof or “sanitary” well cap on top of the well.Extended casings, grout seals and sanitary well capsare required in most states but they are rarely used inPennsylvania because there are no statewide wellconstruction regulations in the state. Recent studies ona small number of wells have demonstrated the impor-tance of a grout seal and sanitary well cap in preventingbacterial contamination but the overall importance ofwellhead versus aquifer-wide contamination remainspoorly understood (Centers for Disease Control, 1998;Zimmerman et al., 2001; Swistock and Sharpe, 2005).Other issues, such as well location relative to sourcesof contamination and land management practices, mayplay an important role in contamination of privatewater systems. For example, a recent study by theprincipal investigators of 50 central Pennsylvaniahomes that were at least 5 years old and had a privatewater supply found that 24 percent had never pumpedtheir septic tank and had experienced higher rates ofwater contamination (Center for Watershed Steward-ship, 2005). This same study also found relationshipsbetween well water quality and well construction, landuse and septic systems.

Homeowners with private wells typically neglectwater supply management unless obvious water quality

1 Based on the average annual increase in drilled wells reportedbetween the 1970 and 1990 U.S. Census. Note: Data on the use ofprivate water systems were not collected during the 2000 U.S.Census.


symptoms occur. Unfortunately, most health-basedpollutants have no obvious tastes or odors in water.Uninformed homeowners may fail to identify danger-ous problems or fall victim to scare tactics used bytreatment vendors and spend thousands of dollars onunnecessary treatment equipment. In fact, a recentsurvey of central Pennsylvania well owners found thatonly 50 percent had ever tested (most just for bacteria)their wells and about 10 percent had purchased unnec-essary water treatment equipment (Center for Water-shed Stewardship, 2005).

Penn State Cooperative Extension has a long historyof educating well owners in Pennsylvania but the largetarget audience has made education a difficult task(Mancl et al., 1989; Swistock et al. 2001). In anattempt to more efficiently educate well owners, theresearchers received federal funding in 2004 to createthe Master Well Owner Network (MWON). Thisprogram provided eight hours of training on properwater well management to over 300 volunteers from 64counties. This program has educated thousands ofprivate well owners in the state resulting in its expan-sion into Virginia, Maryland, Delaware and WestVirginia.

Private wells are pervasive across the landscape ofPennsylvania, serving as important sources of water forrural and suburban homes and farms. However, poorlyconstructed and unmanaged water wells representpotential risks for vital groundwater aquifers and thehomeowners, farmers and businesses that access them.Limited available data suggest that contamination ofthese water supplies is widespread but little is knownabout the magnitude of the problem, the causes ofpollution, and policies or educational efforts that maybest address the problem. The Master Well OwnerNetwork (MWON) volunteers represented a uniquelyefficient opportunity to gather important data aboutprivate water wells throughout the state to betterunderstand the occurrence and sources of contamina-tion and the potential impact of regulation versusvoluntary education.

Goals and ObjectivesThis two-year study, conducted in 2006 and 2007, set

out to:• Determine the occurrence of several health-relatedpollutants in 700 private wells throughout the state.• Determine if well contamination is different be-tween wells with different types of well construction(i.e. buried casing, type of well cap, presence ofgrout, slope around well, etc.).

• Determine how water supply characteristics, such asproximity to polluting activities and nearby septicsystem maintenance, influence groundwater quality.• Determine how frequently private wells are volun-tarily tested and adequately treated for contaminantsand how often homeowners voluntarily followrecommendations to solve well contaminationproblems.

MethodologyA total of 172 MWON volunteers and Cooperative

Extension staff collected samples from 450 privatewells in 2006 and another 251 wells in 2007 for a totalof 701 wells. The approximate locations of the wellsare shown in Figure 1. Each of the MWON volunteersattended eight hours of training on private water systemconstruction, management and testing. Most of theMWON volunteers attended training workshops during2004 or 2005 but some were trained in 2006. Thevolunteers had varied backgrounds including waterresource agency professionals (30 percent), professionsunrelated to water (23 percent), retired (20 percent),educators (7 percent), health care professionals (5percent), local government representatives (5 percent)and farmers (4 percent) (Clemens et al., 2007).

Of the 701 private wells, 79 were owned by existingMWON volunteers and 622 were owned by eithernewly trained volunteers or homeowners. For thepurposes of this report, the 79 MWON volunteer wellswere usually treated differently than the 622 wellsowned by homeowners or newly trained volunteersbecause of their management history. Existing MWONvolunteers would have had at least one year to improvethe management of their private water well based oninformation they learned at their training course.Homeowner or newly trained MWON wells would bemore representative of typical private wells throughoutthe state.

MWON volunteers were recruited to help withsample collection through email and phone calls. Onemonth before each sample collection workshop,interested volunteers were sent sample bottles, surveysand instructions for collecting samples. Approximatelyone week before the respective sample collectionworkshops, the MWON coordinator called eachvolunteer to answer questions and confirm samplecollection criteria and instructions. Volunteers werealso encouraged to contact their MWON regionalcoordinator with any project questions. Coordinatorswere trained in early 2006 on sample and surveyprocedures.


Selection of all study participants was done byMWON volunteers, regional coordinators, CooperativeExtension educators or project staff. To ensure spatialdistribution, volunteers maintained at least a one-miledistance between sampled wells. Volunteers wereencouraged to select private wells that were familiar tothem (friends, family, co-workers, etc.) to facilitatecollection of water samples and water supply informa-tion in a timely and accurate fashion.

Sample and Survey CollectionVolunteers collected water samples from each well on

the morning of the workshop, stored the samples on iceand returned the samples and surveys to the workshoplocation usually no later than noon. The training andexperience of volunteers helped to ensure that sampleswere collected properly and that water supply informa-tion was accurately recorded on the survey.

Two water samples were collected from each home.The first was a sample of “first-draw water” (water thathad been in the plumbing for at least six hours) whichwas tested for lead from corrosion of metal plumbingcomponents. Volunteers collected first-draw samplesfrom the kitchen faucet (78 percent of samples) orother faucets used for drinking water (22 percent ofsamples).

The second sample was analyzed for all other chemi-cal parameters along with coliform and E. coli bacteria.Where only treated water was supplied to the house,samples were collected from the pressure tank or anoutside hose bib. The water was allowed to run for one

to two minutes prior to thesample being collected. Volun-teers reported that 46 percentof the running water samplescame from the kitchen faucet,27 percent was from thepressure tank and 27 percentwas from other taps (mostlyoutside hose bibs).

The survey forms for eachwater supply were divided intoa homeowner and volunteerportion. The volunteers visu-ally inspected the well, plumb-ing, and surrounding landscapeand reported information onwell construction and sourcesof contamination. The home-owner survey included ques-tions about the well age, welldepth, approximate yield, past

water testing, known problems or symptoms, septicsystem characteristics, water treatment equipment, andopinions about the need for private well regulations.The surveys were returned to the drop-off location witheach well sample.

Follow-Up SurveyIn 2007, an additional survey was sent to each of the

450 well owners that participated in the study during2006. The purpose of this survey was to document anyactions taken by the well owner to solve water qualityproblems or better manage their water supply as aresult of participating in the study. This follow-upsurvey was mailed approximately six to 12 monthsafter the wells had been tested in 2006 to givehomeowners the opportunity to take action on theirwell. A reminder postcard was mailed a few weeksafter the survey resulting in an overall return rate of64.2 percent (289 well owners).

Sample AnalysesWater samples were analyzed for the following eight

contaminants that are likely to occur in rural andsuburban areas from activities or surface water con-tamination.

• Total coliform bacteria are “indicators” used todetermine if a pathway exists that might allowdisease-causing bacteria to contaminate the watersupply. E. coli bacteria are a subset of coliformbacteria that only occur in animal or human wastesand indicate more serious contamination.

Figure 1. The approximate locations of the 701 private wells sampled in 2006 and 2007,along with the regional boundaries used for statistical analyses.


• E. coli source tracking was conducted at the E. coliReference Center at Penn State University in anattempt to determine the role of animals versus humansas a source of E. coli contamination in private wells.• pH is a common measure of water quality that isoften related to corrosion of plumbing systemcomponents like lead and copper.• Lead is a dangerous metal with many health effects,especially in children. Lead usually occurs in drink-ing water supplies from corrosion of lead solder orlead impurities in plumbing fixtures.• Nitrate-nitrogen is a health related pollutant thatusually originates from fertilizer, manure or septicsystems.• Arsenic is a relatively new concern in drinkingwater with serious health effects at very low concen-trations. It is thought to most often occur naturallyfrom certain types of rocks but it can also come fromtreated lumber and pesticides.• Triazine pesticides include the most commonpesticides used in Pennsylvania, such as atrazine andsimazine. These pesticides have serious health effects.• Hardness is a measure of the amount of calcium andmagnesium in the water. Although harmless, hardnesscan cause aesthetic effects including stains, damagedhot water heaters and increased soap usage. Hardnesswas measured primarily to determine the appropriateversus unnecessary use of water softener treatmentdevices since softeners are the most common watertreatment device sold in Pennsylvania.

Well Owner NotificationAll participants in the study received their water

quality results within six weeks of their sample submis-sions. Individual water quality results were keptconfidential and were only accessible to project person-nel. All participants with a water system that failed tomeet one or more drinking water standards receivedspecific recommendations and Penn State Fact Sheetsto assist them in taking action to solve their waterquality problem. Homeowners were also given a phonenumber and email address to ask questions about theirresults. During the two-year study, 19 of the 701homeowners called with follow-up questions afterreceiving their water test report.

Statistical AnalysesIn this report, where statistical analyses were not

possible, direct comparison of percentages were usedrather than detailed statistical analyses. Analysis ofcovariance (ANCOVA) and logistic regression modelswere used to determine which well characteristics

(geology, construction, etc.), land uses (distance toseptic, farm fields, etc.) and management activities(septic system pumping frequency, etc.) were importantin explaining the occurrence of the various pollutants.

Data Quality Assurance and Quality ControlThis study analyzed 58 quality control samples,

representing about 8 percent of the 701 private wellsamples analyzed. These samples were submitted to thelaboratory among other samples from private wells tomeasure how accurate, precise and repeatable the waterquality results were from the laboratory.

Overall, the results from the quality control samplesindicated that water quality data collected during thestudy were of excellent quality.

Effect of Location and Sample Method on BacteriaAnother data quality question centered on the loca-

tion and method used to collect water samples forbacteria analysis. Extra care must be taken whensampling water supplies for coliform bacteria and E.coli bacteria because of the risk of contaminating thesample from hands and other surfaces during collectionand subsequent handling. The results of the protocolused in this research suggest that the sampling strate-gies, including washing the bottle three times and notflame sterilizing the end of the faucets, did not causeany systematic bacterial contamination of the samplesfrom the 701 study wells.

ResultsWell Characteristics

The characteristics of the 701 private wells sampledcame from responses on the homeowner and volunteersurveys submitted for each well. Generally, wellowners were very knowledgeable about their wells andmost were able to provide the basic characteristics oftheir well. For example:

• 83 percent were able to estimate the depth of theirwell. The average well depth was 172 feet (mediandepth = 140 feet) with a maximum depth of nearly1,000 feet.• 88 percent knew the approximate year their well wasdrilled. More than 72 percent had been drilled since1970 but only 4 percent had been drilled since 2005.• Well owners knew their well characteristics despitethe fact that very few possessed a copy of their wellcompletion report or “well log.” A copy of the wellcompletion report should be provided to the wellowner to document the well characteristics, such asdepth, construction, geology, and yield.


Water QuantityAbout 51 percent of well owners could estimate the

yield of water coming from their well, with an averageyield of about 18 gallons per minute (gpm). TheMidwest Plan Service recommends a minimum of 6gpm for a home water well (Midwest Plan Service,1992) although most homes will need less than 6 gpm.In this study, nearly 70 percent of the well yields reportedwere greater than 6 gpm. According to well owners, wellyields were not a major problem across the state.

Well ConstructionMWON volunteers evaluated the construction of each

well (Figure 2). Of the 701 wells, 12 percent (84 wells)were completely buried below the ground surface,often in a pit or basement. Nine percent of wells ownedby an MWON volunteer were buried while 13 percentof homeowner wells were buried. This was a commonpractice decades ago to keep water lines from freezing.Of the 85 wells in this study that had a buried casing,most were drilled prior to 1970. Still, there were eightwells with a buried casing that had been drilled after1990. Buried wells were most prevalent in northwestPennsylvania, probably because wells tended to beolder in that region, but there were no statisticallysignificant differences in the occurrence of any wellconstruction features (buried casing, sanitary well cap,grout, etc.) between the six regions of the state.

Of the homeowner wells that were visible above theground, 16 percent had a sealed, sanitary well cap.Sanitary well caps were more than twice as likely onMWON volunteers’ wells, presumably because theyreceived a free sanitary well cap during their MWONtraining. About 9 percent of homeowner wells hadmissing well caps or miscellaneous types of caps(coffee cans, cement, ceramic, etc.). Many of the wellswith miscellaneous caps were very old, hand-dug wellsthat were typically constructed in the early 1900s.Hand-dug wells have a large diameter hole (several feet

wide) lined with stone and usually covered with acement slab cap. Of the 701 wells sampled in thisstudy, 6 percent were hand-dug and the remaining 94percent were drilled.

Volunteers were asked to determine, as best aspossible, if a well was grouted or if the well had avisible cement seal around the casing. Since so fewwells had a well completion report, the presence of aproper grout seal that extended from the surface tobedrock was impossible to determine in most cases.Instead, grout presence had to be based on homeownermemory or visual evidence of cement or grout residualat the surface around the casing by the volunteersampling the well (most cases).

Volunteers reported that grout or a cement sealaround the well casing was present on 18 percent (120wells) of the private wells and this percentage wassimilar between volunteer and homeowner wells. Groutinformation was available from well logs for only 19 ofthese wells. About 5 percent of the homeowner wellshad both a sanitary well cap and evidence of grout,according to volunteers. A follow-up email survey ofwell owners with grouted wells found that only ahandful could confirm for certain that their entire wellcasing was properly grouted at the time the well wasdrilled. Since only eight wells had been drilled afterenactment of county well construction ordinances inChester and Montgomery counties, it seems likely thatthe majority of the wells denoted as “grouted” in thisstudy were only grouted along the first few feet of thecasing below the ground surface instead of the entirelength of the well casing as recommended.

Sloping the ground in the immediate vicinity of thewell casing is recommended to prevent ponding ofsurface water and possible contamination of the well bysurface water. Visual observation by volunteers showedthat 36 percent of homeowner wells had this slopearound the well. A greater percentage of volunteers’wells (54 percent) were sloped, presumably because

volunteers had learned the value of slopingduring their MWON training.

For statistical comparisons, five well con-struction features were combined to providean overall well construction score or“wellscore.” The wellscore was used tocorrelate overall well construction with waterquality parameters in wells. Well constructionfeatures considered in the wellscore included:• Casing – a drilled well with a metal or plasticcasing (casing score = 1) versus a hand-dug wellwithout a casing (casing score = 0).• Buried – the well casing is visible above

Figure 2. Well construction components for homeownerand MWON volunteer wells.


ground (buried score =1) versus a buried well casingnot visible above ground (buried score = 0).• Grout – evidence of grout or cement around the casingat the surface (grout score = 1) versus no evidence ofgrout or cement around the casing (grout score = 0).• Slope – no ground slope around the well headincluding wells in a depression (slope score = 0)versus wells with ground sloping away from thecasing (slope score = 1).• Well cap – a sanitary, sealed, vermin-proof type(well cap score = 1) versus an unsealed, standard ormissing well cap (cap score = 0).The wellscore for each well was the added score of

the five well construction features with scores rangingfrom 0 (very poor construction) to five (very goodconstruction). Of the 701 wells in this study, a break-down of wells in each wellscore category are:

Wellscore 0 = 3 percent (no recommended wellconstruction components);Wellscore 1 = 10 percent (one recommended wellconstruction component);Wellscore 2 = 41 percent (two recommended wellconstruction components);Wellscore 3 = 30 percent (three recommended wellconstruction components);Wellscore 4 = 12 percent (four recommended wellconstruction components); andWellscore 5 = 4 percent (all five recommended wellconstruction components).

Water Treatment EquipmentWater treatment equipment was installed on many of the

private wells, with 53 percent having at least one piece oftreatment equipment. MWON volunteers were onlyslightly more likely to have treatment equipment com-pared to homeowners. Well owners in the northeast regionof the state were significantly less likely to have watertreatment equipment compared to other regions of thestate; presumably because water hardness is much lowerin that region, decreasing the need for water softeners.

The percent of wells that had each type of watertreatment equipment is listed in Figure 3. Water soften-ers were the most common type of water treatmentequipment, occurring in about one-third of homes withprivate wells in this study. Sediment filters were alsoquite common but the majority of sediment filters weresmall, cartridge filters that are installed ahead of larger,expensive units (like softeners, ultraviolet lights, etc.)to remove large sediment particles and prevent wearand tear on other treatment equipment. Only about one-third of the sediment filters appeared to be larger,multilayer (sand and gravel) filters for treating serious

sediment problems. Other types of water treatment,such as oxidizing filters to remove iron, manganeseand/or hydrogen sulfide gas and ultraviolet lights to killbacteria, were far less common, occurring in less than10 percent of the homes with private wells. The aver-age cost of water treatment equipment and installationwas $1,127 with a maximum of $7,000. These valueswere not adjusted for inflation because the year ofinstallation was not measured.

Given that the most common water treatment deviceslisted in Figure 3 are used primarily to treat obviousaesthetic problems in water, it was not surprising thathalf of well owners indicated that they had installedwater treatment equipment because of obvious stains,odors or tastes. The other half of well owners withwater treatment owned it because they had water testreports showing a problem (32 percent) or they inher-ited it from a previous homeowner (20 percent).

Wastewater Treatment CharacteristicsEighty-nine percent of the homes with private wells

used an on-site system for wastewater disposal. Tradi-tional septic tanks and leach fields occurred in 72percent of the homes. The remaining 17 percent usedsand mounds (14 percent), alternative on-lot systems (2percent), or did not know where their wastewater went(1 percent). Of those with on-lot wastewater systems,13 percent reported problems with malfunctions. Septicsystem malfunction rates were similar between stan-dard septic systems, sand mounds and alternative systems,and were also similar between regions of the state.

Figure 3. Occurrence of water treatment equipment onwells in this study.


Septic system maintenance was determined by askinghomeowners about the frequency of pumping theirseptic tank. The generally accepted interval for septictank pumping is every 2 to 4 years, depending onfamily and septic tank size. Septic tanks pumpedinfrequently or not at all may cause groundwater andsurface water contamination from failed leach fields.Of the 625 wells in this study with an on-lot septicsystem and septic tank, 28 percent were never pumped,33 percent were pumped at an interval greater than 4years and 39 percent were pumped at least every 3years. Septic system age could not be used to explainthe number that had not been pumped since only 2percent were less than 4 years old.

Prevalence and Spatial Occurrence of Well Contamination

Of the parameters that were tested on each well watersupply, six had primary drinking water standards (i.e.detrimental health effects are possible if standards arenot met) including coliform bacteria, E. coli bacteria,lead, nitrate, arsenic and triazine pesticides. Hardnessdoes not have a drinking water standard and pH has asecondary or recommended drinking water standard(for aesthetic effects on taste and corrosion).

Contamination rates in raw well water were similarbetween MWON volunteer wells and homeowner wellsso they have been lumped together for reporting here.Overall, approximately 41 percent of the wells testedfailed to meet at least one of the health-based drinkingwater standards. Of these wells, most (89 percent)failed only one of the drinking water standards.

Keep in mind that these results apply to raw waterand do not include the effect of water treatment devicesor bottled water that some well owners used to avoidhealth-related pollutants. In fact, roughly 25 percent ofwell owners with a health-related pollutant was avoid-ing it by using proper water treatment or by drinkingonly bottled water. Actions taken by wells owners, bothbefore this study and as a result of this study, arediscussed later. Details of individual contaminants arediscussed below. For the regional analyses, the totalnumber of wells sampled in each region were: north-west (61 wells), northcentral (115 wells), northeast(167 wells), southwest (98 wells), southcentral (159wells), and southeast (101 wells).

Total Coliform (TC) BacteriaTC bacteria were found in 33 percent of the sampled

wells and were absent in 67 percent. This contamina-tion rate is similar to past national and statewidesurveys of private water systems.

TC bacteria occurred throughout the state but therewere some regional differences. The highest incidenceof TC occurred in the southeast and southwest regionswhile the lowest incidence was observed in the north-west and northeast regions. Statistically, the southeastregion had a significantly higher occurrence of TCbacteria than the northwest or northeast regions. Theseregional trends are consistent with TC contaminationreported by Sharpe et al. (1985).

E. coli BacteriaE. coli (EC) bacteria should be absent from drinking

water for the water to be safe to drink. EC bacteriarepresent a more serious contamination issue thancoliform bacteria since EC bacteria can only originatefrom human or animal waste. In this study, EC bacteriawere detected in 14 percent of the private wells. Noother statewide surveys have been done to documentthe occurrence of EC in private wells. The incidence ofEC bacteria found in this study is greater than reportedby some regional studies (Durlin and Schaffstall, 2001;Zimmerman et al., 2001) but less than values reportedby others (Bickford et al., 1996). EC bacteria showedsimilar regional trends to those found for TC bacteria.

A total of 213 distinct EC bacteria colonies from 79different private wells were analyzed by the E. coliReference Center at Penn State. The DNA fingerprintsof each of these bacteria were statistically compared tothe library of EC bacteria from known animal andhuman sources in an attempt to determine the relativerole of animal versus human wastes in causing ECcontamination of private wells. The results showed thatmost EC colonies were more closely related to animalsources than human sources. This suggests that mostcontamination occurs by surface water from nearbyanimal-related land activites.

pH and LeadThe pH of each sample was determined as a general

indicator of the corrosion potential to correlate to leadcontamination. In general, low pH water (below 7.0)tends to be “corrosive,” which means it can dissolvemetals, such as copper and lead from pipes, solder orfixtures, from the plumbing system. Lead was measured infirst-draw water samples from 251 wells during 2007.

Wells with a pH below the recommended level of 6.5occurred in 18 percent of the wells tested and weremost frequently found in the southeastern region. Wellswith a high pH (above the recommended level of 8.5)were rare (2 percent) and occurred sporadically acrossall regions.


While only 18 percent of private wells were belowthe recommended pH standard of 6.5, it is likely that amuch higher percentage of wells contained “corrosivewater.” Many other water quality parameters, includinghardness and total dissolved solids, are important indetermining the corrosion potential of water. This studydid not test for all of the parameters necessary tocalculate a corrosion index (known as the LangelierSaturation Index or LSI). Sharpe et al. (1985) foundthat, while only 18 percent of private wells were tooacidic (pH < 6.5), a much higher percentage (59percent) were actually corrosive to metal plumbing. Inthat study, corrosive water was the most common waterquality found in Pennsylvania, especially in northcentralPennsylvania where well water tended to be both acidicand soft (low hardness and total dissolved solids).

Elevated lead levels occurred in first-draw water from12 percent of the wells that were tested during 2007. Itappears that corrosion of lead from plumbing compo-nents (rather than lead from groundwater pollution)was the predominant source of lead in these well watersupplies. Ninety-three percent of the wells with highlead levels had acidic water (pH below 7.0) and 80percent also had soft water. As mentioned above, thiscombination of acidic, soft water causes corrosion ofmetals. Lead contamination was most likely in thesouthcentral (17 percent) and southeast (16 percent)regions and least likely in the northeast region (5percent). These regional variations in lead are identicalto those reported by Swistock et al. (1993).

The 12 percent failure rate for lead is much lowerthan the 19 percent reported in a study by Swistock etal. (1993). It is presumed that this reduction is a resultof the 1991 Federal Lead and Copper Rule that re-quired the use of lead-free solder and fixtures in homeplumbing. Seventy percent of the homes with high leadlevels had plumbing systems that were installed prior toenactment of this rule and most also had copperplumbing systems. There was only one private wellwith a high lead level that could not be clearly linked tocorrosion of metal plumbing (i.e. a new home withplastic plumbing and alkaline water).

NitrateNitrate-nitrogen occurred above the drinking water

standard of 10 mg/L in only 2 percent of the privatewells. Nitrate concentrations in private wells variedstrongly between regions. Mean nitrate concentrationswere significantly higher in the southeast andsouthcentral regions compared to the other four re-gions. Still, sporadic nitrate concentrations above 10mg/L were found in the central and northeast regions.

The 2 percent of wells that exceeded 10 mg/L ofnitrate-nitrogen was far below the 14 percent reportedby Sharpe et al. (1985) and the 9 percent reported bySwistock et al. (1993). Both of these past studies had alarger proportion of samples from southeast Pennsylva-nia than this study and both also reported a strongregional influence on nitrate occurrence. Still, thesecomparisons suggest that groundwater nitrates arelower than they have been historically, perhaps due tobetter management of nitrogen on farms and homefertilizers resulting from education and mandatednutrient management plans. Data reported in the 2002Census of Agriculture (the most recent census avail-able) by the U.S. Department of Agriculture (2004)suggest that applications of nitrogen by fertilizer andmanure have dropped in southern Pennsylvania sincethe early 1990s.

ArsenicOnly 2 percent of the wells exceeded the health-

based drinking water standard of 10 mg/L for arsenic.The maximum concentration observed was 35 mg/Lbut the majority of wells (89 percent) had arsenicconcentrations below 6 mg/L. Wells with high arsenicoccurred mostly in northern Pennsylvania regions,presumably due to the geology of these areas. The threenorthern regions of the state had significantly higherarsenic concentrations than the southern regions withthe highest occurring in the northwest region. Theseresults are similar to results reported by the U.S.Geological Survey (2000) for 578 private wells thatwere sampled in southeast and extreme western Penn-sylvania. Arsenic is thought to originate primarily fromnatural geologic sources, thus, it would not be expectedto vary significantly over time.

Triazine PesticidesOnly three wells (less than 1 percent) had unsafe

concentrations of triazine pesticides above 3 mg/L.Two of these wells were located in southcentral Penn-sylvania and one well was located in the northeastregion. Of these three wells, two homeowners werewilling to have their well re-tested in 2007 using moresophisticated analyses. Those re-tests occurred in late2007 and they did not detect pesticides in either well.

The only other survey of pesticides in private wells inPennsylvania was conducted during 1993 in an unpub-lished study by Penn State University. Of 189 privatewells sampled in that study, none had atrazine orsimazine concentrations above the drinking waterstandard. Another recent study by Bartholomay et al.(2007) tested several hundred wells in Pennsylvania


and found none that exceeded thedrinking water standard for atrazine orits breakdown products. It should benoted that these past studies of pesti-cides in Pennsylvania have generallydetected small concentrations of pesti-cides (above the detection level butbelow the drinking water standard).This study’s testing method onlyprovided a result of “present” or “ab-sent” at a level above the drinking waterstandard. So, while the sampling showsthat very few wells were above thedrinking water standard, the results donot allow an estimation of the numberof wells that may have triazine pesti-cides present at lower detectableconcentrations.

HardnessThere are no health effects or drinking

water standards for hardness but hardwater can cause numerous aestheticproblems, especially when water isheated. Because hardness reducescorrosion of household plumbing, alevel of 90 to 100 mg/L is often consid-ered optimum to reduce corrosion whilealso preventing unwanted aestheticeffects. Total hardness is usually reportedin one of four categories as follows: softwater has a hardness concentration of 0 to60 mg/L; moderately hard water has ahardness concentration of 61 to 120 mg/L; hard waterhas a hardness concentration of 121 to 180 mg/L; andvery hard water has a hardness concentration greaterthan 180 mg/L.

There were clear regional differences in hardnessconcentrations across the state. Most wells with veryhard water were located in western counties or incentral and southcentral Pennsylvania. Wells in thenorthcentral and northeast regions had significantlylower hardness concentrations (means = 70 to 90 mg/L) compared to the other four regions (means = 120 to140 mg/L) due to the bedrock geology of theseregions.

Variables Controlling Contaminationof Private Wells

Contamination of private wells can occur throughthe interaction of both natural and human causes.Leaching of arsenic from bedrock is an example of a

Figure 4. Variables used in ANCOVA statistical models to determine causesof contamination of private wells.

natural source while leaching of bacteria from a septicsystem is an example of a human cause. Figure 4 showsall of the variables that were included in the ANCOVAstatistical model in an attempt to explain sources ofcontamination in private wells. The ANCOVA modelwas then able to determine which variables were statisti-cally significant in explaining water quality differences.

Climate and Other Natural FactorsRegional differences were previously reported in the

summary of each water quality parameter but much ofthis variation can be attributed to geologic differencesbetween the regions. In the ANCOVA models, geologywas statistically significant in explaining the variation inall of the water quality parameters with the exception ofarsenic. In most cases, water quality was strongly tied togeology. The geologic controls on water quality are notsurprising considering that groundwater in private wellsis in direct contact with the bedrock for extended periods


of time. However, some of the geologic differences areprobably the result of land uses that are predominant oncertain types of bedrock. For example, higher nitratelevels on carbonate and igneous bedrock are likely dueto the fact that these bedrock types are predominant inthe regions of the state with intensive agricultural landuse rather than actual differences in the bedrock chemistry.

The carbonate rock type produced the most uniquewater quality with significantly higher bacteria levels,pH, nitrate and hardness compared to most other rocktypes. A recent study of private wells in southcentralPennsylvania also reported a higher incidence ofcoliform and E. coli bacteria in wells located in carbon-ate bedrock because of the close connection betweencarbonate aquifers and surface water (Zimmerman etal., 2001). Igneous rock, located in parts of southeastand southcentral Pennsylvania, was more acidic (lowerpH) and lower in hardness compared to other rocktypes. This resulted in generally higher lead concentra-tions although this difference was not statisticallysignificant. Sedimentary and sandstone/shale bedrocktypes, which are both comprised of various types ofsandstone and shale, predictably produced nearlyidentical water quality results. These results suggestthat the natural geology where a private well is drilledplays an important role in the resulting water quality,regardless of well construction and nearby land useimpacts. The overall importance of geology on manywater quality parameters has also been observed intesting of 5,000 private wells in New Jersey (NewJersey DEP, 2004).

Climatic conditions, such as precipitation andtemperature, are thought to be important when testingpollutants in groundwater wells although limitedresearch data are available on this subject. Contami-nants originating from the ground surface, like coliformbacteria and E. coli bacteria, would be expected to bemore prevalent during wet and warm weather sincethese conditions favor the growth of bacteria and wetweather promotes the movement of surface water andsurface contaminants into the ground.

Of the wells that contained coliform bacteria or E.coli bacteria, 84 percent were tested during moistconditions, while only 16 percent were tested duringdry weather. These results agree with those of Swistockand Sharpe (2005) who found that bacterial contamina-tion of private wells in Pennsylvania was greatlyreduced during an extreme drought in 2002. Short-termmoisture conditions were not statistically important inexplaining concentrations of nitrate, lead, pH, hardnessor arsenic in private wells.

The strong relationship between bacteria levels and

moisture conditions has many implications. Results ofbacteria testing will likely be more variable than otherwater quality parameters, depending on the weather atthe time of sampling. As such, the bacteria resultspresented here must be considered a snapshot ofconditions that apply to the weather conditions thatexisted during the study. Conditions during 2006 and2007 throughout Pennsylvania encompassed a range ofshort-term conditions from mild drought (summer2007) to moderate wetness (fall 2006) but, overall,these two years were near climatic norms. Results fromthe homeowner survey and re-sampling of wellsbetween 2006 and 2007 also elude to the variability ofbacteria results. Twenty-six of the 701 well owners (4percent) indicated they had coliform bacteria in theirwater well but this testing found none present. Also,some wells that contained small numbers of coliformbacteria in 2006 did not contain these bacteria whenthey were re-tested in 2007. The variability of bacteriaresults related to weather conditions must be consid-ered when making recommendations to well ownersabout well testing.

Season of the year was also tested as a potentialcause of water quality variations in wells. One mightexpect that warmer and wetter conditions during springand summer would result in a greater likelihood ofbacteria, nitrate, lead and pesticides in wells whilegenerally drier and cooler conditions during fall mightcause higher pH, hardness and arsenic as dilution fromrainwater is reduced. In the ANCOVA models, seasonwas not statistically important for any of the waterquality parameters. Bacterial contamination was nearlyconstant among the three seasons (spring, summer andfall) that wells were tested.

Effect of Well ConstructionWell construction features could play a role in the

entry of some pollutants, especially those generatednear the land surface, into private wells. For example,bacterial contamination of wells may occur throughimproper construction practices that allow surfacewater, insects or small mammals into the well. Sanitarywell caps, grout seals and sloped ground near thecasing are all used to prevent this wellhead contamina-tion. Five separate well construction components(casing present, casing above ground, sanitary cap,grout or cement seal, slope around casing) along withthe overall “wellscore” (the number of proper wellconstruction components on the well) were included inthe ANCOVA models for each water quality parameter.Additional well characteristics, including depth andage, were included in the models.


In the case of coliform and E. coli bacteria, indi-vidual well construction components resulted inslightly reduced contamination rates (for example,wells with sanitary well caps had slightly lower bacte-rial contamination rates than those with standard ormissing well caps) but none of these individual compo-nents produced statistically significant results. But,combinations of well construction features were highlysignificant in reducing total coliform bacteria and E.coli bacteria. Figure 5 illustrates the effect of increas-ing numbers of well construction features on bacterialcontamination. Note that wells with very poor con-struction (zero features) were twice as likely to havecoliform bacteria and five times more likely to have E.coli bacteria compared to wells with excellent con-struction (five features). The contamination rates forwells with a wellscore=5 (28 percent for coliform and 8percent for E. coli) are very similar to data publishedby Swistock et al. (2005) from sampling of 24 wellswith excellent construction in eastern Pennsylvania.Other studies have demonstrated slightly reducedincidence of bacteria from grout (Zimmerman et al.2001), sanitary well caps (Swistock et al., 2005), andcased wells versus hand-dug (Sharpe et al. 1985) but noother studies have shown a clear connection betweenoverall well construction and bacterial contamination.These results suggest that no single well constructionfeature is critical to preventing surface water contamina-tion but, clearly, combinations of features have a signifi-cant effect in preventing bacterial contamination ofprivate wells.

Despite the importance of well construction tobacterial contamination, it was not important in ex-

plaining other water quality problems in private wells.In fact, no single well construction feature, or even theoverall wellscore, was statistically significant in anyother ANCOVA model. In the case of nitrate, the age ofthe well was statistically important and well depth wasmodestly significant. These factors combined suggestthat older wells, which are typically shallower, areperhaps allowing shallow, nitrate-rich water to enterdeeper groundwater aquifers. Other than nitrate, welldepth was not statistically significant in explaining anyof the water quality parameters. The overall lack ofimportance of well characteristics on parametersbesides bacteria is not surprising given the importanceof geology described earlier.

Effect of Nearby ActivitiesThe sample collection survey completed by each

MWON volunteer included estimated distances to 14nearby activities (crop field, dog kennels, septicsystem, etc.) that could influence some water qualityparameters in wells. Each of these distances wasincluded in the ANCOVA models along with wellconstruction, climate, and geology. Overall, few ofthese activities were statistically significant in explain-ing water quality in private wells.

The overall factors that control each well waterquality parameter (natural factors, well constructionfactors, and land activities) are summarized in Figure 6on Page 16. These results do not argue strongly formandated or voluntary wellhead protection areas aroundprivate wells where various activities are prohibited toprotect well water quality. However, these resultsshould be interpreted with some caution. The over-

whelming majority of theprivate wells in this studyalready, in effect, had a well-head protection area of at least50 feet. In fact, more than 70percent of the study wells werelocated at least 50 feet from allof the possible polluting activi-ties. Thus, there were relativelyfew wells that could be used totest the effect of each of theseactivities. These results,therefore, are probably indica-tive of the overall effectivenessof a minimum 50-foot well-head protection area that mosthomeowners and/or welldrillers seem to provide eitherby chance or as common sense.

Figure 5. The effect of increasing numbers of well construction features oncoliform and E. coli bacterial contamination.


Voluntary Water Well ManagementIn the absence of statewide regulations for well

construction and/or maintenance, education plays animportant role in promoting proper management ofprivate water supplies. However, while education cancreate awareness of problems and management strate-gies, it can only be successful if well owners translatethis knowledge into action, such as by testing their wellwater or installing treatment systems, when problemsare identified. If it is found that well owners aregenerally maintaining their wells properly, then it canbe argued that regulations are less necessary to protectthe health of well owners.

Water TestingMost agencies recommend having private wells

tested by a state-certified laboratory annually forcoliform bacteria and every few years for other con-taminants. The 622 homeowners that participated inthis study had rarely followed these recommendations.In fact, three-fourths of homeowners had either testedtheir water quality just once (44 percent) or had neverhad their drinking water tested (30 percent). It isimportant to note that these estimates may be highsince most of these well owners were familiar withMWON volunteers who may have previously encour-aged them to have their water tested. Trained MWONvolunteers themselves were more likely to have theirwater tested more frequently than uneducatedhomeowners but some had still not had their watertested.

Of the 74 percent of homeowners that indicated their

water had been tested at least once in the past, onlyabout half had used a state-certified laboratory (Penn-sylvania Department of Environmental Protection labor state-certified commercial lab). The remainder wastested by either water treatment vendors or the use ofhome water testing kits. MWON volunteers, on theother hand, were much more likely (69 percent) to usecertified water labs. Overall, about half of the home-owner participants in this study had never had theirwater tested properly (either never tested or onlytesting was a non-certified lab) while fewer MWONvolunteers (30 percent) had failed to have their watertested properly.

The frequency of water testing did vary somewhat byregions of the state. Private well owners in the south-east region of the state were significantly more likely tohave their well tested and also to use a certified testinglaboratory compared to other regions of the state. Thereasons for this difference are not clear but could berelated to the better availability of water laboratories orgreater concerns of threats to private water systems inthis region.

Water Quality AwarenessFor the purposes of this discussion, the term “aware-

ness” refers to a well owner’s knowledge of his/herwater well quality prior to participating in this study.Given the low percentage of homeowners that hadpreviously tested their well water quality, it would beexpected that few were aware of the water qualityproblems in their own well. This was generally true forhealth-based pollutants that have no obvious symptom(taste, odor, etc.) in water. Figure 7 illustrates thepercent of homeowners and MWON volunteers thatwere already aware of each pollutant that existed intheir well before participating in this study.

In Figure 7, the 11 percent awareness value forhomeowners with coliform bacteria means that, of the203 homeowner wells that were found to containcoliform bacteria, only 11 percent of those well ownerswere already aware of this contamination prior toparticipating in this study. In all cases, awareness ofcontamination among homeowners was below 31percent, with the highest awareness occurring forparameters that are more often routinely tested bywater laboratories (nitrate, pH, bacteria). MWONvolunteers were generally two to three times morelikely to know about a health-related pollutant in theirwell suggesting that education can greatly improveawareness of problems.

Naturally, homeowners are more aware of waterpollutants that only affect taste, or cause stains or odors

Figure 6. A summary of statistically significant variables(p<0.05) in ANCOVA models for each water quality parameter.

1Wellscore = number of recommended well construction features (0 to5). 2 Well Depth p-value was 0.08 indicating that the Well Agecorrelation is probably related to a combination of age and depth (i.e.older wells were shallower). 3 Lead is highly correlated with plumbingsystem characteristics and water pH (an index to corrosion) which isrelated to geology.


because the effects are obvious. For example, nearly 50percent of homeowners with very hard water identifiedit as a problem with their water supply. Iron was alsoreported as a problem by 21 percent of the studyparticipants. Although this study did not measure ironto confirm it’s presence, the 21 percent that reportedhaving it in their water agrees with the percentagefound to actually have high iron by Sharpe et al. (1985)suggesting a high awareness of this problem.

Solving Water Quality ProblemsThe goal of any drinking water education program is

to facilitate actions taken by the audience tocorrect water quality problems and, ultimately,to increase the percentage of participants thatavoid unsafe drinking water. Actions taken bywell owners to solve problems could includeusing a new water source (bottled water),maintaining the water supply to remove con-taminants (shock chlorinating a well to killbacteria), removing a source of contamination(moving a dog kennel away from a well), orinstalling a continuous water treatment deviceto remove the pollutant. In this study, therewere several measures that provided someevidence of the willingness of well owners totake actions to properly manage their well andthe effectiveness of these actions given someawareness of the problems and potentialsolutions. Actions that homeowners could taketo solve water quality problems were providedin Penn State Cooperative Extension fact sheets

that were enclosed with eachwater test report.

One measure of the successof education is to compareactions taken by MWONvolunteers who had attendedan eight-hour training work-shop with homeowners whodid not receive this training.MWON volunteers weremuch more likely to have asanitary well cap on theirwell and the ground slopingaway from the well casing inall directions to preventsurface water contamination.Both of these actions werestressed during their volun-teer training workshop and, infact, those completing the

training were given a sanitary well cap for their homewell. Additional information on actions taken and theireffect on drinking water quality were measured throughthe follow-up survey and re-testing of a subset of wells.

Follow-Up Survey ResultsMeasurement of actions taken by homeowners to

solve or treat water quality problems was accomplishedthrough a follow-up survey of the 450 well owners thatparticipated in the study in 2006. A summary of thenearly 300 responses to this follow-up survey is shownin Figure 8. About half of the respondents took an

Figure 8. Percent of private well owners that took various actions toimprove their private well as a result of having their well tested(based on the follow-up survey sent to the 450 well owners that

participated in 2006).

1 Includes shock chlorination, sanitary well cap installation, extending wellcasing above ground, sloping ground to prevent contamination, or groutingaround existing well casing. 2 Includes reduced chemical and fertilizer use,removing animals from near well, diverting runoff or maintaining septicsystem. 3Some homeowners took more than one action.

Figure 7. Percent of homeowners versus MWON volunteers who were aware that eachpollutant occurred in their well prior to the study. “N/A” indicates that no MWON

volunteers had wells that tested positive for pesticides.


action to better manage their private well as a result ofparticipating in this study. Homeowners with wells thatfailed at least one drinking water standard were morethan twice as likely to take corrective actions on theirwater supply compared to homeowners that had nowater quality problems. Overall, an impressive 76percent of homeowners with wells that failed at leastone health-based drinking water standard took at leastone action (some took numerous actions) to correct orbetter manage the problem.

Effect of Actions on Bacterial ContaminationSixty wells that tested positive for coliform bacteria

in 2006 were re-sampled in 2007; 32 homeowners hadtaken at least one action, such as shock chlorination,installing a sanitary well cap or removing the source ofcontamination, to solve their wells’ bacteria problem.

While many homeowners took recommended actionsto remove bacteria from their drinking water well, ofgreater interest is how successful these actions were ineliminating bacteria from these wells. Of the 32 wellswhere homeowners took actions, 21 still containedcoliform bacteria representing a 35 percent success rate.

Given the variability of bacteria in wells in responseto moisture conditions discussed earlier, it could beexpected that some wells with bacteria would testnegative one year later simply due to changing weatherconditions or other factors. Of the 28 wells that con-tained coliform bacteria in 2006 where homeownersdid not take action to eliminate bacteria, 21 stillcontained coliform bacteria when re-tested in 2007.Thus, even in wells without any actions taken, 25percent did not contain bacteria one year later. Anotable difference between these wells and the wellswhere actions were successful in eliminating bacteriawas the bacteria concentrations. Wells that tested

negative for bacteria without any actionsby the homeowner had very low pre-existing bacteria concentrations (average= 3 colonies per 100 mL).

The re-sampling of wells withcoliform bacteria between 2006 and2007 results in a very conservativeestimate of the overall success of actions(shock disinfection, sanitary well caps,etc.) in removing bacteria from wells (10percent) as the difference in successrates between those that took actions (35percent) and those that did not takeactions (25 percent). A similar successrate of 10 percent was found for removalof E. coli bacteria from contaminated

wells in this study. The 10 percent success rate shownin the retesting is similar to the 15 percent successreported by Swistock and Sharpe (2005) in a study of17 wells that were shock chlorinated and fitted withsanitary well caps.

The survey data compiled during the overall studyshowed that well owners who were aware of a health-based pollutant in their water were very likely to avoidthe problem. For example, over 75 percent of the wellowners that knew they had E. coli bacteria had alreadyinstalled a disinfection treatment system or were usingbottled water. The barrier to avoidance of contaminatedwell water on a large scale was the large percentage ofwell owners that were not aware of problems due toinadequate water testing.

The well owners that participated in this study weremade aware of problems that occurred in their watersupply. But, would this awareness translate into actionsto avoid exposure to health-based pollutants? The datapresented in Figure 9 illustrate the overall impact ofthis well testing study on reduced exposure ofhomeowners to health-based water pollutants. The firstcolumn shows the percent of well owners that had eachcontaminant and were avoiding the problem prior tothis study. Keep in mind that many of the well ownersthat were avoiding contaminants prior to this studywere only doing so inadvertently. For example, wellowners using only bottled water because their wellwater tasted bad would also be avoiding coliformbacteria even if they did not know it occurred in theirwell. The second column shows the percent of wellowners that started or improved on a method to avoidexposure to each pollutant as a result of this study. Forexample, some well owners that were originally usingbottled water installed water treatment to improvewater quality after receiving their water test report. The

Figure 9. The overall effect of actions in reducing exposure to watercontaminants in this study.

1 Includes use of bottled water and/or installation of proper treatment equipment.Results for coliform bacteria and E. coli include an additional 10 percent removalbased on the effectiveness of other actions (shock chlorination, sanitary well cap,etc.) estimated from the re-sampling of 60 wells. 2 Note that the total avoiding acontaminant (column three) does not sum from columns one and two because somewell owners simply improved on actions that were already having an effect.


final column shows the percent of homeowners thatwere avoiding the problem six to 12 months after thisstudy. Note that the percent avoiding water qualityproblems increased dramatically in each case. Pollut-ants with more severe or better documented impacts onhuman health, like lead, arsenic and E. coli bacteria,had the highest avoidance rates at the end of this study.Pollutants like coliform bacteria and nitrate, whichhave relatively low risk or only affect certain portionsof the population, had slightly lower avoidance rates.Since relative risk information was given to each wellowner in the Penn State Cooperative Extension factsheets included with each water test report, it is notsurprising that well owners responded with greateractions for pollutants with greater risks. For pollutantsthat were not health-related, like hardness, the percentthat chose to avoid the problem generally dropped tonear or below 50 percent (data not shown).

It is important to note that many well owners tookactions in an attempt to avoid exposure to pollutantsthat could not be clearly linked to a reduced exposureto that pollutant. For example, while only 50 percent ofwell owners with coliform bacteria were successfullyavoiding bacterial contamination by the end of thisstudy, many of the remaining 50 percent had at leasttaken some action to try to avoid the problem. As aresult, the percent of homeowners that took no action atall was generally quite low ranging from 8 percent forE. coli bacteria to 20 percent for various other pollut-ants. In most cases, the homeowners that did not takeany action may have made this choice because of lowerperceived risk. For example, those that did not takeaction to remove coliform bacteria nearly always hadlow bacteria counts less than 10 colonies per 100 mLand those that did not take action to remove E. colialways had just 1 colony per 100 mL of water.

Unnecessary Water TreatmentThe principal investigators have worked with private

well owners in Pennsylvania for several decades andoften encounter homeowners that have been soldunnecessary water treatment equipment. Data collectedduring this study provides some insight about the useof water treatment equipment in Pennsylvania.

Of the 701 homes participating in this study, 288employed a total of 372 water treatment systems. Ofthese systems, 18 percent were considered unnecessarybased on water testing conducted during the study andinformation provided by homeowners. This studydemonstrated that 58 homes, or 8 percent, had watertreatment equipment that was apparently not needed.This figure agrees with recent data gathered in central

Pennsylvania (CWS, 2005). Information provided byhomeowners suggests they are often unaware of theirspecific needs with regard to water treatment. Specificexamples of questionable water treatment equipmentinstallations are described below:

• 46 homeowners with “soft” raw well water had awater softener installed. Nineteen of those homes hadzero hardness, indicating that the water sample wastaken post-treatment. Iron was a problem for anadditional nine homes, which may have been thereason for using this type of water treatment equip-ment. Of the remaining 18 homes, three indicated theequipment existed when they purchased the home,nine indicated it was installed because of past watertest results, and six listed various reasons that did notwarrant this type of treatment equipment.• 50 homeowners had an ultraviolet light installed ontheir water system to kill bacteria but 48 percent (24systems) showed no bacteria present in their water.Of the 24 homes that did not have bacteria present,seven had the equipment because it already existed intheir home when it was purchased, seven had itbecause of water testing they did in the past, and theremaining 10 homeowners had various reasons thatdid not justify using an ultraviolet light.• 16 homes treated their water with chlorinationsystems. Most of these homeowners had a bacteriaproblem or indicated their water was high in iron orhydrogen sulfide (which can both be treated withchlorine). The remaining three homeowners did notneed chlorine although one thought he had a bacteriaproblem. Of the three homes with unnecessarychlorination systems, two installed the equipment dueto the results of previous water testing and oneinstalled it due to taste or odor problems.• Another 20 water treatment systems were found tobe unnecessary based on the problems reported bythe homeowner or by the water testing done in thisstudy. Equipment that appeared to be unnecessaryincluded oxidizing filters, nitrate removal systems,acid neutralizing filters and magnetic treatmentdevices.

Well Owner OpinionsA final portion of the survey provided to each well

owner included opinion questions about perceivedthreats to their water supply, opinions of well regula-tions, and use of other sources of drinking water.

Most well owners were very satisfied with theirprivate water well. Eighty-four percent were notwilling to pay even $10 per month to have access topublic water. A total of 16 percent were willing to pay


some monthly fee to havepublic water including $10 (2percent), $20 (6 percent), and$30 or more (8 percent).

Well owners were generallyconcerned about their waterquality but less so about theirwater quantity. Sixty-fourpercent were very or some-what concerned about waterquality in their well whileonly 39 percent were similarlyconcerned about the amountof water their well supplies.Well owners were mostconcerned about new housingdevelopments as a threat totheir water supply. Thirty-five percent of well ownersranked new housing developments as the biggest threat.Agriculture (16 percent), oil and gas drilling (13percent), mining (11 percent) and highways (4 percent)were the other common responses for threats to privatewells. There were obvious regional differences in theseresponses with agricultural concerns centered insouthcentral Pennsylvania, mining concerns mostly inwestern Pennsylvania and oil and gas concerns mostlyin northern and western counties.

Fourteen percent of well owners use bottled water astheir sole source of drinking water. Another 30 percentoccasionally drink bottled water at home. Bottled wateruse was significantly higher in the northcentral regionof the state and significantly lower in the southcentralregion.

Nine percent of well owners use roadside springs fortheir drinking water; 8 percent occasionally and 1percent for all of their drinking water. Roadside springuse was strongly regional with more well owners in thenorthcentral and northwest regions using them com-pared to other regions.

Four questions on the survey asked well owners theiropinions about proposed private well regulations inPennsylvania, as follows:

• (Well Location) There should be a statewideregulation on the location of new water wells thatincludes the minimum distance a new well can bedrilled from existing wells, septic systems and othersources of contamination.• (Well Construction) There should be a statewideregulation that requires certain well constructioncomponents on new wells (such as a sealed well capand cement-like seal or grout around the casing) toprevent well contamination even if it adds $500 to

$1,500 to the cost of a new well.• (Well Testing) There should be a statewide regula-tion requiring private water wells to be tested forcertain pollutants prior to the sale of the home to anew owner even if it adds several hundred dollars tothe cost of a home sale.• (Driller Certification) There should be a statewideregulation requiring that professional water welldrillers be periodically certified by passing a compe-tency exam.A summary of opinions from homeowners is given in

Figure 10. MWON volunteers are not included in thissummary since they were nearly unanimous in theirsupport of these measures as a result of their volunteertraining.

Well owners were generally supportive of all poten-tial well regulations with 63 to 78 percent strongly orsomewhat agreeing with the statements above. Thegreatest support was for proper location of new wellsand well driller certification. In general, the northeastregion of the state had the most favorable opinion ofthe proposed regulations while the northwest andsouthwest regions of the state were the least in favor ofany well regulations.

Well owners that have a question or problem relatedto their well were most likely to seek assistance fromprivate water labs (22 percent), Penn State CooperativeExtension (19 percent), a well driller (11 percent) orthe Pennsylvania Department of Environmental Protec-tion (11 percent).

The well owner opinions, including support forprivate well regulations, found in this study mirrorresults found from an online survey of 865 private wellowners from 63 counties in Pennsylvania conducted bythe researchers in 2006.

Figure 10. Well owner opinions of potential private well regulations in Pennsylvania.


ConclusionsData from this study provide a wealth of information

on the incidence of pollutants in private water wellsthroughout Pennsylvania, the causes of contamination,and the ability of well owners to detect and solve waterquality problems voluntarily.

About 41 percent of samples from the private wellstested in this study failed at least one safe drinkingwater standard. Overall, the prevalence of contamina-tion was stable or declining compared to past resultsfor the parameters measured in this study. Lead con-tamination appeared to be declining in response to the1991 Federal Lead and Copper Rule while nitratecontamination was reduced from the early 1990s,presumably due to reduced applications of nitrogenthrough fertilizers and manures.

Of the 28 variables measured for each well, theresults demonstrated that natural variables, such as thetype of bedrock geology where the well was drilled,were important in explaining the occurrence of mostpollutants in wells. Soil moisture conditions at the timeof sampling were the single most important variable inexplaining the occurrence of bacteria in private wells.Man’s activities, however, were also responsible for theincreased incidence of some contaminants. Inadequatewell construction was strongly correlated with theoccurrence of both coliform and E. coli bacteria inwells. Nearly all lead contamination could be attributedto the historical use of lead plumbing components andthe occurrence of naturally corrosive groundwater.Increased nitrate concentrations were strongly relatedto the location of the well in comparison to nearbyagricultural fields. Overall, these results suggest thatnaturally occurring groundwater is not always safe forhuman consumption and man’s current and pastactivities have worsened the situation for some pollut-ants.

Unsafe levels of these pollutants in wells can beaddressed through maintenance, water treatmentdevices, bottled water or new sources of water. But amajor barrier to successful avoidance of problems thatwas identified in this study was simply creating aware-ness of problems through water testing. Most health-related pollutants do not have obvious symptoms(tastes, odors, etc.) so few wells have been properlytested. The results indicate that only about half of wellowners have had their well water tested by a certifiedlaboratory and, in many cases, they have not beenthoroughly tested. The lack of testing by well owners isnot for a lack of concern over their water quality, butinstead, a lack of awareness and understanding of whattesting should be done.

Overcoming the barrier to water testing is especiallycritical because results from this study have clearlydemonstrated that well owners are willing to takeactions to solve water quality problems. The greatmajority of well owners that were told of health-relatedwater quality issues in their water supply had voluntar-ily solved the problem within one year.

Policy ConsiderationsTo ensure safe drinking water for private water

system owners in Pennsylvania, the state shouldconsider using a combination of educational programsfor homeowners and new regulations. The combinationof these two approaches will make certain that existingwell owners will become aware of water quality issuesand proper management of wells while future wellowners will be protected from poorly protected wells orlack of knowledge. Specific policy recommendationsare addressed below.

Well TestingOne of the most revealing results from this study was

the lack of voluntary water testing done by privatewater well owners and the resulting lack of awarenessof health-related water quality problems. Several statesnow require testing of private water wells during realestate transactions but, in Pennsylvania, water testing isentirely the voluntary responsibility of a well owner.Well owners could benefit from a combination ofeducational programs and regulations targeting watertesting. Requiring well owners to have their watertested routinely is not practical or enforceable given thelarge target audience in Pennsylvania. Instead, a betterapproach would be to require all homes with a privatewater well to be tested by a certified laboratory at thecompletion of new well construction and beforefinalization of any real estate transaction. Additionaleducational programs for the remaining well ownersthat do not buy or sell their home regularly are alsoneeded (see Existing Well Owner Education on the nextpage).

Well ConstructionThe importance of well construction has been consid-

ered self-evident among groundwater experts as ameans to prevent surface water (and associated pollut-ants) from entering groundwater aquifers. In fact, thestate association of professional well drillers (Pennsyl-vania Ground Water Association) has supported wellconstruction standards for many years and most states


currently have well construction standards. By sam-pling a large number of wells with diverse settings andconstruction, the researchers were able to documentthe significant benefit of proper well construction inreducing the incidence of both coliform and E. colibacteria in wells. While this study showed that educa-tion increased the use of sanitary well caps on existingwells, most well construction features need to beincluded at the time the well is drilled. Homeownershaving new wells drilled are difficult to reach witheducational programs and, as a result, the voluntaryapproach to encourage proper well construction haslargely failed. Given the benefits of well constructionand the difficulty in reaching the target audience fornew wells, statewide regulations requiring well con-struction components appear to be warranted. Thisstudy clearly demonstrated that the majority of wellowners (>60 percent) were in favor of well construc-tion and location standards.

Wellhead ProtectionIn comparison to natural influences and well con-

struction features, nearby land-uses were less impor-tant in explaining contamination of private wells. Forexample, the distance to a nearby on-lot septic systemand the frequency of septic tank pumping were notstatistically significant in explaining the occurrence ofbacteria or nitrate in private wells. Instead, nearbysources of animal waste appeared more important incontributing to bacterial contamination of wells. Anotable exception to the lack of importance of nearbyland-uses was the strong correlation between thedistance to nearby agricultural fields and nitrateconcentrations in wells. Given the conservative natureof nitrate (ability to move long distances through soiland rock), this correlation is understandable. Overall,the results of this study do not make a strong case forthe need for wellhead protection areas around privatewells. However, the researchers recommend that thedetermination for the location of new water wellsshould still be considered very important. In mostcases, voluntary wellhead protection areas alreadyexisted around the private wells in this study as a resultof common sense location of wells used by welldrillers and homeowners. The fact that few wells werelocated very close (<50 feet) to sources of contamina-tion made it difficult to determine if minimum well-head protection areas are warranted. The connectionbetween contaminated wells and nearby land-useactivities may be underrepresented in this study sincethe vast majority of homeowners did not have wells

located in close proximity to activities that generatepollution. As a result, the data seem to confirm theimportance and success of de facto wellhead protectionareas of 50 to 100 feet that already exist around mostwells.

Existing Well Owner EducationThe aforementioned regulations and educational

programs are largely focused on new well constructionor change in ownership of existing private wells. Thereis also a clear need for an educational component onimproving drinking water quality for homeowners withexisting private wells. This study has very clearlydemonstrated the effect that education can have toincrease the frequency of water testing, the use ofcertified labs and awareness of water quality problems.Related to this was the impressive percentage (>75percent) of well owners that were willing to take actionto treat their water or use bottled water once they wereaware of a health-related problem in their water and thelarge percentages (50 to 80 percent) that were able toavoid unsafe water through voluntary actions takenduring the study. A well designed and collaborativeeducation program to deliver unbiased education towell owners could be successful if adequately funded.Such a program should be coordinated among govern-ment agencies and businesses that were identified bywell owners as important sources of information (watertesting labs, Penn State Cooperative Extension, welldrillers and Pennsylvania Department of EnvironmentalProtection). A specific component to this educationaleffort should be an emphasis on bacteria testing duringwet weather conditions to improve detection of wellswith bacteria problems.

ReferencesBartholomay, R.C., J.M. Carter, S.L. Qi, P.J. Squillace, and G.L. Rowe. (2007) Summary of Selected U.S. Geological Survey

Data on Domestic Well Water Quality for the Centers for Disease Control’s National Environmental Public HealthTracking Program: U.S. Geological Survey Scientific Investigations Report 2007–5213, p. 57.

Bickford, T.M., B.D. Lindsey, and M.R. Beaver. (1996) Bacteriological Quality of Ground Water Used for HouseholdSupply, Lower Susquehanna River Basin, Pennsylvania and Maryland. U.S. Geological Survey Water-Resources Investi-gations Rep. 96-4212, p. 31.

Centers for Disease Control and Prevention (CDC). (1998) A Survey of the Quality of Water Drawn from Domestic Wells inNine Midwest States.

Center for Watershed Stewardship (CWS). (2005) Spruce Creek Watershed Assessment and Stewardship Plan, KeystoneProject 2004-2005, Phase II. pp. 42-49.

Clemens, S., B.R. Swistock and W.E. Sharpe. (2007) The Master Well Owner Network: Volunteers Educating PennsylvaniaWell Owners. Journal of Extension. 45(4): Article 4RIB7 (http://www.joe.org/joe/2007august/rb7.shtml).

Craun, G.F. (1986) Waterborne Diseases in the U.S. CRC Press, Boca Raton, FL.Durlin, R.R., and W.P. Schaffstall. (2001) Water Resources Data, Pennsylvania, Water Year 2000, Vol. 1. Delaware River

Basin: U.S. Geological Survey Water-Data Report. PA-00-1.Iowa Department of Natural Resources. (2004) Iowa’s Water – Ambient Monitoring Program. Water Fact Sheet 2004-4. p. 4.Lindsey, B.D., J.S. Rasberry, and T.M. Zimmerman. (2002) Microbiological Quality of Water from Noncommunity Supply

Wells in Carbonate and Crystalline Aquifers of Pennsylvania. U.S. Geological Survey, Water Resources InvestigationsReport 01-4268.

Liu, A., J. Ming and R. Ankumah. (2005) Nitrate Contamination in Private Wells in Rural Alabama, United States. Science ofthe Total Environment 346(1-3): pp.112-120.

Mancl, K., W.E. Sharpe, and J. Makuch. (1989) Educating the Rural Public About Safe Drinking Water. Water ResourcesBulletin, 25(1): pp. 155-158.

Midwest Plan Service. (1992) Private Water Systems Handbook. Publication Number MWPS-14, Ames, Iowa: pp. 72.New Jersey Department of Environmental Protection. (2004) Private well testing Act program: Initial Results for September

2002 to March 2003. Division of Science, Research and Technology, and Water Supply Administration, Bureau of SafeDrinking Water: p. 65.

Rowe, B.L., P.L. Toccalino, M.J. Moran, J.S. Zogorski and C.V. Price. (2007) Occurrence and Potential Human-HealthRelevance of Volatile Organic Compounds in Drinking Water from Domestic Wells in the United States. EnvironmentalHealth Perspectives, 115(11): pp. 1539-1546.

Sharpe, W.E., D.W. Mooney and R.S. Adams. (1985) An Analysis of Ground Water Quality Data Obtained from PrivateIndividual Water Systems in Pennsylvania. Northeastern Env. Science, 4(3-4), pp. 155-159.

Swistock, B.R., W.E. Sharpe and P.D. Robillard. (1993) A Survey of Lead, Nitrate and Radon Contamination of PrivateIndividual Water Systems in Pennsylvania. Journal of Environmental Health, 55(5), pp. 6-12.

Swistock, B.R., W.E. Sharpe and J. Dickison. (2001) Educating Rural Private Water System Owners in Pennsylvania UsingSatellite vs. Traditional Programs. Journal of Extension. 39(3).

Swistock, B.R., W.E. Sharpe and P.D. Robillard. (2004) The Influence of Well Construction on Bacterial Contamination. TheCenter for Rural Pennsylvania. May 2004.

Swistock, B.R. and W.E. Sharpe. (2005). The Influence of Well Construction on Bacterial Contamination of Private WaterWells in Pennsylvania. Journal of Environmental Health, 68(2), pp. 17-23.

U.S. Census Bureau. (1970) 1970 Census of Population and Housing, Pennsylvania Housing Characteristics.U.S. Census Bureau. (1990) 1990 Census of Population and Housing, Pennsylvania Housing Characteristics.U.S. Department of Agriculture, National Agricultural Statistics Service. (2004) Pennsylvania State and County Data,

Volume 1: Geographical Area Series, Part 38, AC-02-A-38.U.S. Environmental Protection Agency, Office of Research and Development. (1991), Data User’s Guide to the United States

Environmental Protection Agency’s Long-Term Monitoring Project: Quality Assurance Plan and Data Dictionary, EPA/600/3-91/072, December 1991.

U.S. Environmental Protection Agency, Office of Water. (2003), National Primary Drinking Water Standards, EPA 816-F-03-016, June 2003.

U.S. Geological Survey (2000). Arsenic in Ground-Water Resources of the United States. Fact Sheet FS-063-00. May 2000.Zimmerman, T.M., M.L. Zimmerman and B.D. Lindsey. (2001) Relation Between Well-Construction Practices and Occur-

rence of Bacteria in Private Household-Supply Wells, South-Central and Southeastern Pennsylvania. U.S. GeologicalSurvey Water-Resources Investigations Report 01-4206.

The Center for Rural Pennsylvania200 North Third St., Suite 600Harrisburg, PA 17101Phone: (717) 787-9555Fax: (717) 772-3587www.ruralpa.org1P0109-450

Senator John GordnerChairman

Representative Tina PickettVice Chairman

Senator John WozniakTreasurer

Dr. Nancy FalvoClarion University

Secretary

Representative Tim Seip

Dr. Theodore R. AlterPennsylvania State University

Steve CrawfordGovernor’s Representative

Dr. Stephan J. GoetzNortheast Regional Center for Rural Development

Dr. Keith T. MillerLock Haven University

Dr. Robert F. PackUniversity of Pittsburgh

William SturgesGovernor’s Representative

The Center for Rural PennsylvaniaBoard of Directors

Documents

Drinking Water Quality in Rural Pennsylvania and the