The Science Whore
Climate is too complex for accurate predictions by Jim Giles published in New Scientist:
Quick, call Ripley. Not. Only someone who has never worked with prediction models could possibly take seriously these global warming predictions of temperatures decades out and within a few degrees. It's completely ludicrous. The number of arbitrary assumptions being made in those models are beyond calculation. I remember when one of the space probes first found multiple rings around Uranus, and one question raised was how they interacted. I asked a mathematician friend who did work in astrophysics, and he just laughed: "We have the math for two rings, at three it gets very hard, after that, forget it." (The guy was from India, brilliant, with a wonderful laugh. Once he got a contract from the Defense Department to find a data analysis scheme for something he couldn't tell me about. They gave him a fat contract based on the time their quantitative people thought it would take, which was something like six months. He solved it in two days, and had an enormous laugh about that.)
Global climate change is vastly more complex than the interaction of a few planetary rings -- systems in outer space are relatively simple and isolated, which is why astronomy was the first successful science. In climate systems the number of important causal variables is unknown, but surely very large. Good theories would have to specify all the important ones -- and their interactions! And you'd also have to have good measurement of them. In fact, the theories are primitive, with little agreement even about which variables are most important of the ones that are known -- thus, for example, all the dispute about the influence of the Sun.
(Predicting climate change resembles the simpler case of stock market prediction, which has absorbed tons of computer time at the hands of Ivy League math and physics Ph.D.'s that are hired as Wall Street quants. The return on that investment has been essentially zero. As with climate, you can back-test your models and get them to pretty well fit the old data. It's the data for tomorrow and the next day that turn out to be the problem. "There are more things In heaven and earth, Horatio, than are dreamed of in your philosophy." Is it amazing that the enormous sub-prime mess managed to sneak up on all these gurus? No, it's not amazing at all.)
What we have in the global warming area is the same combination that has pretty well destroyed the social sciences: (1) Strong ideological agendas and (2) weak, non-experimental science. It's an awful mix, especially for science, which ends up whoring for both sides. The final stage is when the science whore is recognized as too weak to be of value to anyone; then she gets kicked into the gutter, and it becomes a contest of naked political agendas. Better to recognize the nature of the controversy at a much earlier point, and leave science out of it.
Byron
...climate policies will necessarily be made in the face of deep, irreducible uncertainties...research dollars invested in ever more sophisticated climate models offer very little marginal benefit to decision makers.
Quick, call Ripley. Not. Only someone who has never worked with prediction models could possibly take seriously these global warming predictions of temperatures decades out and within a few degrees. It's completely ludicrous. The number of arbitrary assumptions being made in those models are beyond calculation. I remember when one of the space probes first found multiple rings around Uranus, and one question raised was how they interacted. I asked a mathematician friend who did work in astrophysics, and he just laughed: "We have the math for two rings, at three it gets very hard, after that, forget it." (The guy was from India, brilliant, with a wonderful laugh. Once he got a contract from the Defense Department to find a data analysis scheme for something he couldn't tell me about. They gave him a fat contract based on the time their quantitative people thought it would take, which was something like six months. He solved it in two days, and had an enormous laugh about that.)
Global climate change is vastly more complex than the interaction of a few planetary rings -- systems in outer space are relatively simple and isolated, which is why astronomy was the first successful science. In climate systems the number of important causal variables is unknown, but surely very large. Good theories would have to specify all the important ones -- and their interactions! And you'd also have to have good measurement of them. In fact, the theories are primitive, with little agreement even about which variables are most important of the ones that are known -- thus, for example, all the dispute about the influence of the Sun.
(Predicting climate change resembles the simpler case of stock market prediction, which has absorbed tons of computer time at the hands of Ivy League math and physics Ph.D.'s that are hired as Wall Street quants. The return on that investment has been essentially zero. As with climate, you can back-test your models and get them to pretty well fit the old data. It's the data for tomorrow and the next day that turn out to be the problem. "There are more things In heaven and earth, Horatio, than are dreamed of in your philosophy." Is it amazing that the enormous sub-prime mess managed to sneak up on all these gurus? No, it's not amazing at all.)
What we have in the global warming area is the same combination that has pretty well destroyed the social sciences: (1) Strong ideological agendas and (2) weak, non-experimental science. It's an awful mix, especially for science, which ends up whoring for both sides. The final stage is when the science whore is recognized as too weak to be of value to anyone; then she gets kicked into the gutter, and it becomes a contest of naked political agendas. Better to recognize the nature of the controversy at a much earlier point, and leave science out of it.
Byron
posted by Byron | Thursday, October 25, 2007
0 Comments:
Post a Comment
<< Home