Computer Model Predictions

[Callipygous]

im curious about the computer models used to predict things like weather patterns and global warming. how accurate are they? how do they work? how much can we rely on them?

 

i need to write a paper on this so detail in how they work is always good. also, any good sources you guys have come across?

[5614]

They use data logging systems to collect data about the current conditions and using pre programmed algorithms they can predict the likeliest possibility as to what the weather will be like in a few days. So on a very basic level if it senses clouds it realises that at some point they will condense/rain (that's very basic though!), it will analyse types of clouds, pressures, temperature, wind strength/direction etc. (probably more that I didn't think of!)

 

As for reliability, well technically rather than telling you what will happen they are predicting the most likeliest outcome of the current conditions (ie. it is cloudy today, it'll rain tomorrow - again that's pretty basic)... so how reliable are they? Well after this many years of being used the programs are pretty complex and reliable, however it is important to note that these are merely probabilities... weather forecasters often say "it will be cloudy tomorrow with the possibility of some rain fall" note the "possibility" meaning it is not certain... if you look at newspapers it often says "possibility of rain, clearing as the day goes on" so that has covered rain (and thus clouds) as well as sun, that's effectively everything, or in the UK it is anyway, where snow is rare.

 

Also you've got to realise that very few people critically analyse weather forecasts, most people see the cloud picture and that's that... if it's cloudy they don't think of it, if there's rain they'd justify 'well it is cloudy and clouds rain' and if it's sunny 'well there can't be clouds everywhere!' and people just pass it off as totaly insignificant.

 

Forecasters often mentioned clouds/rain or clouds/sun in one sentence, which just shows how whilst they are on the right tracks it is not a certainty.

 

They analyse the current situation and use that data to predict the likeliest outcome.

[Callipygous]

hmmm.

 

interesting, but what im looking for is really focussed more long term, like the global warming. for example, one of the sources i have mentions that they have predicted that by 2080 all of the ice in the atlantic ocean will have receded/melted and it will just be all water. things like that are what im more interested in. and how they predict that the overall temperature of the earth will have raised x degrees in y years.

[5614]

I am not reall that familiar with predicting global warming stuff, but one would assume they look at how much has melted over the last 50 years, how much temperature has raised, and why, once they know why (ie. polutant gas) they can start to measure how much is produced and then what effect that will have.... although note the "assume" at the beginning!

[Aeternus]

Not exactly inline with what you want but something related to it none the less and interesting considering the topic - http://www.guardian.co.uk/life/feature/story/0,13026,1496690,00.html?gusrc=rss

 

Something that might be useful (not the actual software but the manual ways of analysing the weather could be useful in your paper for comparison and might help further understand the way in which the trends are analysed on a grander scale using computer software/algorithms) - http://www.theweatherprediction.com/

 

Also you might want to see if you can find this book - http://portal.acm.org/citation.cfm?id=603326 (look in the library, order it in, or otherwise) as it seems to have alot of what your looking for.

[5614]

Now I will honestly say I haven't read more than a paragraph of this but it seems okish if you have some spare time!!!

http://www.gcrio.org/CONSEQUENCES/fall95/mod.html