"During 2009-2010, RTI International developed and published a financial model which calculates the financially optimum level of Non-Revenue Water (NRW). The model computes a target (under future steady-state conditions) for NRW reduction and control programs, based on site conditions and local costs. The model was presented at IWA Water Loss Specialist Conferences in Cape Town and Sao Paulo and international meetings in Morocco, Uganda, Jordan and the UK. It has been applied in over 30 countries using secondary data, and applied at a more detailed level in Brazil, Jordan, Uganda, and Zambia.
During 2010-2011, RTI International worked closely with 2 water supply utilities, in Jordan and Brazil on customized in-depth application of the model. The authors of this paper engaged in extensive discussions and data exchange to adapt, apply, and refine the original model to the two utilities in the two countries. The paper first outlines the water resource, water distribution, economic and social conditions in the locations in Brazil and Jordan. Next, the paper reviews historical information on NRW reduction and control programs, including the level of NRW (using several different indicators) as well as the strategies, actions, programs and management approaches included in the NRW efforts implemented by each utility.
The paper then reviews the process of adapting the model to Jordan and Brazil, including a review of terminology – in order to ensure that data and formulae are consistent across the model and the databases in each utility / country. Data availability was not a major constraint – good estimates of model inputs parameters could be easily made. The paper next reviews the model-predicted optimal NRW levels in the two locations. In both cases the utilities have done a good job reducing losses, but each still have room to reduce losses more to reach the optimal level. The target NRW levels in the two locations were actually quite close to each other – despite considerable variation in certain inputs. The model behavior and results were consistent with previous applications. The model sensitivity and confidence levels are also carefully reviewed in the paper, again with similar results in the two locations.
The paper concludes with an assessment of lessons learned from the field applications and a description of future actions by the 2 utilities, and recommended research and development to further refine the model and develop additional, complementary tools."