June 4, 2008

The Sanctity of Climate Models

Filed under: Surface, Temperature History

Reading between the lines of the new Thompson et al. Nature paper suggests that once they get the details worked out, the “updated” observed global temperature history is going to fit climate model hindcasts even better than it does now, and embolden confidence in their future projections.

The majority of the alarm raised over global warming stems from climate model projections. And thus it is largely inconceivable that there is going to be a major finding published in a journal such as Nature that is going to call into question one of the fundamental results from climate models—that is, climate models accurately simulate the “known” temperature history. The “known history” may change, but you can rest assured that it isn’t going to change in such as way as to make the climate models look like they aren’t doing so well after all.

At first blush, it seems that new paper by Thompson and colleagues published in the current issue of Nature magazine concerning alterations to the “known” temperature history of the world oceans may be an exception to this rule.

Thompson et al. found that irregularities arising from changing observing practices and data availability during World War II produced a discontinuity on the global sea surface temperature record in the middle 1940s. In the abstract they write “We argue that the abrupt temperature drop of ~0.3C in 1945 is the apparent result of uncorrected instrumental biases in the sea surface temperature record. Corrections for the discontinuity are expected to alter the character of mid-twentieth century temperature variability but not estimates of the century-long trend in global-mean temperatures.”

However, while Thompson et al. identify the existence of a significant problem in the sea surface temperature record, they don’t actually fix the problem in this paper. Instead, they leave us with a we-are-working-on-it promise, “The Met Office Hadley Centre is currently assessing the adjustments required to compensate for the step in 1945 and subsequent changes in the SST observing network.”

And tease us with the following:

The adjustments immediately after 1945 are expected to be as large as those made to the pre-war data (~0.3 ºC), and smaller adjustments are likely to be required in SSTs through at least the mid-1960s, by which time the observing fleet was relatively diverse and less susceptible to changes in the data supply from a single country of origin. The new adjustments are likely to have a substantial impact on the historical record of global-mean surface temperatures through the middle part of the twentieth century. The adjustments are unlikely to significantly affect estimates of century-long trends in global-mean temperatures, as the data before ~1940 and after the mid-1960s are not expected to require further corrections for changes from uninsulated bucket to engine room intake measurements.

This statement leaves the door wide open for rampant speculation, for the goings-on in the mid-20th century are critically import for several major issues, 1) they are vital in understanding the impact of sulfate aerosols, 2) they impact attribution of the recent warming, and not to mention that 3) a major change would reflect negatively on the general reliability of the global temperature dataset.

And, not to disappoint, speculation began virtually immediately.

Our first reaction was “wow, this is going to knock the stuffing out of the sulfate hypothesis” since the climate models are specifically tuned to match the 20th century temperature history by fiddling with the cooling impacts of sulfate aerosols. If it really didn’t cool as much during the mid-1940s through the mid-1970s as we thought, then all the models are fit to the wrong temperature history and the big loser is going to be the sulfate aerosols, as they will have a lot less work to do. This has far reaching impacts down the line, for in addition to forcing a cooling from the mid-1940s to the mid-1970s, sulfate aerosols also are required in climate models to keep the 20th century temperatures in check, otherwise, when the climate models are run with greenhouse gas increases alone, they think it should be a lot hotter now than it really is—which means that they would all be wrong. And more, in the future emissions scenarios used by climate modelers, sulfates emissions are presumed to decline (citing current and future air pollution regulations), which means that the world warms up even faster! Sulfate aerosols are like Swiss Army Knives for the climate models—they are used to get the models out of all sorts of inconvenient situations. Just imagine the problems that would arise if this all-purpose tool were suddenly lost. The Thompson et al. paper hinted at just this possibility.

Others, such as Roger Pielke Jr. on his Prometheus blog, focused more on what the Thompson et al. correction could mean for the Intergovernmental Panel on Climate Change’s (IPCC) recent attribution statement that “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.” If the total warming since the mid-century is significantly reduced by the forthcoming corrections, then perhaps the ability to clearly determine the difference between anthropogenic causes and natural causes is blurred more than the IPCC suggests. One of their fundamental conclusions could be in error.

And still others, such as climateaudit.org’s Steve McIntyre, wonder just how much faith we can have in a temperature record that is full of so many measuring inconsistencies that the task of developing a meaning temperature record is so intractable. After a while there are so many different adjustments applied that the final product looks little like the underlying original observations (but, surprise surprise, it looks more and more like what climate models say it should).

But signs abound that the above are simply wishful (or wistful) thinking.

Those who are a closer to knowing the unpublished details of the pending “fix” are suggesting something else entirely.

For instance, RealClimate’s Gavin Schmidt has this reply to a commenter on the realclimate.org blog:

The first cut at the revisions…has effectively the same match to the [climate] model trends as before (maybe a little better) and so no revisions to the models nor to attribution studies are likely.

So much for the death to the sulfate hypothesis. In fact, the “revised” temperature history will probably even make the models look better.

And the last line of their paper, Thompson et al. let loose the following gem:

However, compensation for a different potential source of bias in SST data in the past decade—the transition from ship- to buoy-derived SSTs—might increase the century-long trends by raising recent SSTs as much as ~0.1ºC, as buoy-derived SSTs are biased cool relative to ship measurements.

Ah, yes, not only are they going to tweak the mid-century temperatures, but they are also going to make recent temperatures warmer than they are currently being reported. This will kill two birds with one stone.

It will serve to (more than?) compensate for any mid-century temperature corrections, and it will take some of the wind out of the sails of the good ship “Global Warming Stopped 10 Years Ago.” Again, the models will come out of this like shining stars.

Call us skeptics, but we have grave doubts that the corrections to the observed global temperature history will result in a lessening in the overall confidence that is proclaimed that climate change is manifesting itself even worse than we imagined. After all, there is an overwhelming, odds-busting tendency for publications in the journal Nature to report that things are tending worse (rather than better) than we ever imagined. In an unbiased world, the expectation should be 50-50 that publications in Nature would find things either better or worse than the expectations. In reality, the publication ratio is about 10 to 1 for the worse side. We have a bad feeling, that despite the initial optimism, that the outcome of the Thompson et al. findings will ultimately prove to increase the tally on the worse-than-expected side of things.

Hopefully our expectations will be turned over.


Thompson, D., et al., 2008. A large discontinuity in the mid-twentieth century in observed global-mean surface temperature. Nature, 453, 646-650, doi:10.1038/nature06982

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress