mastodon_Zeke Hausfather

We've already succeeded in making clean energy cheap. Now we need to ensure that we can actually get it built. Thats why I'm supporting the new permitting reform bill, even if its far from perfect. My latest over at The Climate Brink:

I've been away over the past two weeks in Tanzania for an IPCC working group one lead author meeting, and a subsequent climb of Kilimanjaro with some fellow authors and friends that we dubbed "walking group one". We had an amazing adventure and all made it to the top!

Global temperatures have tied their all-time highest value on record at 2.04C above preindustrial levels (measured as a change from the 1850-1900 period) in the Copernicus/ERA5 dataset. The prior record was set back in November 2023.The world has now experienced seven days above 2C.

Join me live tomorrow at 3 PM UK time for an "ask me anything" on my recent State of the Climate Q3 2026 update over at Carbon Brief:

You can check out the full piece, including what to expect with the record setting El Niño and 2026 annual temperatures, as well as a look ahead to 2027 here:

With William Nordhaus' recent passing the field of climate economics has lost one of its giants. Nordhaus published in 1977 what was at the time a rather bold projection: the the world would warm by 1.2C by the year 2026. In reality it warmed by 1.1C.

October 2026 was the warmest October on record in Copernicus' ERA5 dataset, coming in at 1.75C above preindustrial levels.

With 10 months of the year now in, 2026 as a whole is on track to see the second warmest year on record in ERA5. And with a record-setting El Nino we are very likely to see a record warm 2027:

Google has specifically prevented its AI tool from telling users who won the 2020 presidential election, even though it happily tells us that Trump won in 2016 and 2024. This sort of political intervention into LLM responses is deeply problematic. (h/t to Charles Mann)

It will happily tell you the last there presidents and when they were elected, but only breaks when you specifically ask about 2020. Come on @Google, do better.

In the rush to build data centers, companies are starting to use their own off grid gas generators. This results in much higher emissions – 60% to 80% – compared with using their local electricity grid and risks locking in longer-term fossil dependence.
These generators are much less efficient – and have higher methane leakage – compared with modern combined cycle gas turbines. Building off grid also means that data centers won't benefit from the continuing rapid decarbonization of generation occurring in many parts of the US.

In some regions with high gas leakage or cleaner grids the off-grid penalty is even worse. In West Texas it results in around twice the emissions of an on-grid data center.
Its important to try and reduce the barriers to grid connection by making it easier to build more clean energy and transmission quickly to support new loads – from data centers, electric vehicles, heat pumps, and others. Otherwise we will be stuck with a lot more fossil fuel use.

Check out the full piece over at The Climate Brink:

Great discussion with Olya Irzak on just how high the methane leakage may be from off-grid data centers powered by low-efficiency reciprocating gas engines. Another reason to ensure that if data centers are built, they are on the grid!

Join us tomorrow at 8 AM Pacific for the Berkeley Earth monthly press briefing. You can register here:

While there is increasingly strong evidence that global warming is accelerating, we need to be careful about how we detect it. Unfortunately a newly published paper on the topic makes a number of mistakes in it approach, as I discuss over at The Climate Brink:

When we properly run the numbers we end up just shy of statistically significant acceleration in the surface temperature records (at least without correcting for El Niño, solar, and volcanic variability), though with 2026 shaping up to be a top-2 hottest year this will shortly change:

Electric vehicle adoption is accelerating worldwide – but not in the United States:

Will this year's record-shattering El Niño event be the strongest in a millennium? In a new piece over at The Climate Brink I review the paleoclimate literature from tree rings and corals and argue that may well turn out to be the case:
There is a certain irony in spending $276 million on political "free speech".

As always, you can stay up to date on El Niño observations and models over at the Climate Dashboard:

I've also updated the impacts explorer over there to use the latest (October) dynamical model runs for local forecasts:

The October runs for dynamical El Niño models are now all out, and its a mixed bag. Some revise the event downward, some upward, but on average its still expected to peak at ~4C, albeit with the peak more likely to occur in December than November:

There is a bit of a jump between observed September Oceanic Niño Index (ONI – 2.8C) and projected October ONI (~3.6C) in the models, but this isn't too far out of the realm of possibility as its been averaging about 3.2C over the first third of October:

Here is what all the individual models show in October runs compared with September runs:

And the latest projected El Niño peak for each model and across all models:

And the relative Oceanic Niño Index (RONI) version of that figure:

This year's El Niño continues its climb into uncharted waters; Niño3.4 region anomalies are now at a 3.4C above normal (how fitting!), higher than any prior El Niño event on record. September monthly Niño3.4 anomalies also set a record at 2.9C (beating 2015 and 1877's 2.75C).