Sometimes it would be nice to have a crystal ball to predict the future — to know whether to bring an umbrella or sunscreen. And while that’s more a job for meteorologists, Tampa Bay Water also has to understand weather patterns and many other factors to be able to accurately forecast water usage.
We’re proud to say Tampa Bay Water is greatly accurate in predicting water usage, according to a study by the Water Research Foundation. Not only are Tampa Bay Water’s short-term demand forecasting models more accurate in determining one-week-ahead water demand than other utilities’ similar models, our forecasting models outperformed a variety of benchmark statistics. Being able to forecast week-ahead water demand is critical to utility planning; forecasts help reduce costs by streamlining utility operations and improving water efficiency. These models are a major part of the Tampa Bay Water’s weekly source allocation model: Optimized Regional Operations Plan (OROP).
Conducted by researchers at the University of Texas at El Paso, the study looked at demand forecasting models used by four utilities as part of the ongoing Water Research Foundation project, “Analysis of the Effectiveness of Short-Term Demand Forecasting and Recommendations for Improvement.” The study used a variety of benchmark statistical performance metrics to look at forecast and actual observations over a period of four years. Each participating utility provided forecast and actual observation data to the research team.
The independent study showed Tampa Bay Water’s short-term demand models significantly outperformed key benchmark statistics for model accuracy in a variety of categories while also showing which locations in the utility’s system could use enhancement to improve modeling outcomes.
“Studies like these are important because they reinforce the type of computer models and tools we develop to make weekly operational decisions and provide feedback on where we can improve,” said Tirusew Asefa, Ph.D., P.E., D.WRE, Tampa Bay Water’s modeling and systems decision support manager. “We work really hard and gather a great amount of data to predict the water needs in our tri-county service area. It’s gratifying to have our efforts be noticed.”