Incorporating Temperature and Precipitation Trends in Long-Term Planning

This summer, the Western Interstate Energy Board (WIEB) hosted two Shultz Energy fellows from Stanford University to analyze trends in temperature and precipitation data in the West and make recommendations on how utilities, regulators, and policymakers can better account for climate change in long-term planning. The Shultz Energy fellows modeled temperature and precipitation trends and extreme weather events, explored modeling and uncertainty analysis techniques, and carried out expert interviews with utility planners, meteorologists, and load forecasters.

The Stanford University 2022 Shultz Energy fellows are:

• Jake Hofgard, B.S. candidate in Mathematics, Minor in Electrical Engineering
• Evan Savage, M.S. candidate in Atmosphere/Energy in Civil and Environmental Engineering

On August 19, 2022,  the fellows conducted a webinar to share their research findings and provide recommendations on incremental policy steps.

A copy of the slides is available HERE.

The forecasting software can be found HERE.

Description: Our temperature forecasting package generalizes the peak temperature forecasting methodologies from the Northwest Power and Conservation Council (NWPCC) and Puget Sound Energy (PSE) so that planners, forecasters, and other interested parties can produce accurate, long-term temperature forecasts for any location in the United States.

The package, called peaktemp, can produce peak temperature forecasts (the temperature that results in peak load) for any location in the United States using data from any CMIP6 climate model (e.g., MIROC6, CanESM5, etc.) and any CMIP6-supported climate scenario (e.g., SSP5-8.5). Forecasts are supported out to the year 2100 (although the recommended forecast range is 30 years), and in-depth comparisons between different climate scenarios are also supported. In addition to a variety of statistical forecasts for peak temperatures, an experimental machine learning-based forecast for peak temperatures, using XGBoost, is available as part of peaktemp.