It seems like I just got done writing a post which incorporated the point that Real Climate leaves much to be mocked, and low-and-behold Gavin Schmidt deals us a whopper. A fantastic new paper was written as a comment for Nature called “Reconciling Warming Trends”, which proports to explain the lack of observed warming which directly contradicts the bulk of the climate models. The first thing the media should take note of is that these scientists have finally noticed what us evil skeptics have been telling you for several years –the predicted level of warming didn’t happen! It is warming, but not enough to be a problem, and that IS a big problem for the multi-billion dollar climate industry.
As recently as February 2013, Real Climate had their heads in the sand on models with this quote:
The conclusion is the same as in each of the past few years; the models are on the low side of some changes, and on the high side of others, but despite short-term ups and downs, global warming continues much as predicted.
In the meantime, more than this one paper
was being published that claimed the opposite. And recently Roy Spencer made a cute plot for which the only rebuttal I’ve heard is that he chose an inconvenient starting year. Not that it changes the result much:
So for the media who don’t read things like ‘papers’ or data, the blue and green dotted lines have lower slopes than the climate models, therefore the models predicted more warming than was observed. Just like the
Koch funded unfunded skeptics told you.
But this new paper by Gavin A. Schmidt, Drew T. Shindell and Kostas Tsigaridis (Schmidt 14) is a true gem. The crew looked at several observed factors in climate since their last runs and found different values for the years 1990 – 2012. They looked at human aerosols, solar irradiance changes, volcanic aerosols and a “very slightly” modified level of greenhouse gas forcing.
The resulting change in model forcing brought the models in line with observation — almost. Well they still are higher than any actual observation but adjusting moisture feedback (a large and uncertain factor) is not a sanctified IPCC consideration.
Of course they only show the years since 1990 which is hilarious considering that they are addressing a massive failure of the centennial-scale models to predict even a decade into the future. Note that despite the efforts to “find” an explanation, moisture feedback, the greatest unknown in climate modeling, was not even mentioned.
Still, there is one tiny elephant in the Real Climate corner. A claim as specious as the claim Michael Mann makes of being exonerated from wrongdoing by the fake Muir Russel climategate report, yet very often made by the Real Climate crowd.
Climate Models are Not Tuned to Observation
For the heck of it, I searched Real Climate for phrases like – ‘not tuned’.
From RC Frequently asked questions: Are climate models just a fit to the trend in the global temperature data?
No. Much of the confusion concerning this point comes from a misunderstanding stemming from the point above. Model development actually does not use the trend data in tuning (see below).
Gavin comment response: [Response: If you read our papers (and my comments) we are completely up front about what we tune to - the climatology (average annual values), the seasonal cycle, the diurnal cycle and the energy in patterns like the standing wave fields etc. We do not tune to the trends or the sensitivity. - gavin]
Gavin comment response: I’ve said this before, and I’ll say it again, models are not tuned to match long-term time-series data of any sort. – gavin]
Gavin comment response:
[Response: The AR4 runs were done in 2004, using observed concentration data that went up to 2000 (or 2002/3 for a couple of groups). None of them were tuned to the HadCRUT3 temperature data and no model simulations were rejected because they didn't fit. - gavin]
Comment and Gavin response:
It seems clear that each model is tuned to match past temperature trends through individual adjustments to external forcings, feedbacks and internal variability. Then the results from these tuned model are re-presented (via Figure 2 above) as giving strong evidence that nearly all observed warming is anthropogenic as predicted. How could it be anything else ?
[Response: You premise is not true, and so your conclusions do not follow. Despite endless repetition of these claims, models are *not* tuned on the trends over the 20th Century. They just aren't. And in this calculation it wouldn't even be relevant in any case, because the fingerprinting is done without reference to the scale of the response - just the pattern. That the scaling is close to 1 for the models is actually an independent validation of the model sensitivity. - gavin]
What is clear to most of us “skeptics”, and should be very clear to any semi-technical type, is that in modeling, with hundreds of tweakable parameters, if the output doesn’t match the observations, you go back and tweak the input until it does. Gavin’s insistence that models aren’t tuned, is simply his own bias forgetting those hundreds of times when he put CO2 forcing in upside down or with a ridiculous weighting by accident or by test and the result didn’t look at all like he expected, so he adjusted things. He and many others rightfully find it easy to justify the adjustments post hoc - e.g. the paper they just published. It’s not wrong to adjust the model, they should match the data, but they universally, definitely and regularly are adjusted until the output matches some observation.
In this case, the models were so far out of whack, they quietly admitted that the skeptics were right, and adjusted their favorite inputs only. Other inputs were quite thoroughly left out. What is more is that most of the inputs had little effect but by ‘re-analysis’ they made massive corrections to volcanic forcings, only in the recent time-window to correct recent trends.
Oddly enough, I think this sentence from their paper’s conclusion represents my own thoughts best:
Nevertheless, attributing climate trends over relatively short periods, such as 10 to 15 years, will always be problematic, and it is inherently unsatisfying to find model–data agreement only with the benefit of hindsight.
For my own conclusion, I am highly skeptical that they got any model-data agreement if the process is hindcast. I’m also completely unimpressed with the kind of numeric mashing used to claim that models are still somehow ‘on the right track’ but this next sentence in their conclusion is completely unjustified/unsupported/unimagined by any aspect of this paper:
We see no indication, however, that transient climate response is systematically overestimated in the CMIP5 climate models as has been speculated8, or that decadal variability across the ensemble of models is systematically underestimated, although at least some individual models probably fall short in this respect.
There is no analysis in the article of expected short term variance which could possibly explain the models failure. It simply doesn’t exist. This primary aspect of Gavin’s conclusion is much more like a prayer to Gaia than an article of sicence.
As is often the case the Real Climate train-wreck provided us some solid entertainment. I wonder how many more decades will pass before they will figure out that the modeled climate feedback sensitivity looks a little high?