I try not to feed the global warming trolls at the SJR too much, but some people are still swayed by them, so it's important to not ignore them completely. One point that repeatedly comes up is the fact that Mars is warming. (Because, if Mars is warming, and the Earth is warming, clearly they must be warming for the same reasons. These are people who, in other contexts, quite happily trot out the tired warning that correlation is not causation. But whatever.) And, as I mentioned in an earlier post, you also get a lot of kvetching about how climate science is bogus since it's models models models all the way down. Actual Climate Scientist Gavin Schmidt, of Real Climate, has a post up to explain some of the basics of modeling radiative forcing (that is, how solar energy gets trapped by greenhouse gases).
The post itself is interesting, but what really caught my eye was the first comment, which is quibbling with Gavin over the idea of presenting simplified models to the general public, since some simplifications wind up with weird, runaway results.
Now, one of the elements in Gavin's model is that it assumes there's a single layer of atmosphere, topped with the greenhouse gases. He also says that, basically, there are two sources of radiative forcing: increased solar energy and greenhouse gases. (This is where the warming skeptic's Mars stuff comes in--if it's increased solar energy that's causing our warming, then it's no big deal.) The commenter (Spencer Weart) says that it's an important point that the atmosphere isn't a single layer: it stratifies, and each layer acts a little differently to sunlight coming through; in turn, each layer reacts to warming a little differently, and most crucially, these layers react different depending on the source of the warming. If it's warming due to increased solar energy, then you'd see warming at all of these layers of the atmosphere. If it's warming due to greenhouse gases, then you'd see the layers below the greenhouse band warm, and you'd actually see the layers above that band cool off.
So far, so model-ly. But one of the great things about models--which in their feverish contempt for the m-word skeptics never note--is that they give you things that you can empirically test. And guess what the empirical tests show? Yup: lower layers are warming; upper layers are cooling.
Comments (5)
"things you can empirically test". Exactly. If climate scientists claim to understand the climate, they need to make predictions. The true test of a theory is its ability to make predictions. Yet the IPCC doesn't make predictions; they explicitly say that instead of predictions, they make "projections".
If these models are so wonderful, I would like to see some short term predictions, perhaps of hurricane activity over the next 5 years, or sea level rise. Something dead easy for everyone to monitor. If the models are good, they will be accurate. If they are not accurate, then what good are they?
http://www.climatescience.org.nz/assets/20072141112360.SPM07GrayCritique.pdf
Posted by IanRae | April 15, 2007 3:50 PM
Posted on April 15, 2007 15:50
I think we're talking at different scales. You want empirical validation of the entire thing--but that can't happen, because what you're modelling is slow change (in human terms). So, we could wait for fifty years to see if what happens matches what the IPCC is projecting. But then, we've lost the ballgame. (Your call for shorter-term predictions is misplaced, because that's weather, not climate. You're asking to intensify the back-and-forth gotcha of a late season cold snap or a warm winter. If you reject that 5 years is still weather, then why not look at the past twelve years, which have held 9 of the warmest ten years on record?)
Instead, as I understand it, the way the models are tested is each of the components make empirical claims, that are checked against whatever records we have and against the physics and chemistry under question. It's by providing a continuum of empirics, from the small scale up to historical records, that the case for the role of CO2 is made.
Posted by Greg C | April 15, 2007 4:14 PM
Posted on April 15, 2007 16:14
Yes it is a type of double jeopardy. But that's my point; if the science is so uncertain that predictions can't be made...then it's a poor guide to policy. And to say 50 years will be too late is simply to beg the question.
The time scale thing is puzzling. Models can't predict for very long time scales (1000 years from now) because there is too much long-term uncertainty. Models can't predict for short periods (10 years) because there's too much short-term uncertainty. But somehow 100 years is just right.
The climate/weather comment makes little sense. Either the models can model global temperature or they can't. If you look at the reasons they can't model short-term changes (what you call weather), these same reasons (such as LSU -- the limits of scientific understanding) prevent accurate predictions for 100 year changes.
Here's a recent paper "How reliable are climate models?"
http://www.blackwell-synergy.com/doi/abs/10.1111/j.1600-0870.2006.00211.x
The abstract begins: "Overall, the connection between model skill in simulating present-day climate and the skill in simulating future climate changes is poorly known."
Which is precisely my point.
Posted by IanRae | April 15, 2007 6:27 PM
Posted on April 15, 2007 18:27
But somehow 100 years is just right.
Well, there are two different things going on. Short-term uncertainty is due to the variability of day to day weather. Long-term uncertainty is--I imagine, and bearing in mind that I'm a concerned amateur, not a climate scientist--due to the fact that small differences in initial conditions create long-term major differences.
The less uncertain medium-term is a result of two things: 1. you're not dealing with year to year variability (so, no one is projecting that in exactly the year 2107, the weather will be such-and-such), and 2. you're not so far out that the roughness in initial conditions becomes crucial.
But that's all speculation on my part. I've never heard the complaint about long-term uncertainty before.
And to say 50 years will be too late is simply to beg the question.
What question am I begging?
Ultimately, where we're going to differ is that I trust the consensus position of the climate scientists who put together the IPCC reports and on whose research those reports are based.
Posted by Greg C | April 15, 2007 6:46 PM
Posted on April 15, 2007 18:46
>I trust the consensus position of the climate scientists who put together the IPCC reports
Yes, the notion of a funding-driven cabal at the IPCC doesn't seem credible. Personally I think we _will_ know in 5-10 years. Stuff, like satellite-based temperature monitoring, only got started in the 80s, as did serious climate models (due to lack of computing power). Let's hope they figure it out!
Posted by Ian Rae | April 16, 2007 8:53 AM
Posted on April 16, 2007 08:53