Of moles and whacking: “Climate models didn’t predict this lack of warming”
September 10, 2009 ·
Joseph Romm
Or: Journalists should report what climate science actually “says”, rather than what they mistakenly “believe” it to say – Part I
Not everyone who writes misleading or confused stories on climate change does so for the partisan reasons that the Jonah Goldberg’s
Jonah Goldberg’s classic know-nothing, non-denial climate denial
thingsbreak.wordpress.com
of the world do. Sometimes the writer is simply incorrect about what the science says, and his or her errors are made in perfectly good faith. Such problems often arise when the writer mistakes the “conventional wisdom” or individuals’ opinion on climate change for what the science actually says. A case in point is a recent column by the Houston Chronicle’s “SciGuy”, Eric Berger, entitled “Climate scientists should talk about what ‘may’ happen, rather than what ‘will’ happen”. Predictably, the post is being lauded in the denialosphere.
It appears that Mr. Berger has made some unfortunate assumptions about climate science that turn out not to be supportable. Finding these assumptions to be mistaken, Mr. Berger disappointingly chooses to blame climate scientists instead of digging a little deeper into his misconceptions to see where he went awry. Doing so ourselves may help illuminate not only how and why Mr. Berger came to such unsupportable conclusions, but also how we can avoid doing so in the future.
For a long time now, science reporters have been confidently told the science is settled.
This is an immediate red flag. Who is doing the “telling”? What makes up “the science”? What is meant by “settled” if indeed this was even claimed? Searching Google News for the “the science is settled” is quite revealing. Assertions that the claim has been made are legion while the claim itself seems to be apocryphal. This is apparently a “skeptic” canard that appears to have little or no basis in reality.
There are certainly areas where the science can said to be as “settled” as science ever is, in terms of whether questions have been asked and answered at a satisfactory enough level that those working in the field have moved on to other issues. However, there is a real danger in overextending this claim to areas where the science is still very much not “settled”. A good rule of thumb over the past few years has been to use the most recent IPCC Assessment Report as a rough approximation of the “state of the science”. In terms of whether or not the planet is warming, whether or not humans have driven most of it in the past few decades, whether or not the planet will warm to a higher equilibrium in relation to the amount of current and future GHG (greenhouse gas) emissions, and whether or not sea levels will rise appreciably in response, etc.- it’s fairly safe to say that the science is as “settled” as science tends to get. These questions- which I’ll refer to the “broad strokes” of the issue- are, after all, largely rooted in basic physics.
[There is of course a problem with using the AR4 (IPCC 4th Assessment Report) as a benchmark now that we're in the latter part of 2009, however- the AR4, published in early 2007, draws on studies from several years ago which themselves draw on other studies older still. Thus significant conclusions reached subsequent to its writing are by definition excluded from the AR4, and such conclusions are worrisome indeed. Additionally, the requirement that countries with very different political agendas must come to an agreement likewise tends to bias the reports towards the conservative side.]
Questions regarding more complicated issues, such as the rate of ice sheet decay (collapse), the future response of ENSO (the El Niño Southern Oscillation), etc. are touched on by the AR4 but don’t receive anything approaching the confidence enjoyed by the broad strokes of the subject. It is all too common for people to overreach the broad strokes, claim that an area under debate was supposed to be “settled”, and then pounce when they perceive evidence contradicting the tentative conclusions still under discussion- it’s a cheap stunt meant to score rhetorical points and undermine political support for mitigating climate change, while illuminating nothing about the state of the science. A far less cynical corollary to this is a person unintentionally engaging in the same overreach due to his or her own misconceptions about the subject, then becoming confused when reality doesn’t match up to this incorrect expectation. This latter case seems to be what is occurring with Mr. Berger’s column, so let’s see what we can do to clarify things.
That the planet is warming and humans are unquestionably the primary cause.
I would probably amend that to read something like “unquestionably the primary cause since the latter half of the 20th century”, but it’s acceptable as is.
We’ve been told to trust the computer models –
Whoah. There’s another big, big red flag. Computer models certainly play their part in climate science and the modeling community deserves full kudos for its contributions to constraining questions relating to attribution and expected future impacts. However, “computer models” are not necessary to illustrate anthropogenic attribution (i.e. the man-made cause) of the observed warming of the climate system. Likewise, they are not necessary to establish a rough picture of how the planet will respond to future levels of increased concentrations of GHGs- we can look to paleoclimatic evidence (i.e. evidence from our planet’s past). A classic error made by both those attempting to discredit climate science as well as those who simply don’t know any better is to erroneously reduce the entire subject of climate change to computer models.
the models which show a markedly upward trend in temperatures as carbon dioxide concentrations increase. And I’ve trusted the scientists telling me this.
First of all, it’s not simply “computer models” that show this, in terms of what has already occurred. The linear relationship exists in the real world data. [Although the empirical relationship is actually logarithmic, there are additional factors at work that render the relationship linear for our purposes here]. And what does “markedly upward trend” mean here? That over a period of several decades, temperature climbs in response to increased greenhouse gas concentrations? If so, fair enough.
That isn’t what Mr. Berger seems to believe, however:
Below you’ll find the computer model forecasts for the 21st century temperatures from the most recent IPCC summary for policymakers, which call for a 1.8°C to 3.8°C rise in global temperatures by 2100:
It seems pretty clear that the models forecast a steady upward trend in global temperatures as long as carbon dioxide levels rise. (Which they have). Yet according to satellite and surface temperature measurements the global average temperature has essentially remained flat for the last 12 years. This strikes me as somewhat curious.
…And the Earth seems to have, at least temporarily, stopped warming.
Here Mr. Berger is making an all too common and easily correctable error: he is confusing the IPCC projections of the forced component of climate over time with an explicit prediction of future temperature. I realize that this sounds a little confusing, but it’s not. We just need to clarify some terms [there will be some simplification here- if you're already comfortable with this material please feel free to skip it]:
Monotonic warming: Warming that increases over time with no year cooler than the one prior. Much of the confusion- both intentional and unintentional- about how we can expect temperature to behave under global warming comes from the mistaken belief that global average temperature is predicted to warm monotonically. It decidedly is not. Even as man-made influences like increased GHG emissions have greater and greater impacts on the climate system, they will not cancel out all natural variability, and thus we can expect some years to be cooler than the ones before, even as the warming trend continues over the long term.
Natural variability: In terms of climate, this refers to naturally occurring influences upon the climate system that exist independently of human-influence. Two of the largest sources of natural variability in the climate system are ocean circulation changes and changes in solar irradiance.
Individual model run or realization: For our purposes, a single output of global average temperature by a single climate model under given initialization values over time. What actually occurs in the real world is thus a “single realization” of the actual climate system instead of a climate model. Due to natural variability, neither this real world realization nor climate model runs show monotonic warming.
Ensemble average: Simply put, the ensemble average is an average of many different individual model runs. Ensemble averages tend to be more accurate in describing long term trends than individual modeling runs do. This is because the timing if not the magnitude of natural variability- like changes in ocean circulation- is difficult to predict well ahead of time, even though individual runs simulate the general behavior; i.e. climate models can mimic the effect of ENSO changes, but we still have difficulty predicting the onset of an El Niño more than a few months ahead of time. Averaging across several model runs mimicking different expressions of natural variability smooths (or cancels) out these differences.
[For example: if I have a 'Model A Run 1' showing a positive shift in natural variability beginning in 2010, reversing in 2o11, and settling into a neutral mode by 2012, while a Model A Run 2 shows a negative shift in natural variability beginning in 2010, settling into a neutral mode by 2o11 and showing a positive shift in 2012, then the natural variability of the two runs during this time period more or less cancel out.]
Additionally, each model may capture one or more aspects of the climate system’s response to changes more realistically than the others, while capturing other aspects less realistically. Ensemble averaging smooths out these differences as well. So what you’re left with is essentially the “forced” component of temperature.
Projection vs. prediction: To simplify things somewhat, we can call the IPCC ensemble averages “projections”. These projections are by definition not explicit predictions or forecasts- we know that natural variability will still have a large influence on the climate system, especially on the shorter term, and that projections intentionally smooth this influence out. The real world single realization of the future climate system will be much “noisier” than the projections, and so at any given time can be expected to be above or below the projection.
Summary: So what the IPCC offers as a projection is an ensemble average of individual model runs. Such projections illustrate the forced component of climate over time by smoothing out natural variability and modeling errors, and are not the same thing as an explicit prediction of future temperature. If you confuse the IPCC projections with predictions, you will incorrectly assume that “climate models” say temperature should be warming monotonically. An alternative and equally incorrect claim is that “climate models” didn’t predict this “current pause” in warming.
[Note: Those of you who have read my previous post on Jonah Goldberg have already seen the following]
This issue is also explicitly addressed by Easterling and Wehner in their GRL paper Is the climate warming or cooling? They use both real world instrumental data (Fig 1) and an individual model run (Fig 2) to illustrate that natural variability can give the illusion of a pause in warming, while the long term trend remains undeniably one of warming; i.e. warming has not been, nor is it expected to be n the future, monotonic.
The individual runs that comprise the AR4 ensemble average clearly exhibit natural variability and do not predict warming will be monotonic.
Of course this has been covered many, many times elsewhere, notably by Robert Grumbine in determining climatologically significant temperature trends (and here), by Tamino in when to expect new record temperatures, and by RealClimate discussing what the IPCC models actually say.
[End of self-plagiarizing]
Further examination of Mr. Berger’s column, [UPDATE:] hurricanes, ocean circulation changes, and more to follow in Part II. We might move those to a Part III.
thingsbreak.wordpress.com |
