Othello: “But this denoted a foregone conclusion.”
Iago: “Tis a shrewd doubt, though it be but a dream
And this may help to thicken other proofs
That do demonstrate thinly.” Othello, Act 3, Scene 3
Shakespeare’s tragedy centers on the theme of conclusions hastily arrived at, reflecting prior suspicions but unsupported by the available evidence. That same theme runs through much of the last week of snap judgments about the dire events in Texas and their implications for EU power sector reliability and decarbonisation. Headlines such as this, this, this and this make it clear that, while Texas may seem far away, this episode has been aggressively seized upon across Europe, rightly or wrongly, as fodder for multiple European narratives.
We have hesitated until now to join the fray, awaiting at least some minimal level of reliable information from reliable sources before offering diagnoses and recommendations. This has been a traumatic event for many people, and there has been an unseemly rush in some cases to leverage it for political gain. Yet while it remains early days, some things have become clear.
There have been some excellent fact-based analyses released over the past few days (three of the best are here, here and here), and we do not need to cover the same ground other than to highlight the most salient facts.
How extreme was the need for resources?
First, last week’s storms in the central regions of North America were of historic dimensions. For example, between 11 and 16 February, 50 locations across 11 Midwestern states saw the lowest temperatures ever recorded. While we must recognise the strong probability that extreme weather events such as this are likely to be more frequent worldwide, this was a freak storm by any objective standard. Before examining what should have been done better, it is worth reminding ourselves that widespread, long-lasting power system emergencies with staggering consequences have been recorded multiple times just in the past 25 years, including in Europe. Many of these disasters proceeded from weather events that were more historically predictable than the one that hit the Midwest of the U.S. last week.
Texas electricity demand during the storm peaked on the 14th and 15th at about 74GW, versus a normal February peak of about 55GW, a difference greater than total peak demand in The Netherlands in 2020. 74GW is unheard of in winter (the Texas system operator ERCOT’s “extreme” planning scenario assumed 67.2GW, based on the last severe event in February 2011), but it is typical of recent summer peaks, which ERCOT has served reliably for many years. Yet investment in generation was considered to be sufficient to weather the storm. State and federal reliability authorities reported prior to the winter season a quantity of investment in generating capacity sufficient to meet ERCOT’s “extreme” planning case. Actual conditions would have called for some amount of rotating service interruptions, which should be expected for an event of this historical rarity. Unfortunately, much of that planned capacity failed to show up.
What went wrong?
The failure of generating capacity is perhaps the most visible and yet most misrepresented aspect of the catastrophe, with many European politicians rushing to pin the blame for the blackouts on Texas’s significant investment in wind power. The relevant baseline is the resource plan ERCOT had in place for severe winter weather. Wind capacity constituted only a small share of the plan (6GW, or about 7%) and, while some frozen wind turbines were an issue, the shortfall against plan was only about 2GW on average, and wind generation exceeded the “low wind” contingency case in all but a few hours.
This was overwhelmingly a story of the failure of supposedly reliable thermal generation, most of it gas-fired. ERCOT’s resource plan included about 70GW of thermal generation (57.7GW in their “high forced outage” planning case); on the 15th only about 42GW were available. Coal generation was at 60% of planned capacity. One of the state’s four nuclear power trains tripped offline hours into the event, probably due to inadequate freeze protection, and did not return to full capacity until the event was largely over. Gas plants made up 55GW of planned resources, but only 31GW were available on the 15th, representing 80% of the resource shortfall.
It remains to be determined how much of the gas generation unavailability was due to plant freeze-ups and how much to fuel supply issues, but frozen gas infrastructure was a major factor. North American gas production dropped 15% during the storm, and South Central production dropped 20%. Gulf Coast and West Texas wells, pipelines and gas processing facilities, the sole fuel source for most of the 53% of ERCOT resources that are gas-fired, are unprepared for such conditions, lightly regulated and beyond the authority of either ERCOT or the Texas utility regulator. As analysts from the International Energy Agency noted immediately after the event, “Texas has a power shortage because Texas has a gas shortage.”
What are the implications?
Beyond the well-documented human misery endured by many in and beyond Texas, there has also been a seismic financial impact. Reports of some individuals facing incredible bills have been common, but the vast majority of Texas consumers will have been hedged by their suppliers. The toll among suppliers, as a result, is likely to be catastrophic, with few expected to survive and with adverse longer-term consequences for all consumers.
What lessons must be learned? Certainly not that the problem was the lack of a capacity market. Perhaps the best known of the U.S. capacity markets, in the PJM market, experienced a nearly identical failure of supposedly firm fossil generating capacity during a severe winter storm in 2014, largely for the same reason — a failure to properly account for the vulnerability of generator fuel supply chains. ERCOT’s installed reserve margin has been sufficient to reliably serve comparable summer demand peaks year in and year out. Building yet more fossil generating plant with the same vulnerability to the region’s fuel supply chain would have been of virtually no help. Certainly not that “renewables were the problem” – the variability of wind is well understood and had been planned for, and we know how to plan for it in the future. Here are some emerging recommendations of relevance to European policy makers:
- Texas ranks 29th among U.S. states in energy efficiency; aggressive building efficiency standards and efficient electric heating systems can reduce both the demand for energy during extreme events and the speed with which homes become uninhabitable.
- This event revealed an underlying inability for much of potentially flexible demand in Texas to respond to tight supply. It offers a stark reminder that regulators and policy makers should make a concerted effort to enable more demand flexibility through rate designs and technology deployment as a low-cost hedge against extreme volatility.
- As in the recent EU Clean Energy for All package, ERCOT’s market is designed to provide for adequate investment through scarcity pricing, and ERCOT has demonstrated a strong record of its effectiveness in mobilising action and investment to meet demand reliably during its summer peak season. Yet beyond some duration, the actions scarcity pricing is meant to motivate have been largely exhausted, after which it is ineffectual. A pricing circuit breaker such as that employed in Australia would preserve desired functionality while avoiding unintended consequences such as those resulting from this episode.
- Just as payment for reserve capacity in “capacity markets” relies upon an accurate accounting of reliable reserves, so does the administrative scarcity pricing mechanism employed by ERCOT for the same purpose — a mechanism similar to the cash-out mechanism adopted in Great Britain’s market in 2016. For any market design to ensure adequate resources, regulators must set clear standards for reliable capacity value of system resources, starting with the European Resource Adequacy Assessment process.
- There is now ample scientific and empirical evidence for revisiting what sort of climate conditions are considered within, versus beyond, the limits of prudent planning. Whether the severity of this event would have fallen within those limits is a valid question, but the limits must be expanded to encompass the likely effects of climate change, which would have mitigated at least the worst consequences of this episode.
- An emerging strategy for resilience is to enable local communities to isolate from the bulk system and continue to serve at least a minimum level of essential services for some period of time. While net benefits depend on local circumstances, it is an option worth investigating.
It is our hope that, as we digest and act upon lessons from this tragic episode, we strive not “to thicken other proofs that do demonstrate thinly.”
A version of this article originally appeared on Euractiv.