• TABLE OF CONTENTS
HIDE
 Title Page
 Disclaimer
 Table of Contents
 List of Tables
 List of Figures
 Acknowledgement
 Executive summary
 Methodology
 Baseline water use
 Post-retrofit results
 New fixture satisfaction ratin...
 Analysis of costs and benefits
 Conclusions and recommendation...
 Appendix A
 Appendix B
 Appendix C
 References
 Abbreviations
 Mathematical terms






Title: Residential indoor water conservation study: evaluation of high efficience indoor plumbing fixture retrofits…
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/WC05633808/00001
 Material Information
Title: Residential indoor water conservation study: evaluation of high efficience indoor plumbing fixture retrofits…
Physical Description: Book
Creator: Aquacraft, Inc.
Publisher: Aquacraft, Inc.
Publication Date: 2003
 Subjects
Subject: Water conservation
Residential water consumption
 Notes
Abstract: Prepared for and submitted to: East Bay Municipal Utility District and the United States Environmental Protection Agency
 Record Information
Bibliographic ID: WC05633808
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.

Downloads

This item has the following downloads:

WC05633808 ( PDF )


Table of Contents
    Title Page
        Page i
    Disclaimer
        Page ii
    Table of Contents
        Page iii
        Page iv
        Page v
    List of Tables
        Page vi
    List of Figures
        Page vii
    Acknowledgement
        Page viii
    Executive summary
        Page ix
        Page x
        Page xi
        Page xii
        Page xiii
        Page xiv
        Page xv
        Page xvi
        Page xvii
        Page xviii
        Page xix
        Page xx
    Methodology
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
    Baseline water use
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
    Post-retrofit results
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
    New fixture satisfaction ratings
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
    Analysis of costs and benefits
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
    Conclusions and recommendations
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
    Appendix A
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
    Appendix B
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
    Appendix C
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
    References
        Page 165
        Page 166
        Page 167
    Abbreviations
        Page 168
        Page 169
        Page 170
    Mathematical terms
        Page 171
        Page 172
Full Text








RESIDENTIAL INDOOR WATER
CONSERVATION STUDY:


EVALUATION OF HIGH EFFICIENCY INDOOR
PLUMBING FIXTURE RETROFITS IN
SINGLE-FAMILY HOMES IN THE EAST BAY
MUNICIPAL UTILITY DISTRICT SERVICE AREA




July 2003


Prepared for and submitted to:

East Bay Municipal Utility District
and
The United States Environmental Protection Agency


Prepared by:

Peter W. Mayer, William B. DeOreo, Erin Towler, and David M. Lewis
Aquacraft, Inc. Water Engineering and Management
2709 Pine St.
Boulder, Colorado 80302
www.aquacraft.com








DISCLAIMER


This report has been funded in part with a grant from the U.S. Environmental Protection
Agency (EPA). The views expressed by the authors are their own and do not necessarily reflect
those of the EPA. Mention of trade names, products, or services does not convey, and should not
be interpreted as conveying, official EPA approval, endorsement, or recommendation.









TABLE OF CONTENTS



D ISC LA IM ER ............................................................................................................................................................ II

TABLE OF CONTENTS .........................................................................................................................................III

L IST O F T A B L E S.................................................................................................................................................... V I

LIST O F FIG U R ES.................................................................................................................................................V II

ACKNOWLEDGEMENTS.................................................................................................................................. VIII

EXECUTIVE SUMMARY ................................................................................................................................ IX

CHAPTER 2 METHODOLOGY......................................................................................................................... 1
SELECTION OF STUDY PARTICIPANTS 2
F first Team M eeting................................................................... .....................................................................2
Study G group Selection............................................................................... .............................................. 3
Invitation Letter 3
Selection of Participants 3
DATA MANAGEMENT 4
Creation ofR etrofi t D atabase...................................................................... .................................................. 4
B killing D a ta .......................................... ....................................................................................................... 4
INITIAL SITE VISITS 4
Visit Protocol .............................................................................. ........................ 5
Audit Questionnaire 5
Data Logger Installation 5
Hot Water Data Collection 6
Collection of Fixture Traces 6
Participant Agreement 7
Audit Data Entry and QA/QC 7
R etrieval and Ve ficaton ofln tial D ata .................................................................. .................................... 7
P re-R etrofit Logging R eport......................................................................... .......................................... 8
R e-L ogg ing E effort .................................................................................................. ............................ ........... 8
P re-R etrofit F low Trace A analysis .................................................................................... .............................. 8
Quality Assurance and Quality Control 10
PLAN RETROFITS AND INSTALL CONSERVING FIXTURES 12
D raftR etrofi t P lan ...................................................... ....................................................... .......................... 13
F in a l R etrofi t P lan ............................................................................................................................................ 15
Perform Retrofits. ............. ...................................... 16
Validate and Tabulate R etrofits................................................................................................................... 16
Audit Installation Quality Assurance and Quality Control 16
Post Retrofit Logging Quality Assurance and Quality Control 17

CHAPTER 3 BASELINE WATER USE ........................................................................................................... 18
ANNUAL WATER USE 18
Seasonal W after U se ................................................................................. ........................................ ........... 2 0
DEMOGRAPHIC INFORMATION 21
N um ber of residents per household........................................................... ................................................... 21
H household Inform action .................................................................................................. .................. ............ 22
F mixture P perform ance.......................................................... .......................................................................... 23
END USE DATA 25
Daily Household Use ................. ............................... ................................ .....................25
INDOOR PER CAPITAL USE 27









M ean P er C ap ita Indoor U se............................................................................... .........................................28
B ase ne H ot W after Usage............................................................. .............................................................. 31
L e a k s .................................................................................................................................................................. 3 2
FIXTURE USAGE 36
Toilets....................................................................................................................... ........ 36
S h o w e rs .............................................................................................................................................................. 3 8
C loth es W ash ers........................................................................................................................ ....... ........... 4 1
D ish w a sh ers ....................................................................................................................................................... 4 3
F a u ce ts ............................................................................................................................................................... 4 3
B a th s ........................................................................................................................ .................................... 4 4
MAXIMUM DAY DEMANDS 44

CHAPTER 4 POST-RETROFIT RESULTS .................................................................................................... 47
COMPARISON OF WATER USE DURING LOGGING PERIODS 47
D em graphic Inform aton................................................................................. ........................................... 48
DAILY HOUSEHOLD USE 49
INDOOR PER CAPITAL USE 50
P ost-R etrofi t H ot W after Usage ...................................................................................... .............................. 53
Analysis of Water Savings Excluding Leaks......................................................................55
FIXTURE USAGE 56
Toilets...................................... 56
Flushes per capital per day 58
Impact of Toilet Make and Model 59
ULF Toilet Savings from Other Studies 62
Show ers................................... ......................................................................... ......... .............. ................... 63
LF Shower Savings from Other Studies 67
C clothes W ashers........................................................................................................................ . ............. 68
Clothes washer Savings Found in Other Studies 73
Faucets......................................................................................................................................................................... 74
B a th s ........................................................................................................................ .................................... 7 6
Leaks............................................... .............................. 76
MAXIMUM DAY INDOOR DEMANDS 78
BILLING DATA ANALYSIS 80
CHAPTER 5 NEW FIXTURE SATISFACTION RATINGS .....................................................................................84
RESULTS 84
Toilets....................................................................................................................... ........ 84
Niagara vs Caroma 86
C loth es W ash ers........................................................................................................................ ....... ........... 8 7
Clothes washer ratings by brand 89
Sh ow erh eads ........ ........................................................................................................................................... 89
F aucet A erators ........................................................................................ .................................... ............90
Study Participation .............................................. .............................................................. 91
CHAPTER 6 ANALYSIS OF COSTS AND BENEFITS..................................................................................93
TOILETS 93
U t lity C ost S having s....................................................................................... ............................................... 94
CLOTHES WASHERS 95
U t lity C ost S having s....................................................................................... ............................................... 9 7
SHOWERHEADS 97

CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS .........................................................................99
RESEARCH FINDINGS 99
P er C ap ita U se................................................................................................... .......................................... 9 9
Customer Satisfaction with New Products................................................................... ..................... 100
C ost-B benefit A nalysis........................................................... ............... ......................... ...................... 101









Clothes washers 102
Showerheads 103
RECOMMENDATIONS 104
O going R esearch.............................................. ............................................................ .......................... 104

A P P E N D IX A ...........................................................................................................................................................105

IN VITA TIO N LETTER ......................................................................................................................105

INVITATION FACT SHEET .............................................................................................................106

INITIAL PARTICIPATION QUESTIONNAIRE ................................................................................. 107

HOME WATER CONSERVATION STUDY AUDIT FORM......................................................109

PARTICIPATION AGREEMENT ....................................................................................................115

NEW PRODUCT INFORMATION AND SATISFACTION SURVEY .............................................. 120

A P P E N D IX B ...........................................................................................................................................................129

FLOW TRACE ANALYSIS DESCRIPTION ...................................................................................... 129

A P P E N D IX C ...........................................................................................................................................................138

COMPLETE SURVEY RESPONSES PRODUCT INFORMATION AND SATISFACTION
SU R V E Y .................................................................................................................................................... 138

R E FER EN C E S ................................................................................................................................................... 165

A B BR EV IA TIO N S..................................................................................................................................................168

DEFINITION OF MATHEMATICAL TERMS ..................................................................................................171









LIST OF TABLES


TABLE 1 1 END USE DATA TABLE EXAMPLE 9
TABLE 1 2 CONSERVING FIXTURES INSTALLED IN 33 STUDY HOMES 17
TABLE 2 1 ANNUAL WATER USE PER HOUSEHOLD IN 1999 AND 2000, SAMPLE FRAME AND STUDY GROUP 20
TABLE 2 2 PRE-RETROFIT TOILET RATING, NON-ULF TOILETS IN EBMUD AND SEATTLE 24
TABLE 2 3 PRE-RETROFIT RATING OF NON-CONSERVING CLOTHES WASHERS, EBMUD AND SEATTLE 25
TABLE 2 4 INDOOR PER CAPITAL WATER USE COMPARISON 29
TABLE 2 5 AVERAGE BASELINE PER CAPITAL USE AND 95% CONFIDENCE INTERVALS 30
TABLE 2 6 BASELINE PER CAPITAL HOT WATER USE 31
TABLE 2 7 CONFIDENCE INTERVALS FOR BASELINE 32
TABLE 3 1 MEAN DAILY INDOOR PER CAPITAL USE COMPARISON 3 LOGGING PERIODS 48
TABLE 3 2 COMPARISON TESTS OF BASELINE AND POST-RETROFIT INDOOR PER CAPITAL WATER USE 48
TABLE 3 3 MEAN INDOOR PER CAPITAL WATER USE, BASELINE AND POST-RETROFIT 53
TABLE 3 4 POST-RETROFIT PER CAPITAL HOT WATER USE 54
TABLE 3 5 COMPARISON OF BASELINE AND POST-RETROFIT PER CAPITAL HOT WATER USE 55
TABLE 3 6 COMPARISON OF BASELINE AND POST-RETROFIT PER CAPITAL DAILY USE EXCLUDING LEAKS 56
TABLE 3 7 CAROMA VS NIAGARA TOILET WATER USE 61
TABLE 3 8 COMPARISON OF ULF SAVINGS FROM OTHER STUDIES 63
TABLE 3 9 SHOWER USAGE COMPARISON, BASELINE AND POST-RETROFIT 65
TABLE 3 10 COMPARISON OF LF SHOWERHEAD SAVINGS FROM OTHER STUDIES 67
TABLE 3 11 CLOTHES WASHER USAGE COMPARISON 72
TABLE 3 12 COMPARISON OF CLOTHES WASHER SAVINGS FROM OTHER STUDIES 73
TABLE 3 13 FAUCET USE COMPARISONS, BASELINE AND POST-RETROFIT 74
TABLE 3 14 BATH USAGE COMPARISONS, BASELINE AND POST-RETROFIT 76
TABLE 3 15 PRE AND POST-RETROFIT BILLING DATA COMPARISON, STUDY GROUP (N=33) 81
TABLE 3 16 PRE AND POST-RETROFIT BILLING DATA COMPARISON, CONTROL GROUP (N=966) 82
TABLE 4 1 PRE AND POST-RETROFIT TOILET RATING 85
TABLE 4 2 COMPARISON RATINGS OF NEW AND OLD TOILETS 86
TABLE 4 3 CUSTOMER RATINGS OF CAROMA AND NIAGARA TOILETS 87
TABLE 4 4 COMPARISON RATINGS OF NON-CONSERVING AND CONSERVING CLOTHES WASHERS 88
TABLE 4 5 COMPARISON RATINGS OF CLOTHES WASHER BY BRAND 89
TABLE 4 6 SHOWERHEAD SATISFACTION RATINGS 90
TABLE 4 7 FAUCET AERATOR SATISFACTION RATING 91
TABLE 4 8 STUDY PARTICIPATION SATISFACTION RATING 92
TABLE 5 1 WATER REDUCTION AND COST SAVINGS FROM ULF TOILETS 93
TABLE 5 2 COSTS AND PAYBACK PERIOD OF ULF TOILETS 94
TABLE 5 3 UTILITY COST SAVINGS FROM ULF TOILETS 95
TABLE 5 4 WATER REDUCTION AND COST SAVINGS FROM CONSERVING CLOTHES WASHERS 96
TABLE 5 5 COSTS AND PAYBACK PERIOD FOR CONSERVING CLOTHES WASHERS 97
TABLE 5 6 UTILITY COST SAVINGS FROM CONSERVING CLOTHES WASHERS 97
TABLE 5 7 WATER REDUCTION AND COST SAVINGS FROM SHOWERHEADS 98
TABLE 5 8 COST AND PAYBACK PERIOD FOR SHOWERHEADS 98
TABLE 5 9 UTILITY COST SAVINGS FROM CONSERVING SHOWERHEADS 98









LIST OF FIGURES


FIGURE 2 1 FRIGIDAIRE GALLERY CLOTHES WASHER FIGURE 2 2 FISHER & PAYKEL ECOSMART 14
FIGURE 2 3 WHIRLPOOL SUPER CAPACITY PLUS RESOURCE SAVER CLOTHES WASHER 14
FIGURE 3 1 AVERAGE DAILY USE DISTRIBUTIONS FOR 1999 AND 2000, STUDY GROUP (N=33) AND SAMPLE FRAME 19
FIGURE 3 2 BI-MONTHLY WATER USE IN RETROFIT STUDY GROUP (N=33) 21
FIGURE 3 3 HOUSEHOLD SIZE DISTRIBUTION, (N=37) 22
FIGURE 3 4 SCATTER DIAGRAM OF STUDY GROUP PRE-RETROFIT DAILY HOUSEHOLD WATER USE 26
FIGURE 3 5 BASELINE FREQUENCY DIST --TOTAL DAILY HOUSEHOLD WATER USE FOR PRE-RETROFIT GROUP 27
FIGURE 3 6 BASELINE INDOOR PER CAPITAL WATER USE PERCENTAGE INCLUDING LEAKAGE 28
FIGURE 3 7 PERSISTENT FLAPPER LEAK 34
FIGURE 3 8 END OF FLAPPER LEAK WITH TOILET FLUSH (GREEN) 34
FIGURE 3 9 DAILY PER HOUSEHOLD LEAKAGE DISTRIBUTION, PRE-RETROFIT STUDY GROUP 35
FIGURE 3 10 BASELINE TOILET FLUSH VOLUME DISTRIBUTION, PRE-RETROFIT STUDY GROUP 37
FIGURE 3 11 BASELINE TOILET FLUSH FREQUENCY DISTRIBUTION, PRE-RETROFIT STUDY GROUP 37
FIGURE 3 12 BASELINE SHOWER VOLUME DISTRIBUTION, PRE-RETROFIT STUDY GROUP 39
FIGURE 3 13 BASELINE SHOWER DURATION DISTRIBUTION, PRE-RETROFIT STUDY GROUP 40
FIGURE 3 14 BASELINE SHOWER FLOW RATE DISTRIBUTION, PRE-RETROFIT STUDY GROUP 41
FIGURE 3 15 SAMPLE CLOTHES WASHER FLOW TRACE, 1995 MAYTAG SUPER CAPACITY 42
FIGURE 3 16 SAMPLE CLOTHES WASHER FLOW TRACE, 1990 WHIRLPOOL HEAVY DUTY 42
FIGURE 3 17 BASELINE PRE-RETOFIT PEAK DAY DEMAND FOR EACH STUDY HOUSE 45
FIGURE 3 18 BASELINE PRE-RETROFIT PEAK DAILY INSTANTANEOUS DEMAND DISTRIBUTION 46
FIGURE 4 1 SCATTER DIAGRAM OF DAILY INDOOR PER HOUSEHOLD USE, PRE AND POST-RETROFIT 49
FIGURE 4 2 DAILY PER HOUSEHOLD INDOOR WATER USE DISTRIBUTIONS, PRE AND POST-RETROFIT 50
FIGURE 4 3 HOUSE BY HOUSE AVERAGE PER CAPITAL DAILY USE COMPARISON, BASELINE AND POST-RETROFIT 51
FIGURE 4 4 POST-RETROFIT INDOOR PER CAPITAL WATER USE PERCENTAGE INCLUDING LEAKAGE 51
FIGURE 4 5 TOILET FLUSH VOLUME DISTRIBUTION, BASELINE AND POST-RETROFIT 58
FIGURE 4 6 TOILET FLUSHING FREQUENCY DISTRIBUTION, BASELINE AND POST-RETROFIT 59
FIGURE 4 7 SHOWER VOLUME FREQUENCY DISTRIBUTIONS, BASELINE AND POST-RETROFIT 64
FIGURE 4 8 SHOWER DURATION FREQUENCY DISTRIBUTIONS, BASELINE AND POST-RETROFIT 64
FIGURE 4 9 SHOWER FLOW RATE FREQUENCY DISTRIBUTIONS, BASELINE AND POST-RETROFIT 66
FIGURE 4 10 FRIGIDAIRE GALLERY SAMPLE FLOW TRACE 69
FIGURE 4 11 FRIGIDAIRE GALLERY SAMPLE FLOW TRACE 69
FIGURE 4 12 FISHER & PAYKEL ECOSMART WASH CYCLE SAMPLE FLOW TRACE 70
FIGURE 4 13 FISHER & PAYKEL ECOSMART WASH CYCLE SAMPLE FLOW TRACE 70
FIGURE 4 14 WHIRLPOOL SUPER CAPACITY PLUS SAMPLE FLOW TRACE 71
FIGURE 4 15 FAUCET USAGE COMPARISON, BASELINE AND POST-RETROFIT 75
FIGURE 4 16 DAILY PER HOUSEHOLD LEAKAGE DISTRIBUTIONS, BASELINE AND POST-RETROFIT 78
FIGURE 4 17 PEAK DAILY INDOOR USE COMPARISON, BASELINE AND POST-RETROFIT 79








ACKNOWLEDGEMENTS


This research would not have been possible without a grant from the US EPA and the
assistance of John Flowers, Director of the U.S. EPA WAVE program. John immediately saw
the benefits of this research and moved to support the project.
This project was assisted by EBMUD staff members Richard Harris, manager of water
conservation; Richard Bennett, project manager; and Dan Muir, who provided valuable field
support.
Thanks go to the following people who contributed to the research effort and offered
valuable insight and advice: Al Dietemann, Tim Skeel, John Koeller, John Wright, Tiffany
Fulcher, and Leslie Martien.








EXECUTIVE SUMMARY


Residential water conservation retrofits and retrofit rebate programs, often subsidized by
municipal water providers, represent an essential element of water conservation planning and
programs as well as regional best management practices. While many of these programs have
proved popular with customers, questions remain about the actual impact of residential retrofits
on per-capita and per household water use -particularly on individual end uses over time.
Reliable measurements of water savings are essential for long-range projections of the impacts of
conservation projects on urban water demands. As water providers fund water conservation
practices, whether voluntarily or by regulatory requirements, the need for precise measurements
of actual water savings has intensified.
The EBMUD Indoor Residential Water Conservation Study is the second in a series of
three intervention studies that are providing important information on water conserving fixtures
and appliances. The first study was conducted in Seattle, Washington and the third is underway
in Tampa, Florida. The EBMUD Indoor Residential Conservation Study measured the impact of
a variety of water using fixtures and appliances through a before-and-after paired comparison of
water use patterns from a sample of 33 single-family homes in East Bay Municipal Water
District (EBMUD) service area. EBMUD supplies water and provides wastewater treatment for
parts of Alameda and Contra Costa counties on the eastern side of San Francisco Bay in northern
California. EBMUD is a publicly owned utility formed under the Municipal Utility District Act
passed by the California Legislature in 1921 and currently serves 1.3 million people.
The EBMUD Indoor Residential Conservation Study measured the impact of a variety of
indoor water conservation measures on both aggregate and individual water use patterns.
Separate meters were installed to measure hot water usage, adding a valuable new dimension to
the data obtained. Study participants also rated their old fixtures and appliances while they were
still in place, and then rated the new retrofit devices after using them for about six months.
The basic methodology was as follows: two-weeks of specific baseline water use data
were obtained from a sample of 33 homes.1 Next, these homes were retrofit with high efficiency
toilets, clothes washers, showerheads, and faucets. Two weeks of flow trace data were collected
from these homes about a month after the completion of the retrofit and then a second set of








post-retrofit data was obtained about six months later. All of the pre and post-retrofit flow trace
data were disaggregated into relevant end use categories by Aquacraft Inc., using technology
developed for the 1999 Residential End Uses of Water study (AWWA 1999). Paired t-test
analyses were used to evaluate the demands for each end use measured at the study homes in the
pre and post-retrofit periods. This allowed a thorough analysis of the impacts of the retrofits on
the end uses of water in the study homes.
Historic water use patterns of the sample selected for this study (from billing data) tended
to be higher than those from the general population of single-family homes in EBMUD service
area, as shown in Table ES.1. As such, the results should provide a good indication of the
impacts of retrofits on customers who use more water to begin with -similar to a targeted
retrofit program that focuses on larger water users. It is not possible, nor it is the object of this
study to provide conclusive data for the entire nation. The study is a key component in a trio of
studies that will be combined in the near future to provide a more national perspective.2


Table ES.1 Annual water use in 1999 and 2000, sample frame and study group
Average Annual Sample Frame Study Group
n=1000 n=33
Water Use
(gallons per day) (gallons per day)
1999 Total 290.8 386.5
1999 Indoor 195.5 231.2
1999 Outdoor 95.4 155.3
2000 Total 294.6 387.5
2000 Indoor 204.5 237.5
2000 Outdoor 90.1 150.0


RESULTS


Using data collected with the flow recorders, indoor per household and per capital demand
were measured. The logged mean daily indoor demand, which was 191.0 gpd per household
during the baseline period, dropped 35.5 percent to 123.3 gpd after the installation of the new


1 These homes were selected from those customers that expressed a willingness to participate, had not previously
performed extensive retrofits, and who had an average daily per capital use higher than 60 gcd
2 A similar study was conducted m Seattle, Washington and another is planned for Tampa, Florida








devices On an annual basis this equates to an indoor use of 69 7 kgal for baseline conditions

and 45 0 kgal with the retrofit


Per Capita Demand

Indoor water use patterns changed significantly after the conservation retrofit Average daily per

capital use decreased i 31 of the 33 study homes After the retrofit, leakage (17 1 percent), which

had previously been the largest component of indoor use dropped below toilets into fourth place

Toilets (18 6 percent), which have previously been the second largest component of indoor use

moved into third place behind faucets Showers became the largest indoor water use followed by

faucets and toilets Pre and post-retrofit pie charts showing the relative importance of each end

use by percent per capital is shown in Figure ES 1 The combination of showers and baths form

the largest block of indoor use in the post-retrofit era at 25 5 percent


OTHER BATH OTHER BATH
01% 35% 08%
TOILET (0 1 g) (30 gd) TOILET (0 4 gd) (2 8 d)
231% CLOTHES WASHER 18 6%
(19 9 gcd) 16 1% (13 9 gcd) (9 8 go) CLOTHE$WSHER
DISHWASHER 167% (88 g-E)
12% (10 gcd) DISH ASHER
SHOWER FAUCET SHOWER 17% (0 9 gd)
139% 122% 203%
(12 0 god) (10 5 gcd) (10 7 god) FAUCET
19 9%
(10 5gcd)
LEAK LEAK
298% 11le Total: .2 god 1 69%
298% Baselne Total: .2 gd ( Post-Retrofit Total: 52.G gcd
(257 ged) (89 god)


Figure ES.1 Comparing pre-retrofit (on the left) and post-retrofit (on the right) indoor per
capital water use percentage including leakage


Table ES 2 presents a comparison of the mean indoor per capital water use from the

baseline ad post-retrofit data collection periods Overall, indoor water use decreased by 33 9

gcd a 39 4 percent drop A series of unpaired t-tests were performed on each end use in these

two data sets to determine which changes in water use are statistically significant at the 95

percent confidence level Statistically significant changes in water use were detected for clothes

washers, leaks, toilets, and total door use








More than 30 gallons (or 88%) of the 33.9 gcd average saved through the retrofit was the
result of three end uses: toilets, clothes washers, and leaks. Installation of ULF toilets, including

some dual flush models saved an average of 10.1 gcd. The new conserving clothes washers
saved an average of 5.1 gcd. A reduction in leakage resulted in a surprisingly large savings of

16.8 gcd. The leakage savings were almost certainly the result of the toilet retrofit. Toilet leaks,
primarily flapper leaks, are the single largest contributor to household leakage. In this study,

replacing old toilets through the retrofit eliminated almost all of these toilet leaks and resulted in
substantial savings. None of the other measures implemented through this study (clothes

washers, showerheads, or faucet aerators) should have had any impact on the leakage rate.
Statistically significant reductions in water use occurred in most of the end use categories

impacted by the retrofits: toilets, leaks and clothes washers. The shower savings was relatively

small, 1.3 gcd, and only found to be significant at the 90 percent confidence level. Faucets did
not show any significant water use reduction, even though new aerators were installed. The

remaining categories not targeted by the retrofit (baths and dishwashers) also showed no change.


Table ES.2 Mean indoor per capital water use,
Category Baseline Post- Difference
(gcd) Retrofit in Means
(ged) (ged)
Bath 3.0 2.8 -0.2
Clothes washer 13.9 8.8 -5.1
Dishwasher 1.0 0.9 -0.1
Faucet 10.5 10.5 0.0
Leak 25.7 8.9 -16.8
Shower 12.0 10.7 -1.3
Toilet 19.9 9.8 -10.1
Indoor 86.1 52.2 -33.9
Other/Unknown 0.1 0.4 0.3
Total 86.2 52.6 -33.6
Avg. # of 2.56 2.52
Residents per
household
*95 percent confidence level


baseline and post-retrofit3
% Statistically
Change significant
difference?*
-6.6% No
-36.7% Yes
-10.0% No
0.0% No
-65.4% Yes
-10.8% No
-50.8% Yes
-39.4% Yes
75.0% Yes
-39.0% Yes


3 The calculation of per capital per day usage was done on a day by day, house by house basis and then the average
of all individual houses was taken to calculate the overall average per capital per day use This creates a weighted
average where the water use in each household is given equal weight The average number of residents per
household was also calculated on a day by day, house by house basis Multiplying these two weighted averages
together to calculate average daily per household use results in a different value than by taking the average of daily
use for each household










Hot Water Use


Water meters were installed on the hot water heaters of 10 of the 33 study homes and
flow recorders were attached to these meters so that hot water usage could be monitored
alongside overall household usage. Toilet flushing was the only indoor use that had no hot water
component. Only 7 percent of the total leaks were composed of hot water. In the post-retrofit
period, 30 percent of all water used indoors, 16.5 gcd, was hot water. On a daily basis, the most
hot water (83.3 percent) was used for faucets, showers, and baths.
Pre and post-retrofit per capital hot water use are shown in Table ES.3. A statistically
significant difference in mean hot water use before and after the retrofit was detected for the
following categories: clothes washers, faucets, and total indoor use. The total hot water use
dropped by 4.6 gcd after the retrofits, and it appears that nearly all of these savings can be
attributed to the retrofit program. Theoretically, the retrofit program could have impacted
clothes washers, faucets, showers, and total indoor hot water use. Although hot water use
declined in almost all end use categories, the change in shower use was found to be not
statistically significant, but the reductions in clothes washer and faucet use were significant. The
retrofit had no impact on leaks of hot water.


Table ES.3 Comparison of baseline and post-retrofit per capital hot water use
Category Baseline Post-Retrofit Difference % Statistically
Hot Water Hot Water (gcd) change significant
Use Use difference?
(gcd) (gcd)
Bath 1.7 1.5 -0.2 -11.8% No
Clothes Washer 1.9 1.0 -0.9 -47.4% Yes
Dishwasher 1.4 1.0 -0.4 -28.6% No
Faucet 8.6 6.2 -2.4 -27.9% Yes
Leak 0.7 0.7 0.0 0.0% No
Shower 6.9 6.0 -0.9 -13.0% No
Toilet 0.0 0.0 0.0 0.0% na
Other/Unknown 0.02 0.01 -0.01 -50.0% No
Indoor Total 21.1 16.5 -4.6 -21.8% Yes
Avg. # of 2.3 2.3
Residents per
household
*95 percent confidence level








Analysis of Water Savings Excluding Leaks


Because of the high level of leakage found in the study homes both before and after the
retrofit, it was decided to examine the water savings exclusive of leakage One possible

explanation for the high leak rate that was found in some of the study participants' homes could

be traced to the District's change in its water treatment process EBMUD converted from
treating water with chlorine to chloramines (chlorine and ammonia) in 1998 An August 1993

AWWA Journal article reported study results showing that chloramines have a more deleterious

effect on elastomers (products widely used in plumbing distribution, especially for toilet flapper
valves) than does free chlorine When a utility converts from chlorine to chloramine, this

negative effect on the elastomers tends to increase incidents of leaks in the home and in the

distribution system
Leakage accounted for 30 3 percent of indoor per capital use prior to the retrofit and 17 1

percent after the retrofit A specific analysis of leakage is presented later in this report Pre and

post-retrofit pie charts showing the relative importance of each end use by percent per capital,
excluding leaks, is shown in Figure ES 2



OTHER BAT
02% 50% OTHER BATH
02% 50%1) 09% 64%
TOILET gdTOILET (o4l) (28gcd)
32 % CLOTHES WASHER
(199g 230% (139g~d) (98g CLOTHES WASHE
S20 0% (88 gcd)

DISHWASHER DISHWASHER
17% (10 gd) 2 1% (099oa)
FAUCET SHO2W4
-.HO. ^ 1w 244% ^/FAUCET
SHOWER 17 (10 7 d) 3
199% (105 gd) (10(1 5 0d)
(120 gcd)
Baseline Total: 60 4 gcd Post-Retroft Total' 49 gcd


Figure ES.2 Comparing pre-retrofit (on the left) and post-retrofit (on the right) indoor per
capital water use percentage excluding leakage


The average baseline per capital per day indoor use -excluding leaks -was 60 3 gcd and

the post-retrofit average was 43 5 gcd By ignoring leaks both in the baseline and post-retrofit








period, the per capital water savings in indoor use becomes 16.8 gcd -a 27.86 percent reduction
in demand. Results of the analysis excluding leaks are presented in Table ES.4.


Table ES.4 Comparison of baseline and post-retrofit per
Category Baseline Post- Difference
(gcd) Retrofit in Means
(gcd) (gcd)
Bath 3.0 2.8 -0.2
Clothes washer 13.9 8.8 -5.1
Dishwasher 1.0 0.9 -0.1
Faucet 10.5 10.5 0
Shower 12.0 10.7 -1.3
Toilet 19.9 9.8 -10.1
Indoor 60.3 43.5 -16.8
Other/Unknown 0.1 0.4 0.3
Total 60.4 43.9 -16.5
Avg. # of 2.56 2.52
Residents per
household
*95 percent confidence level


capital daily use excluding leaks
% Change Statistically
significant
difference?*
-6.6% No
-36.7% Yes
-10.0% No
0.0% No
-10.8% No
-50.8% Yes
-27.9% Yes
75.0% Yes
-27.3% Yes


Toilet Savings Comparison


A number of studies have measured water savings achievable from installing ULF toilets.
The savings found in each of these studies are shown in Table ES.5. The savings found in the
East Bay study were similar to the REUWS and Seattle study. The highest savings were found
in the statistical models developed for Southern California. The savings from this study were
almost twice as much as those found in the 1991 Stevens Institute study also conducted in the
EBMUD service area. In the 1991 Stevens study, the average flush volume was found to be 1.8
gallons per flush (gpf) and in this study the average flush volume was found to be 1.48 gpf. The
decrease could be attributed to this study's inclusion of newer models and dual flush toilets. In
addition, the Stevens study found the number of daily flushes per person to be 3.7 and this study
found the number of flushes per person per day to be 5.7. These research efforts each
approached the task of calculating savings differently yet their results are reasonably similar.








Table ES.5 Comparison of ULF
Research project


EBMUD Residential Conservation Study (2002)
Seattle Home Water Conservation Study (2000)
REUWS (1999)
MWD (1992 1994)
Tampa, Florida (1993)
East Bay MUD (1991)
Boulder Heatherwood (1996)
*Saturation rate of ULF Toilets after retrofit


savings from
ULF
Flush
Volume
(gal/flush)
1.48
1.38



1.8


other studies

Per capital
savings from
ULF toilets
(gcd)
10.1
10.9
10.5
11.4
6.1
5.3
2.6


Saturation
rate of ULF
toilets in
study homes
85%o
84%*
100%
73%
100%*
100%
50%*


Clothes Washer Savings


A few other studies have measured water savings achievable from installing conserving
clothes washers. The per capital per day clothes washer savings found in these studies is

compared with the EBMUD results in Table ES.6.
The measurements of per capital savings vary in these six studies, although in the three

most recent studies the savings are all quite similar.

The EBMUD and Seattle study both used Frigidaire and Whirlpool clothes washers. The
Maytag Neptune was tested in Seattle and the Fisher & Paykel Ecosmart was tested in the East

Bay. The SWEEP study used Frigidaire clothes washers exclusively. The similarity in water
savings found in the Seattle, SWEEP, and EBMUD studies suggests an approaching agreement

on the impact of these specific machines on per capital water use.


Table ES.6 Comparison of clothes washer savings from other studies
Research project Per capital savings from
conserving clothes washers
(gcd)
EBMUD Residential Conservation Study (2002) 5.2
Save Water & Energy Program -SWEEP (2001) 5.3
Seattle Home Water Conservation Study (2000) 5.6
Westminster water wise homes (1999) 4.6
Bern Kansas (1998) 7.2
Boulder Heatherwood (1996) 10.9
*Estimated from % water reduction reported








Customer Satisfaction Ratings


About six months after installation of the new fixtures and appliances the study
participants were asked to rate their performance. Each participating household was asked to
complete a nine page, 44 question "New Product Information and Satisfaction Survey" that
sought information about customer satisfaction with each of the products installed and with
participation in the study. Many of the questions were intentionally made identical to questions
asked on the initial Audit Survey so that responses could be compared.
The results of the survey were extremely favorable to the high efficiency fixtures and
appliances particularly toilets and clothes washers. This is perhaps surprising given the often
repeated assertions (often based on unscientific anecdotal evidence) that these devices are less
satisfactory.


Toilets
Table ES.7 shows the results of the questions regarding toilet performance. Two trends
are evident in these results: the new ULF toilets were uniformly rated higher in performance than
any of the old toilets and second, looking strictly at the old toilets, the customers preferred the
ULF models to the standard toilets. With respect to the new ULF toilets used for this study, it
should suffice to note that they were rated higher in every category.
Table ES.7 Pre and post-retrofit toilet rating
Rating Category Pre-Retrofit Post-Retrofit
Non-ULF Toilets ULF Toilets
Bowl Cleaning 3.56 3.70
Flushing performance 3.44 4.00
Appearance 3.26 4.58
Noise 3.41 4.42
Leakage 3.59 4.55
Maintenance 3.52 4.58
Overall Average 3.46 4.31
Rating scale from 1 5 where 1 = unsatisfied and 5 = completely satisfied


Clothes Washers

Most of the respondents (70 percent) liked their new clothes washer better than their old
one and only 9.1 percent liked it less. Eighty-five percent said they would recommend the








machine to a friend, six percent would not recommend their new machine, and 9 percent were
unsure. Nearly half of the respondents (48 percent) agreed that if they were in the market for a
washer they would be willing to pay a premium of $150 to get an equivalent quality conserving
washer. Thirty-three percent said they would not be willing to pay the extra money and another
19 percent were unsure.
Study participants rated the performance of their existing clothes washers during the
initial audit interview. As part of the New Product Information and Satisfaction Survey they
were asked to rate their new washer on exactly the same points. The responses to both surveys
are shown in ES.8. Participants rated the new clothes washers higher in every single category.
Of note were the substantially higher ratings of the new machines for noise and moisture content
of the clothes. The new machines scored above 4.5 overall and were particularly praised for
cleaning of clothes, moisture content of clothes, and detergent use. The old machines did not
score above 4.5 in any category. Respondents also expressed satisfaction with the wash cycle
time, cycle selection and reliability of the machines. The new machines scored a 4.33 rating for
noise and capacity.


Table ES.8 Comparison ratings of non-conserving and conserving clothes washers
Rating Category Non-Conserving Conserving
Clothes Washer Clothes Washer
(n=33) (n=33)
Cleaning of clothes 4.23 4.70
Reliability 4.43 4.48
Noise 3.17 4.33
Moisture content of clothes 3.57 4.73
Cycle selection 4.11 4.45
Capacity 4.11 4.33
Wash cycle time NA 4.45
Detergent use NA 4.61
Overall Average 3.94 4.51
Rating scale from 1 5 where 1 = unsatisfied and 5 = completely satisfied


Costs and Benefits


This study was not specifically designed with cost-benefit analysis in mind, but it was
possible to utilize the results to calculate cost of each conservation measure fixture and the value








of the water saved. For this analysis three conservation measures were considered: toilets,
clothes washers, and showerheads.
The payback time for installing 2 ULF toilets (from savings on water and wastewater
charges) was between 3.7 and 6.4 years. The payback time for upgrading to a high-efficiency
clothes washer (from savings on water, wastewater, and energy charges) was between 1.1 and
2.9 years. The payback time for installing a new showerhead (from savings on water and
wastewater charges) was 3.5 years. Cost and benefit analysis was not performed for faucet
aerators.


Conclusions


This study found that significant, verifiable indoor water savings can be achieved through
the installation of high efficiency plumbing fixtures and appliances. Not only did these high
efficiency fixtures save water, but on average, participants reported that they worked better than
their old non-conserving fixtures. An analysis of benefits and costs showed that these products
pay for themselves in water and sewer cost savings within the expect life of the product.

Recommendations
The results from this study make it clear that residential retrofits from the customer
perspective can be a cost-effective tool for saving water and that customers are quite satisfied
with the performance of the new high efficiency toilets and clothes washers currently available.
These results provide powerful evidence of the effectiveness of interior water conservation
measures and justification for continued support of cost-effective programs across the country.
The effects of conservation retrofits is an important area for future research. Clearly, the
more sites that can be included in similar projects, the better and more reliable the results will be
for generalizing to wider populations. Examination of the variability in the reductions in water
use across several cities is an essential part of determining the ability to make generalizations
from the results. A similar study is underway in Tampa, Florida and when that is concluded the
results from all three studies will be combined into a single report document published by the US
EPA.








Ongoing Research
Tracking the consumption of the EBMUD study group via billing data, and collecting
more end use data after 2 years or more time has elapsed is an important to confirm the stability
of the savings. The persistence of water savings over time is a critical component in water
supply planning that includes water efficiency and more research in this area is needed. There is
also interest in conducting more research into the capabilities and accuracy of the flow trace
analysis technology used in this study.
Future studies should also include additional water saving technology such as one gpf
toilets, instant hot water systems, and hands free faucet controllers. While these may not be
economically justified strictly on the basis of water savings, many customers or builders may
wish to include them for their convenience, and their water savings should be evaluated.








CHAPTER 2 METHOD OLOGY


The EBMUD Indoor Residential Water Conservation Study evaluates the impacts and
acceptance of high quality water conservation products in single-family homes. The study was
funded by the U.S. Environmental Protection Agency (EPA) and the East Bay Municipal Utility
District (EBMUD). Aquacraft, Inc. (the consultant) conducted the study with project
management and assistance from Richard Harris, Richard Bennett, and Daniel Muir of EBMUD.
Work on the project commenced in December 2000.
In this study, 33 single-family homes were equipped with new water conservation
fixtures including toilets, clothes washers, showerheads and faucet aerators. Extensive data were
collected before and after the installation of these products so that changes in water use could be
measured. Because the data were collected using flow trace technology it was possible to
disaggregate them into individual end-uses. This allowed the impacts of each fixture and
appliance to be detected directly, including ancillary uses such as leakage. In order to provide
input on user satisfaction the participants were asked to rate both their old fixtures prior to
retrofit and the new products using a consistent set of criteria for both.
This project is the second of three residential retrofit studies conducted for the EPA. The
first was conducted in Seattle, Washington in 1999 and 2000. The third will be conducted in
2002-03 in Tampa, Florida.
The EBMUD Residential Water Conservation Study consisted of five steps:
1. Selection of study participants
2. Initial site visits, audits and data collection
3. Retrofit planning and installation
4. Post-retrofit data collection and customer survey
5. Analysis of results and report writing


This chapter provides an overview of the study group selection methodology used in this
project and the planning and installation of the high efficiency plumbing fixtures.








SELECTION OF STUDY PARTICIPANTS


First Team Meeting

Work on the project began with a project kickoff meeting on December 14, 2000. The
meeting was held at the EBMUD offices in Oakland and was attended by staff members from
Aquacraft, Inc. and EBMUD. A number of decisions were reached at this meeting concerning
how the project should proceed including: the project schedule and timeline, the critical path,
study group selection methodology, selection of which high efficiency fixtures and appliances to
evaluate, and channels of communication. The study team also spoke with Allen Dietemann and
Tim Skeel from Seattle Public Utilities who offered suggestions for fine-tuning based on the
results of the Seattle retrofit study.
The list of fixtures and appliances to be used for the retrofits was finalized at the meeting.
These included 1.5 gpm bathroom faucet aerators, 2.2 gpm kitchen faucet aerators, 2.5 gpm
showerheads, and a variety of ultra low flush (ULF) toilets including the Caroma (1.6/0.8) gallon
per flush (gpf) dual flush model with vitreous tank, the Sloan 1.0 gpf pressure assist model, and
the Niagara flapperless (1.6 gpf) with standard bowl.
An equal number of high efficiency clothes washers were to be purchased from
Frigidaire, Whirlpool and Fisher & Paykel. The only other device considered for inclusion in the
retrofit program was a hands free activator for the kitchen faucet, called the Aqua LeanTM. This
device allows the user to turn the sink on and off without use of hands, by leaning on an
activating bar mounted on the cabinet face under the sink. It was thought that this product
offered the possibility of less continuous faucet use during food preparation and dish washing.
However, this product had not received plumbing code approval hence it was impossible to
include the Aqua Lean in the retrofit program.
The content of the letter of invitation was discussed at the meeting, and the mechanics of
the retrofits were outlined. It was decided to install separate water meters on the supply lines for
the hot water tanks in ten of the study homes to provide a way of quantifying the savings in hot
water use. (This was not part of the original study plan, but was added to provide some insight
into the impacts of retrofits on hot water use.)








Study Group Selection


The goal of the study group selection was to obtain a sample of 33 single-family homes
spread across the EBMUD service area. There were a total of 305,000 single-family homes in the
EBMUD service area at the time of the study, and their average daily indoor use was 190 gallons
per day. To begin the process, EBMUD staff utilized a systematic random sampling procedure
to select a representative sample of 1000 single-family accounts from their entire population of
accounts. To obtain this random sample each single-family account in the EBMUD billing
database was listed in order of their annual water consumption from largest to smallest, and a
random number n was chosen between 1 and 305. Starting with customer n and then proceeding
in even multiples 2n, 3n a list of customers was selected until the end of the list, at which
point customer 1000n represented the 1000t member of the list. The use of a random start
procedure to the sampling process ensured that all members of the sampling frame had an equal
possibility of being selected and hence the list was a true random grouping.

Invitation Letter
An invitation to participate in the retrofit study was sent to approximately 600 of the
1000 selected households in a series of mailings. The invitation packet included a cover letter
from project manager, a description of the study, and a brief questionnaire. These documents are
included in Appendix A.
The questionnaire included questions about the number of people in the household and
the number of fixtures that had previously been retrofit. Approximately 80 questionnaires were
returned to EBMUD. Using the survey data along with historic billing consumption, it was
possible to estimate the average daily per capital use for each responding household.

Selection ofParticipants
Potential participants for the study were selected from those customers that expressed a
willingness to participate, had not previously performed extensive retrofits, and who had an
average daily per capital use higher than 60 gcd. Initially a group of about 40 potential
participants was selected. Each household was contacted by an EBMUD representative to
schedule an initial site visit audit and finalize participation. The final group of 33 participants
was selected so that all geographic regions in the EBMUD service area were equally represented.








DATA MANAGEMENT


Creation of Retrofit Database

To assist with the collection and analysis of the large amounts of data required for this
study, the consultant created a database of participants4. This database contains several
important tables, queries and forms that allow input of information about the customers and
extraction of data needed to meet specific criteria for the selection or analysis processes. All
information generated in the study including the audit surveys and results of the pre and post-
retrofit data collection periods were entered into the database, and it serves as the main
repository of information about the project.

Billing Data

EBMUD provided historic water consumption data from billing records for the 999
homes in their service area selected as the initial sample frame for the study. These billing
records included the name of the current occupant/bill payer, bi-monthly consumption for the
household from January 1999 through February 2001 in gallons, meter read dates, and other
water billing information.
These billing data were organized into a single record per household format to assist with
analysis and the table was included in the EBMUD retrofit database described above.

INITIAL SITE VISITS

The initial site visit was a crucial part of the study. A number of important tasks were
accomplished during these visits. The goals of the visit were to:
Explain the study and the responsibility of participation to the participants
Secure signatures on the participation agreement contract
Complete a detailed customer questionnaire
Inventory all existing water using appliances and fixtures in the house
Determine suitability for installation of new fixtures and appliances
Measure flow rates
Install hot water sub-meters (in 10 of the 33 homes)
Install flow recorders and collect baseline water use data








Visit Protocol


There were at least two people present from the study team for each site visit: one
representative from Aquacraft, one from EBMUD. During the first few days, the visits were
frequently attended by an additional representative from Aquacraft and a plumber hired to install
water meters on the hot water tank.

Audit Questionnaire
An audit questionnaire form was developed by Aquacraft and was based on the survey
form used in the Seattle study. Slight changes were made to obtain additional information and
improve the results. The questionnaire contained approximately 40 questions about the size and
composition of the household -number of adults, teens, and children, year of construction, the
existing water using fixtures in the house, typical water use habits of the residents, satisfaction
ratings of existing fixtures, etc. The questionnaire was reviewed, edited by EBMUD staff. A
copy of this questionnaire is included in Appendix A.

Data Logger Installation
The site visits began on Friday, February 9, 2001. One important goal was to begin
collecting two weeks of pre-retrofit baseline water use data from each participating household.
Upon homeowner agreement a flow recorder (data logger) was placed on the water meter and set
to begin recording water use. These recorders were scheduled to be in place for a total of 15
days each. EBMUD arranged for the old water meter at each home to be replaced with a new
standard magnetic drive meter in order to insure the maximum accuracy of the consumption
data5. Each logger had previously been initialized for local time and synchronized closely to the
auditor's watch. Each data logger was removed only at the conclusion of the 15-day logging
period.
Once the logger was in place, the team began to collect the information from the home.
The Aquacraft staff member administered the audit questionnaire, which involved sitting down
with the customer and asking approximately 40 questions about the home. Each questionnaire
took approximately 20 minutes to complete.


4 Microsoft AccessTM
5 The data loggers are no more accurate than the water meters, so havmg new meters helped insure that the
consumption data would be as accurate as possible








Hot Water Data Collection
At ten homes, the plumber located the hot water tank and installed a 5/8 h-inch water
meter on the cold water feed line during the audit. This turned out to be a fairly simple process,
and in all cases the meter was installed in line above the tank with a few standard fittings.
Normally, the installation of the meter was completed at approximately the same time as the
audit. Each hot water meter was then fitted with a data logger so that simultaneous water use
data could be obtained for both hot and cold water in the home prior to the retrofit.

Collection ofFixture Traces
After completing the survey, the auditor from Aquacraft walked through the home and
operated each fixture in the home and noted the time of each operation. This was intended to
provide a signature trace of each fixture to be captured by the logger. The key to obtaining good
signature traces was to operate each faucet or shower or bath for a long enough time to get a
good sample with the logger that records flow in 10-second intervals. Each fixture was operated
individually for at least 1-3 minutes and each toilet was flushed individually. The next important
step was to allow at least 30 seconds between the operation of each fixture to allow for clear,
discreet water use events. The focus of this process was to get accurate maximum flow rates for
each sink, bathtub and shower so that during the analysis it would be easier to assign fixture
designations for individual events. For example, if the maximum flow rate of the kitchen sink is
2.5 gpm then this fixture can be confidently excluded as the source of any event with peak flows
significantly above 2.5 gpm, even if the volume of the event is comparable with a kitchen sink.
Generally, the more accurate the flow information available the easier it becomes to obtain
accurate disaggregation of water use events.
As a check, the analyst also measured the flow of the fixtures using a calibrated pitcher
that converted a volume captured in 15 seconds to a gallon per minute flow rate. This type of
device can provide flow estimates up to 6 gpm since n quarts per 15 second is equal to n gallons
per minute and one can usually easily handle a 6 quart volume in a calibrated pitcher. It is
important to note, however, that it is not absolutely necessary to perform an in-house audit to
perform the disaggregation. An experienced water use analyst can normally identify the various
fixtures in houses without any flow signatures since they do not vary significantly from house to
house. The audit data and flow rate information from the hand measurement, however, provided
useful information that simplified the analysis.








Participant Agreement
Another task accomplished during the visit was for the EBMUD staff person to explain
the participation agreement to the customer and execute signed contracts. The terms and
conditions of participation in the study were carefully explained to each customer before the
conclusion of the site visit. The key terms included agreement to:
Maintain a log of water use during data collection periods

Not move during study period

Accept fixtures installed for study
It was critical that customers understand and feel comfortable with the participation
agreement to ensure their full participation in the study. A copy of the participation agreement is
included in Appendix A.
The entire audit process typically took between 45 and 90 minutes per household.

Audit Data Entry and QA/QC
Audit data from each household were entered into the database using a customized data
entry form. This form contained numerous data quality checks and data restrictions to prevent
inadvertent data entry errors. Data entry accuracy was independently checked for each
household at the conclusion of the process to ensure the quality of the database.



Retrieval and Verification of Initial Data



The retrieval of the data loggers began 15-days after their installation. EBMUD staff
retrieved the loggers and shipped them back to Aquacraft to be downloaded. As each logger was
retrieved, the ending water meter reading was recorded on a form and the logger was turned off
to end the recording session.
Aquacraft staff members downloaded all the data stored in the loggers to a PC. Each file
was checked for accuracy against the water meter reading to ensure it was operating properly. If
a logger failed to record data or did not record accurate data, a replacement flow trace was
obtained by installing a new data logger on the house and collecting another two weeks of data.








Pre-Retrofit Logging Report


During the first pre-retrofit logging period a total of 43 data loggers were installed at 33
homes -33 loggers on the main water meter outside the home, and 10 loggers on new hot water
meters installed above the hot water tank. From these 45 installed loggers, accurate data were
obtained from 29 of 33 main meter loggers and seven of ten hot water meters. The traces that
could not be used were caused by complete logger failure due to water damage or from electrical
interference in the case of one hot water logger6. These homes were re-logged until good data
were obtained from each home.



Re-Logging Effort



A total of 9 data loggers were installed as part of the re-logging effort. Six loggers were
installed at the homes where good data were not obtained during the first logging period.
Because three homes had also been equipped with hot water meters, three additional loggers
were installed on the hot water meters there, bringing the total to nine.
After the re-logging effort, flow traces were obtained from the main meter at all 33
participating study homes and from the 10 hot water sub-meters. All flow traces included in the
final database met critical accuracy standards (within 5 percent of metered volume).



Pre-Retrofit Flow Trace Analysis



Recorded flow traces from the 33 participating homes (33 main meter traces and 10 hot
water traces) were disaggregated into specific end uses by Aquacraft7. A more detailed
description of the software and the flow trace analysis process is presented in Appendix B.




6 During this logging period there were several heavy rainfalls which caused flooding of most of the meter pits
Several of the data loggers had developed small leaks, which caused them to fall to record when immersed These
loggers were repaired by the manufacturer The electrical interference was remedied by wrapping the sensors in
aluminum foil, which acted as a shield
7 Using the program "Trace Wizard" which was developed by Aquacraft specifically for this purpose









During the flow trace analysis process, the recorded water use flow traces are

disaggregated into water use events using signal processing technology and then fixture

designations are assigned to these events using a pattern recognition algorithm. Specific

statistics are calculated for each individual water use event including its use type, volume, start

time, stop time, duration, peak flow rate, and mode flow rate.

Water use in each flow trace is shown on an interactive graph in Trace Wizard. The

analyst must visually inspect each portion of the trace to ensure proper identification of all water

use events. Analysis for each 15-day flow trace took approximately two hours to complete.

Each analyzed flow trace was reviewed by a senior engineer to ensure accuracy of the analysis

process.

Once the analysis of each trace was completed, two separate water use tables were

created in the database -one for water use recorded from the main water meters and a second for

the water use recorded from the hot water meters. Each water use event (toilet flush, faucet use,

dishwasher cycle, etc.) is included in the database and is associated with a unique keycode

number which identifies the house from with the water use was recorded. An example of this

database is shown in Table 2.1.


KEYCODEJ USETYPE
22010 DISHWASHER
22010 DISHWASHER
22010 LEAK
22010 DISHWASHER
22010 DISHWASHER
22010 LEAK
22010 FAUCET
22010 TOILET
22010 LEAK
22010 FAUCET
22010 LEAK
22010 TOILET


Table 2.1 End use data table example
DATE START DURATION END
2/25/01 1 02 06 PM 50 1 02 56 PM
2/25/01 1 05 46 PM 90 1 07 16 PM
2/25/01 1 07 16 PM 10 1 07 26 PM
2/25/01 1 25 46 PM 90 1 27 16 PM
2/25/01 1 30 46 PM 90 1 32 16 PM
2/25/01 1 32 16 PM 10 1 32 26 PM
2/25/01 2 23 36PM 140 22556 PM
2/25/01 2 24 56 PM 60 2 25 56 PM
2/25/01 2 25 56 PM 10 22606 PM
2/25/01 2 26 16 PM 10 2 26 26 PM
2/25/01 2 39 26PM 20 2 39 46 PM
2/25/01 2 49 26 PM 60 2 50 26 PM


PEAK VOLUME
219 156
1 69 235
0 20 0 03
1 79 229
1 79 229
0 20 0 03
1 99 371
398 289
0 20 0 03
100 017
0 20 0 07
398 299


MODE
1 79
1 66
02
1 66
1 59
02
1 59
1 79
02
1
02
398


MODE NO
2
3
1
3
4
1
7
2
1
1
2
2










Quality Assurance and Quality Control
Numerous quality assurance and quality control measures were taken during the data
collection and analysis process to ensure the quality and accuracy of the data obtained in this
study. These measures included:
Data logger preparation
Water meter calibration

Field verification of data logger operation

Water meter readings upon installation and removal of the data logger

Fixture signature traces

Customer log sheets

Flow rate calibration

Alignment of flow traces with calibrated flow rates

Systematic quality checks of analyzed flow traces

Database integrity testing


These measures are described in detail below.


Data logger preparation Each data logger/flow recorder was charged, initialized, and
tested prior to installation in the field. Loggers were bench tested by Aquacraft by
running a known quantity of water (usually 10 or 20 gallons) through a test meter. This
verified proper operation of the logger and ensured recording accuracy. Loggers that
failed any portion of the testing regime were returned to the manufacturer for repairs.
Only fully charged, fully functional data loggers were installed in the field.
Water meter calibration -The data obtained from the flow recorders is only as accurate
as the water meter it is attached to. EBMUD understood this concept and replaced all
water meters used in that study with brand new Trident T-10 or Badger 25 meters. These
new water meters are tested by the manufacturer and offer high-resolution magnetic pulse
output to the flow recorders.








* Field verification of data logger operation -The Meter-Master flow recorders used in
this study have a special sensor which is strapped to the water meter using a heavy-duty
Velcro strip. Once the sensor and logger are in place it is important to verify that the
logger is installed properly and is picking up the magnetic pulses from the water meter.
This was accomplished by running water through an outside hose bib once the logger was
installed. The Meter-Master logger has an indicator light that flashes after a specific
number of magnetic pulses have been recorded. This flash is repeated 12 times at which
point the light shuts down to conserve battery power. The flashing indicator light signals
that the logger is installed properly and it recording flow through the meter. This proper
functioning was verified during each data logger installation.
* Water meter readings upon installation and removal of the data logger -A key to
verifying the accuracy of the data recorded with the Meter-Master flow recorders is to
compare the volume of water measured by the water meter with the volume measured by
the data logger. These volumes are compared when the data stored in the flow recorder is
downloaded to a PC at the conclusion of the logging period. Careful, accurate water
meter readings were taken when the data logger is installed and again when it is removed.
These readings are recorded on a log sheet and then used as a check against the volume
recorded by the data logger. Agreement between the meter and logger within 5% is the
goal.
* Fixture signature traces -To improve the accuracy of the flow trace analysis process,
fixture signatures were obtained during the audit process in each home. The Aquacraft
auditor intentionally operated each faucet, shower, bath, and toilet fixture individually
during the audit, allowing a minimum of 30 seconds in-between each fixture, and noted
down the time of each operation. During the analysis process these fixture signatures are
carefully examined and used to help identify regular fixture usage during the 15-day
logging period.
* Customer log sheets- Each customer was left with a log sheet on which the number of
persons staying at the house each day was recorded along with the times at which the
clothes washer and dishwasher were operated. This provided better information on
occupancy for calculating per capital consumption, and assisted with recognition of the
two appliances.








Flow rate calibration -During the audits, a special calibrated vessel was use to measure
the maximum flow rate in faucets, showers, and baths or all participants. To use this
device, water is run into the vessel for exactly 15 seconds. The resulting volume
collected in the vessel is then translated into a flow rate in gallons per minutes using
graduated markings. An attempt was made to measure the maximum possible flow rate
from each faucet, shower, and bath in the audited study homes. These flow rates were
noted on the audit form and entered into the Access audit database.
Alignment of flow traces with calibrated flow rates -As a check on the accuracy of
the flow traces, the logged flow rates from the signature draws were compared against the
physical measurements of flow made with the calibrated vessel. Any discrepancy
between the data logged flow rate and the measured flow rate was carefully investigated
and if necessary the flow trace was scaled to match the measurement.
Systematic quality checks of analyzed flow traces -The process of flow trace analysis
with Trace Wizard contains some subjective components. To ensure the accuracy of the
flow trace analysis, two different analysts examined each flow trace in its entirety. The
results from each analysis were compared by summarizing the fixture volumes and
calculating the differences. If any significant discrepancies existed, the traces were re-
examined and finalized in consultation with both analysts.
Database integrity testing -Once all the water use events were assembled into the
database, sorting queries were performed to examine peak flow rate, maximum volume,
and duration for each end use category. A screen for obvious errors is performed (toilets
with a volume of more than 8 gallons, flow rates exceeding meter capacity, etc.). Any
database errors are noted and investigated by reviewing the analyzed flow trace data in
Trace Wizard. Appropriate corrective action is taken.



PLAN RETROFITS AND INSTALL CONSERVING FIXTURES

The goal of the retrofits was to select and install effective high quality indoor water
fixtures on existing housing to determine the water savings that might be available from the use
of these products.








Draft Retrofit Plan


The basic retrofit plan was developed by EBMUD in conjunction with Aquacraft, Inc.
during the initial team meeting in Oakland.
Results from the AWWARF Residential End Uses of Water Study showed that toilets,
clothes washers, showers, and faucets comprise more than 80 percent of indoor water use in a
typical single-family home. Because these are the primary end uses of water it was decided to
focus the retrofits on reducing water use in the following four categories: toilets, clothes washers,
showers, and faucets. This was the same approach taken in the Seattle Home Water
Conservation Study.
The widest variety of conservation options is available for reducing toilet water use. All
toilets currently manufactured in the United States must conform to the Federal Energy Policy
Act standards that mandate a maximum flush volume of 1.6 gallons (6.0 liters). These ultra low-
flush (ULF) toilets represent a substantial reduction in water usage over previous 3.5 and 5.0
gallon per flush (gpf) models. There are toilets on the market that use even less water than the
standard 1.6 gpf ULF models. These include dual flush toilets that offer two flushing modes
one at 1.6 gpf and one at 0.8 gpf, super high-efficiency dual flushing models that flush at 1.0 and
0.5 gpf, and pressure assisted toilets that can flush using 1.1 gpf. Composting toilets that do not
use water are also available, but these are primarily designed for houses without central
plumbing, campers, and boats.
In the 1990s a number of manufacturers began offering high efficiency clothes washers
that use less water and energy than traditional models. Originally, most of the high efficient
washers operated on a horizontal axis (h-axis) and opened on the front of the machine instead of
the top. Manufacturers that offer h-axis machines include Maytag, Frigidaire, and Asko. In the
past few years several companies have entered the high efficiency washer market with top
loading models. Whirlpool and Fisher-Paykel both offer top loading clothes washers that purport
to offer significant water and energy savings. The washers selected for this study are shown in
Figures 1.1 -1.3.s





8 Photos printed with permission courtesy of Frigidaire, Fisher & Paykel, and Whirlpool
























Figure 2.1 Frigidaire Gallery clothes washer Figure 2.2 Fisher & Paykel Ecosmart

ri

















Figure 2.3 Whirlpool Super Capacity Plus Resource Saver clothes washer

Showerheads and faucets, like toilets, are now regulated under the Federal Energy Policy
Act Showerheads must restrict flow to 2 5 gpm and faucets must restrict flow to 1 5 gpm A
wide variety of products are available from numerous manufacturers in the U S
The list of fixtures and appliances to be used for the retrofits was finalized by EBMUD
staff after researching the availability and price of different products These included 1 5 and 2 2
gpm faucet aerators, 2 5 gpm AM Conservation Group showerhead, Niagara Spray Massage
hand held showerhead, 1 6 gallon per flush (gpf) Niagara TM flapperless toilets, and 1 6/0 8 gpf








dual flush Caroma TM toilets Several other toilets were also installed for this study, but not in
sufficient number to use in any meamngful analysis These included the Sloan Flushmate 1 gpf
pressure assist toilet assembly, the Toto Drake ULF toilet, and the Gerber Ultra-Flush pressure
assisted ULF toilet Caroma and Niagara toilets the pnmary fixtures tested in this study, are
shown m Figure 1 4 and Figure 1 5 9 Clothes washers were provided by Fngidaire, Fisher &
Paykel, and purchased from Whirlpool


Figure 2.5 Caroma Caravelle
dual flush ULF toilet


Figure 2.4 Niagara
Flapperless 1.6 GPF toilet


Final Retrofit Plan

A final retrofit plan for each of the participating households was developed based upon
the physical requirements of the home (some houses could only have a specific make and model
toilet installed because of space limitations), requests from the homeowners (some people
expressed a product preference), and the availability of various fixtures EBMUD was in charge
of ordenng the products for the retrofit and finalizing the list of products to be stalled




SPhotos pnnted with permission courtesy of Niagara and Caroma








Some study participants were particularly sensitive about the fixtures they wanted
installed in their home. The staff at EBMUD worked hard to accommodate as many requests as
possible, but in a few cases it proved impossible to get all items required (such as wall mounted
toilets).
During the course of the retrofit, two of the participants withdrew from the study, leaving
33 participants.

Perform Retrofits

EBMUD contracted with a plumber to remove old fixtures and install new ones for this
study. Plumbers were responsible for installing toilets, faucets, and showerheads in study
homes. The appliance dealer who sold the conserving clothes washers also performed the
installation.
Table 2.2 shows the number and make and model of all fixtures installed for this study.

Validate and Tabulate Retrofits

Audit Installation Quality Assurance and Quality Control
To ensure that the proper fixtures were actually installed at the final 33 study homes,
EBMUD completed a product installation form for each house. This survey specified the exact
fixtures (make and model) installed at each home. Initial audit data included a detailed list of
existing fixtures in each study home.
In the follow-up retrofit satisfaction survey, participants were asked to verify the products
installed at their home as part of the study. They were also asked to report on the survey any
installation problems or dissatisfaction.
As a final check, the toilet, faucet, and clothes washer installation was inspected by a
project team member during installation of the hot water meter data loggers. This took place in
only 10 of the 33 homes.










Product

Showerhead


Bathroom
faucet aerator
Kitchen faucet
aerator
Clothes washer
Clothes washer
Clothes washer
Toilet
Toilet
Toilet


Toilet


Table 2.2

Quantity


Conserving fixtures
Make


5 Toto


installed in 33 study homes
Model

Spoiler Showerhead 2.5 gpm


1.5 gpm

2.2 gpm

Ecosmart (GWL10)
Gallery (FWT449)
Super Capacity Plus (LSW9245BQ)
Caravelle 305 (0.8/1.6 gpf dual flush)
Ultimate Flush flapperless 1.6
Flushmate 1.1 gpf (St. Thomas Creations
pottery)
G Max


Post Retrofit Logging Quality Assurance and Quality Control
There were two post retrofit logging sessions. Each involved the same procedures for
installation and calibration of the data loggers. The only homes that the study team entered,

however, were those with the hot water meters. A postage paid mail-back log sheet was left at
the home similar to the initial log sheets. The customers were asked to note some example

operation times of all of the new fixtures and appliances in the home to assist with the post
retrofit analysis.


57 AM
Conservation
Group
79 New Resources
Group
20 New Resources
Group
13 Fisher & Paykel
9 Frigidaire
11 Whirlpool
35 Caroma
32 Niagara
2 Sloan








CHAPTER 3 BASELINE WATER USE


To determine the effect of the conservation retrofit on water use in the study homes,
baseline water use data were collected from the study group. Obtaining high quality baseline use
data was critical for this study because all impacts of the retrofit were measured by comparing
baseline use patterns against water use after the retrofit. Historic billing data were obtained so
that the overall impacts of the retrofit could be measured. Disaggregated flow trace data from
the residential end use study, and a new two-week set of flow trace data were obtained so that the
impacts on each specific end use could be evaluated, a task that would be impossible with only
billing data.
EBMUD provided historic water consumption data from billing records for the 1000
homes in their service area that were selected (as described in Chapter 1) for the initial sample
frame. This random sample was selected to be representative of single family households in the
EBMUD service area. The billing records provided included the name of the current
occupant/bill payer, bi-monthly consumption for the household from January 1999 through
February 2001 in gallons, meter read dates, and other water billing information.

ANNUAL WATER USE

Using the data from the random sample of 1000 homes and the sample of 33 homes
selected for the retrofit study, annual use frequency distributions for the years 1999 and 2000
were developed and plotted on the same axis. These results are shown in Figure 3.1. These
distributions show clearly that annual water use in the retrofit sample was higher than in the
sample frame. This is exactly what the researchers were aiming for because it was desired to
retrofit homes with higher indoor water use to determine the potential for water savings. It can
be seen in Figure 3.1 that water use in the sample frame remained quite consistent over the two-
year period. Average daily water use in the study group10 fluctuated more between 1999 and
2000, but the overall shape of the distribution remained essentially the same.
The average daily demand in the sample frame over the two years was 106.8 kgal (142.8
ccf) or 292.6 gallons per day (gpd) and the average annual demand differed by less than 2
percent between 1999 and 2000. The consistency in demand of this group is important because


10 Calculated as total annual use per household divided by 365 days








the homes that did not participate in the conservation retrofit will be used as a control group.
The control group will be used to test if the expected changes in water use in the study group are
in fact due to the retrofit and not some other factor affecting all households in the East San
Francisco Bay area. Water use in the sample group was also remarkably consistent from 1999 to
2000, differing by less than 1 percent. The average annual demand for the sample group over the
two years (1999 and 2000) was 141.3 kgal (188.8 ccf) or 387.1 gpd.


0% I-ik-- 4- -it


Average Daily Use (gallons)
-4-1999 Sample Frame 2000 Sample Frame -&-1999 Study Group ---2000 Study Group


Figure 3.1 Average daily
sample frame (n=1000)


use distributions for 1999 and 2000, study group (n=33) and


The group of 33 homes, which received new fixtures and appliances in this study (the
study group), was selected from the population of homes that expressed a willingness to
participate in the study. These homes were also selected because they used more than 60 gallons
per capital per day (gcd) for indoor use, as calculated from billing data and survey response
information. Table 3.1 compares the total annual, indoor annual, and outdoor average water use,
over two years of the sample frame (n=1000), the study group (n=33).








Table 3.1 Annual water use per household in 1999 and 2000, sample frame and study group
Average Annual Sample Frame Study Group
n=1000 n=33
Water Use
(gallons per year) (gallons per day) (gallons per year) (gallons per day)
1999 Total 106,155 290.8 141,073 386.5
1999 Indoor 71,347 195.5 84,374 231.2
1999 Outdoor 34,808 95.4 56,698 155.3
2000 Total 107,528 294.6 141,436 387.5
2000 Indoor 74,629 204.5 86,683 237.5
2000 Outdoor 32,899 90.1 54,754 150.0


Seasonal Water Use

Seasonal variations in billed consumption provide an indication of the amount of water
that is used for indoor and outdoor uses. Seasonal water use estimates for the sample frame and
study group in 1999 and 2000 are shown in Table 3.1. Indoor use in this table was calculated
using the minimum bi-monthly billing period for the year and multiplying this value by six to
extrapolate across the entire year. This methodology assumes that all water use in the minimum
bi-monthly billing period is used for indoor purposes. Outdoor use is then calculated by
subtracting indoor use from the total billed consumption for the year.
Indoor water use accounted for 69 percent of total use for the sample frame group and 61
percent of total use for the study group. The bi-monthly consumption records for the retrofit
study group are shown in Figure 3.2 for the period from 1999 to 2000. These data show that the
average bi-monthly minimum demand for the 33-home study group was 14.3 kgal (235.2 gpd). If
one assumed that all use above this was used outdoors, then it establishes an estimate for indoor
use of 85.8 kgal per year (235.1 gpd). Outdoor use is estimated at 55.5 kgal per year (152.1

gpd). Outdoor water use in the East Bay area can occur in almost any month of the year,
depending upon the weather. Most frequently there is little or no outdoor use in January and
February, which are typically the rainiest months.









35,000
Avg bl-monthly winter
demand = 14,255 gallons 5
30,000 (234 5 gal per day) 500


1 25,000 -400
0)

S20,000
300

15,000
0
...... 200 0
m 10,000

4 100 >
5,000 4


0





Figure 3.2 Bi-monthly water use in retrofit study group (n=33)



DEMOGRAPHIC INFORMATION

During the site audits that were performed on the retrofit study group a limited amount of

household level demographic data were collected. These data help describe the households

participating in the study and place them in the context of the population of single family homes

in the EBMUD service area and across the United States.

Number of residents per household

Initial audit survey results indicated that the households in the retrofit study had an

average of 2.74 full time residents. There were an average of 2.20 adults, 0.37 children (0-12

years old), and 0.17 teens (13-19 years old). Figure 3.3 shows the frequency of different numbers

of adults, teens, and children in the 33 retrofit homes. Five households (14 percent) had

teenagers, and nearly 26 percent of the households had children. Only one of the study homes








had more than two children. About 23 percent of the households had more than 2 adults residing
full time. None of the houses had more than four adults.


SAdul ts
* Teens
O Children


0 1 2 3 4
Number of People Per Household


Figure 3.3 Household size distribution, (n=37)


During the data logging period each household was asked to report the number of people
staying at the house during each day. This was done so that more accurate measurements of per
capital water use could be made. The average number of people per household during the
baseline logging period was 2.56.

Household Information

Study participants have lived in their current house about seventeen years on average.
The average move in date was 1984. The earliest reported move-in date was 1950 and the most
recent was 1997. The houses themselves are a mix of older and newer homes. The median age
of the houses was 44 years old (build in 1957). The oldest house was built in 1911 and the
newest house was built in 1990. Since many newer houses are already equipped with conserving








fixtures, it is perhaps not surprising that the owners of older houses volunteered to participate in
the study.
The average floor area of the study houses was 2054 square feet (sf). The minimum size
was 900 sf and the maximum was 6000 sf. The typical study house was 2 stories tall, had a 2-car
garage, 4 bedrooms, one full bathroom, and one 3/ bath. Five of the houses had more than 3
bathrooms. Four of the houses (14 percent) already had one ULF toilet installed. Eight of the
houses (22.8 percent) had hot tubs and four (11.4 Percent) had a swimming pool.
Twenty-seven houses (77 percent) had automatic sprinkler systems, and eight of the
households indicated that they intended to irrigate during the two-week logging period. Five
houses (14 percent) had home offices that were used for telecommuting and other work from
home activities. None of the households reported doing laundry outside the home from time to
time.

Fixture Performance

As part of the audit, homeowners were asked to report on and describe problems with
their existing toilets. They were also asked to rate their primary toilet and clothes washer on the
same performance measures as would be applied to the ULF retrofits later.
Twenty-four households (69 percent) reported experiencing a problem with their primary
toilet within in the past 6 months. Twenty-eight households (80 percent) reported experiencing a
problem with their toilet in the past 2 years. The problems reported ranged from leaking flapper
valves to broken chains to clogs. Only two households reported never experiencing a problem
with their toilet.
Twenty-nine percent of the households reported that their toilets never required plunging.
Fifty-four percent reported infrequent plunging ranging from every other month to every other
year. Seventeen percent of the households reported that more frequent plunging was required for
their toilet.
Double flushing once per week or more was reported by ten households (29 percent).
Another 10 households said they double flushed a few times a year. Ten households indicated
that they never need to double flush their toilets.
Homeowners were asked to rate their non-ULF toilets in the following areas on a scale of
1 to 5 (1 = unsatisfied and 5 = completely satisfied): bowl cleaning, flushing performance,
appearance, noise, leakage, and maintenance. The results of this rating are presented in Table








3.2 and compared against similar baseline ratings from the Seattle retrofit study. It can be seen
from this analysis that the study participants were not particularly enthusiastic about their
existing toilets. The ratings in the East Bay were similar, but slightly lower overall than in
Seattle.


Table 3.2 Pre-retrofit toilet rating, Non-ULF toilets in EBMUD and Seattle
Rating Category EBMUD Seattle
Non-ULF Toilets Non-ULF Toilets
(n=54) 11(n=37)
Bowl Cleaning 3.56 3.76
Flushing performance 3.44 3.54
Appearance 3.26 3.70
Noise 3.41 3.32
Leakage 3.59 3.70
Maintenance 3.52 3.89
Overall Average 3.46 3.65
Rating scale from 1 5 where 1 = unsatisfied and 5 = completely satisfied


Using the same rating system (1 = unsatisfied and 5 = completely satisfied), homeowners
were asked to rate their current non-conserving clothes washer in the following areas: cleaning of
clothes, maintenance/reliability, noise, moisture content of clothes, cycle selection, and capacity.
The results of this rating are presented in Table 3.3 and again compared with the results from
Seattle. The results from both study sites were virtually the same. Homeowners appear to be
satisfied for the most part with their existing clothes washers. Respondents were particularly
pleased with the reliability of the machines, the capacity and the selection of wash cycles
available. They were less satisfied with the noise the machines make and the moisture content of
the clothes after washing.
The participants in this study had the following selection of clothes washer brands: 12
Whirlpool (34 percent), 10 Maytag (29 percent), 7 Kenmore (20 percent), 2 GE (6 percent), 2
Hotpoint (6 percent), 1 Admiral (3 percent), and 1 Norge (3 percent). The average age of the
machines was 11 years old (purchased in 1990). The oldest machine was purchased in 1970 and
the newest in 2001.










Table 3.3 Pre-retrofit rating of non-conserving clothes washers, EBMUD and Seattle
Rating Category EBMUD Seattle
Non-Conserving Clothes Non-Conserving Clothes
Washer (n=33) Washer (n=37)
Cleaning of clothes 4.23 3.86
Reliability 4.43 4.44
Noise 3.17 3.32
Moisture content of clothes 3.57 3.50
Cycle selection 4.11 4.06
Capacity 4.11 4.30
Overall Average 3.94 3.91
Rating scale from 1 5 where 1 = unsatisfied and 5 = completely satisfied



END USE DATA



The water use data collected and analyzed using the flow recorders and Trace Wizard
software contains specific information on the use of water in each study home over the two week
data collection period. These data were analyzed in a variety of ways so that the impacts of the
retrofit can be measured by comparing water use after the retrofit with the baseline demand
patterns. The baseline water use analyses included daily use, daily per capital use, per capital use
for different fixtures, the frequency and intensity of use of various fixtures, and the variability of
water use. Where useful, these baseline results are compared to the findings from the
AWWARF Residential End Uses of Water study national sample.



Daily Household Use



A total of 513 complete days of end use data were recorded from the 33 study homes in
the EBMUD service area for an average of 14.7 days of data per household. A minimum of 14
days of data was collected from each study house, and more than 14 days of data were obtained
from a few houses. Baseline end use data were recorded from February 10h through May 10",
2001, but the majority was obtained between February 10h and April 7 2001.

1 A total of 54 toilets were rated by homeowners









The total daily use (including indoor and outdoor uses) from each logged day is plotted as

a scatter diagram in Figure 3.4. The average daily use for all houses was 210.4 gallons per day

(gpd) and the standard deviation was 162.8 gpd. The median daily use was 182.5 gpd. The

maximum observed daily use during the logging period was 2241.3 gpd, which is actually fairly

low when considering that the maximum average daily use from the REUWS was more than

9000 gpd. These results suggest that there was some outdoor use during the baseline logging

period, but there were also many days without any outdoor use.


Average daily use = 210.4 gal.
Standard Deviation = 162.8 gal.
Median daily use = 182.5 gal.
2000




1500



1, 1000
200 -S-------------------------------------------






500-------






Observations (n=513)



Figure 3.4 Scatter diagram of study group pre-retrofit daily household water use



These same daily use data were used to develop a frequency distribution histogramm),

Figure 3.5. Figure 3.5 shows that nearly 50 percent of daily water use was less than 175 gallons

per day and 90 percent of daily use was less than 350 gallons per household per day.













10% 100%


>% 8% 80% u
u Average daily use = 210.4 gal.
Standard Deviation = 162.8 gal.
Median daily use = 182.5 gal.
lu 6% 60%


) E
W 4% 40% =
0


2% 20%


0% + + + + + +,+0%


Daily Water Use (gal.)



Figure 3.5 Baseline frequency distribution--total daily household water use for pre-retrofit
study group




INDOOR PER CAPITAL USE

Per capital indoor water use was calculated daily for each individual study home using the

daily indoor water use obtained from the flow trace analysis and the day by day reported number

of residents in each house from the log sheets. Averages of per capital indoor use for the group

were made from the individual daily per capital use values calculated for each identified end use.

Days where it was reported that no one was at home were excluded from the set reducing the

number of observed days from 513 to 507.

Leakage was the largest component of baseline indoor per capital water use among the 33

study homes, accounting for 30.3 percent of indoor demand. Toilets were the second largest

component of indoor use at 22.7 percent followed by clothes washers at 16.0 percent, showers at

13.5 percent, faucets at 12.1 percent, baths at 3.9 percent, dishwashers at 1.3 percent, and







other/misc. use at 0.1 percent. Figure 3.6 shows the percentage breakdown of all indoor water
uses collected form the 33 study homes.


OTHER

TrOILT (0.1 god)
23.1%
(19.9 gQ



SHOWER
13.9%
(12.0 ged)


LEAK
29.8%
(25.7 qod)


BATH

S(3.0 god)
CLOTHES WASH ER
16.1% (13-9 gcd)

DISHWASHER
F1.2% (1.0 gcd)
FAUCET
12-2%
(1065g ed)


Baseline Total: 88.2 god


Figure 3.6 Baseline indoor per capital water use percentage including leakage


Mean Per Capita Indoor Use

The baseline average indoor per capital indoor water use in the EBMUD study homes was
86.2 gallons per capital per day (gcd) and the median was 70.2 gcd. The average for the EBMUD
retrofit group was 17 gcd more than was found in the REUWS national sample which was 69.3
gcd.12 Most of the difference was the result of the high leakage rate in the EBMUD study group,
but a significant portion is also attributable to higher toilet use. This can be seen in Table 3.4
which presents the average per capital water use from the EBMUD retrofit study (pre-retrofit),
the Seattle retrofit study (pre-retrofit), and the REUWS national sample.






12 It should be noted that the number of people per household information from the REUWS is much less reliable
than the data collected for the EBMUD retrofit study








Table 3.4 Indoor Per Capita Water Use Comparison1
EBMUD Retrofit Seattle Retrofit All REUWS
Baseline Baseline
Average Percent Average Percent Average Percent
gcd gcd gcd
Bath 3.0 4.0% 3.7 5.8% 1.2 1.7%
Clothes Washer 13.9 16.0% 14.8 23.3% 15.0 21.6%
Dishwasher 1.0 1.3% 1.4 2.2% 1.0 1.4%
Faucet 10.5 12.1% 9.2 14.4% 10.9 15.7%
Leak 25.7 30.3% 6.5 10.3% 9.5 13.7%
Shower 12.0 13.6% 9.0 14.2% 11.6 16.7%
Toilet 19.9 22.7% 18.8 29.5% 18.5 26.7%
Other 0.1 0.1% 0.2 0.3% 1.6 2.3%
Indoor Total 86.2 100.0% 63.6 100.0% 69.3 100.0%
Sample size 33 37 1188
Avg. # of residents 2.56 2.54 2.8
*Number of residents during logging period


The indoor water use in the EBMUD group of 33 homes in the retrofit study was higher
than the indoor use in the Seattle REUWS sample and the REUWS national sample primarily

because of the high rate of leakage. Leaks aside, water use in the EBMUD and Seattle retrofit
groups were quite similar. Combined shower and bath usage was higher in the EBMUD retrofit

study group (15.6 gcd) than the Seattle retrofit group (12.7 gcd), and the entire REUWS study
group (12.8 gcd). Clothes washer, faucet, and dishwasher usage were quite similar between the

groups. The toilet use was slightly higher in EBMUD. The average number of residents per
household in the EBMUD and Seattle retrofit groups were almost identical (2.55 vs. 2.54), but

was higher in the REUWS group (2.8). With the exception of leakage, the per capital usage in
the retrofit group appears quite typical of the demand patterns observed in the earlier REUWS

research, suggesting the group is fairly typical of single family homes in the EBMUD service
area and in other cities across the country.






13 The calculation of per capital per day usage was done on a day by day, house by house basis and then the average
of all individual houses was taken to calculate the overall average per capital per day use This creates a weighted
average where the water use in each household is given equal weight The average number of residents per
household was also calculated on a day by day, house by house basis Multiplying these two weighted averages
together to calculate average daily per household use results in a different value than by taking the average of daily
use for each household








Confidence intervals were calculated for each of the per capital end uses at the 95%
level. These results are shown in Table 3.5. Confidence intervals are calculated using the
equation:

C=Z(a/ [n)

Where Z = confidence value from z-distribution table (1.96 for 95% confidence interval)

a =the standard deviation

n = sample size


This analysis gives a measure of the range within which the true mean occurs. While
not a perfect measure, because it does assume a normal distribution around the mean, the ability
to perform this allows one to estimate the likelihood that changes in the mean are significant or
not.
Based on this analysis the tightest confidence intervals were found for toilets, faucets,
showers, and total indoor use. In all three of these categories the confidence interval was less
than 8 percent of the mean. Not surprisingly these were also the most frequent end uses of water
and occur virtually everyday there are people at home. End uses such as dishwashers, clothes
washers, and baths that do not occur everyday are likely to have wider confidence interval
measures because the standard deviation of demand is higher.


Table 3.5 Average baseline per capital use and 95% confidence intervals
Average 95% Confidence
gcd Interval
gcd
Bath 3.0 +0.63
Clothes Washer 13.9 2.04
Dishwasher 1.0 0.17
Faucet 10.5 0.63
Leak 25.7 5.74
Shower 12.0 0.91
Toilet 19.9 1.02
Other 0.1 +0.08
Indoor Total 86.2 6.83








Baseline Hot Water Usage


An innovative aspect of this study was the logging of both hot and cold water on a

portion of the study group. Water meters were installed on the hot water heaters of 10 of the 33

study homes and flow recorders were attached to these meters so that hot water usage could be
monitored alongside overall household usage. The hot water flow traces were disaggregated into

end uses using Trace Wizard and the data were stored in the EBMUD database.
There was an average of 2.49 residents in the 10 so-called "hot water" homes during the

baseline data collection period, and their average daily per capital use was 70.9 gcd. Over 21 gcd
or 30 percent of the total indoor use was made up of hot water in these homes.

Overall per capital indoor use in the 10 hot water homes was significantly lower than that
of the study group as a whole. Much of this was due to the lower leakage rates, however, the hot

water homes used less water in all categories except toilets, faucets, and dishwashers. A
comparison with the average gcd found in the entire 33 home study group is presented in Table

3.6. We attribute this use pattern to the normal variations in the group since the hot water houses
were chosen in a random manner.


A


Table 3.6 Baseline per capital hot water use
Baseline Hot Water Use Total Use in Hot
Water Homes
average Percent of Use Average
gcd Type that is Hot gcd


Water
Bath 1.7 89.5%
Clothes Washer 1.9 15.3%
Dishwasher 1.4 100.0%
Faucet 8.6 65.2%
Leak 0.7 6.3%
Shower 6.9 71.9%
Toilet 0.0 0.0%
Other 0.02 66.7%
Indoor Total 21.1 29.8%
Sample size 10 10
Avg. # of residents 2.49*
*Residents in homes with hot water monitoring during logging period


EBMUD Retrofit
Group
Average
gcd

3.5
14.3
1.2
10.8
27.1
12.1
20.3
0.1
89.3
33
2.55








Not surprisingly, 100 percent of dishwasher use is made up of hot water this is the only

indoor use that is entire made up of hot water. Only 15.3 percent of clothes washer usage was

hot water while 89.5 percent of baths and 71.9 percent of showers were hot water.
These results suggest that on an annual basis each person in the hot water homes is using

approximately 7,702 gallons of hot water. It is calculated that the annual cost in gas and
electricity charges of heating 7,702 gallons of water is approximately $28.60.4 The calculated

cost of heating 1000 gallons of water with gas is $2.82 and the cost of heating 1000 gallons of
water with electricity is $7.32.
Confidence intervals were calculated for each of the per capital hot water end uses at the

95% level. These results are shown in Table 3.7. The tightest intervals were for faucets, leaks

and total indoor per capital daily use. Because of the smaller sample size, the confidence

intervals for the hot water group were somewhat larger than for the entire study group


Bath
Clothes Washer
Dishwasher
Faucet
Leak
Shower
Toilet
Other
Indoor Total


Table 3.7 Confidence intervals
Baseline Hot Water Use
Average 95% Confidence
gcd Interval
gcd
1.7 0.69
1.9 0.57
1.4 0.33
8.6 0.93
0.7 0.10
6.9 1.23
0.0 NA
0.02 +0.01
21.1 1.98


for baseline
Total Use in Hot Water Homes
Average 95% Confidence
gcd Interval
gcd
1.9 +0.78
12.4 +3.01
1.4 +0.35
13.2 +1.52
11.2 +2.55
9.6 +1.62
21.4 +1.83
0.03 +0.03
70.9 +5.95


Leaks


The leakage rate in the EBMUD retrofit group was substantially above the average found

in the REUWS, and leakage was the largest use of water in the study group. The leakage data

showed a strong positive skew. That is, the mean value is greater than the median, and








consequently, there are more values less than the mean than greater. In this case, the mean per
capital leakage rate was 25.7 gcd and the median leakage rate was only 4.2 gcd. It was found that
10 houses (29 percent) in the study were responsible for more than 86 percent of the total per
capital leakage during the baseline period and four houses (11 percent) were responsible for an
astonishing 64.5 percent of the leakage. The top two leaking homes by themselves were
responsible for nearly 43 percent of the total leakage. These homes each leaked approximately
210 gcd! While leakage is clearly a major problem in this group of homes as a whole, it is really
only significant problem in a small number of homes. This result is quite similar to the findings
from all other study sites in the REUWS -a small number of houses are responsible for most of
the leakage.
The first high leakage homes (296 gcd) had two substantial leaks. The largest was a
persistent flapper leak shown in Figure 3.7. It is believed that this leak was the result of
improperly seated flapper valve on a toilet in a back bathroom that is seldom used. As shown in
Figure 3.8, the leak stops when the toilet is flushed about halfway through the logging period.
The second leak in that home (also evident in Figure 3.7 and Figure 3.8) is a low flow-rate leak
that continues throughout the flow trace. The source of this leak is not known, but a running
toilet, open faucet, leaky sprinkler valve, or even a slow leaking water pipe could cause it.
The second high leakage home also had a continuous low flow leak between 0.07 and
0.11 gpm. This leak resulted in the loss of approximately 144 gallons per day. A distribution of
daily per household leakage is shown in Figure 3.9.


14 Assumptions 80 percent of the homes heat water with gas and 20 percent heat with electricity, the water starts out
at a temperature of 55 F and is heated to 105 F, gas costs $0 44 per Them and electricity costs $0 06 per kWh














I. l iii11 1


Figure 3.7 Persistent flapper leak


Figure 3.8 End of flapper leak with toilet flush (green)









120%


25% -- vieaan = -4 o gpa -100%
Minimum = 00 gpd
Maximum = 662 9 gpd

>, 20% 80% o
u a)


u. 15% 60%


S10% 40%


5% 20%


0% 0%



Daily Per Household Leakage (gal.)

Figure 3.9 Daily per household leakage distribution, pre-retrofit study group


One possible explanation for the high leak rate that was found in some of the study

participants' homes could be traced to the District's change in its water treatment process.

EBMUD converted from treating water with chlorine to chloramines (chlorine and ammonia) in

1998. An August 1993 AWWA Journal article reported study results showing that chloramines

have a more deleterious effect on elastomers (products widely used in plumbing distribution,

especially for toilet flapper valves) than does free chlorine. When a utility converts from chlorine

to chloramine, this negative effect on the elastomers tends to increase incidents of leaks in the

home and in the distribution system. The plumbing industry has responded to this problem by

marketing elastomer products with compounds resistant to attack by chloramines.








FIXTURE USAGE


Toilets

The data set developed for this study made it possible to calculate the number of times
per day each fixture was used and the volume of use per fixture. It is important to compare these
results against the fixture utilization measured after the retrofit to ensure that increased
utilization is not diminishing efficiency savings, which is a commonly expressed concern.
During the baseline data collection period a total of 6,051 individual toilet flushes were
recorded from the 33 study homes on 513 days for an average of 11.7 flushes per household per
day and 5.14 flushes per capital per day.1 The average flush volume across the 33 study sites
was 3.88 gallons per flush (gpf) with a standard deviation of 1.28 gpf. The distribution of toilet
flushing volume of all recorded flushes is shown in Figure 3.10. This distribution shows the
range of flushing volumes found during the baseline data collection period. Only 7 percent of
the baseline recorded flush volumes were in the low flow range (below 2 gpf). About 60 percent
of the flush volumes were above 3.5 gpf and 15 percent were above 5 gpf.
The median value of flushes per capital per day (fpcd) was 4.78. A daily per capital
flushing distribution for all 507 days on which toilet flushes occurred is presented in Figure 3.11.
This distribution shows that 80 percent of the study homes flushed an average of seven times per
person per day or less.



















15 The value of 5 14 flushes per capital per day was derived by averaging the flushes per person per day determined
for each home on each day of the study This is not the same as dividing the 6051 total flushes by the 513 days of
observation and 2 55 persons per home (which gives 4 63 flushes per person per day)























5%


0%
I 'b 4P b 'pt A A? 4e

Flush Volume (gal.)

Figure 3.10 Baseline toilet flush volume distribution, pre-retrofit study group

18% 1


16%

14%


t1 'b N <3 b VA 0

Flushes Per Capita Per Day

Figure 3.11 Baseline toilet flush frequency distribution, pre-retrofit study group


t.J I" h = -" :1 11 :r1-,.- li. *.J















17_,7


, D b b P q 4e










Showers


Showering accounted for 13.6 percent of the average per capital indoor water use during
the baseline period. A total of 806 individual shower events were recorded during the baseline
period for an average of 0.65 showers per person per day. In the hot water study homes (n=10)
71.9 percent of the shower water was hot water. The average shower used 18.4 gallons, lasted
for 8.9 minutes, and was taken at an average flow rate of 2.0 gallons per minute (gpm). This
indicates that on average, the people in the study group already shower at a flow rate
substantially below the national plumbing code standard of 2.5 gpm. A similar result was found
in the Seattle retrofit study and in the REUWS. This suggests that the actual water savings
achievable through a showerhead retrofit may be less than has been traditionally estimated using
standard engineering techniques.
The distribution of showering volumes is shown in Figure 3.12. This distribution shows
that most showers (81 percent) used between 7.5 and 27.5 gallons. These results are quite
similar to the showering volume calculated in the REUWS. In that study the average shower
volume was 17.2 gallons and the average duration was 8.2 minutes.










- J. 1:. -
_hi Le ,i-,, h:.,C, =
I I,-,J 5, =


I': J II.ll.:.CK
Il! .O ,.,l l.:,Z,-,-5
Il I !" ,ihll,',.-,:


2%W I l-IWU

|i0% |iiil, iEES 3.. E._..


Shower Volume (gallons)

Figure 3.12 Baseline shower volume distribution, pre-retrofit study group


The distribution of shower durations for all recorded shower events during the baseline
period is shown in Figure 3.13. In this figure, 77.51 percent of all showers were between 3 and
12 minutes in length with a mean of 8.9 minutes and a standard deviation of 4.4 minutes. Less
than 10 percent of all showers were longer than 15 minutes in duration.


12%


6% -











IAveraae duration = 8 9 minutes
1 2 % -: iJ J .- J i l ,:.,- = ', ,-, ,i =


10%
u
c
U- 8%
LI.
>6%
n-

4%





0%
N% %
Shower Duration (minutes)


Figure 3.13 Baseline shower duration distribution, pre-retrofit study group


The distribution of shower flow rates for all recorded showers during the baseline period

is shown in Figure 3.14. For this chart the mode flow rate statistic generated by Trace Wizard

during flow trace analysis was taken as the actual shower flow rate because it best represents the

flow during the shower itself. An average flow rate might over-estimate shower flows because

many showers start at a high flow rate is water is run through the bathtub spigot and the

temperature adjusted then the flow is restricted when the shower diverter valve is used and flow

is constricted through the shower head.

The mean shower flow rate during the baseline period was 2.0 gpm with a median of 1.89

gpm and a standard deviation of 0.96 gpm. The distribution of shower flow rates drops off

sharply after the 2.25 gpm flow rate and appears less regular than either the distribution of

shower volumes or the distribution of shower durations. More than 84 percent of the showers

recorded during the baseline period were taken at a flow rate below 2.5 gpm, which suggests that

flow rate reductions may only be accomplished on roughly 16 percent of the showers as a result








of the showerhead retrofit. All of the shower distributions shown in this section are compared
against post retrofit distributions later in this report.



25%


o 0


Shower Flow Rate (gpm)
Shower Flow Rate (gpm)


Figure 3.14 Baseline shower flow rate distribution, pre-retrofit study group



Clothes Washers

All of the homes in this study had to have a washing machine as a requirement of
participation. A total of 425 loads of laundry were washed in the 33 study homes during the
baseline period for an average of 0.36 loads of laundry per person per day, a use rate which is
quite consistent with previous findings.
The average volume used to wash a load of clothes was 40.7 gallons with a standard
deviation of 9.8 gallons. Nearly 89 percent of the washer loads used 50 gallons of water or less
and 99 percent of the washer loads used less than 70 gallons of water. In the three loads that
used more than 70 gallons it appeared that one or more extra rinse cycles were used. The largest
individual cycle (wash or rinse) observed was approximately 38 gallons.


20%


10%


Average flow rate = 2.00 gpm
SStd. Deviation = 0.96 gpm
Median flow rate = 1.89 gpm














.ii iilli.. __ I









A few sample clothes washer flow traces from Trace Wizard are presented below Figure

3 15 shows a 1995 Maytag Super Capacity washer The total volume for this load of clothes was

45 0 gallons The four cycles shown include the wash cycle, two nse cycles, and a spin cycle

Tlus machmie often operated with only a single nnse cycle



I I I I


Figure 3.15 Sample clothes washer flow trace, 1995 Maytag Super Capacity


Figure 3 16 shows a 1990 Whirlpool Heavy Duty model clothes washer The total

volume for washing this load of clothes was 40 6 gallons The two sets of short spike-like cycles

are spin cycles


I I ,
I I I


I .. I
S I It


Figure 3.16 Sample clothes washer flow trace, 1990 Whirlpool Heavy Duty









Results from the hot water homes show that 15.3 percent of the water used for clothes
washing is hot water. This was substantially lower than the 27.8 percent found in the Seattle
retrofit study. From the 10 hot water flow traces, it appears that the first cycle (the wash cycle)
is typically the only cycle to contain any hot water. Subsequent cycles (rinse and spin cycles) are
almost exclusively cold water with a few exceptions. In the EBMUD homes many of the clothes
washer cycles contained no hot water at all.
Given the average volume per load of clothes found during the baseline period there
appears to be considerable opportunity for conservation savings with the new conserving
machines if they actually meet their advertised usage of no more than 30 gallons per load.



Dishwashers

A total of 155 loads of dishes were washed by machine during the baseline study period.
On average, dishwashers were used 0.13 times per person per day -about 0.9 times per person
per week. The average load of dishes used 8.9 gallons of water. The maximum amount of water
used by any dishwasher was 18.2 gallons. More than 80 percent of the dishwashers in the study
homes used between 6 and 12 gallons per load.
All of the water used in the dishwashers was hot water. So, while the volume of their use
is relatively small they are significant energy users.



Faucets

Faucet use accounted for 12.1 percent of the total indoor water use during the baseline
study period and 65.2 percent of faucet use was made up of hot water. A useful means of
evaluating faucet utilization is to calculate the duration faucets are utilized per capital per day.
During the baseline period it was found that faucets were used an average of 10.4 minutes per
person per day. In the REUWS, faucet use averaged 8.1 minutes per person per day.
The typical baseline faucet flow rate was 1.2 gpm, based on an analysis of the mode
faucet flow rates calculated using Trace Wizard. Typical peak baseline faucet flow rate was 2.9
gpm. Similar to the situation with showers, it appears that many of the homes in this study either








already have faucet flow restrictors or simply throttle their faucets below the 2.2 gpm federally
establish low flow rate.



Baths

Because baths require a fixed amount of water, this conservation retrofit program is not
expected to reduce bath water usage. With the increasing popularity of larger Jacuzzi tubs in the
past decade, it seems likely that there could be an increase in per capital bath water usage.
During the baseline period, the average bath used 28.5 gallons of water. The maximum
bath usage was 89.9 gallons 1. Baths in the hot water homes were 89.5 percent hot water and
11.5 percent cold water. Study residents took an average of 0.12 baths per person per day or
0.84 baths per person per week. The nine households with children were responsible for more
than half of the baths recorded during the baseline period. Average per capital bath usage in
households without any children under the age of 13 was 3.03 gcd; in households with one child
the bath average was 4.52 gcd; and in households with two or more children the bath average
was 6.17 gcd.
Twenty of the 33 study homes (57 percent) took at least one bath during the baseline data
collection period. Information on bathing habits (time and frequency) was collected during the
audit portion of the study and this information proved useful during the flow trace analysis
process enabling baths to be identified with greater confidence. The hot water traces were also
helpful in accurately identifying baths.



MAXIMUM DAY DEMANDS



Maximum demands are often the driving factor for facility expansions and facility
design. While interior retrofits are not designed with the intention of reducing peak instantaneous
and peak day demand they may achieve this goal anyway.
The maximum day water use for each household in the study was calculated and plotted
from the highest to lowest in Figure 3.17. The average maximum day during the baseline data
collection period was 445 gallons. The largest maximum day household was 2241 gallons.







Because the baseline data collection period extended from winter into spring, outdoor use began
to become more frequent at the end of the data collection period. Typically it is outdoor demand
that drives peak day usage in the single-family sector.
The peak instantaneous demand distribution of all 531 baseline logged days is shown in
Figure 3.18. The average peak instantaneous demand was only 6.1 gpm and the maximum peak
demand was 13.0 gpm. In the REUWS, the average peak instantaneous demand was 8.2 gpm
and the maximum peak was 64.6 gpm. The flows in the EBMUD retrofit study group were
considerably smaller primarily because demand during the baseline period was primarily
indoors. High peak flows are typically observed during automatic irrigation. Baseline peak
flows in the Seattle retrofit study were slightly lower because there was even less outdoor use
during that logging period.


2000 Median = 353 5 gpd
Min= 127 3 gpd
Max = 2241 3 gpd
0 1500


S1000


500 -
I IIIIIIIlIlllIIIIIlll.,.,



Study Household

Figure 3.17 Baseline pre-retofit peak day demand for each study house


16 This house was equipped with a large Jacuzzi tub and water was added several times during the bath












Average = 6 1 gpm
Std deviation = 2 0 gpm
Median = 5 9 gpm
Max = 130 gpm


S6% _______ ----------

4%







Peak Instantaneous Demand (gpm)


Figure 3.18 Baseline pre-retrofit peak daily instantaneous demand distribution








CHAPTER 4 POST-RETROFIT RESULTS


The installation of conserving fixtures in 34 of the 35 participating study homes was
begun in June 2001 and completed in stages by December 2001 and the first set of post-retrofit
end use data was collected during the period from August 8, 2001 to January 10, 2002.17 In June
2002, the participants completed a detailed opinion survey about their experience with the
conserving fixtures and the second and final set of post-retrofit data was collected from May 29
to July 23, 2002. By the conclusion of the study two homes opted to drop out, leaving the final
study group size at 33. Additionally, billing data from the entire sample of 1,000 homes were
obtained through the August 2002 billing period. Aquacraft analyzed all of these data and the
results are presented below.



COMPARISON OF WATER USE DURING LOGGING PERIODS

It was clear that the data analysis would be greatly simplified if the data from the two
post-retrofit logging periods could be combined into a single group. Consequently, one for the
key analyses involved a comparison of the two post-retrofit logging sessions against each other.
If the results of this showed that the two were similar statistically, then it would possible to
combine them for the comparison against the baseline use upon which the assessment of the
retrofits would then be performed.
Table 4.1 shows a comparison of the mean daily indoor per capital water use during the
three logging periods. After the installation of the conserving fixtures, the mean daily indoor per
capital water use dropped from 89.3 gcd to approximately 52 gcd, which represents a 42 percent
reduction. This is based on one baseline and two post-retrofit logging periods in which end use
data were obtained from all 33 participating households for a total of 910 days. It should be
noted that the per capital water use was determined for each home based on the number of people
in the home each day of the study. Days in which no one was home, i.e. days with 0 occupants,
had to be excluded from the data set to avoid the mathematical impossibility of dividing by zero.




17 The residents m one household (Keycode #22014) moved during the retrofit period and subsequently this home
was excluded from all aspects of the study








In order to test whether the observed variation in means is significant, the per capital daily
use from the three data collection periods were compared through a series of paired t-tests at a 95
percent confidence level. The null hypothesis in each test was that the mean daily per capital
indoor use in the compared logging periods were equal; the alternate hypothesis was that they
were not equal. The t-test results are shown in Table 4.2.
The difference between the baseline and post-retrofit water use was found to be
significant at the 95 percent confidence level. This result indicates the observed reductions in
per capital demand were not due simply to random chance. On the other hand, the difference
between the two post-retrofit logging periods was found to be not significant at this confidence
level. This result indicates that there was no statistical difference between the water use during
the two post-retrofit logging periods, and allowed the data sets to be combined into a single data
set, which was used to compare pre and post-retrofit water use.


Table 4.1 Mean daily indoor per capital use comparison 3 logging periods
n Mean daily indoor per Variance Standard 95%
(logged days) capital use (gcd) Deviation Conf Int
Baseline 481 89.3 6156.0 78.5 6.85
Post-Retrofit 1 455 50.3 843.4 29.0 2.70
Post-Retrofit 2 451 54.1 1346.9 36.7 3.32


Table 4.2 Comparison tests of baseline and post-retrofit indoor per capital water use
Mean Degrees t- P- 95% 95% Statistically
difference in of Value Value Lower Higher significant
indoor daily Freedom Limit Limit difference?
per capital
use (gcd)
Baseline vs. Post-Ret. 1 39.3 934 9.35 <.0001 28.8 44.1 Yes
Baseline vs. Post-Ret. 2 35.2 930 8.07 <.0001 24.7 40.6 Yes
Post-Ret. 1 vs. Post-Ret. 2 3.8 904 -1.73 0.0841 -8.1 0.5 No


Demographic Information

In order to improve the accuracy of the per capital measurements each household was
asked to report the number of people in residence during each logging day. The average number
of people per household during the baseline logging period was 2.56, which is lower than the
2.74 full time residents reported during audits. During the post-retrofit logging periods the








average number of people per household was 2.52 -similar to the number reported during the
baseline logging period.

DAILY HOUSEHOLD USE

When the post-retrofit data were used to compare water use before and after the retrofits
a clear pattern of reduction was noted. Figure 4.1 shows a scatter diagram of total indoor water
use before and after the retrofit for each logging day. The figure shows that there were more
high demand days during the baseline period. The mean daily indoor demand, which was 191.0

gpd per household during the baseline period, dropped 35.5% to 123.3 gpd after the installation
of the new devices. On an annual basis this equates to an indoor use of 69.7 kgal for baseline
conditions and 45.0 kgal with the retrofit.



900
800 -
S700 -
L 600 2
S00 Baseline
0 400-
0 0 Post-Retrofit
z 300
> 200
n 100
0
-100
Observations

Figure 4.1 Scatter diagram of daily indoor per household use, pre and post-retrofit


These data are plotted as a histogram (frequency diagram) in Figure 4.2. Here the change
in demand brought about by the retrofit can be seen in the shift of the demand distribution to the
left. The effectiveness of the retrofits in reducing daily demand can be seen from the fact that
while during the baseline period indoor use fell below 300 gallons 85 percent of the time, in the
post-retrofit period it fell below this level 95 percent of the time.









20%
BASELINE POST
18% Avg daily household indoor use 191 0 123 3
Standard deviation 1152 881
16%

> 14%

z 12%








2%
W 6%


40%

0% 1 TT on J-


Daily Indoor Use (gallons)

M Baseline O Post-Retrofit

Figure 4.2 Daily per household indoor water use distributions, pre and post-retrofit




INDOOR PER CAPITAL USE

Indoor water use patterns changed dramatically after the conservation retrofit. Average

daily per capital use decreased in 31 on the 33 study homes. In the two homes where per capital

use increased, the change was less than 2.2 gcd in each case. A comparison of the baseline and

post-retrofit average per capital daily use is shown in Figure 4.3. In this figure the data are sorted

from highest to lowest baseline per capital use. This result shows that a good portion of the water

savings came from a relatively few participants.

After the retrofit, leakage (17.1 percent), which had previously been the largest

component of indoor use dropped below toilets into fourth place. Toilets (18.6 percent), which

have previously been the second largest component of indoor use moved into third place behind

faucets. Showers became the largest indoor water use followed by faucets and toilets. A pie

chart showing the relative importance of each per capital end use is shown in Figure 4.4. The

combination of showers and baths form the largest block of indoor use in the post-retrofit era at

25.5 percent.











350

S300

250
a
200
4->
10.
S150



0
a-
0 100

50

0


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Study Household


U Baseline 0 Post-Retrofit


Figure 4.3 House by house average per capital daily use comparison, baseline and post-
retrofit


OTHER
0.8%

TOILET (04 d)
TOILET
18 6%
(9A god)



SHOWER
20.3%
(10.7 god)


LEAK
16.9%
(8.9 god)


BATH
5,3%
(2,8 god)


CLOTHES WASHER
16.7% (8.8 gcd)

k DISHWASHER


([.9 ged)


)I


1.7%


FAUCET
19.9%
(In nen


x.-1- 5-n


Post-Retrofit Total: 52.6 gcd


Figure 4.4 Post-retrofit indoor per capital water use percentage


including leakage








Table 4.3 presents a comparison of the mean indoor per capital water use from the
baseline and post-retrofit data collection periods. Overall, indoor water use decreased by 33.6
gcd -a 39.0 percent drop. A series of paired t-tests were performed on each end use in these two
data sets to determine which changes in water use are statistically significant at the 95 percent
confidence level. The results of this analysis are also presented in Table 4.3 as the t-Value and
P-Value from the t-test. In order for a difference in means to qualify as statistically significant,
the P-Value must be less than the alpha level of 0.05, the 95 percent confidence level.
Statistically significant changes in water use were detected for clothes washers, leaks, toilets,
indoor use, miscellaneous other use, and total indoor use.
More than 30 gallons of the 33.9 gcd average saved through the retrofit was the result of
three end uses: toilets, clothes washers, and leaks. Installation of ULF toilets, including some
dual flush models saved an average of 10.1 gcd. The new conserving clothes washers saved an
average of 5.1 gcd. A reduction in leakage resulted in big savings of 16.8 gcd. The leakage
savings were almost certainly the result of the toilet retrofit. Toilet leaks, primarily flapper leaks,
are the single largest contributor to household leakage. In this study, replacing old toilets
through the retrofit eliminated almost all of these toilet leaks and resulted in substantial savings.
None of the other measures implemented through this study (clothes washers, showerheads, or
faucet aerators) should have had any impact on the leakage rate. A more detailed analysis of
leakage is presented later in this report.
It is interesting to note that after the retrofit, statistically significant reductions in water
use occurred in many of the end use categories impacted by the retrofits: toilets, leaks and
clothes washers. Showers and faucets did not show any significant water use reduction, even
though new showerheads and faucet aerators were installed. The remaining categories not
targeted by the retrofit (baths and dishwashers) also showed no statistically significant change.
Specific end uses are examined in more detail in the next section.








Table 4.3 Mean indoor per capital water use, baseline and post-retrofit8
Category Baseline Post- Difference % t-Value P-Value Statistically
(gcd) Retrofit in Means Change significant
(gcd) (gcd) difference?*
Bath 3.0 2.8 -0.2 -6.6% 0.578 0.5674 No
Clothes washer 13.9 8.8 -5.1 -36.7% 4.762 <0.0001 Yes
Dishwasher 1.0 0.9 -0.1 -10.0% 1.860 0.0720 No
Faucet 10.5 10.5 0.0 0.0% 0.030 0.9759 No
Leak 25.7 8.9 -16.8 -65.4% 2.158 0.0385 Yes
Shower 12.0 10.7 -1.3 -10.8% 1.959 0.0589 No
Toilet 19.9 9.8 -10.1 -50.8% 9.129 <0.0001 Yes
Indoor 86.1 52.2 -33.9 -39.4% 3.987 0.0003 Yes
Other/Unknown 0.1 0.4 0.3 75.0% -2.614 0.0004 Yes
Total 86.2 52.6 -33.6 -39.0% 3.942 0.0004 Yes
Avg. # of 2.56 2.52
Residents per
household
*95 percent confidence level



Post-Retrofit Hot Water Usage

Water meters were installed on the hot water heaters of 10 of the 33 study homes and
flow recorders were attached to these meters so that hot water usage could be monitored

alongside overall household usage. The hot water flow traces were disaggregated into end uses
using Trace Wizard and the data were stored in the EBMUD database created for this project.

There were an average of 2.29 residents in the 10 so-called "hot water" homes during the
post-retrofit data collection period, and the average daily indoor per capital use in these homes

was 55.8 gcd compared with 52.6 for the larger 33 home retrofit group.
Overall per capital indoor use in the 10 hot water homes was similar to the study group as

a whole, differing by 3.2 gcd. A t-test test performed on the two data sets found no statistical

difference between the means at the 95 percent confidence level.
Toilet flushing was the only indoor use that had no hot water component. Only 7 percent

of the total leaks were composed of hot water. These results and a comparison with the average


18 The calculation of per capital per day usage was done on a day by day, house by house basis and then the average
of all individual houses was taken to calculate the overall average per capital per day use This creates a weighted
average where the water use in each household is given equal weight The average number of residents per
household was also calculated on a day by day, house by house basis Multiplying these two weighted averages
together to calculate average daily per household use results in a different value than by taking the average of daily
use for each household








gcd found in the entire 33 home study group are presented in Table 4.4. In the post-retrofit
period, 30 percent of all water used indoors, 16.5 gcd, was hot water. On a daily basis, the most

hot water (83.3 percent) was used for faucets, showers, and baths.
A comparison of hot water usage during the baseline and post-retrofit periods is shown in

Table 4.4. Here it can be seen that in the post-retrofit period, the study participants used an
average of 4.6 gcd less hot water than during the baseline period. A series of unpaired t-tests

assuming unequal variances were performed to compare the baseline and post-retrofit hot water
end uses. The results of these tests are presented in Table 4.5. These tests showed that a

statistically significant difference in mean water use before and after the retrofit was detected for
the following categories: clothes washers, faucets, and total indoor use.

The total hot water use dropped by 4.6 gcd after the retrofits, and it appears that nearly all

of these savings can be attributed to the retrofit program. Theoretically the retrofit program
could have impacted clothes washers, faucets, showers, and total indoor hot water use. Although

hot water use declined in almost all end use categories, the change in shower use was found to be
not statistically significant, but the reductions in clothes washer and faucet use were significant.

The retrofit had no impact on leaks of hot water.


Bath
Clothes Washer
Dishwasher
Faucet
Leak
Shower
Toilet
Other
Indoor Total
Sample size
Avg. # of residents


Table 4.4 Post-retrofit per capital hot water use
Post-Retrofit Hot Water Use Total Use in Hot
Water Homes
Average Percent of Indoor Post-Retrofit
gcd Use that is Hot Average
Water gcd
1.5 75.0% 2.0
1.0 10.1% 9.1
1.0 100.0% 1.0
6.2 50.0% 12.4
0.7 7.1% 9.9
6.0 60.0% 10.0
0.0 0.0% 11.2
0.01 50.0% 0.2
16.5 29.6% 55.8
10 10 10
2.3 2.3 2.3


EBMUD Retrofit
Group
Post-Retrofit
Average
gcd
2.8
8.8
0.9
10.5
8.9
10.7
9.8
0.4
52.6
33
2.5








Table 4.5 Comparison of baseline and post-retrofit per capital hot water use
Category Baseline Post-Retrofit Difference % t- P- Statistically
Hot Water Hot Water (gcd) change Value Value significant
Use Use difference?*
(gcd) (gcd)
Bath 1.7 1.5 -0.2 -11.8% 0.209 0.8395 No
Clothes Washer 1.9 1.0 -0.9 -47.4% 3.180 0.0112 Yes
Dishwasher 1.4 1.0 -0.4 -28.6% 1.565 0.1520 No
Faucet 8.6 6.2 -2.4 -27.9% 4.502 0.0015 Yes
Leak 0.7 0.7 0.0 0.0% -0.165 0.8723 No
Shower 6.9 6.0 -0.9 -13.0% 1.183 0.2673 No
Toilet 0.0 0.0 0.0 0.0% na na na
Other/Unknown 0.02 0.01 -0.01 -50.0% 0.519 0.6164 No
Indoor Total 21.1 16.5 -4.6 -21.8% 2.891 0.0179 Yes
Avg. # of 2.3 2.3
Residents per
household
*95 percent confidence level


Analysis of Water Savings Excluding Leaks

Because of the high level of leakage found in the study homes both before and after the
retrofit, it was decided to examine the water savings exclusive of leakage. Leakage accounted
for 30.3 percent of indoor per capital use prior to the retrofit and 17.1 percent after the retrofit. A
specific analysis of leakage is presented later in this report. Results of the analysis excluding
leaks are presented in Table 4.6.
The average baseline per capital per day indoor use -excluding leaks -was 60.3 gcd and
the post-retrofit average was 43.5 gcd. By ignoring leaks both in the baseline and post-retrofit
period, the per capital water savings in indoor use becomes 16.8 gcd -a 27.86 percent reduction
in demand. This reduction was found to be significant at the 95 percent confidence level. This
analysis does not change any of the specific fixture results discussed earlier.
Evaluating the savings achieved through the retrofit without reference to leakage was
recommended by EBMUD utility staff since it is anticipated that leakage of the level discovered
in this study group may not be prevalent throughout all single-family customers in the service
area. It is important to keep in mind that the retrofit study group was selected specifically
because they had a relatively high per capital daily usage compared with other customers, likely








in large part due to leaks. As a result, it may be misleading to suggest that the savings through
reduced leakage could be achieved in a more broad-based group of EBMUD residences.


Table 4.6 Comparison of baseline and post-retrofit per capital daily use excluding leaks
Category Baseline Post- Difference % t-Value P-Value Statistically
(gcd) Retrofit in Means Change significant
(gcd) (gcd) difference?*
Bath 3.0 2.8 -0.2 -6.60% 0.578 0.5674 No
Clothes washer 13.9 8.8 -5.1 -36.70% 4.762 <0.0001 Yes
Dishwasher 1.0 0.9 -0.1 -10.00% 1.86 0.072 No
Faucet 10.5 10.5 0 0.00% 0.03 0.9759 No
Shower 12.0 10.7 -1.3 -10.80% 1.959 0.0589 No
Toilet 19.9 9.8 -10.1 -50.80% 9.129 <0.0001 Yes
Indoor 60.3 43.5 -16.8 -27.86% 7.631 <0.0001 Yes
Other/Unknown 0.1 0.4 0.3 75.00% -2.614 0.0004 Yes
Total 60.4 43.9 -16.5 -27.32% 7.471 <0.0001 Yes
Avg. # of 2.56 2.52
Residents per
household
*95 percent confidence level



FIXTURE USAGE

Toilets

During the post-retrofit data collection period a total of 11,859 individual toilet flushes
were recorded from the 33 study homes over 923 days; for an average of 12.8 flushes per
household per day and 5.74 flushes per capital per day19. After the ULF fixtures were installed,
the average flush volume across the 33 study sites for ULF toilets only was 1.48 gallons per flush

(gpf) with a standard deviation of 0.44 gpf. Since not all toilets in the 33 study homes were
replaced there remained a number of non-ULF toilet flushes in the post-retrofit data set. When
these flushes are included the average flush volume increased to 1.65 gpf with a stand deviation
of 0.87 gpf. For purposes of comparing flush volumes, flushing frequency and per capital use the
non-ULF flushes were included in the analysis. During the baseline period the average flush



19 The value of 5 74 flushes per capital per day was derived by averaging the flushes per person per day determined
for each home on each day of the study This is not the same as dividing the 11,859 total flushes by the 923 days of
observation and 2 5 persons per home (which gives 5 12 flushes per person per day)








volume was 3.88 gpf, so the new fixtures reduced the average flush volume by 2.23 gpf, a 57
percent reduction.
The distributions of the volumes of all recorded toilet flushes in the baseline and post-
retrofit periods are shown in Figure 4.5. These distributions shows the range of flushing
volumes found during the baseline and post-retrofit data collection periods. Five households in
the study group kept one (or more) of their old non-conserving toilets and these higher volume
flushes can be seen in the 3 6 gpf range of the post-retrofit distribution.
During the baseline period, only 7 percent of the recorded flush volumes were in the low
flow range (below 2 gpf), about 60 percent of the flush volumes were above 3.5 gpf, and 15
percent were above 5 gpf. During the post-retrofit period 87 percent of the toilet flushes were in
the low flow range. A t-test (two tail) assuming unequal variances was conducted at the 95
percent confidence level to determine with there was a significant difference between the mean
flush volume during the baseline period and the post-retrofit period. The t-Value in this test was
118.25 and the P-Value was <0.00001, clearly indicating a statistically significant difference
between the two mean flush volumes at the 95 percent confidence level. Thus the conclusion
can be made that the installation of the ULF toilets in the study homes resulted in substantial
reduction in toilet water usage in these study homes.











40% Baseline Post-Retrofit
Average = 3 88 gal 1 64 gal
35% Std Deviation = 1 28 gal 0 81 gal
Median = 3 75 gal 1 53 gal
r 3 Flushes/person/day = 5 14 fpcd 5 58 fpcd
c 30%







10%
"25%
Lij
20% -



10%







Flush Volume (gal.)

E Baseline O Post-Retrofit

Figure 4.5 Toilet flush volume distribution, baseline and post-retrofit



Flushes per capital per day
Double-flushing has been a concern about ULF toilets ever since they were introduced.

Critics have charged that it takes two flushes of a ULF toilet to do the job. The data collected in

this study provided another opportunity to test if flushing frequency increases with the
20
installation of ULF toilets .

During the baseline period, residents in the EBMUD retrofit study group flushed the

toilet an average of 5.14 times per capital per day (fpcd). During the post-retrofit period the

average number of flushes per person per day increased to 5.74 fpcd. The daily per capital

flushing frequency distributions for the baseline and post-retrofit periods is shown in Figure 4.6.






20 See Mayer and DeOreo (1999), pg 131, which discusses the lack of evidence for significant double flushing in
the 1188 homes of the REUWS study












14%

12%

10%


Baseline
515
281
478


Post-Retrofit
558 flushes/cap/day
2 71 flushes/cap/day
5 00 flushes/cap/day


\f 1P lC q


Flushes Per Capita Per Day

E Baseline O Post-Retrofit


Figure 4.6 Toilet flushing frequency distribution, baseline and post-retrofit


Flushing frequency increased by an average of 0.6 fpcd after the ULF toilets were
installed, yet the shape of the flushing frequency distributions are remarkable similar. A t-test
(two tail) assuming unequal variances was conducted at the 95 percent confidence level to
determine of the difference between the baseline and post-retrofit mean flushes per capital per
day was significant. This test returned a t-Value of 3.57 and a P-Value of 0.0004. These results
indicate that there was a statistically significant increase in flushing frequency after the retrofit
amounting to a little more than half a flush per person per day. However, this increase in
average flushing frequency did not severely impact the water the water savings accomplished
through the installation of the ULF toilets. Per capital use for toilet flushing reduced by 10.1 gcd
and this decrease was found to be statistically significant.

Impact of Toilet Make and Model
Four types of ULF toilets were installed as part of this study: the Caroma Caravelle (35),
Niagara Flapperless (32), Sloan Flushmate (2), and Toto G-Max (5). The data collected in this


Average =
Std Deviation =
Median =










- -lHlU








study made it possible to evaluate the impacts of both models of toilet to compare water use,
flush volume, and flushing frequency. These results are presented in Table 4.7.
The Caroma Caravelle is a dual flush gravity toilet designed and manufactured in
Australia. The unique feature of this toilet is that it offers two different flush options -a full 1.6
gallon (6 liter) flush and a half flush of 0.8 gallon (3 liter). There are two buttons on the top of
the toilet tank allowing the user to select the size of flush required. Some dual flush devices
have been available in the US, but none offers the reliability and convenience of the Caroma
design.
The Niagara Flapperless toilet offers a dramatic departure from standard residential
toilets that use a flapper valve to release water from the tank into the bowl. The Niagara
Flapperless toilet has a unique "tipping bucket" design that does not use a flapper. Inside the
tank is a plastic trough that is filled with water. When the toilet is flushed, the trough tips and
dumps the water through a larger opening into the bowl creating a strong flushing action. This
design was created to eliminate flapper leaks.
The Toto Drake is a traditional gravity flush toilet with a 3-inch flush valve. This is
substantially larger than the traditional 2-inch valve. The tank is designed to hold up to 3 gallons
of water, but only 1.5 gallons are used in the flush. The extra water in the tank provides more
pressure and during the flush water is rapidly expelled through the 3-inch flush valve into the
bowl.
The Sloan Flushmate is a pressurized flushing mechanism that can be installed in a wide
variety of existing toilet pottery. For this study St. Thomas Creations pottery was used. The
Flushmate vessel traps air, and as it fills with water, it uses the water supply line pressure to
compress the trapped air inside. The compressed air is what forces the water into the bowl, so
instead of the "pulling" or siphon action of a gravity toilet, the pressure-assist unit "pushes"
waste out. This vigorous flushing action is designed to clean the bowl more effectively than other
toilets.
In order to test the water using characteristics of these two devices, daily per capital water
use for toilets was compared for the homes equipped with each of the four toilet models. A total
of 35 Caroma toilets were installed in 17 homes while 32 Niagara toilets were installed in 13
homes. (To avoid confusion, no homes were equipped with both types.) Two Sloan Flushmate
toilets were installed in one home and five Toto toilets were installed in two homes. Because of








the small sample size, it was not possible to compare water use of the Sloan Flushmate and Toto
toilets with the other brands. Results comparing the Caroma and Niagara toilets are shown in
Table 4.7. The homes equipped with Caroma toilets used 9.9 gcd while the homes equipped
with Niagara toilets used 9.0 gcd, a difference of 0.9 gallons per capital per day. The difference
in means was not found to be significant at the 95 percent confidence level.


Table 4.7 Caroma vs. Niagara toilet water use
Caroma Niagara Difference % t- P- Statistically
Homes Homes Difference Value Value significant
Avg. Avg. difference?
Per capital use 9.9 9.0 0.9 9.1% 0.508 0.6158 No
(gcd)
Flush volume 1.34 1.70 0.44 32.8% -46.06 0.00 Yes
(gallons)
Flushing frequency 6.4 5.0 1.4 21.9% 1.782 0.0869 Yes
(fpcd)
Sample size 17 13
(homes)
*95 percent confidence level


As shown in Table 4.7, the average flush volume of the Caroma toilets was 1.34 gallons
while the Niagara toilets used 1.70 gpf a difference of 0.44 gpf. This difference in means was
found to be statistically significant at the 95 percent confidence level. The lower average flush
volume for the Caroma was due to study participants using the 0.8 gpf flush option available on
those fixtures.
Residents equipped with the Caroma toilets flushed an average of 6.4 times per person
per day while residents equipped with Niagara toilets flushed an average of 5.0 times per person
per day. This difference in flushing frequency was found to be significant at the 95 percent
confidence level, but the result was not a strong one.
In this study, the Niagara Flapperless toilets appear to offer increased savings over the
dual flush Caroma -primarily because on decreased flushing frequency. The average volume
per flush for the Caroma was substantially less than the Niagara, but people flushed the Caroma
more frequently, thus negating the savings. It must be noted that the samples used for this
analysis are quite small and further investigation should be made before making final
conclusions about the water saving potential of these two devices.








ULF Toilet Savings from Other Studies
A number of studies have measured water savings achievable from installing ULF toilets.
These studies include the Seattle Home Water Conservation Study (Mayer, DeOreo, & Lewis,
2000), the REUWS (Mayer and DeOreo, et. al. 1999), the Stevens Institute of Technology
micro-metering studies for East Bay MUD and Tampa, Florida (Aher et. al. 1991; Anderson et.
al. 1993), A&N Technical Service's statistical models developed for MWD (Chesnutt et. al.
1992a, 1992b; 1994), and Aquacraft's small scale retrofit study in Boulder, Colorado (DeOreo
et. al. 1996). The per capital per day toilet savings found in these studies is compared with the
EBMUD Indoor Residential Water Conservation Study results in Table 4.8.
The savings found in the East Bay study showed savings similar to the REUWS and
Seattle study. The highest savings were found in the statistical models developed for Southern
California. The savings from this study were almost twice as much as those found in the 1991
Stevens Institute study also conducted in the EBMUD service area. In the 1991 Stevens study,
the average flush volume was found to be 1.8 gallons per flush (gpf) and in this study the
average flush volume was found to be 1.48 gpf. In addition, the Stevens study found the number
of daily flushes per person to be 3.7 and this study found the number of flushes per person per
day to be 5.7.
It should be noted that the REUWS was not a retrofit study and no conserving hardware
was installed as part of this research. Rather, the ULF savings estimates were calculated as the
difference between the mean per capital toilet usage in homes that exclusively used ULF toilets
and homes in the study that did not use a ULF. However, from the similarity of the results in the
2002 East Bay study, the Seattle study, the REUWS, and the MWD studies a more accurate
picture of the per capital savings achievable from ULF toilet retrofits emerges. These research
efforts each approached the task of calculating savings differently yet their results are quite
similar.








Table 4.8 Comparison of ULF savings from other studies
Research project ULF Flush Per capital Saturation
Volume savings from rate of ULF
(gal/flush) ULF toilets toilets in
(gcd) study homes
East Bay Residential Conservation Study (2002) 1.48 10.1 85%
Seattle Home Water Conservation Study (2000) 1.38 10.9 84%
REUWS (1999) 10.5 100%
MWD (1992 1994) 11.4 73%
Tampa, Florida (1993) 6.1 100%
East Bay MUD (1991) 1.8 5.3 100%
Boulder Heatherwood (1996) 2.6 50%


Showers

During the baseline period, the study participants showered an average of 0.65 times per
day. The average shower consumed 18.4 gallons of water, lasted for 8.9 minutes, and was taken
at a flow rate of 2.0 gpm. Thirty of the 33 study homes received one or more AM Conservation
Group Spoiler (2.5 gpm) showerheads as part of the conservation retrofit. In the post-retrofit
period, study residents showered an average of 0.74 times per day an increase of 0.09 showers
per day. The average post-retrofit shower used 15.34 gallons, lasted for 8.2 minutes, and was
taken at flow rate of 1.8 gpm. In the hot water study homes (n=10) the average shower was 60
percent hot water and 40 percent cold water during the post-retrofit period. The frequency
distributions of baseline and post-retrofit shower volumes are presented in Figure 4.7, Figure 4.8
shows the frequency distributions of baseline and post-retrofit shower durations, and Figure 4.9
presents the frequency distributions of baseline and post-retrofit shower flow rates.
From these results, it appears that the LF showerheads installed during the retrofit did
reduce the flow rate at which people shower and consequently reduced the volume of showers.
However, because of an increase in showering frequency no statistically significant savings were
observed in the comparison of baseline and post-retrofit per capital use shown in Table 4.3.
To determine if and how the installation of LF showerheads impacted demand, unpaired
t-tests assuming unequal variance were performed on the baseline and post-retrofit shower usage
data. The results of these analyses are presented in Table 4.9. The mean shower volume
decreased by 3.06 gallons after the retrofit. The change in means in shower volume was found to
be statistically significant at the 95 percent confidence level.










18%

16%

14%
>1
S12%
c
a)
S10%

S8%

6%


(N (N Co C0 O 0- t I n0 (D (D

Shower Volume (gallons)

E Baseline E Post-Retrofit


Figure 4.7 Shower volume frequency distributions, baseline and post-retrofit


o

Shower Duration (Minutes)

E Baseline O Post-Retrofit


Figure 4.8 Shower duration frequency distributions, baseline and post-retrofit










It has been hypothesized that the introduction of LF showerheads and the subsequent
reduction in shower flow rate could cause people to increase the length of time spent in the
shower. Results from this study show that the introduction of LF showerheads actually
decreased the length of time people spend in the shower. The average shower duration decreased
by 41 seconds after the retrofit. Furthermore this change was found to be statistically significant
at the 95 percent confidence level.
LF showerheads are expected to reduce the water flow rate during the shower. These
new fixtures are designed to restrict the flow of water to 2.5 gpm or less. While this may seem
like a low flow rate, results from this study show that most residents showered at a flow rate
below 2.5 gpm prior to the installation of the LF fixtures. Many of the participants already had
LF showerheads installed in their homes. Others simply chose to throttle their showerheads
down, finding that flow rate more comfortable for showering.
After the retrofit, the average flow rate for showers decreased by 0.19 gpm, from 2.00
gpm to 1.81 gpm. The results of the t-test conducted on these data show that this change in mean
showering flow rate is statistically significant at the 95 percent confidence level. This indicates
that the new showerheads successfully reduced shower flow rates in the study homes.


Table 4.9 Shower usage comparison, baseline and post-retrofit
Comparison Baseline Post- Difference % t-Value P-Value Statistically
Category avg. Retrofit change significant
avg. difference?
Shower Volume 18.40 15.34 -3.06 -16.6% 6.398 0.0001 Yes
(gal.)
Shower Duration 8.88 8.20 -0.68 -7.66% 3.534 0.0004 Yes
(min.)
Shower Flow Rate 2.00 1.81 -0.19 -9.50% 5.269 0.0001 Yes
(gpm)
Showers per capital 0.65 0.74 0.09 13.85% -3.057 0.0023 Yes
per day
Baths per capital 0.12 0.10 0.02 -16.67% -2.369 0.0180 Yes
per day
*95 percent confidence level















20 % I- Iv, L ,- I .V III,
Std. Deviation 1.96 0.53 gpm
Median 1.89 1.85 gpm
C
I 15%-

LI.

10%



5% -



0% -
o -, .. I


Shower Flow Rate (gallons per minute)

E Baseline U Post-Retrofit

Figure 4.9 Shower flow rate frequency distributions, baseline and post-retrofit


Although these results indicate the LF showerheads accomplished the task of reducing

shower flows and volumes, the actual per capital usage for showers decreased by only 1.3 gcd

and this change was found to be not statistically significant. These somewhat contradictory

results are explained by the fact that the frequency of showering by study participants increased

in the post-retrofit period. During the baseline period, study participants averaged 0.65 showers

per person per day, but during the post retrofit period this increased to 0.74 -nearly a 14 percent

increase that was shown to be statistically significant. At the same time, the frequency of baths

decreased during the post-retrofit period from 0.12 baths per person per day to 0.10. As shown

in Table 4.9 both of these changes in use were found to be statistically significant.

One possible explanation for the change in showering habits is the change in seasons

from the baseline to the post-retrofit period. Baseline data were obtained primarily during

February and March 2001. The post-retrofit data were obtained during December 2001 and








during June and July 2002. It is possible that different seasonal conditions may have encouraged
the switch towards increased showering and decreased bathing. The impact of seasonal changes
on showering and bathing is an area for further research.
Regardless of the reason, the increase in showering frequency did substantially reduce the
per capital water savings for showering that would have been observed had the number of
showers remained constant. If the number of showers per person per day is held constant at 0.65,
it is anticipated that the introduction of low flow showerheads would save approximately 2.0
gallons of water per person per day or about 726 gallons per person per year.

LF Shower Savings from Other Studies
A number of studies have measured water savings achievable from installing low-flow
shower heads. These studies include the REUWS, the Stevens Institute of Technology micro-
metering studies for East Bay MUD and Tampa, Florida (Aher et. al. 1991; Anderson et. al.
1993) and the 1984 HUD study (Brown & Caldwell 1994). The per capital per day shower
savings found in these studies is compared with the EBMUD retrofit study results in Table 4.10.


Table 4.10 Comparison of LF showerhead savings from other studies
Research project Per capital savings Saturation rate of
from LF LF showerheads in
showerheads study homes
(gcd)
East Bay Residential Conservation Study (2002) 1.3 94%
Seattle Home Water Conservation Study (2000) 0.3 94%
REUWS (1999) 4.5 100%
HUD (1984) 7.2 NA
Tampa, Florida (1993) 3.6 100%
East Bay MUD (1991) 1.7 100%



The showerhead water savings found in this study are most similar to the savings found
in the 1991 EBMUD study. The Seattle study found the lowest level of showerhead savings. The
savings found in the REUWS were higher than found in all the other studies except for the HUD
study. It should be noted that the REUWS was not a retrofit study and no conserving hardware
was installed as part of this research. Rather, the LF showerhead savings estimates were
calculated as the difference between the mean daily per capital shower usage in homes in which








the residents showered exclusively at or below the 2.5 gpm flow rate and homes in which the
residents showered exclusively above the 2.5 gpm flow rate.
The disparate results for showerheads suggest that the savings achieved by the devices
are uncertain at best. Since showerheads are not expensive it probably makes sense to continue
to distribute them broadly, but water planners should be cautious when projecting savings from
showerhead retrofit programs. Many homes are already equipped with LF showerheads through
natural retrofit and many other people simply throttle their showers down below 2.5 gpm for
comfort.



Clothes Washers

As part of the retrofit program, the clothes washer in each of the 33 study homes was
replaced with a new water conserving model. There are now several different water conserving
clothes washers available in the U.S. In this study, three different models were installed and
evaluated: the Frigidaire Gallery, the Fisher & Paykel Ecosmart, and the Whirlpool Super
Capacity Plus. The Frigidaire Gallery is a front loading horizontal axis machine while the Fisher
& Paykel Ecosmart and the Whirlpool Super Capacity Plus are more traditional top loading
washers. In this study 9 Frigidaire, 13 Fisher & Paykel, and 11 Whirlpool machines were
installed.
To depict how these machines differ in operation several sample flow traces from Trace
Wizard were captured and are presented in Figure 4.10 Figure 4.14. These figures show the
different wash and rinse cycles typical of each machine.
The Frigidaire Gallery used an average of 21.8 gallons per load of clothes with a standard
deviation of 4.5 gallons. As shown in Figure 4.10 and Figure 4.11, the Gallery starts with a
wash/fill cycle that is a higher flow rate than subsequent wash and rinse cycles. Sometimes the
first fill is briefly interrupted and then resumed as shown in Figure 4.10. Usually only the first
fill cycle of the Gallery uses any hot water (because most people set the machines to rinse with
cold water). There are typically five wash and rinse cycles that complete the load. These cycles
may be of different volumes and durations based upon the washer settings. It is also possible to
add additional rinse cycles if desired. A single run of the Gallery takes less than one hour.



























Figure 4.10 Frigidaire Gallery sample flow trace


I.. .1, .- I,


Figure 4.11 Frigidaire Gallery sample flow trace


The Fisher & Paykel Ecosmart used an average of 29 2 gallons per load of laundry with a

standard deviation of 8 6 gallons Examples of the standard water use pattern for thus machine is
shown in Figure 4 12, and Figure 4 13 This clothes washer has an automatic sensor that adjusts

the fill volume depending on the size of the load The water usage pattern for the Ecosmart

vanes depending upon the cycle selected and the load size The wash cycles tend to be at higher

flow rates (4 6 gpm) and may include some hot water Typically there are three smaller cold








water rinse cycles at the end that run at a flow rate of approximately 2 gpm A single run of the
Ecosmart takes less than one hour






















Figure 4.12 Fisher & Paykel Eosmart wash cycle sample flow trace























Figure 4.13 Fisher & Paykel Ecosmart wash cycle sample flow trace
Figure 4.13 Fisher & Paykel Ecosmart wash cycle sample flow trace








The Whirlpool Super Capacity Plus clothes washer used an average of 29 0 gallons per
load of laundry with a standard deviation of 7 4 gpl This machine has a traditional top loading
design, but utilzes only one large fill cycle and a series of six short rinse cycles to wash the
clothes A sample for trace from a Super Capacity Plus is shown in Figure 4 14 There is a 15
minute period between the end of the first fill cycle and the first nnse cycle The first fill cycle
typically used about 20 gallons and the subsequent nnse cycles used 1 5 2 gallons each
Typically only the first fill cycle included hot water, but there were examples when the nnse
cycles also had a hot water component It is possible to add a second larger nnse cycle with this
machine using the extra rinse" setting This can effectively negate the water saving benefits of
the Super Capacity Plus An analysis of the hot water consumption of each machine is presented
later in this section A single run of the Super Capacity Plus takes less than one hour

















Figure 4 14 Whirlpool Super Capacity Plus sample flow trace


Prnor to the retrofit the washing machines in the study homes used an average of 407
gallons per load of clothes with a standard deviation of 9 gallons Data from the baseline
period indicated that 15 3 percent (6 2 gallons per load) of the water used for clothes washing
was hot water Dunng this period, participats ran an average of 0 36 loads of laundry per
person per day From the analysis shown in Table 4 3 it is known that a statistically significant
change in per capital water usage was achieved through the installation of the new conserving
clothes washers Table 4 11 shows the relevant usage statistics for each make and model of
washing machine installed for this study








The clothes washers installed for this study used between 21.8 and 29.2 gallons per load
of wash on average. The Frigidaire Gallery averaged 21.8 gallons per load, the Fisher & Paykel
Ecosmart averaged 29.2 gallons per load, and the Whirlpool Super Capacity Plus averaged 29.0
gallons per load.
Hot water usage for clothes washing decreased substantially in the post-retrofit period.
The Frigidaire used an average of only 3.14 gallons of hot water per load (10.7 percent), the
Fisher & Paykel Ecosmart used 3.12 gallons of hot water per load (14.4 percent), and the
Whirlpool averaged 2.74 gallons of hot water per load (6.0 percent).


Table 4.11 Clothes washer usage comparison
Frigidaire Fisher & Whirlpool Entire Study Entire Study
Gallery Paykel Super Group Group
Ecosmart Capacity (post-retrofit) (baseline)
Plus
Volume Per Average 21.82 29.17 29.04 27.20 40.7
Load (gal.) Std. Dev. 4.47 8.57 7.41 7.97 9.8
Hot Water
Volume Per Average 3.14 3.12 2.74 3.00 6.23
Load (gal.)
Loads Per Average 0.32 0.34 0.30 0.32 0.36
Capita Per Day Std. Dev. 0.48 0.50 0.51 0.50 0.61
Per Capita Average 7.53 9.51 9.00 8.78 14.0
Daly Use _(ga.) Std. Dev. 11.51 3.67 16.49 14.26 7.1
Number of
machines in 11 13 9 33 33
study


The frequency of washing machine use decreased slightly after the retrofit. The study
households ran their new clothes washers an average of 0.32 times per person per day and during
the baseline period they ran their washers 0.36 time per person per day -an 11 percent reduction.
This suggests that the capacity of the new machines was sufficient and did not result in any
increased use of the clothes washer. The brand of clothes washer present in the home had little
effect on the washing frequency. The average loads per capital per day decreased for each of the
three brands tested.
All of the clothes washers installed for this study saved water. The mean per capital daily
water use for clothes washing was 7.53 gcd for the Frigidaire homes, 9.51 gcd for the Fisher &








Paykel homes, and 9.00 gcd for the Whirlpool homes. None of these differences was significant
at the 95 percent confidence level. 14.26 gcd. As shown in Table 4.3, the difference in mean per
capital clothes washer usage from the baseline to the post-retrofit period was statistically
significant at the 95 percent confidence level.

Clothes washer Savings Found in Other Studies
A few other studies have measured water savings achievable from installing conserving
clothes washers. These studies include Aquacraft's 1999 study of water wise homes in
Westminster, Colorado (Mayer et. al. 2000), the Bern Kansas study (Tomlinson and Rizy, 1998),
and the small scale Heatherwood retrofit study. The per capital per day clothes washer savings
found in these studies is compared with the EBMUD results in Table 4.12.


Table 4.12 Comparison of clothes washer savings from other studies
Research project Per capital savings from
conserving clothes washers
(gcd)
East Bay Residential Conservation Study (2002) 5.2
Save Water & Energy Program -SWEEP (2001) 5.3
Seattle Home Water Conservation Study (2000) 5.6
Westminster water wise homes (1999) 4.6
Bern Kansas (1998) 7.2
Boulder Heatherwood (1996) 10.9
*Estimated from % water reduction reported


The measurements of per capital savings vary in these six studies, although in the three
most recent studies the savings are all quite similar. The 1996 Heatherwood study only included
four homes and two of those homes received the water efficient Asko washer. The Bern Kansas
study exclusively used Maytag Neptune washers. Both Heatherwood and Bern were true
intervention studies (like EBMUD and Seattle), that measured demand before and after the
installation of conserving washers. The Westminster study compared clothes washer use in a
sample of standard new homes and sample of "water wise" new homes equipped with high
efficiency fixtures including clothes washers.
The EBMUD and Seattle study both used Frigidaire and Whirlpool clothes washers. The
Maytag Neptune was tested in Seattle and the Fisher & Paykel Ecosmart was tested in the East
Bay. The SWEEP study used Frigidaire clothes washers exclusively. The similarity in water








savings found in the Seattle, SWEEP, and EBMUD studies suggests an approaching agreement
on the impact of these specific machines on per capital water use.



Faucets

A total of 79 bathroom faucet aerators and 20 kitchen faucet aerators were installed as
part of the retrofit program in the EBMUD service area. These aerators were manufactured by
New Resources Group and were designed to limit flows to less than 2.2 gpm in the kitchens and
less than 1.5 gpm in the bathrooms. These aerators have a pressure compensating feature so that
flow rates can be maintained under a variety of different water pressures.
Mean per capital faucet was unchanged at 10.5 gcd after the retrofit. As shown in Table
4.3, this was the only end use where a retrofit was accomplished but no change was exhibited.
During the baseline period, faucet usage accounted for 12.2 percent of all indoor use. After the
retrofit it accounted for 20.0 percent of total indoor use because of the large decrease in leaks,
toilet, and clothes washer usage.
In the 10 hot water study homes, mean per capital hot water faucet use went from 13.2
gcd in the baseline period to 12.4 gcd in the post retrofit period, but this change was found not to
be statistically significant. During the baseline period 65.2 percent of all faucet usage was hot
water and after the retrofit the hot water component decreased to 50.0 percent. A summary of
results for faucet usage is presented in Table 4.13.


Table 4.13 Faucet use comparisons, baseline and post-retrofit
Per Capita Daily Use Typical Flow Rates
Volume Duration Average Peak
(gal) (minutes) (gpm) (gpm)
Baseline 10.5 10.4 1.2 2.9
Post-Retrofit 10.5 13.3 0.93 2.6
t-Value 0.030 -6.654 14.123 4.176
P-Value 0.9759 <0.0001 <0.0001 <0.0001
Statistically significant difference? No Yes Yes Yes
*95 percent confidence level
**Not all faucet aerators could be replaced


The faucet results shown in
retrofit faucet flow rate decreased


Table 4.13 bear some important conclusions. First, after the
-both the average and peak. This indicates the aerators did








have an impact. However, the duration of faucet use increased after the retrofit -this increase
was found to be statistically significant. Many faucet uses such as filling a glass or a sink should

be independent of faucet flow rate, meaning that the volume of water used for these purposes is
fixed and that volume will be used regardless of the delivery flow rate of the water. Reducing

the flow rate should result in water users spending more time to fill specific volumes. The net
result in this study was that the average per capital faucet use did not change from the baseline to

the post-retrofit period.
Figure 4.15 shows the baseline and post-retrofit frequency distributions of per capital

faucet use durations. The increase in duration of faucet use can be seen in the shift of the post-
retrofit distribution to the right of the x-axis.



25% -

PRE POST
Avg. Per Cap Daily Faucet Use 10.4 13.3 min.
Standard deviation 6.6 9.6 min.
> Median 8.8 11.0 min.

3 15%






5% -


0%


Per Capita Daily Faucet Usage (minutes)

N Baseline O Post-Retrofit

Figure 4.15 Faucet usage comparison, baseline and post-retrofit








Baths


Because baths require a fixed amount of water, this study did not include any bath
retrofits and therefore a reduction in bath water usage was not expected. During the baseline
period, the average bath used 28.5 gallons of water and during the post-retrofit period the
average bath used 27.5 gallons. This difference was found not to be statistically significant. The
maximum baseline bath usage was 89.9 gallons and the maximum post-retrofit bath usage was
94.3 gallons.
During the baseline period, baths taken in the hot water homes were 89.5 percent hot
water and 11.5 percent cold water. During the post-retrofit period, baths were measured as using
75.0 percent hot water and 25.0 percent cold water.
Study residents took an average of 0.12 baths per person per day or 0.84 baths per person
per week during the baseline and 0.10 baths per person per day or 0.70 baths per person per week
during the post-retrofit period. Comparisons of baseline and post retrofit bath data are shown in
Table 4.14.


Table 4.14 Bath usage comparisons, baseline and post-retrofit
Avg. Bath Max. Bath Hot Water Avg. Baths per
Volume Volume Component Capita Per Day
(gal.) (gal.) %
Baseline 28.5 89.9 89.5% 0.12
Post-Retrofit 27.3 94.3 75.0% 0.10
Statistically
Significant No na Yes Yes
Difference?*
*95 percent confidence level


Leaks

The reduction and elimination of leaks appears to be one important result of the retrofit
program. Mean daily per capital leakage was reduced by 65.4% percent from 25.7 gcd to 8.9 gcd
after the retrofit as shown in Table 4.3. Most of the leaks eliminated during the retrofit were the
result of replacing toilets that had leaky flapper valves. No other leak repair was performed as
part of the retrofit program.
During the baseline period it was discovered that 10 of the households were responsible
for roughly 86 percent of the total leakage in the entire study group. As was the case during the








baseline period and in the results of the REUWS, a few houses accounted for most of the leakage
during the post-retrofit period. Two houses accounted for more than 50 percent of the total
leakage during the two post-retrofit data collection periods and 10 houses accounted for more
than 80 percent of the total. Leaks in the top two leaking homes were re-examined using Trace
Wizard and in both homes the leaks appeared as continuous water uses with very low flow rates
(<0.5 gpm). While it is impossible to determine the exact cause of these leaks, they looked to be
caused by leaking faucets or hose bibs or perhaps by a leak in piping. These leaks did not appear
to be toilet leaks, which are almost always associated with toilet flushes, nor did they appear to
be irrigation leaks, typically caused by broken heads or stuck valves in automatic irrigation
systems.
One possible explanation for the high leak rate that was found in some of the study
participants' homes could be traced to the District's change in its water treatment process.
EBMUD converted from treating water with chlorine to chloramines (chlorine and ammonia) in
1998. An August 1993 AWWA Journal article reported study results showing that chloramines
have a more deleterious effect on elastomers (products widely used in plumbing distribution,
especially for toilet flapper valves) than does free chlorine. When a utility converts from chlorine
to chloramine, this negative effect on the elastomers tends to increase incidents of leaks in the
home and in the distribution system. The plumbing industry has responded to this problem by
marketing elastomer products with compounds resistant to attack by chloramines.











PRE POST
50% Avg. 45.9 19.6 gpd
Std. Dev. 79.3 51.0 gpd
> Median 14.8 4.4 gpd
S40%- Min. 0.0 0.0 gpd
= Max. 662.9 454.8 gpd
o-
L 30% -

4*-'
20%


10% -


0%


Daily per household leakage (gal.)

E Baseline O Post-Retrofit

Figure 4.16 Daily per household leakage distributions, baseline and post-retrofit


Figure 4.16 shows the baseline and post-retrofit distributions of daily household leakage.

The general shape of these distributions is the same, but in the post-retrofit period there were

many more low leakage days. The difference in mean per household leakage (26.3 gpd) was

found to be significant at the 99 percent confidence level. It is interesting to note that the median

leakage rate in both distributions is substantially lower than the mean, belying the positive skew

of the distributions.



MAXIMUM DAY INDOOR DEMANDS



Peak day demands measured for each study house during the post-retrofit period were

typically higher than the peak demands measured during the baseline period. This is not

surprising since there were twice as many post-retrofit data days and some of these data were

collected during the spring and summer when many of the study participants were irrigating








several times per week. Since the retrofit targeted indoor use exclusively it was decided to

compare indoor peak day demands before and after the retrofit.

Figure 4.17 shows the maximum daily indoor demands for each study home during the

baseline and post-retrofit period. Twice as much daily use data were available from the post-

retrofit period, but peak indoor use during the baseline period exceeded peak use during the post-

retrofit period 26 out of 33 study homes. In seven homes, the post-retrofit indoor peak exceeded

the baseline peak. Overall, indoor peak demand decreased by 16 percent after the retrofit. This

reduction was found to be statistically significant at the 95 percent confidence level.


900

800

v 700

w 600
U)
D
o 500
0
o
- 400

0 300

. 200


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Study Household

M Baseline O Post-Retrofit

Figure 4.17 Peak daily indoor use comparison, baseline and post-retrofit


For most systems, peak demand during the summer due to outdoor use is the driving

factor in sizing treatment plants and distribution lines. Reductions in indoor peak demands are

not likely to profoundly impact these design criteria. Nevertheless, a 32 percent reduction in








indoor peak demand will be reflected in a lower demand on the peak day, and will create a
benefit from the perspective of peak reduction as well as volumetric demand reduction.



BILLING DATA ANALYSIS

Billing data from the post-retrofit period were obtained from EBMUD staff as one of the
final data items for the study. It was hoped that billing data could be used to confirm the savings
detected through the flow trace analysis techniques. Ideally such an analysis should be
performed with at least one full year to post-retrofit billing data to use to compare against the
pre-retrofit billing data described earlier in this report. Because of the project time schedule this
was not possible and instead only about eight months of post-retrofit billing data, from January
August 2002, could be obtained. Billing data were obtained for all 999 homes in the original
Q1000 sample frame, which includes all 33 homes in the retrofit study and a 966 home control
group.
The retrofits in the 33 study homes were begun in June 2001 and completed in stages by
December 2001. EBMUD reads customers' water meters on a bi-monthly schedule and different
areas of the service area are read at different times of the month. EBMUD keeps good records of
the date each meter is read so it was possible to screen the billing data from each household to
ensure that billing periods that included both pre and post-retrofit data were excluded.
There are problems inherent in using billing data to evaluate the effectiveness of
conservation measures such as those tested in this study. These problems have been well
documented and include: unequal billing periods, estimated meter readings, unusual usage levels,
meter read errors, rounding of meter reads, changes in customer occupancy, etc. (Dziegielewski
1993a). However, billing data remains a reliable, cheap, and easy way to measure customers'
water use and it should be examined in spite of the inherent shortcomings.
In this study, the limited amount of post-retrofit billing data available made a comparison
of pre and post-retrofit water use more difficult. Much of the available post-retrofit data
included a substantial component of irrigation demand which can easily mask the sought after
effect. To minimize the potential impact of unequal billing periods, the average daily
consumption for each billing period was calculated for the pre and post-retrofit periods, and the




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs