07-15-2009, 08:00 PM
Hi,
In the wake of WW II and the mega research programs instituted then (the Manhattan project being the most renown, but others to develop jet engines, better ciphers, etc. also existed) it became the norm to expect government funding for research. Prior to that, research was mostly the pastime of people who were, for one reason or another, obsessed with knowledge. After that, it became, for many, a big business.
The people in charge of making those grants are not experts in the fields that they are funding -- they aren't completely ignorant, but they aren't active researchers, either. There is much more demand for research funding than there is funding available. To determine who gets the funding, the agencies need to evaluate the importance of the requests. So, if one researcher says "I'd like to be funded to study climatic change because it is cool" and another says "I'd like to be funded to study climatic change because it is taking us to the end of civilization as we know it", guess who gets funded?
Now, if two groups got funded because they each claimed that the end was imminent, and they each developed a model, one of which supported the doomsday scenario and the other which didn't, guess who gets an extension (hint, it's not the group who's report starts with "We were wrong.")
Then there is the fact that scientific computer models are seldom, if ever, written by people who are trained in computer fields. They are written by scientists who have picked up programming as an avocation. The papers generated by these models may be peer reviewed, but the models themselves very seldom are. At most, they are shared among the researchers in a field and some small degree of feedback is given. Usually, those obtaining copies of the code just want it to duplicate work that has been done and look at the code just enough to get it running on their system (often a non-trivial task if the code, like so many scientific codes, is written in a dialect of FORTRAN specific to one machine or class of machines). So misconceptions, poor assumptions, and downright errors are propagated. There is a standard analysis library that has known errors in it, but they are uncorrected so that present and future work will be comparable to that done in the past.
Finally, a model is often judged to be 'good' or 'bad' on the basis of what it predicts. But since there often isn't anything objective to compare it to, then it boils down to the subjective. The modeler often has a general idea of what he expects. If the model more or less gives him that, then it is good, otherwise it isn't. It appears that none of the climate models presently in existence can fully reproduce the *known* history of the planet. Why then believe them for the future?
I don't claim, nor think, that there is a big conspiracy. From my lifetime experience of modeling physical processes, I think that it is just a combination of the pressures of obtaining funding and the natural human inclination to become emotionally attached to our pet theories that leads to the situation we have. The field of climatology is exposed to political scrutiny that could lead to regulations, contracts, etc. And so, a conspiracy explanation is attractive. But look at the situation in cosmology, where there is no external pressures, and yet over the past fifty years various theories have been in vogue, and those researchers not subscribing to the theory du jour have had great difficulty in getting positions, funding, or even published.
Science is a nearly perfect system for obtaining knowledge, but it is conducted by imperfect people. In the long run, it works out. In the short, it is often dominated by personalities, prejudices, and expectations.
--Pete
Quote:It's really that bad?It really is.
In the wake of WW II and the mega research programs instituted then (the Manhattan project being the most renown, but others to develop jet engines, better ciphers, etc. also existed) it became the norm to expect government funding for research. Prior to that, research was mostly the pastime of people who were, for one reason or another, obsessed with knowledge. After that, it became, for many, a big business.
The people in charge of making those grants are not experts in the fields that they are funding -- they aren't completely ignorant, but they aren't active researchers, either. There is much more demand for research funding than there is funding available. To determine who gets the funding, the agencies need to evaluate the importance of the requests. So, if one researcher says "I'd like to be funded to study climatic change because it is cool" and another says "I'd like to be funded to study climatic change because it is taking us to the end of civilization as we know it", guess who gets funded?
Now, if two groups got funded because they each claimed that the end was imminent, and they each developed a model, one of which supported the doomsday scenario and the other which didn't, guess who gets an extension (hint, it's not the group who's report starts with "We were wrong.")
Then there is the fact that scientific computer models are seldom, if ever, written by people who are trained in computer fields. They are written by scientists who have picked up programming as an avocation. The papers generated by these models may be peer reviewed, but the models themselves very seldom are. At most, they are shared among the researchers in a field and some small degree of feedback is given. Usually, those obtaining copies of the code just want it to duplicate work that has been done and look at the code just enough to get it running on their system (often a non-trivial task if the code, like so many scientific codes, is written in a dialect of FORTRAN specific to one machine or class of machines). So misconceptions, poor assumptions, and downright errors are propagated. There is a standard analysis library that has known errors in it, but they are uncorrected so that present and future work will be comparable to that done in the past.
Finally, a model is often judged to be 'good' or 'bad' on the basis of what it predicts. But since there often isn't anything objective to compare it to, then it boils down to the subjective. The modeler often has a general idea of what he expects. If the model more or less gives him that, then it is good, otherwise it isn't. It appears that none of the climate models presently in existence can fully reproduce the *known* history of the planet. Why then believe them for the future?
I don't claim, nor think, that there is a big conspiracy. From my lifetime experience of modeling physical processes, I think that it is just a combination of the pressures of obtaining funding and the natural human inclination to become emotionally attached to our pet theories that leads to the situation we have. The field of climatology is exposed to political scrutiny that could lead to regulations, contracts, etc. And so, a conspiracy explanation is attractive. But look at the situation in cosmology, where there is no external pressures, and yet over the past fifty years various theories have been in vogue, and those researchers not subscribing to the theory du jour have had great difficulty in getting positions, funding, or even published.
Science is a nearly perfect system for obtaining knowledge, but it is conducted by imperfect people. In the long run, it works out. In the short, it is often dominated by personalities, prejudices, and expectations.
--Pete
How big was the aquarium in Noah's ark?