E41
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
Understanding the causes for water system failure
JEFF L. OXENFORD1 AND SHARON I. WILLIAMS2
1Oxenford
2City
Consulting LLC, Golden, Colo.
of Westminster, Colo.
To understand the causes for water system failure and to better
direct limited assistance resources, the Colorado Safe Drinking
Water Program in 2009 conducted an analysis of public water
system data. The goal was to use data that are routinely
collected by the program to determine the root causes for
system failure. Three data sets were analyzed: data tracked by
the program’s Acute Team, data collected during sanitary
surveys, and self-reported compliance data. The results from
this study clearly pointed to a number of issues that most
frequently lead to critical system failure. The data are being
used by the program to focus assistance resources and conduct
additional investigations.
Keywords: chlorine residual, compliance, cross-connections, disinfection, distribution, distribution system, regulations, sanitary survey,
significant and minor deficiencies, violations, water treatment
Public water systems have the responsibility of maintaining
physical and human infrastructure to supply safe drinking water
to the public and to comply with state and federal regulations
that govern drinking water. Colorado is home to more than 2,000
public water systems, and the Colorado Safe Drinking Water
Program (SDWP) is responsible for maintaining the state’s regulatory infrastructure and monitoring and enforcing compliance
with regulations. In addition to conducting regulatory enforcement activities, program staff routinely assist public water systems
in understanding their roles and responsibilities. Staff also provide
capacity-building support through training, technical assistance,
and management support services.
This project was conducted with the Capacity Building Unit of
the Colorado Department of Public Health and Environment
(CDPHE) to identify and illustrate the most common failures
observed by community water systems within Colorado. Failures
were defined as
• the inability to meet maximum contaminant levels (MCLs),
reporting requirements, or other requirements of the Colorado
Primary Drinking Water Regulations (CDPHE, 2010),
• identification of significant and minor deficiencies during
sanitary surveys, or
• events that were brought to the attention of CDHPE that had
the potential for acute health effects.
Failure as defined for this project does not necessarily equate
with a measured public health outcome. However, if these failures were not addressed they had the potential to lead to public health problems. For example, failing to monitor does not
mean that there is contamination in the system, only that if
there was contamination it would go undetected; identification
of a cross-connection does not mean that backflow has
occurred, only that it could occur. Data in this study also
JOURNAL AWWA
included the waterborne outbreak of salmonella that occurred
in Alamosa, Colo. (Falco & Williams, 2009).
It was hoped that by identifying and illustrating the most common failures, significant improvements could be made in public
water system operations, regulatory compliance rates, and overall public health. Also, the unit could focus limited resources on
the most critical issues facing public water systems in Colorado.
APPROACH
This project was initiated in September 2008 and concluded in
June 2009. The project used a modified Pareto analysis to evaluate
water system failures. The Pareto principle, commonly referred
to as the 80/20 rule, has been applied in a variety of industries to
quickly set priorities and focus on the activities that would have
the highest return on investment. The concept is that 80% of the
benefit can be generated through 20% of the effort. To address
the remaining 20% of the benefit would require 80% of the
effort. By applying the Pareto principle, resources could be
targeted to address the most significant issues that cause system
failures and to most effectively use limited resources.
For this study, the Pareto approach was modified such that a
strict adherence to demarcating 80% of all the failures was not
a requirement of the analysis. Instead, the rate of failure was
graphed and the top categories were identified and prioritized.
This simpler approach appeared well-suited for the dataset and
resulted in more focused recommendations. This approach was
used to prioritize resources of the program as a whole. Failures
in the 20% area would still be addressed on a case-by-case basis.
DATA SOURCES
Data analyzed in this study were from a three-year period,
fiscal years 2006–08 (July 1, 2005–June 30, 2008). All data
2014 © American Water Works Association
JANUARY 2014 | 106:1
E42
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
came from existing programs within CDPHE and were used as
provided by the program.
The first data source was from the program’s Acute Team tracking system. The Colorado Drinking Water Program’s Acute Team
comprises key staff members who respond to water system issues
that could pose an acute risk to the public and therefore need an
immediate response from the state. The Acute Team becomes
involved when monitoring data show an acute health risk, when
certain deficiencies are identified in a sanitary survey, when a
public water system asks for help in an emergency, or when the
program learns of an acute risk through other means. Data from
all Acute Team actions were collected and stored in a spreadsheet
that was maintained by the program.
The second set of data consists of significant and minor deficiencies identified by CDPHE inspectors during sanitary surveys.
These deficiencies have been defined by the US Environmental
Protection Agency (USEPA):
• Significant deficiency means any situation, practice, or condition in a public water system with respect to design, operation,
maintenance, or administration that the state determines may
result in or have the potential to result in the production of finished drinking water that poses an unacceptable risk to the health
and welfare of the public served by the water system.
• Minor deficiencies do not pose an immediate risk to human
health but if left unaddressed for a long period of time, could get
worse and cause an adverse health effect. These are often related
to system management and planning.
A sanitary survey is an onsite review of a public water system’s
water source, facilities, equipment, operation, and maintenance.
Eight areas within a water system are evaluated: water sources;
treatment; distribution systems; finished water storage; pumps,
pump facilities, and controls; monitoring, reporting, and data
verification; water system management and operations; and
operator compliance with state requirements.
Sanitary surveys are required by the Safe Drinking Water Act
(SDWA) to be conducted a minimum of every three years for
community surface water systems and every five years for community groundwater systems and noncommunity systems. (This
study was conducted with regulations at the time [2005–08] and
does not account for changes in the Total Coliform Rule, Ground
Water Rule, Surface Water Treatment Rule, and others.) In Colorado, however, many small noncommunity groundwater systems
are inspected more frequently (as often as every year) by local
health departments and SDWP staff.
Data are stored in the Safe Drinking Water Information System
(SDWIS)—a national database administered by USEPA—where
individual states are required to collect and input data (SDWIS/
State). Data are categorized by 80 deficiency codes that are
assigned by the inspector. Until 2008, data collected by local
health departments were not routinely entered in the SDWIS/State
database. Therefore, sanitary survey data from small noncommunity groundwater systems are limited in this study.
The final set of data used consisted of compliance data. These data
are routinely submitted to CDPHE by water systems at frequencies
specified in the regulations. Data on violations and enforcement action
related to the SDWA Regulations are stored in the SDWIS database.
JOURNAL AWWA
APPROACH
The approach for this project consisted of three tasks.
Task 1—Data collection. Data were extracted from the SDWIS
database and Acute Teams spreadsheets from state fiscal years
2006–08 (July 1, 2005–June 30, 2008). At the time this study
was conducted, this represented the latest complete years of
data for analysis. Spreadsheets were prepared for the different
data sources and included
• Acute Team reporting history,
• major deficiencies identified in sanitary surveys,
• minor deficiencies identified in sanitary surveys, and
• compliance history.
Task 2—Data analysis. The second major task was to analyze
the data to identify the most frequent failures and the highest-risk
failures in water systems. The data were compiled in graphs and
tables using Pareto analysis techniques, with frequency of failure
graphed against failure type. The classification of the failures
came directly from the data sources. This information was then
visually assessed to determine categories of failures that contribute most significantly to compliance rates, sanitary survey deficiencies, and Acute Team responses.
Task 3—Recommendations. The final task was to use the results
of the analysis to develop a set of recommendations for future
program efforts to promote improvements in water system
operations and management. Building recommendations based on
the Pareto analysis were designed to reflect a certain transparency
in decision-making so that interested stakeholders and internal staff
have a clear understanding of why and how future program
priorities and operations will be established.
RESULTS
In general, Colorado has a strong history of compliance with
health-based standards, as indicated in annual reports prepared
by the SDWP. This study went beyond compliance rates to include
both health-based and nonhealth-based violations and data
collected about operational failures and system deficiencies. The
goal was to identify types of failures that could potentially lead
to more significant issues in the future.
Events reported in the Acute Team tracking database. The Acute
Team tracking database included 156 acute events reported
during the three-year study period (2006–08). Acute events were
organized according to frequency and the category of failures
observed and then compared with the type of response action
warranted by the failure (Figure 1), type of public water system
affected, and the duration of the event response (Figure 2). The
acute events reported in the Acute Team database were organized
into the following categories:
• Distribution System Deficiency,
• Treatment System Deficiency, Surface Water (SW),
• Bacteriological (bacti)/Microscopic particulate analysis
(MPA) Results Causing Acute Situation,
• Treatment Deficiency, Groundwater (GW),
• Untreated Groundwater (GW) Intended for Drinking,
• Untreated Surface Water (SW) Intended for Drinking,
• Tampering/Security,
• Source Water Supply Disruption.
2014 © American Water Works Association
JANUARY 2014 | 106:1
E43
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 1 Failure classification by response action
50
Boil water order = 47
Bottled water order = 52
Other public notice = 21
No further action = 36
45
Acute Team Incidents*—number
40
9
35
8
6
30
25
9
16
20
4
9
1
2
15
3
10
16
12
5
2
3
9
1
12
9
5
0
Distribution
System
Deficiency
Treatment
Deficiency,
SW
Bacti/MPA
Results
Causing
Acute
Situation
Treatment
Deficiency,
GW
Untreated
GW
Intended
for Drinking
6
1
Untreated
SW
Intended
for Drinking
3
1
2
1
Tampering/
Security
2
1
3
Source
Water
Supply
Disruption
Failure Classifacation
Bacti/MPA—bacteriological/microscopic particulate analysis, GW—groundwater, SW—surface water
*Total = 156
(These categories align in general with failure classifications
used by the Centers for Disease Control and Prevention when
reporting on the underlying causes of waterborne disease and
outbreaks associated with drinking water in biennial
surveillance reports.)
In Figure 1, the number and categories of failures observed
were compared with their response actions. This chart illustrates
the utility of a Pareto analysis. Although the Acute Team responds
to a variety of events, this analysis shows that the top two failures
requiring a response were
• Distribution System Deficiency (such as pipe breaks, leaks,
and power outages) and
• Treatment Deficiency (such as turbidity spikes, chlorinator
failures, and positive coliform results).
Together, these two categories comprised more than half of all
Acute Team responses. By focusing resources on preventing
distribution system and surface water treatment deficiencies, the
SDWP has the potential to reduce internal resource demands and
JOURNAL AWWA
produce a significant and measurable overall benefit to public
water systems in Colorado.
Figure 2 provides a comparison among the categories of failures
observed and the duration of the response as of October 1, 2008.
As shown, systems with untreated groundwater or surface water
intended for drinking had the longest order durations in response
to acute events. The average duration of most bottled water orders
was less than 20 days, whereas the average duration of bottled
water orders for untreated groundwater or surface water was 300
and 362 days, respectively. This was probably because of the long
lead time to design and construct infrastructure improvements to
provide proper treatment where none existed.
In addition to the data analysis just provided, several
observations and insights were noted regarding the Acute Team
tracking database and the nature of acute responses. The
observations and insights listed subsequently show the effect of
acute response events and the usefulness of compliance data in
flagging areas of concern.
2014 © American Water Works Association
JANUARY 2014 | 106:1
E44
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 2
400
Failure classification by order duration
Boil water order
Bottled water order
350
Acute Team Incidents—days
300
250
200
362
150
300
124
100
13
50
9
38
0
Distribution
System
Deficiency
55
Treatment
Deficiency,
SW
64
18
10
Bacti/MPA
Results
Causing
Acute
Situation
5
Treatment
Deficiency,
GW
Untreated
GW
Intended
for Drinking
Untreated
SW
Intended
for Drinking
Tampering/
Security
4
12
Source
Water
Supply
Disruption
Failure Classifacation
Bacti/MPA—bacteriological/microscopic particulate analysis, GW—groundwater, SW—surface water
Data compiled on Oct. 1, 2008, for the period of record of July 1, 2005–June 30, 2008
• One hundred twelve (72%) of all 156 acute responses
required Tier 1 public notification, which involves notifying the
public of the condition as soon as practical but no later than 24
hours after the system operators learn of the condition.
• Ninety nine (63%) of all 156 acute responses required a boil
or bottled water order.
• Nearly 60,000 people were affected by boil or bottle water
orders. The spring 2008 salmonella outbreak in Alamosa, Colo.,
which affected 9,200 people, was the largest boil or bottled water
event during this period.
• Thirteen (68%) of the 19 systems that reported bacti/MPA
acute failures had at least one violation before the acute event
(not necessarily a Total Coliform Rule [TCR] violation).
• Four (21%) of the 19 systems that reported bacti/MPA acute
failures had TCR monitoring and reporting violations following
the event.
• Six (40%) of the 15 systems that had an acute response to
turbidity spikes had reported turbidity violations before acute
violations. Four of these systems were using bag filters.
• Seventy percent of all distribution system deficiencies were
caused by loss of system pressure from main breaks, loss of
power, or other infrastructure failures.
JOURNAL AWWA
Sanitary survey deficiencies. The SDWIS/State included data for
a total of 1,942 sanitary surveys conducted during the three-year
study period (2006–08). Sanitary surveys may have identified
multiple significant or minor deficiencies, and the data was organized into eight categories. These categories were based on the
eight components of a sanitary survey provided in the Colorado
Primary Drinking Water Regulations as follows.
• Treatment
• Distribution System
• Monitoring, Reporting, and Data Verification
• Source
• Finished Water Storage
• System Management and Operation
• Pump/Pumping Facility and Control
• Operator Compliance with State Regulations
Within each category, violation codes were assigned by
inspectors. There are more than 80 potential violation codes.
Significant deficiencies. During the three-year study period,
significant deficiencies were found during 182 (9%) of the
1,942 sanitary surveys reported in SDWIS. In these systems,
there were 353 separate significant deficiency observations, or
an average of nearly two significant deficiencies per deficient
2014 © American Water Works Association
JANUARY 2014 | 106:1
E45
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 3 Sanitary survey significant deficiencies by system type
120
Community = 193
Transient noncommunity = 111
Nonpublic = 25
Nontransient, noncommunity = 24
6
Significant Deficiencies*—number
100
80
7
32
60
5
5
4
1
6
4
40
66
16
14
1
26
20
16
Treatment
Distribution
System
2
13
5
5
5
17
17
20
Source
Finished
Water
Storage
System
Management
and
Operation
29
28
0
3
Operator
Compliance
With State
Requirements
Monitoring and
Reporting and
Data
Verification
Area of Deficiency
Data compiled on Oct. 1, 2008, for the period of record of July 1, 2005–June 30, 2008
*Total = 353
system. Most of the systems that had significant deficiencies
also had minor deficiencies.
For this project, the data on significant and minor deficiencies
were evaluated using a Pareto analysis to identify specific issues
that could be prioritized for maximum potential benefit from the
use of resources.
First, an analysis was conducted to determine the top 281
(80%) of all significant deficiencies identified during the study
period (before grouping by category). The most common
significant deficiency codes were associated with a lack of a
certified operator, failure to obtain design approval, deficient
disinfection equipment, and inadequate disinfection residual in
the distribution system. (Improvements with operator certification
requirements and enforcement were already under way during
the period of this study. These activities have greatly improved
compliance with operator certification requirements. Thus,
operator certification itself was not further addressed in depth in
this work.) In Figure 3 the significant deficiencies were grouped
by category, and then the number and categories of deficiencies
JOURNAL AWWA
were compared with the type of public water system where the
deficiencies were observed. As shown, the most common
significant deficiencies were associated with treatment processes
(111 or 31%), distributions systems (54 or 15%), operator
compliance with the regulations (52 or 15%), and monitoring,
reporting, and data verification (48 or 14%).
The majority of significant deficiencies occurred in community
water systems; however, this may have been in part an artifact of
the data-collection process. Sanitary surveys are conducted a
minimum of every three years for community surface water systems and every five years for community groundwater systems
and noncommunity systems. In addition, although small noncommunity groundwater systems are inspected more frequently (as
often as every year) by the SDWP in partnership with local health
departments, the results of those surveys were not routinely
entered into SDWIS/State during the majority of the study period.
As of 2008, these data were being entered into SDWIS/State;
however, a skew in the data toward community systems still
existed for the study period.
2014 © American Water Works Association
JANUARY 2014 | 106:1
E46
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 4
Sanitary survey deficiency classification by source type
120
Groundwater under the influence of surface water = 50
Groundwater = 194
Surface water = 59
Purchased surface water = 49
1
100
Significant Deficiencies*—number
31
80
60
60
13
6
19
40
7
7
1
20
30
27
29
1
3
21
19
0
4
Treatment
Distribution
System
3
Operator
Compliance
With State
Requirements
7
9
Monitoring and
Reporting and
Data
Verification
Source
4
3
5
17
10
3
Finished
Water
Storage
7
5
System
Management
and
Operation
Deficiency Classification
*Total = 352
Overall, approximately 74% of public water systems in
Colorado are groundwater systems and the remaining percentage
is surface water or groundwater under the direct influence of
surface water (GWUDI). Figure 4 provides a comparison between
the number and categories of significant deficiencies and the
source of the drinking water for the system where the deficiency
was observed. As shown, 194 (55%) of all significant deficiencies
were associated with groundwater systems, 60 of which were
treatment-related deficiencies.
Given that treatment deficiencies were the most common type
of significant deficiency observed (Figures 3 and 4), the nature of
treatment deficiencies was further evaluated. In Figure 5, all
observed treatment deficiencies (111 total) were evaluated based
on unique treatment-related deficiency codes. As shown, the most
common treatment deficiencies were a lack of or inoperable
disinfection equipment (29 or 26%) followed by cross-connection
issues (17 or 15%). Systems with these deficiencies are at a greater
risk of a waterborne disease outbreak.
Because distribution system deficiencies ranked second highest
in the total number of significant deficiencies observed (Figures
JOURNAL AWWA
3 and 4), the nature of distribution system deficiencies was also
further evaluated. In Figure 6 all observed distribution system
deficiencies (54 total) were shown based on their deficiency codes.
The most common distribution system deficiencies were a failure
to maintain disinfection residuals (27 or 50%), followed by crossconnection issues (13 or 12%), and raw or untreated water
service (6 or 11%). These deficiencies have a very high potential
for a significant effect on public health.
Monitoring, reporting, and data verification deficiencies
ranked fourth highest in the total number of significant
deficiencies observed (Figures 3 and 4), and therefore the nature
of these deficiencies was further evaluated. In Figure 7 all
monitoring, reporting, and data verification deficiencies (48
total) are illustrated based on their deficiency codes. As shown,
the most frequent monitoring, reporting, and data verification
deficiency was failure to obtain plans and specifications
approval for the system or renovations to the system (31 or
65%). Anecdotal information suggests that many water systems,
their operators, and their design engineers did not understand
the design criteria or failed to understand the significance of
2014 © American Water Works Association
JANUARY 2014 | 106:1
E47
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 5
Sanitary survey significant deficiencies for treatment
35
25
20
15
29
†
10
‡
10
3
2
T160
T130
T110
T150
T120
T995
T901
T112
1
1
1
1
1
T996
3
T140
4
T133
4
0
T132
9
T131
12
T999
13
5
T113
17
T111
Significant Deficiencies*—number
30
Deficiency Code
CPDWR—Colorado Primary Drinking Water Regulations
*Total = 111
†No disinfection equipment present or equipment not operating. CPDWR 7.1.2 (a),(b) & 7.6.1(b);S (CDPHE, 2010).
‡System has an uncontrolled cross-connection that allows contamination to enter drinking water that will cause an immediate sanitary risk. CPDWR 12.1 (a);S
(CDPHE, 2010).
many of the provisions in the criteria. Failure to obtain proper
approval can lead to other failures at a later date.
Minor deficiencies. Minor deficiencies were found during 477
(25%) of the 1,942 sanitary surveys reported in SDWIS/State
during the three-year study period. In these systems, there were
1,286 separate minor deficiency observations. The top 1,031
(80%) of all the minor deficiencies identified during the study
period (before grouping by category) corresponded to a lack of
or inadequate planning for cross-connection control, bacti
sample-siting, monitoring, and emergency response.
In Figure 8 the minor deficiencies were grouped by category,
and then the number and categories of deficiencies were compared
with the types of public water system where the deficiencies were
observed. As shown, the most common minor deficiencies were
associated with monitoring, reporting, and data verification (425
or 33%) and system management and operation (338 or 26%).
Because monitoring, reporting, and data verification de­­
ficiencies were the most common minor deficiencies observed,
the nature of these deficiencies was further evaluated. In Figure
9, all observed monitoring, reporting, and data verification
JOURNAL AWWA
deficiencies (425 total) are shown based on their deficiency codes.
The most common monitoring, reporting, and data verification
deficiency codes were a lack of or inadequate bacti sample-siting
plan (130 or 31%), a lack of or inadequate general monitoring
plan (120 or 28%), and inadequate record keeping or document
retention (36 or 8%).
Regulatory violations. Reports from SDWIS/State included more
than 16,000 violations (largely associated with monitoring and
reporting) during the three-year study period. Of the 2,023 public
water systems in Colorado, a total of 1,286 (approximately 63%)
reported at least one violation. These water systems collectively
provide water to 47% of the population served by all public
water systems. A total of 292 water systems (14% of all systems
in Colorado) reported at least one violation in the high-priority
categories of MCL or treatment technique violations; these
violations do indicate an increased public health risk. This
information is summarized in Table 1.
In Figure 10, all violations reported in SDWIS for the threeyear study period were grouped according to the violation
categories shown in Table 1 for 10 Colorado drinking water
2014 © American Water Works Association
JANUARY 2014 | 106:1
E48
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 6
Sanitary survey significant deficiencies for distribution systems
35
30
Significant Deficiencies*—number
25
20
15
27
†
10
13
‡
5
6
§
4
2
0
D230
D901
D210
D997
D260
1
1
D280
D999
Deficiency Code
CPDWR—Colorado Primary Drinking Water Regulations
*Total = 54
†System is not maintaining the required disinfection residuals in the distribution system. CPDWR 7.1.2(c)(3)(i) & 7.6.1(c);S (CDPHE, 2010).
‡System has an uncontrolled cross-connection that allows contamination to enter drinking water that will cause an immediate sanitary risk. CPDWR 12.1 (a);S
(CDPHE, 2010)
§System has customer service connections that serve untreated drinking water. CPDWR 7.1.2(c) & 7.6.1(b);S (CDPHE, 2010).
rules. These groupings were then plotted to show the types (i.e.,
category and rule) of violations most commonly reported by
public water systems. The drinking water regulations used in
this evaluation pertain to
• inorganic contaminants (IOC),
• radionuclides,
• synthetic organic chemicals (SOC),
• volatile organic chemicals (VOC),
• consumer confidence reports,
• disinfection by-products (DBPs),
• Lead and Copper Rule,
• public notification,
• Surface Water Treatment Rule, and
• TCR.
As shown in Figure 10, the largest number of monitoring and
reporting violations occurred with the VOC, SOC, and IOC
JOURNAL AWWA
regulations. However, each of these contains a suite of
contaminants that must be monitored at the same time, and
therefore any single failure to sample/report may result in
multiple violations, one for each contaminant in the suite. No
public water systems reported MCL violations for VOCs and
SOCs; 169 MCL violations were reported for IOCs.
Figure 11 compares the number and categories of violations
with the number of public water systems in which the violations
were reported. As shown, the greatest number of individual
public water systems reported monitoring and reporting
violations of the TCR (714 or 56% of all systems with violations),
followed by the IOC violations (481 or 37% of all systems), and
the Disinfection/Disinfectants Byproducts Rule (335 or 26% of
all systems with violations).
MCL violations were most frequently associated with the TCR
(121 or 9% of all systems with violations), followed by the
2014 © American Water Works Association
JANUARY 2014 | 106:1
E49
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 7
Sanitary survey significant deficiencies for monitoring, reporting, and data verification
35
Significant Deficiencies*—number
30
25
20
15
31
†
10
5
6
4
0
R540
R997
R510
3
R514
1
1
1
1
R521
R527
R530
R999
Deficiency Code
CPDWR—Colorado Primary Drinking Water Regulations
*Total = 48
†System has not received plans and specifications approval for the system or for renovations to the system, including the addition of new sources, changes
in treatment or changes in the distribution system. CPDWR 1.11.2;S (CDPHE, 2010).
Radionuclide Rule (45 or 3% of the systems with violations), and
IOC violations (40 or 3% of the systems).
KEY FINDINGS
A modified Pareto analysis of the failure data for a threeyear study period (ending June 30, 2008) from the division’s
Acute Team tracking database and the SDWIS/State database
showed important trends in compliance failures at public
water systems in Colorado. The following are key findings
from analysis of this data.
• Acute Team data show that approximately 80% of the system failures resulting in a response were associated with
o distribution system deficiencies and
o bacteriologic result indicating an acute health concern
(such as an Escherichia coli–positive sample).
The majority of acute responses occurred at community water
systems, although this may be the result of transient and nontransient water systems failing to report to the Acute Team during
JOURNAL AWWA
the study period. Systems with failures associated with untreated
groundwater and surface water intended for drinking resulted in
the longest duration of boil or bottled water orders. Although the
full extent of the public health effect is unclear, more than 60,000
people were affected by boil or bottled water orders during acute
events during the three-year study period.
• SDWIS/State data from sanitary surveys showed that individual significant deficiencies were most commonly associated with
the lack of a certified operator, failure to obtain design approval,
deficient disinfection equipment, and failure to maintain an adequate disinfection residual in the distribution system. When
grouped into categories, significant deficiencies were most commonly associated with
o treatment (primarily inadequate disinfection equipment) and
o distribution systems (primarily failure to maintain disinfection residuals and cross-connection control).
• SDWIS/State data showed that minor deficiencies were most
commonly associated with nonexistent or inadequate cross-
2014 © American Water Works Association
JANUARY 2014 | 106:1
E50
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 8
Sanitary survey minor deficiency classification by public water system type
450
Community water system = 859
Transient noncommunity = 314
Nonpublic = 35
Nontransient, noncommunity = 78
1
400
2
350
Minor Deficiencies*—number
9
20
300
76
250
13
200
44
150
5
233
100
57
170
13
50
11
21
0
Treatment
Distribution
System
63
62
Source
Finished
Water
Storage
31
Operator
Compliance
With State
Requirements
Monitoring
and
Reporting
and Data
Verification
System
Management
and
Operation
Pump/
Pumping
Facility and
Control
Deficiency Classification
*Total = 1,286
connection control plan, bacti sample–siting plan, or monitoring
plan. When grouped into categories, minor deficiencies were most
commonly associated with
o monitoring, reporting, and data verification (primarily
lack of or inadequate bacti sample–siting plan and general
monitoring plan) and
o system management and operation (primarily failure to
maintain cross-connection control).
SDWIS/State compliance data also showed that although 63%
of all systems in Colorado reported at least one violation during
the study period, the majority of violations were monitoring and
reporting violations, which are not necessarily associated with an
increased public health risk. About 11% of the systems reported
at least one violation in the high-priority category of MCL violations, which do indicate an increased public health risk. MCL
violations were most frequently associated with the TCR, followed
by the radionuclide and inorganic contaminant violations.
JOURNAL AWWA
In summary, the data collected over the three-year period suggested the following primary reasons for observed failures in
public water systems:
• Disinfection
o Failure to maintain disinfection residuals
o Inadequate disinfection equipment
o Positive total coliform samples
• Cross-connection control
o Presence of cross-connections
o Lack of cross-connection control planning
• Distribution system operations
o Failure resulting in loss of system pressure
o Failure to maintain disinfection residuals
o Failure to maintain cross-connection control
• Failure to plan
o Lack of or inadequate bacti sample–siting plan
o Lack of or inadequate general monitoring plan
2014 © American Water Works Association
JANUARY 2014 | 106:1
E51
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 9
Minor deficiencies for monitoring, reporting, and data verification
140
130
120
120
80
†
60
‡
40
9
9
9
7
7
6
6
5
3
3
3
3
2
2
2
1
R513
9
R535
12
R533
12
R512
14
R534
§ 15
R524
20
R523
36
R517
Minor Deficiencies*—number
100
R999
R998
R522
R525
R516
R531
R527
R526
R521
R997
R511
R540
R530
R520
R510
R514
0
Deficiency Code
CPDWR—Colorado Primary Drinking Water Regulations
*Total = 425
†System lacks a properly designed bacteriological sampling plan; plan is inadequate for the system. Not having a bacteriological sampling plan is an alleged
violation of the CPDWR 1.12.1(e), 5.1.1(a);M (CDPHE, 2010).
‡System lacks a monitoring plan detailing the system’s background information, sources, treatment, and distribution system or plan has not been properly
maintained. CPDWR 1.12;M (CDPHE, 2010).
§System does not maintain records according to the minimum requirements, including five years bacteriological analysis, including turbidity, 10 years
chemical analysis, 10 years sanitary surveys. CPDWR 1.6.3;M (CDPHE, 2010).
o Failure to obtain design approval
RECOMMENDATIONS
• Management and operation
o Operator compliance with state requirements
o Failure to monitor and report water quality in compliance
with regulations
o Failure to obtain design approval before construction
The sanitary survey results indicated a lack of adequate planning, and compliance data in turn reflected a failure to execute
planned or required activities. Minor deficiencies were most
frequently related to lack of or inadequate plans for monitoring
cross-connection control. A lack of planning can create the conditions that lead to treatment upsets, distribution failures, and
contaminated water.
JOURNAL AWWA
The recommendations from the modified Pareto analysis
described in this article provided the basis for the Capacity
Building Unit of the Colorado SDWP to effectively target
resources to address the most common failures of public water
systems. The key recommendations from this project follow.
• Focus training, technical assistance, and management support
services on the key weaknesses and compliance failures identified
in this study. The failures identified can be used to prioritize and
target training, technical assistance, and other support activities.
• Develop programs to further assess the root cause of monitoring violations and to reduce them. The most common failure
among public water systems is failure to comply with monitoring
2014 © American Water Works Association
JANUARY 2014 | 106:1
E52
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
TABLE 1
Summary of violations data
Violation
Category
Population
Affected
by Violations
Population Affected
by Violations—%
Population
Unaffected
by Violations—%
Number of Water
Systems Affected
by Violations
Water Systems
Affected
by Violations—%
Water Systems Unaffected
by Violations—%
All
2,604,165
46.91
53.09
1,286
63.57
36.43
MON
2,514,412
45.30
54.70
1,198
59.22
40.78
RPT
451,828
8.14
91.86
249
12.31
87.69
MCL
263,604
4.75
95.25
216
10.68
89.32
TT
89,592
1.61
98.39
76
3.76
96.24
PN
543
0.01
99.99
3
0.15
99.85
SS
65
0.00
100.00
1
0.05
99.95
MCL—maximum contaminant level, MON—monitoring, PN—public notice, RPT—reporting, SS—sanitary survey, TT—treatment technique
FIGURE 10 Violation classification by rule
MCL violations = 691
TT violations = 192
M&R violations = 15,037
CCR+PN violations = 346
5,000
4,561
4,500
4,100
4,000
3,498
Violations—number
3,500
3,000
2,500
2,000
1,500
1,330
950
1,000
500
207
169
110
0
IOC
365
341
RAD
142
0
0
SOC
VOC
CCR
18
0
DBP
LCR
5
PN
174 123
SWTR
173
TCR
Rule or Regulation
CCR—consumer confidence reports, DBP—disinfection by-products, IOC—inorganic contaminants, LCR—Lead and Copper Rule, MCL—maximum contaminant
level, M&R—monitoring and reporting, PN—public notification, RAD—radionuclides, SOC—synthetic organic chemical, SWTR—Surface Water Treatment Rule,
TCR—total coliform rule, TT—treatment technique, VOC—volatile organic chemical
JOURNAL AWWA
2014 © American Water Works Association
JANUARY 2014 | 106:1
E53
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
FIGURE 11 Number of PWSs with violations classified by rule
MCL systems
TT systems
M&R systems
CCR+PN systems
800
714
700
600
PWS—number
500
481
400
335
300
263
247
200
161
149
121
100
45
40
0
32
29
0
IOC
RAD
0
SOC
VOC
CCR
56 52
9
0
DBP
LCR
2
PN
SWTR
TCR
Rule or Regulation
CCR—consumer confidence reports, DBP—disinfection by-products, IOC—inorganic contaminants, LCR—Lead and Copper Rule, MCL—maximum contaminant
level, M&R—monitoring and reporting, PN—public notification, RAD—radionuclides, SOC—synthetic organic chemical, SWTR—Surface Water Treatment Rule,
TCR—total coliform rule, TT—treatment technique, VOC—volatile organic chemical
and reporting regulations. Although the health effects of inadequate monitoring are difficult to quantify, monitoring is clearly
key to ensuring safe water. Although this study shows that failure
to monitor is an important issue, it does not identify the root
cause for these failures. Additional studies may be needed.
• Continue improving data consistency and use the resulting
data to develop benchmarks for improvement. This study was
made possible by the extensive database within the program.
Improving data consistency and availability is an ongoing effort,
and should continue. Benchmarks should be established to track
and assess improvements in public water system performance and
to evaluate the effectiveness of capacity-building services and
other programs.
• Use these and future report results to promote and enhance
communication with stakeholders. The project results provide
a foundation for improved communication with stakeholders
and for focused development of capacity-building services. The
results provide a certain transparency in decision-making so
JOURNAL AWWA
that interested stakeholders as well as internal staff have a clear
understanding of why and how program priorities and operations are established.
The program has already addressed many of the recommendations
of this study. These have included an increased emphasis on
improving disinfection practices, working with systems to develop
monitoring plans, and incorporating the results of this analysis into
training content statewide. The program is actively engaged with
the AWWA Rocky Mountain Section and other training providers
in the state to develop a statewide training strategy.
Through analysis such as this study and continual data quality
improvements, Colorado has an opportunity to continue being
one of the leaders in the nation in providing safe water.
ACKNOWLEDGMENT
The authors would like to acknowledge Dean Vlachos, Jon
DeBoer, Dave Rogers, Richard Summerville, and Margo Griffin
for their assistance on this project.
2014 © American Water Works Association
JANUARY 2014 | 106:1
E54
Oxenford & Williams | http://dx.doi.org/10.5942/jawwa.2014.106.0006
Peer-Reviewed
REFERENCES
ABOUT THE AUTHORS
Jeff Oxenford is principal of Oxenford
Consulting LLC, 2100 Washington
Ave., Golden, CO 80401 USA;
[email protected]. He is currently
leading training of operators on a wide
range of topics that include distribution
system water quality, main location, leak
detection, and board and council training.
Oxenford has worked for the New Jersey Department of
Environmental Protection, Awwa Research Foundation, and
Stratus Consulting. He received his bachelor’s degree in
chemistry and environmental management from Warren Wilson
College in Swannanoa, N.C., and his master’s degree in
environmental engineering from the University of North
Carolina at Charlotte. Sharon I. Williams is the water resources
engineering coordinator for the city of Westminster, Colo.
PEER REVIEW
Date of submission: 01/25/2013
Date of acceptance: 10/29/2013
JOURNAL AWWA
CDPHE (Colorado Department of Public Health and Environment), 2010. Colorado
Primary Drinking Water Regulations. 5 CCR 1003-1. Water Quality Control
Commission, CDPHE, Denver.
Falco, R. & Williams, S.I., 2009. Waterborne Salmonella Outbreak in Alamosa,
Colorado March and April 2008. Safe Drinking Water Program, Water Quality
Control Division, CDPHE. www.colorado.gov/cs/Satellite?blobcol=urldata&bl
obheadername1=Content-Disposition&blobheadername2=Content-Type&blo
bheadervalue1=inline%3B+filename%3D%22Alamosa+Outbreak+Investigati
on+Report+.pdf%22&blobheadervalue2=application%2Fpdf&blobkey=id&
blobtable=MungoBlobs&blobwhere=1251807327157&ssbinary=true
(accessed November 2013).
Oxenford, J.L. & Williams, S.I., 2009. Failure and Root Cause Analysis Project
Report. Capacity Building Unit, Colorado Safe Drinking Water Program,
CDPHE. www.colorado.gov/cs/Satellite?blobcol=urldata&blobheadername1
=Content-Disposition&blobheadername2=Content-Type&blobheadervalue1=
inline%3B+filename%3D%22Failure+and+Root+Cause+Analysis+Project+Re
port+.pdf%22&blobheadervalue2=application%2Fpdf&blobkey=id&blobtable
=MungoBlobs&blobwhere=1251807346928&ssbinary=true (accessed
November 2013)
USEPA (US Environmental Protection Agency), 2013. Sanitary Survey. Office of
Water, Washington. www.epa.gov/safewater/dwa/sanitarysurvey/index.html
(accessed November 2013).
2014 © American Water Works Association
JANUARY 2014 | 106:1