Paleo tropical temperatures and climate modeling

A very large conundrum in climate modeling appears to have been resolved.  Details here:

http://www.nature.com/nature/journal/v457/n7230/full/457669a.html

Of course, skeptics will argue that once again, scientists have adjusted data to fit the models.  But the point is that if models are incorporating physics correctly, there is enough experience with them now that if the data and models disagree, it is wise to look for problems in the data first. 

Argh!  I forgot the Nature link wouldn't work. 

The conundrum relates to paleoclimate simulations and time periods when the arctic was sub-tropical.  Under that climate regime, models predict that the tropical temperatures were on average 10-15 C warmer than today (since the equatorial-pole temperature gradient remains roughly the same as today).  However, available temperature proxies (mostly involving leaf size) indicate that equatorial tropical temperatures were not a whole lot larger than today, maybe a couple of degrees celsius.  This led climate modelers to assume that something was missing from the physics of the models that would regulate temperature in the tropics, preventing it from getting so hot.  Likely candidates were deep convection and cloud formation (Sound familiar?  Lindzen wasn't really a genius, he just picked up on known issues with modeling heat transfer in the tropics and amplified the uncertainty) but the modelers and climate dynamicists just weren't sure how to do a better job, or what precisely was missing in the physics. 

Anyway, the existence of the huge freaking snake (HFS, thanks Dana!) gives a second temperature proxy for that era (no, I didn't know this before I read the article, biologist I am not).  It turns out snake size correlates with environmental temperature, since it is easier to keep a larger body warm enough in a hotter climate.  If you develop a relationship between snake size and temperature, and plot the HFS on that curve, the predicted environmental temperature comes out to be around 10 C warmer than the tropics are today.  In other words, if the HFS was really that big and snake size is a good proxy for temperature, the models are actually doing a reasonable job both at high latitudes and in the tropics.  This implies the parameterizations of deep convection and tropical clouds are not so bad. 

Unfortunately, it's just one more reason why reducing CO2 is a really good idea. 

<sigh>  I could have saved myself some time and just said:  "Exactly Gerda." 

Beats me.  Both methods were published in very respectable journals (Phil Trans. Royal Soc. B for the snake work, for example).  I would bet on the snakes, but that's just because my opinion is that the models are not that bad. 

As a general rule of thumb in research, before you start invoking new physics to explain an anomalous result, it's always a good idea to check the data one more time. 

Hey smart guy, stop reading my mind.  :-)

Yeah, precisely!  Once you decide to start invoking new physics before rigorously checking your own data, you have ceased being a scientist. 

Another great example of this is Pons and Fleischman and cold fusion.  Tens of millions of dollars in research funding and hundreds of man-years of effort have been squandered proving that Pons and Fleischman didn't do a very good job of verifying their own results.