Comments on: Climate models fail to predict short-term shifts http://www.futurity.org/earth-environment/climate-models-fail-to-predict-short-term-shifts/ Research news from leading universities Sun, 19 May 2013 18:07:41 +0000 hourly 1 http://wordpress.org/?v=3.0.4 By: nerf herder http://www.futurity.org/earth-environment/climate-models-fail-to-predict-short-term-shifts/comment-page-1/#comment-155202 nerf herder Tue, 25 Sep 2012 19:42:00 +0000 http://www.futurity.org/?p=145762#comment-155202 You set up a model using a set of parameters for the present and existing conditions and see how well it predicts what happens. There's no way you can (or need to) incorporate all the historical data in making the model, so it's entirely valid to test it in various historical situations. Volcanoes and such are disruptive events, but we certainly can't rely on volcanoes to change the climate for us in the next 30-60 years, we need to take our own actions now, if possible. Adapt our response later, if needed. In the end though, economics is 80% of decision making, cultural/historical factors are another 15%, global warming and other environmental concerns add up to about 5%, so all the models in the world will have very little input to the actual policy decisions we make. (As you probably know, 68% of all statistics are made up on the spot, but I think these are reasonable estimates). You set up a model using a set of parameters for the present and existing conditions and see how well it predicts what happens. There’s no way you can (or need to) incorporate all the historical data in making the model, so it’s entirely valid to test it in various historical situations.

Volcanoes and such are disruptive events, but we certainly can’t rely on volcanoes to change the climate for us in the next 30-60 years, we need to take our own actions now, if possible. Adapt our response later, if needed. In the end though, economics is 80% of decision making, cultural/historical factors are another 15%, global warming and other environmental concerns add up to about 5%, so all the models in the world will have very little input to the actual policy decisions we make. (As you probably know, 68% of all statistics are made up on the spot, but I think these are reasonable estimates).

]]> By: Les McNamara http://www.futurity.org/earth-environment/climate-models-fail-to-predict-short-term-shifts/comment-page-1/#comment-138882 Les McNamara Tue, 18 Sep 2012 23:46:32 +0000 http://www.futurity.org/?p=145762#comment-138882 If the IPCC models are tested using historical data, isn't there a high likelihood that the 'skill' of the models will be over-estimated because the calibration and validation data are essentially the same? Also, if we are looking for temporal precision of less than thirty years and spatial precision smaller than continents, aren't we evaluating these global-scale, long-term climate models at a higher resolution than intended by the developers? For shorter timeframes, I imagine there will always be limitations on the reliability of models due to our inability to forecast events - such as major volcanic eruptions - even if we can collect and process the amount of data needed to simulate local and regional repsonses and interactions. If the IPCC models are tested using historical data, isn’t there a high likelihood that the ‘skill’ of the models will be over-estimated because the calibration and validation data are essentially the same?

Also, if we are looking for temporal precision of less than thirty years and spatial precision smaller than continents, aren’t we evaluating these global-scale, long-term climate models at a higher resolution than intended by the developers?

For shorter timeframes, I imagine there will always be limitations on the reliability of models due to our inability to forecast events – such as major volcanic eruptions – even if we can collect and process the amount of data needed to simulate local and regional repsonses and interactions.

]]>