Yardx: Yield and Reliability Demonstrated in Xeriscape - Colorado Water Wise

Yield and Reliability Demonstrated in Xeriscape

Final Report

Metro Water Conservation, Incorporated (MWCI)

December 2004

Cover: Photo courtesy of David Winger, Denver Water, Denver, Colorado.

YARDX
YIELD AND RELIABILITY DEMONSTRATED IN XERISCAPE
Final Report

Prepared for

Bureau of Reclamation

And

Metro Water Conservation, Incorporated (MWCI)
Under Cooperative Agreement 6-FC-60-06670

by

Jonnie G. Medina (Bureau of Reclamation)
and
Julia Gumper (MWCI)

December 2004

Citation information:
Medina, Jonnie G., and Julia Gumper, 2004: YARDX: Yield And Reliability Demonstrated in
Xeriscape: Final Report. Metro Water Conservation, Incorporated, Littleton, CO, 140 p. For
copies of the full report, see www.coloradowaterwise.org or contact the Office of Water
Conservation at Denver Water.

Abstract

Metro Water Conservation, Inc. (MWCI) of Denver, Colorado, in partnership with the Bureau of
Reclamation, conducted a water conservation study known as the Yield And Reliability Demonstrated in
Xeriscape (YARDX) project to estimate the benefits of water-conserving landscaping known as Xeriscape.
Benefits to be assessed were seasonal water savings, landscape installation, and annual maintenance
costs. Seven municipalities from Fort Collins, Colorado, to Colorado Springs, Colorado, participated in
the study. The YARDX project is one of five field projects of Reclamation’s National Xeriscape
Demonstration Program (NXDP) established to study the benefits of installing Xeriscape under differing
climatic and other potentially impacting conditions.

YARDX was conducted from 1997 through 2002. The project included seven field demonstrations, each
with some differing attributes, including Xeriscape application type (retrofits or new starts), application
level (high or moderate water savings designs), yard size, irrigation method, socio-economic level, and
soil type. In seven demonstrations, control groups of traditional high water use turf were established
with similar characteristics to Xeriscapes, except for landscape type. Participants had to install
Xeriscapes, except at one demonstration. All participants had to maintain them with no major revision
during the study period. Xeriscape participants were provided a small rebate to join the study, and were
given education on installing and maintaining Xeriscapes. One demonstration involved comparison of
older, established Xeriscapes, with comparable control landscapes.

Data analysis of historical water use established the need for sample sizes of approximately 30 properties
studied over 4 growing seasons, to have at least a 90-percent chance of detecting a 30-percent change in
water use at the 5-percent significance level. The demonstrations yielded high-quality data that generally
enabled the estimation of water savings and annual maintenance costs.

Xeriscape installation costs ran a modest $0.90 to $1.45 per square foot, with homeowners in the project
contributing a substantial amount of labor. Demonstrations obtained water savings from 18 to more than
50 percent over control samples. Results indicated that relatively consistently, water savings in the
30-percentile range could be obtained for properly designed and maintained Xeriscapes. Annual
maintenance costs ranged from $0.34 to $1.33 per square foot. For cost estimation, homeowner labor was
computed at $18 per hour. Generally, the maintenance cost of the Xeriscapes sampled, compared to the
non-Xeriscaped properties, was found to be less than controls during the plant establishment years, but
somewhat more during the plant maturation years. This suggests that as Xeriscapes age, they gradually
require more maintenance, compared to traditional landscapes.

Xeriscape participants overwhelmingly expressed satisfaction with their landscapes and would freely
recommend this type of landscaping to others. The information gained in the YARDX project should
provide an additional alternative in dealing with water conservation needs in the Colorado Front Range.

Yardx: Yield and Reliability Demonstrated in Xeriscape - Colorado Water Wise

Acknowledgements

The nonprofit Metro Water Conservation, Inc. (MWCI), located in Denver and consisting of members
from several water utilities from the Colorado Front Range, in partnership with the Bureau of
Reclamation, U.S. Department of Interior, funded and conducted the Yield And Reliability Demonstrated
in Xeriscape (YARDX) project. The YARDX project is one of five field projects that have contributed
information to Reclamation’s National Xeriscape Demonstration Program (NXDP).

The authors appreciate the MWCI Board of Directors, Mr. Paul Lander of the City of Boulder,
Ms. Elizabeth Gardener of Denver Water, Ms. Deb Pilon, Willows Water District, and Ms. Sally Dale
(former Board member), Nautilus Resources, for their tireless dedication to the project. For their
contribution of data and support to the YARDX project, we wish to thank Mr. David Winger of the
Denver Water, Ms. Jeanie Sims and Ms. Ann Seymour of the City of Colorado Springs Utilities, Ms.
Laurie D’Audney of the Fort Collins Water Utility, Mr. John Hendrick and Ms. Diane Schorege of
Highlands Ranch Metropolitan Districts, Mr. Webb Jones of the East Larimer County Water District, Ms.
Sue Vest and Mr. Mike DiTullio of the Fort Collins-Loveland Water District, Mr. Chris Koerner and Mr.
Ken Peterson of the City of Arvada Utilities Division, Mr. Walt Pettit of the Wheat Ridge Water District,
and Ms. Ruth Quade and Mr. Phil Carter of the City of Greeley. Many thanks go to Mr. Steffen Meyer of
Reclamation for his tireless assistance to Jon Medina and Julia Gumper in formatting of the report text
and to Ms. Elizabeth Gardener for her thorough review of the draft report. Mr. Lonnie Lewis of
Reclamation provided editing for the report. The authors appreciate the support provided to YARDX by
Ms. Avra Morgan, Dr. David Matthews, Mr. Jim Pierce, Mr. Luis Maez, Ms. Christy Bridges, Ms. Paula
Sunde, Ms. Julie Swanda (formerly of Reclamation), Mrs. Susan Meyer, and Mr. Tom Phillips all of
Reclamation for their support to the YARDX project.

Contents

Abstract...................................................................................................................................... iii
Acknowledgements..................................................................................................................... v
Executive Summary................................................................................................................... xi

Chapter One—Introduction ......................................................................................................... 1-1
General................................................................................................................................. 1-1
Study Location ................................................................................................................. 1-2
Study Goals...................................................................................................................... 1-2
Previous Studies............................................................................................................... 1-3
Why YARDX in the Front Range.................................................................................... 1-3

Chapter Two—Project Design..................................................................................................... 2-1
Introduction.......................................................................................................................... 2-1
Demonstration Characteristics ............................................................................................. 2-2
Study Variables................................................................................................................ 2-2
Sampling Plan .................................................................................................................. 2-3
Sample Size Estimation ................................................................................................... 2-5
Field Data............................................................................................................................. 2-6
Cost Estimation................................................................................................................ 2-6
Xeriscape Installation................................................................................................... 2-6
Maintenance................................................................................................................. 2-6
Annual Surveys................................................................................................................ 2-6
Example of Design........................................................................................................... 2-7

Chapter Three—Promotion.......................................................................................................... 3-1
Introduction.......................................................................................................................... 3-1
Promoting Xeriscape Participation ...................................................................................... 3-2
Xeriscape Seminars.............................................................................................................. 3-4
Restricting YARDX Promotion........................................................................................... 3-4
Project Benefits.................................................................................................................... 3-5
Oberservations ..................................................................................................................... 3-6

Chapter Four—Installation Cost of Xeriscape............................................................................. 4-1
Introduction.......................................................................................................................... 4-1
Cost Data Quality................................................................................................................. 4-1
Cost Estimation.................................................................................................................... 4-2
Average Installation Costs ............................................................................................... 4-3
Installation Labor ............................................................................................................. 4-5
Costs of Plants and Hardscape......................................................................................... 4-6

Chapter Five—Water Use Comparisons...................................................................................... 5-1
Introduction.......................................................................................................................... 5-1
Water Data ........................................................................................................................... 5-2

vii

Water Use Comparisons ...................................................................................................... 5-3
Precipitation ......................................................................................................................... 5-7

Chapter Six—Maintenance Costs ................................................................................................ 6-1
Introduction.......................................................................................................................... 6-1
Data Quality ......................................................................................................................... 6-1
Maintenance Cost Calculations............................................................................................ 6-2
Maintenance Cost Results................................................................................................ 6-2
Maintenance Education........................................................................................................ 6-3

Chapter Seven—Final Survey...................................................................................................... 7-1
Introduction.......................................................................................................................... 7-1
Specific Survey Questions Analyzed................................................................................... 7-4

Chapter Eight—Summary and Conclusions ................................................................................ 8-1
Preface.................................................................................................................................. 8-1
Results.................................................................................................................................. 8-2
Enlisting Participants ....................................................................................................... 8-2
Xeriscape Installation Costs............................................................................................. 8-2
Water Use Results............................................................................................................ 8-2
Maintenance Costs ........................................................................................................... 8-3
Final Survey Results ........................................................................................................ 8-4
Reflections ........................................................................................................................... 8-4
Recommendations................................................................................................................ 8-5
Conclusions.......................................................................................................................... 8-5

References

Glossary

viii

Appendices
(available as a separate file)

Appendix A: Xeriscape Participant Dropouts
Appendix B: Landscape Installation Audit Forms
Appendix C: Landscape Revisit Audit Forms
Appendix D: Inflation Rate Estimate
Appendix E: Landscape Maintenance Form
Appendix F: Final Survey of YARDX Participants
Appendix G: Final Survey Results
Appendix H: Environmental Orientation Query
Appendix I: Sample Newsletter
Appendix J: Before and After Photographs

Tables
Table 2–1: YARDX demonstrations and primary characteristics .............................................. 2-4
Table 3–1: Number of participants who signed up, completed or dropped from the
YARDX project ................................................................................................ 3-2
Table 3–2: Example of Xeriscape signup results from a direct mail campaign ......................... 3-3
Table 5–1: Water use comparisons between Xeriscape and traditional landscape samples....... 5-5

ix

Figures

Figure 1-1: Location of cities participating in the YARDX study.............................................. 1-2
Figure 2-1: Xeriscape design, Colorado Springs, CO................................................................. 2-7
Figure 2-2: Photograph of the Xeriscape 4 years after installation in Colorado Springs ........... 2-8
Figure 4-1: Number of new start (NS) and retrofit (RT) Xeriscapes providing installation
costs. ................................................................................................................. 4-2
Figure 4-2: Average landscape size by demonstration. .............................................................. 4-4
Figure 4-3: Average Xeriscape installation cost per square foot by demonstration. .................. 4-4
Figure 4-4: Installation hours by homeowner labor only............................................................ 4-6
Figure 5-1: Denver water use per unit area (ft2) versus landscape area in Denver..................... 5-3
Figure 5-2: Water use for the demonstration sample.................................................................. 5-4
Figure 5-3: Demonstration water use savings by Xeriscape group over respective control
group. ................................................................................................................ 5-6
Figure 5-4: Winter water use. ..................................................................................................... 5-6
Figure 5-5: Denver water use in time for April 1999 – January 2002........................................ 5-7
Figure 5-6: Colorado Springs water use in time for April 1999 – January 2002........................ 5-7
Figure 5-7: Annual precipitation at demonstration projects. ...................................................... 5-8
Figure 6-1: Demonstration maintenance cost and number of maintenance reports from all
homeowners per each demonstration................................................................ 6-4
Figure 7-1: Response rate to Final Survey of all participants per sample group........................ 7-2
Figure 7-2: Pro-environmental responses by demonstration. ..................................................... 7-3

x

Executive Summary

PURPOSE
In an effort to study the regional effects of water-conserving landscaping called Xeriscape, Metro Water
Conservation, Inc. (MWCI) and the Bureau of Reclamation (Reclamation) partnered with nine water
utilities along the Colorado Front Range. The study was entitled Yield And Reliability Demonstrated in
Xeriscape or YARDX. The YARDX study results were compared to similar studies in four other arid or
semi-arid communities in the western United States. The intent was to provide to utility managers a
good basis for future decision making about landscape water efficiency programs.

The Colorado project goal was to estimate Xeriscape benefits over a range of landscape and urban
environments and within the local climate and weather. Determining benefits involved the study and
assessment of landscape water use, installation costs, and annual maintenance expenses. Each of the nine
utilities in seven municipalities (Fort Collins, Greeley, Arvada, Wheat Ridge, Denver, Highlands Ranch
and Colorado Springs) hosted one or more demonstrations of Xeriscape application. There were 357
landscapes for single-family residential homeowner customers in the study. The study design called for
assessment of benefits in retrofits (existing landscapes retrofitted with Xeriscape), new starts (newly
constructed homes), and pre-existing Xeriscapes (previously planted Xeriscape). To assess water savings
and annual maintenance costs, the new Xeriscape sites were compared with nearby traditional landscape
sites, either Xeriscape or traditional landscapes. Project participants installed landscapes from 1997 to
early 1999. Operation and monitoring of the project demonstrations occurred over the period 1997
through 2002.

RESULTS
Establishment of study participant samples proved to be challenging. This was due largely to the cost of
landscapes, tight schedules (people needed more time to install landscapes and deal with associated
expenses), and the necessity to remain within project guidelines for the duration of the project. While
installation of Xeriscapes by participants was a lengthy process, installation costs appeared relatively
modest, ranging from about $0.90 to $1.45 per square foot (in 1999 dollars). Homeowners contributed all
or at least a substantial amount of the labor. Landscapes were completed for as low as $676 and as high
as $25,451. Estimated installation labor was an average of 50 to 60 hours per 1000 square feet of
landscape with an automatic irrigation system.

Xeriscape annual maintenance costs submitted by homeowners ranged from $0.34 to $1.33 per square
foot. For cost estimation, homeowner labor was computed at $18 per hour. Generally, maintenance costs
for the Xeriscape sites were less during the plant establishment years, but somewhat more during the
plant maturation years, compared to traditional landscapes. This suggests that as Xeriscapes age, they
gradually require somewhat more maintenance.

YARDX demonstrated that properly planned and installed Xeriscapes save water. The project Xeriscapes
saved from 18 to over 50 percent of the water when compared with paired traditional landscape control
groups. On new properties, YARDX results indicate that water savings in the 30-percentile range can
routinely be achieved, assuming the property owners are committed to maintaining the savings. New
property owners obtained their savings with a design scheme of approximately ¼ the area with low
water use plants, ¼ with moderate water use plants, and up to ½ the area with traditional turf. Higher
water savings could possibly be obtained with a ⅓-⅓-⅓ design scheme.

xi

The YARDX water savings from retrofits were slightly less than for new properties (generally 28 to 32
percent). Water savings in retrofits appear to vary with the amount of turf that remains in landscapes.
Although YARDX retrofit participants were guided toward the ⅓-⅓-⅓ design scheme, the actual water
savings did not reach the anticipated savings of 50 percent.

The YARDX project involved the installation and monitoring of landscapes, except for one demonstration
(Arvada/Wheat Ridge). In that case, older installed landscapes were only monitored during YARDX.
This older landscape study did not yield water savings, and it is not apparent why this occurred. A
number of differences other than age of the landscape were apparent. This demonstration consisted of
obtaining and comparing the water use of older pre-existing Xeriscapes, with a selected peer control
group, as in the other demonstrations. Their watering systems were a mix of manual (also called hose
drag) and automated systems. Over the years, there was likely the usual turnover in ownership, so
commitment to maintain the existing Xeriscape design might not have persisted with new owners.
YARDX did not provide education about Xeriscape as was accomplished with the other demonstrations.
More study of this data set is recommended.

The actual data collection for YARDX ended in 2002. A Final Survey was mailed to all project
participants to sample their attitudes on their landscapes, installation and annual maintenance costs, and
orientation on environmental issues. Overwhelmingly, Xeriscape owners in all demonstrations indicated
they were very satisfied with their landscapes, and that they would recommend this type of landscaping
to others. Interestingly, Xeriscape owners felt they spent less time on maintenance than with previous
traditional landscapes, which did not entirely agree with YARDX maintenance data results.

The information gained in the YARDX project should provide Front Range water managers an alternative
in dealing with current and future water demand. The general consistency of water savings in YARDX
demonstrations, and the lengthy data collection that occurred, should lend confidence that YARDX
results could be consistently achieved.

xii

1.CHAPTER ONE

Introduction

GENERAL
Population growth on the Colorado Front Range
continues at a high level. As a consequence of
the semi-arid climate, occasional drought and
the continued high growth, Front Range water
utilities are accelerating their planning for
increasing future water demand.

The Bureau of Reclamation (Reclamation) has a
responsibility to help improve water resource
management and the efficiency of water use in
the western United States. Reclamation
recognizes that cooperative efforts with partners
facing similar challenges can produce solutions
more efficiently to benefit all parties.
Consequently, Reclamation pursued several
cooperative demonstration projects of landscape
water conservation, collectively called the
National Xeriscape Demonstration Program
(NXDP). The NXDP cooperative studies were
conducted at locations in the western United
States that experience different climates,
including the Colorado Front Range centered at
Denver, Colorado; Phoenix, Arizona; Austin,
Texas; the Las Vegas area of southern Nevada;
and Fargo, North Dakota. Xeriscape ™ 1
landscaping is defined as a set of landscaping
principles, including low-water-using plants,
efficient watering systems, soil amendments,
and proper maintenance practices to create an
aesthetically pleasing landscape, while
maintaining desired attributes, such as reduced
water use, recreation, and cooling.

In January 1996, Reclamation partnered with a
Colorado Front Range nonprofit organization,
Metro Water Conservation, Inc. (MWCI), to co-

1
Denver Water, Denver, CO, holds the trademark for
Xeriscape.

Chapter One 1-1

sponsor the project known as YARDX (Yield
And Reliability Demonstrated in Xeriscape),
aimed at assessing the benefits of water
conserving landscaping. MWCI consists of a
group of Colorado water supply agencies and
interested stakeholders cooperating to aid
planning for future water demand and supply,
and to promote water conservation programs.
Reclamation, MWCI, and nine water utilities in
the Colorado Front Range comprised the group
participating in and sponsoring the YARDX
project.

STUDY LOCATION
The YARDX project became a 5-year
demonstration and evaluation study based on
Xeriscapes installed in nonrental, single-family
homes along the Colorado Front Range. The
area targeted in the Colorado Front Range is the
most heavily populated in the state. The study
includes landscapes in neighborhoods of Fort
Collins and Greeley in the north, Denver and
several suburbs in the middle Front Range
corridor, and Colorado Springs in the south. Figure 1-1: Location of cities
Figure 1-1 presents the study locations. This participating in the YARDX study.
region is considered high plains with a semi-
arid climate that receives on average 15 inches of
yearly rainfall in Fort Collins and Denver, and Specifically, the primary study goals of YARDX
17.5 inches in Colorado Springs were:
(http://www.weatherbase.com).
• Conduct consistent investigations at
Specifically, the nine utilities and 357 of their multiple sites in different geographic
single-family customers that elected to join the and municipal settings.
YARDX project are the Fort Collins Utilities,
East Larimer County Water District (Fort • Collect data that are uniform in content
Collins), Fort Collins-Loveland Water District, and method.
City of Greeley, City of Arvada, Wheat Ridge
• Quantify the range of water savings
Water District, Denver Water, Highlands Ranch
annually and seasonally when Xeriscape
Metropolitan Districts with Centennial Water
is properly installed and maintained.
District, and Colorado Springs Utilities.
• Determine the reliability of Xeriscape
STUDY GOALS water savings (Do water consumption
patterns change with the age of the
The YARDX study goals were to develop data
landscape or varying human factors?).
and provide evaluation and estimates on the
water savings, installation, and annual • Calculate the cost of installing Xeriscape
maintenance costs in implementing Xeriscape. landscapes for new construction and for
The project was to assess the reliability of retrofits to existing traditional
landscape water conservation from the landscapes (mostly high-water turf
application of Xeriscape landscaping. landscapes).

1-2 Introduction

• Analyze the cost of maintaining evaluations dealt with water use in single-family
Xeriscape landscapes for new residences, as did YARDX.
construction, for retrofits to formerly
traditional landscapes, and for pre- WHY YARDX IN THE FRONT RANGE
existing (installed prior to YARDX)
Xeriscapes compared to pre-existing Motivation for incorporating the Colorado Front
traditional Kentucky bluegrass Range in the NXDP national study included:
landscapes. (1) the high-growth characteristic of the area,
(2) concerns over prolonged drought in the state,
• Identify what marketing strategies (3) urban landscapes accounting for
affected the implementation of approximately 50 percent or more of the water
Xeriscape in this study. used by residences in this region (Winger), and
(4) the citizenry’s familiarity with Xeriscape.
PREVIOUS STUDIES As people continue to select Colorado for its
moderate climate, variety of outdoor recreation,
Some results from the other NXDP field studies jobs availability, and retirement appeal, water
are now available. The Phoenix project demand will increase. Front Range corridor
(Stinnett) obtained water savings of 53 percent growth is exemplified by “six counties in the
over control properties with traditional state [making] the census’ list of the 100 fastest-
landscaping. This project was similar to a growing counties in the nation” and Highlands
YARDX pre-existing Xeriscape landscape study Ranch in Douglas County, a southern suburb of
conducted in Arvada/ Wheat Ridge Denver, Colorado, and part of the YARDX
neighborhoods, except for lack of an observation study, “was the third-fastest-growing county in
period. The Phoenix study involved selecting, the nation from 2000 to 2003 “ (Siebert,
acquiring, and evaluating landscape historical Sinisi, B1).
water use data.
The U.S. Housing Markets, a research firm,
The Southern Nevada project (Sovocool and announced in 1997 that the Fort
Rosales) obtained water savings of 39 percent Collins/Loveland area (part of the YARDX
(summer) over control properties. Xeriscape study) ranked number 9 in the country in
maintenance was estimated to be about ⅓ less residential construction, with the Greeley area
than for control properties. (also part of YARDX) holding the number 18
spot in the United States (Cornelius C1+). This
The Austin Xeriscape project (Gregg, 1994)
Front Range metropolis, from Weld County in
obtained water savings of 31 percent during
the north to Pueblo County in the south,
summer months. Nelson (1994), in preliminary
includes about 3,733,308 residents (Romine).
results of the North Marin Study, obtained a 25-
percent water savings from Xeriscape. Testa The high growth and difficulty of executing new
and Newton (1993) obtained a 33 percent water traditional water projects has motivated water
savings in a Mesa, Arizona, study of Xeriscape. utilities to consider water conservation
alternatives to water supply development
Xeriscape participants in all projects of the
options, in part, because of the economic and
NXDP have overwhelmingly expressed
environmental concerns associated with
satisfaction with their landscapes and indicated
traditional water projects. Landscape
they would freely recommend this type of
irrigation’s water use near 50 percent of
landscaping to others.
residential use is an ideal target for water
The above studies have provided evidence conservation, and offers a potential source for
suggesting that Xeriscaping can reduce dealing with a major challenge of future Front
landscape water usage by 20 to 50 percent Range growth.
during peak irrigation months, as compared to
Front Range landscapes with predominantly
traditional turf landscapes. Those water-use
thirsty Kentucky bluegrass typically need nearly
30 inches additional water over average rainfall

Chapter One 1-3

(Winger) to maintain their health. As suggested and mulched areas without plantings. This
in the previous studies discussed above, lower landscaping approach is particularly being
water demand plantings and more efficient utilized in moderate and higher priced
irrigation methods could potentially decrease properties. Positive results from YARDX and
outdoor watering by 30 to 50 percent, or about readily available educational materials from
10 inches less irrigation. Noticeable water water utilities (for example) could further
savings are obtained when viewed over the promote the installation of water conserving
many landscapes of the Front Range. plantings.

The extreme Front Range drought of 2002 Following in this report are descriptions of the
(exceeding 100-year records) and the continuing YARDX project and the results obtained.
dry years have contributed to increasing interest Chapter Two discusses the project design,
in water conservation approaches. The 2002 Chapter Three covers the promotion efforts to
drought persuaded utilities to enact watering recruit homeowners to join YARDX and install a
restrictions. These restrictions apparently Xeriscape. A summary of the Xeriscape
impacted the 2002 water savings by YARDX installation costs is given in Chapter Four.
Xeriscapes in Colorado Springs (discussed in Chapter Five presents water use results, and
Water Use Results). Drought remains an ongoing Chapter Six covers landscape maintenance costs.
concern in Colorado. Chapter Seven presents the YARDX
participants’ responses to a Final Survey that
Xeriscapes are a familiar sight in Colorado Front included questions on satisfaction with their
Range municipalities, rendering the area a likely landscape and query for pro-environmental
candidate for Xeriscape research. In new tendencies that may have influenced their water
neighborhoods, there are indications of use. Finally, Chapter Eight presents a summary
increasing use of lower water demand plantings and conclusion of primary results.

1-4 Introduction

2.CHAPTER TWO

Project Design

INTRODUCTION
A proper evaluation plan and data sampling
design was required to accomplish the goals
established for YARDX. However, design features
would need to function within demonstration
logistics. The desired sampling scheme was
random allocation of the treatment (Xeriscape)
and control (traditional landscape) to lessen
possible bias. From a large pool of willing
participants, homeowners would be assigned to
treatment and control groups at random.
Treatment participants would install Xeriscapes
and control participants would install traditional
landscapes. It was also determined to sample only
owner-occupied, single-family properties to
reduce noise that may occur from rentals or
multiple-family housing. To encourage
participation, a rebate was established of $300 to
new starts (newly constructed homes), and $600 to
retrofits (Xeriscaping a portion of an established
high-water use landscape). Also, Xeriscape
participants would be provided 2.5 hours of
professional landscape design assistance,
educational seminars on Xeriscaping, and one-on-
one consultation upon request during the study
period.

In the initial stages of the project, it became
apparent that establishing a treatment group with
random selection would not be possible, despite
the benefits provided by YARDX. There was
inadequate participation to install Xeriscapes. The
primary barrier appeared to be Xeriscape
installation cost and time, though other concerns
voiced included conforming to the rigor of
participation (e.g., reporting requirements) and
structured Xeriscape design features and irrigation
type selected for particular demonstrations.

Elimination of random allocation to the treatment
group forced development of a new Xeriscape

Chapter Two 2-1

sampling plan, possibly introducing more data 2. Xeriscape application type (retrofit, new
noise. start or pre-existing)

That nine utilities expressed interest in joining 3. Xeriscape application level (landscape
YARDX also impacted the study design. Large designed for 30-40 percent or
participation presented an opportunity to develop 60-70 percent water savings)
seven demonstrations, each with at least one 4. Yard size
unique variable factor setting (such as large yard
or sandy soil). 5. Irrigation method (manual hose drag or
automatic sprinkling system)
The project covered approximately 6 years of
6. Family income level (home values were
information gathering, as a consequence of the
used to approximate income levels)
landscape enlisting and installation process
becoming rather time consuming. The YARDX 7. Soil type (loam, sand, or clay)
project began enlisting properties in 1997.
8. Precipitation
Xeriscapes were installed from 1997 through the
first half of 1999. Field data collection continued The first variable, water use, is the key response
until the end of 2002. Data responses from (dependent) variable of the study. The other listed
participants continued well into 2003. variables, operating in combination, are
considered to substantially impact water use.
DEMONSTRATION YARDX collected data on some secondary
variables, including exposure of a landscape to
CHARACTERISTICS wind (a categorical variable as defined by
Consideration was given to the various variable YARDX), degree of shadiness (more important in
factor influences (Gregg, et al.) on study sample retrofits that had pre-existing plants remaining
establishment, including project duration and and providing shade), and area with 15-degree
available funding. Table 2-1 lists the YARDX slope or greater. A variable not listed above that
seven demonstrations along with their primary certainly impacts water use, but is difficult to
influence factor and logistic characteristics. In assess, is landscape maintenance. YARDX
general, a mixture of three retrofit, three new collected related information via an assignment of
starts and one pre-existing (Xeriscape landscapes a landscape or turf health score (see appendix C).
installed prior to YARDX, by the spring of 1996) While the health score assignment was subjective
demonstrations was developed for YARDX. (by the person conducting the field audit), there
Marketing failed to convince more homeowners to may be some useful maintenance level assessment
participate in additional demonstrations, except in the scores.
possibly for additional high-priced properties that
may have joined if pursued. The table shows the Table 2-1 shows the various influences of factor
demonstration in Greeley as a mix of retrofits and settings by demonstration. Because landscape
new starts. This occurred because few people water use is the end result of a number of
joined the project. operating and interacting factors such as weather
and climate, soil type, yard size, and home
Demonstrations were designed to address occupant attitude, individual neighborhoods of
variables that appear to impact water use. differing demographics may yield different water
Previous studies (Gregg, et al.) have identified use savings.
variables that are correlated with water use.
Accordingly, the YARDX field demonstrations Xeriscape application type refers to a Xeriscape
were structured to yield data on eight seemingly new start, retrofit, or a pre-existing type.
correlated variables, including the primary
Xeriscape application level relates to the presumed
response variable, water use.
water savings level of the Xeriscape design, based
on the area ratio landscaped in high-, moderate-,
STUDY VARIABLES or low-water plants (see Glossary) plus the area
1. Water use (primary response variable) amount of hardscape. For YARDX, two

2-2 Project Design

application levels were selected: 30-40 percent whatever type of irrigation system had been
water savings design (used in the new start installed, and thus a mixture occurred in this
demonstrations) and the higher water savings of group.
60-70 percent (used in the retrofit demonstrations).
The 30-40-percent design consisted of ¼-¼-½ area Home value was used as a surrogate for
ratio of low, moderate, and high water use. The attitudinal variables of family income levels that
60-70-percent design used corresponding area some previous studies (Gregg, et al.) suggested
ratios of ⅓-⅓-⅓. The pre-existing landscapes were could affect household water use. The home value
taken “as is.” Consequently, a combination of representation varied by region. One value that
different potential water savings landscape plans may seem low in one locale could reflect a high-
existed in these yards. priced home in another city. A relative
comparison system on a local basis was used.
The water savings estimation levels (Xeriscape Three general classes of home values are identified
application levels) could possibly be achieved on Table 2-1: low, medium, and high.
under proper designing and installing, watering
and performing general maintenance (such as The soil types also fell into three main groups:
thinning plants). Landscape designs were loam, clay, and sand. In some demonstrations,
structured to specify certain water-hardy plants there were some properties with mixes, but
and grasses, density level of installation, generally there was a predominant soil type.
proportion of turf to nonturf area, and landscape
Lastly, three groupings of relative precipitation
preparations that included soil preparation,
were prevalent during the growing season of
mulches, and hardscape.
April through October in YARDX demonstrations.
The variable yard sizes were categorized into The relative code of ”low“ indicated about
small, medium, and large. The classification was 8 inches or less, “moderate” reflected more than
relative to each city, so that category size cut-offs, 8 and up to about 11 inches, and “wet” indicated
from city to city, varied minor amounts. Thus, 12.5 inches or more of rainfall.
large yards in Fort Collins tended to be larger than
large yards in the Denver demonstration area. SAMPLING PLAN
The study in Fort Collins was structured to
Because random allocation of the treatment was
include larger properties and more affluent
not possible, a new sampling plan was needed, at
owners. Table 2-1 presents the yard size cut-offs.
least for establishing the Xeriscape sample. For
Approximately, the three size categories that
each demonstration, the goal was still to have the
described one-half of the “landscapable area”
treatment and control group properties as similar
(square footage to be formally landscaped) were:
as possible, except for random assignment of the
low-sized yards up to about 2750 square feet,
treatment. After some consideration, the sampling
medium sizes from approximately 1000 to 4000
plan was developed that involved:
square feet, and large sizes with over 2500 square
feet. The Xeriscape landscape had to encompass • Only certain households were willing to
50 percent or more of the landscapable square participate in the project. The greatest
footage. This area did not include dryland area difficulty in obtaining participants was in
(natural vegetation surviving solely on rainfall). the Xeriscape group.
Only one YARDX property involved dryland area.
• Willing Xeriscape participants were
Only two types of irrigation systems were studied surveyed to conform to settings of the
in the project: automatic sprinklers on a clock and relevant demonstration. Participants (and
timer that would start and stop the watering as properties) that satisfied their
programmed, and hose dragging. In each demonstration settings were invited to
demonstration, all participants (both treatments join the study.
and controls) had the same type of irrigation
system with the exception of the pre-existing
demonstration. In this case, YARDX accepted

Chapter Two 2-3

2-4

Table 2–1: YARDX demonstrations and primary characteristics.
Xeriscape Periods of
Project Design

Demo Level ½ Yard Water Data
Demo Type 1 (%)2 Soil Rainfall 3 Size (ft2) Home Value Irrigation Analyzed
Arvada/ PX Any 4 Clay 8" 1000-4000 $150-275,000 As Is 9/99 -- 12/02
Wheat Ridge (Moderate) (Medium) (Medium)
Colorado RT 60-70 Clay 12.5" 1000-3750 $150-200,000 Auto 2/99 -- 12/02
Springs (Wet) (Medium) (Medium)
Colorado NS 30-40 Sand / 12.5" 1000-3750 $150-200,000 Auto 2/99 -- 12/02
Springs sandy loam (Wet) (Medium) (Medium)
Denver RT 60-70 Clay 9" up to 2750 up to $200,000 Manual 5/98 -- 12/02
(Moderate) (Low) (Low)
Ft. Collins / NS 30-40 Clay 9" >2500 >$175,000 Auto 9/98 -- 12/02
Loveland (Moderate) (High) (High)
Greeley RT & 60-70 Loam 8" 1000-3750 $80-200,000 Auto 2/98 -- 12/02
NS (Low) (Medium) (Medium)
Highlands NS 30-40 Clay / 11" 1000-4000 $150-250,000 Auto w/ 2/98 -- 12/02
Ranch clay-sand (Wet) (Medium) (Medium) submeter
1 NS= New Start; RT= Retrofit; PX=Pre-existing
2 Xeriscape application level was designed for a 30-40% or 60-70% water savings over traditional landscapes.
3 May to September average rainfall
4 The Arvada/Wheat Ridge demonstration was not a YARDX-designed project, and Xeriscape application levels were not
specified.

• With knowledge of Xeriscape toward pro-environmental orientation over
participant characteristics, a pool of control groups. However, this result may have
properties with traditional landscapes occurred as a result of several circumstances. It
was surveyed for matching the is not clear that pro-environmentalism always
treatment characteristics. A random leads to higher water conservation.
sample was selected of the acceptable
control pool, in excess of treatment SAMPLE SIZE ESTIMATION
numbers.
Sample sizes must be large enough to enable the
• Those randomly selected were contacted detection of a treatment effect (Xeriscaping in
for joining the control group of the this study) on water savings anticipated from
particular demonstration. Not all Xeriscape application. The natural “noise”
contacted agreed to join the study. levels of the local water-use data must be
Upon enlisting the desired number of assessed to estimate sample sizes. Data noise
control participants, the contacting sources can occur from influential elements
ceased. changing with time (such as effects of weather)
and influences varying across households. Real
• Information conveyed to control water-use historical data offer an opportunity to
participants consisted of only stating estimate sample size requirements.
that they would belong to a local
landscape water-use study of traditional Several years of water-use data were obtained
landscapes (to avoid potential influence from the city of Boulder, Colorado, for some
on water use). No education on 26,000 homes. The project team agreed that the
landscaping was provided to control Boulder data were well representative of wide
participants by YARDX. socio-economic, educational, and other water-
use-impacting variables. Given this assumption,
The restricted population of willing Xeriscape the Boulder water data would be a suitable
participants could lead to some bias in the candidate for sample size estimation for YARDX
samples obtained. The project also used a rebate demonstrations.
to aid in obtaining adequate numbers of
participants, which could introduce some bias. Computer software was developed that utilized
The rebate undoubtedly motivated some a resampling technique similar to bootstrap
participation (see Chapter Seven, Final Survey). (Manly) that uses random selection of samples,
Finally, new home construction leads to the which can be compared. This technique was
need for new landscapes (neighborhood used to study the detection of different water
covenants usually require homeowners to install amount savings in different sample sizes from
a landscape). The type and amount of bias samples without water savings. For example, a
resulting from these initial conditions was 30-percent increase in water use was applied to
difficult to foresee. In particular, the rebate each of 100 randomly selected samples of
could lead to specific types of water users. residential water use and compared with 100
However, the project team realized that some unaltered samples, using a nonparametric test at
participants may simply want to conserve water. the 5-percent significance level. By studying the
Others seemed attracted to the aesthetics of proportion of comparisons over and under the
Xeriscape. 5-percent significance level, an estimate of
probability of detecting the inserted water
Efforts made to investigate bias consisted of amount change was obtained.
studying responses in a survey administered at
the end of the project by YARDX (see Chapter Results of applying different water amount
Seven, Final Survey). Specific questions on the changes in different sample sizes indicated that
survey suggested the environmental inclination to achieve the 90-percent detection level, water
by participants and the primary motivation for use from about 30 homes for approximately 25
joining the study. Results in Chapter Seven growing season months would be required to
suggest some bias of Xeriscape participants detect a 30-percent water savings. This meant

Chapter Two 2-5

that water data should be collected from about season water use, with the exception of the
30 homes per sample for four growing seasons Highlands Ranch demonstration, had to be
in the YARDX demonstrations. The growing derived from readings from the utility-provided
season selected was April through October. home water meter. In the case of Highlands
This sample size included about 10 percent Ranch, an additional meter was installed to
overage for participant dropout during the measure only the landscape water use. These
study. The sample estimate would data were available for the last two growing
accommodate typical data noise caused by seasons only (2001 – 2002).
varying climate effects, property characteristics,
irrigation systems and management, and COST ESTIMATION
homeowners’ varying behavior.
Xeriscape Installation
FIELD DATA One of the primary goals of YARDX was to
For their customers in the YARDX study, the estimate the cost of installation of Xeriscapes.
participating utilities routinely collected Data were obtained from six demonstrations.
monthly or bimonthly water-use information. The seventh demonstration, the Wheat Ridge /
This information was provided to the YARDX Arvada study, did not involve the installation of
project for analysis over the course of the study. landscapes, but rather the study of existing
Table 2.1 shows the periods of water data for Xeriscapes and traditional landscapes. See
each utility. The water data required substantial Chapter Four for a detailed discussion of
manipulation for the final tabulation. The final installation costs.
tabulation consisted of calendar monthly totals
(obtained from reporting period average daily Maintenance
usage) per participating property. The Annually, both the treatments and controls
estimated monthly totals became the water-use received a maintenance log to record their time
basic database for treatment and control sample and expenses on their landscapes. Segregated
comparisons. by major landscape area (e.g., turf, groundcover
plants and perennials, irrigation, etc.) and by
Generally, water-use data were available from
three seasons (spring, summer/fall, and winter),
mid-1998 or early 1999 through 2002. Water-use
these data were tabulated annually from 1997
data files were constructed with complete
through 2002. A discussion of maintenance
records through 2002 for all participants. A few
costs can be found in Chapter Six, Maintenance
properties underwent major landscape revisions
Costs, and the form can be found in Appendix E.
no longer in character with the particular
demonstrations. These records were adjusted In addition to studying the maintenance costs,
for a partial study time period and included in the effect of providing landscape maintenance
the analysis. education was also considered. Xeriscape
maintenance training was provided to the new
To estimate outdoor use, winter use was
start and retrofit treatment groups but not to the
subtracted from summer meter readings.
pre-existing sample. By isolating the pre-
Winter use was estimated by averaging water
existing group, the effects of maintenance
use for January, February, November, and
training could be better observed between the
December. These were considered the core
“trained” and “untrained” groups.
winter months most likely to include little or no
outdoor use. The project team decided that
using the average of “encircling” months best ANNUAL SURVEYS
represented indoor use. For winter comparisons Also annually (from 1997 through 2002), both
between samples, the water use in March was the treatments and controls received a survey.
included with the above 4 months. The purpose of this survey was to identify the
age of the property and indoor water-usage
The growing season months were designated as
factors such as the number of adults and
April through October. Estimation of growing

2-6 Project Design

children, the number of low-flow toilets and EXAMPLE OF DESIGN
showers, etc. At the conclusion of the study in
2002, the survey was expanded to determine Figure 2-1 presents a sample plan designed for a
attitudes toward their landscapes and study participant. Figure 2-2 depicts the
environment. The results of this survey can be completed Xeriscape 4 years after installation.
found in Chapter Seven, Final Survey. Other homeowners’ landscape photographs,
before and after installation, can be found in
Appendix J.

Figure 2-1: Xeriscape design, Colorado Springs, CO. Design created by Joelle Dunaetz,
Designs by Dunaetz.

Chapter Two 2-7

Figure 2-2: Photograph of the Xeriscape 4 years after installation in Colorado Springs.
Photo courtesy of participant.

2-8 Project Design

3.CHAPTER THREE

Promotion

INTRODUCTION
The YARDX project was designed to study
owner-occupied, single-family residential
homes. Project participants would be
homeowners willing to install a new Xeriscape
landscape (treatment) or maintain a traditional
landscape (control) throughout the study.
YARDX offered a rebate to treatments to
participate in the study (except for the pre-
existing demonstration). Most of the cost of
installing Xeriscape would be borne by those
willing to participate.

In each demonstration, approximately 45 to 55
homeowners were targeted for each of the
treatment and control groups. The team
planned to enlist about 40 participants, nearly
20-percent over subscription, to deal with
anticipated attrition. The sample size estimation
of Chapter Two indicated the need for about 30
participants per sample group completing the
study. Accordingly, five of the seven
demonstrations achieved the desired number of
signups at the onset.

Table 3-1 indicates a total of 294 Xeriscape
participants signed up and paid a $100
commitment check (where required).
Participants were given time schedules for
completing their Xeriscapes. A total of 170
Xeriscape participants of the original 294 that
initially joined YARDX (58 percent of 294)
completed the study.

Thus, 124 Xeriscape participants failed to
complete YARDX. Each demonstration was
designed with pre-specified criteria of
participation. Failure was caused by violation of
criteria, such as failing to install a required
automatic watering system. Even with high
attrition, all but the Greeley demonstration had

Chapter Three 3-1

Table 3–1: Number of participants who signed up, completed or dropped from the YARDX
project.
No. that Completed Project to
No. of the End 1 (% of Signups) No. of Treatments
Treatment Dropped from Project
Demonstration Signups Treatments Controls (% of Signups)

Arvada/Wheat Ridge 30 26 (87%) 28 4 (13%)
(PX)

Colorado Springs (NS) 52 27 (52%) 37 25 (48%)

Colorado Springs (RT) 64 27 (42%) 32 37 (58%)

Denver (RT) 63 34 (54%) 31 29 (46%)

Fort Collins (NS) 46 34 (74%) 38 12 (26%)

Greeley (NS & RT) 13 6 (46%) 4 7 (54%)

Highland Ranch (NS) 26 16 (62%) 17 10 (38%)

Total 294 170 (58%) 187 124 (42%)

Legend:
RT = Retrofit Xeriscape NS = New Start Xeriscape PX = Pre-existing Xeriscape
1
The project was able to use only partial data on 12 treatments and 8 controls due to changes to their landscapes late in the study.

adequate participation for conducting data used at times. Included in the mailing was a
analysis. colored flyer with Xeriscape landscapes, a brief
letter, and an application form. When a
Appendix A presents drop-out numbers with homeowner called or returned an application
breakdown by city and reason for leaving the form, YARDX responded with an additional
project. letter covering more details about the project,
including costs of installing a Xeriscape.
PROMOTING XERISCAPE Applicants were then asked to confirm their
PARTICIPATION interest with a mail-back form. Lastly, the
project team prequalified the interested
The YARDX Project used several different homeowners with a drive-by audit checking for
avenues to elicit homeowner participation in the approximate home value and age, irrigation
Xeriscape group. The primary approach was a system, yard size, soil type, etc., for matching
direct mailing campaign to preselected the demonstration’s study settings.
neighborhoods that fit the criteria for the
demonstration (home value, soil type, etc.), and At each level of scrutiny, homeowner response
to new water tap permittees (for homes recently levels dropped. Participation confirmation
built for the new start demonstrations). Over requests were below expectations.
19,600 letters were mailed to prospective Approximately 234 homeowners (24%) of the
participants by 9 utilities on utility letterhead. original 965 who expressed interest in installing
a Xeriscape responded favorably to our
Table 3-2 shows homeowner approximate confirmation request. Of the original 19,600
response to the Xeriscape direct mail campaign. homeowners that were contacted, 1.2 percent
About 5 percent or 965 prospective participants replied that they wanted to proceed with
responded with interest. Second requests were

3-2 Promotion

Table 3–2: Example of Xeriscape signup results from a direct mail campaign.
Signups
No. of No. of Qualified
No. of Applications Homes
Invita- Received (% of % of % of % of
tions (% of Invitations / Invita- Appli- Quali-
Demonstration Mailed Invitations) % of Apps) No. tions cations fied
Arvada (PX) 664 54 (8%) 24 (4%) / (44%) 20 3 37 83
Boulder (NS) 143 14 (10%) cancelled cancelled demonstration
demonstration
Colorado Springs 7,305 360 (5%) 280 (4%) / (78%) 81 1 23 29
(NS & RT)
Denver (NS) 586 23 (4%) cancelled 2 0.3 9 —
demonstration
Denver (RT) 1,524 246 (16%) not available 63 4 26 —
Ft. Collins (NS):
Ft. Collins Utility 355 31 (9%) 28 (8%) (90%) 13 4 42 46
Ft. Collins- 160 27 (17%) not available 10 6 37 —
Loveland
East Larimer 38 16 (42%) not available 6 16 38 —
County
Ft. Collins Total 553 74 (13%) not available 29 5 39 —
Greeley 7,058 194 (3%) 135 (2%) / (70%) 13 0.2 7 10
(NS & RT)
Highlands Ranch 544 not available not available 26 5 — —
(NS)
Subtotal 18,377 965 (5%) -- 234 1 24 —
Various 1,223 not available not available not available
Grand Totals 19,600 965 (5%)1 -- 234 1 24 —
Legend:
RT =Retrofit Xeriscape NS =New Start Xeriscape PX =Pre-existing Xeriscape
1 Calculated without Highlands Ranch results

installing a Xeriscape. These results forced a educate the citizenry about the overall
prolonged marketing effort. benefits and aesthetics of Xeriscape, but
did not specify information on YARDX.
Besides mailing invitations to prospective
participants, several other marketing campaigns • The government channel on cable TV
were added to help enhance YARDX credibility promoted the project in Colorado
and circulate availability more widely. Springs and Greeley.
Additional marketing included:
• Newsletter articles were published
• Approximately 19 newspaper articles and/or presentations made to such
highlighting the project and its benefits organizations as the Home Builders
were circulated in various papers and Association, real estate agent groups,
cities. homeowner associations, environmental
groups, garden centers, Colorado State
• About seven paid newspaper ads were University Extension, the Horticulture
printed in local newspapers. Arts Society, the Associated Landscape
Contractors of Colorado, and the
• Denver Water sponsored several TV
Nurserymen’s Association. These
news bites about Xeriscaping on a local
groups were in contact with the public
TV news station. This publicity helped

Chapter Three 3-3

and likely enhanced promotion of demonstration. The control sign-up process
YARDX. started in 1998 and continued into 1999.

• Letters were sent regarding YARDX to
XERISCAPE SEMINARS
more than 100 builders and Home
Builder Association members in the Homeowners who confirmed their interest to
Colorado Front Range. In addition, participate in YARDX were subsequently
YARDX representatives visited or called invited to a 2½-hour Xeriscape seminar in their
several homebuilder sales offices to neighborhoods. At these seminars, members of
utilize their contacts with the public. the project team discussed the requirements of
These efforts produced varying degrees the project, the benefits, preparation for a
of support for YARDX. private design session with a landscape
architect/designer, the seven principles of
• Flyers were deposited at single-family Xeriscape, and the procedures and costs
residential homes, at libraries, and at involved with installing a landscape. At the
homeowner associations in Fort Collins, seminar conclusion, the team signed up
Greeley, and Highlands Ranch. The participants, which required a $100 commitment
Colorado Springs Xeriscape check. These monies helped to subsidize the
Demonstration Gardens passed out homeowners’ private design sessions. In some
YARDX flyers to their visitors. cases, YARDX representatives met with
homeowners who were unable to attend
• Some utilities mailed inserts with water
scheduled seminars.
bills.

• The Fort Collins Utility taped a video of RESTRICTING YARDX
the YARDX introductory Xeriscape PROMOTION
seminar for distribution to interested
homeowners. YARDX established standards and related
expectations of treatment and control groups.
• For locating pre-existing Xeriscapes in These demands may have contributed to the
Arvada, a mailing was conducted to slow signup rate. However, obtaining high
participants in the city’s previous quality data demanded rigor in the conduct of
Xeriscape rebate program. Also, the project.
Arvada and Wheat Ridge
representatives assisted in identifying Standards required of new start and retrofit
areas of existing Xeriscapes in their Xeriscapes were similar. The pre-existing
neighborhoods. Xeriscapes had less demands, as those
Xeriscapes were already established and the
The Xeriscape marketing effort started late in interest was to evaluate the water savings of
the summer of 1996 and continued through the older Xeriscapes. The following lists some
winter of 1998, exceeding the original time landscape requirements established for YARDX:
schedule and plans. YARDX spent at least an
additional year on marketing due to initial low • The retrofit and new start homeowners
sign-up rates. had to commit with a $100
nonrefundable check, which would be
Mass marketing techniques were not used to returned to them through project
elicit homeowners for the control role. Chapter benefits. This fee was enacted to help
Two presents procedures for selecting control determine participants likely to
properties. Promotion consisted of personal complete their installations.
door-to-door contact on randomly selected
properties that were determined to closely • Retrofit and new start homeowners also
match treatment properties of the had to have their landscape plans
approved and installed by June 1, 1999.
The plan had to encompass over

3-4 Promotion

50 percent of their landscape. For new 100 percent of those completing their
starts, they could not have an existing installations received the maximum
landscape older than 6 months in at rebate allowed.
least 50 percent of the yard. Conversely,
the retrofit demonstrations had to have • Participate in YARDX-exclusive
an existing landscape. Both groups had discounts negotiated with 25 vendors in
to track their installation work time and 5 cities throughout the study region.
expenses and submit copies of all These discounts averaged 10 to
invoices to YARDX. 20 percent and included nurseries and
garden centers, irrigation
• Annually, the homeowners were to suppliers/contractors, soil amendment
report landscape maintenance including providers, hardware stores, and
tracking their time and expenses, hardscape suppliers.
completing an annual survey, and
providing YARDX with photos of their • Receive a 2½-hour private session with
yard. The project supplied a a landscape architect or designer paid
maintenance tracking form. The pre- for by the YARDX Project. Participants
existing Xeriscapes were not required to were given a simplistic homework
submit photos. assignment to complete before their
design session: photograph their yard,
• Homeowners were asked not to complete a questionnaire about their
significantly alter their landscape desires, and draw a schematic of their
during the study (concluding property with measurements.
December 31, 2002) and to allow project
staff to occasionally view the progress of • The YARDX Project provided one free
the Xeriscape. maintenance seminar in four
demonstration cities (Greeley, Fort
• Participants had to own their home (no Collins, Denver, and Colorado Springs).
renters); not be an employee of their Additionally, an annual newsletter with
utility, Metro Water Conservation, Inc., maintenance tips was mailed to all the
or Reclamation; and have no expectation Xeriscape participants except for the
of moving in the next 3 to 5 years. pre-existing group in Arvada and
Wheat Ridge. The plan was to observe
• For the Highlands Ranch demonstration what the pre-existing demonstration
only, Xeriscape participants were would do without any formal
required to install a submeter on their maintenance education from the project.
irrigation line to measure outdoor
watering. The submeter was provided • At the conclusion of the study, the
by the utility, but the homeowners were YARDX project team is to provide each
responsible for its installation. homeowner with a personalized
analysis of their water use, installation
PROJECT BENEFITS costs, and maintenance expenses,
comparing their results to those of other
The total benefit package offered to the project participants.
treatment group encompassed:
• Some general benefits include increased
• Receive a rebate of $0.45/ft2, up to a home value through an attractive
maximum of $300 for new starts and landscape, and reduced water bills by
$600 for retrofits, depending on the size saving water. Homeowners would be
of the Xeriscape installed. No rebate helping the environment by creating
was disbursed to pre-existing Xeriscape microclimates and reducing energy
homes, since they had already needs, reducing wastewater through
completed their installations. Almost reduced water applied to landscapes,

Chapter Three 3-5

maintaining higher water quality return • The utilities with dedicated staff on the
flows (assumed less pesticides and Project had better signup results.
fertilizers), and assisting utilities to
better deal with future water demand • Signups for the retrofit treatment group
planning and demands. Water progressed better than for the new starts
conservation can provide immediate (about 13 percent better) despite
benefits without requiring marketing for only two retrofit
environmentally sensitive water demonstrations compared to three new
projects. start groups. The smaller population of
new homeowners (compared to the
The promotion campaign advertised project number of homeowners with existing
benefits to entice homeowners to join YARDX. landscapes) indicated they were
The value of the provided landscape design, overburdened with new home expenses
installation discounts, and project rebate were and tasks and thus were reluctant to
assessed as high as $1000 for the higher rebated commit to a new type of landscape.
properties. Often their builders provided Kentucky
bluegrass in parts of their yard,
The main benefits for the control group were the diminishing the need for a complete
personalized analysis of their water use and landscape. Additionally, many
maintenance expenses, and the comparison to homeowners had small children who
other participants in the project. They also needed grassy areas for play and chose
gained the satisfaction that they were helping mostly turf landscapes.
their utility better determine the water use of
traditional landscapes. • After newspaper articles were published
about YARDX, the utilities were
Two pre-existing Xeriscape demonstrations inundated with inquiries (Denver about
were dropped: Boulder and Colorado Springs. 150 phone calls and Colorado Springs
Boulder switched from being a new start about 1400). Most of these callers lacked
demonstration to a pre-existing demonstration. sufficient interest to proceed.
They received a poor response from letters sent
to homeowners who had previously been • The Fargo, North Dakota, Xeriscape
identified as having a Xeriscape landscape. study also experienced new start signup
Since Colorado Springs was already hosting two rates less than initially anticipated,
demonstrations, a retrofit and new start, they despite using a higher rebate amount of
decided not to pursue a pre-existing $1200, or four times the YARDX new
demonstration as well. start rebate of $300 and twice the retrofit
rebate of $600. However, the Fargo
OBSERVATIONS rebate did contribute to an increased
interest in the retrofit demonstration.
During the promotion campaign, some success
and lack thereof became apparent. The
following lists some observations on these
issues:

3-6 Promotion

4.CHAPTER FOUR

Installation Cost
of Xeriscape

INTRODUCTION
Estimation of the cost of installing Xeriscape was
one of three major goals of the YARDX study.
The cost of Xeriscape will likely impact its level
of implementation. Participants in Xeriscape
samples of six demonstrations (excluding the
pre-existing group) were asked to submit
receipts of costs of materials and labor
pertaining to installation of their landscapes.
Because many homeowners provided some or
all of the installation labor, they were asked to
track their work hours as well as expenses.
YARDX tabulated participant information.
Similar information was not requested from
control properties, so as not to potentially
influence their water use. As a result, no
treatment/control installation cost comparison
was conducted.

COST DATA QUALITY
To ensure a high response and data quality,
Xeriscape participants were not given a cash
rebate until they had submitted a list of all of
their expenses along with receipts and a total of
their related work hours broken down by
irrigation and other construction work. YARDX
reviewed all participant-submitted paperwork
and categorized expenses into 11 groups,
including plants, hardscape, mulch, walls or
drainage, irrigation, home labor hours,
contractor hours, contractor costs, non-Xeriscape
costs, other Xeriscape expenses, and discounts
received. A total of 143 participants from 6
demonstrations submitted landscape cost data.
Figure 4.1 shows data contribution by
demonstration.

Chapter Four 4-1

40

35

30
Starts / Retrofits

25

20

15
COS-NS = Colorado Springs new start
10
COS-RT = Colorado Springs retrofit
5
DEN-RT = Denver retrofit
0
GRL-RT HGR-NS COS-RT COS-NS DEN-RT FTC-NS FTC-NS = Fort Collins new start
Homes 6 15 27 27 34 34
GRL-RT = Greeley combined new start & retrofit
HGR-NS = Highlands Ranch new start

Figure 4-1: Number of new start (NS) and retrofit (RT) Xeriscapes providing installation
costs.

computed to compensate for area differences.
COST ESTIMATION Figure 4-2 presents the average area of
landscapes in each Xeriscape sample. This
Determining the average cost of a Xeriscape was figure shows the relative size of landscapes on
a goal of YARDX. Categorization of cost data average. Property landscape size was reduced
also enabled determination of the costs of by “dryland” area where appropriate. Dryland
different landscape components. Studying cost is defined as an area with sparse, native
data suggested that extraordinary costs vegetation that receives only natural
associated with substantial land preparation be precipitation and has no irrigation system. Only
analyzed separately from other invoice costs. one retrofit property area (in Colorado Springs)
These include costs such as grading that was impacted by this adjustment.
generally is conducted by builders, retaining
walls, drainage issues, and non-Xeriscape The retrofit landscapes of demonstrations in
expenses such as fences, dog runs, playgrounds, Greeley, Colorado Springs, and Denver
birdbaths, storage sheds, and large equipment contained areas of pre-existing plants not
or tools. replanted. Consequently, nonreplanted areas
did not contribute to landscape costs. This
Homeowners received substantial discounts suggested adjustment of area used in
from nurseries and other providers of landscape computations, for nonreplanted area. However,
materials as a result of participation in YARDX. only the Denver demonstration pre-existing area
As a consequence, one-half of vendor discounts measurements were consistently logged,
were considered extraordinary and added to adequate for making adjustments. Some retrofit
invoice costs to simulate typical cost. The homes retained existing turf in their landscapes
special vendor discounts had been negotiated in small, separate areas difficult to accurately
specifically for YARDX participants and were measure. No adjustment was made in Table 4-3
not available to the general public. The vendor for nonreplanted areas in the remaining two
discounts were an additional benefit to retrofit demonstrations, Colorado Springs and
participants for joining YARDX. One-half of Greeley. The adjustment effect on installation
discounts for YARDX were considered a cost per square foot for the Denver
reasonable estimate of discounts obtainable by demonstration is an increase of 7.9 percent. An
the general public. estimate of the underestimate of cost per square
foot for the Colorado Springs and Greeley
YARDX properties varied in area.
retrofit demonstrations is Denver’s 7.9 percent
Consequently, cost per square foot was

4-2 Installation Cost of Xeriscape

increase. The average area given for Denver in irrigation. It is not clear what caused
Table 4-2 does not contain the correction of this noticeable difference.
222 square feet (reduction) for removing
nonreplanted area to obtain cost per square foot. • New starts with automatic irrigation
cost about $1.36 per square foot.
The majority of the installations were completed
in 1998 with a few extending into 1999. For an • Irrigation system costs in retrofits
estimate of installation costs in 2005, an inflation averaged $0.14 per square foot
adjustment would need to be applied. (30 homes).
Appendix D discusses an estimate of the
• New start irrigation systems averaged
multiplication factor, 1.1543, that can be applied
$0.29 per square foot (74 homes).
to current estimates to obtain approximate 2005
installation costs. Overall, lower cost averages for the retrofits
suggest some financial benefit by having an
AVERAGE INSTALLATION COSTS existing landscape. The retrofitted homes may
have modified an existing irrigation system,
Figure 4-3 presents the average Xeriscape
kept part of their turf, or transplanted some
installation costs per square foot for each new
existing plants to save costs. Also, retrofits only
start and retrofit demonstration. The figure
occasionally needed to deal with retaining walls
indicates a relatively narrow installation cost
and drainage issues. Hardscape may already
range from $0.83 to $1.43 per square foot. The
have been in place, or could have been enhanced
figure shows a $0.60-per-square-foot difference
inexpensively.
between the costliest demonstration average in
Highlands Ranch (a new start demonstration) The Denver retrofit group did not install
and the least expensive average for the Colorado irrigation systems, yet they had higher install
Springs retrofit. It is noted that applying the costs than their retrofit counterparts with
Denver factor estimate of 7.9 percent (more automatic irrigation. The Denver landscapes
costly) for excluding pre-existing area to the were smaller than those in other
Colorado Springs and Greeley retrofit costs of demonstrations, and perhaps this caused some
Figure 4-3, would adjust their costs to $0.90 and loss of economy.
$1.06 per square foot, respectively.
The Fort Collins new start average installation
For comparisons, homes were grouped into cost of $1.34 fell within the average cost of the
three categories: new starts with automatic other new starts, despite involving large
irrigation, retrofits with automatic irrigation and properties and higher family affluence. Perhaps
the Denver retrofit group with manual homeowners in Fort Collins benefited from
irrigation. Regarding the two groups with some economy of large properties. On the other
automatic irrigation, the difference in hand, Highlands Ranch, with smaller properties,
installation cost is about $0.38 per square foot may have lost economy, yielding the highest
less costly for the retrofits. This difference is average install cost of $1.43 per square foot.
about 25 to 30 percent less expensive for
installing retrofits. Other average cost outcomes When installation costs were stratified by
include: considering only properties classified as full
Xeriscapes, differences between retrofits and
• The Colorado Springs new start cost is new starts narrowed. Data were pooled across
$0.40 per square foot more than its demonstrations for these computations.
companion retrofit (adjusted by
excluding pre-existing area). • Based on 22 properties, the retrofit
average cost per square foot was $1.26.
• The hose-drag Denver retrofit cost is
$1.12 per square foot, $0.22 more per
square foot than the Colorado Springs
retrofit (adjusted) with automatic

Chapter Four 4-3

Average Landscape Size by Demonstration

10,000
Area in Square Feet (SF)

8,000

6,000

4,000 DEN-RT = Denver retrofit

2,000 HGR-NS = Highlands Ranch new start
0 All
DEN-RT HGR-NS COS-NS COS-RT FTC-NS GRL-RT COS-RT = Colorado Springs retrofit
Demo
FTC-NS = Fort Collins new start
Area (SF) 3049 4328 4895 5506 7658 9290 5353

Figure 4-2: Average landscape size by
demonstration.

Average Xeriscape Installation Cost by
Demonstration
$2.00

$1.50
Cost ($/SF)

$1.00

$0.50

$0.00
COS- GRL- DEN- COS- FTC- HGR- DEN-RT = Denver retrofit
RT RT RT NS NS NS COS-NS = Colorado Springs new start

Average $/SF $0.83 $0.98 $1.12 $1.30 $1.34 $1.43 FTC-NS = Fort Collins new start

Figure 4-3: Average Xeriscape installation cost
per square foot by demonstration.

• Considering 53 properties, the new start
average cost per square foot was $1.36.
Xeriscaped, the $1.26 is considered a more
Because of the difficulty in estimating the area realistic estimate of cost per square foot for
not planted for YARDX in properties, estimates retrofits. This retrofit cost and the $1.36 per
given in Figure 4-3 for retrofits are smaller than square foot for new starts show minor cost
the $1.26 given above estimated from full differences. This outcome is plausible and not
Xeriscapes. Because of the estimation difficulty surprising.
for those properties not fully (newly)


The project’s range of average Xeriscape • Twenty-seven percent of homeowners
installation costs is considered relatively performed all the work (38 homes).

Costs were determined for extraordinary • Sixty-eight percent of homeowners
landscape items for 40 participants who hired contractors to do some of the work
incurred these costs. These numbers show that (98 homes).
retrofits spent almost double the average
amount spent by the new start owners The average installation cost per square foot was
(87 percent more). Nevertheless, these numbers calculated for the above three labor breakdowns:
provide useful guidance. mostly contractor installed, homeowner
installed, and mixed labor installed.
• Considering retrofits, the average Considering only nonextraordinary installation
extraordinary Xeriscape expenses per costs, contractor-installed landscapes were most
home were $1,432 (12 homes). expensive, followed by mixed labor, followed by
homeowner installed, as would be expected.
• Considering new starts, the average Costs increased approximately 85 percent per
extraordinary Xeriscape expenses per square foot from homeowner to mixed labor and
home were $768 (28 homes). about a 66-percent increase from mixed labor to
mostly contractor-provided labor. Costs nearly
Over twice the number of new start homes
tripled from homeowner-provided labor to
incurred extraordinary costs than did the
mostly contractor labor.
retrofits. Presumably, many of the retrofits had
already addressed these issues in their pre- • Landscapes installed mostly by
existing landscape. There is the possibility that contractors cost $2.16 per square foot
sample size effects of 12 homes may have partly (7 homes).
led to the higher value of $1,432. Also, the
retrofit yards tended to be larger than the new • Mixed-labor-installed landscapes cost
starts, possibly increasing costs for more $1.30 per square foot (98 homes).
materials.
• Homeowner-installed landscapes cost
INSTALLATION LABOR $0.70 per square foot (38 homes).

The cost figures in Figure 4-3 and of Further analysis of homeowner labor input can
extraordinary costs discussed do not include a be seen in Figure 4-4. This figure shows the
cost estimate for homeowner labor input to relative homeowner labor contribution to
installations (but do include any out-of–pocket landscape construction and irrigation system
labor costs). installation, stratified by demonstration retrofit
and new starts. The chart statistics represent
The relatively modest cost per square foot may only installations by the homeowner. Because
be due to substantial homeowner labor the data would not allow separation of
contribution to installation. Labor costs were contractor labor hours from total contractor
analyzed when homeowner receipts allowed costs, properties were selected without
categorization. Labor costs were designated as contractor costs to estimate the number of hours
mixed if (1) the homeowner hours or contractor to install a Xeriscape.
hours and cost information were not discernable
from the paperwork, or (2) the contractor costs • New start homeowners spent about
were less than 75 percent of the total installation 40-percent less time on construction of
cost and no homeowner hours were logged. landscapes than retrofits.

• Five percent of homeowners hired
contractors to do 75 percent or more of
the installation (7 homes).

Chapter Four 4-5

calculate the average installation cost per square
Home Labor Average
foot.
237
250
• The retrofits spent an average of
59 hours per 1000 square feet to install
200 their Xeriscape (18 homes).

150
143
Retrofits
• The new starts spent an average of
Hours

New Starts
50 hours per 1000 square feet to install
100 their Xeriscape (11 homes).

45 48 These numbers provide useful estimates of
50
homeowner Xeriscape installation time.

0
Construction Irrigation
COSTS OF PLANTS AND HARDSCAPE
The installation costs were stratified into
11 major landscaping categories. Two categories
Figure 4-4: Installation hours by
homeowner labor only. consisted of plants and hardscape. The average
cost per square foot was determined for each
• New start owners spent an average category, by retrofit and new start stratum. The
143 hours per property in construction plant calculations included all types of plants
(11 homes). (including turf).

• Retrofit owners averaged 237 hours per Documenting cost per square foot was the
property in construction (27 homes). desired goal. To calculate average cost of plants
per square foot, the project team subtracted
• Retrofits (18 homes) and new starts from the total landscape area these components:
(11 homes) spent about the same the area covered by pure mulch, all hardscape,
amount of time installing their and any dryland. The averages were computed
automatic irrigation systems (45 and for only those homes that had identifiable
48 hours). expenses for the category of interest. Data were
pooled across demonstrations.
• The new start group spent an average of
25 percent of their total installation time Plants
on irrigation systems.
• For retrofits, the average cost was
• The retrofit group spent an average of $0.38 per square foot (65 homes).
16 percent of their total installation time
on irrigation systems. • For new starts, the average cost was
$0.49 per square foot (75 homes).
The difference in construction time between
retrofits and new starts is not easily reconciled. Hardscape
Perhaps retrofits spent more time removing the
old landscape and in the land preparation • Average cost was $1.74 per square foot
phase. Also, retrofit groups may have included (38 homes) for retrofits.
more retired homeowners with more time to
• Average cost for new starts was $4.63
spare.
per square foot (51 homes).
It is instructive to estimate the Xeriscape
On average, the cost per square foot for plants
construction time per 1000 square feet of
and hardscape was noticeably higher in the case
landscape. Estimates were made solely for the
of new starts.
homeowner-provided labor cases. The
landscape area used is that used previously to


5.CHAPTER FIVE

Water Use
Comparisons

INTRODUCTION
A prominent goal of the YARDX project was to
assess whether Xeriscape landscapes save water
over traditional landscapes. Determining this
involved comparing water use between
treatment (TR) samples and their counterpart
control (CN) samples. Comparisons were
needed for each demonstration. Cross-
demonstration comparisons were also of
interest.

Demonstration sample selection procedures
were aimed at reducing within-demonstration
variation, except for landscaping type (Xeriscape
or high-water-use turf). Cross-demonstration
differences involved different settings of some
demonstration characteristics, such as soil type.
Within-demonstration comparisons of TR and
CN samples, or within-sample comparisons
between individual landscapes, were less likely
to have water difference estimates masked by
high variability. Impacts on water use
assessment caused by differing weather, soil
type, and family affluence were lessened by the
substantial efforts expended by YARDX to enlist
control properties with characteristics similar to
the companion Xeriscape yards.

Efforts were made to minimize companion
sample differences in several subtler variables,
such as landscape exposure (effects of wind),
shadiness, and area of landscape on steep slopes
(15 degrees and steeper). These data were
collected on all YARDX properties.

Several water use relationships were also of
interest. These included water use amount
versus Xeriscape age, or whether water
application per unit area of landscape seemed to

Chapter Five 5-1

differ according to landscape size or family adjustments, the indoor use average was
affluence. computed from January, February, November,
and December monthly computed values
WATER DATA encompassing each growing season. The project
team felt that using the average of “encircling”
Utilities routinely collect water use information months best represented indoor use for each
monthly or bimonthly. Water meter readings growing season.
were provided to the YARDX project for
analyses. Table 2-1 shows the periods of water The water utility at Highlands Ranch offered a
data obtained by YARDX. second water meter for their demonstration
participants, to assess the outdoor water use.
The water data possessed a number of Homeowners of TR and CN landscapes installed
differences. Landscapes differed in month and the meters. For the growing seasons of 2001 and
day-of-month of initiation of irrigation. The 2002, outdoor water use readings were
differing number of properties per sample, incorporated into the database for Highlands
using water in the initial months of each Ranch. The outdoor readings more accurately
demonstration, was important. Because the determined the landscape water use than
monthly water use per square foot was the basic winter-corrected house meter readings. The
assessment of interest, the initial month was data were incorporated with winter corrected
eliminated for a property when more than the data for the 1998 through 2000 growing seasons.
initial 4 days (of the month) had missing data. If As the number of participants in the Highlands
the number of properties in the initial months of Ranch demonstration was less than the desired
incomplete samples (CN and TR) of a 30, the more accurate summer data for 2 years
demonstration, differed by more than three seemed to somewhat offset the impact on data
properties beyond full sample difference, variability caused by the smaller data samples.
months were truncated until samples were
within three properties difference of their full Yearly winter water use for comparisons
sample configuration. between samples was estimated by the average
of monthly estimated values for January
Meter readings occurred on differing days of through March and November and December.
months. Consequently, average water use per In making winter use comparisons, the water
day was computed for each metering period, data were not divided by landscape area,
using the water use total and the day count per because outdoor watering was not involved. In
period. Water use rates were used to compute YARDX demonstrations, March can be a
monthly total use. Monthly water use totals transition month, during which minor outdoor
were then divided by landscapable area to watering may occur in some years. In the Front
obtain monthly water use per square foot. Range, March is, on average, a snowy month.
These values became the basic data analyses
unit for conducting TR and CN sample Other adjustments were applied to water use
comparisons. data samples as necessary. Data truncation
occurred when a participating property was
Generally, water use data were available from sold and the new owner altered the landscape
mid-1998 or early 1999 through 2002. The with revisions no longer in character with the
growing season months were selected as April particular demonstration. In a few cases,
through October. Estimation of growing season owners abandoned care of the landscape. The
water use, with the exception of the Highlands level of landscape revisions generally
Ranch demonstration, had to be derived from considered excessive for YARDX purposes was
readings from the single utility-provided home established at 30 percent of area of the water use
water meter. The home meter registers indoor zone in question. Such revision would likely
plus outdoor water use. Outdoor monthly impact landscape installation costs, maintenance
water use was estimated by subtracting the and water use. Water use data sets were
average winter monthly use from growing truncated at the estimated date of revisions.
season monthly computed values. For these

5-2 Water Use Comparisons

WATER USE COMPARISONS 6

Water Use Per Unit Area (gal/SF)
Water use data for an individual homeowner 5
typically consisted of 45 to 65 monthly values,
depending on when landscape irrigation and 4
meter readings began. Thus, the database for a
3
demonstration sample of 25 participants could
contain over 1200 monthly water use readings. 2

Pooling the monthly values for TR and for CN 1 DEMO
separately, and computing basic statistics for RT-CN
each group, indicated the data to be positively 0 RT-TR
1000 2000 3000 4000 5000 6000
skewed. In such circumstances, averaging may Property Area (SF)
provide an unrealistically large estimate of
centrality. An alternative estimate, the median, Figure 5-1: Denver water use per unit
is often preferable. Consequently, the medians area (ft2) versus landscape area in
of the pooled samples are used here for Denver. (RT-CN = Retrofit Control, RT-TR
comparisons. = Retrofit Treatment).
Figure 5-1 presents a plot of Denver homeowner • Control median water use ranged from a low
average monthly water use during summer, of 1.8 gal/ft2 in the Denver and Colorado
versus landscapable area, over all growing Springs retrofits, to a high of 2.9 gal/ft2 in
the Greeley demonstration of 4 control
seasons. This figure shows the within-
properties.
demonstration variation of Denver participant
water use with corresponding landscapable • The new start CN traditional landscapes
area. The curve represents the application of the used more water than the established older
Lowess (Cleveland, 1979, 1981) moving-window CN landscapes selected for the retrofits.
smoother. The figure is an example of how • Xeriscape median water use ranged from
water use may vary with landscape area. The 0.8 gal/ft2 at the Highlands Ranch new start
plot shows the tendency of higher water demonstration to 1.9 gal/ft2 at the
application in the smaller landscapes regardless Arvada/Wheat Ridge pre-existing
of CN or TR properties. Computing water use demonstration.
per unit area for comparisons was aimed at •
TR median water use was consistently lower
removing some effects of differences in watering than the corresponding CN use, except in the
area. Demonstration sample average property case of the Arvada/Wheat Ridge pre-
area is given in Table 5-1. existing study.
Table 5-1 gives percentages of growing season
Figure 5-2 presents table values and a pictorial water use savings using sample group, median
comparison of growing season median monthly monthly water use computed per square foot of
water use in gal/ft2, applied in outdoor use for landscapable area, using growing season values
the TR and CN groups for each demonstration. adjusted for winter. Figure 5-2 presents a
pictorial of the growing season results. Sample
The median is obtained from pooled April
medians were computed from all seasonal
through October monthly, indoor-adjusted
monthly values from all properties per sample.
water use for all years of data. The water use
units are monthly values, so the median For the 2002 growing season, Xeriscapes in the
represents the middle monthly value (similar to Colorado Springs demonstrations indicated
the average monthly use during the growing greater seasonal water use in Xeriscapes than
season). The results, either evident or respective control properties. As a consequence,
suggested, in Figure 5-2 are: year 2002 single season results, and project
period results without 2002 data were separately

Chapter Five 5-3

DEMONSTRATION SAMPLE GROWING SEASON
LANDSCAPE WATER USE

WATER USE (gal/sq ft)
MEDIAN MONTHLY 4.00

3.00
CN
2.00
TR
1.00
0.00
ARVWH-
DEN-RT COS-RT GRL-RT COS-NS FTC-NS HGR-NS
PRE

CN 1.77 1.85 2.86 2.10 2.44 2.46 1.91
TR 1.28 1.42 1.31 1.73 1.56 0.85 1.90

SAMPLE

Figure 5-2: Water Use for the demonstration sample.

computed and are given in Table 5-1 and homes) and differences versus controls were
Figure 5-3. large, the versatile, nonparametric, Wilcoxon
(1947) two-sample rank test was applied to
Plausible explanations for the Colorado Springs determine P-values.
year 2002 results include unrestricted watering
regulation of Xeriscapes during the drought Because the probability of detection of
year. Xeriscape owners apparently hand differences in small samples of water use data is
watered their plants or used their drip irrigation undesirably small, P-values were not calculated
systems allowable under the restrictions. for 2002 data solely, or for Greeley data
Watering turf was more restricted and thus comparisons.
affected the controls more severely.
Computation of water use savings between
Table 5-1 and Figure 5-4 present winter median Xeriscape samples and respective control
daily water use in gallons per day, computed properties ranged from 18 percent savings in the
from winter months January, February, March, Colorado Springs new start demonstration
November, and December. These values are not (28 percent in data without year 2002 water use
adjusted for differences in landscape area, as values) to 63 percent water savings in the
outdoor watering is very low during winter. Highlands Ranch new start demonstration. All
Figure 5-4 indicates somewhat higher control comparisons tested with the Wilcoxon method
winter use, but only minor treatment-control yielded P-values less than 0.01. The retrofit
differences except for Arvada / Wheat Ridge demonstration in Colorado Springs yielded a
and Highlands Ranch. water savings of 23 percent (32 percent
excluding year 2002 data) while the Denver
Table 5-1 also contains P-values, the probability retrofit demonstration (hose drag) produced a
that sample differences could have occurred by 28-percent savings. The new start
chance. The generally accepted significance demonstration in Fort Collins, consisting of
value range is 0.05 or smaller for so-called large area, and more expensive properties,
significant difference. Because samples were produced a 36-percent savings.
large (except for 2002 water use examined
separately and data from too few Greeley


Table 5–1: Water use comparisons between Xeriscape and traditional landscape samples.

Winter1 median daily
Seasonal (Apr-Oct) water use Seasonal water use (gal/day)
savings (%) per square foot of savings
N=number of properties
landscapable area, using without
growing season values adjusted adjustment Landscape average area
for winter1 average use2 for area (A) in square feet
All Years
project except Year 2002 Control
Demonstration
years 2002 only All years Xeriscapes group
Colorado Springs 18 28 50 22 savings 169 (N=27) 182 (N=37)
savings savings increase P<0.01 A=4608 A=5190
new start P<0.01 P<0.01
Colorado Springs 23 32 34 24 savings 194 (N=27) 191 (N=32)
savings savings increase P<0.01 A=5374 A=5472
retrofit P<0.01 P<0.01
Denver retrofit 28 26 33 17 savings 146 (N=34) 152 (N=31)
savings savings savings P<0.01 A=3047 A=2537
P<0.01 P<0.01
Fort Collins new start 36 39 29 32 savings 191 (N=34) 182 (N=38)
savings savings savings P<0.01 A=7583 A=8859
P<0.01 P<0.01
Highlands Ranch new 63 69 63 50 savings 165 (N=13) 218 (N=17)
start savings savings savings P<0.01 A=3971 A=3576
P<0.01 P<0.01
Greeley retrofit 54 56 30 40 savings 130 (N=6) 185 (N=4)
savings savings savings A=9290 A=9348
Wheat Ridge & Arvada 1 savings 1 3 savings 25 savings 164 (N=26) 198 (N=28)
pre-existing Xeriscapes P>0.05 savings A=4068 A=5558
P>0.05
1 Winter months included January, February, March, November and December.
2P-values (probability that sample differences could have occurred by chance) are not given for small
samples.

The water savings in the Arvada/Wheat Ridge The retrofit demonstrations did not achieve
demonstration was less than 1 percent (not water savings levels consistent with expected
statistically significant). The limited data savings from the Xeriscape design developed for
samples from Greeley suggested water savings retrofits. This design called for Xeriscape plant
of 54 percent. The estimated high water savings area according to ⅓ low (or no), ⅓ moderate,
of 63 percent in the Highlands Ranch and ⅓ high water use, and overall water savings
demonstration appeared out of range with the of about 60 percent. Water savings may have
other new start comparisons. The results could been negatively impacted by some participants
have been somewhat influenced by the employing greater than ⅓ high water use area
occurrence of an imbalance in samples of zero (of total landscape).
water monthly values possibly caused by the
drought. Determination of the possible
influence of this feature called for in-depth
investigation and analysis outside the scope of
this study.

Chapter Five 5-5

WATER USE SAVINGS

80
60
Savings (%)

40
20
0
-20
-40 ARW-PRE = Arvada/Wheat Ridge
-60 COS-NS = Colorado Springs new start
ARW- COS- COS- DEN- FTC- GRL- HGR-
PRE NS RT RT NS RT NS
DEN-RT = Denver retrofit
All Years 1 18 23 28 36 54 63
FTC-NS = Fort Collins new start
Non-Drought Years 1 28 32 26 39 56 69 GRL-RT = Greeley combined new start
& retrofit
Drought Year 2002 3 -50 -34 33 29 30 63
Demonstration

Figure 5-3: Demonstration water use savings by Xeriscape group
over respective control group.

WINTER WATER USE
250
Water Use (gal/day)

Treat ment s
Controls

200
150
ARW-PRE = Arvada/Wheat Ridge
100
50
0 DEN-RT = Denver retrofit
ARW-
COS-NS COS-RT DEN-RT FTC-NS GRL-RT HGR-NS FTC-NS = Fort Collins new start
PRE
GRL-RT = Greeley combined new start
Treatments 164 169 194 146 191 130 165 & retrofit
Controls 198 182 191 152 182 185 218 HGR-NS = Highlands Ranch new start

Demonstration

Figure 5-4: Winter water use.


The apparent lack of water savings in the 4
Arvada/Wheat Ridge demonstration was

surprising. These landscapes were estimated to
3
be 5 years or older than landscapes of other
YARDX demonstrations. Lack of water savings
suggests savings may eventually decrease with 2
landscape aging, unless proper maintenance and
homeowner education is maintained. Plots of
1
monthly water use with time, of other YARDX DEMO
Xeriscape samples did not indicate a reduction RT-CN
in savings with time over the duration of 0
10 20 30 40 50 60 70 80
RT-TR
YARDX. Figures 5-5 and 5-6 show, for the Month
Denver and Colorado Springs demonstrations,
the water use per square foot over time, month Figure 5-5: Denver water use in time for April
by month. The Lowess curves for water use in 1999 – January 2002. (MONTH = Months since
start of data [April 1999], RT-CN = Retrofit Control,
Denver show the water savings (amount of
RT-TR = Retrofit Treatment).
separation of curves) and the downturn in water
use in 2002 by both TR and CN. The Lowess 4
curves for Colorado Springs present the pattern

of higher water use by TR in 2002.
3

Possible causes for nil (1%, statistically
insignificant) water savings in the Arvada / 2
Wheat Ridge demonstration include inadequate
thinning of plants, occurrence of increasing
plant root structure that required additional 1
DEMO
watering, and lack of knowledge by new owners RT-CN
on proper maintenance. Such detriments can be 0 RT-TR
20 30 40 50 60 70 80
overcome by careful selection of plants upon Month
installation, proper landscape maintenance, and
the continuation of education programs on Figure 5-6: Colorado Springs water use in time
water-conserving landscaping. for April 1999 – January 2002. (MONTH = Months
since start of data [April 1999], RT-CN = Retrofit
Control, RT-TR = Retrofit Treatment).
PRECIPITATION
years is about one-third of the growing season
It is well known that precipitation usually
requirement of Kentucky bluegrass. Average
impacts landscape irrigation, particularly in
growing season precipitation varies from about
properties using manual irrigation. Often,
9 inches in Arvada and Greeley to about
automated system properties do not benefit as
13.5 inches in Colorado Springs, a relatively
well, because most irrigation systems do not
narrow range. Precipitation extremes of 1999
include a rain sensor that can reduce irrigation
and 2002 and water restrictions due to drought
when rainfall occurs, or property owners fail to
influenced irrigation water demand.
temporarily shut off automated systems.
Figure 5-7 presents annual rainfall at locations of Without question, the YARDX Xeriscapes saved
the YARDX demonstrations (or nearby as in water over their counterpart controls. These
Cherry Creek and Lakewood). The figure results suggest water savings in the 30s
indicates 1999 as the wettest year with most percentage should be readily achievable if
locations receiving 20 inches of precipitation or Xeriscapes are properly planned, installed, and
more (as in Colorado Springs). The driest year maintained. New start water savings appeared
was 2002 with all locations receiving less than greater than those for retrofits, but attention to
10 inches. The difference between these extreme amount of turf in retrofits and appropriate
irrigation could overcome this difference.

Chapter Five 5-7

Annual Precipitation at Demostration Sites
Wheat Ridge Cherry Crk ColoSprgs Denver Ft Collins Greeley Lakewood

30

25

20
Precipitation (in)

15

10

5

0
1998 1999 2000 2001 2002
Year

Figure 5-7: Annual precipitation at demonstration projects.


6.CHAPTER SIX

Maintenance
Costs

INTRODUCTION
Estimation of landscape annual maintenance
costs was one of three major goals of YARDX.
Whereas installation costs were incurred up
front, maintenance costs accrued over the
duration of the project. YARDX did not have
adequate funding of rebates to encourage
tracking and periodic submission of costs.
YARDX developed the forms shown in
Appendix E for participants to tabulate costs
and submit information after each growing
season, November through February. Control
participants were also provided cost data forms
and encouraged to log and provide maintenance
information. The control form was identical to
that for treatments, other than the word
“YARDX” was removed. Again, control
participants were not told their information
would be compared to that of Xeriscape samples
to lessen the possibility of influencing control
water use. YARDX mailed a second request for
information in January to those homeowners
who had not yet responded.

DATA QUALITY
Overall, response from Xeriscape samples for
maintenance cost information was less than
desired. Xeriscape participants had a 46-percent
response rate, while controls managed only a
23-percent response. In combination, the
response rate was 34 percent.

Reporting of maintenance information during
the project was erratic. Participants reported at
some times and not others. Some never turned
in a report. Very few turned in all requested
reports.

Chapter Six 6-1

MAINTENANCE COST As a final adjustment, maintenance surveys
were eliminated that came from participants
CALCULATIONS
who were dropped from the project. Drops
As with installation costs, it was desirable to generally occurred because of noncompliance
factor out the disparity in landscape area with project guidelines, such as property
differences by estimating the annual conversion to a rental, or substantial alteration
maintenance cost per unit area. Participant of the landscape or irrigation type. There were
landscape area was reduced by dryland area 12 drops in 7 demonstrations.
and “pure mulch” area (generally a rock area
without vegetation and irrigation). Little or no To calculate the demonstration’s average annual
maintenance is applied to dryland areas, maintenance cost per square foot, the following
especially if a fabric weed barrier is installed steps were performed:
(recall that dryland impacted only one retrofit
• A participant’s annual maintenance cost
home in Colorado Springs). The pure mulch
per adjusted square foot was
area was subtracted, because minimal or no
determined for each year reported. A
maintenance is required, especially during the
time average of interest, such as the first
first several years, and pure mulch areas are
3 years and the last 3 years, was
sometimes quite sizeable (as in area between
obtained by averaging over the years
houses), which could improperly distort average
within the specific time period of
annual maintenance cost per square foot.
interest. Two time periods of interest,
Remaining landscape area was that area
1997-1999 and 2000-2002, were
requiring routine maintenance and irrigation.
individually averaged to obtain two
Table 5-1 gives the adjusted average area per
values for each participant. These two
demonstration for comparison with average
periods allowed comparison of plant
areas of Figure 4-2 that include pure mulch.
establishment maintenance versus the
To accurately calculate all maintenance, an mature plant phase.
accounting procedure of time spent was
• A group or demonstration sample
required. After some discussion with YARDX
average was obtained by averaging over
utility representatives, the project team decided
the individual time average that was
to apply a rate of $18 per hour for noncontract
each “report weighted.” The individual
labor provided by homeowners. The $18 rate
weights were determined by dividing
was intermediate between an unskilled labor
an individual’s number of reports in a
rate and a professional contractor rate.
period of interest by the total number of
All maintenance reports were scanned for reports for all in that period. The
comments that suggested an inappropriate “report weighted” averages for a
entry. Occasionally, an expense was claimed demonstration sample were summed
that was clearly a capital expenditure, such as a over the individuals in that group to
new fence. These were removed when they obtain the (weighted) average annual
could be clearly identified. Also filtered out maintenance cost per square foot per
were installation costs, mostly recorded in the demonstration.
early years (1997-1999) while many homeowners
No adjustments were made for inflation.
were still completing their landscapes. Only
12 homes were found to have apparent
installation costs recorded on their maintenance MAINTENANCE COST RESULTS
reports. Most of these costs were either already Maintenance data were stratified according to
included in the installation costs or were capital the periods 1997-1999 and 2000-2002 to examine
expenditures that occurred after installations whether maintenance costs increased from the
were completed and thus were not routine plant establishment period to the mature plant
maintenance items. phase. Figure 6-1 presents average annual
maintenance costs for each demonstration,

6-2 Maintenance Costs

stratified by timeframe. It should be noted that These results show that treatment maintenance
the Greeley demonstration consisted of too few costs were lower than those for traditional
participants to conduct desired data landscapes during the establishment period, but
comparisons. become comparable or greater thereafter. In
three of the treatment demonstrations (both
The following discusses the results given in Colorado Springs groups and Fort Collins), the
Figure 6-1: maintenance costs increased from the first
period to the second period, possibly reflecting
• Treatment average annual maintenance
an increasing need for weeding and thinning of
costs other than Greeley ran $0.34 to
those Xeriscapes. Only one control group
$1.33 per square foot, yielding a range of
showed an increase (Highlands Ranch).
about $1.00.
There are two other interesting trends of note.
• Controls average annual maintenance
First, the older, pre-existing Xeriscapes in
costs other than Greeley ran $0.27 to
Arvada/ Wheat Ridge tended to have relatively
$1.44 per square foot for a range of
high average annual maintenance costs
$1.17.
compared to most other treatment groups with
• During the first time period (1997 – the exception of Denver. The Denver Xeriscapes
1999), five of six treatment and controls had the highest maintenance costs
demonstrations yielded average annual of any group during the early and later years.
maintenance costs less than their control The Denver treatment group had high
counterparts. environmental orientation as indicated in Figure
7-2. Denver treatments also indicated spending
• In the latter time period (2000 – 2002), more time on maintaining their Xeriscapes over
three of seven treatment demonstrations previous traditional landscapes (Appendix G,
yielded average annual maintenance question 20).
costs less than their control
counterparts. Two demonstration design features were shared
by the Denver and Arvada/Wheat Ridge
On the other hand, comparing the treatment and demonstrations: “hose drag” (manual)
control groups within their own peer group watering, and older landscapes. Denver was the
(treatment groups, control groups), yielded the only hose drag watering system demonstration
following: in YARDX. Both treatments and controls had to
manually water their landscapes. In the
• During the second time period (2000 – Arvada/Wheat Ridge study, the project team
2002), four out of seven treatments accepted whatever type of watering system was
experienced a decline in their average in place. A few homes had manual systems,
annual maintenance costs from the first while a larger number had automated spray
period. This is a fairly likely outcome if heads in their turf, but hand watered their
in fact no difference exists between flower beds and vegetable gardens. Still others
period maintenance (similar to had fully automated systems for their entire
obtaining four heads in seven tosses of a landscape. After reviewing their installation
fair coin). audits, about 50 percent (14 out of 28) of the
Arvada/Wheat Ridge controls had some aspect
• Control demonstrations experienced a
of manual watering, as did the Xeriscapes (13
decline in annual maintenance costs in
out of 26). However, high annual maintenance
five of six demonstrations, from the first
costs are only seen in the Arvada/Wheat Ridge
period through the second period. This
treatment group.
is a far less likely outcome if controls do
not, in fact, experience declining The Arvada/Wheat Ridge controls had costs
maintenance after plant establishment. comparable to other control demonstrations in
the study. In fact, they had lower maintenance

Chapter Six 6-3

6-2
Demonstration Annual Maintenance Cost

$1.60
CN=Control

Maintenance Costs
ARW=Arvada/Wheat Ridge
$1.40 TR=Treatment COS=Colorado Springs
NS=New Start DEN=Denver
RT=Retrofit FTC=Ft. Collins/Loveland
$1.20 PRE=Preexisting GRL=Greeley
HGR=Highlands Ranch
$1.00

$0.80

Cost / SF
$0.60

$0.40

$0.20

$-
ARW ARW COS COS COS COS DEN DEN FTC- FTC- GRL- GRL- HGR- HGR-
PRE-CN PRE-TR NS-CN NS-TR RT-CN RT-TR RT-CN RT-TR NS-CN NS-TR RT-CN RT-TR NS-CN NS-TR
Cost for 1997-1999 $0.45 $1.19 $0.52 $0.45 $0.58 $0.38 $1.44 $1.33 $0.37 $0.34 $- $0.54 $0.62 $0.54
Cost for 2000-2002 $0.37 $0.95 $0.44 $0.60 $0.54 $0.46 $1.27 $1.04 $0.27 $0.44 $0.21 $0.33 $0.81 $0.51
Reports 1997-1999 10 22 14 24 14 33 17 47 14 38 0 7 9 14
Reports 2000-2002 24 28 25 37 23 44 17 61 26 33 5 8 16 12

Figure 6-1. Demonstration maintenance cost and number of maintenance reports from all homeowners per each demonstration

costs than the Colorado Springs retrofit controls each fall to this same group of Xeriscape
(also with somewhat older landscapes but with homeowners.
automated irrigation) and several of the new
start control groups as well (Colorado Springs Three maintenance seminars were held in
and Highlands Ranch). YARDX municipalities after landscape
installation. One seminar was held in Fort
It is not clear why there are higher maintenance Collins for the Fort Collins and Greeley
costs in only the pre-existing treatment group. homeowners, a second one was presented in
One possibility is that the care of older Denver for the Denver and Highlands Ranch
Xeriscapes takes more time. Perhaps, more groups, and one was held in Colorado Springs
weeding and thinning of plants, transplanting, for the two demonstrations there. Overall, the
and pruning add up to more maintenance. This maintenance seminars were poorly attended.
could have also been a contributing factor in the
Denver demonstration’s high costs, since the The seminars lasted approximately 2 hours and
landscapes were older and some pre-existing covered topics such as lawn care, irrigation,
landscape was left unchanged. However, in the trees and shrubs, and perennials. Homeowners
Colorado Springs retrofit demonstration, were also given handouts, including those
homeowners did not report these high compiled by the Colorado State University
maintenance costs. Further study may be able to (CSU) Cooperative Extension on various
shed light on these discrepancies. landscape-related subjects. The local CSU
Cooperative Extension phone numbers were
Yet another possibility is that data quality may also made available for future homeowner
have played a role in the maintenance cost questions.
results of Denver and Arvada/Wheat Ridge. It
is possible that some hose drag homeowners The annual newsletter included an update on
recorded the time their sprinklers operated the project, but most importantly, landscape
rather than their personal time in accomplishing maintenance tips. These tips were focused on
the watering. One may speculate that problems that the field auditors saw during
homeowners willing to install Xeriscape as a their site visits. A sample newsletter can be
retrofit or those who installed Xeriscape years found in Appendix I.
ago, tended to be more environmentally
No discernible trend in maintenance costs was
conscious and more prone to enjoy gardening.
apparent between the YARDX-trained Xeriscape
Cross-referencing results here to the Final
participants and the untrained Arvada/Wheat
Survey outcomes are discussed in Chapter
Ridge Xeriscape homeowners. Several items
Seven. Study of homeowner maintenance labor
may have contributed to these indistinct results
hours, versus contractor time, versus material
on maintenance training. The training effect
costs compared with Final Survey responses
signal may be small and be easily confounded
may shed more light on environmental
by other influential factors. How many
inclination.
participants read the newsletters and/or
training materials is unknown.
MAINTENANCE EDUCATION
The person(s) actually performing the
Two types of maintenance education were
maintenance may have had little exposure to
conducted throughout the study timeframe.
education on maintenance. The poor attendance
First, maintenance seminars were held for all the
at the YARDX maintenance seminars suggests
Xeriscape homeowners except those in the
less than desired training was absorbed. Lastly,
Arvada/Wheat Ridge pre-existing study.
response to the maintenance information
Secondly, an annual newsletter was sent out
requests was low.

Chapter Six 6-3

7.CHAPTER SEVEN

Final Survey

INTRODUCTION
A Final Survey was mailed to all participants of
the YARDX project. The survey was conducted
during the October 2002 to February 2003
period. Its purpose was to sample participants'
attitudes toward Xeriscape, the environment in
general, and water conservation at project end.
The Final Survey developed for controls did not
include questions relating to Xeriscape
landscapes (installation costs, maintenance,
training, rebates, etc.). Also, the control form
did not reference “YARDX”. In the Final Survey
for pre-existing Xeriscapes, two questions were
eliminated regarding project rebates and
instructional seminars in which they did not
participate.

Figure 7-1 presents participant response rates to
the Final Survey. Response rates varied from 16
percent to 58 percent of total participants in the
various sample groups. These overall results
excluded the Greeley control group, which was
very small and replied 100 percent to the Final
Survey.

Generally, the controls were less faithful at
responding to the Final Survey. The controls'
response rates were lower than their treatment
group in all demonstrations except two, Greeley
and Highlands Ranch. The overall response rate
for all the controls in YARDX was 33 percent,
and for the treatments it was 50 percent.

Summarized Final Survey results for all
demonstrations can be found in Appendix G.

Questions 1 through 8 included queries on
water-saving appliances and number of people
living in the home, etc., which were not used to
gauge attitudes but did assist in the property's
water use analysis. These eight questions were
sent to all participants annually. Questions 28

Chapter Seven 7-1

Percent Response to Final Survey
120%
Treatments
Survey Responses 100%
Controls
80%
60%
40% AWR=Arvada / Wheat Ridge
20% COS-NS=Colorado Springs, New Start
0% COS-RT=Colorado Springs, Retrofit
COS- COS-
AWR DEN FTC GRL HGR Total DEN=Denver
NS RT
58% 41% 56% 53% 47% 50% 44% 50%
FTC=Fort Collins
Treatments
Controls 50% 16% 19% 26% 37% 100% 53% 33% GRL=Greeley
Demonstration HGR=Highlands Ranch

Figure 7-1: Response rate to Final Survey of all participants per sample group.

and 30 requested verbal responses. Responses prioritize their responses to question 10, each
are given in Appendix G. Survey question 30 selected choice was given equal value.
requested that participants provide comments
that they considered helpful to improving water Figure 7-2 presents the estimated environmental
conservation through lawn and landscaping characterization of participants by sample
practices. group. Environmental orientation was
determined by reviewing responses to survey
The percentage response rates noted for each questions 10, 19, 22, and 25 through 29. For each
question relate to the number of homeowners participant, eight possible points were assigned
who responded to the survey question, as a (nine points for the new start and retrofit
percentage of all who responded to a question treatments) from responses to the eight
(not a percentage of all study participants). questions to determine an environmental score
Homeowners who were dropped retroactively computed as a percentage of the maximum eight
to the beginning of the study were deleted from points. A pro-environmental orientation was
the survey results. Homeowners who had at determined if the score was 0.50 or greater.
least partial involvement in the project were Figure 7-2 presents the number of participants of
kept in the Final Survey. the responding total whose scores were 0.50 or
greater. This procedure was applied to all
On scoring the Final Survey, each response was responders.
worth one point, except in the case of question
10 (see Appendix F). Question 10 queried Figure 7-2 shows that the number of responders
participants on why they chose to join YARDX. varied from 3 at Greeley to 18 for the Denver
If a participant selected multiple answers where Xeriscape group. Overall, pro-environmental
only one was supposed to be designated, scores (listed as percentages in the figure) varied
fractional values of equal weighting were from 42.9 to 100 percent.
assigned to each answer so their total equaled
one. For question 10, multiple answers could be Only for the Fort Collins controls and the
selected and prioritized as to the level of Highlands Ranch treatments were scores below
importance to the responder. A percentage 50 percent. Clearly, in this scoring system, there
response for each priority was calculated based was a strong pro-environmental bent in
on the number of selections of that level divided treatment and control participants.
by the number of participants who responded to
This should not be surprising, given that the
this question. If the homeowner did not
Final Survey was given in the end of 2002, a
severe drought year.

7-2 Final Survey

Pro-Environmental Responses
ARW=Arvada/Wheat Ridge
COS=Colorado Springs
DEN=Denver
100%
FTC=Ft. Collins/Loveland
GRL=Greeley
HGR=Highlands Ranch

75%
CN=Control
TR=Treatment
NS=New Start
RT=Retrofit
PRE=Pre-Existing 50%

than 50%
Response (%)
Pro-Environmental

Fraction with Greater
25%

0%
ARW ARW
COS COS COS COS DEN DEN FTC- FTC- GRL- GRL- HGR- HGR-
PRE- PRE-
NS-CN NS-TR RT-CN RT-TR RT-CN RT-TR NS-CN NS-TR RT-CN RT-TR NS-CN NS-TR
CN TR

Percent 57.1% 66.7% 50.0% 63.6% 50.0% 60.0% 62.5% 88.9% 42.9% 68.8% 50.0% 100.0% 55.6% 42.9%
Number Participants With Pro-Environmental Responses 8 10 3 7 3 9 5 16 6 11 2 3 5 3
Greater than 50%
Total Number of Responding Participants 14 15 6 11 6 15 8 18 14 16 4 3 9 7

Figure 7-2: Pro-environmental responses by demonstration.

Chapter Seven
7-3

The environmental orientation of participants the rebates may have helped sway their
was important, because it could possibly impact decisions to go forward. The final answer
water savings as well as other aspects of the suggests that rebates are helpful, but not critical,
study, such as maintenance time in the yard. and they may not have to be large dollar
Environmentally inclined participants might be amounts.
high water savers without Xeriscape and also
enjoy working in their landscapes. When the treatments were asked if they would
recommend Xeriscape to others based on their
Figure 7-2 results indeed suggest more current knowledge and experience, the majority
environmentally inclined participants in of each demonstration answered “definitely”
treatment groups with values of 43 to 100 (question 15). The answers ranged from 80 to
percent. All of the treatment demonstrations 100 percent, which are extremely strong
show higher pro-environmental response rates responses, implying a sound satisfaction with
than their control group except for Highlands the results of their landscape. Curiously, the
Ranch. Controls showed a lower tendency to be demonstrations on the low end of this response
pro-environmental (42 to 63 percent). Generally, were two of the three retrofits (the small Greeley
this result would cause some concern that water group excluded) and the pre-existing
savings' comparisons may have been impacted. demonstration (Colorado Springs and
Several attending factors diminish this Arvada/Wheat Ridge at 80 percent and Denver
possibility. Xeriscape training provided to at 83 percent). This could insinuate that a small
treatment participants could have increased fraction of retrofits and pre-existing participants
environmental awareness. A limited number of were less happy with the results of their
participants responded to the Final Survey, landscapes compared to what existed before.
quite possibly the more environmentally However, these satisfaction numbers, while
oriented (Colorado Springs' two control groups lower, are still very high. Although, it should be
had less than 20 percent response). noted that not one Xeriscape participant selected
the answer “not at all.” The synthesis of these
SPECIFIC SURVEY QUESTIONS results shows that homeowners were overall
quite satisfied with their landscapes.
ANALYZED
Several survey questions are analyzed in more When question 15 above is examined with
detail below. Some of these questions could be question 16, “do you like your landscape now?”
helpful to a utility trying to gauge whether a there are similar compelling, positive responses
Xeriscape rebate is helpful and whether from the homeowners with Xeriscapes. The
homeowners are pleased with their landscapes. treatment groups ranged from 61 to 88 percent
for a “very much so” answer, outperforming
Question 12 asked if the homeowner would each of their control groups. The controls only
have installed a Xeriscape without the rebate. responded 13 to 57 percent with “very much
This question was only presented to the new so.” This suggests there may be potential for
start and retrofit Xeriscape properties, because swaying those with traditional yards into
they received a project rebate. changing to a more interesting Xeriscape
landscape.
Overwhelmingly, the majority from each
demonstration answered that they would have The topic of landscape maintenance came up in
installed their Xeriscape without a rebate. The several questions. Question 18 queried, “What
percentage who responded affirmatively to this is your opinion on the cost of maintenance of
question ranged from 56 to 83 percent. your landscape?” Unanimously, the majority
Interestingly, two of the new start for each Xeriscape group said “not at all
demonstrations, Colorado Springs and expensive” with the majority in each control
Highlands Ranch, had the highest favorable group saying “moderately expensive” with one
responses (80 and 83 percent). One might read control group’s majority saying “noticeably
into this that the retrofits need a little more expensive” (Colorado Springs new start control).
prodding to rip out an existing landscape, and

7-4 Final Survey

Therefore, in each demonstration, the treatment Even though the perception of the Xeriscapes
group thought their maintenance was less was that they spent less time and money, which
expensive than did their control peer group. was generally true in the early years (1997 –
1999), it was not always true in the latter years
Question 20 asked homeowners to compare (2000 – 2002). Eighty-three percent of the
maintenance time on Xeriscape with treatment groups did indeed have lower costs
maintenance time on their previous, traditional than their control counterparts in the plant
landscapes. Only the Xeriscape homeowners establishment period, but 57 percent of the
were given this question, and the majority for control groups actually had lower annual
each demonstration said “somewhat less” to maintenance costs than did their treatments in
“substantially less” was spent on their the maturing period. Perhaps the treatments
Xeriscapes than on a traditional landscape. The enjoyed working in their yards more than the
answers to questions 18 and 20 show strong controls and saw more positive benefits from
evidence that the Xeriscape homeowners believe their labor, so that possibly the time and money
their maintenance is less than that of a spent did not seem as costly.
traditional landscape.
In retrospect, more information could have been
The next step is to compare these maintenance obtained by interviewing all participants in
questions with the actual maintenance time and person. However, this was beyond the scope of
expense results stated previously in Chapter Six. YARDX.

Chapter Seven 7-5

8.CHAPTER EIGHT

Summary and
Conclusions

PREFACE
Metro Water Conservation Inc. (MWCI) and the
Bureau of Reclamation (Reclamation) formed a
temporary partnership in 1996 to create the
YARDX (Yield And Reliability Demonstrated in
Xeriscape) project. The main purpose of the
project was to study the benefits of water-
conserving landscaping called Xeriscape. The
primary goals of YARDX were to develop
reality-based estimates of water savings, and
installation and annual maintenance costs in
implementing Xeriscape.

The YARDX project is one of five cooperative
demonstration projects on landscape water
conservation pursued by Reclamation,
Photo: Courtesy of David Winger, Denver Water, Denver, collectively called the National Xeriscape
Colorado.
Demonstration Program (NXDP). The NXDP
cooperative studies were conducted at locations
in the western United States that experience
different climate, including the Colorado Front
Range, centered at Denver, Colorado; Phoenix,
Arizona; Austin, Texas; the Las Vegas area of
southern Nevada; and Fargo, North Dakota.

The Colorado Front Range’s high population
growth, concerns over prolonged drought, and
urban landscapes accounting for 50 percent or
more of residential water use, have raised the
level of concern of meeting current and future
water demand. Water conservation offers an
alternative to new traditional water
development projects that are difficult to
execute because of economic and environmental
concerns. Water conservation through
Xeriscape may offer at least a partial solution to
the growing municipal water demand on the
Colorado Front Range. In addition, the
citizenry’s familiarity with Xeriscape made the

Chapter Eight 8-1

Front Range a likely candidate to conduct (including completion of installation by set
YARDX. date). One demonstration (Greeley) was unable
to obtain adequate numbers of participants.
Nine water utilities along the Colorado Front Another demonstration (Highlands Ranch)
Range, and their 357 single-family customers enlisted about 60 percent of desired numbers,
chose to participate in YARDX. Seven somewhat less than desired. Of Xeriscape
demonstrations of Xeriscape, using a mix of enlistees, 58 percent completed the study.
settings of eight impacting variables, were Attrition appeared mostly caused by financial
initiated and conducted for nearly 6 years. The and time constraints. Sample sizes of the
demonstration sites were located in remaining demonstrations were adequate to
Arvada/Wheat Ridge, Colorado Springs, support preferred data analysis.
Denver, Fort Collins, Greeley, and Highlands
Ranch. The demonstrations consisted of three
XERISCAPE INSTALLATION COSTS
landscape retrofits; three new starts, and one
pre-existing Xeriscape. The eight variables Participants of YARDX were required to submit
included water use, Xeriscape application receipts and a listing of hours of labor expended
(retrofit, new start, or pre-existing), Xeriscape in installing their Xeriscapes. Costs were
application level (landscape designed for 30-40 stratified by demonstration and estimated per
percent or 60-70 percent water savings), yard square foot.
size, irrigation method (manual or automatic),
family income, soil type, and precipitation. Overall, the installation costs ran $0.83 to $1.43
Over the 6 years of fieldwork (1997 – 2002), per square foot, a rather narrow range. A
YARDX developed and analyzed data on the prominent factor in costs was that participants
water savings, and installation and annual provided most of the installation labor. In
maintenance costs in implementing Xeriscape. general, homeowner-installed landscapes cost
about one-third the cost of contractor largely-
installed Xeriscapes. The Highlands Ranch
RESULTS
demonstration logged the highest cost per
square foot, but exceeded others (new starts) by
ENLISTING PARTICIPANTS only $0.10 to $0.15 per square foot. Automated
Analysis on real water use data had shown that watering systems added about $0.14 to $0.29 per
about 30 participants per sample, monitored square foot. Installation costs for new properties
over 4 growing seasons, were required to detect ran $1.36 per square foot and $1.26 for retrofits.
a 30-percent water savings. This suggested that The lower retrofit averages suggest some
seven demonstrations would require over 400 financial benefit by having an existing landscape
participants (treatments and controls). and being able to modify irrigation systems and
However, it was not easy to enlist 400 simply replant some plants.
homeowners to participate. Although
Estimates were obtained of labor hours required
homeowners (treatments, except pre-existing)
for Xeriscape installation, from properties for
were charged a $100 commitment fee up front to
which owners conducted installations. The
participate in the project, they were
labor effort to install Xeriscapes was estimated
compensated with a $300 rebate for a new
at 50 to 60 hours per 1000 square feet of
landscape or $600 for a retrofit. All homeowners
landscape, including time to install automatic
(except the control groups and the pre-existing
irrigation systems. Automatic irrigation system
demonstration) received design and educational
installation required 46 hours per property on
support, as well as discounts on Xeriscape
average.
plants and other landscape materials.

YARDX encountered substantial difficulty in WATER USE RESULTS
completing sampling needs. The main obstacles
A major goal of the YARDX project was to
appeared to be cost of installing a landscape,
estimate the water use savings of Xeriscape over
and meeting requirements of participation
traditional landscape water use. The study was

8-2 Summary and Conclusions

designed with each demonstration consisting of The Denver retrofit demonstration produced a
treatment pairs (Xeriscape/traditional 28-percent water savings. The retrofit
landscape) of similar property and landscape demonstration in Colorado Springs yielded a
features, except for landscape type. water savings of 23 percent (32 percent
excluding year 2002 data). The new start
Water utilities provided the meter readings of demonstration in Fort Collins, consisting of
participants. YARDX converted the water data large area and more expensive properties,
to monthly amounts applied per square foot of produced a 36-percent savings. The limited
landscapable area. To estimate the outdoor use, samples from Greeley suggested water savings
average winter monthly (January, February, of 54 percent. The high water savings indicated
March, November, and December) use was in the Highlands Ranch demonstration may
subtracted from each growing season (April have been caused by the occurrence of an
through October) month’s water. The imbalance in samples of some zero water use
Highlands Ranch demonstration added water monthly values.
meters to measure strictly the outdoor use. This
feature was expected to lower water data noise,
MAINTENANCE COSTS
so that fewer properties (than 20) could produce
useful assessments. The study focus was on Another goal of the YARDX project was the
comparing the growing season water use. estimation of landscape annual maintenance
costs. After each growing season, YARDX
Comparing Xeriscape samples with respective participants were asked to report their
control properties yielded water savings ranging maintenance-related expenses and time
from 18 percent in the Colorado Springs new expended. As a means of developing a cost for
start demonstration (28 percent in data without homeowner labor input, their hours were
year 2002 values), to 63 percent savings in the charged at $18 per hour. To estimate
Highlands Ranch new start demonstration. All maintenance costs over time, results were
comparisons tested with the Wilcoxon method estimated separately for the periods 1997-1999,
yielded P-values (see Glossary) less than 0.01, a and 2000-2002. The low and high average
value generally considered highly significant. annual maintenance costs per square foot of all
Xeriscape samples and separately, the control
The study period included the severe drought
samples, yielded the treatment range of $0.34
year of 2002. Front Range water utilities
per square foot to $1.33 per square foot and a
imposed watering restrictions that varied from
mid-range of $0.84, and similarly for the control
municipality to municipality. Some
the range of $0.27 per square foot to $1.44 per
municipalities allowed Xeriscape owners to
square foot and a mid-range of $0.86.
water as needed. In extreme summer conditions
similar to a 2002 drought year, traditional The Xeriscape maintenance costs generally
landscapes as well as Xeriscapes would likely tended to be lower than their controls during the
require more water than in years with normal plant establishment period (first time period),
rainfall. but somewhat higher during the plant maturing
years. This suggests that older Xeriscapes take
The demonstrations in Colorado Springs were
more work to maintain than for a traditional
most impacted by the drought and the imposed
landscape. And, the two Xeriscape
watering restrictions. During the growing
demonstrations that most dealt with hose
season of 2002, YARDX Xeriscapes in Colorado
dragging (Arvada/Wheat Ridge and Denver)
Springs used twice the amount of water used by
had the highest annual maintenance costs per
the corresponding control groups. Other
square foot, possibly indicating the extra effort
demonstrations appeared with little impact,
needed to hand water. The Denver control
except that of Arvada/Wheat Ridge. This
group that also hose dragged, had the highest
demonstration did not yield water savings,
cost of any demonstration group, suggesting the
according to calculations.
traditional landscapes took more watering
effort, perhaps needing more frequent watering.

Chapter Eight 8-3

However, the Arvada /Wheat Ridge controls true. Other customers may have other
that also did some hose dragging did not follow priorities than spending time
this pattern. maintaining their landscapes. Thus
these people may choose to keep a
There was no discernable difference in predominantly turf landscape for the
maintenance costs between those who received simplicity of maintenance. If water
education on good maintenance practices and utilities are to achieve savings from
those who did not. these customers, there must be
incentives, or perhaps disincentives,
FINAL SURVEY RESULTS stronger than current values.
At the project’s conclusion in 2002, a Final • Since so few participants attended the
Survey was mailed to all participants to sample maintenance seminars, water utilities
their attitudes toward Xeriscape, maintenance, need to find ways to make learning
landscape installation and their environmental about maintenance more enticing.
orientation, which could possibly impact their Scintillating presentations or seminars
water savings and maintenance efforts. on weekday afternoons or evenings may
The Final Survey results show a somewhat help. Perhaps a credit on the customer’s
stronger pro-environmental orientation in water bill for attending and
Xeriscape groups than control groups. demonstrating water savings over time
However, many Xeriscape participants indicated might bring in more people.
they joined the project because they needed a • Since there are few professional
landscape. landscape contractors in Colorado who
Overall, the Xeriscape groups in all maintain Xeriscapes, perhaps utilities
demonstrations were very satisfied with their could join with landscape trade
landscapes (more so than controls) and would associations in creating an incentive
recommend Xeriscape to others. The treatment program for the contractors. Otherwise,
groups thought their maintenance was less those who cannot maintain landscapes
expensive than did their control peer groups. for themselves will have strong
Homeowners with Xeriscapes thought they disincentive to change to Xeriscape.
were contributing less maintenance time and • MWCI is committed to making all the
money to their landscapes than with prior data from the study available for further
traditional landscapes. This perception, though, research.
did not completely match up with the actual
maintenance costs logged in the project. • One big question not addressed in the
YARDX study, but needing to be
REFLECTIONS addressed soon, is “How does a water
provider motivate more customers to
• For a water utility to motivate more
have Xeriscape on their properties and
customers to save water through
maintain it?” The answer to this
Xeriscape, a variety of measures may
question will be the key to water
need to be used to target different types
savings sustainable in the future.
of customers based on their values. For
Results of the YARDX approach are
example, low-income customers may
given in Table 3-2. The table suggests a
not be able to afford to change out a
YARDX success enrollment rate of 5
landscape, even with the benefit of a
percent. However, this enrollment rate
small to medium rebate. Customers
may not be transferable to other areas,
who cannot physically handle the rigors
given the changes in the inflation rate,
of maintaining a Xeriscape may not
water rates in general, and other factors.
choose to embrace it. Some water
utilities are already finding that to be

8-4 Summary and Conclusions

• Overall, the YARDX study shows that
CONCLUSIONS
Xeriscape saves water, and saves the
most water when all seven steps are • Xeriscape installations cost $0.90 to
included: $1.45 per square foot if homeowners are
willing to supply part of the labor.
o Design About 50 to 60 labor hours per
1000 square feet are required for
o Soil amendment
installation.
o Limited turf areas
• Xeriscapes save water. YARDX results
o Mulches indicated a savings from 18 to over
50 percent in comparison to traditional
o Appropriate irrigation landscapes.
o Appropriate plant selection • One demonstration on pre-existing
older landscapes indicated no water
o Appropriate maintenance
savings.
MWCI recommends that water providers and
• Xeriscape annual maintenance costs
other water-related organizations continue to
generally tended to be lower than their
look at Xeriscape as one of the many water-
control counterparts during the early
savings tools in an effective water management
establishment period, but somewhat
program.
higher during the maturing years.

RECOMMENDATIONS • Overall, Xeriscape owners in all
demonstrations were very satisfied with
• Additional investigation of the water
their landscapes, and would
savings by older Xeriscapes is needed.
recommend this type of landscaping to
• YARDX produced a rich data set. others.
Additional studies could reveal
additional beneficial information.
Current resources did not allow
additional study.

Chapter Eight 8-5

References

Cleveland, W.S. “Robust locally weighted regression and smoothing scatterplots.” Journal of the
American Statisticians Association, 74, 829-836. 1979.

Cleveland, W.S. “LOWESS: a program for smoothing scatterplots by robust locally weighted
regression.” The American Statistician, 35, 54. 1981.

Coloradoeconomy.com. June 2004. The Adams Group, Inc. U.S. Department of Commerce, Bureau
of Labor Statistics. 15 Oct 2004 <http://www.coloradoeconomy.com/data/dinflate.html>.

Cornelius, Coleman. “Fort Collins’ housing sizzles.” Denver Post. 20 Nov 1997: C1+.

Gregg, Tony, J. Curry, and C. Grigsby. Xeriscaping: Promises and Pitfalls. Report for the Texas
Water Development Board. City of Austin, 1994.

Manly, F.J. Brian. Randomization, Bootstrap and Monte Carlo Methods in Biology. 2nd ed. New
York: Chapman and Hall, 1997.

Nelson, John. "Water Saved by Single Family Xeriscapes." Preliminary Draft for AWWA National
Conference. New York, June 1994.

Romine, Jeff. Regional Economist, Denver Regional Council of Governments. “RE: Population
Figure.” E-mail to Julia Gumper. 24 July 2004.

Siebert, Trent, and J. Sebastian Sinisi. “Growing through the pains.” Denver Post. 9 Apr 2004: B1+.

Sovocool, A. Kent, and J.L. Rosales. A Five-Year Investigation into the Potential Water and Monetary
Savings of Residential Xeriscape in the Mojave Desert. Las Vegas, Nevada: Southern Nevada
Water Authority, 2003.

Stinnett, Robin. Landscape Trends and Water Use in the Phoenix Metropolitan Area. Available from
the Bureau of Reclamation, Phoenix Area Office, 2003.

Testa, A., and A. Newton. “An Evaluation of a Landscape Rebate Program.” AWWA Conserv’93
Proceedings. December, 1993.

Weatherbase.com. 2004. Canty & Associates LLC. 15 July 2004 <http://www.weatherbase.com
/weather/city.php3?c=US&s=CO&refer=>.

Wilcoxon, Frank. Individual Comparisons by Ranking Methods. Biometrics Bulletin, 1, 1947.

Winger, David. Certified Landscape Irrigation Auditor and Water Conservation Specialist, Denver
Water. Telephone interview. 15 July 2004.

References RF-1

Glossary

TERM DEFINITION
Control (CN) = Properties in the study with mostly bluegrass turf in the
landscape
Dryland = Area in a landscape that is planted in sparse, native vegetation
that exists on natural precipitation and is not irrigated
ft2 = Square feet
Hardscape = 1) Patios and decks that are not attached to the house or
attached patios and decks that have plants growing within the
structure (e.g., ground cover planted between the flagstone in
a flagstone patio)
2) Retaining walls, decorative pathways, and sitting areas.
Hardscape does not include playgrounds, dog runs, or vehicle
parking areas for purposes of YARDX calculations.
Landscapable = Area that is to be formally landscaped. It does not include
dryland area.
New start = Single-family residences in the study with a newly installed
Xeriscape landscape. At the start of the study, these properties
did not have any prior landscape.
P-value = Probability that differences in samples could have occurred by
chance.
Plants, high water-use = Plants that require moisture similar to Kentucky bluegrass turf
Plants, low water-use = Plants that require ¼ or less of the moisture of Kentucky
bluegrass turf
Plants, moderate water-use = Plants that require ½ the moisture of Kentucky bluegrass turf
Pre-existing = Single-family residences in the study with a pre-existing
had a Xeric landscape that was installed by the spring of 1996.
Pure mulch = A rock area, or infrequently a wood-chip area without
vegetation and irrigation
Retrofit = Single-family residences in the study with a newly installed
had a previous non-Xeric landscape.
Submeter = A separate water meter from the whole-house water meter
that monitors water used outdoors, primarily to irrigate the
landscape. Sometimes called an irrigation meter.
Traditional landscape = A landscape with mostly high water turf such as Kentucky
bluegrass.
Treatment (TR) = Properties in the study with Xeriscape landscaping in at least
50 percent of the landscapable area.

Glossary GL-1

TERM DEFINITION
Xeriscape = Xeriscape is a set of seven principles of sound landscaping for
water conservation. The seven principles are:

• Plan and design for water conservation and beauty
from the start.
• Create practical turf areas of manageable sizes, shapes,
and appropriate grasses.
• Select low-water-requiring plants and group plants of
similar water needs together. Then experiment to
determine how much and how often to water the
plants.
• Use soil amendments like compost or manure as
needed by the site and the type of plants used.
• Use mulches such as woodchips, to reduce
evaporation and to keep the soil cool.
• Irrigate efficiently with properly designed systems
(including hose-end equipment) and by applying the
right amount of water at the right time.
• Maintain the landscape properly by mowing,
weeding, pruning and fertilizing properly

GL-2 Glossary

Yardx: Yield and Reliability Demonstrated in Xeriscape - Colorado Water Wise

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to Yardx: Yield and Reliability Demonstrated in Xeriscape - Colorado Water Wise (20)

More from Eric851q (20)

Recently uploaded (20)

Yardx: Yield and Reliability Demonstrated in Xeriscape - Colorado Water Wise