Thread by @AndrewDessler

More on @ERCOT_ISO and the Texas grid. In a previous 🧵, we talked about how the supply of energy on the TX grid is very tight. This is not ERCOT's fault — it's a fault of the way the market is set up.

But that doesn't mean ERCOT is blameless. A 🧵:

https://twitter.com/AndrewDessler/status/1439349679247011841?s=20

ERCOT makes seasonal forecasts in order to ensure that supply is sufficient for the demand. You can find them here: ercot.com/gridinfo/resou…

For the last winter, we can compare these forecasts to forecasts we make from a large ensemble of climate model runs and to reality. More info can be found in the preprint written by my grad student, Jangho Lee (eartharxiv.org/repository/vie…).

Here is the money plot. It is a plot of the wintertime seasonal maximum power demand (the hour of highest demand during winter). The blue bars show a probability distribution we created from a climate model ensemble + a statistical model. See our paper for more info.

Prior to the winter, ERCOT estimated a base-case (most likely) maximum demand of 57 GW (the red line). That's very close to our model's estimates of the most likely value. And with 83 GW of power available (black dashed line), there is a comfortable margin.

They also estimated an "extreme peak load scenario", which is the gray dotted line — 67 GW. This is based on 2011 temperatures. They did not consider the possibility that it could get colder than that year.

It's time for a digression here: ERCOT only uses historical temperatures in its forecasts. They do not consider climate change or variability outside of the historical record when making forecasts.

This is from an interview with @TexasMonthly:

https://twitter.com/Forrest4Trees/status/1406322654559223818?s=20

I'll let y'all speculate about the answer ERCOT's former CEO Bill Magness gave to that question. But I will note he acknowledges ignoring climate change but leaves hanging the question of why.

Back to the figure. The climate model produces many colder winters than 2011. In fact, 19% of the winters have a seasonal max power demand higher than 67 GW. The model's coldest winter has a demand of about 90 GW.

Nature provides support for our model: winter 2021 was much colder than winter 2011. We estimate that demand during winter reached 82 GW (the green line in the figure above) — 15 GW above ERCOT's extreme estimate.

ERCOT estimates that best-case supply is 83 GW (the black dashed line in the figure), so there was (theoretically) enough power to keep the grid up, but failure of the natural gas energy infrastructure reduced supply and doomed the grid.

https://twitter.com/MichaelEWebber/status/1426877858409033730?s=20

We spent a long time trying to find someone at ERCOT to talk to about our predictions. It was difficult, but we finally did find someone who told us the following: they estimate demand during the winter storm peaked at 76 GW, 6 GW below our estimate.

This seems like a small difference but it's actually huge. If they're right, then there's 7 GW of spare capacity on the grid. If we're right, then there's 1 GW of spare capacity.

In other words, ERCOT's estimate is good for ERCOT because they can say that the TX grid does not have a supply problem.

If we're right, then the supply of power on ERCOT's grid is tight enough that there is zero margin for error on the grid.

We've gone over our calculations and just don't see how ERCOT's estimate of 76 GW can possibly be right. However, they are experts in this and so it's certainly possible that there's something we've overlooked in our calculation.

But their model/calculations are not public so we cannot examine them and resolve the difference between the estimates.

Given the high stakes of this issue, ERCOT needs to improve the transparency of their estimates. They need to make their calculations publicly available.