Models - What Do They Represent?

[Norway Nut]

I understand that the models are mathematical i.e. numerical, representations of the chaotic atmosphere. There are competing "brands" ( GFS, METO, GEM, etc) and each one of the brands has its own atmospheric model giving rise to different forecasts. However, what do we get to see? Do we get a person moderated output or does the forecast come straight out of a zx80?

Could someone define all the different outputs we see and tell us the difference ?

[Gibli]

NN - they are not brands and are not really competing. The names of each model are usually abreviated, so we have :

GFS = Global Forecast System (A USA model)

GEM = Global Environmental Multiscale (An Canadian model)

NOGAPS = Navy's Operational Global Atmospheric Prediction System (A US Navy model)

Google 'weather model GFS' for instance and you'll see the full names for other models. The UK Met Office uses several models and you may see two, the Global (i.e. covers the whole world) and the NAE (North Atlantic and Europe). Think of it in the same way Ford produces different models Focus, Fiesta and Ka for instance, each with different dimensions and different engine capacities.

Similiarly, different models use slightly different configurations in the equations they solve and at different resolutions - maybe 60km, 30km or 10km. They may have slightly different starting conditions as well (e.g. those that start to solve equations at 1300 will have less data than those starting at 13:30) so they will come up with slightly different answers (represented as pressure charts for instance) ,

Specialists at main weather centres can and sometimes do intervene to correct things, but you would not be aware of this. Essentially the data / charts you see on the internet come straight out of the computer - no human intervention.

I would say models are not really competing. Some models tend to do 'better' than others in certain weather situations but all have known problems and no model is 'better' than the others all the time. Upper troughs advancing from the Atlantic for instance cause most models some problems - forecasters will know about this and make allowances in the forecasts. The ECMWF model is thought to be slightly 'better' than the others on average (depends partly on how you define better), but it starts solving the equations some hours later than the other models, so the information is not available so soon - swings and roundabouts

Hope that helps a bit.

Gibli

[knocker]

To be specific and just taking the METO this may help.

http://www.metoffice.gov.uk/research/modelling-systems/unified-model/weather-forecasting