Water Journal : Water Journal May 2011
water MAY 2011 83 refereed paper demand management inconvenience for a particular behaviour change and any constraints preventing a choice (eg. water restrictions). Persistence describes how long an individual retains a particular behaviour and relates to the reversion/bounce-back in behaviour when influences are reduced. Persistence may be influenced by the extent to which an individual is inconvenienced with its current behaviour, the duration that an individual has endured a particular behaviour (ie. becomes accustomed to a behaviour, resulting in reduced inconvenience), and the level of influence still bearing on that behaviour. Validating the Model Methodology Extensive validation of the simulation model was conducted to prove the model and ensure confidence in the simulation results (forecasts). Validation is conducted by predicting forward from the past, where actual results (customer billing data) have not been explicitly used as an input to the model. The model can be validated at various levels (region, sub- region, household area and product type). The CHW model was validated to a high level of accuracy for Ballarat. A more simplified version of the model was also validated in the unrelated Bendigo region, demonstrating the transferability of the model. Validation of the ability for the model to accurately forecast bounce- back in demand was also performed. Validating demand forecasts in Ballarat Figure 1 shows a comparison of the simulated output versus the actual monthly water consumption of households in the Ballarat region. The graphs are closely aligned and show similar trends over an eight-year period from July 2001 to June 2009 where different restrictions (Stages 1 to 4) and influences were present. The average accuracy for each monthly time point between the simulated and actual water consumption is 95.1%. Consistent with simulation results for Bendigo (see Figure 2), simulation forecasts show lesser water consumption when severe outdoor restrictions were first introduced, potentially indicating a high level of non-compliance at this time and/or a time lag for customers to become aware of restrictions and adjust their behaviour accordingly. As an extra point of validation, initial simulation results showed lower water consumption over the period from August 2003 to February 2005 where Stage 3 restrictions were in place. It was identified that the current Stage 3 water restrictions differed from Stage 3 restrictions in the past where greater garden watering was permitted. Therefore, the simulation model was able to identify this discrepancy, and as such the simulation results in Figure 1 use the actual Stage 3 water restriction definitions that applied. Validating bounce-back and transferability In order to validate bounce-back and transferability of the model, in 2009 the model was applied to the Bendigo region which had already commenced easing water restrictions. Results shown in Figure 2 show a high level of accuracy -- greater than 96% in any given year. Validation results in 2009 when restrictions were eased show an accuracy of greater than 99% in forecasting demand. As with the Ballarat validation in the previous section, the forecasts were able to identify discrepancies in watering times with current and past water restrictions. Validating bounce-back -- Part 2 As further validation of the model to predict bounce-back, a scenario was run in the Ballarat region using restrictions that were implemented in Horsham and Ararat, to determine if forecasts show a similar increase in water consumption as water restrictions are eased. Horsham and Ararat moved from Stage 4 to Stage 1 in October 2009, and observed a 35% and 28% increase in water consumption, respectively, between November 2009 and April 2010. Forecasts applying the same restrictions in Ballarat show a similar increase (34%) in water consumption as seen in Horsham and Ararat. Evaluating Past Conservation Measures Overview The CHW "forecasting" model was first used to retrospectively assess and quantify the effectiveness of past conservation measures that were implemented in the Ballarat region, including the Project Aquarius retro-fit program, Target 150, which provides households with feedback on their water usage compared with the target of 150L/p/d, and other influences from media and marketing communicating the severity of the water situation. The model was also used to assess past consumer behaviour and water demand in Ballarat (from July 2006 to June 2009) if particular water conservation measures had not been implemented. This allowed evaluation of the effectiveness of individual water conservation programs, using savings in water, energy, carbon and household bills as measures of effectiveness. Quantifying Past Conservation Measures Results revealed that the Project Aquarius retro-fit program provided the greatest individual household savings (37.4KL), but provided the lowest overall region- wide savings (200ML over three years) as only a small proportion of households participated in the program. The greatest region-wide savings of 1.9GL over three years resulted from voluntary behaviour change of consumers in response to communications about the severity of the water situation. Feedback about water usage also contributed to 312ML of water savings, providing consumers greater awareness of their water usage. The water conservation programs also provided energy and carbon savings by households, and thus provided savings to both their water and energy bills. Figure 2: Historical demand forecasts (red) versus actual demand (grey) in Bendigo.
Water Journal April 2011
Water Journal July 2011