Thread by @AndrewDessler

To follow up on @douglewinenergy's tweet about @ERCOT_ISO chronically underestimating demand, let's talk about ERCOT's predictions for this summer.

A 🧵:

https://twitter.com/douglewinenergy/status/1536139939242463232

ERCOT predicted seasonal maximum power (SMP, the hour of highest demand during a season) during June-July-August (JJA) of 77.3 GW. They predict an "extreme" max demand of 81.6 GW.

You can see details of ERCOT's forecast here: ercot.com/files/docs/202…

Of particular importance, ERCOT's forecasts are based only on historical temperatures, which we know are not good indicators of the future.

My student Jangho Lee and I have developed power forecasts that use climate models to incorporate the impact of climate change, as well as allow us to take a more statistical view of the future. Here is the probability distribution of JJA 2022 seasonal maximum power (SMP).

Our median estimate is (78.4 GW) is very close to ERCOT's base case (77.3 GW).

We predict that their "extreme" scenario (81.6 GW) is slightly below our 90th percentile, so we should expect to exceed that about every decade or so.

Note that ERCOT does not provide any probability for their "extreme" scenario, but in a conversation with one of their analysts, he agreed that their "extreme" scenario was about a 90th-percentile event.

We can argue about what an "extreme" scenario is, but you cannot argue that ERCOT has failed to tell people what their "extreme" scenario represents. This seems to be a pretty big oversight on ERCOT's part and it's hard not to speculate on why they do that.

In a 1-in-20 year event, demand could reach 83.7 GW and in a 1-in-100 year event (like winter storm Uri), demand could reach 86.4 GW.

The good news is that ERCOT estimates total available power is 91.4 GW. So if all generators produce as expected, there should be enough power.

But, of course, "as expected" does not always occur on the ERCOT grid. Their worst-case scenario for unplanned thermal outages (coal, gas, nuclear) is 13.7 GW and their worst-case scenario for underproduction of renewable power is 9.1 GW.

A loss of 22.8 GW would cause blackouts even on the median SMP. Less severe loss of generation could easily cause blackouts during a 1-in-10 year SMP.

And loss of thermal power seems to be happening more and more frequently.

https://twitter.com/douglewinenergy/status/1536110319394869248?s=20&t=CZNJisenDSsKTKbx-GfKYg

If you look critically at the numbers, you cannot help but feel that our grid is on the ragged edge. It seems like ERCOT should do a better job explaining to people the situation rather than try to tell people that everything is fine.

To do that, though, ERCOT needs to incorporate climate change into their forecasts. @TexasObserver
texasobserver.org/texas-insane-h…

This also underscores the fact that there is NO MARGIN for error on the ERCOT grid. If things go even slightly wrong on the grid, Texans are going to be without air conditioning. We need to ask ourselves if that's the world we want to live in.

Our forecast methodology is described here: Lee & Dessler, The impact of neglecting climate change and variability on ERCOT’s forecasts of electricity demand in Texas, Weather, Climate, and Society, 14, doi:10.1175/WCAS-D-21-0140.1, 2022.
Reprint: drive.google.com/file/d/10jVEDb…

The forecast above uses temperatures come from the NCAR CESM-LE between 2016-2025; non-climate factors (P_ref) are extrapolated to 2022 from historical data (Fig. 2b of the paper).

Explainer video of the methodology: