Mitigating the Impact of Weather Forecast Error in Electric and Natural Gas Load Forecasting

Premium quality weather information, specifically, hourly temperature forecasts, is one of the most important factors in electric and natural gas load forecasting models operated by utility companies. Up to 90% of the error in load forecasts can be attributed to an imperfect deterministic weather forecast. Utilizing a Bayesian Model Averaging technique, a new generation of forecast content was created, specialized for the utility industry and comprised of hourly probability of exceedance temperature forecasts. This probabilistic dataset helps account for the uncertainty inherent within the current deterministic weather forecasts deployed in load forecasting applications. The methodology and advanced meteorological datasets behind the enhanced forecast technology will be outlined in addition to end-user functionality and results.