I guess this isn't too far off of a statement as models are *abstractions* of reality.
However, it seems like you mention the word "model" to a "skeptic" and it doesn't matter what follows, but it will be immediately ignored.
So, since models are supposedly unreliable, what other methods of quantifying the climate should we use? Is there a better way the current hierarchy of a nested mathematical framework?
Should we continue to refine models with further experimentation at the cost of (gasp!) taxpayer funded research money?
What if those future experiments resulting in decreases in the confidence for catastrophic projections?
Or should we just give up on understanding them completely because our models aren't perfect?
The first person to mention UN, socialism, communism or Al Gore will have a curse put upon them by my gypsy grandmother.
Anonymous2012-06-27T15:57:44Z
Favorite Answer
My best understanding of climate models that we use them not to predict future global temperatures, but to understand how climate works. Some of the most interesting results of climate models is not a projected temperature of 2100, but how hindcasting shows that we can not explain climate trends without including both natural and anthropogenic factors. http://www.grida.no/publications/other/ipcc_tar/?src=/climate/ipcc_tar/wg1/figspm-4.htm
Do we need climate models to tell us that carbon dioxide is a greenhouse gas? No. That is basic physics. http://www.youtube.com/watch?v=q0kIaCKPlH4
Do we need climate models to tell us that adding greenhouse gases to the atmosphere causes Earth to warm? No. The laws of thermodynamics say that they do.
Spider Boy
<would you call "sensitivity" one of those parameters, or is sensitivity supposed to be predicted by OTHER parameters>
Sensitivity is empirical. Climate models, particularly hindcasting, are used to calculate climate sensitivity.
Models are very scary. They work like magic. You take some data (the ingredients) and mix them according to a spell (the mathematics) and then run hundreds, thousands or millions of simulations and draw conclusions based on the probabilities of the various outcomes. A few decades ago this was impossible because computers weren't powerful enough, but these days complex bayesian analyses can use even small numbers of observations to produce simulations giving ideas of outcomes which are not possible to observe.
EG I had to design an experiment last year looking at how grazing cattle influences fire in alpine environments of southern australia. There was no baseline data so we had to use parameters of fire effects from a similar ecosystem and modify them repeatedly to produce multiple simulations, then run these simulations in the model. On the one extreme you could say it was an extremely clever exercise which utilised cutting edge technology to estimate the risks. Or you could go to the other extreme and say garbage in garbage out, the model is worthless until falsified by actually placing cattle into the ecosystem to see what happens (by which time it will be too late if they do more harm than good). But in either case you need a good understanding of the system to have an informed opinion. I've yet to encounter a denialist here who does, yet most of them are happy to disregard modelling without really understanding the specifics.
I mean, it IS possible that some models will be bad and others will be good AT THE SAME TIME. But if you aren't qualified to tell the good from the bad, a subjective opinion is the best you can come up with. Personally I think this is why it is great that we have experts, they are the ones qualified to make judgements and fortunately there are enough of them in science so numerous models can be made, compared and discussed. If there are bad methods or inductive conclusions in a scientific report, it will be discussed in the literature (and NOT just by denialists). I feel like we could quite easily disregard the opinion of any sufficiently qualified individual who has received monies from fossil fuel industries and related "charities" without any worry that bad science will not be identified and dissected by unbiased scientists. But I suppose I should be less sceptical of denialists and assume that the coal oil and gas monies are not influencing their opinions.
All models are not 'wrong'. Some are better at predicting than others though.
Regarding climate models: They require dozens of parameters. We don't even know what all of the climate driver parameters are. Most of the ones we DO know are loosely quantified. What anyone who has dabbled in modeling knows is that such a model can be expected to diverge in a hurry. This is born out by the dismal record of climate models up to this point. This won't change no matter how much effort we put into them.
"Give me four parameters and I can fit an elephant. Give me five and I'll make him wag his tail."-- Enrico Fermi
So, Mr. gcnp58, would you call "sensitivity" one of those parameters, or is sensitivity supposed to be predicted by OTHER parameters?
Are you implying that we get to make judgments here and there wrt these models? Who makes these judgments? As you note, these sensitivity judgments have varying effects on models.
Models are just that. They are a representation of what we know or think we know about a system. They are constantly being changed because we do not know as much about things as we think we know. When ever a model does not predict what we observe, we have to change the model.
The basic idea that you can model any extremely complex and interactive chaotic system and come out with anything of any value has not been proven. It sounds like a worthy goal, but it is really just a waste of time and money. Psychologists try to model human behavior with the same results. Economist have tried to model the world economy with the same result. Models are not science. They are speculation based on past history, and will by definition always be wrong. The real world and real life will always present surprises that models cannot predict.