Drinking Water Quality in Rural Pennsylvania

and the Effect of Management Practices

Drinking Water Quality in Rural Pennsylvania and

the Effect of Management Practices
By
Bryan 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

This research was sponsored by a grant from the Center for Rural Pennsylvania, a legislative agency of the
Pennsylvania General Assembly.
The Center for Rural Pennsylvania is a bipartisan, bicameral legislative agency that serves as a resource for
rural policy within the Pennsylvania General Assembly. It was created in 1987 under Act 16, the Rural Revitalization Act, to promote and sustain the vitality of Pennsylvanias rural and small communities.
Information contained in this report does not necessarily reflect the views of individual board members or
the Center for Rural Pennsylvania. For more information, contact the Center for Rural Pennsylvania, 200
North Third St., Suite 600, Harrisburg, PA 17101, telephone (717) 787-9555, fax (717) 772-3587, email:
[email protected].

EXECUTIVE SUMMARY
Millions of rural and suburban Pennsylvania residents rely on private wells for drinking water,
and, each year, 20,000 new wells are drilled. While research has shown that many private 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 of
TABLE OF CONTENTS
regulatory versus voluntary management of private wells. It set out
Introduction................................ 5
to determine if specific indicators, including natural factors, well
Goals and Objectives................. 6
construction features, and nearby land uses, could be correlated with
water quality parameters in private wells.
Methodology............................... 6
To complete the study, the researchers enlisted the help of more
Results ........................................ 8
than 170 Master Well Owner Network (MWON) volunteers to collect
samples from 701 private wells statewide.
Conclusions............................... 21
The findings indicated that bedrock geology was statistically sigPolicy Considerations.............. 21
nificant in explaining variations in all of the water quality parameters,
References................................. 23
with the exception of arsenic. Soil moisture, at the time of sampling,
was the most important factor in explaining the occurrence of bacteria
in wells. Individual well construction features were not statistically
important in explaining well water quality but combinations of well
construction components were statistically significant in explaining both coliform and E. coli
bacterial contamination. No specific land activities were correlated with bacterial contamination, but DNA fingerprinting of E. coli bacteria from wells found that the majority were from
animal sources. Nitrate concentrations in wells were statistically correlated with the distance
to the nearest cornfield and other crop fields. Lead contamination was found to be largely from
metal plumbing components that were exposed to acidic and soft raw groundwater.
About half of the homeowner participants in this study had never had their water tested
properly, which resulted in low awareness of water quality problems. MWON volunteers were
generally 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 percent
of the well owners that were shown to have unhealthy drinking water took steps to successfully
avoid the problem within one year after having their water tested.
Results from this study suggest a combination of educational programs for homeowners
and new regulations to overcome the largest barriers to safe drinking water. Regulations are
warranted to increase mandatory testing of private water wells at the completion of new well
construction and before finalization of any real estate transaction. For existing well owners, this
study demonstrated the effect education can have to increase the frequency of water testing, the
use of certified labs and awareness of water quality problems.
While this study showed that education increased the use of sanitary well caps on existing wells, 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, as
a result, the voluntary approach to encourage proper well construction has largely failed. Given
the benefits of well construction and the difficulty in reaching the target audience for new wells,
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 wellhead protection areas around private wells. In most cases, voluntary wellhead protection areas already
existed around private wells in this study. As a result, the data seem to confirm the importance
and success of de facto wellhead protection areas of 50 to 100 feet that already exist around
most wells.
Overall, 63 to 78 percent of well owners were supportive of potential regulations targeting
well construction, well location and well driller certification.
4

The Center for Rural Pennsylvania

INTRODUCTION
More 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 the
state1 (U.S. Census Bureau, 1970, 1990). Only Michigan has a larger population served by private water
supplies. Unlike residents that use community water
systems, homeowners with private wells are not protected by any statewide regulations.
Despite many past attempts, (the most recent were
House Bill 1591 and Senate Bill 870 in 2001), Pennsylvania remains one of the few states where well location,
construction, testing and treatment are the voluntary
responsibility of the homeowner. Some counties and
townships have passed ordinances, and considerable educational efforts have been made to meet the demands
of private well owners interested in properly managing
their water supply (Mancl et al., 1989; Swistock et al.,
2001).
The voluntary management of private wells is a
problem because most health-related pollutants in water
are symptomless. As a result, homeowners with private
water supplies may be exposed unknowingly to health
related pollutants unless they voluntarily have their
water tested for the correct water quality parameters.
Several studies have documented the occurrence of
various 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). Largescale national or statewide studies typically report that
about 15 to 50 percent of private water systems fail at
least one safe drinking water standard. Smaller, regional studies often report much lower or higher contamination rates. Only a few of these studies have made any
attempt to determine the causes, such as natural geology, land use and well construction, of contamination
of private water systems.
A significant portion of the rural population may be
exposed to unhealthy drinking water unless it properly
treats the water or uses bottled water. However, documenting the impact of polluted drinking water on the
health of residents using private water supplies is difficult because most pollutants require long-term exposure
and mimic the effects from other air- or food-borne pollutants. Those that create acute effects, such as bacteria,
have symptoms similar to common viral or bacterial

Based on the average annual increase in drilled wells reported

between the 1970 and 1990 U.S. Census. Note: Data on the use
of private water systems were not collected during the 2000 U.S.
Census.
1

Drinking Water Quality in Rural Pennsylvania

illnesses. A comprehensive study of 228 waterborne

illnesses in the 1970s by Craun (1986) determined
that residents with private or small, semi-public wells
were most vulnerable to waterborne illnesses. Recent
research in Pennsylvania has further documented the
occurrence of disease-causing bacteria in private wells
(Lindsey et al., 2002; Swistock et al., 2004).
Contamination of groundwater wells can occur from
both above and below the surface. Pollution of entire
groundwater aquifers may occur from failing septic
systems, manure and fertilizer applications, mining, or
other land uses. Individual water supplies may also be
contaminated around the exposed well casing (wellhead) from surface water flowing along the well casing
and/or from a loose fitting or absent well cap that allows insects, animals or surface water to directly enter
the well. Thus, the wellhead area, especially in poorly
constructed wells, represents a very sensitive area that
can serve as an open conduit to underground aquifers
that threaten nearby private and public water supplies.
Surface contamination can be prevented by extending the well casing above the ground surface, installing
a cement-like grout seal around the casing, and fitting
a vermin-proof or sanitary well cap on top of the
well. Extended casings, grout seals and sanitary well
caps are required in most states but they are rarely used
in Pennsylvania because there are no statewide well
construction regulations in the state. Recent studies on
a small number of wells have demonstrated the importance of a grout seal and sanitary well cap in preventing
bacterial contamination but the overall importance of
wellhead versus aquifer-wide contamination remains
poorly understood (Centers for Disease Control, 1998;
Zimmerman et al., 2001; Swistock and Sharpe, 2005).
Other issues, such as well location relative to sources
of contamination and land management practices, may
play an important role in contamination of private water
systems. For example, a recent study by the principal
investigators of 50 central Pennsylvania homes that
were at least 5 years old and had a private water supply
found that 24 percent had never pumped their septic
tank and had experienced higher rates of water contamination (Center for Watershed Stewardship, 2005).
This same study also found relationships between well
water quality and well construction, land use and septic
systems.
Homeowners with private wells typically neglect
water supply management unless obvious water quality symptoms occur. Unfortunately, most health-based

pollutants have no obvious tastes or odors in water.

Uninformed homeowners may fail to identify dangerous problems or fall victim to scare tactics used by
treatment vendors and spend thousands of dollars on
unnecessary treatment equipment. In fact, a recent
survey of central Pennsylvania well owners found that
only 50 percent had ever tested (most just for bacteria)
their wells and about 10 percent had purchased unnecessary water treatment equipment (Center for Watershed Stewardship, 2005).
Penn State Cooperative Extension has a long history
of educating well owners in Pennsylvania but the large
target audience has made education a difficult task
(Mancl et al., 1989; Swistock et al. 2001). In an attempt
to more efficiently educate well owners, the researchers
received federal funding in 2004 to create the Master
Well Owner Network (MWON). This program provided eight hours of training on proper water well management to over 300 volunteers from 64 counties. This
program has educated thousands of private well owners in the state resulting in its expansion into Virginia,
Maryland, Delaware and West Virginia.
Private wells are pervasive across the landscape of
Pennsylvania, serving as important sources of water for
rural and suburban homes and farms. However, poorly
constructed and unmanaged water wells represent
potential risks for vital groundwater aquifers and the
homeowners, farmers and businesses that access them.
Limited available data suggest that contamination of
these water supplies is widespread but little is known
about the magnitude of the problem, the causes of pollution, and policies or educational efforts that may best
address the problem. The Master Well Owner Network
(MWON) volunteers represented a uniquely efficient
opportunity to gather important data about private water
wells throughout the state to better understand the occurrence and sources of contamination and the potential
impact of regulation versus voluntary education.

Goals and Objectives

This two-year study, conducted in 2006 and 2007, set

out to:
Determine the occurrence of several health-related
pollutants in 700 private wells throughout the state.
Determine if well contamination is different between wells with different types of well construction (i.e. buried casing, type of well cap, presence of
grout, slope around well, etc.).
Determine how water supply characteristics, such
as proximity to polluting activities and nearby septic
system maintenance, influence groundwater quality.
6

Determine how frequently private wells are voluntarily tested and adequately treated for contaminants
and how often homeowners voluntarily follow recommendations to solve well contamination problems.

Methodology

A total of 172 MWON volunteers and Cooperative

Extension staff collected samples from 450 private
wells in 2006 and another 251 wells in 2007 for a total
of 701 wells. The approximate locations of the wells
are shown in Figure 1. Each of the MWON volunteers attended eight hours of training on private water
system construction, management and testing. Most
of the MWON volunteers attended training workshops
during 2004 or 2005 but some were trained in 2006.
The volunteers had varied backgrounds including water
resource agency professionals (30 percent), professions
unrelated to water (23 percent), retired (20 percent),
educators (7 percent), health care professionals (5 percent), local government representatives (5 percent) and
farmers (4 percent) (Clemens et al., 2007).
Of the 701 private wells, 79 were owned by existing
MWON volunteers and 622 were owned by either newly trained volunteers or homeowners. For the purposes
of this report, the 79 MWON volunteer wells were
usually treated differently than the 622 wells owned by
homeowners or newly trained volunteers because of
their management history. Existing MWON volunteers
would have had at least one year to improve the management of their private water well based on information they learned at their training course. Homeowner
or newly trained MWON wells would be more representative of typical private wells throughout the state.
MWON volunteers were recruited to help with
sample collection through email and phone calls. One
month before each sample collection workshop, interested volunteers were sent sample bottles, surveys and
instructions for collecting samples. Approximately one
week before the respective sample collection workshops, the MWON coordinator called each volunteer to
answer questions and confirm sample collection criteria
and instructions. Volunteers were also encouraged to
contact their MWON regional coordinator with any
project questions. Coordinators were trained in early
2006 on sample and survey procedures.
Selection of all study participants was done by
MWON volunteers, regional coordinators, Cooperative
Extension educators or project staff. To ensure spatial
distribution, volunteers maintained at least a one-mile
distance between sampled wells. Volunteers were
encouraged to select private wells that were familiar to
The Center for Rural Pennsylvania

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

The survey forms for each

water supply were divided into
a homeowner and volunteer
portion. The volunteers visually
inspected the well, plumbing,
and surrounding landscape
and reported information on
well construction and sources
of contamination. The homeowner survey included questions about the well age, well
depth, approximate yield, past
water testing, known problems
or symptoms, septic system
characteristics, water treatment equipment, and opinions
about the need for private well
regulations. The surveys were
returned to the drop-off location with each well sample.

Follow-Up Survey

them (friends, family, co-workers, etc.) to facilitate collection of water samples and water supply information
in a timely and accurate fashion.

Sample and Survey Collection

Volunteers collected water samples from each well on

the morning of the workshop, stored the samples on ice
and returned the samples and surveys to the workshop
location usually no later than noon. The training and
experience of volunteers helped to ensure that samples
were collected properly and that water supply information was accurately recorded on the survey.
Two water samples were collected from each home.
The first was a sample of first-draw water (water that
had been in the plumbing for at least six hours) which
was tested for lead from corrosion of metal plumbing
components. Volunteers collected first-draw samples
from the kitchen faucet (78 percent of samples) or other
faucets used for drinking water (22 percent of samples).
The second sample was analyzed for all other chemical 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 an
outside hose bib. The water was allowed to run for one
to two minutes prior to the sample being collected.
Volunteers reported that 46 percent of the running water
samples came from the kitchen faucet, 27 percent was
from the pressure tank and 27 percent was from other
taps (mostly outside hose bibs).
Drinking Water Quality in Rural Pennsylvania

In 2007, an additional survey was sent to each of the

450 well owners that participated in the study during
2006. The purpose of this survey was to document any
actions taken by the well owner to solve water quality
problems or better manage their water supply as a result
of participating in the study. This follow-up survey was
mailed approximately six to 12 months after the wells
had been tested in 2006 to give homeowners the opportunity to take action on their well. A reminder postcard
was mailed a few weeks after the survey resulting in an
overall return rate of 64.2 percent (289 well owners).

Sample Analyses

Water samples were analyzed for the following eight

contaminants that are likely to occur in rural and suburban areas from activities or surface water contamination.
Total coliform bacteria are indicators used to determine if a pathway exists that might allow diseasecausing bacteria to contaminate the water supply. E.
coli bacteria are a subset of coliform bacteria that
only occur in animal or human wastes and indicate
more serious contamination.
E. coli source tracking was conducted at the E. coli
Reference Center at Penn State University in an attempt to determine the role of animals versus humans
as a source of E. coli contamination in private wells.
pH is a common measure of water quality that is
often related to corrosion of plumbing system components like lead and copper.
7

 Lead is a dangerous metal with many health effects,

especially in children. Lead usually occurs in drinking water supplies from corrosion of lead solder or
lead impurities in plumbing fixtures.
Nitrate-nitrogen is a health related pollutant that
usually originates from fertilizer, manure or septic
systems.
Arsenic is a relatively new concern in drinking
water with serious health effects at very low concentrations. It is thought to most often occur naturally
from certain types of rocks but it can also come from
treated lumber and pesticides.
Triazine pesticides include the most common
pesticides used in Pennsylvania, such as atrazine and
simazine. These pesticides have serious health effects.
Hardness is a measure of the amount of calcium and
magnesium in the water. Although harmless, hardness
can cause aesthetic effects including stains, damaged
hot water heaters and increased soap usage. Hardness
was measured primarily to determine the appropriate
versus unnecessary use of water softener treatment
devices since softeners are the most common water
treatment device sold in Pennsylvania.

Well Owner Notification

All participants in the study received their water

quality results within six weeks of their sample submissions. Individual water quality results were kept confidential and were only accessible to project personnel.
All participants with a water system that failed to meet
one or more drinking water standards received specific recommendations and Penn State Fact Sheets to
assist them in taking action to solve their water quality
problem. Homeowners were also given a phone number
and email address to ask questions about their results.
During the two-year study, 19 of the 701 homeowners
called with follow-up questions after receiving their
water test report.

Statistical Analyses

In this report, where statistical analyses were not possible, direct comparison of percentages were used rather than detailed statistical analyses. Analysis of covariance (ANCOVA) and logistic regression models were
used to determine which well characteristics (geology,
construction, etc.), land uses (distance to septic, farm
fields, etc.) and management activities (septic system
pumping frequency, etc.) were important in explaining
the occurrence of the various pollutants.

Data Quality Assurance and Quality Control

This study analyzed 58 quality control samples,

representing about 8 percent of the 701 private well
samples analyzed. These samples were submitted to the
laboratory among other samples from private wells to
measure how accurate, precise and repeatable the water
quality results were from the laboratory.
Overall, the results from the quality control samples
indicated that water quality data collected during the
study were of excellent quality.

Effect of Location and Sample Method on Bacteria

Another data quality question centered on the location and method used to collect water samples for bacteria analysis. Extra care must be taken when sampling
water supplies for coliform bacteria and E. coli bacteria
because of the risk of contaminating the sample from
hands and other surfaces during collection and subsequent handling. The results of the protocol used in this
research suggest that the sampling strategies, including
washing the bottle three times and not flame sterilizing
the end of the faucets, did not cause any systematic
bacterial contamination of the samples from the 701
study wells.

Results

Well Characteristics

The characteristics of the 701 private wells sampled

came from responses on the homeowner and volunteer
surveys submitted for each well. Generally, well owners
were very knowledgeable about their wells and most
were able to provide the basic characteristics of their
well. For example:
83 percent were able to estimate the depth of their
well. The average well depth was 172 feet (median
depth = 140 feet) with a maximum depth of nearly
1,000 feet.
88 percent knew the approximate year their well
was drilled. More than 72 percent had been drilled
since 1970 but only 4 percent had been drilled since
2005.
Well owners knew their well characteristics despite
the fact that very few possessed a copy of their well
completion report or well log. A copy of the well
completion report should be provided to the well
owner to document the well characteristics, such as
depth, construction, geology, and yield.

The Center for Rural Pennsylvania

Water Quantity
About 51 percent of well owners could estimate the
yield of water coming from their well, with an average
yield of about 18 gallons per minute (gpm). The Midwest Plan Service recommends a minimum of 6 gpm
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 reported
were greater than 6 gpm. According to well owners,
well yields were not a major problem across the state.

lined with stone and usually covered with a cement slab

cap. Of the 701 wells sampled in this study, 6 percent
were hand-dug and the remaining 94 percent were
drilled.
Volunteers were asked to determine, as best as possible, if a well was grouted or if the well had a visible
cement seal around the casing. Since so few wells had
a well completion report, the presence of a proper grout
seal that extended from the surface to bedrock was
impossible to determine in most cases. Instead, grout
presence had to be based on homeowner memory or visual evidence of cement or grout residual at the surface
Well Construction
MWON volunteers evaluated the construction of each around the casing by the volunteer sampling the well
well (Figure 2). Of the 701 wells, 12 percent (84 wells) (most cases).
Volunteers reported that grout or a cement seal
were completely buried below the ground surface, often
around the well casing was present on 18 percent (120
in a pit or basement. Nine percent of wells owned by
wells) of the private wells and this percentage was
an MWON volunteer were buried while 13 percent of
similar between volunteer and homeowner wells. Grout
homeowner wells were buried. This was a common
practice decades ago to keep water lines from freezing. information was available from well logs for only 19 of
these wells. About 5 percent of the homeowner wells
Of the 85 wells in this study that had a buried cashad both a sanitary well cap and evidence of grout, acing, most were drilled prior to 1970. Still, there were
cording to volunteers. A follow-up email survey of well
eight wells with a buried casing that had been drilled
owners with grouted wells found that only a handful
after 1990. Buried wells were most prevalent in northcould confirm for certain that their entire well casing
west Pennsylvania, probably because wells tended to
was properly grouted at the time the well was drilled.
be older in that region, but there were no statistically
Since only eight wells had been drilled after enactment
significant differences in the occurrence of any well
of county well construction ordinances in Chester and
construction features (buried casing, sanitary well cap,
Montgomery counties, it seems likely that the majorgrout, etc.) between the six regions of the state.
ity of the wells denoted as grouted in this study were
Of the homeowner wells that were visible above the
only grouted along the first few feet of the casing below
ground, 16 percent had a sealed, sanitary well cap.
the ground surface instead of the entire length of the
Sanitary well caps were more than twice as likely on
well casing as recommended.
MWON volunteers wells, presumably because they
Sloping the ground in the immediate vicinity of the
received a free sanitary well cap during their MWON
well casing is recommended to prevent ponding of
training. About 9 percent of homeowner wells had
surface water and possible contamination of the well by
missing well caps or miscellaneous types of caps (coffee cans, cement, ceramic, etc.). Many of the wells with surface water. Visual observation by volunteers showed
that 36 percent of homeowner wells had this slope
miscellaneous caps were very old, hand-dug wells that
around the well. A greater percentage of volunteers
were typically constructed in the early 1900s. Handdug wells have a large diameter hole (several feet wide) wells (54 percent) were sloped, presumably because
volunteers had learned the value of sloping
during their MWON training.
Figure 2. Well construction components for homeowner
For statistical comparisons, five well conand MWON volunteer wells.
struction features were combined to provide an
overall well construction score or wellscore.
The wellscore was used to correlate overall
well construction with water quality parameters in wells. Well construction features
considered in the wellscore included:
Casing a drilled well with a metal or plastic
casing (casing score = 1) versus a hand-dug
well without a casing (casing score = 0).
Buried the well casing is visible above
Drinking Water Quality in Rural Pennsylvania

ground (buried score =1) versus a buried well casing

not visible above ground (buried score = 0).
Grout evidence of grout or cement around the casing at the surface (grout score = 1) versus no evidence
of grout or cement around the casing (grout score = 0).
Slope no ground slope around the well head including wells in a depression (slope score = 0) versus
wells with ground sloping away from the casing
(slope score = 1).
Well cap a sanitary, sealed, vermin-proof type
(well cap score = 1) versus an unsealed, standard or
missing well cap (cap score = 0).
The wellscore for each well was the added score of
the five well construction features with scores ranging
from 0 (very poor construction) to five (very good construction). Of the 701 wells in this study, a breakdown
of wells in each wellscore category are:
Wellscore 0 = 3 percent (no recommended well
construction components);
Wellscore 1 = 10 percent (one recommended well
construction component);
Wellscore 2 = 41 percent (two recommended well
construction components);
Wellscore 3 = 30 percent (three recommended well
construction components);
Wellscore 4 = 12 percent (four recommended well
construction components); and
Wellscore 5 = 4 percent (all five recommended well
construction components).
Water Treatment Equipment
Water treatment equipment was installed on many of the
private wells, with 53 percent having at least one piece of
treatment equipment. MWON volunteers were only slightly more likely to have treatment equipment compared to
homeowners. Well owners in the northeast region of the
state were significantly less likely to have water treatment
equipment compared to other regions of the state; presumably because water hardness is much lower in that region,
decreasing the need for water softeners.
The percent of wells that had each type of water
treatment equipment is listed in Figure 3. Water softeners were the most common type of water treatment
equipment, occurring in about one-third of homes with
private wells in this study. Sediment filters were also
quite common but the majority of sediment filters were
small, cartridge filters that are installed ahead of larger,
expensive units (like softeners, ultraviolet lights, etc.)
to remove large sediment particles and prevent wear
and 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
10

Figure 3. Occurrence of water treatment equipment on wells

in this study.

sediment problems. Other types of water treatment,

such as oxidizing filters to remove iron, manganese
and/or hydrogen sulfide gas and ultraviolet lights to kill
bacteria, were far less common, occurring in less than
10 percent of the homes with private wells. The average
cost of water treatment equipment and installation was
$1,127 with a maximum of $7,000. These values were
not adjusted for inflation because the year of installation was not measured.
Given that the most common water treatment devices
listed in Figure 3 are used primarily to treat obvious
aesthetic problems in water, it was not surprising that
half of well owners indicated that they had installed
water treatment equipment because of obvious stains,
odors or tastes. The other half of well owners with
water treatment owned it because they had water test
reports showing a problem (32 percent) or they inherited it from a previous homeowner (20 percent).
Wastewater Treatment Characteristics
Eighty-nine percent of the homes with private wells
used an on-site system for wastewater disposal. Traditional septic tanks and leach fields occurred in 72
percent of the homes. The remaining 17 percent used
sand mounds (14 percent), alternative on-lot systems (2
percent), or did not know where their wastewater went
(1 percent). Of those with on-lot wastewater systems,
13 percent reported problems with malfunctions. Septic
system malfunction rates were similar between standard
septic systems, sand mounds and alternative systems,
and were also similar between regions of the state.
The Center for Rural Pennsylvania

Septic system maintenance was determined by asking

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

Prevalence and Spatial Occurrence

of Well Contamination

Of the parameters that were tested on each well water

supply, six had primary drinking water standards (i.e.
detrimental health effects are possible if standards are
not met) including coliform bacteria, E. coli bacteria,
lead, nitrate, arsenic and triazine pesticides. Hardness
does not have a drinking water standard and pH has a
secondary or recommended drinking water standard
(for aesthetic effects on taste and corrosion).
Contamination rates in raw well water were similar
between MWON volunteer wells and homeowner wells
so they have been lumped together for reporting here.
Overall, approximately 41 percent of the wells tested
failed to meet at least one of the health-based drinking water standards. Of these wells, most (89 percent)
failed only one of the drinking water standards.
Keep in mind that these results apply to raw water
and do not include the effect of water treatment devices
or bottled water that some well owners used to avoid
health-related pollutants. In fact, roughly 25 percent of
well owners with a health-related pollutant was avoiding it by using proper water treatment or by drinking
only bottled water. Actions taken by wells owners,
both before this study and as a result of this study, are
discussed later. Details of individual contaminants are
discussed below. For the regional analyses, the total
number of wells sampled in each region were: northwest (61 wells), northcentral (115 wells), northeast
(167 wells), southwest (98 wells), southcentral (159
wells), and southeast (101 wells).
Total Coliform (TC) Bacteria
TC bacteria were found in 33 percent of the sampled
wells and were absent in 67 percent. This contamination rate is similar to past national and statewide surveys of private water systems.
Drinking Water Quality in Rural Pennsylvania

TC bacteria occurred throughout the state but there

were some regional differences. The highest incidence
of TC occurred in the southeast and southwest regions
while the lowest incidence was observed in the northwest and northeast regions. Statistically, the southeast
region had a significantly higher occurrence of TC
bacteria than the northwest or northeast regions. These
regional trends are consistent with TC contamination
reported by Sharpe et al. (1985).
E. coli Bacteria
E. coli (EC) bacteria should be absent from drinking water for the water to be safe to drink. EC bacteria
represent a more serious contamination issue than
coliform bacteria since EC bacteria can only originate
from human or animal waste. In this study, EC bacteria
were detected in 14 percent of the private wells. No
other statewide surveys have been done to document
the occurrence of EC in private wells. The incidence of
EC bacteria found in this study is greater than reported
by some regional studies (Durlin and Schaffstall, 2001;
Zimmerman et al., 2001) but less than values reported
by others (Bickford et al., 1996). EC bacteria showed
similar regional trends to those found for TC bacteria.
A total of 213 distinct EC bacteria colonies from 79
different private wells were analyzed by the E. coli Reference Center at Penn State. The DNA fingerprints of
each of these bacteria were statistically compared to the
library of EC bacteria from known animal and human
sources in an attempt to determine the relative role of
animal versus human wastes in causing EC contamination of private wells. The results showed that most EC
colonies were more closely related to animal sources
than human sources. This suggests that most contamination occurs by surface water from nearby animalrelated land activites.
pH and Lead
The pH of each sample was determined as a general
indicator of the corrosion potential to correlate to lead
contamination. In general, low pH water (below 7.0)
tends to be corrosive, which means it can dissolve
metals, such as copper and lead from pipes, solder or
fixtures, from the plumbing system. Lead was measured
in first-draw water samples from 251 wells during 2007.
Wells with a pH below the recommended level of
6.5 occurred in 18 percent of the wells tested and were
most frequently found in the southeastern region. Wells
with a high pH (above the recommended level of 8.5)
were rare (2 percent) and occurred sporadically across
all regions.

11

While only 18 percent of private wells were below

the recommended pH standard of 6.5, it is likely that a
much higher percentage of wells contained corrosive
water. Many other water quality parameters, including
hardness and total dissolved solids, are important in
determining the corrosion potential of water. This study
did not test for all of the parameters necessary to calculate a corrosion index (known as the Langelier Saturation Index or LSI). Sharpe et al. (1985) found that,
while only 18 percent of private wells were too acidic
(pH < 6.5), a much higher percentage (59 percent) were
actually corrosive to metal plumbing. In that study, corrosive water was the most common water quality found
in Pennsylvania, especially in northcentral Pennsylvania where well water tended to be both acidic and soft
(low hardness and total dissolved solids).
Elevated lead levels occurred in first-draw water from
12 percent of the wells that were tested during 2007. It
appears that corrosion of lead from plumbing components (rather than lead from groundwater pollution) was
the predominant source of lead in these well water supplies. Ninety-three percent of the wells with high lead
levels had acidic water (pH below 7.0) and 80 percent
also had soft water. As mentioned above, this combination of acidic, soft water causes corrosion of metals.
Lead contamination was most likely in the southcentral
(17 percent) and southeast (16 percent) regions and
least likely in the northeast region (5 percent). These regional variations in lead are identical to those reported
by Swistock et al. (1993).
The 12 percent failure rate for lead is much lower
than the 19 percent reported in a study by Swistock et
al. (1993). It is presumed that this reduction is a result
of the 1991 Federal Lead and Copper Rule that required
the use of lead-free solder and fixtures in home plumbing. Seventy percent of the homes with high lead levels
had plumbing systems that were installed prior to enactment of this rule and most also had copper plumbing
systems. There was only one private well with a high
lead level that could not be clearly linked to corrosion
of metal plumbing (i.e. a new home with plastic plumbing and alkaline water).
Nitrate
Nitrate-nitrogen occurred above the drinking water
standard of 10 mg/L in only 2 percent of the private
wells. Nitrate concentrations in private wells varied
strongly between regions. Mean nitrate concentrations
were significantly higher in the southeast and southcentral regions compared to the other four regions. Still,
sporadic nitrate concentrations above 10 mg/L were
found in the central and northeast regions.
12

The 2 percent of wells that exceeded 10 mg/L of

nitrate-nitrogen was far below the 14 percent reported
by Sharpe et al. (1985) and the 9 percent reported by
Swistock et al. (1993). Both of these past studies had a
larger proportion of samples from southeast Pennsylvania than this study and both also reported a strong
regional influence on nitrate occurrence. Still, these
comparisons suggest that groundwater nitrates are
lower than they have been historically, perhaps due to
better management of nitrogen on farms and home fertilizers resulting from education and mandated nutrient
management plans. Data reported in the 2002 Census
of Agriculture (the most recent census available) by
the U.S. Department of Agriculture (2004) suggest
that applications of nitrogen by fertilizer and manure
have dropped in southern Pennsylvania since the early
1990s.
Arsenic
Only 2 percent of the wells exceeded the health-based
drinking water standard of 10 ug/L for arsenic. The
maximum concentration observed was 35 ug/L but the
majority of wells (89 percent) had arsenic concentrations below 6 ug/L. Wells with high arsenic occurred
mostly in northern Pennsylvania regions, presumably
due to the geology of these areas. The three northern
regions of the state had significantly higher arsenic
concentrations than the southern regions with the highest occurring in the northwest region. These results
are similar to results reported by the U.S. Geological
Survey (2000) for 578 private wells that were sampled
in southeast and extreme western Pennsylvania. Arsenic
is thought to originate primarily from natural geologic
sources, thus, it would not be expected to vary significantly over time.
Triazine Pesticides
Only three wells (less than 1 percent) had unsafe
concentrations of triazine pesticides above 3 ug/L. Two
of these wells were located in southcentral Pennsylvania and one well was located in the northeast region. Of
these three wells, two homeowners were willing to have
their well re-tested in 2007 using more sophisticated
analyses. Those re-tests occurred in late 2007 and they
did not detect pesticides in either well.
The only other survey of pesticides in private wells
in Pennsylvania was conducted during 1993 in an
unpublished study by Penn State University. Of 189
private wells sampled in that study, none had atrazine
or simazine concentrations above the drinking water
standard. Another recent study by Bartholomay et al.
(2007) tested several hundred wells in Pennsylvania
The Center for Rural Pennsylvania

and found none that exceeded the drinking water standard for atrazine or its
breakdown products. It should be noted
that these past studies of pesticides in
Pennsylvania have generally detected
small concentrations of pesticides
(above the detection level but below the
drinking water standard). This studys
testing method only provided a result of
present or absent at a level above
the drinking water standard. So, while
the sampling shows that very few wells
were above the drinking water standard,
the results do not allow an estimation of
the number of wells that may have triazine pesticides present at lower detectable concentrations.

Figure 4. Variables used in ANCOVA statistical models to determine causes

of contamination of private wells.

Hardness
There are no health effects or drinking
water standards for hardness but hard
water can cause numerous aesthetic
problems, especially when water is
heated. Because hardness reduces corrosion of household plumbing, a level
of 90 to 100 mg/L is often considered
optimum to reduce corrosion while also
preventing unwanted aesthetic effects.
Total hardness is usually reported in
one of four categories as follows: soft
water has a hardness concentration of 0
to 60 mg/L; moderately hard water has
a hardness concentration of 61 to 120
mg/L; hard water has a hardness concentration of 121 to 180 mg/L; and very hard water
has a hardness concentration greater than 180 mg/L.
There were clear regional differences in hardness
concentrations across the state. Most wells with very
hard water were located in western counties or in
central and southcentral Pennsylvania. Wells in the
northcentral and northeast regions had significantly
lower hardness concentrations (means = 70 to 90
mg/L) compared to the other four regions (means =
120 to 140 mg/L) due to the bedrock geology of these
regions.

Variables Controlling Contamination

of Private Wells

Contamination of private wells can occur through

the interaction of both natural and human causes.
Leaching of arsenic from bedrock is an example of a

Drinking Water Quality in Rural Pennsylvania

natural source while leaching of bacteria from a septic

system is an example of a human cause. Figure 4 shows
all of the variables that were included in the ANCOVA
statistical model in an attempt to explain sources of contamination in private wells. The ANCOVA model was
then able to determine which variables were statistically
significant in explaining water quality differences.
Climate and Other Natural Factors
Regional differences were previously reported in the
summary of each water quality parameter but much of
this variation can be attributed to geologic differences
between the regions. In the ANCOVA models, geology
was statistically significant in explaining the variation in
all of the water quality parameters with the exception of
arsenic. In most cases, water quality was strongly tied to
geology. The geologic controls on water quality are not
surprising considering that groundwater in private wells
is in direct contact with the bedrock for extended periods
13

of time. However, some of the geologic differences are

probably the result of land uses that are predominant
on certain types of bedrock. For example, higher nitrate
levels on carbonate and igneous bedrock are likely due
to the fact that these bedrock types are predominant
in the regions of the state with intensive agricultural
land use rather than actual differences in the bedrock
chemistry.
The carbonate rock type produced the most unique
water quality with significantly higher bacteria levels,
pH, nitrate and hardness compared to most other rock
types. A recent study of private wells in southcentral
Pennsylvania also reported a higher incidence of coliform and E. coli bacteria in wells located in carbonate
bedrock because of the close connection between carbonate aquifers and surface water (Zimmerman et al.,
2001). Igneous rock, located in parts of southeast and
southcentral Pennsylvania, was more acidic (lower pH)
and lower in hardness compared to other rock types.
This resulted in generally higher lead concentrations
although this difference was not statistically significant.
Sedimentary and sandstone/shale bedrock types, which
are both comprised of various types of sandstone and
shale, predictably produced nearly identical water quality results. These results suggest that the natural geology where a private well is drilled plays an important
role in the resulting water quality, regardless of well
construction and nearby land use impacts. The overall
importance of geology on many water quality parameters has also been observed in testing of 5,000 private
wells in New Jersey (New Jersey DEP, 2004).
Climatic conditions, such as precipitation and temperature, are thought to be important when testing pollutants in groundwater wells although limited research
data are available on this subject. Contaminants originating from the ground surface, like coliform bacteria
and E. coli bacteria, would be expected to be more
prevalent during wet and warm weather since these
conditions favor the growth of bacteria and wet weather
promotes the movement of surface water and surface
contaminants into the ground.
Of the wells that contained coliform bacteria or
E. coli bacteria, 84 percent were tested during moist
conditions, while only 16 percent were tested during
dry weather. These results agree with those of Swistock
and Sharpe (2005) who found that bacterial contamination of private wells in Pennsylvania was greatly
reduced during an extreme drought in 2002. Short-term
moisture conditions were not statistically important in
explaining concentrations of nitrate, lead, pH, hardness
or arsenic in private wells.
The strong relationship between bacteria levels and
14

moisture conditions has many implications. Results of

bacteria testing will likely be more variable than other
water quality parameters, depending on the weather
at the time of sampling. As such, the bacteria results
presented here must be considered a snapshot of conditions that apply to the weather conditions that existed
during the study. Conditions during 2006 and 2007
throughout Pennsylvania encompassed a range of shortterm conditions from mild drought (summer 2007)
to moderate wetness (fall 2006) but, overall, these
two years were near climatic norms. Results from the
homeowner survey and re-sampling of wells between
2006 and 2007 also elude to the variability of bacteria
results. Twenty-six of the 701 well owners (4 percent)
indicated they had coliform bacteria in their water well
but this testing found none present. Also, some wells
that contained small numbers of coliform bacteria in
2006 did not contain these bacteria when they were retested in 2007. The variability of bacteria results related
to weather conditions must be considered when making
recommendations to well owners about well testing.
Season of the year was also tested as a potential cause
of water quality variations in wells. One might expect
that warmer and wetter conditions during spring and
summer would result in a greater likelihood of bacteria,
nitrate, lead and pesticides in wells while generally
drier and cooler conditions during fall might cause
higher pH, hardness and arsenic as dilution from rainwater is reduced. In the ANCOVA models, season was
not statistically important for any of the water quality
parameters. Bacterial contamination was nearly constant among the three seasons (spring, summer and fall)
that wells were tested.

Effect of Well Construction
Well construction features could play a role in the
entry of some pollutants, especially those generated
near the land surface, into private wells. For example,
bacterial contamination of wells may occur through improper construction practices that allow surface water,
insects or small mammals into the well. Sanitary well
caps, grout seals and sloped ground near the casing are
all used to prevent this wellhead contamination. Five
separate well construction components (casing present,
casing above ground, sanitary cap, grout or cement seal,
slope around casing) along with the overall wellscore
(the number of proper well construction components
on the well) were included in the ANCOVA models for
each water quality parameter. Additional well characteristics, including depth and age, were included in the
models.

The Center for Rural Pennsylvania

In the case of coliform and E. coli bacteria, individual

well construction components resulted in slightly
reduced contamination rates (for example, wells with
sanitary well caps had slightly lower bacterial contamination rates than those with standard or missing well
caps) but none of these individual components produced statistically significant results. But, combinations
of well construction features were highly significant in
reducing total coliform bacteria and E. coli bacteria.
Figure 5 illustrates the effect of increasing numbers of
well construction features on bacterial contamination.
Note that wells with very poor construction (zero features) were twice as likely to have coliform bacteria and
five times more likely to have E. coli bacteria compared
to wells with excellent construction (five features). The
contamination rates for wells with a wellscore=5 (28
percent for coliform and 8 percent for E. coli) are very
similar to data published by Swistock et al. (2005) from
sampling of 24 wells with excellent construction in
eastern Pennsylvania. Other studies have demonstrated
slightly reduced incidence of bacteria from grout (Zimmerman et al. 2001), sanitary well caps (Swistock et al.,
2005), and cased wells versus hand-dug (Sharpe et al.
1985) but no other studies have shown a clear connection between overall well construction and bacterial
contamination. These results suggest that no single well
construction feature is critical to preventing surface
water contamination but, clearly, combinations of
features have a significant effect in preventing bacterial
contamination of private wells.
Despite the importance of well construction to bacterial contamination, it was not important in explaining other
water quality problems in private wells. In fact, no single

well construction feature, or even the overall wellscore,

was statistically significant in any other ANCOVA model. In the case of nitrate, the age of the well was statistically important and well depth was modestly significant.
These factors combined suggest that older wells, which
are typically shallower, are perhaps allowing shallow,
nitrate-rich water to enter deeper groundwater aquifers.
Other than nitrate, well depth was not statistically significant in explaining any of the water quality parameters.
The overall lack of importance of well characteristics on
parameters besides bacteria is not surprising given the
importance of geology described earlier.

Effect of Nearby Activities

The sample collection survey completed by each
MWON volunteer included estimated distances to 14
nearby activities (crop field, dog kennels, septic system,
etc.) that could influence some water quality parameters in wells. Each of these distances was included
in the ANCOVA models along with well construction,
climate, and geology. Overall, few of these activities
were statistically significant in explaining water quality
in private wells.
The overall factors that control each well water quality parameter (natural factors, well construction factors,
and land activities) are summarized in Figure 6 on Page
16. These results do not argue strongly for mandated
or voluntary wellhead protection areas around private
wells where various activities are prohibited to protect well water quality. However, these results should
be interpreted with some caution. The overwhelming
majority of the private wells in this study already, in
effect, had a wellhead protection area of at least 50 feet.
In fact, more than 70 percent
Figure 5. The effect of increasing numbers of well construction features on coliform
of the study wells were located
and E. coli bacterial contamination.
at least 50 feet from all of the
possible polluting activities.
Thus, there were relatively
few wells that could be used
to test the effect of each of
these activities. These results,
therefore, are probably indicative of the overall effectiveness
of a minimum 50-foot wellhead protection area that most
homeowners and/or well drillers seem to provide either by
chance or as common sense.

Drinking Water Quality in Rural Pennsylvania

15

Figure 6. A summary of statistically significant variables

water had been tested at least once in the past, only
(p<0.05) in ANCOVA models for each water quality parameter. about half had used a state-certified laboratory (Penn-

Wellscore = number of recommended well construction features

(0 to 5). 2 Well Depth p-value was 0.08 indicating that the Well Age
correlation is probably related to a combination of age and depth
(i.e. older wells were shallower). 3 Lead is highly correlated with
plumbing system characteristics and water pH (an index to corrosion) which is related to geology.
1

Voluntary Water Well Management

In the absence of statewide regulations for well

construction and/or maintenance, education plays an
important role in promoting proper management of
private water supplies. However, while education can
create awareness of problems and management strategies, it can only be successful if well owners translate
this knowledge into action, such as by testing their well
water or installing treatment systems, when problems
are identified. If it is found that well owners are generally maintaining their wells properly, then it can be
argued that regulations are less necessary to protect the
health of well owners.
Water Testing
Most agencies recommend having private wells tested
by a state-certified laboratory annually for coliform
bacteria and every few years for other contaminants.
The 622 homeowners that participated in this study
had rarely followed these recommendations. In fact,
three-fourths of homeowners had either tested their
water quality just once (44 percent) or had never had
their drinking water tested (30 percent). It is important
to note that these estimates may be high since most of
these well owners were familiar with MWON volunteers who may have previously encouraged them to
have their water tested. Trained MWON volunteers
themselves were more likely to have their water tested
more frequently than uneducated homeowners but some
had still not had their water tested.
Of the 74 percent of homeowners that indicated their

16

sylvania Department of Environmental Protection lab

or state-certified commercial lab). The remainder was
tested by either water treatment vendors or the use of
home water testing kits. MWON volunteers, on the
other hand, were much more likely (69 percent) to use
certified water labs. Overall, about half of the homeowner participants in this study had never had their
water tested properly (either never tested or only testing
was a non-certified lab) while fewer MWON volunteers (30 percent) had failed to have their water tested
properly.
The frequency of water testing did vary somewhat by
regions of the state. Private well owners in the southeast region of the state were significantly more likely to
have their well tested and also to use a certified testing
laboratory compared to other regions of the state. The
reasons for this difference are not clear but could be
related to the better availability of water laboratories or
greater concerns of threats to private water systems in
this region.
Water Quality Awareness
For the purposes of this discussion, the term awareness refers to a well owners knowledge of his/her
water well quality prior to participating in this study.
Given the low percentage of homeowners that had
previously tested their well water quality, it would
be expected that few were aware of the water quality
problems in their own well. This was generally true for
health-based pollutants that have no obvious symptom
(taste, odor, etc.) in water. Figure 7 illustrates the percent of homeowners and MWON volunteers that were
already aware of each pollutant that existed in their well
before participating in this study.
In Figure 7, the 11 percent awareness value for homeowners with coliform bacteria means that, of the 203
homeowner wells that were found to contain coliform
bacteria, only 11 percent of those well owners were
already aware of this contamination prior to participating in this study. In all cases, awareness of contamination among homeowners was below 31 percent, with
the highest awareness occurring for parameters that are
more often routinely tested by water laboratories (nitrate, pH, bacteria). MWON volunteers were generally
two to three times more likely to know about a healthrelated pollutant in their well suggesting that education
can greatly improve awareness of problems.
Naturally, homeowners are more aware of water pollutants that only affect taste, or cause stains or odors
because the effects are obvious. For example, nearly 50
The Center for Rural Pennsylvania

Figure 7. Percent of homeowners versus MWON volunteers who were aware that each
pollutant occurred in their well prior to the study. N/A indicates that no MWON volunteers had wells that tested positive for pesticides.

percent of homeowners with very hard water identified

it as a problem with their water supply. Iron was also
reported as a problem by 21 percent of the study participants. Although this study did not measure iron to
confirm its presence, the 21 percent that reported having it in their water agrees with the percentage found to
actually have high iron by Sharpe et al. (1985) suggesting a high awareness of this problem.

One measure of the success

of education is to compare
actions taken by MWON
volunteers who had attended
an eight-hour training workshop with homeowners who
did not receive this training.
MWON volunteers were
much more likely to have a
sanitary well cap on their well
and the ground sloping away
from the well casing in all
directions to prevent surface
water contamination. Both of
these actions were stressed
during their volunteer training
workshop and, in fact, those
completing the training were
given a sanitary well cap for
their home well. Additional
information on actions taken
and their effect on drinking water quality were measured through the follow-up survey and re-testing of a
subset of wells.
Follow-Up Survey Results
Measurement of actions taken by homeowners to
solve or treat water quality problems was accomplished
through a follow-up survey of the 450 well owners
that participated in the study in 2006. A summary of
the nearly 300 responses to this follow-up survey is
shown in Figure 8. About half of the respondents took

Solving Water Quality Problems

The goal of any drinking water education program is to
facilitate actions taken by the audience to correct water
quality problems and, ultimately, to increase
Figure 8. Percent of private well owners that took various actions
the percentage of participants that avoid unsafe
to improve their private well as a result of having their well tested
drinking water. Actions taken by well owners to
(based on the follow-up survey sent to the 450 well owners that
solve problems could include using a new water
participated in 2006).
source (bottled water), maintaining the water
supply to remove contaminants (shock chlorinating a well to kill bacteria), removing a source
of contamination (moving a dog kennel away
from a well), or installing a continuous water
treatment device to remove the pollutant. In this
study, there were several measures that provided
some evidence of the willingness of well owners to take actions to properly manage their well
and the effectiveness of these actions given some
awareness of the problems and potential solu1
Includes shock chlorination, sanitary well cap installation, extending well
tions. Actions that homeowners could take to
casing
above ground, sloping ground to prevent contamination, or grouting
solve water quality problems were provided in
around existing well casing. 2 Includes reduced chemical and fertilizer use,
Penn State Cooperative Extension fact sheets that removing animals from near well, diverting runoff or maintaining septic
were enclosed with each water test report.
system. 3Some homeowners took more than one action.
Drinking Water Quality in Rural Pennsylvania

17

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

concentrations. Wells that tested negative

for bacteria without any actions by the
homeowner had very low pre-existing
bacteria concentrations (average = 3
colonies per 100 mL).
The re-sampling of wells with coliform bacteria between 2006 and 2007
results in a very conservative estimate
of the overall success of actions (shock
disinfection, sanitary well caps, etc.)
1
in removing bacteria from wells (10
Includes use of bottled water and/or installation of proper treatment equipment. Results for coliform bacteria and E. coli include an additional 10 percent removal based
percent) as the difference in success
on the effectiveness of other actions (shock chlorination, sanitary well cap, etc.) estirates between those that took actions (35
mated from the re-sampling of 60 wells. 2 Note that the total avoiding a contaminant
percent) and those that did not take ac(column three) does not sum from columns one and two because some well owners
tions (25 percent). A similar success rate
simply improved on actions that were already having an effect.
of 10 percent was found for removal of
E. coli bacteria from contaminated wells
an action to better manage their private well as a result
in
this
study.
The
10 percent success rate shown in the
of participating in this study. Homeowners with wells
retesting is similar to the 15 percent success reported by
that failed at least one drinking water standard were
Swistock and Sharpe (2005) in a study of 17 wells that
more than twice as likely to take corrective actions on
were shock chlorinated and fitted with sanitary well
their water supply compared to homeowners that had
caps.
no water quality problems. Overall, an impressive 76
The survey data compiled during the overall study
percent of homeowners with wells that failed at least
showed that well owners who were aware of a healthone health-based drinking water standard took at least
based pollutant in their water were very likely to avoid
one action (some took numerous actions) to correct or
the problem. For example, over 75 percent of the well
better manage the problem.
owners that knew they had E. coli bacteria had already
installed a disinfection treatment system or were using
Effect of Actions on Bacterial Contamination
bottled water. The barrier to avoidance of contaminated
Sixty wells that tested positive for coliform bacteria
well water on a large scale was the large percentage of
in 2006 were re-sampled in 2007; 32 homeowners had
well owners that were not aware of problems due to
taken at least one action, such as shock chlorination,
installing a sanitary well cap or removing the source of inadequate water testing.
The well owners that participated in this study were
contamination, to solve their wells bacteria problem.
made
aware of problems that occurred in their water
While many homeowners took recommended actions
supply. But, would this awareness translate into acto remove bacteria from their drinking water well, of
tions to avoid exposure to health-based pollutants?
greater interest is how successful these actions were in
The data presented in Figure 9 illustrate the overall
eliminating bacteria from these wells. Of the 32 wells
where homeowners took actions, 21 still contained coli- impact of this well testing study on reduced exposure
of homeowners to health-based water pollutants. The
form bacteria representing a 35 percent success rate.
first column shows the percent of well owners that
Given the variability of bacteria in wells in response
had each contaminant and were avoiding the problem
to moisture conditions discussed earlier, it could be
prior to this study. Keep in mind that many of the well
expected that some wells with bacteria would test
negative one year later simply due to changing weather owners that were avoiding contaminants prior to this
study were only doing so inadvertently. For example,
conditions or other factors. Of the 28 wells that contained coliform bacteria in 2006 where homeowners did well owners using only bottled water because their
well water tasted bad would also be avoiding coliform
not take action to eliminate bacteria, 21 still contained
bacteria even if they did not know it occurred in their
coliform bacteria when re-tested in 2007. Thus, even
well. The second column shows the percent of well
in wells without any actions taken, 25 percent did not
owners that started or improved on a method to avoid
contain bacteria one year later. A notable difference
exposure to each pollutant as a result of this study. For
between these wells and the wells where actions were
example, some well owners that were originally using
successful in eliminating bacteria was the bacteria
bottled water installed water treatment to improve water
18

The Center for Rural Pennsylvania

quality after receiving their water test report. The final

column shows the percent of homeowners that were
avoiding the problem six to 12 months after this study.
Note that the percent avoiding water quality problems
increased dramatically in each case. Pollutants with
more severe or better documented impacts on human
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, which have
relatively low risk or only affect certain portions of the
population, had slightly lower avoidance rates. Since
relative risk information was given to each well owner
in the Penn State Cooperative Extension fact sheets
included with each water test report, it is not surprising
that well owners responded with greater actions for pollutants with greater risks. For pollutants that were not
health-related, like hardness, the percent that chose to
avoid the problem generally dropped to near or below
50 percent (data not shown).
It is important to note that many well owners took
actions in an attempt to avoid exposure to pollutants
that could not be clearly linked to a reduced exposure
to that pollutant. For example, while only 50 percent of
well owners with coliform bacteria were successfully
avoiding bacterial contamination by the end of this
study, many of the remaining 50 percent had at least
taken some action to try to avoid the problem. As a
result, the percent of homeowners that took no action at
all was generally quite low ranging from 8 percent for
E. coli bacteria to 20 percent for various other pollutants. In most cases, the homeowners that did not take
any action may have made this choice because of lower
perceived risk. For example, those that did not take action to remove coliform bacteria nearly always had low
bacteria counts less than 10 colonies per 100 mL and
those that did not take action to remove E. coli always
had just 1 colony per 100 mL of water.
Unnecessary Water Treatment
The principal investigators have worked with private
well owners in Pennsylvania for several decades and
often encounter homeowners that have been sold unnecessary water treatment equipment. Data collected
during this study provides some insight about the use of
water treatment equipment in Pennsylvania.
Of the 701 homes participating in this study, 288
employed a total of 372 water treatment systems. Of
these systems, 18 percent were considered unnecessary based on water testing conducted during the study
and information provided by homeowners. This study
demonstrated that 58 homes, or 8 percent, had water
treatment equipment that was apparently not needed.
Drinking Water Quality in Rural Pennsylvania

This figure agrees with recent data gathered in central

Pennsylvania (CWS, 2005). Information provided by
homeowners suggests they are often unaware of their
specific needs with regard to water treatment. Specific
examples of questionable water treatment equipment
installations are described below:
46 homeowners with soft raw well water had a
water softener installed. Nineteen of those homes
had zero hardness, indicating that the water sample
was taken post-treatment. Iron was a problem for
an additional nine homes, which may have been the
reason for using this type of water treatment equipment. Of the remaining 18 homes, three indicated the
equipment existed when they purchased the home,
nine indicated it was installed because of past water
test results, and six listed various reasons that did not
warrant this type of treatment equipment.
50 homeowners had an ultraviolet light installed on
their water system to kill bacteria but 48 percent (24
systems) 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
in their home when it was purchased, seven had it
because of water testing they did in the past, and the
remaining 10 homeowners had various reasons that
did not justify using an ultraviolet light.
16 homes treated their water with chlorination
systems. Most of these homeowners had a bacteria
problem or indicated their water was high in iron
or hydrogen sulfide (which can both be treated with
chlorine). The remaining three homeowners did not
need chlorine although one thought he had a bacteria
problem. Of the three homes with unnecessary chlorination systems, two installed the equipment due to
the results of previous water testing and one installed
it due to taste or odor problems.
Another 20 water treatment systems were found to
be unnecessary based on the problems reported by the
homeowner or by the water testing done in this study.
Equipment that appeared to be unnecessary included
oxidizing filters, nitrate removal systems, acid neutralizing filters and magnetic treatment devices.

Well Owner Opinions

A final portion of the survey provided to each well

owner included opinion questions about perceived
threats to their water supply, opinions of well regulations, and use of other sources of drinking water.
Most well owners were very satisfied with their private water well. Eighty-four percent were not willing to
pay even $10 per month to have access to public water.
A total of 16 percent were willing to pay some monthly
19

fee to have public water

Figure 10. Well owner opinions of potential private well regulations in Pennsylvania.
including $10 (2 percent), $20
(6 percent), and $30 or more
(8 percent).
Well owners were generally
concerned about their water
quality but less so about their
water quantity. Sixty-four percent were very or somewhat
concerned about water quality
in their well while only 39
percent were similarly concerned about the amount of
water their well supplies. Well
owners were most concerned
about new housing developments as a threat to their water
(Well Testing) There should be a statewide regulasupply. Thirty-five percent of well owners ranked new
tion requiring private water wells to be tested for
housing developments as the biggest threat. Agriculture
certain pollutants prior to the sale of the home to a
(16 percent), oil and gas drilling (13 percent), mining
new owner even if it adds several hundred dollars to
(11 percent) and highways (4 percent) were the other
the cost of a home sale.
common responses for threats to private wells. There
(Driller Certification) There should be a statewide
were obvious regional differences in these responses
regulation requiring that professional water well drillwith agricultural concerns centered in southcentral
ers be periodically certified by passing a competency
Pennsylvania, mining concerns mostly in western Pennexam.
sylvania and oil and gas concerns mostly in northern
A summary of opinions from homeowners is given in
and western counties.
Figure 10. MWON volunteers are not included in this
Fourteen percent of well owners use bottled water as
their sole source of drinking water. Another 30 percent oc- summary since they were nearly unanimous in their
support of these measures as a result of their volunteer
casionally drink bottled water at home. Bottled water use
training.
was significantly higher in the northcentral region of the
Well owners were generally supportive of all potenstate and significantly lower in the southcentral region.
tial
well regulations with 63 to 78 percent strongly or
Nine percent of well owners use roadside springs
somewhat
agreeing with the statements above. The
for their drinking water; 8 percent occasionally and 1
greatest
support
was for proper location of new wells
percent for all of their drinking water. Roadside spring
use was strongly regional with more well owners in the and well driller certification. In general, the northeast
region of the state had the most favorable opinion of the
northcentral and northwest regions using them comproposed regulations while the northwest and southwest
pared to other regions.
Four questions on the survey asked well owners their regions of the state were the least in favor of any well
regulations.
opinions about proposed private well regulations in
Well owners that have a question or problem related
Pennsylvania, as follows:
to
their well were most likely to seek assistance from
(Well Location) There should be a statewide regulaprivate
water labs (22 percent), Penn State Cooperative
tion on the location of new water wells that includes
the minimum distance a new well can be drilled from Extension (19 percent), a well driller (11 percent) or the
Pennsylvania Department of Environmental Protection
existing wells, septic systems and other sources of
(11 percent).
contamination.
The well owner opinions, including support for pri (Well Construction) There should be a statewide
vate well regulations, found in this study mirror results
regulation that requires certain well construction
found from an online survey of 865 private well owncomponents on new wells (such as a sealed well cap
ers from 63 counties in Pennsylvania conducted by the
and cement-like seal or grout around the casing) to
researchers in 2006.
prevent well contamination even if it adds $500 to
$1,500 to the cost of a new well.
20

The Center for Rural Pennsylvania

Conclusions

Data from this study provide a wealth of information

on the incidence of pollutants in private water wells
throughout Pennsylvania, the causes of contamination,
and the ability of well owners to detect and solve water
quality problems voluntarily.
About 41 percent of samples from the private wells
tested in this study failed at least one safe drinking
water standard. Overall, the prevalence of contamination was stable or declining compared to past results for
the parameters measured in this study. Lead contamination appeared to be declining in response to the 1991
Federal Lead and Copper Rule while nitrate contamination was reduced from the early 1990s, presumably due
to reduced applications of nitrogen through fertilizers
and manures.
Of the 28 variables measured for each well, the
results demonstrated that natural variables, such as the
type of bedrock geology where the well was drilled,
were important in explaining the occurrence of most
pollutants in wells. Soil moisture conditions at the time
of sampling were the single most important variable in
explaining the occurrence of bacteria in private wells.
Mans activities, however, were also responsible for the
increased incidence of some contaminants. Inadequate
well construction was strongly correlated with the occurrence of both coliform and E. coli bacteria in wells.
Nearly all lead contamination could be attributed to the
historical use of lead plumbing components and the occurrence of naturally corrosive groundwater. Increased
nitrate concentrations were strongly related to the
location of the well in comparison to nearby agricultural fields. Overall, these results suggest that naturally
occurring groundwater is not always safe for human
consumption and mans current and past activities have
worsened the situation for some pollutants.
Unsafe levels of these pollutants in wells can be addressed through maintenance, water treatment devices,
bottled water or new sources of water. But a major
barrier to successful avoidance of problems that was
identified in this study was simply creating awareness
of problems through water testing. Most health-related
pollutants do not have obvious symptoms (tastes, odors,
etc.) so few wells have been properly tested. The results
indicate that only about half of well owners have had
their well water tested by a certified laboratory and,
in many cases, they have not been thoroughly tested.
The lack of testing by well owners is not for a lack of
concern over their water quality, but instead, a lack of
awareness and understanding of what testing should be
done.
Drinking Water Quality in Rural Pennsylvania

Overcoming the barrier to water testing is especially

critical because results from this study have clearly
demonstrated that well owners are willing to take actions to solve water quality problems. The great majority of well owners that were told of health-related water
quality issues in their water supply had voluntarily
solved the problem within one year.

Policy Considerations

To ensure safe drinking water for private water system owners in Pennsylvania, the state should consider
using a combination of educational programs for homeowners and new regulations. The combination of these
two approaches will make certain that existing well
owners will become aware of water quality issues and
proper management of wells while future well owners
will be protected from poorly protected wells or lack
of knowledge. Specific policy recommendations are
addressed below.

Well Testing

One of the most revealing results from this study

was the lack of voluntary water testing done by private
water well owners and the resulting lack of awareness
of health-related water quality problems. Several states
now require testing of private water wells during real
estate transactions but, in Pennsylvania, water testing is entirely the voluntary responsibility of a well
owner. Well owners could benefit from a combination
of educational programs and regulations targeting water
testing. Requiring well owners to have their water
tested routinely is not practical or enforceable given the
large target audience in Pennsylvania. Instead, a better
approach would be to require all homes with a private
water well to be tested by a certified laboratory at the
completion of new well construction and before finalization of any real estate transaction. Additional educational programs for the remaining well owners that do
not buy or sell their home regularly are also needed (see
Existing Well Owner Education on the next page).

Well Construction

The importance of well construction has been considered self-evident among groundwater experts as a
means to prevent surface water (and associated pollutants) from entering groundwater aquifers. In fact, the
state association of professional well drillers (Pennsylvania Ground Water Association) has supported well
construction standards for many years and most states
currently have well construction standards. By sam-

21

pling a large number of wells with diverse settings and

construction, the researchers were able to document the
significant benefit of proper well construction in reducing the incidence of both coliform and E. coli bacteria
in wells. While this study showed that education increased the use of sanitary well caps on existing wells,
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, as a result, the voluntary approach to
encourage proper well construction has largely failed.
Given the benefits of well construction and the difficulty in reaching the target audience for new wells,
statewide regulations requiring well construction
components appear to be warranted. This study clearly
demonstrated that the majority of well owners (>60
percent) were in favor of well construction and location
standards.

Wellhead Protection

In comparison to natural influences and well construction features, nearby land-uses were less important
in explaining contamination of private wells. For example, the distance to a nearby on-lot septic system and
the frequency of septic tank pumping were not statistically significant in explaining the occurrence of bacteria or nitrate in private wells. Instead, nearby sources of
animal waste appeared more important in contributing
to bacterial contamination of wells. A notable exception to the lack of importance of nearby land-uses
was the strong correlation between the distance to
nearby agricultural fields and nitrate concentrations in
wells. Given the conservative nature of nitrate (ability to move long distances through soil and rock), this
correlation is understandable. Overall, the results of
this study do not make a strong case for the need for
wellhead protection areas around private wells. However, the researchers recommend that the determination for the location of new water wells should still be
considered very important. In most cases, voluntary
wellhead protection areas already existed around the
private wells in this study as a result of common sense
location of wells used by well drillers and homeowners. The fact that few wells were located very close
(<50 feet) to sources of contamination made it difficult
to determine if minimum wellhead protection areas are
warranted. The connection between contaminated wells
and nearby land-use activities may be underrepresented

22

in this study since the vast majority of homeowners did

not have wells located in close proximity to activities
that generate pollution. As a result, the data seem to
confirm the importance and success of de facto wellhead protection areas of 50 to 100 feet that already exist
around most wells.

Existing Well Owner Education

The aforementioned regulations and educational

programs are largely focused on new well construction or change in ownership of existing private wells.
There is also a clear need for an educational component
on improving drinking water quality for homeowners with existing private wells. This study has very
clearly demonstrated the effect that education can have
to increase the frequency of water testing, the use of
certified labs and awareness of water quality problems.
Related to this was the impressive percentage (>75
percent) of well owners that were willing to take action
to treat their water or use bottled water once they were
aware of a health-related problem in their water and
the large percentages (50 to 80 percent) that were able
to avoid unsafe water through voluntary actions taken
during the study. A well designed and collaborative
education program to deliver unbiased education to
well owners could be successful if adequately funded.
Such a program should be coordinated among government agencies and businesses that were identified by
well owners as important sources of information (water
testing labs, Penn State Cooperative Extension, well
drillers and Pennsylvania Department of Environmental
Protection). A specific component to this educational
effort should be an emphasis on bacteria testing during
wet weather conditions to improve detection of wells
with bacteria problems.

The Center for Rural Pennsylvania

References

Bartholomay, 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 Controls National Environmental Public
Health Tracking Program: U.S. Geological Survey Scientific Investigations Report 20075213, p. 57.
Bickford, T.M., B.D. Lindsey, and M.R. Beaver. (1996) Bacteriological Quality of Ground Water Used for Household
Supply, Lower Susquehanna River Basin, Pennsylvania and Maryland. U.S. Geological Survey Water-Resources
Investigations Rep. 96-4212, p. 31.
Centers for Disease Control and Prevention (CDC). (1998) A Survey of the Quality of Water Drawn from Domestic
Wells in Nine Midwest States.
Center for Watershed Stewardship (CWS). (2005) Spruce Creek Watershed Assessment and Stewardship Plan, Keystone
Project 2004-2005, Phase II. pp. 42-49.
Clemens, S., B.R. Swistock and W.E. Sharpe. (2007) The Master Well Owner Network: Volunteers Educating Pennsylvania Well 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) Iowas 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 Investigations Report 01-4268.
Liu, A., J. Ming and R. Ankumah. (2005) Nitrate Contamination in Private Wells in Rural Alabama, United States. Science of the 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 Resources
Bulletin, 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 Safe Drinking Water: p. 65.
Rowe, B.L., P.L. Toccalino, M.J. Moran, J.S. Zogorski and C.V. Price. (2007) Occurrence and Potential Human-Health
Relevance of Volatile Organic Compounds in Drinking Water from Domestic Wells in the United States. Environmental Health 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 Private
Individual 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 Private
Individual 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 Using Satellite 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.
The Center for Rural Pennsylvania. May 2004.
Swistock, B.R. and W.E. Sharpe. (2005). The Influence of Well Construction on Bacterial Contamination of Private Water Wells 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 Users Guide to the United
States Environmental Protection Agencys 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 816F-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 Occurrence of Bacteria in Private Household-Supply Wells, South-Central and Southeastern Pennsylvania. U.S. Geological Survey Water-Resources Investigations Report 01-4206.

The Center for Rural Pennsylvania

Board of Directors
Senator John Gordner
Chairman
Representative Tina Pickett
Vice Chairman
Senator John Wozniak
Treasurer
Dr. Nancy Falvo
Clarion University
Secretary
Representative Tim Seip
Dr. Theodore R. Alter
Pennsylvania State University
Steve Crawford
Governors Representative
Dr. Stephan J. Goetz
Northeast Regional Center for Rural Development
Dr. Keith T. Miller
Lock Haven University
Dr. Robert F. Pack
University of Pittsburgh
William Sturges
Governors Representative

The Center for Rural Pennsylvania

200 North Third St., Suite 600
Harrisburg, PA 17101
Phone: (717) 787-9555
Fax: (717) 772-3587
www.ruralpa.org
1P0109-450