The California Single-Family Home Water Use Efficiency Study

March 01, 2013

WEB ONLY

This online-only article is a supplement to the March/April 2013 print edition of Home Energy Magazine.

Click here to read more articles about Utility Programs

This article deals with a simple subject: how water is used in single-family homes in California. This topic has important consequences for the future of the state of California. The state’s official goal is to reduce per capita water use by 20% by 2020.

The study upon which this article is based is a bottom-up approach to the subject. Rather than trying to infer customers’ water use patterns from gross production data and various other sources—such as surveys and census information conducted on whole populations of customers—we have collected highly detailed information at the water meter on random samples of customers chosen from billing databases, with the goal of projecting patterns in the populations from these samples. The overall period covered by our investigation ranges from 2005 to 2010, and the bulk of the water use data were collected from 2005 through 2008.

Figure 1. Changes in patterns of indoor use between 1997 and 2005.

Figure 2. Toilet flush volumes in the California study and the REUWS study.

Figure 3. Distribution of clothes washer load volumes in the California study.

Figure 4. Distribution of application ratios in California study homes.

We believe that the results of the study shed light on how California single-family customers are currently using water; on how their water use patterns have changed over the ten-year period since the Residential End Uses of Water Study (REUWS) was conducted in 1997; and on how future water use patterns might be modified in order to increase the efficiency of use and modify demands to moderate the need for raw water withdrawals from increasingly overextended supplies.

The California Single-Family Home Study

The California Single-Family Home Water Use Efficiency study includes data from many traditional sources—such as billing data, survey data, weather data, and aerial photo data— to analyze the water use patterns of a sample of over 700 single-family homes across ten water agencies throughout the state of California. Detailed flow trace data were obtained from portable data loggers, which were attached to the water meters of each of the study homes. These flow traces provided flow readings at ten-second intervals from the magnetic pickup, which generates 80 to 100 pulses per gallon. These highly detailed flow data make it possible to identify individual water use events and to categorize them by their end use. The flow trace data tell not just how many gallons per day the home used, but how many gallons per day were used for individual end uses, such as toilet flushing, clothes washing, dishwashers, showers, irrigation, faucets, and leaks.

Detailed use information can be pulled from the trace, giving, for example, a count of toilet flushes and toilet flush volumes during a logging period. Researchers used flow trace data to determine levels of daily use in the homes, and the efficiency of that use. Although the flow trace technique is subject to marginal error, such as error from the miscategorization of some events, it provides information on end uses that is not available from any other source. This article summarizes the results of the study, which began in 2005 and was completed in 2010. Water use patterns found during the 2007 logging period were analyzed to determine the potential for conservation savings from both indoor and outdoor efforts.

Information on Current Water Use

Assessing the efficiency of water use in single-family homes implies having a standard upon which to base the comparison. The efficiency of the homes can then be described as a numerical value based on the chosen standard. For single-family homes, it is necessary to have two standards: one for indoor use and one for outdoor use.

Determining efficiency standards. The standard used in this study for indoor use was the household water use for a home employing best available technology for all fixtures and appliances, and with less than 25 gallons per household per day (gphd) of leakage. In effect, the indoor standard was based on the EPA WaterSense specifications for indoor devices. In the report, the data from the 2000 study of a group of 100 homes that had been retrofitted with high-efficiency devices—the EPA Post Retrofit Group—were used as the benchmark for what we referred to as efficient homes. For indoor uses, it was possible to have a single number that represented the number of gallons per day of use expected for efficient homes.

While indoor uses are relatively consistent from home to home, outdoor uses are much morevariable, and it is really not possible to have a single number that tells how many gallons per year should be used for outdoor purposes. What served the purpose for an outdoor standard were two values referred to in the study as the application ratio and the volume of excess use. The application ratio is the ratio of actual outdoor water use to the theoretical requirement for outdoor use; this theoretical requirement is based on the size and type of landscape, the local evapotranspiration (ET), and whether there is a swimming pool on the property. An application ratio of 100% indicates that precisely the correct amount of water is being used outdoors at the home. The volume of excess use is the difference between the actual outdoor use and the theoretical requirement (in thousands of gallons, or kgal). Using these parameters, an efficient home will have an application ratio of 100% or less, and will have no excess outdoor use.

A total of 735 homes from ten water agencies were included in the indoor analysis for this study. The weighted average annual total water use of these homes was 132 kgal per year, or 362 gphd. The average daily indoor use for the homes, as determined by the flow trace analysis, was 171 gallons per household per day (gphd). By subtracting the indoor water use from the annual use, the outdoor use can be estimated. The weighted average daily outdoor use for the group was190 gphd. Approximately 53% of the annual total water use appears to be for outdoor uses and 47% appears to be for indoor uses, based on billing data analysis.

Indoor efficiencies. When the indoor use (plus leakage) was analyzed from the flow trace data, it suggested that the indoor use for households has declined since 1997, when the RUEWS group conducted its study. However, it is still significantly greater than the indoor use for the EPA Post Retrofit Group. Table 1 compares the indoor use of the study group to that of these two benchmark groups.

When the indoor uses are disaggregated, the results are more revealing. The disaggregated data (Figure 1) show that, as one would expect, there have been significant reductions in indoor use for toilets and clothes washers in California between 1997 and 2005. At the same time, the indoor uses attributed to the other categories have stayed the same or have increased in a way that has masked the savings from the toilets and clothes washers. This pattern is especially true for events classified as leaks. The analysis showed significantly more long- duration or continuous flows that get classified as leaks. These continuous events, which are found in a small number of homes, raise the average volume of water attributed to leaks for the study group from around 22 gphd to 31 gphd. This finding calls for further investigation to determine whether these truly are leaks, or whether they are due to devices that actually create a continuous demand for water. This information is important, because if the leakage, faucet use, and shower use were all brought down to the levels shown in the REUWS study, the average indoor use for the group would be around 150 gphd—which would have been a significant improvement from the 1997 data.

The data show a major improvement in the water use efficiency of toilets. There were a total of 122,869 flushes recorded during the data-logging period. The average flush volume was 2.76 gallons, and 64% of all flushes were less than 2.75 gallons. The one negative finding on toilets was that apparently many toilets that are designed to meet the ultralow-flow (ULF) standard of 1.6 gallons per flush (gpf) are flushing at significantly larger volumes. This helps explain why the study found that only 30% of the homes were at average flush volumes of 2 gpf or less, while all of the program data, confirmed by survey data from this study, suggest that over 60% of the toilets in the population are ULF or better models.

Figure 2 compares the distribution of toilet flush volumes in the California Single-Family Home study and the REUWS study. This shows a dramatic shift in the bins containing the largest percentage of flushes. In the 1997 sample, the largest percentage of flushes were between 3.75 and 4.25 gpf, but as of 2007 they were between 1.25 and 2.25 gpf. As more of the toilets on the right side of the distribution are replaced with high-efficiency models, the overall demands for toilet flushing will drop well below the current levels, and the percentage of homes meeting the 2 gpf efficiency criterion used for this study will increase.

Figure 3 shows the distribution of clothes washer load volumes from the data in the California study. As of 2007, approximately 30% of homes were using 30 gallons per load or less for clothes washing. At the time of the REUWS study, only around 1% of the clothes washers used less than 30 gallons per load, so the current data represent a major advance; but these data also show that there is still significant potential for savings in clothes washer use.

There was little change in shower use between 1997 and 2007. The average was just over 18 gallons per shower (around the same volume that is required to fill up an occupied bath tub), and the duration of showers was just under nine minutes. Nearly 80% of all showers were flowing at 2.5 gpm or less. Reducing flow rates and duration of showers remain the methods available for conservation in showers. These are behavioral changes that people can make to save water during periods of drought or all the time.

The average leakage rate in the California study homes was 31 gphd. The median rate was 12 gphd. The wide disparity between these values shows that a small group of homes are leaking at a very high rate, which increases the average for the entire study group. Leakage is complicated by the fact that some events that Trace Wizard categorizes as leaks may be due to devices such as water treatment systems that create a continuous demand for water. The research team does not believe that this occurs very often. Leaks from very short-duration events, such as drips or occasional toilet flapper malfunctions, usually amount to 10 gpd or less of household demand. The leaks that contribute heavily to household demand are those that continue for many hours or days. Most of these leaks are caused by continuously running toilets, broken valves, or leakage from pools and irrigation systems. We need to learn more about these continuous events, so that we can deal with them appropriately. The sample group used an average of 33 gpd of water for miscellaneous faucet use. Average volume was less than 1 gallon per use; average duration was 37 seconds. The average home recorded over 57 faucet events per day. Faucet use represents a category of growing importance as toilets and clothes washers become more efficient. The key to improving the efficiency of faucet use is to decrease the flow rates and the duration of the events.

Outdoor use efficiencies. Only 87% of the homes in the study group were assumed to be irrigating. This assumption was based on the fact that some lots had no irrigable area, and some homes showed little or no seasonal water use. Only around 54% of the homes that irrigate are doing so to excess. Overall, outdoor use efficiency is fairly good. Figure 4 shows the distribution of application ratios in the study homes.

The average outdoor use volume for the irrigating homes as a group is very close to the average theoretical requirement. The average annual outdoor use for the group as a whole is 92.7 kgal. The average theoretical irrigation requirement (TIR) for the group is 89.9 kgal. So, taken as a whole, there is only 2.8 kgal of excess use per lot occurring in the group. Another way of looking at this is that the underirrigation in the less-than-TIR group just about balances the overirrigation in the more-than-TIR group

The data indicate that there is a substantial savings potential on the lots of customers who are overirrigating. From the perspective of water conservation, the customers who are deficit irrigating need to be set aside, and attention should be focused on the overirrigators.

The excess-use statistics shown in the full report show that the average excess use on the lots that are irrigating is approximately 30 kgal per year. Since only 87% of the lots were irrigators, the average excess use for all single-family accounts is estimated at 26.2 kgal per year. Approximately 62% of this excess use is occurring on 18% of the irrigating lots, or 15% of all lots. This is critical for water management, because it shows that in a typical system most of the savings from outdoor use will be derived from around 15% of the customers.

Estimating Conservation Potential

The question of how much potential water savings we can achieve in California is closely related to how well we determine the levels of efficiencies. The study used models of indoor and outdoor water use developed from the data collected in the study homes to predict the impact of making specific changes in indoor and outdoor water use. These models were used to predict conservation potential in the study homes, and by inference, in the state.

The data and models show that average indoor household water use could be reduced from the 2007 level of 175 gphd to 120 gphd if

the maximum clothes washer volume were 20 gallons per load;
the volume of water used by miscellaneous faucets could be reduced by 10% from 2007 levels;
leakage could be reduced to a maximum of 25 gphd; and
the maximum toilet flush volume could be set at 1.25 gpf.

This amounts to a potential of 55 gphd of indoor savings per household, or 20 kgal per year. The report did not discuss precisely how these goals were to be met, and these changes could occur gradually over many years. The key is to retain building codes and regulations that require the standards be met in new and remodeled construction.

The data and models show that the conservation potential remaining in the system from outdoor use is greater than the potential from indoor use. The study found that there are three key parameters for modifying outdoor use: the irrigated area, the water demands of the landscaping, and the percentage of homes that are overirrigating. Data in the full report show that, according to the outdoor use relationships observed in this study, if the average irrigated area were decreased by 15%, the average landscape ratio decreased by 35%, and the percentage of homes that are overirrigating reduced from 50% to 20%, it would be possible to reduce outdoor use to an average of 40 kgal per household from its 2007 level of 90 kgal. The low-end estimate is that by simply reducing the percentage of overirrigators to 20%, and leaving all of the other parameters intact, outdoor use could be reduced by 28%, saving approximately 0.6 million acre-feet (MAF).

Our study identified three levels of potential conservation savings for the single-family sector. The indoor savings potentials are based on the end point chosen for indoor household use. A potential average savings of 20 kgal per home was estimated assuming an indoor use benchmark of 120 gphd. The estimate could be raised to 30 to 40 kgal per household assuming that benchmarks of 105 gphd could be achieved and more aggressive indoor technologies used. Consequently, we can conceive of three levels of indoor water conservation benchmark: a low, medium, and high level at 20, 30, and 40 kgal per year per home. Total indoor estimates statewide are based on the estimate of 9.5 million single-family households in the state.

The study estimated outdoor potential conservation savings at a low of 0.6, a medium of 0.80, and a high of 1 MAF per year. The savings in all three ranges are deemed technically achievable, but achieving them would require significant and increasing work over time. It would also require innovations in preventing overirrigation and changes to both irrigated areas and plants used for landscaping. It is encouraging, however, that the low-end savings would more than achieve the desired 20% reduction in use. The practicality of achieving savings in the high range is less clear, and is closely related to less clear; it depends on the value placed on the saved water (or costs for agencies to develop new supplies as alternatives). Table 2 provides a summary of the estimated potential conservation savings derived from this study.

Guidance for Allocation of Resources

Since the signing of a memorandum of understanding (MOU) in 1991, water conservation efforts have been focused on implementation of the Best Management Practices. These are mainly programs that lend themselves to tracking on the basis of activities performed and fixtures replaced. The most convincing argument for the effectiveness of water conservation efforts, however, is one that is backed up by hard data that show reductions in household water use. The California study demonstrated techniques of sampling and data collection that can be used to collect these data. Detailed analyses of household and per capita water use on representative samples of customers can provide valuable information that will complement the agencies’ other tracking and evaluation efforts.

The degree to which both excess use and potential savings are skewed in the population should be considered when designing programs. Programs that aim to control leakage or excess use of water for irrigation should not be targeted to the entire population, since most of this leakage and excess use is associated with only a few customers. It would be better to design programs targeted to just these customers. Water budgets, smart meters, leak detection devices, and better customer education will all prove useful.

The information on toilets should also help water agencies to design better programs. The data in the California study showed, first, that even though a high percentage of toilets appear to have been replaced with ULF models, the percentage of homes that are flushing at 2 gpf or less is lagging what it should be, and second, that the actual flush volumes of ULF models range well above 1.6 gpf. If future retrofits are focused on newer high-efficiency toilets (those using 1.28 gpf or less), and work continues to replace all of the remaining high-volume toilets in the homes upgraded to the high-efficiency toilets, the percentage of complying homes will increase rapidly over time, and household water use devoted to toilet flushing will decrease.

The data show that reducing the percentage of homes that overirrigate is the single most important factor in reducing outdoor use. The report, however, does not support making weather-based irrigation controllers mandatory. The data show that these devices would cause irrigation to increase in about as many homes as it would be reduced. The key to controlling outdoor use is to design programs that discourage excess irrigation use while allowing customers who prefer to underirrigate to continue to do so. This requires targeting overirrigators. This in turn requires having a rough estimate of the irrigated areas and outdoor water use for each customer, and comparing this information to the customer’s actual seasonal use.

The report highlighted the importance of leaks and other unexplained continuous uses in raising average use for the entire population. Rather than have general programs targeted to all customers, the report suggests that it would be better to have systems that can alert customers to the existence of a use pattern that suggests a leak, so that the leak, if any, can be remedied immediately. In every group of houses that were logged as part of the study, there were several that showed these long-duration and high-volume leak-like use patterns. Having programs in the billing system that detect increases in use and then send a text message, phone call, or e-mail to the customer is one possibility. Having in-home monitors that read data from the AMR meters directly is another. Having water rates that seriously penalize excess water use would provide an economic incentive for customers to save water.

learn more

Download for free the full report upon which this article is based.

Get more information about the U.S. EPA WaterSense specifications.

The report suggests that developing better customer information and water use tracking systems would greatly improve water agencies’ ability to establish better water management programs. As the old saying goes, you can’t manage what you don’t measure. Key information that would assist in water management would include the number of residents in the home, annual and winter month water consumption, the size of the lot, and size of the irrigated area, and the local ET for the lot. Such information would be invaluable for planning and evaluation purposes. Systems that provide customers with real-time information on water use, along with targets for use, enlist the customer as an active partner in water management. Having the customers as partners should greatly enhance the response of the entire system.

William B. DeOreo is president of Aquacraft, Incorporated, Water Engineering and Management in Boulder, Colorado.

The author wants to acknowledge Fiona Sanchez and Amy McNulty, of Irvine Ranch Water District, for applying for and managing the grant, and the financial and project-related aspects of this study, and for acting as liaison between the participating agencies, the California Department of Water Resources, and the consultants.

The California Department of Water Resources provided the Proposition 50 Grant that funded most of this study.

The hundreds of water customers in the participating agencies who allowed their water use data to be included in the study, filled out surveys, and allowed us to data log their homes and analyze their landscapes, and the participating water agencies who contributed cash, data, in-kind services, and moral support to the project, made the study possible.

Research partners and study coauthors are Jim Henderson and Bob Raucher of Status Consulting, and Peter Gleick, Matt Heberger, and Heather Cooley of Pacific Institute.

Special thanks to Bill Gauley, Veritec Consulting, Incorporated, for assistance with the double-blind evaluation of flow trace analysis discussed in this report.