Accuracy of weather predictions

[jake williams]

Is there any interesting or readable information (ie. not scientific papers, I'm more interested in things like colourful three dimensional graphs and cleverly written blogs by amiable specialists) regarding the accuracy of weather prediction over time?

So, for example, the average error of predictions of temperature or precipitation on the Nth day after as a prediction is made? (Or even better, information about how this accuracy itself changes over time?)

Just curious.

[Rocky Rococo]

I heard the prediction for tonight was widely scattered satellite.

[CleverTitle]

I don't have anything immediate, but I'd assume that the SPC (http://spc.noaa.gov/) and CPC (http://www.cpc.ncep.noaa.gov/) (and maybe NWS (http://www.crh.noaa.gov/crh/) or HPC (http://www.hpc.ncep.noaa.gov/)) sites have some info on the topic.

EDIT: Now that I think about it, the EMC (http://www.emc.ncep.noaa.gov/) site might be your best bet, though I'm not sure.

[Q]

After some googling, this might help (http://www.esrl.noaa.gov/psd/spotlight/12012001/).

I also made a post that I could recycle in the "predict a revolution" thread:



I don't believe it'll be able to ever predict the future revolution any more than perhaps a few weeks in advance; basically for the same reason that we can't predict the weather with any reliability more than three weeks in advance. The reason is scientifically explained in chaos theory (http://en.wikipedia.org/wiki/Chaos_theory).

In the case of the weather: all laws and mechanics involved are well understood, yet we can't predict what will happen when. This leads to what is often popularised as the "butterfly effect": The wing flapping of one butterfly in the US can cause and contribute a mighty storm in China weeks later.

So we can predict the weather to some reliability in advance, but the further into the future we look, the more important tiny factors can become. I don't see how any "predict the revolution" computer would escape such limits.

Furthermore, DNZ is quite right that what we need is conscious political action of the masses, something that goes beyond "mere" labour struggles (which stay within the logic of the system). Essential for this is the organisation of the class as a class for its own and political programme that is aimed at going beyond the system. So, we need more than a computer to know when to launch the "spark", what is needed is "revolutionary patience" in building mass organisations that are armed with a clear historical mission to change society.
Not exactly an answer to your question, but I don't know how familiar you are with the problems involved.

[jake williams]

I don't have anything immediate, but I'd assume that the SPC (http://spc.noaa.gov/) and CPC (http://www.cpc.ncep.noaa.gov/) (and maybe NWS (http://www.crh.noaa.gov/crh/) or HPC (http://www.hpc.ncep.noaa.gov/)) sites have some info on the topic.

EDIT: Now that I think about it, the EMC (http://www.emc.ncep.noaa.gov/) site might be your best bet, though I'm not sure.

After some googling, this might help (http://www.esrl.noaa.gov/psd/spotlight/12012001/).
Those links look useful but would either of you happen to have anything aimed at non-specialists? There is a lot of info on those pages and I'm not exactly sure how to find what I'm looking for (how much predictions vary depending on the timescale over which they're made).


Not exactly an answer to your question, but I don't know how familiar you are with the problems involved.
I know virtually nothing about meteorology but quite a bit as an uneducated science geek. I can understand the general sorts of principles involved and what sorts of relationships and processes could generate error in prediction, but I have no idea what the actual relationships are processes are. So, I get the idea of how complex systems work in general, but not with weather specifically, so I have no way of judging quantitatively for how long predictions would be reliable, and to what extent.

[MarxSchmarx]

Check this out:

http://www.scientificamerican.com/article.cfm?id=weather-forecasters-set-s

off-topic rant:
Non-meteriological scientists routinely make claims in the popular press about whether their modeling efforts amount to much.

For example, you can search the bbc archives to find examples of biologists and economists saying things like "Well it's like the weather, we don't know what we're doing but it's better than a random guess" or some such. ON some level this is (comparatively educated) non-specialists commenting on the ability of meteoroligists to predict the weather.