Skip to main content

Full text of "Water Savings Resulting From Turf Removal and Irrigation Equipment Rebates"

See other formats


December 14, 2018 


WATER SAVINGS RESULTING FROM TURF REMOVAL AND IRRIGATION EQUIPMENT REBATES 

Neeta Bijoor, Ph.D., Water Conservation Specialist, Santa Clara Valley Water District 

Abstract 

The State of California and several local and regional water agencies offer rebate programs for removing 
turf and replacing it with low-water use landscapes, and/or rebate programs for replacing conventional 
irrigation equipment with high-efficiency irrigation equipment. Confirming and quantifying program 
water savings would enable agencies to more thoroughly assess the benefit provided by these 
programs. A study was conducted at the Santa Clara Valley Water District (SCVWD) to determine the 
water savings associated with turf removal and irrigation equipment rebates offered through its 
Landscape Rebate Program (LRP). The study focused on single-family residences, which comprise the 
largest LRP participant type. Water use billing data for LRP participants was obtained from 10 retailers 
within the SCVWD service area. Water use up to three years prior to installation of a new landscape or 
irrigation equipment was compared to water use after installation, for a period up to five years. The 
following rebate types were studied: (1) conversion of turf area to a landscape of low-water use species 
in conjunction with drip irrigation (LND), (2) conversion of timer-based automatic irrigation controller to 
weather based irrigation controller with a rain shut-off device (WBIC+RN), (3) conversion of 
conventional nozzles to high-efficiency nozzles (HEN), and (4) conversion of conventional nozzles and 
sprinkler bodies to high efficiency nozzles and spray bodies with pressure regulation and/or check valves 
(HEN+BOD). LND savings were marginally significant the first year following conversion, and statistically 
significant for each study year afterwards. LND average water savings incrementally increased each year 
following conversion. Annual average LND savings were on average 31 gallons per square foot per year 
(g/ft 2 /y) for years 2 to 5 when savings were significant, and were 48 g/ft 2 /y during the fifth year 
following conversion. The annual water savings for WBIC+RN were statistically significant each year 
following conversion, incrementally increased each year following conversion, and were on average 9 
g/ft 2 /y. The annual water savings for HEN were marginally significant in the first year, significant in the 
following two years, and were 1,243 g/unit/y on average. Annual savings for HEN+BOD were significant 
in the first year following conversion, and were 1,661 g/unit/y on average. This study shows that the turf 
removal and irrigation equipment rebate programs can be successful in achieving water conservation 
goals. 

Introduction 

California experienced the most severe drought in the past 1,200 years from 2012-2014 (Griffin and 
Anchukaitis, 2014), and this drought continued until 2016. California's water storage and distribution 
systems are critically dependent on the Sierra Nevada snowpack, which is projected to decline by 64 
percent by the end of this century (Reich et al., 2018). The snowpack reached a startling 5 percent of 
normal on April 1, 2015. In response, Governor Brown issued an executive order for the first ever 
statewide mandatory water use reductions, along with other measures to reduce water use (Executive 
Order No. B-29-15, 2015). Among the other measures was a requirement to replace 50 million square 


Page 1 of 12 



feet of turf with drought-tolerant landscapes to reduce water use in the urban sector. The State of 
California and local and regional water agencies implemented this plan through rebate programs to 
incentivize property owners to replace turf, and spent over $350 million on these rebate programs 
during the last two years of drought (Knickmeyer, 2016). As these programs were carried out by various 
agencies, they varied in implementation. However, all programs essentially shared the same 
requirement regarding the removal of turf, and typically offered between $0.50 - $4/square foot for its 
removal. 

There is a critical need to quantify the benefit of conservation savings provided by these programs, 
relative to the cost. This will assess the effectiveness of these strategies to allow for science-based 
planning and guidance for future programs. While removal of turf, a water-intensive plant, might seem 
to obviously result in substantial water savings, this may not necessarily occur if irrigation practices are 
not changed concurrently. Though irrigation demand may decrease following turf removal, property 
owners may not reduce irrigation rates, either due to lack of awareness on the need to change irrigation 
practices, lack of knowledge regarding irrigation controllers or systems, or lack of willingness to change 
behaviors. Most programs have enforcement in place to ensure adherence to program requirements, 
for example, a post-inspection to ensure that turf has been removed and that replacement 
requirements have been met. However, property owners could evade program requirements following 
the post-inspection period (e.g. by planting high water use species). In addition, it is unknown how many 
years it would take for water savings related to turf removal to be realized, as replacement species may 
require substantial irrigation in the early years to establish. 

Studies of the water savings associated with turf removal are impeded by lack of sufficient years of post¬ 
conversion data, the difficulty of obtaining water billing data, which may originate at several agencies, 
and the challenge of mining voluminous water billing data to analyze water savings and inform program 
management. 

The Santa Clara Valley Water District (SCVWD), a wholesale water supply and groundwater management 
agency in Santa Clara County, CA, is in a unique position to conduct this study, as it has had a long¬ 
standing turf removal rebate program since 2007. The current version of the program has operated 
since 2010, allowing for the unique opportunity to examine the long-term water savings associated with 
the program. Data for this research study was obtained from 10 retailers through a voluntary research 
partnership with SCVWD. Examination of SCVWD's program also offers the opportunity to quantify 
savings from irrigation equipment rebates, as its program also offers rebates for weather-based 
irrigation controllers installed in conjunction with rain shut-off devices, high-efficiency nozzles, and 
sprinkler bodies. 

This study focuses on evaluating the turf replacement and irrigation equipment rebate program offered 
by the SCVWD, also known as the Landscape Rebate Program (LRP). In this study, the water savings 
associated with the various elements of the LRP are assessed - the replacement of turf with low-water 
use species, automatic timer-based controllers with weather-based irrigation controllers, and 
conventional sprinkler nozzles with high-efficiency nozzles, both independently and in conjunction with 
sprinkler bodies. 


Page 2 of 12 



Methods 


Description of Study System 

The SCVWD service area covers nearly 3,400 km 2 (1,300 mi 2 ) and includes 15 cities within Santa Clara 
County, comprising one of the state's largest urban centers (Silicon Valley). SCVWD provides water to 13 
major retailers. The SCVWD service area experiences a Mediterranean climate with seasonal winter 
rainfall, and has a reference evapotranspiration of approximately 49 inches per year (California Irrigation 
Management Information System, www.cimis.water.ca.gov). During the period of this study, 2010-2016, 
the region experienced drought conditions during the periods from 2006-2010 and 2012-2016. 

The SCVWD Landscape Rebate Program 

The SCVWD LRP offers four rebate types which were analyzed in this study and are described below. A 
more detailed description of each program and the requirements can be found at 
https://www.valleywater.org/landscaperebateprogram. The programs involve pre- and post¬ 
inspections (typically on-site) to ensure that program requirements are met. The programs offer 
educational and technical assistance for meeting requirements via a hotline, plus videos and educational 
resources online. 

Landscape (LND) - This program offers rebates for conversion of turf or pool areas to a landscape of 
low-water use species in conjunction with drip irrigation. In this study, all subjects converted live (green) 
turf areas only. Replacement of a live lawn to a low water-use landscape was a requirement for this 
rebate program during the time of this study. In 2015, replacement of dead (brown) lawns were also 
permitted for this rebate program; these sites were not included in the study as they would confound 
water saving calculations. The species selected for planting in the conversion area are required to be 
selected from the SCVWD's list of qualifying plants, which is adapted from the Water Use Classification 
of Landscape Species (WUCOLS IV) plant list ( http://ucanr.edu/sites/WUCOLS/) . 

Equipment rebate types are described as follows. Program participants must select equipment from the 
SCVWD qualifying list in order to receive a rebate. 

Weather based irrigation controller with rain shut-off device (WBIC+RN) - replacement of conventional 
automatic irrigation controller with "smart" irrigation controller that sets and adjusts water application 
in response to changes in the weather. These controllers are programmed to calculate plant irrigation 
based on weather parameters, typically measured on-site. These controllers are termed as "weather- 
based," "smart," or "ET," and are collectively referred to by the irrigation industry as Smart Water 
Application Technology, or SWAT. These controllers are eligible for rebate if they contain or are 
installed in conjunction with on-site rain shut-off devices. 

High efficiency nozzles (HEN) - conversion of conventional nozzles to high-efficiency nozzles. 

High efficiency nozzles and spray bodies with pressure regulation and/or check valves (HEN+BOD) - 

conversion of conventional nozzles and spray bodies to high efficiency nozzles and spray bodies with 
pressure regulation and/or check valves. 


Page 3 of 12 







Data Collection 


This study focused on single-family residences (SFR), which comprise the largest LRP participant type. 

The participating SFRs included for this study were restricted to those that obtained a single rebate type 
(LND, WBIC+RN, HEN, or HEN+BOD), as the inclusion of combined rebates would restrict the ability to 
differentiate the savings due to different programs. In addition, sites were required to have at least one 
year of data in the pre-conversion period and one year of data in the post-conversion period. 

LRP participants sign an agreement permitting their water retailers to share their water consumption 
data with the SCVWD for study purposes. A partnership was established with ten retailers in the SCVWD 
service area to obtain water billing data for this study. Copies of participant rebate agreements were 
provided to each participating retailer. Water billing data in units of 100 cubic feet (ccf) were obtained 
from each retailer for SFRs that participated in LRP from 2011-2015. Data were uniquely identified by 
street address. The following participating retailers provided water consumption data in a monthly 
format: California Water Service Company, City of Morgan Hill, City of Santa Clara, Purissima Hills Water 
District, and Stanford University. These participating retailers provided bimonthly water consumption 
data: San Jose Municipal Water System, Great Oaks Water Company, City of Milpitas Community 
Services, City of Mountain View Public Works, and San Jose Water Company. Each participating retailer 
also provided the average single family household water use in their service area in units of 100 cubic 
feet (CCF) for comparison to participant water use. As each retailer provided a different data format, a 
standard format was adopted prior to analysis. 

Post-inspection dates varied between 2011-2013 for LND, 2011-2014 for WBIC+RN, 2011-2015 for HEN, 
and 2013-2015 for HEN+BOD, and 5, 4, 3, and 1 year(s) of post-conversion data were available for each 
of these programs respectively. Sample size varied by year, with the highest sample size being one year 
following conversion and declining each year thereafter. In other words, the largest group of subjects 
within each rebate category was in the first year following conversion. 

Data Analyses 

Water savings were determined in this study based on the difference between pre- and post- conversion 
water use for SFRs that participated in LRP from 2011 to 2015. The water billing data used in this study is 
based on meters that measure both indoor and outdoor water use cumulatively. Thus, an assumption of 
this study is that pre- and post- water use differences are due to changes in outdoor water use as a 
result of participation in LRP. To exclude the possibility of indoor water use changes influencing the 
water use differences, each SFR in this study was checked for participation in the SCVWD's water audit 
program known as Water Wise House Call. Through this program, inspectors replace showerheads, 
aerators, or toilet flappers if efficiency may be improved. If an SFR had any of these items replaced 
during the study period, the SFR was excluded from the study. If items were replaced after participation 
in LRP, the years following replacement were excluded from the study. This study does not account for 
other possible indoor water use changes or possible non-rebate related changes in outdoor water use. 

Data were analyzed using Microsoft Excel software v. 2016 (Microsoft, Redmond, WA, USA). Pre¬ 
conversion water use of program participants was determined by averaging available water use prior to 


Page 4 of 12 



conversion. The majority of participants had 3 years of pre-conversion water use. Paired t-tests were 
used to analyze pre- vs. post-treatment differences. For all analyses, P-values less than 0.05 were 
considered significant, and p-values less than 0.1 were considered marginally significant. 

Billing intervals by retailers are staggered differently for different accounts; thus, billing data were 
linearly interpolated at monthly intervals for individual accounts in order to obtain water use at 
consistent time intervals. The monthly water use values were summed to provide annual sums of water 
use before and after landscape or equipment conversion. 

Water savings per unit was determined for irrigation equipment (WBIC + RN, HEN, and HEN + BOD), and 
per square foot for LND and WBIC+RN. For LND, water savings was divided by conversion area, which 
was measured on-site by inspectors. For WBIC+RN, water savings was divided by irrigated area, which 
was determined using GIS and Google Maps. 

Average water use of the retailers' total SFR sector was compared to average water use of program 
participants to help determine whether average sector-wide data could be used to help inform study 
findings. Since participation in the rebate programs is low (less than 0.5% of retailers' total SFR 
customer base), the average SFR use is not expected to be impacted by participation. Average SFR use 
was expected to decrease during the study period as an increasing number of county residents reduced 
their outdoor water use in response to drought and allowed their lawns to die. Thus, average SFR use 
was expected to signal an increasing number of dead lawns and was used in this study for the purpose 
of comparison, rather than an experimental control. For each account, data was paired with the 
average SFR water use from the same retailer service area for comparison. 

Results 

Water savings in each year following installation are shown in Tables 1 through 4. For each year, sample 
size, and p-values of paired t-tests between pre-installation and post-installation samples are also shown 
for the years in which a suitable sample size exists. The water savings for LND were marginally significant 
in the first year following conversion, and were statistically significant in Years 2 through 5 following 
conversion. The average water savings increase incrementally each year following conversion (Figure 1). 
The water savings for WBIC + RN were statistically significant in the first year following conversion, and 
continue each year through 4 years following conversion. Similar to LND, the average water savings for 
WBIC + RN incrementally increases each year following conversion (Figure 2). The water savings for HEN 
was marginally significant in the first year following conversion, and was statistically significant 1-3 
Years following conversion (Figure 3). The water savings for HEN + BOD was statistically significant a year 
following conversion (Figure 4). 


Page 5 of 12 



Table 1: Water savings resulting from turf removal rebate program (LND), based on the difference 
between post-conversion water use and average of 1 - 5 years of pre-conversion water use. 


Year following installation 

Year 1 

Year 2 

Year 3 

Year 4 

Year 5 

Water savings (gal/ft 2 /y) 

8 ±4 

22 ±5 

26 ± 6 

28 ± 6 

48 ±9 

Sample size 

142 

137 

106 

64 

30 

p-value of t-test between pre¬ 
installation and post¬ 
installation sample 

0.0738* 

<0.0001** 

<0.0001** 

<0.0001** 

0.0003** 


*marginally significant (p<0.1) 
^^significant (p<0.05) 


Table 2: Water savings resulting from weather based irrigation controller with rain sensor rebate 
program (WBIC + RN), based on the difference between post-conversion water use and average of 1 - 
5 years of pre-conversion water use. The water savings is shown for square footage of area irrigated 
by WBIC, as well as per WBIC unit. 


Year following installation 

Year 1 

Year 2 

Year 3 

Year 4 

Water savings (g/ft 2 /y) 

4 ± 2 

7 ± 2 

11 ±3 

14 ±4 

Water savings (g/unit/y) 

22,724 ± 5839 

26,403 ± 8,067 

32,712 ± 9,008 

38,440 ± 11,647 

Sample size 

84 

60 

35 

16 

p-value of t-test between pre¬ 
installation and post¬ 
installation sample 

0.0002** 

0.0013** 

0.0010** 

0.0049** 


^^significant (p<0.05) 


Table 3: Water savings resulting from the conversion of conventional nozzles to high-efficiency nozzles 
(HEN), based on the difference between post-conversion water use and average of 1 - 5 years of pre¬ 
conversion water use. 


Year following installation 

Year 1 

Year 2 

Year 3 

Water savings (g/unit/y) 

541± 233 

1,536 ±337 

949 ± 412 

Sample size 

52 

40 

11 

p-value of t-test between pre¬ 
installation and post¬ 
installation sample 

0.0626* 

0.0001** 

0.0291** 


*marginally significant (p<0.1) 
^^significant (p<0.05) 


Page 6 of 12 






Table 4: Water savings resulting from the conversion of conventional nozzles and bodies to high- 
efficiency nozzles and bodies (HEN + BOD), based on the difference between post-conversion water 
use and average of 1 - 5 years of pre-conversion water use. 


Year following installation 

Year 1 

Water savings (gal/unit/y) 

1,661 ±701 

Sample size 

17 

p-value of t-test between pre¬ 
installation and post¬ 
installation sample 

0.0384** 


^^significant (p<0.05) 


Figure 1: Landscape Conversion Program 
Annual Water Savings 


60 

50 

40 

30 

20 

10 

0 


Year 1 



Year 2 Year 3 Year 4 


Years following installation 



Year 5 


Page 7 of 12 







g/sq.ft. 


Figure 2: Weather-Based Irrigation Controller 
Program Annual Water Savings 


20 
18 

16 1 
14 

12 

10 J 

8 I 
6 
4 

2 J 

o -L 

60,000 . 


ifi 

Year 1 



Year2 Year3 Year4 


50,000 

40,000 



Yearl Year 2 Year 3 Year 4 


Years following installation 


2,000 

1,500 

+-J 

§ 1,000 

oo 

500 

0 


Figure 3: High-Efficiency Nozzle Program 
Annual Water Savings 


i 

Year 1 



Year 2 

Years following installation 



Year 3 


Page 8 of 12 












Figure 4: Sprinklers with Check Valves and High- 
Efficiency Nozzles Program - Annual Water 
Savings 


2,000 

1,500 

4-J 

I 1,000 

tSo 

500 

0 



Year 1 

Year following installation 


As noted above, average water use by program participants was compared to average SFR use. The 
purpose of this analysis was to determine whether average SFR use could be used to determine water 
savings associated with program participation. Tables 5 through 8 show that pre-installation water use 
differed between average SFRs and program participants in every case for LND, WBIC + RN, HEN, and 
HEN + BOD (p = 0.0007, p = 0.0004, p = 0.0023, p = 0.0044, respectively). For turf removal rebates, 
participants had lower pre-installation water use. For equipment rebates, participants had higher pre¬ 
installation water use. The average pre-installation SFR water use varies between Tables 5 through 8 
because the pre-installation periods varied. During the period analyzed, average SFR water use declined 
along with declines in participant water use. The author attributes the decline in average SFR use to 
residents' response to the 2012-2016 drought, which included an increasing number of county residents 
letting their lawn brown in order to conserve water. Some may have responded to the local "brown is 
the new green" media campaign during the time, which encouraged people to conserve by cutting back 
on outdoor irrigation. Given the differences in pre-installation water use between the program 
participants and average SFR water use, as well as the variation in average SFR water use, average SFR 
use is not appropriate as an experimental control. Thus, average SFR use is shown only for comparison. 
Water savings was determined as the difference between pre- and post- LRP project water use. 


Table 5: Pre-installation annual water use (average water use 1-5 years prior to conversion) and 
percentage decrease in water use for LND participants and average SFRs. Participants had lower 
average pre-installation water use than average single family residences within the same retailer area 
(p = 0.0007). 



Pre-installation annual 
water use (CCF) 

Percent decrease in water use 

Year 1 

Year 2 

Year 3 

Year 4 

Year5 

Participants 

161 ±9 

6% 

13% 

18% 

23% 

32% 

Average SFR 

191 ±5 

-2% 

7% 

11% 

19% 

28% 


Page 9 of 12 












Table 6: Pre-installation annual water use (average water use 1-5 years prior to conversion) and 
percentage decrease in water use for WBIC + RN participants and average SFRs. Participants had 
higher average pre-installation water use than average single family residences within the same 
retailer area (p = 0.0044). 



Pre-installation annual 
water use (CCF) 

Percent decrease in water use 

Year 1 

Year 2 

Year 3 

Year 4 

Participants 

250 ± 17 

15% 

17% 

20% 

27% 

Average SFR 

205 ± 6 

9% 

16% 

20% 

29% 


Table 7: Pre-installation annual water use (average water use 1-5 years prior to conversion) and 
percentage decrease in water use for HEN participants and average SFRs. Participants had higher 
average pre-installation water use than average single family residences within the same retailer area 
(p = 0.0004). 



Pre-installation annual 

Percent decrease in water use 


water use (CCF) 

Year 1 

Year 2 

Year 3 

Participants 

254 ± 19 

8% 

20% 

18% 

Average SFR 

180 ±6 

10% 

23% 

22% 


Table 8: Pre-installation annual water use (average water use 1-5 years prior to conversion) and 
percentage decrease in water use for HEN + BOD participants and average SFRs. Participants had 
higher average pre-installation water use than average single family residences within the same 
retailer area (p = 0.0023). 



Pre-installation annual 
water use (CCF) 

Percent decrease in 

water use 

Year 1 

Participants 

305 ± 43 

18% 

Average SFR 

186 ± 15 

20% 


Discussion 

Analysis indicated water savings for each post-conversion year examined in this study for WBIC (1-4 
years) and HEN+BOD (1 year). LND and HEN savings were statistically significant after the first post¬ 
conversion year, up to five years for LND and up to three years for HEN. For LND and WBIC, there was an 
incrementally increasing trend in average annual water savings. Further study would be needed to 
determine how many years the savings would continue to increase. Plant water use is expected to 
decrease with maturity and eventually stabilize. Thus, the water savings for LND may plateau after a 


Page 10 of 12 






certain number of years, corresponding with the time needed for the new landscape to fully establish 
and for maintenance practices to regularize. Annual average LND savings were up to 48 g/ft 2 (during the 
fifth year following conversion), and on average 31 g/ft 2 for years 2 through 5 when savings were 
significant. For comparison to LND savings (Table 1), the estimated annual savings in the Alliance for 
Water Efficiency Tracking Tool model was 36 gal/ft 2 , 

(http://www.allianceforwaterefficiency.org/Tracking-Tool.aspx) and the estimated savings by the 
California Data Collaborative in Moulton Niguel are 24.6 gal/ft 2 . 

Surprisingly, HEN were only marginally significant in the first year following conversion, and were 
significant thereafter. This may be because landscapers or householders who install high efficiency 
nozzles may not be aware of how to properly adjust the scheduling for the new nozzles during their first 
year. It is recommended that with HEN rebates, the recipients should be provided information or 
assistance to adjust their irrigation run times following conversion to achieve the immediate savings that 
would be expected from HEN. Once savings from HEN were significantly achieved in year 2 (1,536 ± 337 
g/unit/y) and year 3 (949 ± 412 g/unit/y), the savings do not differ much from high-efficiency nozzles in 
conjunction with bodies, indicating that there is not additional savings to be achieved by adding 
sprinkler bodies to the high-efficiency nozzle rebate. 

Further study would help determine the lifetime for the water savings for each rebate type, or whether 
the rebates lead to a permanent reduction in water use. There was a large decline in average SFR use 
during these drought years. It is expected that average SFR water use would rebound in the years 
following this study, while participant water use remains lower. 

Installation of WBICs has greater savings per unit as compared to HEN and HEN + BOD, but these savings 
are dependent on area utilized by the WBIC and are expected to vary based on the household's 
landscape area. 

Interestingly, participants in the LND had lower pre-installation water use than average SFR user, which 
suggests that program uptake is by households already conserving water relative to others in their 
service area, and may be motivated by environmental ideology. For equipment rebates, pre-conversion 
water usage was higher for participants as opposed to the average SFR. This suggests that equipment 
rebates may be pragmatic or financially motivated, as these users are on average spending more on 
their water bills than others in their service areas. 

Controlled experiments, primarily in humid climates, have shown water savings of 40-70% when using 
weather-based irrigation devices, but large real-world studies have shown savings less than 10% (Dukes 
2012). Smart controllers may be programmed incorrectly, causing over-irrigation (Pittenger et al. 2004, 
Bijoor et al, 2014). They cannot reduce the irrigation rate unless sprinklers irrigate uniformly, or are 
adjusted to irrigate uniformly. The percentage of savings reported in this study is higher than other 
studies for several possible reasons. First, WBICs must be installed with rain-sensors, and this is likely to 
increase the likelihood of success. In addition, most previous studies of WBICs are in regions with lower 
reference evapotranspiration. 


Page 11 of 12 



Conclusions 


The study results indicate that significant water savings were achieved by SCVWD conservation 
programs involving rebates for turf removal, weather-based irrigation controllers, high-efficiency 
nozzles, and high efficiency nozzles with sprinkler bodies. For turf removal rebates, savings are marginal 
in the first year, significant thereafter, and incrementally increase. The study shows that the rebate 
programs offered by SCVWD have been successful in achieving significant water conservation. Beyond 
the adoption of new landscape and technology, other factors that may contribute to program success 
are SCVWD's stringent requirements (requirements for plant list, drip irrigation, and pre- and post¬ 
inspection verification) and SCVWD's capacity-building efforts (program and educational information 
provided by inspectors on-site, the availability of a conservation hotline to provide participants with 
program assistance, and detailed online program information, educational outreach, and instructional 
videos). 

References 

Bijoor, N., Pataki, D., Haver, D., Famiglietti, J. (2014) A comparative study of the water budgets of lawns 
under three management scenarios. Urban Ecosystems, 1-23. 

Diffenbaugh, N., Scherer, M., Ashfaq, M. (2012) Response of snow-dependent hydrologic extremes to 
continued global warming. Nature Climate Change, doi:10.1038/nclimatel732. 

Dukes, M. D. 2012. Water conservation potential of landscape irrigation smart controllers. Transactions 
oftheASABE, 55:563-569. 

Executive Order No. B-29-15 (2015), accessed from 
https://www.gov.ca.gOv/docs/4.l.15_Executive_Order.pdf 

Knickmeyer, E. (2016) In California, a $350 million social experiment over lawns. AP News, 

https://apnews.com/c7acl74c2aec4470ba4c416944864d01/california-350-million-social-experiment- 

over-lawns 

Griffin, D., Anchukaitis, K.J. (2014) How unusual is the 2012-2014 California drought? Geophysical 
Research Letters, 41, 9017-9023. 

Pittenger, D. R., D. A. Shaw, and W. E. Richie. 2004. Evaluation of weather-sensing landscape irrigation 
controllers. University of California Cooperative Extension: Riverside, CA; 26 pp. 

Schmitt, E., Tull, C., Atwater, P. (2016) How Much Water Does Turf Removal Save? Applying Bayesian 
Structural Time-Series to California Residential Water Demand. Paper presented to SIGKDD: Data Science 
for Food, Energy and Water, San Francisco, August 2016. 


Page 12 of 12