Improving Load Forecasting Techniques: Adapting to Climate Change

Abstract

With the looming threat of climate change, electric utilities need to adapt their current load forecasting techniques so as to generate climate-sensitive load forecasts. This study investigates potential improvements in hourly and monthly load forecasting models by incorporating weather variables. While the hourly models show mixed results across seasons and regions, the monthly model shows marked improvement over a purely auto regressive approach to load forecasting. In light of climate change, electric utilities can avail of economic benefits from minimizing their exposure to the volatile spot market prices and significant losses through inaccuracies in predictions. Moreover, decision-making based on more climate-sensitive forecasts will result in reduction in the carbon footprint of the electric utilities and improvements in their investment strategies for renewable energy technologies for the future.