HomeMy WebLinkAboutxeri_study_final
Xeriscape Conversion Study
Final Report
2005
By
Kent A. Sovocool
Senior Conservation Programs Analyst
Southern Nevada Water Authority
With data processing assistance from
Mitchell Morgan
Assistant Management Analyst
Southern Nevada Water Authority
Funded in part by a grant from
Bureau of Reclamation
U.S. Department of Interior
SNWA Member Agencies
• Big Bend Water
• City of Boulder City
• City of Las Vegas
• City of Henderson
• City of North Las
Vegas
• Clark County Water
Reclamation District
• Las Vegas Valley
Water District
i
Table of Contents
TABLE OF TABLES ................................................................................................................................ii
TABLE OF FIGURES..............................................................................................................................iii
ABSTRACT................................................................................................................................................ 4
ACKNOWLEDGEMENTS........................................................................................................................ 6
INTRODUCTION AND BACKGROUND ............................................................................................... 7
XERISCAPE AND WHAT IT MAY MEAN FOR WATER CONSERVATION..........................................................7
NEVADA’S COLORADO RIVER RESOURCES AND THE SPECIAL IMPORTANCE OF OUTDOOR WATER ............8
CONSERVATION ...................................................................................................................................8
THE RESEARCH STUDY ........................................................................................................................ 10
METHODOLOGY ................................................................................................................................... 11
STUDY GROUPS AND MONITORING....................................................................................................... 11
GENERAL DATA METHODS, STRATEGIES, AND STATISTICS..................................................................... 13
PRE/POST ANALYSES ........................................................................................................................... 13
ANALYSES OF SAVINGS OVER TIME AND SEASONS ................................................................................. 14
COMPARATIVE PER-UNIT AREA IRRIGATION ANALYSES ......................................................................... 18
MULTIVARIATE ANALYSES TO IDENTIFY SIGNIFICANT SOURCES OF VARIABILITY..................................... 20
ECONOMIC ANALYSES.......................................................................................................................... 21
RESULTS AND DISCUSSION............................................................................................................... 22
REDUCTION IN TOTAL HOUSEHOLD WATER CONSUMPTION FOLLOWING CONVERSION .......................... 22
TO XERISCAPE ..................................................................................................................................... 22
ASSESSMENT OF SAVINGS POTENTIAL ACROSS TIME AND SEASONS ........................................................ 24
COMPARISON OF PER UNIT AREA WATER APPLICATION BETWEEN TURFGRASS AND ............................... 30
XERIC LANDSCAPE .............................................................................................................................. 30
Annual application........................................................................................................................ 30
Monthly Application..................................................................................................................... 32
SOURCES OF SIGNIFICANT VARIABILITY IN SINGLE-FAMILY RESIDENTIAL ............................................... 40
CONSUMPTION .................................................................................................................................... 40
Variability in Annual Residential Consumption........................................................................... 41
Variability in Annual Consumption for Irrigation of Monitored Xeric Landscape...................... 46
FINANCIAL SAVINGS ASSOCIATED WITH CONVERSION PROJECTS AND COST EFFICIENCY........................ 50
ESTIMATED APPROPRIATE LEVEL OF FINANCIAL INCENTIVE ................................................................. 54
Homeowner Perspective............................................................................................................... 54
SNWA Perspective....................................................................................................................... 58
EXECUTIVE SUMMARY AND CONCLUSIONS................................................................................ 60
REFERENCES ......................................................................................................................................... 64
APPENDICES.......................................................................................................................................... 66
Appendix 1: Attachment A (Scope of Work) for BOR ...................................................................... 67
Agreement 5-FC-30-00440 as Revised 11/19/98 .......................................................................... 67
Appendix 2: Multivariate Model Details ......................................................................................... 73
ii
Appendix 3: Raw Data ..................................................................................................................... 74
Appendix 4: Information on Single-Family Residential Water Bill Model...................................... 93
Appendix 5: Information on Homeowner Perspective Model.......................................................... 94
Appendix 6: Information on SNWA’S Water Smart Landscapes Program....................................... 96
Table of Tables
TABLE 1: Planned Pre-/Post-Retrofit Analyses for XS Group............................................................... 14
TABLE 2: Planned Pre-/Post-Retrofit Analyses for TS Group............................................................... 14
TABLE 3: Planned Post-Retrofit Analyses for XS Group Across Time................................................. 15
TABLE 4: Planned Post-Retrofit Analyses for TS Group Across Time................................................. 15
TABLE 5: Enhanced Post-Retrofit Analyses for XS Group Across Time.............................................. 16
TABLE 6: Enhanced Post-Retrofit Analyses for TS Group Across Time .............................................. 16
TABLE 7: Planned Summer Post-Retrofit Analyses for XS Group........................................................ 17
TABLE 8: Planned Summer Post-Retrofit Analyses for TS Group........................................................ 18
TABLE 9: Planned Comparative Analysis of Turf and Xeric Per Unit .................................................. 19
TABLE 10: Planned Comparative Analysis of Turf and Xeric Per Unit ................................................ 19
TABLE 11: Pre-/Post-Retrofit Analyses for XS Group .......................................................................... 22
TABLE 12: Pre-/Post-Retrofit Analyses for TS Group........................................................................... 23
TABLE 13: Enhanced Post-Retrofit Analyses for XS Group Across Time............................................ 26
TABLE 14: Summer Post-Retrofit Analyses for XS Group.................................................................... 27
TABLE 15: Enhanced Post-Retrofit Analyses for TS Group Across Time ............................................ 28
TABLE 16: Summer Post-Retrofit Analyses for TS Group.................................................................... 29
TABLE 17: Annual Per-Unit Area Application to Turf and Xeriscape.................................................. 30
TABLE 18: Monthly Per-Unit Area Application to Turf and Xeriscape ................................................ 33
TABLE 19: Multivariate Regression Model of Annual Single-Family Residential................................ 73
TABLE 20: Multivariate Regression Model of Annual Xeric Study Area Consumption....................... 73
iii
Table of Figures
FIGURE 1: Pre-/Post-Retrofit Consumption for XS and Comparison Groups....................................... 22
FIGURE 2: Pre-/Post-Retrofit Consumption for XS Group Across Time .............................................. 25
FIGURE 3: Annual Per Unit Area Application to Turf and Xeriscape................................................... 31
FIGURE 4: Distribution of Annual Per Unit Area Application Data for Turf and Xeriscape ................ 32
FIGURE 5: Monthly Per-Unit Area Application for Turf and Xeric Areas............................................ 34
FIGURE 6: Monthly Per-Unit Area Savings (Turf Area Application– Xeric Area Application)........... 35
FIGURE 7: Monthly Per-Unit Area Application to Turf and Reference................................................. 36
FIGURE 8: Monthly Per-Unit Area Application to Xeric Areas ............................................................ 36
FIGURE 9: Absolute Departure in Irrigation Application from Derived Respective............................. 37
FIGURE 10: Relative Departure in Irrigation Application from Derived Respective ............................ 38
FIGURE 11: Relative Departure in Irrigation Application from Derived Respective ............................ 39
FIGURE 12: Average Monthly Maintenance Time and Annual Direct Expenditures............................ 51
FIGURE 13: Modeled Monthly Water Bill for a Typical Las Vegas Area Home and ........................... 52
FIGURE 14: Modeled Monthly Water Bill Savings for A Typical Las Vegas Area.............................. 53
FIGURE 15: Summary of Economics of Typical Single-Family Xeriscape........................................... 57
4
Abstract
The authors present a manuscript covering the Southern Nevada Water Authority’s
(SNWA) multi-year Xeriscape Conversion Study, which was funded in part by the
Bureau of Reclamation - Lower Colorado Regional Area1.
Xeriscape (low-water-use landscaping) has held the promise of significant water savings
for a number of years, but how much exactly it can save, especially as a practical
residential landscape concept has been a point of debate and conjecture. Lacking to date
has been a truly experimental quantitative study in which per-unit area application data
has been gathered to quantify savings estimates (for a variety of reasons, most research
has been limited to the total household level, with comparisons involving homes that are
mostly xeriscape or traditional landscaping). Recognizing the need for more exacting
(and locally applicable) savings estimates, SNWA conducted a study that could yield
quantitative savings estimates of what a xeriscape conversion facilitation program could
yield under real world conditions.
The experimental study involved recruiting hundreds of participants into treatment
groups (a Xeric Study and a Turf Study Group and control groups), as well as the
installation of submeters to collect per unit area application data. Data on both household
consumption and consumption through the submeters was collected, as well as a wealth
of other data. In most cases, people in the xeric study group converted from turf to
xeriscape, though in some cases recruitment for this group was enhanced by permitting
new landscapes with xeric areas suitable for study to be monitored. Portions of xeric
areas were then submetered to determine per-unit area water application for xeric
landscapes. The TS Group was composed of more traditional turfgrass-dominated
landscapes, and submeters were installed to determine per-unit area application to these
areas as well. Submeter installation, data collection, and analysis for a small side-study
of multi-family/commercial properties also took place.
Results show a significant average savings of 30% (96,000 gallons) in total annual
residential consumption for those who converted from turf to xeriscape. The per-unit
area savings as revealed by the submeter data was found to be 55.8 gallons per square
foot (89.6 inches precipitation equivalents) each year. Results showed that savings
yielded by xeriscapes were most pronounced in summer. A host of other analyses
covering everything from the stability of the savings to important factors influencing
consumption, to cost effectiveness of a xeriscape conversion program are contained
within the report.
An abbreviated summary of the report’s findings appears as the Executive Summary
and Conclusions section (pg. 60).
1This report with written and electronic appendices satisfies a deliverables requirement pursuant
the applicable funding agreement with the Bureau of Reclamation (Cooperative Agreement #5-
FC-30-00440). SNWA gratefully acknowledges BOR for its funding assistance with this project.
5
Keywords: water conservation, xeriscape, xeric, landscape conversion, desert landscape
low-water-use, plants, landscape, irrigation, residential water consumption, outdoor
water use, submeter, irrigation controller, resource conservation, incentive programs,
utility, water resources.
6
Acknowledgements
There have over the course of this study been so many contributors to this research that
thanking each of them individually is an impossibility; however. the authors would like to
express their gratitude to individuals in the following groups and organizations without
whom this research would not have been possible.
• The study participants without whom no data could have been collected.
• The Southern Nevada Water Authority (SNWA) and the numerous personnel who
have supported and promoted this research. The authors especially thank
members of the Conservation Division, present and past, which have supported
this research over the years.
• The member agencies of the SNWA listed on the cover. Special thanks go to
Boulder City, the cities of Henderson and North Las Vegas, and the Las Vegas
Valley Water District for the reading of meters and submeters related to the study.
• The SNWA Xeriscape Conversion Study Team and all those who have
contributed to study recruitment, data collection, and data processing.
• Aquacraft, Inc. for preliminary analyses and collection of trace flow data.
• The Bureau of Reclamation Lower Colorado River Office and members of this
office who helped facilitate funding for the research.
• Those whom have reviewed and commented on the Xeriscape Conversion Study
and the results. The authors graciously acknowledge the contributions of
reviewers of the final manuscript. Final manuscript reviewers included SNWA
personnel in the Conservation and Resources Divisions, members of the Las
Vegas Valley Water District’s Resources Division, a contractor to the SNWA,
and outside reviewers at the City of Austin, Texas. The Bureau of Reclamation
also had the opportunity to comment on a draft prelude to final submission.
• Those organizations that have acted as a forum for exhibiting this research.
• Those who have worked for the study in the past and have gone on to other
things.
• Those who have supported the research and the research personnel in other
capacities.
7
Introduction and Background
XERISCAPE AND WHAT IT MAY MEAN FOR WATER CONSERVATION
In the Mojave Desert of the southwestern United States, typically 60 to 90% of potable
water drawn by single-family residences in municipalities is used for outdoor irrigation.
Thus, in this region, and indeed most of the entire Southwest, the most effective
conservation measures are oriented towards reducing outdoor water consumption. A
commonly considered method for accomplishing water conservation is to use xeriscape
(low-water-use landscaping) in place of traditional turf. Xeriscape is based on seven
principles:
• Sound Landscape Planning and Design
• Limitation of Turf to Appropriate Areas
• Use of Water-efficient Plants
• Efficient Irrigation
• Soil Amendments
• Use of Mulches
• Appropriate Landscape Maintenance
The term “xeriscape” was invented by Nancy Leavitt, of Denver Water (a public utility)
in the early 1980s as a fusion of the Greek word “xeros” (meaning dry or arid) and
landscape. Denver Water trademarked the term shortly thereafter though it has entered
the English vernacular over the last 20 years as the concept has spread globally.
So promising was xeriscape, that water purveyors and others interested in conservation
began actively promoting xeriscape in place of traditional landscape as early as the
mid-80s as part of water conservation strategies. The need to better understand its true
effectiveness as a conservation tool led to a host of studies being conducted in the 1990s,
which have generally pegged savings associated with xeriscape at between 25% and 42%
for the residential sector (Bent1 1992, Testa and Newton2 1993, Nelson3 1994, Gregg4
et al. 1994). The variation in savings estimates is due to a large number of factors
ranging from the different climates of each study locality, different local definitions of
xeriscape, and different study methodologies employed.
The work done to this point has greatly advanced the water conservation community’s
ability to evaluate, modify, and justify programs to encourage the use of xeriscaping as an
integral component of water conservation plans. Utilities, water districts, cities, counties,
and states are beginning to promote xeriscape as a cost-effective, mutually beneficial
alternative to traditional turfgrass-dominated landscapes. Recently, this interest has
increased at the national level, and this study is part of that evolution. Interest is further
enhanced at the time of publication of this report due to a significant drought impacting
the Colorado River Basin and much of the western United States.
8
NEVADA’S COLORADO RIVER RESOURCES AND THE SPECIAL IMPORTANCE OF
OUTDOOR WATER CONSERVATION
The Colorado River serves as the lifeblood for many of the communities of the
southwestern United States, permitting society to flourish, despite the harsh, arid
conditions that often define it. It serves the needs of millions within the region and its
yearly volume is entirely divided up by the Colorado River Compact5 and subsequent
legislation and legal decisions, known as the “Law of the River” that specify allocations
for each of the states (and Mexico) through which it flows. Among other things, the
Bureau of Reclamation – Lower Colorado Region (BOR-LCR) is charged with
maintaining an adequate and established allocation of water for each of the states in the
arid Lower Basin. Since water demand management is ultimately accomplished at local
levels, BOR-LCR actively partners with entities that divert Colorado River water to
encourage conservation. In southern Nevada, the major regional organization meeting
this criterion is the Southern Nevada Water Authority (SNWA).
In 1991 the SNWA was established to address water on a cooperative local basis, rather
than an individual water purveyor basis. The SNWA is committed to managing the
region’s water resources and developing solutions that ensure adequate future water
supplies for southern Nevada. The member agencies are the cities of Boulder City,
Henderson, Las Vegas, North Las Vegas, the Big Bend Water District, the Clark County
Water Reclamation District, and the Las Vegas Valley Water District. As southern
Nevada has grown into a metropolitan area and a world-famous vacation destination, so
too have its water needs. The SNWA was created to plan and provide for the present and
future water needs of the area.
Five different water purveyors provide potable water to most of Clark County. Big Bend
Water District provides water to the community of Laughlin; the cities of Boulder City
and Henderson provide water to their respective communities. The Las Vegas Valley
Water District provides water to the City of Las Vegas and portions of unincorporated
Clark County; the City of North Las Vegas provides water within its boundaries and to
adjacent portions of unincorporated Clark County and the City of Las Vegas. The
SNWA member agencies serve approximately 96% of the County’s population.
Southern Nevada’s climate is harsh. The Las Vegas Valley receives only 4.5 inches of
precipitation annually on average, has a yearly evapotranspirational (ET) water
requirement of nearly 90 inches, and it is one of the fastest growing metropolitan areas in
the United States. Clark County, the southernmost county in Nevada, has a population in
excess of 1.6 million people and has been experiencing extremely strong economic
growth in recent years with correspondent annual population growth averaging in excess
of 5% percent. The primary economic driver of Clark County’s economy is the tourism
and gaming industry, with an annual visitor volume in excess of 30 million people per
year. Today more than 7 out of every 10 Nevadans call Clark County home.
Consumptive use (use where Colorado River water does not return to the Colorado River)
is of paramount interest to SNWA (specifically, consumptive use is defined by SNWA as
9
the summation of yearly diversions minus the sum of return flows to the River). A 1964
Supreme Court Decree in Arizona v. California verified the Lower Basin apportionment
of 7.5 million acre feet (MAF) among Arizona, California, and Nevada, including
Nevada’s consumptive use apportionment of 300,000 acre feet per year (AFY) of
Colorado River water as specified initially in the Colorado River Compact5 and
Boulder Canyon Project Act6. Return flows in Nevada consist mainly of highly treated
Colorado River wastewater that is returned to Lake Mead and to the Colorado River at
Laughlin, Nevada. With return flow credits, Nevada can actually divert more than
300,000 AFY, as long as the consumptive use is no more than 300,000 AFY (see diagram
below). Since Colorado River water makes up roughly 90% of SNWA’s current
water-delivering resource portfolio, it means that in terms of demand management,
reduction of water used outdoors (i.e., water unavailable for accounting as return flow) is
much more important in terms of extending water resources than reduction of indoor
consumption at this point in time.
Diagram Showing Dynamic of Diversions, Return Flow Credits (from indoor uses)
and Consumptive Use
Since most of the SNWA (Authority) service area contains relatively scarce local
reserves (there are little surface or groundwater resources) and since, as explained above,
its Colorado River apportionment is limited, the organization has an aggressive
conservation program that began in the 1990s. The Authority has been committed to
achieving a 25% level of conservation (versus what consumption would have been
without conservation) by the year 2010 (note though that soon this goal will be revised to
probably be even more aggressive in the immediate future due to the drought). In 1995,
the SNWA member agencies entered into a Memorandum of Understanding (MOU)
regarding a regional water conservation plan. The MOU, updated in 1999, identifies
specific management practices, timeline, and criteria the member agencies agree to
follow in order to implement water conservation and efficiency measures.
10
The programs or Best Management Practices (BMPs) listed in the MOU include water
measurement and accounting systems; incentive pricing and billing; water conservation
coordinators; information and education programs; distribution system audit programs;
customer audit and incentive programs; commercial and industrial audit and incentive
programs; landscape audit programs; landscape ordinances; landscape retrofit incentive
programs; waste-water management and recycling programs; fixture replacement
programs; plumbing regulations, and water shortage contingency plans. The BMPs
provide the framework for implementing the water conservation plan and guidance as to
the methods to be employed to achieve the desired savings.
THE RESEARCH STUDY
The potentially large water savings attainable with the broad-scale use of xeriscaping and
the fact that associated reductions are in consumptive-use water makes xeriscape of
paramount interest for both BOR and SNWA. For this reason, a partnership between
BOR and SNWA was formed to investigate the savings that could be obtained with a
program to encourage converting traditional turfgrass landscape to xeriscape. This was
formally implemented as a Cooperative Agreement7 in 1995. With its signing, a
multi-year study of xeriscape was born, which has come to be known as the SNWA
Xeriscape Conversion Study (XCS). As delineated in the most recent version of the
Scope (Appendix 1) for this agreement, the objectives of the Study are to:
• Objective 1: Identify candidates for participation in the Study and monitor their
water use.
• Objective 2: Measure the average reduction in water use among Study
participants.
• Objective 3: Measure the variability of water savings over time and across
seasons.
• Objective 4: Assess the variability of water use among participants and to identify
what factors contribute to that variability.
• Objective 5: Measure the capital costs and maintenance costs of landscaping
among participants.
• Objective 6: Estimate incentive levels necessary to induce a desired change in
landscaping.
SNWA assembled a team to support the XCS, and field data was collected through 2001
with the draft final report finished in 2004 (intermediate reports outlined some of the
major conclusions). By agreement, the SNWA agreed to provide the raw data collected
for possible use in national research efforts by BOR (data was included with the final
version of this manuscript submitted to BOR).
11
Methodology
STUDY GROUPS AND MONITORING
The study team recruited participants who live in single-family residences within the
following entities’ water jurisdictions: The Las Vegas Valley Water District (77% of the
participants in the entire study group), Henderson (12%), North Las Vegas (9%), and
Boulder City (2%).
There are a total of three groups in the XCS, the Xeriscape Study (XS) Group, the Turf
Study (TS) Group, and a non-contacted Comparison Group. The XS Group is composed
of residents who converted at least 500 square feet (sqft) of traditional turfgrass to xeric
landscape as well as residents who installed new xeric landscaping. To clarify, in this
region, xeric landscaping is principally composed of a combination of desert-adapted
shrubs, trees, some ornamental grasses, and mulch (often rock). A $0.45 per square foot
incentive helped the property owner by absorbing some, but not the majority, of the cost
of the conversion. Homeowners were required to plant sufficient vegetation so that the
xeric landscape would at a minimum have 50% canopy coverage at maturity. This
avoided the creation of unattractive “zeroscapes” composed exclusively of rocks, which
could potentially act as urban heat islands. The incentive was capped for each residence
at $900 for 2,000 sqft; however, many residents converted more landscape than that
which qualified for the incentive with the cap. Indeed, the average area converted in this
study group was 2,162 sqft. A total of 472 properties were enrolled in the Study as
XS Group participants. Aerial photographs, supported by ground measures, were used
for recording areas. As a supplement to the main experimental group, 26 multi-family
and commercial properties were submetered as well.
In return for the incentive, XS Group residents agreed to ongoing monitoring of their
water consumption. This was accomplished in two ways. First, mainmeter data was
taken from standard monthly meter reading activity (this was for assessing water use at
the entire single-family residence level). Second, residents agreed to installation of a
submeter that monitored irrigation consumption on a portion of the xeric landscape.
Submeters were typically read monthly, as with mainmeters and were used to study
per-unit area application of water comparatively. The area monitored by the submeter
was called the Xeric Study Area. Study areas were tied to irrigation zones and stations.
Virtually all study properties have in-ground irrigation systems and controllers to avoid
the presence or absence of these as a major confounding factor. This experimental
control is important because it has been noted that the presence of automated irrigation is
highly associated with increased water usage for residential properties (Mayer and
DeOreo8 et al. 1999) apparently because such systems make irrigation more likely to
occur regularly versus hand-watering. Having participants in both groups possess
automated systems also avoids the potential bias of more heavily turf-covered properties
being more likely to be fully automated, thus having higher consumption as was the case
for Bent1 1992 (as identified in Gregg4 et al. 1994). All areas of each property were
broken down into landscape categories. For example, a XS Group property might have
monitored (via the submeter) xeric landscape and unmonitored xeric, turf, garden, and
12
other (non-landscaped) areas. Square footages were recorded for each of these respective
area types.
In addition to water-consumption monitoring, residents agreed to a yearly site visit for
data-collection purposes. During site visits, information was collected on the xeric
species present, plant canopy coverage at the site, components of the irrigation system,
and per-station flow rates.
Staff trained in the identification of locally used landscape plants collected data on plant
size and species present.
Plant canopy coverage was calculated by first taking the observed plant diameters,
dividing this number by two to get radius, then applying the formula for getting the area
of a circle (A=πr2). This area result was then multiplied by the quantity of those species
of plants observed to be at that size. The summation of all areas of all plants of all size
classes in the study area is the total canopy coverage for that area.
Data on the components of irrigation systems was collected by staff trained in the
different types of irrigation emitters available (ex. drip, microsprays, bubblers, etc.).
Staff then ran individual stations and watched meter movement to get the per-station flow
rates.
The Turf Study (TS) Group is composed of properties of more traditional landscape
design, where an average 2,462 sqft of the landscaped area was of traditional turfgrass
(most commonly Fescue). Mainmeter data was collected in the same manner as for the
XS Group. Due to design challenges, the submeter was more commonly hooked to
monitor a mixed type of landscape rather than just turf, though many did exclusively
monitor turf (only “exclusively turf” monitoring configurations were used in per-unit area
landscape analyses). TS participants enrolled voluntarily, without an incentive and
agreed to yearly site visits as above. Other data on irrigation systems was collected in a
manner similar to that for the XS Group properties. A total of 253 residences were
recruited into the TS Group.
The enrollment of participant residences into the XS and TS Groups was directly
dependent on homeowners’ willingness to participate in this study. For this reason,
sampling bias was of reasonable concern to SNWA. To address this, a third subset of
non-contacted Comparison Groups was created to evaluate potential biases. Comparison
properties were properties with similar landscape footprints and of similar composition to
the TS group and pre-conversion XS Group and were in the same neighborhoods as these
treatment properties. This control group was also subject to the same water rates,
weather, and conservation messaging as the treatment groups. Having this group also
permitted SNWA to evaluate the combined effects of submetering and site visits on the
treatment groups.
13
GENERAL DATA METHODS, STRATEGIES, AND STATISTICS
Several different data analysis methods were applied in the course of the study. Details
of each can be found in the corresponding subsections below. Broadly, analysis methods
fell into the categories of pre- vs. post-treatment evaluations, comparative analyses of
different treatment groups, analyses to determine variables associated with consumption,
and assorted cost-benefit analyses. Statistical methods employed include descriptive
statistics (ex. means, medians, etc.), tests for differences in means assuming both
normally distributed data (t-tests) and non-normally distributed (i.e., non-parametric) data
(Mann-Whitney U-tests), as well as techniques employing established economic
principles and multivariate regression (some details of regression models are included in
Appendix 2). In means comparisons, statistical significance was determined to occur
when the probability of a Type I error was less than 5% (α=0.05). Presentation of data
involving calculations of differences in values (for example, means differences) may not
appear to add up in all cases, due to rounding. Types of data analyzed include mainmeter
consumption data, submeter consumption data combined with area data (i.e., application
per unit area data), flow-rate data, cost data, survey responses, and assorted demographic
and Clark County Assessor’s Office data. Consumption data was gathered by the
aforementioned purveyor entities and assembled by SNWA. Most other data was
collected by SNWA (Aquacraft Inc. also performed some analyses on consumption and
data logger collected data under contract to SNWA). In many analyses, data was
scatterplotted and objective or subjective outlier removal done as deemed appropriate.
Finally, in some cases, data analysis was expanded upon to include attempts at modeling.
These endeavors are elaborated on in other parts of the manuscript.
PRE/POST ANALYSES
For each property and year where complete monthly consumption records were available,
these were summed to provide yearly consumption. Data for each XS Group property
was assembled from the five years before conversion (or as many records as were
available; only properties having converted from turf to xeriscape were in this analysis
sample) and as many years post-conversion as records permitted up through 2001. These
data sets permitted comparison of total yearly consumption before and after the landscape
conversion. The impact of submetering and site visits could also be evaluated by
comparing mainmeter records for the TS Group pre- and post-installation of landscape
submeters. Differences could be further confirmed by comparing the change in total
household consumption following the conversion or submetering event for the XS and
TS groups respectively against the change in consumption for non-contacted, non-
retrofitted properties of similar landscape composition. The general analysis strategy for
Objective 2 of the approved Scope (Appendix 1) is summarized in the following tables
(Tables 1 and 2):
14
TABLE 1: Planned Pre-/Post-Retrofit Analyses for XS Group
Pre-retrofit
(kgal/yr)
Post-retrofit
(kgal/yr)
Difference in
Means (kgal/yr)
Xeriscape
Treatment
Comparison
Difference in
Means (kgal/yr)
TABLE 2: Planned Pre-/Post-Retrofit Analyses for TS Group
Pre-retrofit
(kgal/yr)
Post-retrofit
(kgal/yr)
Difference in
Means (kgal/yr)
Submetered
Conventionally
Landscaped
Treatment
Comparison
Difference in
Means (kgal/yr)
ANALYSES OF SAVINGS OVER TIME AND SEASONS
Objective 3 directs SNWA to measure the variability of water savings over time and
across seasons. In the approved Scope, this was anticipated to involve comparing the XS,
TS, and Comparison Groups to derive savings estimates in the manner specified in the
tables that follow (Tables 3 and 4):
15
TABLE 3: Planned Post-Retrofit Analyses for XS Group Across Time
First Year’s
Consumption (Y1)
Third Year’s
Consumption (Y3)
Difference in
Means (kgal/yr)
Xeriscape
Treatment
Comparison
Difference in
Means (kgal/yr)
TABLE 4: Planned Post-Retrofit Analyses for TS Group Across Time
First Year’s
Consumption (Y1)
Third Year’s
Consumption (Y3)
Difference in
Means (kgal/yr)
Submetered
Conventionally
Landscaped
Treatment
Comparison
Difference in
Means (kgal/yr)
Since in most cases, meters were read monthly or at least bimonthly, SNWA is able to
provide an analysis exceeding the level of detail originally specified in the Scope.
Specifically, the longevity of savings from conversions for each year following the
conversion could be evaluated, thus the following new table specifies the more in-depth
level for the “over time” analyses called for in Objective 3:
16TABLE 5: Enhanced Post-Retrofit Analyses for XS Group Across Time Mean Post-retrofit Consumption First Year Post-retrofit (Y1) Second Year Post-retrofit (Y2) Third Year Post-retrofit (Y3) Fourth Year Post-retrofit (Y4) Fifth Year Post-retrofit (Y5) Xeriscape Treatment (kgal/year) Comparison Group (kgal/year) Difference in Means (kgal/year) TABLE 6: Enhanced Post-Retrofit Analyses for TS Group Across Time Mean Post-retrofit Consumption First Year Post-retrofit (Y1) Second Year Post-retrofit (Y2) Third Year Post-retrofit (Y3) Fourth Year Post-retrofit (Y4) Fifth Year Post-retrofit (Y5) Submetered Conventionally Landscaped Treatment (kgal/year) Comparison Group (kgal/year) Difference in Means (kgal/year)
17
Recruitment of properties for the XCS spanned a couple of years. For this reason, in order to
evaluate true changes over time, the first year after each conversion was designated as Y1, the
second as Y2, and so forth. As such, consumption data for a property starting in, for example,
1995, was designated as belonging to Y1, but for a different property starting in 1996, 1996 was
Y1. In this way, the impact of different start years was corrected for and multiyear analyses could
be considered on a more common basis. This permits inferences to be made about how landscape
water consumption and savings change over time as plants in the xeric areas mature. It is also the
reason the sample size appears to diminish for the XS Groups from Y1 to Y5. It is not that there
was heavy loss of sample sites, rather that fewer sites were in existence for a total of five years
owing to early enrollment. A similar effect is seen in the TS Group. There is no data for Y5 for
the TS Group because enrollment for that Group started later than for the XS Group.
Savings from xeriscape may be greatest in summer when evapotranspirational demand is greatest
for all plants, but so to an extreme degree in southern Nevada for turfgrasses (Source: University
of Nevada Cooperative Extension). In considering how savings may be different across seasons,
the Scope (Appendix 1) directs the SNWA to certain prescribed analyses (Tables 7 and 8):
TABLE 7: Planned Summer Post-Retrofit Analyses for XS Group
Pre-Retrofit
Summer
Consumption
(kgal/month)
Post-Retrofit
Summer
Consumption
(kgal/month)
Difference in
Means
(kgal/month)
Xeriscape
Treatment
Comparison
Group
Difference in
Means
(kgal/month)
18
TABLE 8: Planned Summer Post-Retrofit Analyses for TS Group
Pre-Retrofit
Summer
Consumption
(kgal/month)
Post-Retrofit
Summer
Consumption
(kgal/month)
Difference in
Means
(kgal/month)
Submetered
Conventionally
Landscaped
Treatment
Comparison
Group
Difference in
Means
(kgal/month)
Because of the resolution available by submetering, even more detailed data pertaining to
application of water to turf and xeriscape through seasons is available in the comparative per-unit
area irrigation analyses (see following section and Comparison of Per-Unit Area Water
Application between Turfgrass and Xeric Landscape in Results and Discussion).
COMPARATIVE PER-UNIT AREA IRRIGATION ANALYSES
Submeter consumption data combined with measurement of the irrigated area permitted
calculation of irrigation application on a per-unit area basis (gallons per square foot, which can
also be expressed as precipitation inches equivalents) for most study participants. In this way,
exacting measures of consumption for irrigation of xeric and turf landscape types could be
measured. The sample size (Ns) is the product of the number of months or years of data and the
number of valid submeter records analyzed. Sample sizes for specific analyses appear in Results
and Discussion. Only records for submeters that monitored turf exclusively were included in
per-unit area analyses involving the TS Group so that other landscape types would not confound
calculation of results.
No prescribed analyses of submeter consumption data appear in the Scope. The two basic sets of
analyses selected by SNWA were (i.) a comparative analysis of annual application to xeric and
turf areas and (ii.) a comparative analysis of monthly application to xeric and turf areas. The
analytical setup of these appears in Tables 9 and 10 respectively. Secondary analyses comparing
usage to theoretical reference ET demand projections follow the basic comparisons and appear in
Results and Discussion.
19TABLE 9: Planned Comparative Analysis of Turf and Xeric Per Unit Area Annual Application Per Unit Area Application (gallons/square foot/year) Submetered Turf (TS Group) Submetered Xeriscape (XS Group) Difference (gallons/square foot/year) TABLE 10: Planned Comparative Analysis of Turf and Xeric Per Unit Area Application for Each Month Jan Gal/SqFt Feb Gal/SqFt Mar Gal/SqFtApr Gal/SqFtMay Gal/SqFtJun Gal/SqFtJul Gal/SqFtAug Gal/SqFtSep Gal/SqFtOct Gal/SqFtNov Gal/SqFtDec Gal/SqFt Submetered Turf (TS Group) Submetered Xeriscape (XS Group) Difference (gallons/square foot/month)
20
MULTIVARIATE ANALYSES TO IDENTIFY SIGNIFICANT SOURCES OF VARIABILITY
Objective 4 of the Scope (Appendix 1) directs SNWA to assess variability of water use amongst
the study participants and identify what factors contribute to that variability. Potential sources of
variability originally specified for investigation in the Scope included the following:
• Number of members in the household
• Age of occupants
• Number of bathrooms
• Income
• Home value
• Percentage of xeriscaping
• Xeriscape density
• Turf type
• Type of irrigation
• Lot size
• Landscapeable area
• Existence of a pool
• Flow rates
• Water use factors
As the XCS developed, additional potential factors were assessed. A complete listing of data
recorded is included in Appendix 3 (not all data was collected for all properties in the study).
Preliminary investigations focused on some of the above variables from a principally univariate
analysis perspective (DeOreo9 et al. 2000, Sovocool10 et al. 2000, Sovocool and Rosales11 2001,
Sovocool12 2002). The advantage of this was that it permitted rapid quantification and association
of target variables’ influences on participant water use, especially at the per-unit area scale.
However, the most sophisticated way to deal with a study of this type where there are a number of
potential independent associations of several predictor variables to a dependent variable is by the
application of multivariate regression analysis methods. This permits so-called “partial
regression” of independent variables to the target dependent one, here water consumption.
Multiple regression for estimation can be expressed in the general multiple regression equation as
follows:
Ŷi = â + b1*X1i + b2*X2i + ... + bni*Xni + ∈
Where Ŷ is the estimated dependent variable, â is the y-axis intercept, b is each estimated beta
partial regression coefficient representing the independent contribution of each independent
variables’ influence on Ŷ, X is each independent variable up to the nth variable, i is the time
period and ∈ is the error term for the model.
Multicollinearity between X variables violates the underlying assumptions of regression models
and can be dealt with by setting limiting tolerance thresholds of similarity in contribution of
variability to a regression model. This, in turn, permits identification and possible exclusion of
such highly collinear and possibly inappropriate independent variables. The most significant
variables can then be quantified and their relative vector and magnitude of association on the
21
dependent variable can be deduced, ultimately yielding an explanatory multivariate model of how
such variables may contribute to water consumption. Such variables are explored for association
to total household consumption and xeric landscape submeter consumption in the results section
in two distinct modeling exercises.
ECONOMIC ANALYSES
Objective 5 of the Scope mandates quantification and measurement of capital costs and
maintenance costs of the conversion. In the summer of 2000, data on landscape maintenance
economics was obtained via surveys sent to study participants. The survey helped quantify both
labor hours and direct costs associated with landscape choices. For details on the survey and
methodology, consult Hessling13 (2001). Three hundred surveys were returned for analysis.
Results of these were tabulated and compiled, and analyses proceeded from there.
By the very nature of the study methodology, it was recognized at the outset that a simple
comparison of the XS and TS groups would likely fail to demonstrate the economic
considerations with respect to maintenance of the whole landscape level as most residents’
landscapes were composed of multiple landscape types (at the least, both xeric and turfgrass
areas). This led to an analytical method of comparing the costs of landscape maintenance based
on the relative percentages residents had of turf and xeric areas respectively.
The water bill savings associated with conversion projects were calculated based on the
Las Vegas Valley Water District’s water rates as they currently stand (in early 2004). Savings
were calculated by modeling bills for a typical fifth decile (midrange in consumption) home
where the average yearly consumption is 208,057 gallons and for such a home doing an average
(according to data collected for the Water Smart Programs single-family sector in early 2004)
1,615.8-sqft-conversion from turfgrass to xeric landscape (note the difference in this average size
conversion relative to that of the XS Study Group; conversion sizes, along with lot sizes, have
diminished over time in this area). Bills were modeled on a monthly basis and all charges were
applied that actually appear for customers. An example output of this model appears in
Appendix 4.
As directed in the Scope (Appendix 1), the financial viability of xeriscape conversions was
explored. This necessitated looking at the economics of conversions from the homeowner and
SNWA perspectives. Hessling13 (2001) attempted some of these initially. A follow-up analysis
from these same perspectives was performed in the writing of this report and is included in
Results and Discussion. The homeowner perspective included an estimative Net-Present-Value
(NPV)-based modeling approach to determine when return on investment (ROI) was achieved
and details on this model appear in Appendix 5. This same model is used to determine the
incentive level necessary to induce change (Objective 6) by making some assumptions about what
timeframe is acceptable for owners to achieve ROI. The approach used for the SNWA
perspective is to consider alternative sources of water and use the cost associated with these to
determine the maximum amount SNWA should pay to help convert grass to xeric landscape.
22
Results and Discussion
REDUCTION IN TOTAL HOUSEHOLD WATER CONSUMPTION FOLLOWING CONVERSION
TO XERISCAPE
Results for the XS Group pre/post-conversion comparisons are shown in Table 11 and Figure 1.
TABLE 11: Pre-/Post-Retrofit Analyses for XS Group
Pre-retrofit
(kgal/yr)
Post-retrofit
(kgal/yr)
Difference in
Means (kgal/yr)
t-tests
(* denotes
significance)
Xeriscape
Treatment
n=321
Mean=319
Median=271
Mean=223
Median=174
96*
(30% reduction
from pre-retrofit)
t=16.8*
p<0.01
Comparison
n=288
Mean=395
Median=315
Mean=382
Median=301
13
(3% reduction from
pre-submetering)
t=1.85
p=0.07
Difference in
Means (kgal/yr)
76* 159*
t-tests (* denotes
significance)
t=4.32*
p<0.01
t=9.69*
p<0.01
FIGURE 1: Pre-/Post-Retrofit Consumption for XS and Comparison Groups
Mean monthly consumption for the residences dropped an average of 30% following conversion.
A dependent t-test demonstrates that the reduction in usage is highly significant (t=16.8; p<0.01).
23
Though individual performance may vary greatly, the overwhelming majority of homes in the
study saved water following the conversion (285 out of 321 analyzed). This finding of about a
third reduction in consumption is nearly identical to findings from a study of residences in Mesa,
Arizona (Testa and Newton2 1993). It may be that a reduction of about this percentage may be
anticipated to occur when the average single-family residence built in the late 20th century does an
average-size conversion in the southwestern United States. The large savings are likely in part
because the great majority of water consumption goes to outdoor irrigation in this region. In this
study, the average savings realized was 96,000 gallons per year per residence.
The difference in consumption of the pre-retrofit homes to the non-contacted comparison homes
is shown in Table 11 and Figure 1. As demonstrated, a t-test of consumption between these two
groups shows there was significant difference in initial consumption between the two groups
(t=4.32; p<0.01), suggesting self-selection bias. This is not surprising since recruitment of study
participants was voluntary. People who were already conserving more were apparently more
likely to enroll and agree to convert a portion of their respective properties. This does not
however invalidate the results, as (i.) this incentive-based approach is essentially the same as the
approach used for enrolling people in the actual program SNWA has (see Appendix 5) and, more
importantly (ii.), there is no compelling evidence that the Comparison Group experienced
significant reduction over the same time period so the savings are likely attributable exclusively
to the landscape conversion.
The analysis procedures in the Scope (Appendix 1) suggest that the impact of submetering on
outdoor irrigation may be revealed by comparing consumption at the conventionally landscaped
properties with submeters (the TS Group) to that for the associated comparisons for that Group.
The data appearing in Table 12 fulfill this prescribed Scope treatment.
TABLE 12: Pre-/Post-Retrofit Analyses for TS Group
Pre-submetering
(kgal/year)
Post-submetering
(kgal/year)
Difference in
Means (kgal/yr)
t-tests
(* denotes
significance)
Submetered
Conventionally
Landscaped
Treatment
n=205
Mean=352
Median=303
Mean=319
Median=268
34*
(10% reduction
from pre-retrofit)
t=5.08*
p<0.01
Comparison
n=179
Mean=364
Median=314
Mean=347
Median=296
17*
(5% reduction over
timeframe)
t=2.08*
p<0.05
DIFFERENCE IN
MEANS
(KGAL/YR)
12 28
T-TESTS (*
DENOTES
SIGNIFICANCE)
t=0.52
p=0.60
t=1.41
p=0.16
24
There are two potential issues though with trying to consider this analysis an evaluation of the
effectiveness of submetering. First, submetering is typically studied where the scenario is one
where water consumption through the submeter is relayed to end-use customers and where the
customers are billed for it. Without consumption data and billing, the residents in this study have
received no price signal to encourage them to read the meter or reduce consumption. This theory
corresponds with what staff members have observed in the field with respect to the behavior of
customers. Most participants apparently did not even think about the meter until it was time for
their yearly site review and often they stated they had forgotten it was even there. So here, the
dynamic of submetering is rather unique and the impact most likely minimal.
The second consideration, at least as potentially significant, is the fact that participants had been
exposed to annual site visits, which is likely a more important variable in terms of modifying
behavior (no conservation training or formal education took place at site visits, though staff
members did answer questions posed to them). Indeed, the Comparison Group provides for a
good gauge of the impacts on treatment groups due to site visits. Initially, results seem to suggest
a reduction of possibly up to 34,000 gallons annually associated with visits and submetering
(t=5.08; p<0.01) though, as revealed in the next analyses, this impact appears to be only
temporary (seen only in the first year, Table 15) and is probably in actuality much more negligible
given half the “reduction” also appears to have taken place in the control group (t=2.08, p<0.05).
The control group reduction may be due to background conservation at the community level.
With respect to understanding how submetering with consumption billing may be of conservation
benefit, a national research effort (Mayer et al. 200414), supported in part by SNWA, has just
been completed which provides much more insight into the benefits of submeters for water
conservation purposes (also see Rosales15 et al. 2002).
ASSESSMENT OF SAVINGS POTENTIAL ACROSS TIME AND SEASONS
For the XS Group, significant reduction in total yearly consumption took place immediately
following conversion and remained relatively stable at that decreased level through subsequent
years, showing no erosion with time (Table 13 and Figure 2). In every year, the XS Group
consistently had lower consumption than the Comparison Groups, and this was statistically
significant (Table 13). This suggests that conversions are a viable way to gain substantial water
savings over at least a medium-term timeframe and quite possibly over a long one as well. It also
resolves questions about whether or not xericape takes more water in the first year following
conversion (apparently the answer is no) and it suggests that, at least over the medium-term, there
is no erosion of savings obtained from conversions due to residents’ response to growth of plants
in their xeric areas.
For the XS Group, the relative reduction in consumption became even more pronounced in the
summer (Table 14) where, savings averaged 13,000 gallons per summer month (Table 14:
t=18.5;p<0.01) versus an average of 8,000 per month over the entire year. It should be noted that
a very small, but statistically significant reduction of 1,600 gallons per month appears to have also
taken place in the Comparison Group during the summer (in a pre- vs. post-comparison of the
study timeframe, Table 14: t=1.98; p<0.05). Overall, the results are consistent with the theory
that xeric landscapes save the most during the summer. The comparative per-unit analyses that
follow reveal why this is the case.
25
In considering savings stability over extended time, it was found that the submetered TS group
only demonstrated significantly decreased consumption for the first year following retrofit, after
which savings were not significant (Table 15; statistics in table). This initial reduction might be
due to residents’ interest in the research and in conservation when new to the study, this wearing
off with time. Again, it is important to recall that in no single year was the consumption
statistically different from the comparison group properties. The submetered TS Group did have
significantly lower consumption in the summer, with a savings of 3,300 gallons per month
(Table 16: t=3.78;p<0.01) whereas the comparison group to the TS Group showed no such
reduction (Table 16: t=1.03;p=0.31). However, there was no difference in average monthly
summer consumption between the submetered properties and the controls after the retrofit
(Table 16: t=1.03;p=0.31). Overall the results in Table 15 seem to reflect the finding that little
enduring change in consumption was achieved by the TS Group over time despite submeter
installation.
FIGURE 2: Pre-/Post-Retrofit Consumption for XS Group Across Time
0
50
100
150
200
250
300
350
Pre-Conversion Y1 Y2 Y3 Y4 Y5
Yearkgal/year
26TABLE 13: Enhanced Post-Retrofit Analyses for XS Group Across Time Post-retrofit Consumption First Year Post-retrofit (Y1) Second Year Post-retrofit (Y2) Third Year Post-retrofit (Y3) Fourth Year Post-retrofit (Y4) Fifth Year Post-retrofit (Y5) Xeriscape Treatment (kgal/year) 214∆ (32% reduction from pre-retrofit) n=320 220∆ (30% reduction from pre-retrofit) n=318 227∆ (28% reduction from pre-retrofit) n=306 211∆ (33% reduction from pre-retrofit) n=211 202∆ (36% reduction from pre-retrofit) n=61 Comparison Group (kgal/year) 372 n=280 387 n=275 383 n=260 362 n=183 345 n=54 Difference in Means (kgal/year) 158 167 156 151 143 t-tests (* denotes significance) t=9.98* p<0.01 t=9.29* p<0.01 t=9.08* p<0.01 t=8.02* p<0.01 t=4.85* p<0.01 Treatment group values with a ∆ are significantly lower than pre-retrofit value.
27TABLE 14: Summer Post-Retrofit Analyses for XS Group Pre-Retrofit Summer Consumption (kgal/month) Post-Retrofit Summer Consumption (kgal/month) Difference in Means (kgal/month) t-tests (* denotes significance) Xeriscape Treatment n=321 Mean=38 Median=31 Mean=25 Median=19 13* t=18.5* p<0.01 Comparison Group n=288 Mean=47 Median=38 Mean=46 Median=35 1.6* t=1.98* p<0.05 Difference in Means (kgal/month) 9* 21* t-tests (* denotes significance) t=4.23* p<0.01 t=10.1* p<0.01
28 TABLE 15: Enhanced Post-Retrofit Analyses for TS Group Across Time Post-submetering Consumption First Year Post-submetering (Y1) Second Year Post-submetering (Y2) Third Year Post-submetering (Y3) Fourth Year Post-submetering (Y4) Fifth Year Post-submetering (Y5) Submetered Conventionally Landscaped Treatment (kgal/year) 291∆ (6% decrease from pre-submetering) n=228 312 (1% increase from pre-submetering) n=229 317 (2% increase from pre-submetering) n=228 315 (2% increase from pre-submetering) n=146 No Data Available Comparison Group (kgal/year) 332 n=170 357 n=173 351 n=167 351 n=108 No Data Available Difference in Means 41 45 34 36 t-tests (* denotes significance) t=2.28 p=0.02 t=2.39 p=0.02 t=1.65 p=0.10 t=1.40 p=0.16 Treatment group values with a ∆ are significantly lower than pre-submetering value.
29 TABLE 16: Summer Post-Retrofit Analyses for TS Group Pre-Submetering Summer Consumption (kgal/month) Post-Submetering Summer Consumption (kgal/month) Difference in Means (kgal/month)t-tests (* denotes significance) Submetered Conventionally Landscaped Treatment n= 205 Mean=41.7 Median=34.0 Mean=38.5 Median=31.0 3.3* t=3.78* p<0.01 Comparison Group n=179 Mean=42.0 Median=36.0 Mean=41.0 Median=34.7 1.0 t=1.02 p=0.31 Difference in Means (kgal/month) 0.3 2.5 t-tests (* denotes significance) t=0.97 p=0.92 t=1.03 p=0.31
30
COMPARISON OF PER-UNIT AREA WATER APPLICATION BETWEEN TURFGRASS AND
XERIC LANDSCAPE
Annual application
Annual per unit area irrigation application data summaries are found in Table 17 and Figures 3
and 4. There was a great difference in the annual water application to turf and xeric landscape
areas (Table 17 and Figure 3). Turf received an average of 73.0 gallons per square foot annually
(117.2 inches), while xeriscape received on average, just 17.2 gallons (27.6 inches) each year
(only 23.6% of the amount of water applied for turfgrass maintenance). The difference was thus
55.8 gallons per square foot per year (89.6 inches), and this was found to be highly significant
assuming a normal distribution of data (t=27.0; p<0.01).
TABLE 17: Annual Per-Unit Area Application to Turf and Xeriscape
Per Unit Area
Application
(gallons/square
foot/year)
Per Unit Area
Application
(inches/year)
Sample Distribution Statistics
Submetered
Turf
(TS Group)
ns=107
Mean=73.0
Median=64.3
Mean=117.2
Median=103.2
Standard Deviation=40.0
Skewness=1.17
Kurtosis=1.36
Submetered
Xeriscape
(XS Group)
ns=1550
Mean=17.2
Median=11.5
Mean=27.6
Median=18.5
Standard Deviation=18.6
Skewness=3.14
Kurtosis=14.9
Difference
(gallons/square
foot/year)
Mean=55.8 Mean=89.6
t-tests (* denotes
significance)
t=27.0*
p<0.01
Levene’s Test
(* denotes
significance)
F(1, 1655)=130.3*
p<0.01
Mann-Whitney U
Test (* denotes
significance)
U=10177
z=15.2*
p<0.01
Detailed statistics were not generated for the small set of multifamily and commercial sites;
however, the average consumption on those xeric areas where viable data could be collected was
16.7 gallons per square foot per year (ns=22). This suggests the use of xeric landscape in these
sectors may result in similar savings as that observed above on a comparative landscape basis
31
(i.e., savings of ca. 55.8 gallons per square foot annually versus what application would have been
for turf).
FIGURE 3: Annual Per Unit Area Application to Turf and Xeriscape
Distinct differences in the sample distributions for the XS and TS irrigation data were of concern
from a statistical analysis perspective. Both distributions had features strongly suggesting data
was not distributed homogenously across the two groups (Table 17 and Figure 4). In particular,
the XS Group data was heavily skewed with the vast majority of participants using very little
water. Turf application, while indeed skewed, appears almost normal compared to xeric
application, which is very heavily skewed (skewness = 3.14) and peaks sharply (kurtosis=14.9) at
the lower end of the distribution. This is because the vast majority of XS participants used a very
small amount of water to irrigate their xeric areas, while a handful used greatly more volume on
theirs. Because t-tests assume normality, the atypical and non-congruent distributions were of
sufficient concern to justify running a Levene’s Test simultaneous with the t-tests to assess the
potential need to apply non-parametric analytical techniques (though in practice the need for
normality is lessened with large sample sizes due to the tendency of such a collection of data to
mimic a normal distribution; aka. the central limit theorem). Indeed, the Levene’s Tests
demonstrated significant differences in the distributions [Levene F(1,1655) = 130.3; p<0.01].
This suggested the need to backup the findings with non-parametric approaches. Mann-Whitney
U (a summation and ranking based approach to the problem) was chosen as a good backup test.
Associated z statistics for this test with corresponding probabilities are thus reported with the
results in Table 17 as supporting evidence for statistical difference in irrigation application
between the groups.
32
FIGURE 4: Distribution of Annual Per Unit Area Application Data for Turf and Xeriscape
0.00
5.00
10.00
15.00
20.00
25.00
0102030405060708090100110120130140150160170180190200Gallons/Square Foot/Year% of Sample in Each CategoryPercent of XS Group Percent of TS Group
Higher Water
Application
Lower Water
Application
Monthly Application
Monthly submeter data summaries for the XS Group and exclusively monitored turf TS Group
participants appear in Table 18. It should be noted that at times the interval between reads
stretched over more than one month and thus the dataset for the monthly data is slightly different
than that for the above annual comparison as only consumption data deemed complete and
assignable to a given month could be included (sometimes consumption across a two-month gap
was averaged to fill the gap). There were issues with resolution in monitoring because typically
at least a thousand gallons had to pass through the meter between reads in order for the
consumption figure to be advanced and registered by the reader, and sometimes this did not
happen for XS Group submeters monitoring relatively small areas due to low consumption. Both
these factors likely result in slight inflation of monthly consumption values for both groups and
this indeed appears to be manifest if monthly averages are summed across the year (i.e., this per
unit area consumption figure is slightly higher than the annual one calculated in the previous
section). Still, on a monthly basis the data is generally valid and valuable in comparative analyses
and in comparing water application to irrigation requirements. Per-unit area application data is
displayed graphically in Figure 5.
33 TABLE 18: Monthly Per-Unit Area Application to Turf and Xeriscape Jan Gal/SqFt Feb Gal/SqFt Mar Gal/SqFtApr Gal/SqFtMay Gal/SqFtJun Gal/SqFtJul Gal/SqFtAug Gal/SqFtSep Gal/SqFtOct Gal/SqFtNov Gal/SqFtDec Gal/SqFt Submetered Turf (TS Group) 2.97 2.11 ns=85 2.96 2.06 ns=85 3.44 3.29 ns=85 6.07 4.85 ns=88 9.37 7.86 ns=93 10.79 9.38 ns=93 11.86 10.50 ns=95 10.23 8.71 ns=96 8.47 7.15 ns=99 6.20 5.29 ns=105 4.37 3.50 ns=107 2.47 1.96 ns=106 Submetered Xeriscape (XS Group) 1.16 0.46 ns=1291 0.87 0.43 ns=1337 0.99 0.57 ns=13771.43 0.83 ns=14091.64 1.08 ns=14122.01 1.30 ns=14212.24 1.40 ns=14312.27 1.39 ns=14562.22 1.27 ns=14961.66 1.02 ns=15191.35 0.77 ns=15340.91 0.48 ns=1534 Difference (Gallons/Sqft) 1.81 2.09 2.45 4.64 7.74 8.78 9.62 7.96 6.25 4.54 3.02 1.56 t-tests (* denotes significance) t=73.36* p<0.01 t=7.52* p<0.01 t=13.33*p<0.01 t=9.92* p<0.01 t=29.87*p<0.01 t=27.7* p<0.01 t=26.22*p<0.01 t=21.96*p<0.01 t=13.15*p<0.01 t=17.59*p<0.01 t=13.45*p<0.01 t=9.39* p<0.01 Mann-Whitney U Tests (* denotes significance) U=23499 z=8.84* p<0.01 U=18127 z=10.54* p<0.01 U=15959z=11.27*p<0.01 U=14225z=12.14*p<0.01 U=6824z=14.49*p<0.01 U=4415z=15.10*p<0.01 U=6062z=14.89*p<0.01 U=9776z=14.13*p<0.01 U=12307z=13.91*p<0.01 U=14501z=14.04*p<0.01 U=25290z=11.98*p<0.01 U=31202 z=10.62* p<0.01 Note: bold gal/sqft values are means; regular font gal/sqft values are medians
34
The first, most obvious finding from the graph is that, turf application exceeds xeric application
by a large statistically significant margin in every month. Ultimately, this is what constitutes the
large annual savings seen at the annual landscape application and total home consumption levels.
FIGURE 5: Monthly Per-Unit Area Application for Turf and Xeric Areas
The data also suggests, among other things, that the reason for the aforementioned enhancement
of savings during the summer is because turf application peaks drastically in the summer whereas
application to xeriscape does not. A graph of the difference between the groups (Figure 6)
demonstrates this is the case, and the observed pattern in savings obtained each month parallels
the pattern observed for turfgrass application (Figure 5). It appears that the reason xeriscape
saves so much water in this climate is related as much to the high demand of turfgrasses vs.
plantings of most other taxa as it is to any inherent aspect of xeric landscape per se. Furthermore,
inefficiencies in spray irrigation system design, installation, and operation further contribute to
the savings of having xeric landscape in place of turf because these inefficiencies even further
drive up application to the turfgrass to the point that it is much higher than the rate of
evapotranspiration over the same timeframe (Figure 7).
Additional inferences can be made about the application of water to turfgrass areas by the
participants. Specifically, on average, whereas they irrigated relatively efficiently in the spring,
with the onset of summer temperatures in May, residents quickly increased their application,
ultimately going way above ETo. Moreover, they tended to stay well above ETo through
November. While it is expected that due to system inefficiencies, a high Kc for Fescue
(Source: Cooperative Extension Office), leaching fraction considerations, and other factors,
application usually would tend to exceed ETo for turfgrass locally, the pattern suggests that
35
overall people irrigate relatively efficiently in spring as the weather warms and ETo rises,
probably due to the immediate feedback they receive as the grass yellows in response to moisture
deficits. As they observe their landscape beginning to show visible signs of stress due to deficit
irrigation, they increase their application accordingly. However, in May, they appear to start
overreacting to the increasing stress and increase irrigation to well over the requirement. In fall,
they do not however appear to respond in a correspondent way “coming down the curve,”
probably because they do not have the same sort of visual feedback mechanism as they do in
spring (i.e., they do not view the grass being “too green,” wet, nor the occurrence of runoff as
something amiss). The result is a long lag in returning to application rates more closely
approximating ETo in the fall and early winter (Figure 7).
FIGURE 6: Monthly Per-Unit Area Savings (Turf Area Application– Xeric Area
Application)
It is more difficult to make similar types of inferences with respect to xeric area application.
While there is research under way on a variety of desert taxa to attempt to quantify irrigation
demand and there have been generalized attempts to model or approximate xeriscape need based
on observations and fractions of reference ETo, at this time it would be risky to make highly
specific inferences. The relative flatness of the xeric curve in Figure 5 does though seem to
suggest that residents may irrigate xeric areas inefficiently as they seem to show little response to
demands of different seasons.
36
FIGURE 7: Monthly Per-Unit Area Application to Turf and Reference
Evapotranspirational Demand
FIGURE 8: Monthly Per-Unit Area Application to Xeric Areas
and 1/3 of Reference Evapotranspirational Demand
37
If one does assume a sometimes-used local “rule-of-thumb” which states that xeriscape requires
about a third of what turf needs, one can compare per-unit area application for xeriscape and this
modified reference value (Figure 8). Using a one-third ETo value is not out-of-line with
modification approaches employed by the Irrigation Association16 (2001) or WUCOLS17 (2000)
for estimating needs of low-water-use woody taxa in high-temperature southwestern regions. It is
quite noteworthy that the summation of monthly xeric-area application values yields a yearly
xeric-area application usage of 30.1 inches per year - nearly identical to the summation of
monthly .33(ETo) values, which is 30.5 inches. This would appear, initially at least, to suggest
that this rule of thumb may work quite well on average for approximating xeric landscape usage
over broad spatial and long temporal scales, even if it may not precisely work in a given month.
Normalizing these aforementioned potential reference values and the absolute departure from
these in observed water application may reveal insights about when during the year the greatest
absolute potential savings can be obtained. In Figure 9, this is done such that the absolute
difference between mean application and respective references is quantified and displayed. Here,
“0” (reference) is ETo for turf and .33(ETo) for xeric landscape respectively.
FIGURE 9: Absolute Departure in Irrigation Application from Derived Respective
Reference ET0 Values (Turf and Xeric Areas)
Even with the xeric reference but a third of ET0, it appears that, in addition to the differences due
to plant usage, much more water is wasted in application to turfgrass than to xeric landscape. The
38
greatest waste for turfgrass occurs in the period of May through November. Thus, any
improvements in turfgrass irrigation efficiency during this timeframe will have the greatest
absolute impact in terms of water conservation. Interestingly, the greatest absolute potential for
savings for xeric areas is not during this period, but rather from September thru January. Indeed
to look upon the graph, one might initially conclude that residents under-irrigate xeric areas in
spring and summer. Caution should be observed though in this type of reasoning as the .33(ET0)
reference is only theoretical and developed here as a guideline. That stated, the findings may
suggest that, on average, little potential exists during the spring and summer for significant water
savings by irrigation improvements to xeriscape. Finally, on an absolute basis, little total
potential appears to exist for squeezing additional conservation out of xeric landscapes as,
considered over the span of an entire year, xeric area irrigation appears to be efficient.
In contrast, opportunities to save great volumes of water appear to exist for turf areas throughout
most of the year. Significant overwatering appears to occur May through November; efficiency
improvements will yield the most absolute benefit during this period of the year. But how does
the issue appear when one considers the problem through the perspective of when can the most
readily obtainable savings be achieved?
Considering absolute irrigation departure from reference as above gives insights into the total
potential to save water through a variety of irrigation improvements. However, there is also the
question of how much water could be saved principally by relatively simple improvements in
controller management. Figure 10 is such an attempt to view the problem through this
framework,where the blue line is ETo for turf and .33(ETo) for xeric areas respectively, and is
equivalent to 100% of each respective types reference value or “perfect efficiency.” Absolute
values for inches application were normalized by converting them to percent departure from
normalized respective reference values. In this way the relative departure from these
aforementioned references is displayed as a percent value.
FIGURE 10: Relative Departure in Irrigation Application from Derived Respective
Reference ET0 Values (Turf and Xeric Areas)
39
Figure 10 may suggest that there are specific times of the year when people irrigate both turf and
xeric landscapes more or less efficiency than the ideal. As interpreted from Figure 10, the most
inefficient irrigation, in a relative sense, may actually occur during non-peak months if efficiency
is defined to be the difference between theoretical requirement and application. Expanding on
this type of analysis and breaking the above relative departure values into efficiency classes
yielded a summary of when people appear to irrigate most and least efficiently (Figure 11).
FIGURE 11: Relative Departure in Irrigation Application from Derived Respective
Reference ET0 Values (Turf and Xeric Areas)
It is well understood that, in practice, there is no such thing as a perfectly efficient irrigation
system and, for this reason, the green designation in Figure 11 includes relative applications
ranging from subreference values to those up to 20% above reference (this allows that there is
typically a need in practice to compensate for lacking distribution uniformity in irrigation
systems).
Interpretation of Figure 11 suggests that both xeric and turf areas are irrigated relatively
efficiently in the spring. Irrigation efficiency for turfgrass areas starts to decline in May to the
point where significant waste starts to occur and this continues until about September. In contrast
xeric irrigation continues to be quite efficient during this time. Around September, turf is starting
to be very inefficiently watered, in a relative sense, owing to residents’ failure to respond to the
lower rate of evapotranspiration and decrease irrigation accordingly. A similar, if less severe,
pattern is observed for xeric area irrigation, where at this time, these areas are also beginning to
be irrigated inefficiently, probably for the same reason. By November, both xeric and turfgrass
areas are, on average, being severely over-irrigated and this pattern continues through the cool
season until February. Finally, efficiency starts to recover and both areas are actually being
irrigated under suggested reference values by the end of March.
It needs to be acknowledged that some of this conclusion includes theoretical and speculative
reasoning, especially considering the lack of data on xeric landscape water requirements and the
fact that in actuality stress impacts, including those from water stress, lag in woody vegetation
(Kozlowski et al. 199018) so efficiency as considered here is much harder to gauge.
Nevertheless, again, failure of residents to more effectively tie controller management (irrigation
frequency and duration) to the changing environmental conditions appears to be one of the most
pressing reasons for efficiency losses in both study groups, it is just to a lesser extent (and much
lesser absolute impact in gallons) for those with more xeriscape.
Irrigation application 0-20% over reference
Irrigation application 20-50% over reference
Irrigation application >50% over reference
Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Landscape Type
Turf Area
Xeric Area
40
This set of analyses provides SNWA with quantitative data on what parts of the year it should
focus its strongest controller-management-oriented conservation messaging. This could be
considered the “low hanging fruit” in terms of water conservation; it is where messaging to effect
changes that may not require significant work and monetary investments on the part of residents
may produce significant water conservation results. To recap, the findings in this section suggest
the most value can be obtained by targeting controller-management messaging to the late summer
and early fall as people begin to depart from “reasonable” efficiency values owing to their
collective failure to adjust irrigation down for the cooler, low ET season. Reemphasis of this
messaging should continue all winter long.
The exploration of application per-unit area vs. reference values is important for making
inferences about management efficiency of water application. This; however, should not obscure
the result that on average, per-unit area, xeric landscapes in this study received much less water in
totality (Figures 3 and 4) and the pattern of received irrigation showed much less tendency
towards “peaking” (Figure 5) than those areas planted with turf.
SOURCES OF SIGNIFICANT VARIABILITY IN SINGLE-FAMILY RESIDENTIAL
CONSUMPTION
As explained in Methodology, multivariate regression analyses were employed to identify and
quantify sources of variability of mainmeter and xeric submeter data. Specifically, variables in
the combined study groups are explored for association to total household consumption and, for
the XS Group, to xeric landscape submeter consumption. Regression modeling proceeded with
the goal being to yield an optimum combination of the highest reasonable R-squared value with
due consideration given to maximizing the degree to which the model was “complete” (to the
extent possible given the available collected data). Details of the final selected multivariate
regression models appear in Appendix 2. Explanation and discussion of each variable included
follow for each of the respective models.
Presented models are only designed to broadly assess variables’ impacts. The models presented
here are “estimation” models as defined (see Methodology). These models are not intended for
use as “engineering” or “computational” type model applications whereby collecting certain data
one could be reasonably certain that the answer yielded would closely approximate the real
consumption at a given property.
41
Variability in Annual Residential Consumption
Discussions of the selected independent variables included in the annual consumption model for
the dependent variable annual residential consumption (labeled MAINMETE) follow. Overall,
the annual consumption model appears to be a very good “fit” (adjusted R2= 0.80) for this type of
work (Nelson3 1994, Gregg4 et al. 1994, Gregg19 et al. 1999). This is likely due as much to the
strong tie between outdoor usage (and the ability of independent variables associated with outdoor
use to be practically measured) as to any design elements or analytical methods associated with
the study. While relatively strong for the sample size, it must be stressed that this model’s utility
is mostly in terms of helping to uncover and, to some extent, explain variables discreet
associations with consumption at single-family residences. Quantifications of these associations
in the multivariate context are limited to only those variables deemed significant.
TOTALTUR
Definition of Variable:
The total amount of turf at a residence in square feet as determined by research personnel. This
includes all turf regardless of whether it is part of a submetered area and regardless of what type
of grass it is.
Results and Discussion:
This was the most significant variable by far (t=14.86), and was found to be strongly positively
associated with single-family residential consumption. It is a principal component of the model,
contributing the bulk of its strength (β=0.622). The results suggest that consumption increases
roughly 59.1 gallons annually for each square foot of turf at the average home. It then increases
further if the grass is Fescue (the impact of Fescue vs. other grasses is further explored below).
Since the alternative grass is almost always Bermuda, the result suggests the average application
rate for this warm-season grass by the study participants is about 59 gallons per square foot (see
variable FESCUE for more discussion on this).
It should be noted that earlier multivariate work attempted to deduce the influence of landscape
type by scrutinizing how much xeric landscape was found at a residence (DeOreo8 et al 2000).
While this is an acceptable approach, the amount of turfgrass present appears to be much more
closely correlated with total annual consumption and, when included, typically displaces xeric
area as a significant variable in the final models developed. Furthermore, since the amount of
xeriscape was not significant in multivariate context (nor were other individual landscape types) it
should be understood that the savings developed by SNWA’s Water Smart Landscapes program
are mostly due to it, in essence, being a turf-removal program more than an alternative-landscape-
promotion program. The results also suggest further significant lowering of household
consumption probably would not be yielded by permitting the owner to get a rebate for turf
removal at the expense of a quality landscape (for example, incentivizing the aforementioned
“zeroscapes” at a higher SNWA incentive rate since they have no vegetation and theoretically
require no water – this has been suggested by some). Since the xeric area contribution to annual
consumption is so small, the substantial loss in quality of life yielded for the small gains in
42
conservation realized by effectively hardscaping landscape areas makes the argument for
choosing hardscape in place of xeriscape for water conservation a position difficult to defend.
TOTVAL
Definition of Variable:
The dollar value of the single-family residential study property as specified in the Clark County
Assessor’s Office database. This should not be considered to equate to a home’s market value.
Results and Significance:
The assessed monetary value of the property, like the amount of turf at a residence, was a very
highly significant variable in the model (t=5.45). It is reasonable to assume that higher value
properties are associated with higher consumption because (i.) they are likely to contain larger
homes with typically larger, possibly more extravagant water-intensive landscapes and (ii.) they
are, by nature, likely to be inhabited by people of greater wealth who are less sensitive to the price
of water and thus more likely to use a greater volume of it. In a multivariate context, annual
water consumption on average increases ca. 2.1 gallons alongside each dollar increase in
Assessor’s Office property value.
That increased wealth is associated with greater individual consumption is a well-understood
tenant of economics and is a well-established concept in understanding persons’ household utility
consumption patterns. The impact of wealth in a similar context was explored by Gregg19 et al.
(1999) where the impact of neighborhood wealth was a significant factor in determining water
usage.
NLTHOMEA
Definition of Variable:
The age of the residence is calculated as the difference between the analysis year (2004) and the
year of construction as recorded in the Clark County Assessor’s Office database. This should not
automatically be taken to be the age of the landscape or even, necessarily, the exact age of the
specific study residence due to the way many developments are built as components of phases in
this community.
Results and Significance:
This was a quite significant variable (t=2.67) and one easily worthy of inclusion in the model. On
average, consumption increased ca. 1600 gallons for each additional year older the property was.
There are several potential reasons for this. First, older properties in the Las Vegas area tend, on
average, to be larger and the ratio of hardscape footprint to landscapeable area is lower. Next,
older properties are more likely to incorporate landscape elements heavy on traditional themes
(i.e., large areas of turfgrasses) in contrast to newer residences with landscapes built in a time
where water conservation began to be a significant consideration (in the 1990s restrictions on the
amount of turfgrass that could be installed at single-family residences were passed). Older
properties are more likely to have irrigation systems that incorporate lower-efficiency devices and
43
fixtures (ex. brass spray heads). Finally, as irrigation systems age they inevitably become less
efficient and more likely to leak.
Aspects of indoor use also likely contribute to the pattern. The installation of high-efficiency,
low-flow fixtures and appliances after being legally mandated is anticipated to have contributed to
newer properties having, on average, lower consumption. Also, as fixtures wear they may leak
for some time without notice (toilet flappers for example) so, without timely maintenance, older
properties are more likely to have continuous indoor leaks further contributing to higher
consumption. The increased efficiency gains in homes with newer fixtures have been well
documented (see Mayer and DeOreo8 et al. 1999) and the overall finding that older homes tend to
have higher water consumption is not surprising.
APROXINC
Definition of Variable:
Approximate total household income as revealed by 2001 survey data. To make the income
survey question less intimidating, and more likely to generate valid, significant numbers of
responses, the potential answers were categorical with ranges and it was explicitly stated that this
question was optional. Analysis proceeded based on the mean values of response ranges. While
a great number of participants did respond, many of course did not and income is, unsurprisingly,
the most limiting of independent variables in the multiple regression.
Results and Significance:
It is to be expected that, everything else being equal, increasing household income would on
average be associable with higher per-household consumption of all commodities. This is the
case for water as well in this multivariate model, which suggests that, on average, annual
consumption may increase on average ca. 3000 gallons for every $10,000 rise in income level
(t=2.16). Some may be surprised this should be given the fact that indoor water use is relatively
constant per capita across a range of conditions and thus the sensitivity of the relationship
between water consumption and price is usually considered to be rather muted. But, while water
is indeed inelastic by common economic standards, in the Southwest, where a high proportion is
used outdoors, it may be considered to be more discretionary in nature, especially when that
outdoor use is for irrigation of landscapes (instead of crops), which are after all just ornamental.
Certainly this study suggests that income is an important consideration in water consumption, as
have others. Furthermore, higher incomes could be considered to be well correlated with large
houses, large landscapeable areas, and more lush landscapes, all of which further drive up
consumption in their own right.
There was considerable discussion between the principal author and some reviewers as to whether
or not the income data should be included in the model. The arguments for inclusion were that it
was found to be a significant variable in most comparisons, it is a different indicator than home
value in that the former is more indicative of wealth and the latter is more indicative of actual
disposable income (which could be spent on water use beyond necessity), and that removing it
significantly weakens the model. The arguments for removing it include the supposition that
often people give erroneous or fictional answers to questions about income, that income is
potentially highly covariate with home value, that home value is really a better proxy variable for
44
income (and indeed in many studies using multiple regression it has been used for this purpose),
and that its deletion does not weaken models such as this. Finally significant improvement in
model sample size would be obtained by removing income as many people opted not to report it
and thus it is very limiting to the model’s available degrees of freedom.
The author considered the arguments for and against inclusion of income data carefully and
proceeded to investigate the relationship between income and home value. The results of a
correlation analysis between these two variables showed relatively little correlation (R2= 0.288) as
did a scatterplot of the data. Nonetheless, the concern was valid enough (and the possibility of
significantly more degrees of freedom of sufficient interest) to justify creation of an incarnation of
the model without income as an independent model variable. This exercise however resulted in
an increase in the standard error of the estimate (i.e., an increased error of over 7,000 gallons per
year) and a drop in the overall model fit (adjusted R2= 0.740). However, most tellingly, the
B values were off significantly from what one would expect (ex. Variable POOL B= 27.8; yearly
evaporation in gallons per year is far in excess of this). Based on these findings it was decided
that the APPROXINC variable should remain in the model.
FESCUE
Definition of Variable:
Whether or not the turfgrass present at a residence is Fescue or an alternative turfgrass. This is
a binary (i.e., “dummy” in the vernacular) variable indicating presence (1) or absence (0) of a
variable’s specified condition.
Results and Significance:
Fescue grasses (which are widely popular cool-season grasses found in local landscapes) have
been observed to require large volumes of water in the Las Vegas area (ca. 91 inches), over 62%
more annually than the other somewhat less popular warm-season Bermuda grass (requiring ca.
56 inches; calculations for both grasses are based on data from the local Cooperative Extension
Office). Locally, Fescue is much less drought tolerant than Bermuda and has a correspondingly
higher Kc value (the July Kc value for Fescue is calculated to be a very high 1.10 whilst only
being ca. 0.71 for non-overseeded Bermuda; Source: University of Nevada Cooperative
Extension Office).
Furthermore, being a cool-season grass, Fescue is capable of active photosynthesis all year long
with sufficient irrigation and management, which is no doubt the reason for its desirability; it can
yield an attractive green year round. Bermuda on the other hand usually goes into dormancy in
the winter and it is likely many people curtail irrigation at dormancy so its total yearly application
is even further reduced relative to Fescue. While there are of course different requirements for
different types and morphologic forms of grasses (ex. tall vs. short fescue), the general finding
that the cool-season grasses require more water than the warm season ones is well understood and
this apparently translates into residences with Fescue having, on average, higher annual
consumption at the household level (t=2.09) (note: most residences had at least some turfgrass
integral to their landscapes). Based on the multivariate analysis, a residence with Fescue may on
average use more than 25,000 gallons more annually than one with a lower-water-use grass.
45
There is another possible inference that may be made. The submeter data is heavily dominated by
Fescue landscapes and thus the highlighted gallons-per-square-foot application rates are probably
at or near the actual for Fescue. It should be noted though that from the model, one might infer
that in situations where there is not Fescue at the site, the B value of 59.1 may be the typical
application rate, in gallons per square foot per year, for Bermuda installed at a residence. Though
this derived value of 59.1 gallons per square foot per year (94.9 inches precipitation equivalents)
is somewhat suppositional, and no doubt not exact given the standard error of the model, it
appears to be a very reasonable average application rate that could be expected locally for
Bermuda grass.
PARCEL SIZE
Definition of Variable:
The size, in square feet, of the parcels of study residences as specified in the Clark County
Assessor’s Office database.
Results and Significance:
In the final version of the model, parcel size was technically not significant (t=1.79); however, it
was positively correlated with higher residential consumption in most multiple regressions
developed so it is included here. It is reasonable to assume that, on average, residences associated
with larger parcels are more likely to have higher consumption because they would be expected to
have (i.) more, possibly lusher, landscape (they are also more likely to have a pool) and
(ii.) typically larger homes situated on them. Both of these would be anticipated to raise
consumption due to larger residential landscapes having higher total outdoor irrigation
requirements and larger houses being more likely to be inhabited by more or, perhaps, simply
more heavily consuming, residents.
POOL
Definition of Variable:
The total water surface area of pools and spas in square feet at residences as measured by
research personnel. For residences without pools this variable equates to zero.
Results and Significance:
As with parcel size, pool surface area was not significant in the final most complete version of the
model (t=1.70), but often cropped up as significant in alternative models as being positively
correlated with higher consumption. It is reasonable to include this variable as it is to be expected
that the more evaporative water surface area outside at a residence owing to a pool and/or spa, the
higher the evaporative water loss at the residence and the greater the need, in gallons, to replenish
it.
TOTALOCC
Definition of Variable:
The total number of occupants at each study property in the analysis year (2001) as determined
by survey.
46
Results and Significance:
Though not a statistically significant independent variable in the final model (t=1.62), and only
occasionally significant in alternatives, the number of people living at the residences was
ultimately included, as it lends explanatory strength to the model (β=0.524) and it is logical to
assume that consumption does increase with more people living at a location. That it is not
statistically significant is actually a testament to the dominance of outdoor end uses in
determining total yearly consumption at single-family properties in this region.
TOTALLAN
Definition of Variable:
The total landscapeable area at a property. This includes areas with landscape as well as areas
potentially landscapeable.
Results and Significance:
This variable is difficult to interpret and was not significant in this particular model (t=-1.41).
The only reason for its inclusion is the sheer number of times it cropped up as significant in
different alternative models. Here, however its sign is inverse of what would be anticipated (that
greater landscapeable area would lead to higher consumption). It may be that it captures the
inverse of the building and hardscape footprints, but this is only theory.
check from here on…
Variability in Annual Consumption for Irrigation of Monitored Xeric Landscape
A model of yearly consumption for the monitored xeric component of landscapes for XS Group
homes was also developed to attempt to evaluate the impacts of variables listed in the Scope
(Appendix 1). The developed model has a much lesser fit than the total consumption model
(adjusted R2= 0.40), in part, one speculates, because other important but non-quantified or hidden
variables are not included (one possible example – detailed data on controller management which
may be more associated with management of turf rather than xeric areas). For this reason, no
attempt is made to quantify impacts in a multivariate context as above, but rather the goal is to
identify variables likely associated with xeric area consumption (for some attempts at
quantification using univariate approaches consult Sovocool and Rosales11 2001).
Despite the limitations due to the weaker model, many variables did appear significant in most if
not all modeling attempts, and these are discussed below in a format similar to the above
discussion on annual consumption. The same strength of association denotation as used for the
annual consumption model is applied to the xeric areas variable discussion as well. See
introduction to Sources of Significant Variability in Single-Family Residential Consumption for
more information.
47
TOTALCAN
Definition of Variable:
The total canopy coverage in the monitored xeric area of the XS Group properties, in square feet.
This is calculated by first taking the observed plant diameters from the 2001 site review, dividing
this number by two to get radius, then applying the formula for getting the area of a circle
(A=πr2). This area result is then multiplied by the quantity of those plants observed to be at that
size. The summation of all areas of all plants of all size classes in the study area is the total
canopy coverage.
Results and Significance:
It is reasonable to expect that total plant canopy coverage within the monitored xeric area would
positively correlate to the total amount of water applied to that area as plant leaf surface area
(evapotranspirational area) is the principal locale of water loss from vegetation. To replace this
loss, areas with higher plant coverage should theoretically require more water and it should be
expected that residents would respond by irrigating these more (via both longer run times and
having irrigation systems of greater application capacitance). Examination for a link between
total canopy coverage and total yearly consumption for xeric areas in a multivariate context
confirms a significant association (t=4.31; the relationship between coverage and per unit area
consumption was also noted and explored in Sovocool and Rosales11 2001). One
acknowledgement; this is a relatively simplistic finding, which does not fully explain the
relationship between consumption and the taxa present and species’ specific water use
characteristics (this was beyond the practical scope of this investigation). Data on specific xeric
species’ water requirements is needed for this and this area remains worthy of more in-depth
research.
AVGFLOWR
Definition of Variable:
The arithmetic average flow rate, in gallons per minute, of all irrigation stations servicing
monitored xeric landscape for each of the XS Group properties.
Results and Significance:
It has long been suspected that within the range of lower flow types of irrigation systems used to
irrigate xeric areas, those capable of delivering water relatively faster via high-flow emitters may
contribute to higher water consumption, especially when used by someone less knowledgeable
about how to irrigate with different types of emitters. For this reason, SNWA’s current Water
Smart Landscapes program limits individual emitters to a maximum output of 20 gph as part of
the program requirements (Appendix 5). Based on this research, this concern appears well-placed
as the model shows stations with higher average flow rates are indeed associated with higher
consumption in this study (t=4.14). Typically, such station configurations may have one or more
of the following conditions: sprays used for xeric-area irrigation, incorporation of high-flow
emitters (such as turf bubblers), use of microsprays, stations composed of mixed types of
irrigation emitters, and individual stations irrigating large and/or lush expansions of xeriscape (an
exploration of how emitter class relates to average flow rates also appears in Sovocool and
48
Rosales11 2001; this manuscript suggested a strong association between irrigation system design
and xeric area consumption as well).
STUDYA
Definition of Variable:
The xeric study land area (in square feet ) monitored via submeter for XS Group properties.
Results and Significance:
It is logical to assume that, on average, the more area monitored by the submeter, the greater the
consumption through that meter, and the significant association between monitored xeric-study
area and total yearly consumption (t=3.08) is consistent with this expectation (for further
exploration of per-unit area savings, see Comparison of Per-Unit Area Water Application between
Turfgrass and Xeric Landscape).
TOTVAL
Definition of Variable:
The dollar value of the residence as specified in the Clark County Assessor’s Office database.
This should not be considered the same as the home’s market value.
Results and Significance:
There was a positive association between the total value of the property and total consumption for
xeric area consumption (t=2.94). A discussion of how this variable tends to be positively
associated with water consumption appears above in the discussion of the annual consumption
model. It is worthwhile to again emphasize that given water use for residential landscapes can
ultimately be considered discretionary, higher homeowners’ wealth (here, evidenced by higher
property value) may be anticipated to lead to greater consumption for landscape irrigation.
PARCEL SIZE
Definition of Variable:
The size, in square feet, of the parcel of a study residence as specified in the Clark County
Assessor’s Office database.
Results and Significance:
The parcel size of the residence was significantly inversely associated with consumption for xeric
area irrigation (t=-2.78). This result was unexpected, as a relationship or mechanism acting to
result in a link between parcel size and the irrigation of xeric areas on that parcel is not
immediately obvious. The possibility that there is an inverse relationship between xeric study
area and parcel area was examined, but this is not the case (rather, as would be expected, larger
properties tended to be positively correlated with larger study areas, though this relationship is
weak; R2=0.064). Likewise, the theory that perhaps larger parcels had xeric areas that might be
sparser in terms of canopy was examined and rejected (the data does not support this).
49
Discussion and consideration of other findings led to some other possible explanations. One
possibility is that those residences with larger parcels were more likely to incorporate native,
lower-water-requirement plants in their landscapes. Some data supports the theory that owners of
large properties may indeed make more use of native taxa, but only marginally so (the properties
in the top 10% in parcel size had an average of 10.9% of their plant palette composed of native
vegetation; the average for the rest of the properties was 6.9%).
Another theory is that larger xeriscape installations may be more likely to necessitate the need for
a contractor, who is more likely to install a properly designed drip system and, as suggested by
the findings linking station flow rate to consumption and (as revealed below) “drip-only” systems
are more likely to result in lower total yearly consumption than those piecemealed together with
multiple types of emitters. Since those residents doing larger xeriscape conversion projects were
found to be more likely to use a contractor, there is some evidence supporting this second theory.
Perhaps the most likely reason for this finding is that people with smaller parcels are more able to
afford to consume more water outdoors. To understand this reasoning better, consider an
example of two sets of land, one acre each, in a similar area and climate each with all
landscapeable area landscaped. One has a single residence upon it, the other acre, more
subdivided, supports five homes (and thus is composed of five parcels). One would conclude,
usually correctly, that the outdoor consumption for the total area would be greater for the one-
home case, owing to its maintaining a greater amount of landscaped area (more of the available
area is consumed as development in the five-homes case). But what about total water
consumption for irrigation on a per-parcel basis? Each of the family income streams in the five-
homes-per-acre case support less irrigated area than that for the single home on the one acre.
Thus, each of these five owners can afford to support more discretionary water use as their
respective landscape irrigation “shares” are less than for the one owner supporting all of that area.
As a result, the owners of the smaller parcels may use more irrigation water per parcel than in the
alternative case, and this may be what is being observed here (internal research by SNWA has
shown that subdivision tends to result in higher per-parcel usage while decreasing usage for the
total equivalent area).
Without more information, these are only hypotheses. At this time, while the inverse relationship
between parcel area and xeric area consumption stands, the mechanism behind the relationship is
not completely understood.
DRIP
Definition of Variable:
Presence (1) or absence (0) of an exclusively drip irrigation system irrigating the xeric study
area. This is a binary variable.
Results and Significance:
This is a different type of measure of the influence of irrigation system design on total xeric area
water application. Specifically evaluated was whether the presence of a “true” drip system (no
bubblers, microsprays, mixed systems) was associated with xeriscapes with lower consumption
than others. The model does support this theory, with a significant finding that such “drip-only”
50
xeriscapes do have lower consumption (t=-2.27). As suggested by Sovocool and Rosales11 2001,
such systems typically have the lowest flow rates (average per-station flow rate = 4.0 gpm) of the
types used to irrigate xeric landscape, so if run similar amounts of time to other systems, it would
be expected that these would output lower total volume over a year. Based on the data, it does
seem likely that many residents run their systems without careful consideration as to which kind
of emitters they actually have, in turn resulting in systems composed exclusively of true drip
emitters being associated with the least amount of water consumed over the year. Since slow
application rates are generally the preference in irrigating drought-tolerant vegetation (this is
especially the case with woody taxa) and because landscapes with “true” drip systems had the
lowest consumption, this finding may be worthy of future considerations relevant to SNWA’s
Water Smart Landscapes program.
DONTKNOW
Definition of Variable:
Whether or not the respondent was knowledgeable about the level of enforcement of local
restrictions designed to reduce water waste. This binary variable indicating presence (1) or
absence (0) of understanding was adapted from part of an alternative answer to a question asking
respondents if they felt enforcement of water waste provisions was "too lax," "good," or "too
strict." In addition to these responses, residents taking the survey were also given the option of
answering “Don’t Know” if they did not have any sense of how aggressively water waste
regulations in the area were enforced.
Results and Significance:
Theoretically a person’s viewpoints on water waste enforcement could tie into how diligently they
practice good irrigation management. Recognizing this, the study staff formulated a question
addressing this for the survey implemented in 2001. In preliminary analyses (Sovocool12 2002)
there really was not a difference in per-unit area irrigation for xeriscapes between those
respondents answering “too lax” and “good” (only two people said enforcement was “too strict”
resulting in no ability to tie this to consumption with any statistical precision, though this is quite
telling of how the community viewed enforcement in 2001). However, interestingly there was a
difference between respondents with any kind of an opinion and respondents who had no sense of
the issue. This suggested at the time that awareness of enforcement of water waste regulations
may be a principal motivator to conserve, regardless of one’s viewpoint on how appropriate the
level of enforcement is. The recurrence of this basic result, here in a multivariate scheme (i.e.,
those answering “don’t know” were associated with higher consumption in the regression model;
t=2.13) suggests that sensitizing the public about enforcement of water waste restrictions may be
a powerful motivator for achieving outdoor water conservation.
FINANCIAL SAVINGS ASSOCIATED WITH CONVERSION PROJECTS AND COST EFFICIENCY
As explained in the methods section, the research scope included a mandate to study some of the
economics of xeriscape conversions, as this has been left relatively uninvestigated to date.
Specifically, the directives were to quantify costs associated with the conversion and the
subsequent maintenance of the xeriscape and to develop estimates as to what incentive level
51
would theoretically be necessary to entice people into doing conversion projects. Collection and
analysis of this data is explained in Methodology, below, and in Appendices 5 and 6. Results are
as follows below, starting with the conversion costs findings.
The average cost of the conversion for those converting in the XS Group was obtained via data
collected on parts and materials, as well as contractor receipts. The average cost for all
participants was $2,881.21 for 1,862.1 sqft converted ($1.55 per square foot for 91 participants
sampled). The average cost for those who did the conversion themselves was $2,428.31 for
1,766.22 sqft ($1.37 per square foot), and the cost for those hiring a contractor was $4,076.88 for
2,115.22 sqft ($1.93 per square foot). These dollar amounts for costs and dollar valuations are as
they stood in the late 1990s and have likely climbed slightly by today. As might be anticipated, it
appears that residents may on average be more likely to hire a contractor for larger conversion
projects.
Landscape maintenance requirements constitute a significant cost in labor and dollars directly
spent. The relative amount of xeriscape at a residence figured prominently in landscape
maintenance reductions for both these costs (Figure 12). For those who had at least 60% of their
landscapeable area as xeric landscaping, maintenance savings of about one-third were realized
versus those with 60% or more turf. The average difference is 2.2 hours/month in labor and $206
per annum in direct expenditures (N=216). Landscape maintenance savings are value added on
top of water bill savings. This serves to greatly enhance the attractiveness of xeriscape to the
customer. Hessling12 (2001) provides a detail of the capital costs and savings obtained.
FIGURE 12: Average Monthly Maintenance Time and Annual Direct Expenditures
for Participants Having At least 60% Turf or Xeriscape
8.17
5.95
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
MaintenanceHours / monthTurf Maintenance Time Xeriscape Maintenance Time
$680.03
$473.93
$0
$100
$200
$300
$400
$500
$600
$700
$800
Maintenance Spending Dollars / yearTurf Maintenance costs Xeriscape Maintenance costs
52
Bill savings for a typical mid-consumption range customer were modeled as explained in
Methodology and in Appendix 4. Results show that there is a large difference in the monthly bills
between a modeled residence with and without the conversion throughout the majority of the year
(Figure 13). The total difference in the annual cost for water between these two homes using the
current (2004) rate structure is $239.92 - a significant savings attributable to the conversion
(nearly $0.15 per square foot converted per annum). It should be noted that this savings of 54%
in total annual water charges is greater than would initially be anticipated from consumption
savings data (Figure 6). This is because the Las Vegas Valley Water District, as well as the other
SNWA member agencies, uses a tiered, increasing block rate structure.
Increasing block rate structures (also called conservation rate structures) are setup such that the
more a user consumes on an average daily basis within a cycle, the more expensive, per unit
(i.e., per gallon), water becomes. The high per-unit area application to turfgrass results in
residences with more grass typically crossing thresholds into higher billing rate strata much more
frequently and this in turn exacerbates their water costs per unit and, ultimately, their total costs.
In this case, the difference in per-unit water charges for the two fifth-decile homes, with all
charges considered over the entire year is about $0.28 per thousand gallons (i.e., there is a 13%
difference; effective prices of $1.85 vs. $2.13 per thousand gallons, respectively). The
comparison highlights the utility of tiered rate structures as a conservation tool and for promotion
of xeriscape as a conservation tactic.
FIGURE 13: Modeled Monthly Water Bill for a Typical Las Vegas Area Home and
The Same Home with an Average-Size Conversion
53
The expected water bill savings a resident of a typical home would realize from doing an average-
size conversion of turfgrass to xeriscape (anticipated monthly savings – including tier rate
impacts) is thus as illustrated in Figure 14. As can be seen, the typical monthly water bill savings
range from a low of $5.68 (25%) in December to a high of $40.84 (70%) in July, again
reemphasizing that the greatest savings obtained by having xeric landscape are realized in the
extremes of summer in this area. The savings obtainable serves to create a strong price signal to
convert, especially when coupled with the incentive offered by SNWA currently ($1.00 per
square foot for qualifying residential conversions).
As suggested by Figures 13 and 14, on average xeriscape not only results in significant savings in
water utility charges, it also makes the charges more manageable as they no longer “peak” to
anywhere near the extent they did under the “no-conversion” condition. For the “no-conversion”
model, the low-consumption month to high-consumption month ratio is 1:2.93 (the peak month is
July). For the same house with the conversion, the ratio is 1:1.58 and the peak is pushed out to
September owing to the difference in xeric irrigation pattern (Figure 8). For homes proximal to
the modeled condition, xeriscape conversions appear to make paying monthly bills easier because
the peak is (i.) greatly attenuated and (ii.) potentially pushed out until later in the year, so it does
not parallel other local utility bills which peak in the summer (power, for example).
FIGURE 14: Modeled Monthly Water Bill Savings for A Typical Las Vegas Area
Home Completing an Average Size Conversion
54
ESTIMATED APPROPRIATE LEVEL OF FINANCIAL INCENTIVE
Homeowner Perspective
Hessling13 (2001) performed analyses of the financial viability of SNWA’s xeriscape conversion
program, “Southern Nevada Xeriscapes” (since revised and renamed to “Water Smart
Landscapes”). It should be noted that at the time Hessling did his analysis, the program paid
recipients an incentive of $0.40 per square foot. He presented a Net Present Value (NPV)
analysis demonstrating that, from the homeowner perspective, the return on investment by
SNWA’s conversion facilitation program is two to three years for a resident and that the incentive
is not really required to induce change as the NPV is positive, even when no incentive is
rewarded. See Hessling’s manuscript for additional details.
A constructed model (Appendix 5) using a similar approach (and more recent data) seems to
support the finding that no incentive is theoretically necessary for a typical do-it-yourself xericape
conversion where subsequent financial savings in landscape maintenance are realized. However,
the incentive may be important in a variety of other situations. The scenario, similar to the one
used by Hessling as well as others, was explored by the model developed by SNWA
(Appendix 5). Some of the most common scenarios explored, with findings from model outputs,
are summarized in Figure 15.
In Figure 15, there are four different scenarios modeled (see explanation below), and each
scenario has four associated results (Methodology and Appendix 5). The outputs associated with
each exercise are: the average payback time (at a dollar per square foot) for a typical home doing
a typical conversion (see Appendix 5), the average payback time without an incentive, the
incentive required for a 3-year return on investment (ROI), and the incentive required for a 5-year
ROI. Special note should be made regarding the expression of payback times. The display is not
the range of payback times given the combination of scenario conditions, rather, it reflects that the
expected average payback time falls sometime between the years shown. The model is based on
annual, not monthly data thus the need to display outputs in this manner. The “incentive
required” outputs, are simply average model outputs and are not to be considered appropriate for
any one condition; their value is principally in comparative analyses between scenarios and in
broad generalizations.
The summary (Figure 15) is designed to facilitate inferences about the economics of the
conversion project. Along the horizontal axis are the “Only Conversion Material Costs” and
“Conversion Material Costs + Labor” titles. The first scenario condition refers to situations
where only the direct costs for materials, supplies, rentals, and other such items are considered.
Residents doing their own xeriscape conversion might consider this to be their scenario if they
consider only the real financial outlays paid and don’t consider their time spent on the conversion
to be a real financial cost. In contrast, the “Conversion Material Costs + Labor” condition
includes a valuation of the time to actually do the conversion, which naturally lengthens the
payback time. This perspective is more appropriate for those who consider the labor outputted by
55
themselves to be a true financial expenditure. It is also the appropriate model perspective to
consider if the project is performed by a contractor.
Along the vertical axis of Figure 15, are the titles “Only Maintenance Goods Conserved” and
“Conserved Maintenance Goods and Labor.” Similar to above, the “Only Maintenance Goods
Conserved” condition reflects consideration of savings associated with only direct expenditures
on things such as fertilizer, replacement irrigation parts, occasional replacement of capital items
such as shovels, etc. (so long as the conversion is significant enough to yield savings in these
areas; see the discussion surrounding Figure 12). The category “Only Maintenance Goods
Conserved” would be most appropriate for people who do not consider the savings in labor
obtained by having some of their area as xeriscape to be equivalent to a monetary outlay,
situations where not enough of the total landscape area is converted to obtain this type of savings,
or when a landscape maintenance company, which may or may not realize the savings, is either
unwilling or unable to pass on labor savings to the customer as realized by the landscape retrofit.
Again, there is an alternative category for the consideration of realized maintenance savings in
labor costs resulting from the conversion. The maintenance savings plus labor savings category,
“Conserved Maintenance Goods and Labor,” is most appropriate when enough of the yard has
been converted that real savings in maintenance labor can be realized and when the owner
attaches value to this. It would also be appropriate when the homeowner’s landscape company
passes on realized labor savings to him or her.
The matrix of results developed (Figure 15) permits some inferences to be made about what
scenarios turn around financially the fastest and are thus most readily facilitated by a landscape
conversion incentive. In increasing order of time to payback (i.e., the first bulleted scenario is the
most readily facilitated) these are:
• Situations where only the material costs of the conversion are valued and where savings in
both maintenance goods and labor can be realized (in fact, this scenario theoretically may
not even require an incentive to generate financial savings in an acceptable investment
timeframe).
• Situations where both the material costs of the conversion and the labor cost of the
conversion are valued and where savings in both maintenance goods and labor can be
realized.
• Situations where only the material costs of the conversion are valued and where only
savings in maintenance goods (not labor) can be realized.
• Situations where both the material costs of the conversion and the labor cost of the
conversion are valued and where only savings in maintenance goods (not labor) can be
realized.
Considering that the optimal price point for the first three of these scenarios is probably covered
by the current incentive level, but not the old $0.40-per-square-foot incentive, it may be that the
SNWA hit upon a critical threshold value in stimulating the marketplace when it went to the
Shorter Time to Investment Return
Longer Time to Investment Return
56
$1.00 per-square-foot level in 2003. A recent surge in program interest in the residential sector is
consistent with this (Appendix 5). Even in the fourth scenario, the current incentive level
shortens the payback time such that the project might be deemed affordable by many people (see
the associated 5-yr ROI). While few, if any, residents do a detailed economic assessment of the
payback time for their respective xeriscape conversion projects, the dollar per square-foot is
almost certainly perceived to make conversion projects much more “affordable,” plus there is
significant symbolic value to the $1.00-per-square-foot figure versus the past sub-dollar incentive
levels.
If the payback time outputs presented in this model are close to reality, it may be that SNWA’s
Water Smart Landscapes program will continue to experience high interest until a point where
materials, supply (i.e., practically convertible turf), or services associated with the conversion
project come to be in short supply and/or become expensive enough to cause feedback such that
program enrollment is slowed.
57
FIGURE 15: Summary of Economics of Typical Single-Family Xeriscape
Conversion Projects Under Different Scenarios
Only Conversion Material Costs Conversion Material Costs + Labor
Avg. Payback Time at $1.00 per SqFt:
Avg. Payback Time Without Incentive:
Incentive Required for 3-Year ROI:
Incentive Required for 5-Year ROI:
Avg. Payback Time at $1.00 per SqFt:
Avg. Payback Time Without Incentive:
Incentive Required for 3-Year ROI:
Incentive Required for 5-Year ROI:
Avg. Payback Time at $1.00 per SqFt:
Avg. Payback Time Without Incentive:
Incentive Required for 3-Year ROI:
Incentive Required for 5-Year ROI:
Avg. Payback Time at $1.00 per SqFt:
Avg. Payback Time Without Incentive:
Incentive Required for 3-Year ROI:
Incentive Required for 5-Year ROI:
Only Maintenance Goods Conserved (or when labor savings not realizable) Conserved Maintenance Goods and Labor $1.03/SqFt
$0.14/SqFt
5-6 Years
3-4 Years
$2.23/SqFt
$1.34/SqFt
8-9 Years
5-6 Years
None Req.
None Req.
2-3 Years
1-2 Years
$0.91/Sqft
None Req.
4-5 Years
2-3 Years
58
SNWA Perspective
The financial viability of SNWA’s Water Smart Landscapes Program is difficult to assess as
resource alternatives available to the Authority against which this “water option” may be
measured are diverse and have widely divergent respective costs (SNWA20 2003). Furthermore,
availability of water resources is not constant and shortage or surplus conditions can exist which
can make using these as standards against which conservation programs can be measured again
difficult. A prime and current example of this is the drought that the Lower Colorado River Basin
is experiencing which is currently impacting SNWA (SNWA Drought Plan21 2003). In these
types of situations, the economics of conservation programs are obviously enhanced, and it is
against this backdrop that the economics of the Water Smart Landscapes Program is being
considered in this study.
In Hessling’s analyses13, the drought had not yet been recognized and designated as such and
SNWA had no drought policies in place at the time of the analysis. He grounded his analysis in
comparing the marginal cost of water in the Southwest to the marginal benefit realized by the
incentive program. In doing so, he concluded that the cost of the incentive program at the time
was just offset by its resource value, and the program was thus a worthwhile initiative (see
analysis for details).
In 2004, a reanalysis of the Water Smart Landscapes Program was done to consider the economic
viability of it in the face of the drought and the current resource and program incentive values.
Given the current scarcity of local water resources, the drought, and the fact that SNWA is now
approaching the point of withdrawing its full Colorado River allotment (SNWA20 2003), the
Las Vegas Valley Water District has recently paid $9,500 per acre-foot for undeveloped
groundwater rights in the local basin and, furthermore, views this purchase as a bargain
(LVVWD22 2003). Because the largest purveyor member in the SNWA is willing to pay this
amount currently for undeveloped, non-administered water rights, this should be a good
alternative price for comparing the cost effectiveness of the program on a per-square-foot basis
(not including administrative and advertising costs).
It follows that to estimate the savings yielded by the program in dollars per square foot, the above
marginal cost of water, converted to a square-foot basis, can be multiplied by the savings per
square foot yielded by the conversion as below:
$9,500 per acre-foot X 325851 gallons per acre-foot X 55.8 gallons per sqft yield = $1.627 per sqft
The cost calculation is slightly more complex, as the SNWA not only spends the $1.00 per square
foot to incentivize the conversion, but it also forgoes the yield it would have claimed on this
amount had it invested it. The mature yield of municipal bonds in February 2004 is used as this
alternative rate. Thus the true cost per square foot for SNWA can be estimated as:
$1.00 per sqft expended + ($1.00 + 4.65% mature interest yield if invested instead) = $1.047
The cost-effectiveness of the program can then be calculated as the difference between these
values:
59
$1.627 per sqft saved - $1.047 per sqft saved = $0.58 per sqft net positive value to SNWA
The analysis suggests that for each dollar the SNWA is spending for the incentive, it is bringing
in $1.58 and that the program appears as such to be a good deal from a financial perspective for
SNWA. The ca. 37% net positive value means the program should be financially advantageous
even with addition of the program advertising and administration costs which have not to date
been quantified.
In considering the theoretical maximum that SNWA could pay for the program (a component of
Objective 6), it should be noted that $1.627 is not the maximum as, again, the yield of the
alternative investment must be considered. Subtracting out this missed or forgone yield results in
a figure of $1.55 and this is the theoretical maximum price SNWA could currently justifiably
sustain. Again, the actual maximum would be anticipated to be lower due to program
administration costs.
60
Executive Summary and Conclusions
The major conclusions of this research are as follows:
1. Xeriscape conversion projects can save vast quantities of water at single-family
residences. Homes in this study saved an average of 96,000 gallons annually following
completion of an average-size conversion project. This is a savings of 30% in total annual
consumption; a finding in line with those yielded by other research studies in this region.
2. Over the long timeframe of this study, total yearly savings have neither eroded nor
improved across the years. On average, household consumption drops immediately and
quickly stabilizes.
3. There is an enormous difference in application of water to locally used turfgrasses and
xeric landscape by residents. On average, each year residents applied 73.0 gallons per
square foot (117.2 inches) of water to grow turfgrass in this area and just 17.2 gallons per
square foot (27.6 inches) to xeric landscape areas. The difference between these two
figures, 55.8 gallons per square foot (89.6 inches) is the theoretical average savings
yielded annually by having xeriscape in lieu of turf in this area. This is a substantial
savings (76.4%) when considered in the context of the available residential water
conservation measures. A sub-study of other commercial properties with xeriscape found
the average application to xeric areas by these customers to be essentially equivalent to
that observed for the residential customers.
4. Over the course of a year, the difference in application between turf and xeric areas varies
in a predictable bell-shaped-curve manner, with the greatest difference occurring in
summer. This is because turf irrigation peaks to a much greater extent in summer than
xeric irrigation. The difference in irrigation between these two types of landscape varies
from as little as 1.56 gallons per square foot for the month of December, on up to
9.62 gallons per square foot for the month of July.
5. In comparing irrigation application to the reference evapotranspirational rate (ETo), it was
found that on average application to turf exceeded ETo in every month except March,
exceeding it the most May through November. In contrast, xeric application remained
well below ETo year round.
6. The author experimented with using a locally invoked “rule-of-thumb” which holds that
xeric plantings require about a third of the evapotranspirational rate as needed for turf. In
comparing this developed reference, 0.33(ETo), to application, it was found that these
values were, in absolute terms, somewhat close month to month and very close over the
entire year. In comparing this developed reference to application, it was found that xeric
application was below 0.33(ETo) half the year and above it the other half of the year
(September-February).
61
7. Relative to questions about irrigation management and the potential for further efficiency
gains, findings associated with conclusions 4 through 6 and subsequent analyses led the
author to the suggest that (i.) the greatest absolute savings from assorted improvements in
irrigation will be realized in the summer, but (ii.) the most readily obtained efficiency
improvements (i.e., not requiring capital outlays) yielded from better controller
management may be obtained September through January, as this is the period when a lot
of residents fail to successfully decrease irrigation in response to lower irrigation
requirements (for both types of landscape).
8. Multivariate regression modeling was used to help discover some of the factors associated
with variability in water consumption at single-family residences. These are:
i. The amount of turf at the residence (positive correlation).
ii. The property value of the residence (as indicated by the local assessor’s office;
positive correlation).
iii. The age of the residence (positive correlation).
iv. The total income of the property’s residents (positive correlation).
v. Whether or not the turfgrass present at the residence is Fescue (a locally popular
cool-season grass; positive correlation). As a side-result from one of the
multivariate analyses, Bermuda grass may be receiving approximately 59 gallons
per square foot per year – certainly less than the application for the much more
common cool-season grass in this study.
Some variables which were significant in many other incarnations of the model (but not
the final model) include parcel size, surface area of open water for pools and spas, the
total number of occupants living at the residence, and total landscapeable area.
9. A similar approach was used to identify some of the factors associated with variability in
irrigation application to monitored xeric areas. These are:
i. The total canopy coverage within the xeric area (positive correlation).
ii. The average per-station flow rate of the installed irrigation system serving the
xeric area (positive correlation).
iii. The size of the xeric area (positive correlation).
iv. The property value of the residence (positive correlation).
v. Parcel size (inverse correlation).
vi. Whether or not the irrigation system was exclusively a drip irrigation system (i.e.,
not composed of microsprays, bubblers, other higher flow emitters, or
combinations of emitters; inverse correlation).
vii. Whether or not the resident responsible for managing irrigation at the home is
knowledgeable about enforcement of local provisions prohibiting outdoor water
waste (inverse correlation).
10. Tracking of the costs residents incurred when converting their landscapes from turf to
xeric landscape revealed that at the time of the study, the average conversion cost was
$1.55 per square foot across all of the conversion projects for which data was available.
The average cost for those who did the work themselves was $1.37 per square foot, and
for those employing a contractor, it was $1.93 per square foot. All of these costs are
probably higher today due to inflation and a strong market for conversion projects.
62
11. In comparing those with 60% or more of their landscape as xeric landscaping and those
whose landscape was 60% or more turf, it was found that those with the majority as
xeriscape condition enjoyed a 2.2 hrs-per-month reduction in landscape maintenance and
an additional $206 per annum savings in direct maintenance expenditures as well. This
represented a savings of about a third in total landscape labor and maintenance
expenditures, respectively.
12. A model of two identical homes, one near the average for consumption (technically in the
fifth decile for consumption), the other the same, but having completed an average-size
conversion, revealed the following:
i. The annual water bill savings yielded by landscape conversion projects can be
large. For the Las Vegas Valley Water District customer modeled, the annual
financial savings was $239.92 (figure includes all applicable surcharges). This
equates to a savings of nearly $0.15 per square foot.
ii. This is a large savings of 54% in total annual charges for water consumption. This
level of savings is elevated over what might have been initially anticipated due to
an aggressive tiered water rate structure. The effective average fifth-decile annual
water charges with all surcharges added would be $2.13/kgal for the typical
traditional home and $1.85/kgal for the one having completed the average-size
conversion.
iii. The savings vary by season as expected by the findings associated with the
submeter data. Whereas the bill payer of the home having done the conversion
saved 25% ($5.68) in charges for December vs. the typical homeowner, the same
individual would realize an enormous savings of 70% ($40.84) for July. One of
the great benefits of xeriscape is that it drastically mediates “peaking” in summer,
making summer bills much more affordable for households, especially since power
bills also peak in summer.
13. A model was also created to explore payback time and the appropriateness of the financial
incentive. This revealed that payback time varies in part on whether or not homeowners
do the work themselves or enlist the assistance of a contractor and whether or not savings
in maintenance labor is actually realized. Modeling proceeded such that different
combinations of these scenarios were explored. The results suggest that in most situations
the current SNWA incentive is sufficient to help facilitate conversions such that there is an
acceptable time to return on investment (ROI) for the homeowner. In order of increasing
time to ROI from the point of conversion, with a dollar-per-square foot incentive from
SNWA, these are as follows:
• Situations where only the material costs of the conversion are valued and where
savings in both maintenance goods and labor can be realized (average payback time of
one to two years).
• Situations where both the material costs of the conversion and the labor cost of the
conversion are valued and where savings in both maintenance goods and labor can be
realized (average payback time of two to three years).
63
• Situations where only the material costs of the conversion are valued and where only
savings in maintenance goods (not labor) can be realized (average payback time of
three to four years).
• Situations where both the material costs of the conversion and the labor cost of the
conversion are valued and where only savings in maintenance goods (not labor) can be
realized (average payback time of five to six years).
14. An economic analysis of the cost-efficiency of SNWA’s Water Smart Landscapes
Program suggests that in theory the program is cost-efficient and could be bringing in the
equivalent of $1.58 for each $1.00 spent on rebate incentives (a 37% positive return) by
way of effectively freeing up local water resources for immediate use. When the
opportunity cost is included in the calculation, it is determined that the theoretical
maximum incentive SNWA should be currently willing to pay in 2004 dollars is $1.55 per
square foot (the actual maximum is less due to program administration costs). In practice,
this means it is probably not cost-effective to raise the incentive further at this time as the
level necessary to yield a 3-yr ROI for those not yet facilitated to convert (i.e., the final
bulleted scenario in Conclusion 13) equates to $2.23, an incentive level far in excess of the
theoretical top-out point for an incentive provided by SNWA. Furthermore, in the
majority of modeled scenarios, the dollar per-square-foot is sufficient incentive for
homeowners to justify the landscape conversion project.
64
References
1. Bent, 1992. East Bay Municipal Utility District. CA (data recapitulated from citation in
Reference 4).
2. Testa, A. and Newton, A., 1993. An Evaluation of a Landscape Rebate Program. AWWA
Conserv’93 Proceedings, December. 1763 – 1775. Mesa, AZ.
3. Nelson, J., 1994. Water Saved By Single Family Xeriscapes. 1994 AWWA Annual
Conference Proceedings, June. 335-347. North Marin Water District, Novato, CA
4. Gregg, T. et al., 1994. Xeriscaping: Promises and Pitfalls. City of Austin. Austin, TX.
5. Colorado River Compact of 1922. United States of America. H. Doc 605, 67th Cong.,
4th Sess.
6. Boulder Canyon Project Act. 1928. United States of America. H. Doc 642, 70th Cong.,
2nd Sess.
7. Cooperative Agreement No. 5-FC-30-00440, 1995. Cooperative Agreement Between
Bureau of Reclamation and Southern Nevada Water Authority for Research Investigation
of Implementation of Xeriscape.
8. Mayer, P. and DeOreo, W. et al., 1999. Residential End Uses of Water. AWWA
Research Foundation. U.S.A.
9. DeOreo, W., Mayer, P., Rosales, J., 2000. Xeriscape Conversion for Urban Water
Conservation. National Irrigation Symposium Proceedings, November. Phoenix, NM.
10. Sovocool, K., Hessling, M., and Rosales, J., 2000. Residential Xeriscape: Answering
Some Common Questions on Water and Landscape Maintenance Savings by Field
Research. New Mexico Xeriscape Conference 2000 Proceedings, October. Albuquerque,
NM.
11. Sovocool, K. and Rosales, J., 2001. A Five-year Investigation into the Potential Water
and Monetary Savings of Residential Xeriscape in the Mojave Desert. 2001 AWWA
Annual Conference Proceedings, June. Washington D.C. Southern Nevada Water
Authority, NV.
12. Sovocool, K., 2002. Subtle Mediators of Water Savings in Xeriscape Conversion
Projects. AWWA 2002 Water Sources Conference, January. Las Vegas, NV. Southern
Nevada Water Authority, NV.
13. Hessling, M., 2001. Turf Landscapes versus Xeriscapes: Analysis of residential
landscapes in the Las Vegas Valley, Nevada. Master’s Project. Duke University, NC.
65
14. Mayer, P., Towler, E., and DeOreo, W., et al., 2004. National Multiple Family
Submetering and Allocation Billing Study. Aquacraft, Inc., and East Bay Municipal
Utility District, CA.
15. Rosales, J., Weiss, C., and DeOreo, W., 2002. The Impacts of Submetering on Water
Usage at Two Mobile Home Communities in Las Vegas, Nevada. AWWA 2002 Water
Sources Conference, January. Las Vegas, NV. Southern Nevada Water Authority, NV.
16. Predicting and Estimating Landscape Water Use. The Irrigation Association.
April 2001.
17. A Guide to Estimating Irrigation Water Needs of Landscape Plantings in California –
The Landscape Coefficient Method and WUCOLS III. University of California
Cooperative Extension and California Department of Water Resources. August 2000.
18. Kozlowski, T., Kraper, P., and Pallardy, S., 1990. The Physiological Ecology of Woody
Plants. Academic Press, Inc., San Diego.
19. Gregg, T. et al., 1999. Xeriscaping: Sowing the Seeds for Reducing Water Consumption.
City of Austin. Austin, TX.
20. SNWA 2003 Water Resources Plan. Southern Nevada Water Authority. 2003.
21. SNWA Drought Plan. Supplement to SNWA 2003 Water Resources Plan. Southern
Nevada Water Authority. 2003.
22. Agreement Between Ruffin Gaming, LLC, Ruffin Vegas, LLC and the District for the
Purchase of 300 Acre-feet of Certificated and Permitted Groundwater Rights for An
Amount Not To Exceed $2,850,000. Las Vegas Valley Water District Board of Directors
Agenda Item #8. Las Vegas Valley Water District. April 2003.
66
Appendices
67
APPENDIX 1: ATTACHMENT A (SCOPE OF WORK) FOR BOR
AGREEMENT 5-FC-30-00440 AS REVISED 11/19/98
73
APPENDIX 2: MULTIVARIATE MODEL DETAILS
Note: Detailed definitions of variables and units for with each variable for both of the below
models appear in the corresponding sections within Sources of Significant Variability in Single-
Family Residential Consumption.
TABLE 19: Multivariate Regression Model of Annual Single-Family Residential
Consumption
Regression Summary
Dependent Variable: MAINMETE (i.e., annual consumption registered through mainmeter)
R2=0.80889235; Adjusted R2=0.80046113
F(9,204) = 95.940; p<0.0001
Std. Error of Estimate=76890
Variable Beta Std. Error
of Beta
B Std. Error
of B
t(204) p - level
Intercept -90852.6 25413.77 -3.57494 0.000437
POOL 0.060698 0.035627 51.3 30.13 1.70371 0.089959
TOTALTUR 0.622464 0.041887 59.1 3.98 14.86045 0.000000
TOTALLAN -0.145252 0.102765 -5.5 3.90 -1.41344 0.159051
APPROXINC 0.073217 0.033839 0.3 0.14 2.16370 0.031649
FESCUE 0.068672 0.032854 25756 12322.71 2.09020 0.037839
TOTVAL 0.281661 0.051686 2.1 0.39 5.44950 0.000000
PARCELSI 0.214206 0.119536 5.9 3.28 1.79197 0.074620
NLTHOMEA 0.117091 0.043809 1600.6 598.85 2.67274 0.008132
TOTALOCC 0.52416 0.032356 8860.4 5469.42 1.61999 0.106780
TABLE 20: Multivariate Regression Model of Annual Xeric Study Area Consumption
Regression Summary
Dependent Variable: SUBMETER (i.e., annual consumption registered through submeter)
R2=.64787230; Adjusted R2=.41973852
F(7,178) = 18.394; p<0.0001
Std. Error of Estimate=32272
Variable Beta Std. Error
of Beta
B Std. Error
of B
t(178) p - level
Intercept -7697.6 8973.436 -0.85782 0.392144
STUDYA 0.211132 0.068633 6.4 2.087 3.07623 0.002427
TOTALCAN 0.299352 0.069467 9.2 2.126 4.30934 0.000027
DONTKNOW 0.122082 0.57381 10922.2 5133.663 2.12756 0.034750
TOTVAL 0.213746 0.072592 0.4 0.137 2.94447 0.003667
PARCELSI -0.211758 0.076239 -1.5 0.524 -2.77756 0.006064
AVGFLOWR 0.265679 0.064116 3637.4 877.802 4.14372 0.000053
DRIP -0.133730 0.058997 -13615 6006.406 -2.26674 0.024609
74
APPENDIX 3: RAW DATA
Raw data for possible further analysis is included in the file “BORdata.mdb.” A copy of this
Microsoft Access database file is being included on disk with submission of this report to BOR.
Below is the data description and dictionary for the file (this is also saved on disk).
Xeriscape Conversion Study Data Description
1. tblCustomerList – 716 Records, basic customer information.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Program – Indicates if the property is a xeriscape or turf study site
i. Text – 50
ii. XS = Xeriscape Study, TS = Turf Study
c. FirstName – Property occupant’s first name
i. Text – 50
d. LastName – Property occupant’s last name
i. Text – 50
e. Address – Address of property
i. Text – 50
f. City
i. Text – 50
g. Zip – Postal code
i. Text – 5
h. HomePhone
i. Text – 50
i. WorkPhone
i. Text – 50
j. Comments – Optional comment field
i. Memo
k. OwnerChange – Indicates if there has been a change in the ownership of the
property.
i. Boolean
l. FollowupMonth – Number of the month the property has been assigned to
schedule an annual follow-up site visit.
i. Text – 2
m. AccountNum – LVVWD / SNWA account number assigned to the property
i. Number – Long Integer
n. ServiceArea – Indicates if the customer receives service from LVVWD or one of
the other entities.
i. Text – 50
ii. S = LVVWD Service, O = Outside Entity.
75
o. Agreement – Date the customer signed the agreement to become a participant in
the study.
i. Date/Time
p. FinalReview – Date final inspection site visit was conducted after the installation
of the submeter.
i. Date/Time
q. Status – File quality status indication.
i. Text – 50
r. FileQuality – Quality rating of file information
i. Text – 50
2. tblAllSubmeterData – 2667 Records, customer’s submetered consumption data.
a. nltClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. nitYear
i. Number – Integer
ii. Primary Key
c. txtEntity – Indicates which water provider services the customer
i. Text – 5
d. txtProgram – Indicates if the property is a xeriscape or turf study site
i. Text – 2
ii. XS = Xeriscape Study, TS = Turf Study
e. nstJan – January submeter consumption in gallons
i. Number – Single Precision
f. nstFeb – February submeter consumption in gallons
i. Number – Single Precision
g. nstMar – March submeter consumption in gallons
i. Number – Single Precision
h. nstApr – April submeter consumption in gallons
i. Number – Single Precision
i. nstMay – May submeter consumption in gallons
i. Number – Single Precision
j. nstJun – June submeter consumption in gallons
i. Number – Single Precision
k. nstJul – July submeter consumption in gallons
i. Number – Single Precision
l. nstAug – August submeter consumption in gallons
i. Number – Single Precision
m. nstSep – September submeter consumption in gallons
i. Number – Single Precision
n. nstOct – October submeter consumption in gallons
i. Number – Single Precision
o. nstNov – November submeter consumption in gallons
i. Number – Single Precision
76
p. nstDec – December submeter consumption in gallons
i. Number – Single Precision
q. nstTotal – Total yearly submeter consumption in gallons
i. Number – Single Precision
3. tblAOX2 – 702 Records, parcel information from Assessor’s database
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. PLDECKSQF – Pool decking square footage
i. Number – Single Precision
c. STORAGESQF – Storage area square footage
i. Number – Single Precision
d. PAVE1SQF – Paved area one square footage
i. Number – Single Precision
e. PAVE2SQF – Paved area two square footage
i. Number – Single Precision
f. PATIO1SQF – Patio one square footage.
i. Number – Single Precision
g. PATIO2SQF – Patio two square footage
i. Number – Single Precision
h. PATIO3SQF – Patio three square footage
i. Number – Single Precision
i. GARAGE1SQF – Garage area 1 square footage
i. Number – Single Precision
j. GARAGE2SQF – Garage area 2 square footage
i. Number – Single Precision
k. CARPORTSQF – Carport area square footage
i. Number – Single Precision
l. FIRSTFLSQF – First floor footprint square footage
i. Number – Single Precision
m. TOTALHS – Total of all hardscape areas
i. Number – Single Precision
n. PARCEL – Assessor’s parcel number
i. Text – 11
4. tblETDatawithCustomerIDs – 716 Records, total monthly and annual
evapotranspiration rates for 2001 by month correlated with SNWA client identification
numbers.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. ETType
i. Text - 50
c. JanET
i. Number – Single Precision
77
d. FebET
i. Number – Single Precision
e. MarET
i. Number – Single Precision
f. AprET
i. Number – Single Precision
g. MayET
i. Number – Single Precision
h. JunET
i. Number – Single Precision
i. JulET
i. Number – Single Precision
j. AugET
i. Number – Single Precision
k. SepET
i. Number – Single Precision
l. OctET
i. Number – Single Precision
m. NovET
i. Number – Single Precision
n. DecET
i. Number – Single Precision
o. TotalET
i. Number – Single Precision
5. tblETDatawithCustomerIDsAvg – 716 Records, average monthly and annual
evapotranspiration rates for 2001 by month correlated with SNWA client identification
numbers.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. ETType
i. Text – 50
c. JanAvgET
i. Number – Single Precision
d. FebAvgET
i. Number – Single Precision
e. MarAvgET
i. Number – Single Precision
f. AprAvgET
i. Number – Single Precision
g. MayAvgET
i. Number – Single Precision
h. JunAvgET
i. Number – Single Precision
78
i. JulAvgET
i. Number – Single Precision
j. AugAvgET
i. Number – Single Precision
k. SepAvgET
i. Number – Single Precision
l. OctAvgET
i. Number – Single Precision
m. NovAvgET
i. Number – Single Precision
n. DecAvgET
i. Number – Single Precision
o. TotalAvgET
i. Number – Single Precision
6. tblInstalledCanopy – 447 Records, total of square feet of plant coverage of xeriscape
participants upon installation of the landscape.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. InstCanopyArea – Installed plant canopy square feet.
i. Number – Single Precision
7. tblParcelInfo – 702 Records, Information from Clark County Assessor’s office database
extracted November 2002.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. ParcelNum – Assessor’s office parcel number
i. Text – 11
c. ParcelSize – Size of parcel in square feet
i. Number – Single Precision
d. CONSTYR – Construction year
i. Number – Integer
SALEPRICE – Last Sales price
ii. Number – Long Integer
e. LYTOTAL – Last years assessed value land and improvement
i. Number – Long Integer
f. SALEDATE – Last sales date (Year)
i. Text - 6
g. nltHomeAge – Age of home calculated by construction year from the year 2001.
i. Number – Long Integer
79
8. tblResults – 603 Records, collection of landscape areas, yearly water consumption data,
other site, and customer information
a. nltClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Program – (TS = Turf Study Participant, XS = Xeriscape Study)
i. Text – 50
c. Converted – Area converted if XS participant
i. Number – Single Precision
d. Pool – Square footage of pool surface if present
i. Number – Single Precision
e. GardenMon – Square footage of garden area where the irrigation is monitored by
the submeter
i. Number – Single Precision
f. GardenUnmon – Square footage of garden area where the irrigation is not
monitored by the submeter
i. Number – Single Precision
g. Other – Square footage of other undeveloped property area. No irrigation, plants,
or hardscape present.
i. Number – Single Precision
h. Study – Total xeriscape area where irrigation is monitored by the submeter.
Applies to XS participant only.
i. Number – Single Precision
i. TurfMon – Square footage of turf grass where irrigation is monitored by the
submeter.
i. Number – Single Precision
j. TurfUnmon – Square footage of turf area where the irrigation is not monitored by
the submeter
i. Number – Single Precision
k. XeriMon – Square footage of xeriscape where irrigation is monitored by the
submeter. (Applies to Turf Study Group)
i. Number – Single Precision
l. XeriUnmon – Square footage of xeriscape area where the irrigation is not
monitored by the submeter.
i. Number – Single Precision
m. TotalLandscape – Total of all landscapable area on the property.
i. Number – Single Precision
n. TotalEvaporative – Total of all landscapable area with pool area added.
i. Number – Single Precision
o. dtt2001SR – Date of final annual visit conducted in 2001.
i. Date/Time
p. AgeOfXeriscape – Age of xeriscape in days calculated by the difference in days
between the post submeter installation inspection and the final 2001 follow-up site
visit.
i. Number – Long Integer
80
q. TotalXeriArea – Total of all xeriscape areas, monitored and unmonitored.
i. Number – Single Precision
r. Status – File quality status indication.
i. Text - 50
s. TotalCanopy – Total of all plant canopy areas as of the 2001 annual site visit.
i. Number – Single Precision
t. nitYear
i. Number – Integer
u. txtEntity – Water agency that services the customer.
i. Text - 5
v. Submeter2001 – Total gallons used in the year 2001 through the submeter
i. Number – Single Precision
w. Mainmeter2001 – Total gallons used in the year 2001 through the main meter
i. Number – Single Precision
x. pctGarden – Percent of total landscape area in garden
i. Number – Single Precision
y. pctXeri – Percent of total landscape in xeriscape
i. Number – Single Precision
z. pctTurf – Percent of total landscape area in turf
i. Number – Single Precision
aa. pctOther – Percent of total landscape in other non-landscaped area
i. Number – Single Precision
bb. pctPool – Percent of total landscape area in pool
i. Number – Single Precision
cc. pctXeriRank – Xeriscape study participants were divided into ten percent ranges
based upon percentage of landscape in xeriscape and given a ranking.
i. Number – Single Precision
dd. XeriDensity – Percent of plant coverage per square foot of xeriscape.
i. Number – Single Precision
ee. TurfType – Type of turf (Bermuda, Fescue, etc.) on property if present.
i. Text – 50
ff. BarrierType – Type of weed barrier present if Xeriscape study participant.
i. Text – 50
9. tblSurveyInfoOfInterest – 603 Records, Responses to survey questions. Each possible
response is listed as a separate field. The responses are grouped where appropriate.
a. CLIENTID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. SurveyAnswered – “Yes” or “No” Indicates if the customer answered any of the
questions on the survey.
i. Text – 3
c. CLOCKADJ – How many times per year the irrigation clock was adjusted
i. Number – Byte
81
d. INCBILL – How much of an increase in the monthly bill would produce
conservation?
i. Number – Integer
e. RESPAGE – Respondent’s age
i. Number – Byte
f. Respondent’s gender
i. MALE
1. Number – Byte (1 = Yes, 0 = No)
ii. FEMALE
1. Number – Byte (1 = Yes, 0 = No)
g. Respondent’s marital status
i. MARRIED
1. Number – Byte (1 = Yes, 0 = No)
ii. SINGLE
1. Number – Byte (1 = Yes, 0 = No)
iii. WIDOWED
1. Number – Byte (1 = Yes, 0 = No)
h. RETIRED – Indicates if respondent is retired or not
i. Number – Byte (1 = Yes, 0 = No)
i. NATIVE – Native to southern Nevada?
i. Number – Byte (1 = Yes, 0 = No)
j. AGE65PLS – Number of residents at the property age 65 and older
i. Number – Byte
k. APROXINC – Median of a range of household income
i. Number – Long Integer
l. Respondent’s opinion on Water Waste enforcement
i. DONTKNOW
1. Number – Byte (1 = Yes, 0 = No)
ii. GOOD
1. Number – Byte (1 = Yes, 0 = No)
iii. LAX
1. Number – Byte (1 = Yes, 0 = No)
iv. STRICT
1. Number – Byte (1 = Yes, 0 = No)
m. Highest Education Level
i. ASSOCDEG – Associate’s degree
1. Number – Byte (1 = Yes, 0 = No)
ii. BACHDEG – Bachelor’s degree
1. Number – Byte (1 = Yes, 0 = No)
iii. GRADDEG – Graduate degree
1. Number – Byte (1 = Yes, 0 = No)
iv. HSDEG – High school degree
1. Number – Byte (1 = Yes, 0 = No)
82
v. SOMECOLL – Some College
1. Number – Byte (1 = Yes, 0 = No)
vi. SOMEGRAD – Some graduate education
1. Number – Byte (1 = Yes, 0 = No)
vii. TECHTRAD – Technical or trade school
1. Number – Byte (1 = Yes, 0 = No)
viii. ADTECTRN – Advanced technical training
1. Number – Byte (1 = Yes, 0 = No)
n. Type of Grass at residence
i. BERMUDA
1. Number – Byte (1 = Yes, 0 = No)
ii. FESCUE
1. Number – Byte (1 = Yes, 0 = No)
iii. BUFFALO
1. Number – Byte (1 = Yes, 0 = No)
iv. BFMIX – Bermuda / Fescue Mix
1. Number – Byte (1 = Yes, 0 = No)
v. UNKNOWN
1. Number – Byte (1 = Yes, 0 = No)
vi. NONE
1. Number – Byte (1 = Yes, 0 = No)
10. tblSurveryTotBath – 623 Records, total number of bathrooms on the property
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Bathrooms
i. Number – Single Precision
11. tblSurveyTotOccupants- 341 Records, total number of occupants in the household at the
time of the survey.
a. nltClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. TotalOccupants
i. Number – Integer
12. tblIrrigationData – 355 Records, Irrigation system components for each property were
assessed, and each property assigned to one of the following categories.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. AvgFlowRate – Average flow rate of all stations
i. Number – Single Precision
c. BubblerDrip – Irrigation system is composed of bubbler and drip systems
i. Number – Integer (1 = Yes, 0 = No)
83
d. BubblerDripSpray – Irrigation system is composed of bubbler, drip, and spray
systems
i. Number – Integer (1 = Yes, 0 = No)
e. Bubblers – Irrigation system is composed of bubblers
i. Number – Integer (1 = Yes, 0 = No)
f. BubblerSpray – Irrigation system is composed of bubbler and spray systems
i. Number – Integer (1 = Yes, 0 = No)
g. Drip – Irrigation system is composed of drip systems
i. Number – Integer (1 = Yes, 0 = No)
h. DripOff – Irrigation system is composed of drip systems with one or more other
irrigation zones turned off
i. Number – Integer (1 = Yes, 0 = No)
i. DripMicro – Irrigation system is composed of drip and micro-spray systems
i. Number – Integer (1 = Yes, 0 = No)
j. DripPopup – Irrigation system is composed of drip and popup spray systems
i. Number – Integer (1 = Yes, 0 = No)
k. DripSpray – Irrigation system is composed of drip and spray systems
i. Number – Integer (1 = Yes, 0 = No)
l. Hose – Irrigation is done with a hose
i. Number – Integer (1 = Yes, 0 = No)
m. Microspray – Irrigation system is composed of micro-spray systems
i. Number – Integer (1 = Yes, 0 = No)
n. Sprays – Irrigation system is composed of spray systems
i. Number – Integer (1 = Yes, 0 = No)
o. BubblerDripPopup – Irrigation system is composed of bubbler, drip, and popup
spray systems
i. Number – Integer (1 = Yes, 0 = No)
p. DripMicroPopup – Irrigation system is composed of drip micro-spray and popup
spray systems
i. Number – Integer (1 = Yes, 0 = No)
q. DripPopupSpray – Irrigation system is composed of drip, popup spray, and spray
systems
i. Number – Integer (1 = Yes, 0 = No)
r. DripPopupRotor – Irrigation system is composed of drip, popup spray, and rotor
systems
i. Number – Integer (1 = Yes, 0 = No)
s. DripLaser – Irrigation system is composed of drip and laser tube systems
i. Number – Integer (1 = Yes, 0 = No)
t. DripSoaker – Irrigation system is composed of drip and soaker hose systems
i. Number – Integer (1 = Yes, 0 = No)
u. DripTurfBubbler – Irrigation system is composed of drip and turf bubbler systems
i. Number – Integer (1 = Yes, 0 = No)
v. DripFountain – Irrigation system is composed of drip systems, and a fountain refill
is controlled with the irrigation clock
i. Number – Integer (1 = Yes, 0 = No)
84
13. tblMulches – 715 Records, mulch and weed barrier information
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. txtMulch – Typical type of mulch
i. Text - 18
c. txtMulchSize – Typical size of mulch
i. Text - 50
d. txtMulchColor – Typical color of mulch
i. Text - 6
e. nstMulchDepth – Depth of mulch in inches
i. Number – Single Precision
f. yntWeeds – Indicates if excessive weeds are present
i. Boolean
g. yntSlope – Is a steep slope present?
i. Boolean
h. yntTraffic – Is there heavy traffic in landscape?
i. Boolean
i. yntAlkali – Indicates if excessive alkali deposits present at surface.
i. Boolean
j. txtBarrierType – Type of weed barrier
i. Text – 20
k. txtBarrierColor – Color of weed barrier
i. Text – 6
l. yntBarrierShowing – Is the barrier showing at surface?
i. Boolean
m. txtWear – Extent of wear
i. Text – 6
n. txtLocationType – Wear location type
i. Text – 16
14. tblMainmeterConsumption – 4318 Records, Gallons used per customer per month as
measured by the property’s main service meter.
a. nltClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. nitYear
i. Number – Integer
ii. Primary Key
c. txtEntity – Indicates which water provider services the customer
i. Text – 5
d. nstJan – January consumption in gallons
i. Number – Single Precision
e. nstFeb – February consumption in gallons
i. Number – Single Precision
85
f. nstMar – March consumption in gallons
i. Number – Single Precision
g. nstApr – April consumption in gallons
i. Number – Single Precision
h. nstMay – May consumption in gallons
i. Number – Single Precision
i. nstJun – June consumption in gallons
i. Number – Single Precision
j. nstJul – July consumption in gallons
i. Number – Single Precision
k. nstAug – August consumption in gallons
i. Number – Single Precision
l. nstSep – September consumption in gallons
i. Number – Single Precision
m. nstOct – October consumption in gallons
i. Number – Single Precision
n. nstNov – November consumption in gallons
i. Number – Single Precision
o. nstDec – December consumption in gallons
i. Number – Single Precision
p. nstTotal – Total annual consumption in gallons
i. Number – Single Precision
15. tbl2001PropAreasOK4 – 673 Records, Property area information as recorded for the
year 2001.
a. nltClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Converted – Area converted from turf to xeriscape. Refers to “XS” Participants
only.
i. Number – Single Precision
c. Pool – Pool area if applicable
i. Number – Single Precision
d. GardenMon – Garden area where irrigation is being monitored by the submeter
i. Number – Single Precision
e. GardenUnmon – Garden area where irrigation is unmonitored by the submeter
i. Number – Single Precision
f. Other – Square footage of other undeveloped property area. No irrigation, plants
or hardscape present.
i. Number – Single Precision
g. Study – Total xeriscape area where irrigation is monitored by the submeter.
Applies to XS participant only.
i. Number – Single Precision
h. TurfMon – Square footage of turf grass where irrigation is monitored by the
submeter.
i. Number – Single Precision
86
i. TurfUnmon – Square footage of turf area where the irrigation is not monitored by
the submeter
i. Number – Single Precision
j. XeriMon – Square footage of xeriscape where irrigation is monitored by the
submeter. (Applies to xeriscape study Group)
i. Number – Single Precision
k. XeriUnmon – Square footage of xeriscape area where the irrigation is not
monitored by the submeter.
i. Number – Single Precision
l. TotalEvaporative – Total of all landscape areas plus pool area.
i. Number – Single Precision
m. TotalLandscape – Total of all landscape areas.
i. Number – Single Precision
n. dtt2001SR – Date of 2001 follow-up site visit
i. Date / Time
o. AgeOfXeriscape – Age of xeriscape in days calculated by the difference between
the post submeter installation inspection and the final 2001 follow-up site visit.
i. Number – Long Integer
p. TotalXeriArea – Total of all xeriscaped areas
i. Number – Single Precision
q. TotalGarden – Total of all garden areas
i. Number – Single Precision
r. TotalTurf – Total of all Turf areas
i. Number – Single Precision
s. PctGarden – Percent of total landscape area in garden
i. Number – Single Precision
t. PctXeri – Percent of total landscape in xeriscape
i. Number – Single Precision
u. PctTurf – Percent of total landscape area in turf
i. Number – Single Precision
v. PctOther – Percent of total landscape in other non-landscaped area
i. Number – Single Precision
w. PctPool – Percent of total landscape in pool
i. Number – Single Precision
x. PctXeriRank – Xeriscape study participants were divided into ten percent ranges
based upon percentage of landscape in xeriscape and given a ranking.
i. Number – Long Integer
16. tblTurfOnlySubMonthly – 107 Records, monthly submeter consumption data and per
square foot usage for turf study group of participants. Note – this usage is limited to those
TS participants where ONLY turf was irrigated with submeter-monitored usage.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Year
i. Number – Integer
87
c. Entity – Water purveyor that serves the customer
i. Text – 5
d. FileQuality – Quality rating of file information
i. Text – 10
e. Status – Customer status
i. Text – 7
f. TurfMon – Square feet of grass where irrigation is monitored by the submeter
i. Number – Single
g. JanCons – January submeter consumption in gallons
i. Number – Single
h. FebCons – February submeter consumption in gallons
i. Number – Single
i. MarCons – March submeter consumption in gallons
i. Number – Single
j. AprCons – April submeter consumption in gallons
i. Number – Single
k. MayCons – May submeter consumption in gallons
i. Number – Single
l. JunCons – June submeter consumption in gallons
i. Number – Single
m. JulCons – July submeter consumption in gallons
i. Number – Single
n. AugCons – August submeter consumption in gallons
i. Number – Single
o. SepCons – September submeter consumption in gallons
i. Number – Single
p. OctCons – October submeter consumption in gallons
i. Number – Single
q. NovCons – November submeter consumption in gallons
i. Number – Single
r. DecCons – December submeter consumption in gallons
i. Number – Single
s. JanGalSF – Gallons used per square foot of turf in January
i. Number – Single
t. FebGalSF – Gallons used per square foot of turf in February
i. Number – Single
u. MarGalSF – Gallons used per square foot of turf in March
i. Number – Single
v. AprGalSF – Gallons used per square foot of turf in April
i. Number – Single
w. MayGalSF – Gallons used per square foot of turf in May
i. Number – Single
x. JunGalSF – Gallons used per square foot of turf in June
i. Number – Single
y. JulGalSF – Gallons used per square foot of turf in July
i. Number – Single
88
z. AugGalSF – Gallons used per square foot of turf in August
i. Number – Single
aa. SepGalSF – Gallons used per square foot of turf in September
i. Number – Single
bb. OctGalSF – Gallons used per square foot of turf in October
i. Number – Single
cc. NovGalSF – Gallons used per square foot of turf in November
i. Number – Single
dd. DecGalSF – Gallons used per square foot of turf in December
i. Number – Single
17. tblTurfOnlySubYearly – 107 Records, yearly submeter consumption data and per
square foot usage for turf study group of participants. Note – this usage is limited to those
TS participants where ONLY turf was irrigated with submeter-monitored usage.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Year
i. Number – Integer
ii. Primary Key
c. Entity – Water purveyor that serves the customer
i. Text – 5
d. TurfMon – Square feet of grass where irrigation is monitored by the submeter
i. Number – Single
e. GalSqFt – Gallons used per square foot of turf per year
i. Number – Single
f. YearlyCons – Total submetered consumption for the year.
i. Number – Single
g. FileQuality – Quality rating of file information
i. Text - 8
h. Status – Customer status
i. Text – 7
18. tblXeriOnlySubMonthly – 1550 Records, monthly submeter consumption data and per
square foot usage for xeriscape study group of participants. Note – this usage is limited to
those XS participants where ONLY xeriscape was irrigated with submeter-monitored
usage.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Year
i. Number – Integer
ii. Primary Key
c. Entity – Water purveyor that serves the customer
i. Text – 5
89
d. ConvNew – Indicates if the property’s xeriscape was a new installation or a
conversion of grass to xeriscape.
i. Text – 4
e. Status – Customer status
i. Text – 7
f. FileQuality – Quality rating of file information
i. Text – 10
g. XeriMon – Square feet of xeriscape where irrigation is monitored by the submeter
i. Number – Single Precision
h. JanCons – January submeter consumption in gallons
i. Number – Single Precision
i. FebCons – February submeter consumption in gallons
i. Number – Single Precision
j. MarCons – March submeter consumption in gallons
i. Number – Single Precision
k. AprCons – April submeter consumption in gallons
i. Number – Single Precision
l. MayCons – May submeter consumption in gallons
i. Number – Single Precision
m. JunCons – June submeter consumption in gallons
i. Number – Single Precision
n. SepCons – September submeter consumption in gallons
i. Number – Single Precision
o. OctCons – October submeter consumption in gallons
i. Number – Single Precision
p. NovCons – November submeter consumption in gallons
i. Number – Single Precision
q. DecCons – December submeter consumption in gallons
i. Number – Single Precision
r. JanGalSF – Gallons used per square foot of xeriscape in January
i. Number – Single
s. FebGalSF – Gallons used per square foot of xeriscape in February
i. Number – Single
t. MarGalSF – Gallons used per square foot of xeriscape in March
i. Number – Single
u. AprGalSF – Gallons used per square foot of xeriscape in April
i. Number – Single
v. MayGalSF – Gallons used per square foot of xeriscape in May
i. Number – Single
w. JunGalSF – Gallons used per square foot of xeriscape in June
i. Number – Single
x. JulGalSF – Gallons used per square foot of xeriscape in July
i. Number – Single
y. AugGalSF – Gallons used per square foot of xeriscape in August
i. Number – Single
90
z. SepGalSF – Gallons used per square foot of xeriscape in September
i. Number – Single
aa. OctGalSF – Gallons used per square foot of xeriscape in October
i. Number – Single
bb. NovGalSF – Gallons used per square foot of xeriscape in November
i. Number – Single
cc. DecGalSF – Gallons used per square foot of xeriscape in December
i. Number – Single
19. tblXeriOnlySubYearly – 1550 Records, yearly submeter consumption data and per
square foot usage for xeriscape study group of participants. Note – this usage is limited to
those XS participants where ONLY xeriscape was irrigated with submeter-monitored
usage.
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. Year
i. Number – Integer
ii. Primary Key
c. Entity – Water purveyor that serves the customer
i. Text – 5
d. ConvNew – Indicates if the property’s xeriscape was a new installation or a
conversion of grass to xeriscape.
i. Text – 4
e. XeriMon – Square feet of xeriscape where irrigation is monitored by the submeter
i. Number – Single Precision
f. YearlyCons– Total submetered consumption for the year.
i. Number – Single
g. GalSqFt – Gallons used per square foot of monitored xeriscape per year
i. Number – Single
h. FileQuality – Quality rating of file information
i. Text – 10
i. Status – Customer status
i. Text – 7
20. tblPlantList – 538 Records, list of plants used to verify xeriscape participant’s
compliance with minimum canopy standards for program participation and classification
of landscape plants in subsequent follow-up visits.
a. PlantID
i. Number – Long Integer
ii. Primary Key
b. Genus
i. Text - 50
c. Species
i. Text - 50
91
d. Var/Cult – Variety or cultivar of plant
i. Text - 50
e. Common Name
i. Text - 50
f. Width – Expected mature width of the plant
i. Number - Single
g. Height – Expected mature height of the plant
i. Number - Integer
h. Plant Habit – Type of plant (shrub, tree, etc.)
i. Text - 50
i. H20Use – Rated plant water needs.
i. Text – 50
21. tbl2001HomeSales – 45 Records, data provided by SalesTraq. Information related to
home sales in Southern Nevada area in the year 2001 by zip code.
a. Zipcode
i. Text – 5
ii. Primary Key
b. NumberSold – Number of homes sold in zip code
i. Number – Single Precision
c. MedianPrice – Median price of homes sold in zip code
i. Number – Single Precision
d. AvgPrice – Average price of homes sold in zip code.
i. Number – Single Precision
e. AvgPricePerSqFt – Average Price per square foot of homes sold in zip code.
i. Number – Single Precision
f. AvgSize – Average size of homes sold in zip code.
i. Number – Single Precision
g. Volume – Total value of homes sold in zip code
i. Number – Single Precision
h. AvgAge – Average age of homes sold in zip code
i. Number – Single Precision
22. tblMeterInfo – 716 Records
a. ClientID – SNWA Customer identification number
i. Number – Long Integer
ii. Primary Key
b. MeterNum – Serial number stamped on submeter by manufacturer
i. Text – 50
c. Installed – Date submeter was installed by contractor
i. Date/Time
d. Cost – Cost of meter installation
i. Number – Single Precision
e. RetrofitNum – AS/400 account number assigned to submeter
i. Number – Long Integer
92
f. Location – approximate location of submeter on site.
i. Memo
93
APPENDIX 4: INFORMATION ON SINGLE-FAMILY RESIDENTIAL WATER BILL MODEL
A model was used to explore the differences in water consumption charges for a typical fifth
decile in consumption LUC 110 property (single-family home) and one doing an average-size
conversion. The model assumes the properties are in the Las Vegas Valley Water District’s
service area and subject to its regular service rules. A typical 5/8-inch-meter size was assumed
(meter size in large part determines rate per consumption unit). Rates for each tier and the size of
the tier rate block appear below in the screen shot of the actual modeling processes for the model
used in this report. As demonstrated, within a given billing cycle the rate for the first
5,000 gallons is $1.05/kgal, the next 5,000 gallons after the initial 5000 costs $1.75/kgal, the next
10,000 gallons after these first 10,000 gallons is $2.38/kgal and so on (for billing purposes, the
utility rounds to the nearest thousand gallons). In addition to the direct charges for the water,
SNWA purveyor members bills commonly include a service charge, a commodity charge, and a
reliability charge and these are reflected in the model below, so that the outputs are reflective of
actual bills. A 30-day billing cycle was assumed.
In practical terms, the calculation of outputs in the model and the savings is derived by
multiplying the expected average savings per square foot per month that would be yielded by a
conversion (as calculated from Table 18) by the average-size conversion and then subtracting this
from the fith-decile consumption level. This yielded the costs with having done the conversion
(below called “Total Bill). In contrast, the cost without doing the conversion (i.e., “Average
Fifth-Decile bill without reduction”) is shown under the “did conversion” scenario. The
difference between these, highlighted in red, is the anticipated monthly bill savings yielded from
having completed the conversion project.
Water Bill Calculator Screen Shot
94
APPENDIX 5: INFORMATION ON HOMEOWNER PERSPECTIVE MODEL
The model is a dynamic Net Present Value Model that calculates the NPV of the project in future
years. It does this by computing the difference in the yield by converting to xeriscape to the costs
(water and maintenance) incurred by keeping turfgrass over the years.
“Conversion cost” and “awarded incentive” are products of the associated rates and the square
feet converted. These are onetime costs. The “interest rate” is the discount or alternative rate
(i.e., the rate associated with the loss incurred by spending money on the conversion rather than
placing it in an interest-bearing account). The “average yearly rate increase” is the long-time
average increase in water costs. “Yearly maintenance savings” is a product of the “Labor
Savings” and “Direct Maintenance” variables (which are themselves calculated in a manner
similar to “awarded incentive,” however, these savings are yielded each year). “Average total bill
savings for a year” is not automatically calculated, but entered either by use of real data or
modeled bill savings (see Appendix 4). Model Outputs are “NPV” and “Year.” Year 0 is the
year of the conversion.
This model can directly yield the payback time with and without the incentive. By iterative
process one can then develop what the input variables values would need to provide for a positive
NPV at a given year. This is how the values for the third and fifth-year ROIs were developed for
Figure 15. Example inputs and outputs are given below. In this case, at $1.00 per square foot, the
conversion reached a positive NPV between years one and two.
In terms of yielding the actual data in Table 15, the following were used as data sources:
“Square Feet Converted” – This was the average conversion size for SNWA’s Water Smart
Landscapes Program in early 2004.
“Incentive Level” – This was the $1.00 per square foot incentive level for almost all single-family
conversion projects in SNWA’s Water Smart Landscapes Program in early 2004 (also see
Appendix 5).
“Conversion cost” – This was the conversion cost as revealed by survey. This was one of the
variables that were modified to reflect whether or not one did the conversion themselves or
utilized contract assistance. Rates for each of these scenarios were developed based on
compilation of receipts from both types of installations.
“average total bill savings for a year” – This was the yearly savings as provided by a model of the
Las Vegas Valley Water District for a LUC 110 property in the fifth decile (mid-range) of
consumption (see Appendix 4 for details on this model).
“interest rate” – This was the interest rate of a home equity loan in early February 2004.
“average yearly rate increase” – This is the average yearly rate increase for the Las Vegas Valley
Water District over the long term. In practice, the District has often gone significant periods of
95
time without a rate increase and then increased them much more than 3%, but this was the most
practical method of doing the calculation for purposes of creating the model.
“Labor Savings” – This was adapted from Hessling12 (2001). This savings was effectively turned
on or off to see the impacts of the situations when labor savings are and are not realized. See text
for additional information.
“Direct Maintenance” – This rate was derived from the maintenance survey data and is per
Hessling12 (2001).
Examples of Homeowner Perspective Model Inputs and Outputs
NPV Year
($2,070.88) 0
($636.58) 1
$751.63 2
$2,095.24 3
$3,395.67 4
$4,654.31 5
Inputs:Type
Square Feet Converted 1616
Incentive level $1.00
Conversion cost: $1.37
conversion cost: $2,213.92
average total bill savings for a year: $240.00
awarded incentive: $1,616.00
interest rate: 6.32%
average yearly rate increase 3.00%
Labor Savings $0.20
Labor Savings $323.20
Direct Maintenance $0.11
Direct Maintenance $177.76
Yearly maintenance savings $500.96
96
APPENDIX 6: INFORMATION ON SNWA’S WATER SMART LANDSCAPES PROGRAM
Growth of Program:
See Program Application (following)
97