Water efficiency, and changes to water demand, associated with implementing specific water conservation measures and programs can only be assessed by comparing changes to water use over time. In that water conservation measures and programs that a utility may select to implement require some investment of resources – either funding or labor or both – monitoring and verification of changes in water use related to any specific investment is vital to understanding the value of the investment and the need for program revisions or adjustments. Decision makers within the organization require information related to the costs and benefits of any implemented program to base future program assessments and ultimately support program implementation. In addition, it is increasingly important that water utilities that are responsible to maintain public trust be transparent in how they utilize public funds to support the sustainability and efficiency of the organization. For these reasons, monitoring and verification of programs is a requisite component of overall water resource management and water conservation planning.
Tracking water use over time requires comparing before to after, where the “before” case is called the baseline; the “after” case is referred to as the post-implementation or performance period. Baseline data can include any number of pieces of information such as individual customer water use, total system water use, number of customers participating in a particular program (with the baseline being zero), total real water loss, total water loss (real plus apparent), etc. Baseline data is collected to support assessments related to the goals that the water utility selected to address with water conservation programs, water resource management objectives, or any other number of reasons or justifications. Post-implementation data is the same data collected over time for use in comparative analyses. Any of those data listed in the data collection and warehousing BMP could be used for both baseline and post-implementation cases.
Challenges exist when comparing data overtime that is influenced by outside forces. For example, outdoor irrigation rates for residential or commercial customers are influenced by ambient weather and precipitation. Therefore, care must be taken when doing comparative analyses to isolate and/or account for external influences on time series of data such that accurate assessment of any specific program can be measured.
Generally speaking, data collection is best for shorter intervals over a long time period– so either daily or monthly intervals depending on the characteristic parameter being measured over a two to five year period. For example, total water production is typically measured on a daily basis, and therefore the comparative analyses should use daily data. Customer water use is typically obtained monthly to support water billing. Therefore, customer water use should be compared using monthly data. However, it is vital that the comparative analyses occur over a period of time, for example on an annual basis using daily or monthly data over more than one year, to identify trends and to allow noise or static in the data to be reduced.
Once the data of interest have been collected, statistical comparison, or other modeling maybe needed to perform the assessment and identify impacts of implementation on water demand over time. Given that comparative analyses can be complex to assess, it may be valuable for a water utility to seek support from a trained professional to assist with its monitoring and verification needs.
The water utility must exercise diligence to ensure that the data collection, warehousing, and assessment efforts are fully incorporated into the water resources management and/or water conservation program such that the appropriate level of performance verification for the specific conservation measures is measured and verified. To accomplish this, the data collection and assessment efforts should include:
The following is a listing of data assessment techniques for water conservation programs based on manufacturer specifications and estimates, field data, and analysis. The techniques vary in their level of sophistication and likely error, from simplest to complex (based on based on “Measurement and Verification for Federal Energy Projects,” U.S. Department of Energy Office of Energy Efficiency and Renewable Energy, 2000).
Stipulated and measured factors – Water savings are based on a combination of measured and estimated (stipulated) factors. Measurements are collected in the field from specific locations at the component or system level for a single (i.e., spot) or short-term period. A stipulation factor, or estimate, is supported by historical or manufacturer’s data (for example, estimating water savings from high efficiency toilets may involve using a combination of the number of new toilets installed multiplied by the difference in flushing volume between the old toilets (measured in the field) and the new toilets (based on manufacturers specifications) multiplied by the expected number of uses per unit time). These savings are supported by engineering calculations, or component, or system models. Costs to collect the data and perform these calculations are estimated to range from 1%-3% of project costs, depending on number of points measured.
Measured factors - Water savings are based on spot or short-term measurements taken at the component or system level when variations in factors that may create noise or interference regarding estimating water use reductions are not expected, or based on continuous measurements taken at the component or system level when variations are expected (for example, taking continuous, or daily readings of customer water use at their meter after a water audit is performed to identify savings in daily water use). These saving estimates are supported by engineering calculations, or component, or system models. Costs to collect the data and perform these calculations are estimated to range from 3%-15% of project costs, depending on number of points measured and term of metering.
Utility billing data analysis – Estimate water savings based on long-term, whole-building (or home) utility meter, facility level, or sub-meter data. These saving estimates are supported by regression analysis of utility billing meter data. It is important that for many water conservation programs that involve rebates or incentives – for both indoor and outdoor water efficiency – it is important to compare homes and/or business that participate in the program with a control population that is similar to the participants, but did not participate in the subject program. This is performed by developing databases that track customer water use with and without program participation for customers of similar type (e.g., for residential programs include control population based on age of home, home size in square feet, irrigated acres, etc.). In this way, the effect of organic changes to customer water use can be distilled from the effects of any specific program.
It is also important to note that tracking substantial amounts of customer data will require a data validation and quality control step to an ensure that the customer water use records are accurate and can be tracked to specific street addresses (for mapping analyses are an important data assessment method). Therefore, some type of customer data preprocessing software may be needed to develop sets of water use over time for appropriate customer sets. Costs to collect the data and perform these calculations are estimated to range from 5%-12% of project costs depending on complexity of the billing analysis.
Calibrated computer simulation - Computer simulation inputs may be based on several of the following: engineering estimates; spot, short-, or long-term measurements of system components; and long-term, whole-building utility meter data. These savings are supported by computer simulation model calibrated with whole-building and/or customer water use segment end-use data. Estimated range is 5%-12% depending on complexity and the number systems modeled.