Published online 20 January 2010 | Nature 463, 284-287 (2010) | doi:10.1038/463284a
News Feature
Like any other field, research on climate change has some fundamental gaps, although not the ones typically claimed by sceptics. Quirin Schiermeier takes a hard look at some of the biggest problem areas.
The article can be found here (Google is your friend):
http://www.nature.com/news/2010/100120/full/463284a.html
What did you find as "great" there? These are all well known problems. The article focuses on four aspects.
(1) "Regional climate prediction"
The reporter, Quirin Schiermeier, as usual for general ignoramuses, misrepresents the problem. He skews it into more favorable light, to make an impression that only "regional details" might be missing, and are of secondary importance. In fact, the lack of convergence of GCMs is a fundamental flaw in the entire approach. In exact sciences, it is customary to shrink calculation grid down to a point where all features of solutions converge, become stable. Only then the model is considered as relatively reliable, and numerical experiments may be conducted to explore various dependencies. In climatology, this is not a case. Climatology uses an excuse that since we do not have sufficient computational resources, this is "good enough". No wonder that details of GCMs fail apart as grid getting smaller. But what he forgot to say is that global climate model is a simple average of regional pieces, and if the pieces are wrong, there is no chance that the sum of them is right. To think otherwise is a fallacy that "two wrongs make one right".
(2) "precipitation"
This is the same problem as above. For example, in the climateprediction.net experiment (where Lenny Smith is one of principals), it can rain where there are no clouds, and massive arrays of clouds could appear or disappear in a matter of one-two time steps (15 or 30 minutes). Obviously this kind of processes could not possibly represent local energy balances correctly. This is a classic problem of "parameterization" in a system that has continuous spectrum of scales. Obviously, the rain cannot be without clouds, and clouds can block 90% or more of solar radiation at surface, which in turn affects evaporation and supply of humidity for cloud formation. It is a closed loop of an internal self-consistent prognostic variable, and parameterizations just cannot do the job of a closed-loop feedback system.
(3) "Atmospheric aerosols"
Again, same thing, an attempt to parameterize internal prognostic variable that is critically important for process of cloud formation.
(4) "Tree-ring reconstructions"
This is a total joke. It is obvious that previous three parts were a smokescreen for this apologism. BTW, Quirin Schiermeier seems to be a mouthpiece for the ClimateGate scandal, and his name is mentioned in the leaked mail. He was trained by Phil Jones on how to spin climate news down to "right message": " We simply want to do our best to help make sure that the right message is emphasized." [1211462932.txt]
So, nothing is new in this article.
The article can be found here (Google is your friend):
http://www.nature.com/news/2010/100120/full/463284a.html
What did you find as "great" there? These are all well known problems. The article focuses on four aspects.
(1) "Regional climate prediction"
The reporter, Quirin Schiermeier, as usual for general ignoramuses, misrepresents the problem. He skews it into more favorable light, to make an impression that only "regional details" might be missing, and are of secondary importance. In fact, the lack of convergence of GCMs is a fundamental flaw in the entire approach. In exact sciences, it is customary to shrink calculation grid down to a point where all features of solutions converge, become stable. Only then the model is considered as relatively reliable, and numerical experiments may be conducted to explore various dependencies. In climatology, this is not a case. Climatology uses an excuse that since we do not have sufficient computational resources, this is "good enough". No wonder that details of GCMs fail apart as grid getting smaller. But what he forgot to say is that global climate model is a simple average of regional pieces, and if the pieces are wrong, there is no chance that the sum of them is right. To think otherwise is a fallacy that "two wrongs make one right".
(2) "precipitation"
This is the same problem as above. For example, in the climateprediction.net experiment (where Lenny Smith is one of principals), it can rain where there are no clouds, and massive arrays of clouds could appear or disappear in a matter of one-two time steps (15 or 30 minutes). Obviously this kind of processes could not possibly represent local energy balances correctly. This is a classic problem of "parameterization" in a system that has continuous spectrum of scales. Obviously, the rain cannot be without clouds, and clouds can block 90% or more of solar radiation at surface, which in turn affects evaporation and supply of humidity for cloud formation. It is a closed loop of an internal self-consistent prognostic variable, and parameterizations just cannot do the job of a closed-loop feedback system.
(3) "Atmospheric aerosols"
Again, same thing, an attempt to parameterize internal prognostic variable that is critically important for process of cloud formation.
(4) "Tree-ring reconstructions"
This is a total joke. It is obvious that previous three parts were a smokescreen for this apologism. BTW, Quirin Schiermeier seems to be a mouthpiece for the ClimateGate scandal, and his name is mentioned in the leaked mail. He was trained by Phil Jones on how to spin climate news down to "right message": " We simply want to do our best to help make sure that the right message is emphasized." [1211462932.txt]
So, nothing is new in this article.
