4 ms·

99.999% certainty humans are driving global warming

9 points by rblion 12y ago

todd8 12y ago
Is anyone else skeptical of the approach taken in the OP's cited analysis? I believe in anthropomorphic global warming in the sense that human generated CO2 is going to affect the climate. But I'm not sure that I would go as far as to say that humans are "driving" global warming with "certainty". The reasons I say this are that the global circulation models used by climate scientists are tweaked and tweaked until they match up with historical climate figures, but seem to have trouble with predicting the future well. Less violent tornados, less hurricane activity, puzzling ice melting behavior, and a hiatus in global temperature increase have made me doubt the accuracy of the models (and the interpretation of their predictions). I'm not saying that I doubt that global warming is a problem, just that the models seem to be incomplete or that the problem of detailed climate modeling is somehow too chaotic to be computationally tractable. The questions that I feel are important, the size of the global warming problem and how it should be addressed are not the point of the OP citation [Kokic, et. al]. Given the failures of climate model predictions (not the ability to tune models to historical data) it seems that we should have at least a modicum of skepticism for the way they are used in this paper. I don't have the energy myself to go over another climate paper in detail, but from looking over it quickly it appears that the basic argument is: (1) Pick a time interval that includes time before significant human contribution to greenhouse gases (GHGs). In this case the 20th century. (2) Pick a number of interesting time intervals (3) Pick a model for GW. (3) Adjust the coefficients on the model to completely exclude contribution of GHGs. (4) Run simulations (including bootstraps on top of bootstraps) (5) Conclude that out of 100,000 runs of the simulation only 1 run would account for the climate behavior actually observed for the intervals. (6) Conclude that there is 99.999% certainty that humans are driving global warming. I am not surprised that current GW models (built on the historical record) don't reflect historical behavior when GHGs are taken out of the equation. The paper just doesn't provide enough detail and justification for me to be comfortable with the 99.999% number. From now on people will be quoting this five 9's certainty, but is it really 99.999% or is it 99.9% or 99%? The whole point of this paper seems to be to win some sort of straw-man argument. Of course, there are people that believe that human GHGs play no part in climate, but is it really important to provide Chris Mathews with the 99.999% number instead of the 99% number to use when mocking Republicans? This distorts any intelligent discussion about energy and climate. There are large number of assumptions and seemingly arbitrary choices made by the authors of the paper in question. There is the possibility (at least at the 0.0001% level) that their programs have bugs or unjustified assumptions or simplifications that affect the conclusions. In my opinion, and take this with a grain of salt since I am not expert in this area, this paper is one where the data, statistics, and simulation programs should be opened up to examination by outside researchers.

9 points by rblion 12y ago

todd8 12y ago

Is anyone else skeptical of the approach taken in the OP's cited analysis? I believe in anthropomorphic global warming in the sense that human generated CO2 is going to affect the climate. But I'm not sure that I would go as far as to say that humans are "driving" global warming with "certainty". The reasons I say this are that the global circulation models used by climate scientists are tweaked and tweaked until they match up with historical climate figures, but seem to have trouble with predicting the future well. Less violent tornados, less hurricane activity, puzzling ice melting behavior, and a hiatus in global temperature increase have made me doubt the accuracy of the models (and the interpretation of their predictions). I'm not saying that I doubt that global warming is a problem, just that the models seem to be incomplete or that the problem of detailed climate modeling is somehow too chaotic to be computationally tractable. The questions that I feel are important, the size of the global warming problem and how it should be addressed are not the point of the OP citation [Kokic, et. al]. Given the failures of climate model predictions (not the ability to tune models to historical data) it seems that we should have at least a modicum of skepticism for the way they are used in this paper. I don't have the energy myself to go over another climate paper in detail, but from looking over it quickly it appears that the basic argument is: (1) Pick a time interval that includes time before significant human contribution to greenhouse gases (GHGs). In this case the 20th century. (2) Pick a number of interesting time intervals (3) Pick a model for GW. (3) Adjust the coefficients on the model to completely exclude contribution of GHGs. (4) Run simulations (including bootstraps on top of bootstraps) (5) Conclude that out of 100,000 runs of the simulation only 1 run would account for the climate behavior actually observed for the intervals. (6) Conclude that there is 99.999% certainty that humans are driving global warming. I am not surprised that current GW models (built on the historical record) don't reflect historical behavior when GHGs are taken out of the equation. The paper just doesn't provide enough detail and justification for me to be comfortable with the 99.999% number. From now on people will be quoting this five 9's certainty, but is it really 99.999% or is it 99.9% or 99%? The whole point of this paper seems to be to win some sort of straw-man argument. Of course, there are people that believe that human GHGs play no part in climate, but is it really important to provide Chris Mathews with the 99.999% number instead of the 99% number to use when mocking Republicans? This distorts any intelligent discussion about energy and climate. There are large number of assumptions and seemingly arbitrary choices made by the authors of the paper in question. There is the possibility (at least at the 0.0001% level) that their programs have bugs or unjustified assumptions or simplifications that affect the conclusions. In my opinion, and take this with a grain of salt since I am not expert in this area, this paper is one where the data, statistics, and simulation programs should be opened up to examination by outside researchers.