Conversation

I object to this chart, because it turns out that the reconstruction of a single location based on an ice core drilled 40 years ago in Greenland that ends in 1885 is not a good representation of global temperatures. If we use hundreds of sites and proxy records worldwide we get:

Quote Tweet

Dr Jordan B Peterson

@jordanbpeterson

Jan 15

Anyone object to this chart? twitter.com/ClimateThere/s…

545

5,231

Zeke Hausfather

@hausfath

Jan 15

Here are the climate proxies used in more recent holocene reconstructions (vs a single dot in Greenland that Peterson is showing): nature.com/articles/s4159

385

Zeke Hausfather

@hausfath

Jan 15

More broadly, even if we just focus on Greenland and combine all the different ice core records available there we get a different picture than Peterson portrays, as I discuss in this

@CarbonBrief

analysis: carbonbrief.org/factcheck-what

302

Zeke Hausfather

@hausfath

But no need to take my word on it; listen to Dr. Richard Alley, who put together the record Peterson is sharing: dotearth.blogs.nytimes.com/2010/02/08/ric

5:20 PM · Jan 15, 2023

22.8K

Views

Replying to

. I'm sure you read this thread ☝️. Now like a good scientist, you are going to post a retraction of any implied endorsement of the aforementioned chart, put that data into context, weigh ALL available data, and revise your views? Right? eh.... right?

Show replies

Replying to

We already discussed this. Those hundreds of proxies are calibrated against each other. What is the margin of error? There may be a systematic error from the calibration process. You cannot simply assume that if you take N proxies, the error goes as ~ 1 / sqrt(N).

Replying to

The composite proxy record is low resolution & screened to remove proxies showing no signal. You can not reliably compare this with the instrumental data:

climateaudit.org

The IPCC AR6 Hockeystick

Although climate scientists keep telling that defects in their “hockey stick” proxy reconstructions don’t matter – that it doesn’t matter whether they use data upside …