New Climate Normals for Electricity Demand Forecasting

Hydro-Québec’s distribution division, in collaboration with the Ouranos Consortium, implemented ways to incorporate recent changes in average temperature into energy and peak demand forecast. Outside temperatures are an important factor in electricity demand in Quebec, because many households and businesses rely on electricity for heating and cooling. Hydro-Québec’s distribution division noted back in 2001 that average temperatures during 1991-2000 were clearly warmer than they had been between 1961 and 2000. In the forecasting horizon, demand forecast models should use normal weather assumptions consistent with economic and demographic assumptions. In a warming climate with higher average temperatures than those of a few decades ago, observed records no longer provide the best statistical basis for demand forecasts. To adjust to this change, Hydro-Québec developed methods to improve demand forecasts that incorporate climate change trends.

Understanding and Assessing Impacts

Hydro-Québec’s distribution division noted back in 2001 that average temperatures during 1991-2000 were clearly warmer than they had been between 1961 and 2000. A study of the effects of higher temperatures on electricity demand revealed corresponding shifts in consumption and sales patterns during both winter and summer. The distribution division wanted to better understand these findings and explore the potential relationship with climate change. As a founding member of the Ouranos Consortium on Regional Climatology and Adaptation to Climate Change, Hydro-Québec could readily provide its demand-forecast team with access to climate scenarios and services. In response to the team’s request in 2004, Ouranos scientists developed a regionalized scenario of the evolution of monthly average temperatures using a single climate model simulation. At the distribution division, experts began to explore how to incorporate these long-term monthly temperature trends into demand forecasts, and whether the results would be useful. To develop a way to use climate-model temperature data in Hydro-Québec demand forecasts, a more sophisticated assessment was completed in 2007 using the CMIP3 climate simulations. Based on this analysis, 1970 was defined as a starting point for warming and used in the simplification of warming trends through the 2030s. The distribution division at Hydro-Québec used the results of the climate projections analysis to update the normal weather used in operational forecasting.

Identifying Actions

To keep its database current, Hydro-Québec requested an update of the climate assessment in 2012, this time based on a larger ensemble of 137 climate simulations. The update confirms the robustness of the earlier assessment conducted in 2004, with only minor differences in monthly trends. Hydro-Québec then studied the impact of projected changes of extreme temperatures on the utility’s ability to forecast winter peak demand. For this, Hydro-Québec again relied on the climate experts at Ouranos to explore this topic. Ouranos dedicated several months to studying extreme weather events and shared the results with the scientific community. The research confirmed the hypothesis of faster warming of extreme cold temperatures in Quebec compared to mean winter temperatures.

Implementation

Hydro-Québec now incorporates the findings of research on changes of extreme cold temperatures into its peak forecasting. The experts at Hydro-Québec expect that the model will improve the accuracy of their forecasts during winter, just as the model based on projected warming trends improved their forecasting during the rest of the year.

Next Steps

Hydro-Québec’s distribution division continuously update their climatic normal to reflect impacts of climate change in demand forecast.

Resources