OK, so I'm looking at WUWT,

Ah, I think I see where you went wrong..

But seriously, yes this is an interesting paper, however it seems to be a synthesis of some things we already know, for example that measurement coverage is incomplete, a factor which lead to the Cowtan and Way dataset (you'll see that Cowtan is an author of this paper), and that over the oceans, the global mean temperature series (eg HADCRUT) use sea surface temperatures while the models simulate surface air temperature.

Models suggest that air-temperature warming is 24% greater than observed by HadCRUT4 over 1861–2009 because slower-warming regions are preferentially sampled and water warms less than air5. Correcting for these biases and accounting for wider uncertainties in radiative forcing based on recent evidence, we infer an observation-based best estimate for TCR of 1.66 °C, with a 5–95% range of 1.0–3.3 °C, consistent with the climate models considered in the IPCC 5th Assessment Report.

http://www-users.york.ac.uk/~kdc3/papers/coverage2013/background.html
http://www-users.york.ac.uk/~kdc3/papers/robust2015/background.html

So, real measurements are scare, combined together in bizarre ways to give a disputed result.

Answer: make same errors/deficiencies in the model to see if it matches!

Do two wrongs make a right? I don't think so.

If your methods of manipulating data can give you different results then the choice of method is paramount.

We know many of the methods used to collect and process temperature data is suspect to say the least.

Phil's already covered this, but the basic result of this paper is pretty straightforward. If you want to do a proper comparison of models and observations, you should aim to compare like with like. The temperature observations suffer from coverage bias (not all regions are equally sampled with some faster warming regions - Arctic - undersampled) and the observations tend to be air temperature over land and sea surface temperatures over the oceans.

Often what is presented from the models is air temperatures everywhere and this is then compared with observations and found to be somewhat discrepant. What this new paper is showing is that if you do a like-for-like comparison (by using air temperatures over land, sea surface temperatures over the oceans, and accounting for coverage bias) then the models-obs discrepancy largely goes away.

This has a few basic consequences. It suggests that coverage bias plus using sea surface temperatures over the oceans under represents warming by maybe as much as 24% compared to what would be the case if we had air temperatures everywhere. It also means that if you correct for this you get an air temperature only TCR of about 1.7K with a 5-95% range from 1K - 3.3K, pretty similar to what the models suggest. Hence there is no real indication that the models are projecting more warming than is likely.

Can the models do sea temps? Do they get them right on an areal basis, never mind predicting? Are the models in the paper tuned to match past observations identical in composition to what they are predicting, IOW did they redo the whole thing with the whole ensemble or did they look at different parts of already-existing output?

And why don't they chuck out that Canadian one?

As I understand it, the models do sea surface temperatures, surface temperatures (which is sea surface over the oceans and land surface on land) and air temperatures (which is meant to represent the air temperature just above the surface). I do not think they reran the models. They simply used the output from the CMIP5 ensemble runs.

Models which were able to hindcast the observed GMT the old way would need to be tuned to match the new criteria for comparison. Once the hindcast is right you can look at the forecast. I don't think you can shortcut it.

Look at Figure 1a in their paper.

Paywalled. I did find this:

Models suggest that air-temperature warming is 24% greater than observed by HadCRUT4 over 1861–2009 because slower-warming regions are preferentially sampled and water warms less than air.

This is no different to the old statement that the warming is where the thermometers ain't. And it always is. This reeks of cherry-picking and a paper designed to achieve a conclusion. Now, why weren't we doing it properly before? Why are we still using totally unsatisfactory averages as a yardstick rather than more precise measurements?

Here is Figure 1a.

Thank you. A lot of excursions pre-1920. So we are using CMIP5. Itself a problem because of the ensemble thing. Where are the other temp series? Did they try them all and pick the most suitable or is only HADCRUT4 appropriate and if so why?

What we have here, it seems to me (usual disclaimers) is a correctly stated problem but a totally inadequate methodology aimed at doing the job with what is to hand rather than do it properly, as if the aim was to get a figure to counter Nic Lewis's work

Jul 25, 2016 at 9:15 PM | Unregistered Commenter...and Then There's Physics

As you know the models, what actually caused the 2 dips seen in the HADCRUT4 record in the 1860+ record??

In relation to real global warming, I am in Texas right now and it is damn hot!!!

Dr. Kevin Cowtan has added some commentary on his blog

When we treat the models like the observations, we get a lower estimate of climate sensitivity. When we treat the observations like the models, we get a higher value. In both cases the models and the observations agree. But which is right?

[…]

The largest contribution to the low observational estimates of TCR is the incomplete global coverage of historical temperature observations (Figure 2). Applying historical coverage to climate model outputs reduces the temperature change by about 15%. (This is larger than the change estimated by infilling the unobserved regions as in Cowtan and Way 2013, because the early record is too incomplete for infilling to be fully effective.)

The next largest effect is the use of sea surface temperatures rather than air temperatures in the observational record. If the climate models are analyzed using both sea and air temperatures rather than air temperatures alone (as required by the formal definition of TCR), the temperature change is reduced by a little under 5%.

The final effect arises from the blending of air and sea temperatures in regions where the sea ice edge has changed. This effect is the most uncertain, but has the smallest effect; less than 5%.
When combined, these three factors reduce the temperature change in the climate model outputs by about a quarter. The different handling of the temperature data between the models and observations therefore explains almost all of the difference between the estimates of climate sensitivity from models and observations.

http://www-users.york.ac.uk/~kdc3/papers/reconciled2016/background.html

As has Ed Hawkins

http://www.climate-lab-book.ac.uk/2016/reconciling-estimates-of-climate-sensitivity/

Yep, the heating is where the thermometers ain't. QED. But surely there is no disaster at CS=1.66? (If you accept CS as a predictive tool, which I don't.)

rhoda, they fantasize about Procustes, but their featherbed sticks together with wax.
========