mhaze
Banned
- Joined
- Jan 10, 2007
- Messages
- 15,718
mhaze, you brought this up and claimed it meant that "climate models have no predictive value". So you've obviously read the paper and understand why the problem I raised isn't an issue.
Care to comment?
I'm not sure what your comments add up to. They are more complaints than anything else: You don't like that he discussed Albany for a page and a half, that he only used 8 stations, that local results are not regional or global, that some of the runs were not initialized to start conditions.
We've got close to two decades of actual weather and we've got predictions from models that were published, let's go compare them and see what if any accuracy the models had. That is basically what Koutsoyiannis et al. did.
Now you don't like the "local or regional" because it is not global? Last I heard, all climate was local or regional. India, for example, is concerned about monsoon dynamics; South Texas about Gulf hurricanes and the interaction between the moist air coming in from the gulf with dry air from the West, on and on. Well, what scale of results, either spacial or geographical, should models be quantifiable examined on?
- A 10km sided area? 100km? 1000km? 2x2 Grid points? 5x5? 20x20?
- What timeframe? 10 years? 20? 50? 100?
- What accuracy?
- Should these (indelicate, politically incorrect) questions not be asked?
Who decided the model runs would not be initialized to known states? Is that a complaint you have against the arthur of the paper in question here? Might want to rethink that. Where did they get their (published in the past, right? ) model runs?
Do climate models have accuracy and skill, over a defined geographical scale and time frame? What, then are these parameters, and what is the skill, the accuracy, and the proof of it ? A lot of people would like to here that one.
Only 8 stations? They said they would have liked to do more, time and money premitting. Maybe other people will pick up the ball and examine 100. I'm curious - do you seriously think the results would differ?
Koutsoyiannis et al. go further: stating not just that local predictions of the models have been falsified, but presuming that they work over longer distances and/or timeframes is unsupported.
Well informed people have noted this or made similar comments. (Trenberth, IPCC lead arthur: "IPCC does not do predictions"). Is Koutsoyiannis saying anything that has not been said before, or only trying to quantify the extent of the problem? Quantifying the extent of a problem is advancing the state of scientific understanding. If you are not just complaining about the paper, then perhaps you believe the models have a skill and predictive capability that in fact they do not have, that no knowledgable people have or should claim.
Clearly I differ your comment -
I would never let something like that go by as a referee. It's either dishonest, or merely strongly seems so - either way, it stinks. If this is the level at which this kind of science gets done, no wonder it's so hard to determine the facts.
Last edited:
