Historical Reconstruction of the Water-Distribution System … · 2001. 11. 18. · 1995 1988 1971 1962 1996 1996 1995 1988 1971 1962. Front Cover Illustrations: Upper Images: Maps

Summary of FindingsHistorical Reconstruction of the Water-Distribution System

Serving the Dover Township Area, New Jersey:January 1962–December 1996

Atlanta, Georgia–September 200101-0489

MA

MJ

JA

SO

ND

FJ

MONTH

MO

NT

HLY

PR

OD

UC

TIO

N, I

N M

ILLI

ON

GA

LLO

NS

1962 1

964196

6 1968

1970 1

972197

4 1976

1978 1

980198

2 1984

1986 1

988199

0 1992

1994 1

996

500

450

400

350

300

250

200

150

100

50

0

1995

1988

1971

1962

1996

1996

1995

1988

1971

1962

Front Cover Illustrations:

Upper Images: Maps showing configuration and expansion of the historical water-distribution system networks serving the Dover Township area, New Jersey:

1962, 1971, 1988, 1995, and 1996.

Lower Image: Plot showing three-dimensional representation of monthly water-supply well production for the Dover Township area, New Jersey, January 1962–December 1996.

S

UMMARY

OF

F

INDINGS

H

ISTORICAL

R

ECONSTRUCTION

OF

THE

W

ATER

-D

ISTRIBUTION

S

YSTEM

S

ERVING

THE

D

OVER

T

OWNSHIP

A

REA

, N

EW

J

ERSEY

:

J

ANUARY

1962 – D

ECEMBER

1996

Agency for Toxic Substances and Disease RegistryU.S. Department of Health and Human Services

Atlanta, Georgia

September 2001

ii Summary of Findings: Historical Reconstruction of the Water-Distribution System Serving the DoverTownship Area, New Jersey: January 1962–December 1996

F

OREWORD

The New Jersey Department of Health and SeniorServices (NJDHSS), with support from the Agency forToxic Substances and Disease Registry (ATSDR), isconducting an epidemiologic study of childhood cancersin Dover Township, Ocean County, New Jersey. In 1996,ATSDR and NJDHSS developed a Public HealthResponse Plan in cooperation with the Ocean CountyHealth Department and the Citizens’ Action Committeeon Childhood Cancer Cluster. The plan outlines a seriesof public health activities including assessments ofpotential environmental exposures in the community. In1997, ATSDR and NJDHSS determined that anepidemiologic study was warranted, and that the studywould include assessments of the potential for exposureto specific drinking-water sources.

To assist the epidemiologic efforts, ATSDR developed awork plan to reconstruct historical characteristics of thewater-distribution system serving the Dover Townshiparea by using water-distribution system modelingtechniques. The numerical model chosen for this effort,EPANET 2, is available in the public domain and isdescribed in the scientific literature. To test thereliability of model simulations, water-distributionsystem data specific to the Dover Township area wereneeded to compare with model results. Lacking suchdata, a field-data collection effort was initiated to obtainpressure measurements, storage-tank water levels, andsystem operation schedules (the on-and-off cycling ofwells and pumps) during winter-demand (March 1998)and peak-demand (August 1998) operating conditions.

Using these data, the water-distribution system modelwas calibrated to present-day (1998) conditions. ATSDRreleased a report and a technical paper in June 2000describing the field-data collection activities and modelcalibration results.

Having established the reliability of the model and themodeling approach, the model was used to examine (orreconstruct) historical characteristics of the water-distribution system. For this purpose, monthlysimulations were conducted from January 1962 throughDecember 1996 to es t imate the proport ionatecontribution of water from points of entry (well or wellfields) to various locations throughout the DoverTownship area.

This summary of findings was developed to provide anoverview of the historical reconstruction analysisconducted by ATSDR and NJDHSS. A full descriptionof the analysis is forthcoming in a comprehensivereport. For the historical period, the following topics arepresented in the full report: (1) data sources andrequirements, (2) methods of analysis, (3) simulationapproaches, (4) selected simulation results of thehistorical reconstruction analysis, and (5) the use ofsensitivity analysis to address issues of uncertainty andvariability of historical system operations. Readersinterested in details of the historical reconstructionmethodology, simulation approaches, or results forspecific years and locations for the Dover Township areashould refer to the full report that is available over theIn t e rne t a t t he ATSDR Web s i t e a t URL:

www.atsdr.cdc.gov

.

CONTENTS

Foreword ii

Contents iii

List of Illustrations iv

List of Tables iv

Glossary and Abbreviations v

Disclaimer vi

Background 1

Methods and Approach 3

Specific Data Needs 3

Data Availability, Quality, Methods, and Sources 10

Examples of Simulation Results 16

Sensitivity Analysis 24

References 27

Principal Investigators, Contributors, and Acknowledgements 28

Principal Investigators 28

Contributors 28

Acknowledgements 28

Questions and Answers 29

What is in the ATSDR report? 29

What is a water-distribution system model and which one did ATSDR use? 29

How are ATSDR and NJDHSS using water-distribution system modeling? 29

What type of information did ATSDR and NJDHSS need to conduct the historical reconstruction analysis? 30

What procedures did ATSDR and NJDHSS use to reconstruct historical water-distribution system conditions? 30

What type of information did ATSDR and NJDHSS have regarding historical operations of the water-distribution system serving the Dover Township area? 30

Given the lack of historical system operating information, could the system have operated in a vastly different manner than was used in the model? 31

Has the proportion of water contributed by different well fields to my street remain about the same over the years? 31

What kind of oversight and input from independent experts did ATSDR have for the historical reconstruction approach it used and for review of its findings? 32

Where and how can I obtain a copy of the ATSDR report? 32

Contents iii

CONTENTS—CONTINUED

List of Illustrations

Figures 1–5. Maps showing:

1. Investigation area, Dover Township, Ocean County, New Jersey 2

2. Water-distribution system serving the Dover Township area, New Jersey, 1962 5




Figure 6. Plot showing three-dimensional representation of monthly water-supply well production, Dover Township area, New Jersey, 1962-1996 9

Figures 7–9. Graphs showing annual variation of simulated proportionate contribution of water from wells and well fields to selected locations in the Dover Township area, New Jersey:

7. Minimum-demand months, 1962-1996 18

8. Maximum-demand months, 1962-1996 20

9. Average-demand months, 1962-1996 22

Figure 10. Graph showing results of sensitivity analyses using the manual adjustment process and Genetic Algorithm (GA) optimization for maximum-, minimum- and average-demand months, 1978 25

List of Tables

Table 1. “Master Operating Criteria” used to develop operating schedules for the historical water-distribution system, Dover Township area, New Jersey 10

2. Water-distribution system operating schedule, Dover Township area, New Jersey, May 1962 11

3. Water-distribution system operating schedule, Dover Township area, New Jersey, July 1971 11

4. Water-distribution system operating schedule, Dover Township area, New Jersey, July 1988 12

5. Water-distribution system operating schedule, Dover Township area, New Jersey, June 1996 13

6. Data availability and method of obtaining data for historical reconstruction analysis, Dover Township area, New Jersey 15.

iv Executive Summary: Historical Reconstruction of the Water-Distribution System Serving the DoverTownship Area, New Jersey: January 1962-December 1996

GLOSSARY AND ABBREVIATIONS

Definition of terms and abbreviations used throughout this report are listed below:

Term orAbbreviation Definition

ATSDR Agency for Toxic Substances and Disease Registry

Consumption The use of water by customers of a water utility; is also known as demand. In a water-distribution system, consumption should equal production if there are no losses through leaks or pipe breaks

Direct measurement or A method of obtaining data that is based on measuring or observing observation the parameter of interest.

EPA U.S. Environmental Protection Agency

EPANET 2 A water-distribution system model developed by the EPA

Epidemiologic study A study to determine whether a relation exists between the occurrence and frequency of a disease and a specific factor such as exposure to a toxic compound found in the environment

GA Genetic Algorithm; a method of optimization that attempts to find the most optimal solution by mimicking the mechanics of natural selection and genetics

Historical reconstruction A diagnostic analysis used to examine the historical characteristics of a water-distribution system

Manual adjustment process A modeling approach whereby a balanced flow condition is achieved through the repeated modification and refinement of modeling parameters by the analyst

Master Operating Criteria Guidelines developed for operating a water-distribution system that are based, in part, on hydraulic engineering principles

Maximum-demand month A time during a prescribed year when water usage is greatest; is also known as a peak- or summer-demand period

Minimum-demand month A time during a prescribed year when water usage is least; is alsoknown as a low- or winter-demand period

NJDHSS New Jersey Department of Health and Senior Services

NPL National Priorities List; the EPA’s official list of hazardous waste sites which are to be cleaned up under the Superfund

Point of entry The location where water enters a water-distribution system from asource such as an aquifer, lake, stream, or river. For the DoverTownship area, the points of entry are the wells and well fields

Production The processing of potable water by a water utility and the delivery of the water to locations serviced by the water-distribution system. In a water-distribution system, production should equal consumption if there are no losses through leaks or pipe breaks

Contents v

Proportionate contribution The derivation of water from one or more sources in differingproportions. The sum of the proportionate contribution at anylocation in the water-distribution system should equal 100%

Qualitative description A method of estimating data that is based on inference or issynthesized using surrogate information

Quantitative estimate A method of estimating data by using computational techniques

Sensitivity analysis A method of characterizing or quantifying uncertainty and variability.This involves conducting a series of model simulations, changing specific parameter values, and comparing the effect of the changed parameter(s) with reference to a base condition

Source-trace analysis A method used to identify the source of delivered water using a water-distribution model. A source-trace analysis can be used to track thepercentage of water reaching any point in a distribution system overtime from a specified location or source

System operations The on-and-off cycling of wells and high-service and booster pumps, and the operational extremes of water levels in storage tanks over a 24-hour period

TIGER Topologically integrated, geographic encoding and referencing system; a database developed by the U.S. Department of Commerce thatdescribes in a digital format the locations of roadways, hydrography,landmarks, places, cities, and geographic census boundaries

Water-distribution system A water-conveyance network consisting of hydraulic devices such as wells, reservoirs, storage tanks, and high-service and booster pumps;and a series of pipelines for delivering potable water

DISCLAIMER

Use of trade names and commercial sources is for identification only and does not imply endorsement bythe Agency for Toxic Substances and Disease Registry or the U.S. Department of Health and HumanServices.

For additional information, write to:

Project OfficerExposure-Dose Reconstruction ProjectDivision of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry1600 Clifton Road, Mail Stop E-32Atlanta, Georgia 30333

vi Summary of Findings: Historical Reconstruction of the Water-Distribution System Serving the DoverTownship Area, New Jersey: January 1962–December 1996

eheoge

esr-

e

aa

a tohady

alm2

SUMMARY OF FINDINGS

HISTORICAL RECONSTRUCTION OF THE WATER-DISTRIBUTION SYSTEM SERVING THE

DOVER TOWNSHIP AREA, NEW JERSEY:JANUARY 1962–DECEMBER 1996

BACKGROUND

Contamination of groundwater resourcesin Dover Township, Ocean County, NewJersey (Figure 1), including the contaminationof water-supply wells, was identified in the1960s (Toms River Chemical Corporation1966) and subsequently documented in the1970s (ATSDR 2001a,b,c,d). Based on publichealth assessments conducted for the DoverTownship area, the Agency for Tox icSubstances and Disease Registry (ATSDR)and the New Jersey Department of Health andSenior Services (NJDHSS) have determinedthat completed human exposure pathways togroundwater contaminants have occurredthrough private and community water supplies(ATSDR 2001a,b,c,d). As a result, NJDHSSand ATSDR are conducting an epidemiologicstudy of childhood leukemia and nervoussystem cancers that occurred in DoverTownship. The epidemiologic study isexploring a variety of possible risk factors,including environmental exposures. To assistNJDHSS with the environmental exposureassessment component of the epidemiologicstudy, ATSDR developed a water-distribution

mode l us ing the EPANET 2 so f twar(Rossman 2000). Results obtained from tmodel will be used to assess exposure tdr ink ing water sources that are beininvestigated as potential risk factors in thepidemiologic investigation.

Because of the lack of appropriathistorical data, the EPANET 2 model wacalibrated to the present-day (1998) watedistribution system characteristics using datacollected during March and August 1998. Threliabi l i ty of the cal ibrated model wasdemonstrated by successfully conductingwater-quality simulation of the transport of naturally occurring conservative element—barium—and comparing results with datcollected at 21 schools and 6 points of entrythe water-distribution system during Marcand April 1996. Results of the field-datcollection activities, model calibration, anreliability testing were described previousl(Maslia et al. 2000a,b). Following calibration,the model was used to simulate historiccharacteristics of the water-distribution systeserving the Dover Township area from 196through 1996.

Background 1

2 Summary of Findings: Historical Reconstruction of the Water-Distribution System Serving the Dover Township Area, New Jersey: January 1962–December 1996

0 5 10 15 MILES

0 1 2 3 MILES

N

Water body

Dover Township

Ciba–Geigy NPL Site

Reich Farm NPL Site

Hydrography

Major road

Street0 20 40 MILES

NEW JERSEY Ocean County

BA

RN

EG

AT

BA

Y

Silver Bay

JACKSONTOWNSHIP

LAKEWOODTOWNSHIP BRICK

TOWNSHIP

BERKELEYTOWNSHIP

South T

oms

River B

orough

Island HeightsBorough

Sea

sid

e H

eig

hts

Bo

rou

gh

EXPLANATION

Green

Branch

Davenpor

t Branch

Jakes

Branc

h

Toms River

WrangleBrook

To

ms

River

Tom

s

River

MANCHESTERTOWNSHIP

OceanCounty

DoverTownship

State Highway

70

Church Road

State Highway 37

Avenue

Hoo

per

Bay B

ay Avenue

FischerB

oulevard

Avenue

Washington St

State

Highway 37

Route

9

Gar

den

Stat

eP

arkw

ay

Figure 1. Investigation area, Dover Township, Ocean County, New Jersey.

Roads, hydrography, and boundaries based on 1995 TIGER/Line data

hel

f

seed-

eisngn.

lmdenfd

,e-

hek(5)ticn

ofndin

This document is a summary of a detailedreport that describes the historical recon-struction analysis. This summary and the fullreport are viewed as companion documents toMaslia et al. (2000a) which describes theanalysis of the 1998 water-distribution systemserving the Dover Township area. The fullreport focuses on the historical reconstructionanalysis of the water-distribution systemincluding: (1) data sources and requirements,(2) methods of analysis, (3) simulationstrategies, (4) selected simulation results, and(5) the use of sensitivity analysis to addressissues of uncertainty and variabil i ty ofhistorical system operations.

METHODS AND APPROACH

Given the pauc i t y o f h is to r i ca lcontaminant-specific concentration data duringmost o f the per i od re levant to theepidemiologic study, ATSDR and NJDHSSdec ided that model ing e ffo r ts shou ldconcentrate on estimating the percentage ofwater that a study subject might have receivedfrom each point of entry (well or well fields) tothe water-distribution system (Figure 2). Thisapproach uses the concept of “proportionatecontribution” described in Maslia et al.(2000a, p. 4) wherein at any given point in thedistribution system, water may be derived fromone or more sources in differing proportions.

Databases were developed from diversesources of information and were used todescribe the historical distribution-systemnetworks specific to the Dover Township area.These data were applied to EPANET 2 andsimulations were conducted for each month of

the historical period—January 1962 througDecember 1996 (420 simulations or “modruns”). After completing the 420 monthlyanalyses, source-trace analysis simulationswere conducted to determine the percentage owater contributed by each well or well fieldoperating during each month. Results of theanalyses—the percentage of water derivfrom the different sources that historically supplied the water-distribution system—werprovided to health scientists for their analysin assessing the environmental factors beiconsidered by the epidemiologic investigatio

SPECIFIC DATA NEEDS

A simulation approach to the historicareconstruction of the water-distribution systein the Dover Township area requ i reknowledge of the functional as well as thphysical characteristics of the distributiosystem. Accordingly, six specific types oinformation were required: (1) pipeline annetwork configurations for the distributionsystem; (2) potable water-production dataincluding information on the locat ioncapaci ty, and t ime of operat ion of thgroundwater production wells; (3) consumption or demand data at locations throughout td is t r ibu t ion sys tem; (4) s torage- tancapacities, elevations, and water-level data; high-service and booster pump characteriscurves; and (6) system-operations informatiosuch as the on-and-off cycling schedule wells and high-service and booster pumps, athe operational extremes of water levels storage tanks.

Methods and Approach 3

tedsd,f-nd

ofhee-hege-

d,oreye),P.,bys

t

Examples of the histor ical networkconfigurations for 1962, 1971, 1988, and 1996are shown in Figures 2 through 5, respectively.(Yearly historical network configurations mapsfor the period 1962 through 1996 are presentedin the full report.) Figures 2 through 5 showthe complexity of the system increasedsignif icantly over the t ime span of thehistorical period. For example, the 1962 water-distribution system served nearly 4,300customers from a population of about 17,200persons (Board of Public Utilities, State ofNew Jersey 1962) and was characterized formodeling by (Figure 2):

• approximately 2,400 pipe segmentsranging in diameter from 2 to 12 inchesand comprising a total service length of 77miles;

• 3 groundwater extraction wells with arated capacity of 1,900 gallons per minute;

• 1 elevated storage tank and standpipe witha combined rated storage capacity of 0.45million gallons; and

• production of about 1.3 million gallonsper day during the peak-production monthof May.

By contrast, in 1996—the last year of thehistorical reconstruction period—the water-distribution system served nearly 44,000customers from a population of about 89,300persons (Board of Public Utilities, State ofNew Jersey 1996) and was characterized formodeling by (Figure 5):

• more than 16,000 pipe segments rangingin diameter from 2 to 16 inches andcomprising a total service length of 482miles;

• 20 groundwater extraction wells with arated capacity of 16,550 gallons perminute;

• 12 high-service or booster pumps;

• 3 elevated and 6 ground-level storagetanks with a combined rated capacity of7.35 million gallons; and

• production of about 13.9 million gallonsper day during the peak-production monthof June.

Analysis of production data indicates thathe historical distribution systems could bcharacterized by three typical demand perioeach year: (1) a low- or winter-demand periogeneral ly represented by the month oFebruary—designated as the minimumdemand month; (2) a peak- or summer-demaperiod, represented by one of the monthsMay, June, July, or August—designated as tmaximum-demand month; and (3) an averagdemand period, generally represented by tmonth of October—designated as the averademand month.

Water-production data were gathereaggregated, and analyzed for each well fevery month of the historical period. Thesdata were obtained from the water utilit(Flegal 1997), Board of Public Utilities, Statof New Jersey, Annual Reports (1962–1996and NJDHSS data searches (Michael McLinden, written communication, August 281997). The production data were measured using in-line flow meters at water-supply well(George J. Flegal, Manager, United WaterToms River, Inc., oral communication, Augus28, 2001).



South T

oms

River B

orough

BA

RN

EG

AT

B

AY

Silver Bay

Toms River

MANCHESTERTOWNSHIP

BERKELEY TOWNSHIP

LAKEWOOD TOWNSHIP

BRICK TOWNSHIP

Figure 2. Water-distribution system serving the Dover Township area, New Jersey, 1962.

Wells 13,14

C

A

D

E

B

Well15

Horner Streettank and standpipe

N

Water body

Dover Township


Reich Farm NPL Site

Hydrography

Major road

Water pipeline Municipal well

Storage tank

EXPLANATION

Notes: (1) Water pipelines range in diameter from 2 inches to 12 inches(2) Roads, hydrography, and boundaries based on 1995 TIGER/Line data

D Pipeline location and letter. Percent contribution is reported in text

0 1 2 3 MILES


South T

oms

River B

orough

BA

RN

EG

AT

B

AY

Silver Bay

Toms River

MANCHESTERTOWNSHIP

BERKELEY TOWNSHIP

LAKEWOOD TOWNSHIP

BRICK TOWNSHIP

Well20

Well15

Wells 14,16,18,19,21

Well 17

Wells 22,23,26,27

Indian Hillelevated

Parkwayground level

Silver Bay Well

AnchorageWell

Holly Plantground level

South Toms Riverelevated

C

A

D

E

B

N

EXPLANATION



Water body

Dover Township


Reich Farm NPL Site

Hydrography

Major road


Storage tank


0 1 2 3 MILES


South T

oms

River B

orough

BA

RN

EG

AT

B

AY

Silver Bay

Toms River

MANCHESTERTOWNSHIP

BERKELEY TOWNSHIP

LAKEWOOD TOWNSHIP

BRICK TOWNSHIP

Well 20

Well31

Well15

Wells 33,34,35

Wells21,30

Wells 32,38

Wells 22,23,24,26,28,29

Indian Hillelevated

Parkwayground level

Holiday Cityground level

Route 37ground level

Holly Plantgroundlevel


Windsorground level

C

A

D

E

B

N

EXPLANATION



Water body

Dover Township


Reich Farm NPL Site

Hydrography

Major road


Storage tank


0 1 2 3 MILES


South T

oms

River B

orough

BA

RN

EG

AT

B

AY

Silver Bay

Toms River

MANCHESTERTOWNSHIP

BERKELEY TOWNSHIP

LAKEWOOD TOWNSHIP

BRICK TOWNSHIP

Well 20

Well 31

Well15

Wells 33,34,35

Wells21,30,37

Wells 32,38

Wells 22,24,26, 28,29,39,41,42

Indian Hillelevated

Parkwayground level

Holiday Cityground level

Well40

Route 37ground level

Holly Plantgroundlevel


Windsorground level

North Doverelevated

C

A

D

E

B

N

EXPLANATION



Water body

Dover Township


Reich Farm NPL Site

Hydrography

Major road


Storage tank


0 1 2 3 MILES

i-of,

rere-

ntoi-

ead

Month ly p roduc t i on da ta can berepresented graphically as shown in a three-dimensional plot (Figure 6). Referring to thisplot, the x-axis is the year (1962–1996), the y-axis is the month (January–December), and thez-axis is the total monthly production inmillion gallons. Maximum production isshown to occur in the months of May, June,July, or August. In addition, considerableproduction increases occurred in 1971, 1988,and 1995. These years are characterized on theplot by sharp peaks.

As noted previously, to simulate the distrbution of water for each of the 420 months the historical period, network configurationdemand, and operational information werequired. Before 1978, operational data weunavailable requiring development of systemoperation parameters—designated as “MasterOperating Criteria.” These are based ohydraulic engineering principles necessary successfully operate distribution systems simlar to the one serving the Dover Township ar(Table 1). From 1978 forward, for selecte


MAM

JJA

SO

ND

FJ

MONTH

MO

NT

HLY

PR

OD

UC

TIO

N, I

N M

ILLI

ON

GA

LLO

NS

1962 1

964196

6 1968

1970 1

972197

4 1976

1978 1

980198

2 1984

1986 1

988199

0 1992

1994 1

996

500

450

400

350

300

250

200

150

100

50

0

1988

1971

1962

Figure 6. Three-dimensional representation of monthly water-supply well production, Dover Township area, New Jersey, 1962–96.

y x

z

1996

1995

ly

nreatepsm5);e 5

drnt

tbed of

tainct,

calornto

e

years, operators of the water utility providedinformation on the generalized operating prac-tices for a typical “peak-demand” (summer)and “non-peak demand” (fall) day. Theseguidelines were used in conjunction with the“Master Operating Criteria” to simulate a typi-cal 24-hour daily operation of the water-distri-bution system for each month of the historicalperiod.

Examples of historical water-distributionsystem operating schedules for the maximum-demand months of May 1962, July 1971, July1988, and June 1996 are shown in Tables 2through 5, respectively. These tables indicatethe hour-by-hour operation of wells and high-service and booster pumps during a typical dayof the maximum-demand month for the givenyear. Note that in 1962 (Table 2), high-serviceand booster pumps were not part of thedistribution system and, therefore, only

groundwater wells were operated to suppdemand by discharging water directly into thedistribution system (wells 13–15, Figure 2). I1968, high-service and booster pumps weadded to the distribution system. From thyear forward, some wells supplied storagtanks, then high-service and booster pumwere operated to meet distribution-systedemands (wells 21–30, 40, and 42, Figure other wells still discharged directly into thdistribution system (refer to Tables 2 throughfor details).

DATA AVAILABILITY, QUALITY, METHODS, AND SOURCES

In this type of study, the ideal or desirecondition is to obtain all data required fomodel simulations through direct measuremeor observation. In reality, however, necessarydata are not routinely available by direcmeasurement or observation and must synthes ized us ing general ly accepteengineering analyses and methods. Issuesdata sources and the methods used to obdata that cannot be directly measured refleultimately, on the credibility of simulationresults. To address these issues for historireconstruction analysis, the methods fobtaining the necessary data were grouped ithree categories (Table 6):

• Direct measurement or observation—Data included in this category wereobtained by direct measurement orobservation of historical data and areverifiable by independent means. Of thethree data categories, these data were thmost preferred in terms of reliability andleast affected by issues of uncertainty.

Table 1. “Master Operating Criteria” used to develop operating schedules for the historical water-distribution system, Dover Township area, New Jersey

Parameter CriteriaPressure1

1Generally, for residential demand, minimum recommended pres-sure is about 20 pounds per square inch. However, for some locations in the Dover Township area (mostly in areas near the end of distribution lines) lower pressures were simulated.

Minimum of 15 pounds per square inch, maximum of 110 pounds per square inch at pipeline locations, including network end points

Water level Minimum of 3 feet above bottom elevation of tank; maximum equal to elevation of top of tank; ending water level should equal the starting water level

Hydraulic device on-line date

June 1 of year installed to meet maximum-demand conditions

On-and-off cycling: Manual operation

Wells and high-service and booster pumps cannot be cycled on-and-off from 2200 to 0600 hours

On-and-off cycling: Automatic operation

Wells and high-service and booster pumps can be cycled on-and-off at any hour

Operating hours Wells should be operated continuously for the total number of production hours, based on production data2

2See full report for historical monthly production data.



Tab

le 2

.W

ater

-dis

trib

utio

n sy

stem

ope

ratin

g sc

hedu

le, D

over

Tow

nshi

p ar

ea, N

ew J

erse

y,

Wat

er-d

istr

ibut

ion

syst

em o

pera

ting

sche

dule

, Dov

er T

owns

hip

area

, New

Jer

sey,

May

196

2[M

ay is

max

imum

-dem

and

mon

th fo

r 19

62; h

our

of d

ay in

col

or m

eans

wel

l ope

ratin

g;

G

roun

dwat

er w

ell]

Wel

l ID

1

1 Wel

ls d

isch

arge

dire

ctly

into

the

dist

ribut

ion

syst

em.

Ho

ur

of

the

Day

2

2 Hou

r of

the

day:

0 is

mid

nigh

t; 12

is n

oon,

res

pect

ivel

y.

01

23

45

67

89

1011

1213

1415

1617

1819

2021

2223

Gro

un

dw

ater

Wel

l

Hol

ly (

13)

Hol

ly (

14)

Bro

oksi

de (

15)

Tab

le 3

. Ju

ly 1

971

[Jul

y is

max

imum

-dem

and

mon

th fo

r 19

71; h

our

of d

ay in

col

or m

eans

wel

l or

pum

p op

erat

ing;

Gro

undw

ater

wel

l;

H

igh-

Ser

vice

or

Boo

ster

pum

p]

Wel

l ID

1

1 Wel

ls d

isch

arge

dire

ctly

into

the

dist

ribut

ion

syst

em.

Ho

ur

of

the

Day

2

2 Hou

r of

the

day:

0 is

mid

nigh

t; 12

is n

oon,

res

pect

ivel

y.

01

23

45

67

89

1011

1213

1415

1617

1819

2021

2223

Gro

un

dw

ater

Wel

l

Bro

oksi

de (

15)

Sou

th T o

ms

Riv

er (

17)

Indi

an H

ead

(20)

Anc

hora

ge3

3 Anc

hora

ge a

nd S

ilver

Bay

do

not h

ave

wel

l num

bers

des

igna

ted

by w

ater

util

ity.

Silv

er B

ay3

Hig

h-S

ervi

ce o

r B

oo

ster

Pu

mp

Pu

mp

IDH

olly

pum

p 14

4 Hol

ly p

ump

1, H

olly

pum

p 2,

and

Hol

ly p

ump

3 su

pplie

d by

Hol

ly g

roun

d-le

vel s

tora

ge ta

nks

and

Hol

ly w

ells

14,

16,

18,

19,

and

21.

Hol

ly p

ump

24 4

Hol

ly p

ump

3

Par

kway

pum

p 15 5

5 Par

kway

pum

p 1

and

Par

kway

pum

p 2

supp

lied

by P

arkw

ay g

roun

d-le

vel s

tora

ge ta

nk a

nd P

arkw

ay w

ells

22,

23,

26,

and

27.

Par

kway

pum

p 2


, Ju

ly 1

988

[Jul

y is

max

imum

-dem

and

mon

th fo

r 19

88; h

our

of d

ay in

col

or m

eans

wel

l or

pum

p op

erat

ing;

Gro

undw

ater

wel

l;

H

igh-

Ser

vice

or

Boo

ster

pum

p]

Wel

l ID

1

1 Wel

ls d

isch

arge

dire

ctly

into

the

dist

ribut

ion

syst

em.

Ho

ur

of

the

Day

2

2 Hou

r of

the

day:

0 is

mid

nigh

t; 12

is n

oon,

res

pect

ivel

y.

01

23

45

67

89

1011

1213

1415

1617

1819

2021

2223

Gro

un

dw

ater

Wel

lB

rook

side

(15

)

Indi

an H

ead

(20)

Rou

te 7

0 (3

1)

Sou

th T

oms

Riv

er (

32)

Ber

kele

y (3

3)

Ber

kele

y (3

4)

Ber

kele

y (3

5)

Sou

th T

oms

Riv

er (

38)

Hig

h-S

ervi

ce o

r B

oo

ster

Pu

mp

Pu

mp

ID

Hol

ly p

ump

13

3 Hol

ly p

ump

1, H

olly

pum

p 2,

and

Hol

ly p

ump

3 su

pplie

d by

Hol

ly g

roun

d-le

vel s

tora

ge ta

nks

and

Hol

ly w

ells

21

and

30.

Hol

ly p

ump

23

Hol

ly p

ump

33

Par

kway

pum

p 14

4 Par

kway

pum

p 1

and

Par

kway

pum

p 2

supp

lied

by P

arkw

ay g

roun

d-le

vel s

tora

ge ta

nk a

nd P

arkw

ay w

ells

22,

23,

24,

26,

28,

and

29.

Par

kway

pum

p 24

Hol

iday

City

Pum

p5

5 Pum

p op

erat

ed fr

om 0

630

to 1

130

hour

s an

d 16

30 to

203

0 ho

urs.

St.

Cat

herin

e’s

pum

p 6, 7

6 Pum

p op

erat

ed fr

om 0

930

to 1

445

hour

s.7A

lso

know

n as

Rou

te 3

7.

Sou

th T

oms

Riv

er p

ump

1

Sou

th T

oms

Riv

er p

ump

2

Win

dsor

pum

p 18 8

8W

inds

or p

ump

1 and

Win

dsor

pum

p 2 su

pplie

d by

Win

dsor

gro

und-

leve

l tan

k.

Win

dsor

pum

p 2

Tab

le 4

.W

ater

-dis

trib

utio

n sy

stem

ope

ratin

g sc

hedu

le, D

over

Tow

nshi

p ar

ea, N

ew J

erse

y,


June

199

6[J

une

is m

axim

um-d

eman

d m

onth

for

1996

; hou

r of

day

in c

olor

mea

ns w

ell o

r pu

mp

oper

atin

g;

G

roun

dwat

er w

ell;

Hig

h-S

ervi

ce o

r B

oost

er p

ump]

Wel

l ID

1

1 Wel

ls d

isch

arge

dire

ctly

into

the

dist

ribut

ion

syst

em.

Ho

ur

of

the

day

2

2 Hou

r of

the

day:

0 is

mid

nigh

t; 12

is n

oon,

res

pect

ivel

y.

01

23

45

67

89

1011

1213

1415

1617

1819

2021

2223

Gro

un

dw

ater

Wel

lB

rook

side

(15

)

Rou

te 7

0 (3

1)

Sout

h To

ms

Riv

er (

32)

Ber

kele

y (3

3)

Ber

kele

y (3

4)

Ber

kele

y (3

5)

Sout

h To

ms

Riv

er (

38)

Pu

mp

IDH

igh

-Ser

vice

or

Bo

ost

er P

um

p

Hol

ly p

ump

13

3 Hol

ly p

ump

1, H

olly

pum

p 2,

and

Hol

ly p

ump

3 su

pplie

d by

Hol

ly g

roun

d-le

vel s

tora

ge ta

nks

and

Hol

ly w

ell 3

0.

Hol

ly p

ump

23 3H

olly

pum

p 3

Park

way

pum

p 14

4 Par

kway

pum

p 1

and

Par

kway

pum

p 2

supp

lied

by P

arkw

ay g

roun

d-le

vel s

tora

ge ta

nk a

nd P

arkw

ay w

ells

22,

24,

26,

28,

29,

and

42.

Park

way

pum

p 24

Hol

iday

City

pum

p

St. C

athe

rine’

s pu

mp5

5 Als

o kn

ow a

s R

oute

37.

Sout

h To

ms

Riv

er p

ump

1

Sout

h To

ms

Riv

er p

ump

2

Win

dsor

pum

p 16 6 6

6W

inds

or p

ump

1, W

inds

or p

ump

2, a

nd W

inds

or p

ump

3 su

pplie

d by

Win

dsor

gro

und-

leve

l sto

rage

tank

and

Win

dsor

wel

l 40.

Win

dsor

pum

p 2

Win

dsor

pum

p 3

Tab

le 5

.W

ater

-dis

trib

utio

n sy

stem

ope

ratin

g sc

hedu

le, D

over

Tow

nshi

p ar

ea, N

ew J

erse

y,

aledofgst

efee

ofdo

nsde)d-e

alr

h,

ps).ekesesey

edtol

• Quantitative estimates—Data included inthis category were estimated or quantifiedusing computational methods.

• Qualitative description—Data included inthis category were based on inference orwere syn thes ized us ing sur rogateinformation. Of the three data categories,data derived by qualitative descriptionwere the least preferred in terms ofreliability and the most affected by issuesof uncertainty.

Of the six specific types of informationrequired for the historical reconstruction anal-ysis, the network pipeline data, groundwaterwell-location data, groundwater well-produc-tion data, and storage-tank data were obtainedby direct measurement or observation (Table6). These data were available throughout theentire historical period and they could beassessed for quality and verified by indepen-dent means such as state reports or field obser-vations. For example, groundwater well-production data were available for every wellfor every month of the historical period andthese data were measured by the water utilityusing in-line flow-metering devices at ground-water wells (George J. Flegal, Manager,United Water Toms River, Inc., oral communi-cation, August 28, 2001).

Data for historical consumption (ordemand) consisted of two components—monthly volumes (quantity) and spatialdistribution (location). The monthly volumeswere obtained by using a quant i tat iveestimation method. Data were available frommetered billing records for October 1997through April 1998 and verified through thecalibration process described in Maslia et al.

(2000a,b); the magnitude of monthly historicproduction was known based on measurflow data. Using these data, estimates historical demand were quantified by imposinthe requirement that total consumption muequal total production.

Direct measurement or quanti tat ivest imates of the spatial distr ibut ion ohistorical demand were not available for thDover Township area. Therefore, qualitativdescription methods were used to estimatehistorical data values. In doing so, estimatesthe spatial distribution of historical deman(demand pat terns) were based on twassumptions: (1) historical demand patterwere similar to the present-day demanpatterns which are known from availablmetered billing records (Table 6); and (2demand patterns could be inferred from lanuse classification using historical land-usclassification as a surrogate indicator. Toassess the validity of this approach, historicland-use classification or zoning maps foDover Township were used in conjunction witdistribution-system network maps for 19621967, 1978, 1990, and 1996 (network malike the ones shown in Figures 2 through 5Using information obtained from the land-usclassification and distribution-system networmaps, geospatial and comparative analyswere conducted. Results of these analysindicated that the distribution of land-usclassification in Dover Township was relativelstatic and changed little during the historicalperiod. These analyses substantially validatthe qualitative description method used estimate the spatial distribution of historicademand.



Tab

le 6

. D

ata

avai

labi

lity

and

met

hod

of o

btai

ning

dat

a fo

r hi

stor

ical

rec

onst

ruct

ion

anal

ysis

, Dov

er T

owns

hip

area

, Ne

w J

erse

y

Dat

a Ty

pe

Tim

e F

ram

e o

f A

vaila

bili

tyS

ou

rce

Met

ho

d o

f O

bta

inin

g D

ata

1

1 Dire

ct m

easu

rem

ent o

r ob

serv

atio

n—m

easu

red

or o

bser

ved

data

ava

ilabl

e fo

r so

me

or th

roug

hout

his

toric

al p

erio

d; d

ata

verifi

able

by

inde

pend

ent m

eans

; Q

uant

itativ

e es

timat

e—di

rect

mea

sure

men

t or

obse

rvat

ion

of h

isto

rical

dat

a no

t ava

ilabl

e fo

r so

me

or m

ost o

f his

toric

al p

erio

d; d

ata

estim

ated

by

com

puta

tiona

l met

hods

; Q

ualit

ativ

e de

scrip

tion—

dire

ct m

easu

rem

ent o

r ob

serv

atio

n of

his

toric

al d

ata

not a

vaila

ble

for

mos

t or

all o

f hi

stor

ical

per

iod

; dat

a ba

sed

on in

fere

nce

or s

ynth

esiz

ed

No

tes

Dir

ect

Mea

sure

men

t o

r O

bse

rvat

ion

Qu

anti

tati

veE

stim

ate

Qu

alit

ativ

e D

escr

ipti

on

Pipe

line

loca

tion

and

geom

etry

1962

–199

6W

ater

util

ity d

atab

ase

2

2 Fle

gal (

1997

).

xIn

-ser

vice

dat

e as

sign

ed to

be

Janu

ary

1 of

in

-ser

vice

yea

r

Gro

undw

ater

wel

l lo

catio

n19

62–1

996

Wat

er u

tility

dat

abas

eB

oard

of P

ublic

Util

ities

rep

orts3 ,

ATS

DR

and

NJD

HS

S fi

eld

verifi

catio

n

3 B

oard

of P

ublic

Util

ities

, Sta

te o

f New

Jer

sey,

Ann

ual R

epor

ts (

1962

–199

6).

xIn

-ser

vice

dat

e as

sign

ed to

be

June

1 o

f in

-ser

vice

yea

r

Gro

undw

ater

wel

l pr

oduc

tion

1962

–199

6W

ater

util

ity d

atab

ase

2 ,2 ,

Boa

rd o

f Pub

lic U

tiliti

es A

nnua

l Rep

orts3 ,

NJD

HS

S d

ata

sear

ch4

4 Mic

hael

P. M

cLin

den,

writ

ten

com

mun

icat

ion,

Aug

ust 2

8, 1

997.

xW

ater

util

ity d

ata,

196

2–19

96; N

JDH

SS

da

ta, 1

962–

1979

; dat

a fr

om in

-line

flow

met

ers5 ,

hou

rly v

alue

s av

aila

ble

for

1996

;pr

ior

to 1

996,

mon

thly

val

ues

avai

labl

e;

aver

age

daily

ope

ratio

n es

timat

ed fr

om

mon

thly

dat

a an

d w

ell c

apac

ity

5 Geo

rge

J. F

lega

l, M

anag

er, U

nite

d W

ater

Tom

s R

iver

, Inc

., or

al c

omm

unic

atio

n, A

ugus

t 28,

200

1.

Stor

age

tank

ge

omet

ry, c

apac

ity,

and

loca

tion

1962

–199

6W

ater

util

ity d

atab

ase

2 ,

Boa

rd o

f Pub

lic U

tiliti

es A

nnua

l Rep

orts3 ,

ATS

DR

and

NJD

HS

S fi

eld

verifi

catio

n

xIn

-ser

vice

dat

e as

sign

ed to

be

June

1 o

f in

-ser

vice

yea

r

Est

imat

ion

of

cons

umpt

ion

(dem

and)

(dem

and)

Oct

ober

199

7–A

pril

1998

Wat

er u

tility

bill

ing

reco

rds,

AT

SD

R c

alib

rate

d m

odel6

6 Mas

lia et

al.

(200

0a, b

).

xP

rior

to O

ctob

er 1

997,

dat

a no

t ava

ilabl

e to

in

vest

igat

ors;

qua

ntita

tive

estim

ate

base

d on

ass

umpt

ion

that

dem

and

mus

t equ

al

prod

uctio

n

Spat

ial d

istr

ibut

ion

of c

onsu

mpt

ion

Oct

ober

199

7–A

pril

1998

Wat

er u

tility

bill

ing

reco

rds,

ATS

DR

cal

ibra

ted

mod

el6x

Prio

r to

Oct

ober

199

7, d

ata

not a

vaila

ble

to

inve

stig

ator

s; e

stim

ates

bas

ed o

n qu

alita

tive

asse

ssm

ent o

f lan

d-us

e an

d ge

ospa

tial a

naly

sis

Pum

p-ch

arac

teris

tic

curv

es19

98W

ater

util

ity d

ata2 ,

ATS

DR

cal

ibra

ted

mod

el6x

Syst

em o

pera

tions

,

196

2–19

77N

one

Non

e

“Mas

ter

Ope

ratin

g hy

drau

lic e

ngin

eerin

g pr

inci

ples

, w

ater

util

ity o

pera

ting

prac

tices

x

Syst

em o

pera

tions

,

197

8–19

87Ty

pica

l pea

k da

y (s

umm

er)

and

non-

peak

day

(fa

ll)

for

sele

cted

yea

rs

Wat

er u

tility

ope

ratio

nal n

otes7 ,

“M

aste

r O

pera

ting

Crit

eria

,”

Crit

eria

,”

Crit

eria

,”

hydr

aulic

eng

inee

ring

prin

cipl

es

hydr

aulic

eng

inee

ring

prin

cipl

es,

7 Ric

hard

Otte

ns, J

r., P

rodu

ctio

n M

anag

er, U

nite

d W

ater

Tom

s R

iver

, Inc

., w

ritte

n co

mm

unic

atio

n, 1

998.

xx

Hig

h-se

rvic

e an

d bo

oste

r pu

mp

disc

harg

e es

timat

ed fr

om w

ater

util

ity n

otes

Dai

ly h

ours

of o

pera

tion

for

wel

ls fr

ompr

oduc

tion

data

Syst

em o

pera

tions

,

198

8–19

96W

ater

util

ity o

pera

tiona

l not

es7 ,“M

aste

r O

pera

ting

obse

rved

wat

er-u

tility

op

erat

ing

prac

tices6

xx

xH

igh-

serv

ice

and

boos

ter

pum

p di

scha

rge

estim

ated

from

wat

er u

tility

not

es; h

ourly

op

erat

ions

dat

a fo

r 19

96

usin

g su

rrog

ate

info

rmat

ion.

Typi

cal p

eak

day

(sum

mer

)an

d no

n-pe

ak d

ay (

fall)

fo

r se

lect

ed y

ears

; all

of 1

996;

and

Mar

ch a

nd

Aug

ust 1

998

t

ro-n

ern-n

ntedn,ip-tangheer

.,

esg

dtoh

dad

s

The high-service and booster pump-characterist ic data were derived usinginformation obtained from the water utility(Flegal 1997). This information consisted ofhead values versus flow values which wererefined during the model calibration process(Maslia et al. 2000a,b).

The historical system-operation data wereobtained using each of the three methods ofobtaining data descr ibed previously—depending on the time frame (Table 6). For theear l y h is to r ica l pe r i od (1962-1977) ,investigators relied on hydraulic engineeringprinciples and the “Master Operating Criteria”(Table 1). Because data describing specificoperational practices were not available,operating schedules developed for these earlyhistorical networks (for example, Table 2 andTable 3) were based on qualitative descriptionsof system operations. To maintain a balancedflow condition however, water-distributionsystems of similar configuration and facilitiesas the historical Dover Township area system,generally operate using on-and-off cyclingschedules of limited variability. That is, wellsand high-service and booster pumps must becycled on-and-off within a limited or narrowoperating range. Simulations conducted on thewater-distribution system serving the DoverTownsh ip area conf i rmed the l im i tedvariability of the on-and-off cycling operatingschedule.

For the 1977–1987 period, system-opera-tion data were developed from quantitativeestimates and qualitative descriptions of theoperating schedules (Table 6). These data werederived using hydraulic engineering principles,

the “Master Operating Criteria,” and frominformation provided by the water utility thadescribed the general operations of the water-distribution system for a typical “peak” day(summer) and a “non-peak” (fall) day. Fosome of the years, the water utility also prvided estimates of discharge to the distributiosystem from the high-service and boostpumps (Richard Ottens, Jr., Production Maager, United Water Toms River, Inc., writtecommunication, 1998).

System-operation data for the most recehistorical systems (1988–1996) were obtainfrom direct measurement or observatioquantitative estimates, and qualitative descrtions of operating schedules (Table 6). Dasources used to develop these operatischedules (for example, Table 5) included tgeneralized operating notes from the watut i l i ty (Richard Ottens, Jr., ProductionManager, United Water Toms River, Incwri t ten communication, 1998), hour lyoperations data for 1996 (Flegal 1997), nottaken by ATSDR and NJDHSS staff durinfield-data collection activities in March andApril 1998 (Maslia et al. 2000a), and theobservation that the distribution system hapreviously operated in a manner very similar the present-day system (1998) for whicdetailed information was available.

EXAMPLES OF SIMULATION RESULTS

Analysis of the proportionate contributionof water from wells and well fields to selectenetwork locations in the Dover Township areillustrates the increasing complexity anoperational variability of the distributionsystem throughout the historical period. A


e

rte

ee

r

nd

nge

alice,hs

-lyD

es

previously described, these results wereobtained by conducting source-trace analysissimulations. The annual variation of thesimulated proportionate contribution of waterfrom all active wells and well fields to selectedlocations in the Dover Township area is shownfor the minimum-demand month of February(Figure 7), the maximum-demand months ofMay, June, July, or August (Figure 8), and theaverage-demand month of October (Figure 9).For each o f t hese examples , f i vegeographically distinct pipeline locations wereselected from the historical networks torepresent the spa t ia l d i s t r i bu t ion o fproportionate contribution results. Theselocations are identified on Figures 2 through 5,and Figures 7 through 9 as locations A, B, C,D, and E.

Comparison of the May 1962 results withthe June 1996 results (Figure 8), indicates theincreasing complexity of the network anddistribution-system operations and how suchoperations influenced the proportionatecontribution of water to specific locations. InMay 1962, only two well fields (Holly andBrookside) provided water to any one location;whereas, in June 1996, as many as seven wellfields provided water to the distribution system(for example, pipeline location E in Figure 8).

In Figures 7 through 9, the sum of theproportionate contribution of water from allwells and well fields to any pipeline locationshould be 100%. Because of numericalapproximation and roundoff, however, the totalcontribution from all wells and well fields maysum to slightly less or slightly more than 100%at some locations. This is expected when using

numerical simulation techniques. In thhistorical reconstruction analysis conductedfor the distribution system serving the DoveTownship area, the sum of the proportionacontribution at any location ranges from 98%to 101%.

In reviewing the simulation results, thannua l and seasona l va r ia t ion o f thproportionate contribution of water is evidentby inspecting, for example, the results fopipeline location D. Annual variation isdetermined by selecting a certain demacondition (minimum, maximum, or average—Figures 7, 8, or 9, respectively) and comparithe proportionate contribution results over thhistorical period (1962–1996). Seasonvariation is determined by choosing a specifyear and comparing the propor t ionatcontr ibution results for the minimum-maximum-, and average-demand mont(Figures 7, 8, and 9, respectively).

Simulation results for the maximumdemand months of May 1962, July 1971, Ju1988, and June 1996 for pipeline location exempl i fy the annual var ia t ion in thecontribution of water to this location andindicate the following (see Figure 8 for thproportionate contribution results and Figure2 through 5 for well and well field locations):

• May 1962—100% of the water wasprovided by the Brookside well (15);

• July 1971—30% of the water wasprovided by the Holly wells (14, 16, 18,19, and 21); 54% by the Brookside well(15); 3% by the Indian Head well (20);and 14% by Parkway wells (22, 23, 26,and 27);



No

pipe

line

to lo

catio

n B

No

pipe

line

to lo

catio

n B

100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION A

PIPELINE LOCATION B

PIPELINE LOCATION C

MINIMUM-DEMAND MONTH (FEBRUARY)

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996


100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION D

PIPELINE LOCATION E

Holly Brookside (15) South Toms River Indian Head (20) Parkway Route 70 (31) Berkeley

Figure 7. Annual variation of simulated proportionate contribution of water from wells and well fields to selected locations in the Dover Township area, New Jersey, minimum-demand months, 1962–96.

Well or well field

No

pipe

line

to lo

catio

n E

No

pipe

line

to lo

catio

n E

See full report for information on specific wells in operation. Number in paren-thesis is individual well number

MINIMUM-DEMAND MONTH (FEBRUARY)

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996


100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION A

PIPELINE LOCATION B

PIPELINE LOCATION C

No

pipe

line

to lo

catio

n B

No

pipe

line

to lo

catio

n B

MAXIMUM-DEMAND MONTH

MAY

1962

JULY

196

3

JUNE

196

4

JUNE

196

5

JULY

196

6

JUNE

196

7

JULY

196

8

JUNE

196

9

JULY

197

0

JULY

197

1

AUG

1972

JULY

197

3

JULY

197

4

AUG

1975

JUNE

1976

JULY

197

7

JULY

197

8

JULY

197

9

JUNE

198

0

JULY

198

1

JULY

198

2

JULY

198

3

JUNE

198

4

JULY

198

5

JUNE

198

6

AUG

1987

JULY

198

8

AUG

1989

JULY

199

0

JULY

199

1

JULY

199

2

JULY

199

3

JUNE

199

4

AUG

1995

JUNE

199

6


100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION D

PIPELINE LOCATION E

Holly Brookside (15) South Toms River Indian Head (20) Parkway Route 70 (31) BerkeleyWindsor (40)AnchorageSilver Bay

Figure 8. Annual variation of simulated proportionate contribution of water from wells and well fields to selected locations in the Dover Township area, New Jersey, maximum-demand months, 1962–96.

Well or well field

No

pipe

line

to lo

catio

n E

No

pipe

line

to lo

catio

n E

MAXIMUM-DEMAND MONTH

MAY

1962

JULY

196

3

JUNE

196

4

JUNE

196

5

JULY

196

6

JUNE

196

7

JULY

196

8

JUNE

196

9

JULY

197

0

JULY

197

1

AUG

1972

JULY

197

3

JULY

197

4

AUG

1975

JUNE

1976

JULY

197

7

JULY

197

8

JULY

197

9

JUNE

198

0

JULY

198

1

JULY

198

2

JULY

198

3

JUNE

198

4

JULY

198

5

JUNE

198

6

AUG

1987

JULY

198

8

AUG

1989

JULY

199

0

JULY

199

1

JULY

199

2

JULY

199

3

JUNE

199

4

AUG

1995

JUNE

199

6

See full report for information on specific wells in operation. Num-ber in parenthesis is individual well number. Anchorage and Silver Bay wells not assigned a number by water utility


100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION A

PIPELINE LOCATION B

PIPELINE LOCATION C

AVERAGE-DEMAND MONTH (OCTOBER)

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

No

pipe

line

to lo

catio

n B

No

pipe

line

to lo

catio

n B


100

90

80

70

60

50

40

30

20

10

0

100

90

80

70

60

50

40

30

20

10

0

PE

RC

EN

TAG

E O

F W

AT

ER

FR

OM

WE

LL

OR

WE

LL

FIE

LD

PIPELINE LOCATION D

PIPELINE LOCATION E

Holly Brookside (15) South Toms River Indian Head (20) Parkway Route 70 (31) BerkeleyAnchorageSilver Bay

Figure 9. Annual variation of simulated proportionate contribution of water from wells and well fields to selected locations in the Dover Township area, New Jersey, average-demand months, 1962–96.

Well or well field

No

pipe

line

to lo

catio

n E

No

pipe

line

to lo

catio

n E

AVERAGE-DEMAND MONTH (OCTOBER)

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

See full report for information on specific wells in operation. Num-ber in parenthesis is individual well number. Anchorage and Silver Bay wells not assigned a number by water utility

henress

all

ndd

sas antalinger-s

ad

ral

ollon

eglhednses

• July 1988—49% of the water wasprovided by Holly wells (21 and 30); 26%by the Brookside well (15); 11% by theSouth Toms River wells (32 and 38); 14%by the Parkway wells (22, 23, 24, 26, 28,and 29); and 1% by the Berkeley wells(33-35); and

• June 1996—66% of the water wasprovided by the Holly well (30); 2% bythe Brookside well (15); 9% by the SouthToms River wells (32 and 38); 2% by theParkway wells (22, 24, 26, 28, 29, and42); 4% by the Berkeley wells (33-35),and 17% by the Windsor well (40).

The simulation results shown in Figures 7,8, and 9 demonstrate that the contribution ofwater from wells and well fields varied by timeand location. However, the results also showthat certain wells provided the predominantamount of water to locations throughout theDover Township area. Readers who areinterested in the proportionate contribution ofwater from specific water sources at specifiedtimes during the historical period of 1962through 1996 should refer to the full report.

SENSITIVITY ANALYSIS

The proportionate contribution resultsdescribed above were obtained from trace-analysis simulat ions conducted on thehistorical distribution-system networkswhereby balanced flow conditions wereachieved through the manual refinement ofmodeling parameters. The adjusted parameterswere the on-and-off cycling pattern values ofwells (pattern factor values assigned inEPANET 2) and the operational extremes ofwater levels in the storage tanks. This

modeling approach was designated as t“manual adjustment process.” Simulatioresults presented in this summary weobtained using the manual adjustment proceand were the bases of comparisons for sensitivity analyses.

To address the issue of uncertainty avar iab i l i ty o f sys tem operat ions, anspecifically to test the sensitivity of theproportionate contribution results to variationin model-parameter values, a technique wrequired that would “search” for and selectset of alternate operating conditions differefrom those determined using the manuadjustment process. These alternate operatcondit ions needed to also result in thsatisfactory operation of the historical watedistribution system. Such a technique wafound in the Genetic Algorithm optimization(GA) method. Simply put, a GA refers to method of optimization that attempts to finthe most optimal solution by mimicking (in acomputational sense) the mechanics of natuselection and natural genetics. (Aral et al.[2001] discuss GAs and their application twater-distribution system analysis; the fureport also presents additional references the development and application of GAs.)

Changes in simulated proportionatcontribution results were compared usinresu l ts ob ta ined th rough the manuaadjustment process and the GA approach. Tsensitivity analysis simulations were groupeinto three categories (Figure 10): (1) variatioo f pa t te rn fac tors ass igned to we l l(operational variation in the value and the timof day—designated as sensitivity simulation


alyrer8,s,re,s.tois

SENS0, SENS1, SENS2, and SENS3); (2)variation in the operational minimum pressurecriteria at pipeline locations (designated assensitivity simulations SENS4 and SENS5);and (3) variation in the operational storagetank water-level differences between thestarting time (0 hours) and ending time (24hours) o f a s imulat ion (des ignated assensitivity simulations SENS6 and SENS7).Sensitivity analysis simulation SENS0 andSENS1 applied the GA approach to everyhistorical network (420 simulations) using thebalanced flow conditions obtained from the

manual adjustment process as the initistarting conditions. The remaining sensitivitanalysis simulations (SENS2–SENS7) weconducted for distribution-system networks foselected years of 1962, 1965, 1971, 1971988, and 1996. For these historical networkthe previously described parameters wevar ied wi th respect to the min imum-maximum-, and average-demand monthReaders desiring specific details pertaining the definition of each of the sensitivity analyssimulations should refer to the full report.

Sensitivity Analysis 25

0 5 10 15 20 25 300

10

20

30

40

50

60

70

80

90

100

PE

RC

EN

T O

F S

TU

DY

LO

CA

TIO

NS

LE

SS

TH

AN

ABSOLUTE VALUE OF DIFFERENCE IN SIMULATED CONTRIBUTION OF WATERBETWEEN MANUAL ADJUSTMENT PROCESS AND GENETIC ALGORITHM

OPTIMIZATION, IN PERCENT CONTRIBUTION

Figure 10. Results of sensitivity analyses using the manual adjustment process and Genetic Algorithm (GA) optimization for maximum-, minimum- and average-demand months, 1978.

SENSITIVITY ANALYSIS SIMULATIONS, 1978

(1) n=number of study locations where the contribution of water is greater than 0 percent

(2) See complete report for specific details on definitions of sensitivity analysis simulations

(3) Study locations provided by New Jersey Department of Health and Senior Services without personal identifiers and status

SENS0 6,004

SENS1 6,061

SENS2 5,963

SENS3 5,950

SENS4 5,988

SENS5 6,021

SENS6 6,102

SENS7 6,138

Notes

Minimum pressure criteria

Difference between startingand ending storage-tank water-level in a 24-hour period

Well-pattern factors

nSimulation

identification Parameters varied

lal6d

20e

ed

se-ine)e,

fre

rsy

mddr

rtr-r

leye

ot

Figure 10 shows examples of results forthe sensitivity analysis simulations represent-ing 1978 conditions. These results indicatesmall variations when comparing the propor-tionate contribution results from the manualadjustment process to results obtained usingthe GA approach. Figure 10, however, alsoshows that the simulated proportionate contri-bution of water from wells and well fields isrelatively insensitive to changes in systemoperational parameters. For a 24-hour period,the average percentage of water over all studylocations derived from all wells or well fieldsusing either the manual adjustment process orany of the GA simulations does not varyappreciably. For example, the results in Figure10 indicate that more than 90% of study loca-tions show a difference of 10% or less in thesimulated proportionate contribution resultsderived from either the manual adjustment pro-cess or any of the GA simulations. Theseresults (Figure 10) indicate that there was anarrow range within which the historicalwater-distribution system could have success-fully operated to maintain a balanced flow con-dition and satisfy the “Master OperatingCriteria” previously described. Results forother historical networks (such as 1988 and1996) show less variation when comparingsimulated proportionate contribution resultsobtained using either the manual adjustmentprocess or any of the GA optimizationapproaches.

For the historical reconstruction analysis,investigators assumed that daily systemoperations over a period of 1 month could berepresented by a “typical” 24-hour day foreach month of the historical period. To test the

val idi ty of this assumption, addit ionasensitivity analyses using hourly operationdata obtained from the water utility for 199were conducted. For the maximum-demanmonth of June 1996, a 30-day analysis—7hours—was conducted by simulating thhourly operation according to the data suppliby the water utility. When results for thehourly operation simulation for 30 day(average over the 30-day period) wercompared with results from the “typical” 24hour day for the month of June, differences the proportionate contribution of water to thfive pipeline locations (A, B, C, D, and Eshowed only slight variations. As an examplthe difference in the contribution of water fromthe Parkway well field for the two methods osimulating the daily system operations we0% for location A, 1% for location B, 4% forlocation C, 2% for location D, and 3% folocation E. Therefore, sensitivity analysiassisted in confirming that the day-to-daoperations of the water-distribution systewere highly consistent over a 30-day perio(based on available 1996 hourly data) ancould be represented by a “typical” 24-houoperational pattern.

The sensitivity analysis conducted as paof the historical reconstruction of the watedistr ibut ion system serving the DoveTownship area indicate that: (1) there was anarrow range within which the historicawater-distr ibut ion systems could havsuccessful ly operated and st i l l sat isfhydraulic engineering principles and th“Master Operating Criteria,” and (2) dailyoperational variations over a month did nappreciably change the propor t ionate


. acfl-

.r2.

.s1,

c.l-).

sg-

n

,r-tfnt

l.

ch

.ers

contribution of water from specific sourceswhen compared to a typical 24-hour dayrepresen t ing the mon th . Thus , t hereconstructed historical water-distributionsystems and operating criteria—based onapplying the “Master Operating Criteria” andusing generalized water-utility information—are believed to be the most probable andrealistic scenarios under which the historicalwater-distribution systems were operated.

REFERENCESAgency for Toxic Substances and Disease Registry

(ATSDR). 2001a. Health consultation, drinkingwater quality analyses, March 1996 to June1999, Uni ted Water Toms River, DoverTownship, Ocean County, New Jersey. Atlanta:Agency for Toxic Substances and DiseaseRegistry.

Agency for Toxic Substances and Disease Registry(ATSDR). 2001b. Public health assessment forCiba-Geigy Corporation, Dover Township,Ocean County, New Jersey, EPA facility ID:NJD001502517. Atlanta: Agency for ToxicSubstances and Disease Registry.

Agency for Toxic Substances and Disease Registry(ATSDR). 2001c. Public health assessment forDover Township Municipal Landfill (a/k/aDover Township Landfill), EPA facility ID:NJD980771570 and Silverton private wellcontamination investigation (a/k/a Silvertonwells), EPA facility ID: NJD981877780, DoverTownship, Ocean County, New Jersey. Atlanta:Agency for Toxic Substances and DiseaseRegistry.

Agency for Toxic Substances and Disease Registry(ATSDR). 2001d. Public health assessment forReich Farm, Dover Township, Ocean County,New Jersey, EPA facility ID: NJD980529713.Atlanta: Agency for Toxic Substances andDisease Registry.

Aral MM, Guan J, Maslia ML, Sautner JB. 2001Reconstruction of hydraulic management ofwater-distr ibution system using genetialgorithms. Atlanta: Georgia Institute oTechnology, Mult imedia EnvironmentaSimulations Laboratory Report No. MESL0101.

Board of Public Utilities, State of New Jersey1962. Annual report of Toms River WateCompany, year ended December 31, 196Newark (NJ): Board of Public Utilities.

Board of Public Utilities, State of New Jersey1996. Annual report of United Water TomRiver Company, year ended December 31996. Newark (NJ): Board of Public Utilities.

Flegal GJ. 1997. United Water Toms River, InLetter to Barker Hamill from George J. Flegaconcerning hydraul ic model and waterdistribution system data. Toms River (NJMarch 27, 1997.

Maslia ML, Sautner JB, Aral MM. 2000a. Analysiof the 1998 water-distribution system servinthe Dover Township area, New Jersey: fielddata collection activities and water-distributiosystem modeling. Atlanta: Agency for ToxicSubstances and Disease Registry.

Maslia ML, Sautner JB, Aral MM, Reyes JJAbraham JE, Williams RC. 2000b. Using wated is t r ibu t ion sys tem model ing to assisepidemiologic investigations. ASCE Journal oWater Resources Planning and Manageme126(4):180-98.

Rossman LA. 2000. EPANET 2 users manuaCincinnati: US Environmental ProtectionAgency, National Risk Management ResearLaboratory.

Toms River Chemical Corporat ion. 1966Memorandum to R.K. Sponagel from P. Wehnconcerning the water supply situation. TomRiver (NJ). March 21, 1966.

References 27

atllick...

V.

dey);

ipnlf

tal.

d;

eE.he,g.

kithr-asalngforon.

.terd

t.ed

d,

nd

PRINCIPAL INVESTIGATORS, CONTRIBUTORS, AND ACKNOWLEDGEMENTS

Principal Investigators

MORRIS L. MASLIA, MSCE, PEResearch Environmental Engineer and

Project OfficerExposure-Dose Reconstruction ProjectDivision of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry

JASON B. SAUTNER, MSCE, EITEnvironmental Health Scientist

Division of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry

MUSTAFA M. ARAL, Ph.D., PE, PHyDirector, Multimedia Environmental Simulations

LaboratorySchool of Civil and Environmental EngineeringGeorgia Institute of Technology

Contributors

RICHARD E. GILLIG , MCPSupervisory Environmental Health Scientist

Chief, Superfund Site Assessment BranchDivision of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry

JUAN J. REYES, MSPAProgram Manager

Director, Office of Regional OperationsOffice of the Assistant AdministratorAgency for Toxic Substances and Disease Registry

ROBERT C. WILLIAMS , PE, DEEAssistant Surgeon General

Director, Division of Health Assessment and Consultation

Agency for Toxic Substances and Disease Registry

AcknowledgementsThe authors acknowledge their colleagues

ATSDR for providing assistance and advice with aaspects of this investigation: John E. Abraham, PatrBrady, James A. Clark, Amy B. Funk, Amanda JGonzalez, Virginia Lee, Susan Metcalf, Thomas AMignone, Jennifer Noack, Mary G. Odom, Sven ERodenbeck, Kevin A. Ryan, Kathy Skipper, Gregory Ulirsch, and Ann Walker.

The following organizations and their staff providesuggestions, assistance, or information: New JersDepartment of Health and Senior Services (NJDHSSOcean County Health Department; Dover TownshMunicipal Government; Citizens Action Committee oChildhood Cancer Cluster; Multimedia EnvironmentaSimulations Laboratory at the Georgia Institute oTechnology; New Jersey Department of EnvironmenProtection; New Jersey Geological Survey; U.SEnvironmental Protection Agency, Region II; UniteWater Toms River, Inc.; Union Carbide Corporationand Ciba-Geigy Corporation.

The authors would like to specifically acknowledgJerald A. Fagliano, Michael Berry, and Jonathan Savrin of NJDHSS for suggestions and advice from tepidemiologic perspective; Michael J. McLindenformerly with NJDHSS, for assistance in obtaininhistorical production data; Lewis A. Rossman, U.SEnvironmental Protection Agency, Nat

Documents

Historical Reconstruction of the Water-Distribution System … · 2001. 11. 18. · 1995 1988 1971 1962 1996 1996 1995 1988 1971 1962. Front Cover Illustrations: Upper Images: Maps