Hot and Cold: The Varying Effects of Extreme Weather Events on Battery Revenues

Key Points

• Gas prices in southern California rose sharply in the final weeks of 2022, owing to several supply- and demand-side conditions.
• Because power prices saw upward movement across all hours of the day during the December cold weather event, the added benefit to battery energy arbitrage was muted.
• Extreme weather events and uncertainty around resource availability highlight the growing challenge around seasonal resource planning.

Ascend's Analysis

In November 2022, gas prices began to rise sharply in California. In the final days of December, Winter Storm Elliott brought extremely cold weather to much of the United States, increasing heating demand and putting upward pressure on already elevated power and gas prices. California gas markets were particularly strained, with PG&E day-ahead gas prices rising by over 210% from November to December, eventually peaking over $56/MMBtu (Figure 1). These high prices were the result of several coincident circumstances, including:

• Extremely cold temperatures and related increased gas demand for heating
• Pipeline disruptions and reduced gas imports
• Low gas storage levels in California
• Low hydro imports from the Pacific Northwest

One significant driver of increased gas prices was natural gas import capacity to California, which has been limited by a constraint in the El Paso gas network. In 2021, an explosion at the Coolidge pipeline stopped operations and required the line to operate at reduced capacity while undergoing maintenance. The limited throughput and California’s geographic position near the end of the gas network tightened supply in the state and put additional upward pressure on prices. Additionally, monthly underground gas storage levels were at their lowest in years in the Pacific region (Figure 2), which includes California, Oregon, and Washington. While most of the underground gas storage in the Pacific region is in California, the state’s storage capacity decreased in 2017 after the California Public Utilities Commission (CPUC) lowered the maximum limit of the Aliso Canyon natural gas storage facility from 86 billion cubic feet (bcf) to 34 bcf in response to a large gas leak in 2015. The CPUC increased the limit to about 41 bcf in 2021, but the loss of significant storage capacity has left the state more vulnerable to gas price volatility and shortage events, which contributed to historically high prices.

Hydropower production in California and the Pacific Northwest (PNW) also contributed to the sharp increase in gas prices. With abundant water resources, the PNW generates most of its power from hydroelectric facilities and often sends excess power south to California when price signals incentivize it. This winter, however, cold weather and high local demand caused power prices in the PNW to rise to keep power in the region. Power imports to California averaged about 4.5 GW during the storm, a significant decrease from the approximately 9.5 GW of imports during the same period in 2020 and 2021. As a result, California load-serving entities were forced to dip into the already thin gas storage to generate electricity.

Load serving entities (LSE) operating in California are already working to decarbonize on the timeline mandated by state policy or faster, so elevated gas prices are unlikely to significantly impact their operations or long-term resource plans. However, recent market volatility has re-emphasized the importance of hedging load to ensure that LSEs and their customers are adequately insulated from high prices when they occur. CCAs do not usually hedge their full load, and the remaining unhedged portion can increase costs when exposed to high gas and power prices.

Despite high power prices, batteries in CAISO did not profit as much during the winter storm as they did during the September heat wave, which lasted from September 5-10. Energy arbitrage opportunities depend on price volatility and daily spreads. The average daily real-time price volatility during the September heat wave was more than 230% higher than during the winter storm(measured by RTB120, Ascend’s metric for real-time power price volatility),leading to significantly higher battery revenues. As shown in Figure 3, power price differentials between the mid-day and evening hours were stretched during the heatwave. Solar, hydro, and imports kept prices lower during the day, but record-setting load and the evening solar drop-off led to high prices in the evening hours, served largely by gas units. During winter storm Elliott, less abundant solar and hydro, along with high import prices from the PNW, pushed gas onto the margin across all hours, causing a significant increase in power prices throughout the day and diminishing the price differentials.

The price shapes in Figure 3 highlight the impact of gas prices and renewable generation on power prices. The figure also shows ‘levelized’ prices, which are shifted to set their minimum value at zero to better compare the absolute price spread. Price spreads soared during the September heat wave as the system transitioned from low-cost resources on the margin during daylight hours to high-cost gas on the margin during the evening hours, which saw record-levels of demand. During the December event, limited solar generation coupled with low imports from PNW pushed gas onto the margin during solar hours more frequently, muting the price spreads during those days. Power price dynamics during these extreme weather events highlight that gas prices alone do not dictate battery revenues. The market dynamics of solar, hydro, imports, and load contribute to a price differential between the highest- and lowest-priced hours of the day.  

The September heat wave and December cold snap highlight how critical hours on the grid will increasingly be defined not only by the absence of renewable production, but the duration of their absence. While gas currently serves a critical role in meeting demand during these peak hours, the state’s decarbonization efforts heighten the need to introduce incentives that can bring clean, longer-duration resources online.

___________________________________________________

¹ Normalized price shapes were made by subtracting the lowest price of each day from each hourly price of that day. These hourly prices were then averaged for the duration of the weather event.