Summary: Weather regimes – i.e., recurring large-scale circulation patterns in the atmosphere – have long been used for subseasonal (10-30 day) forecasts of power generation and demand in Europe. For Europe, the focus is on at least four weather regimes spanning the Atlantic and European continent. Similarly, recently-derived North American winter weather regimes can be incorporated into subseasonal forecasts for North America to better predict winter peak loads, as we will show is possible for the Southwest Power Pool (SPP). Salient improves its forecasting by continuously incorporating the latest scientific research such as my work.

Why Subseasonal Forecasts Matter for the Energy Industry

Winter brings several different kinds of weather extremes — snowstorms, ice storms, and extreme cold air outbreaks. On the demand side, extreme cold snaps or sudden warm-ups lead to sharp swings in electricity consumption, creating challenges for decision-makers managing peak loads and price volatility. For power producers, severe winter weather not only threatens physical assets like power lines and substations but also makes maintaining consistent supply difficult. 

According to the North American Electric Reliability Corporation (NERC), extreme winter events pose significant risks to power generation. NERC recently flagged the Southwest Power Pool (SPP)—which covers the Central U.S.—as especially vulnerable for the winter of 2023-2024 and facing elevated risks for threats to electricity supplies over the next decade. Managing this growing risk demands smarter forecasting to optimize power trading strategies and grid operations.

This is where subseasonal weather forecasts come in. Skillful and reliable forecasts that anticipate extreme weather events 10 to 30 days in advance (i.e., the subseasonal-to-seasonal, or S2S, range) can be a game-changer. They enable power traders, utilities, and regional operators to make better-informed decisions around energy purchases, demand planning, and risk hedging strategies—especially in regions prone to high winter weather variability like SPP.

This is not news. Citadel and other power and gas traders understand that skillful, reliable forecasts leveraged in a strong decision-support system may be the most valuable commodity of all. So let’s dig into some of the science that informs skill and reliability. 

Salient Incorporates Cutting-Edge Research Into Their Subseasonal Models

Salient Predictions is integrating my meteorological research from the University of Oklahoma (OU) alongside other R&D into its forecasting models to help the energy industry predict and manage winter demand more effectively. This research on electricity demand predictions and winter weather strongly leverages knowledge of weather regimes—i.e., recurring large-scale atmospheric patterns that provide clues about upcoming temperature patterns and, subsequently, electricity demand shifts.

While weather regimes have long been used for power generation and demand in Europe, they are now gaining traction in the U.S.

What Are Weather Regimes, and Why Do They Matter?

Weather regimes simplify complex atmospheric behavior by identifying large-scale circulation patterns that persist over days to weeks. Think of these regimes as the atmosphere’s “moods”—patterns that tend to repeat and signal certain types of weather. For example, a weather regime where the polar jet stream shifts south could indicate colder temperatures across the U.S., increasing heating demand and pushing electricity loads higher.

The OU research team identified five (5) key winter regimes, each associated with distinct jet stream patterns and temperature outcomes across the U.S. For example:

• Alaska Ridge (AkR) & Arctic High (ArH): These regimes drive cold air south, increasing the likelihood of extreme cold events and surging electricity demand across the Central US.‍
• Pacific Trough (PT) & Arctic Low (ArL): These regimes bring milder air across much of the Lower 48, reducing the need for heating and keeping power demand low.

Predicting Winter Peak Load Demand in the Southwest Power Pool

Temperature is a significant driver for electricity usage. The relationship between the two during the cold season is fairly straightforward (albeit non-linear): as temperatures drop below freezing, electricity demand increases rapidly. Conversely, milder winter weather (i.e., temperatures below 18°C) reduce demand, as heating or cooling needs are minimal.

With this relationship illustrated, we now show how integrating weather regimes into probabilistic forecasting models can provide traders with actionable forecasts of peak demand spikes or dips. For example:

• AkR and WCR regime days are more prone to above average, or even extremely high, peak loads within the Southwest Power Pool. In fact, on AkR and WCR regime days, the Southwest Power Pool has a 50-100% higher risk than climatology of exceeding extreme peak loads.
• PT and ArL regime days feature reduced risk for excessive peak load (i.e., below normal peak demand) - i.e, a 50-90% lower than normal chance.

Altogether, these findings, and current efforts to integrate these findings into Salient’s models, go beyond just predicting temperature—they provide a probabilistic framework that assesses the likelihood of extreme load events on specific days.

So what? Once a decision-maker has a skillful, reliable probabilistic forecast in hand that effectively incorporates weather regimes to inform potential peak load days, what do they do? How to drive business value is user-dependent.

• For utilities, decision-makers may prepare by adding additional staff and line workers, and coordination with neighboring utilities. 
• Traders will build an appropriate hedging strategy for their position
• Industrial customers in the affected areas may shift production or begin preparing their factories/buildings for potential extended outages. 
• The oil and gas industry might harden operations or plan to shut down production.

Conclusion

As the Southwest Power Pool and other regions face growing winter demand, investing in reliable S2S forecasts becomes essential. While on my sabbatical from OU with Salient Predictions (through the end of 2024), I am working with Salient’s team to integrate weather regime analysis into its subseasonal (10-30 day) forecasting models, providing a valuable tool for energy industry leaders. With these forecasts, decision-makers can anticipate peak load risks more accurately, optimize hedging strategies, and reduce exposure to unexpected price swings. 

Dig deeper by reading our paper: Millin, O. T., J. C. Furtado, and C. Malloy, 2024: The impact of North American winter weather regimes on electricity load in the Central United States. npj Climate Atmos. Sci., 7, 254, https://doi.org/10.1038/s41612-024-00803-1.