General Discussions: ASK John: Models

By coldbear on Monday, December 22, 2008 - 08:24 pm: Edit Post

When speaking of weather,many meteroligist refer to their findings as a model.Will you explain what this termonoligy represents? An early thanks for the reply.

By frnash on Tuesday, December 23, 2008 - 03:14 am: Edit Post

"… many meteroligist (sic) refer to their findings as a model."

While John is cutting z's at this hour, let me lend a hand.

Actually the term "model" does not refer to the meteorologists' findings, but to the resources they use in developing their forecasts.

This involves the arcane art of Numerical weather prediction, using Supercomputers to run (i.e.compute) very complex mathematical simulations (Mathematical Models) of the atmosphere.

There are a number of different models in use, for example:
Global Weather Models:
 The European Centre for Medium-Range Weather Forecasts Model (ECMWF),
 Global Environmental Multiscale Model (GEM),
 Global Forecast System (GFS),
 Unified Model (UM),
 Japan Meteorological Agency (JMA).

Regional and mesoscale models:
 The Fifth-Generation NCAR / Penn State Mesoscale Model (MM5),
 North American Mesoscale Model (NAM),
 Rapid Update Cycle (RUC),
 Weather Research and Forecasting Model (WRF),
 The High Resolution Limited Area Model (HIRLAM).

Here's more info on Atmospheric models.

See also: Use of forecast models, quoted here in part [emphasis is mine]:


In the past, the human forecaster was responsible for generating the entire weather forecast based upon available observations. Today, human input is generally confined to choosing a model based on various parameters, such as model biases and performance. Using a consensus of forecast models, as well as ensemble members of the various models, can help reduce forecast error. However, regardless how small the average error becomes with any individual system, large errors within any particularly [sic] piece of guidance are still possible on any given model run. Humans are required to interpret the model data into weather forecasts that are understandable to the end user. Humans can use knowledge of local effects [Such as Lake Effect Snow (LES)!] which may be too small in size to be resolved by the model to add information to the forecast.

Here's a picture of just one of the two IBM POWER5+ Cluster 1600 systems in the High Performance Computer Facility (HPCF) at The European Centre for Medium-Range Weather Forecasts, where a new milestone was reached just yesterday.

The various models will often yield differing results, due largely to the limited number of both surface observations and upper atmosphere data obtained from radiosondes on weather balloons, and in recent years, data transmitted from commercial airplanes through the Aircraft Meteorological Data Relay (AMDAR) system.

In my humble opinion, the interpretation of that diverse model output, among other factors, and the translation of that data into accurate weather forecasts that are understandable to the end user represents the true art of weather forecasting!

That's the "short answer" to your question from the perspective of a "Mathemagician" and Software Engineer, and I will now hand this off to the Meteorologist. :-O

By admin on Tuesday, December 23, 2008 - 09:51 am: Edit Post

Looks like FRNash pretty much covered all the bases.

To summarize. The models we speak of are just computer programs that take data collected from the current state of the atmosphere and apply mathmatical computations on that data to try and emulate what the atmosphere is going to do.

Those computer progams, or models, then spit out what they think the atmosphere will look like at differing time periods in the future. Meteorologists then look at the output from these computer models and try and figure out what is the most likely scenario for the future.


By canon on Tuesday, December 23, 2008 - 02:04 pm: Edit Post

The usage for 2008 has surpassed 100 million units. This corresponds to 7.5 million CPU hours on the HPCF POWER5+ systems (see here for a description of these systems). The previous milestone (over 10 million units) occurred in the fourth quarter of 2004.

Im a little confused what a unit is represents, is this the
amount of users or the data collection devices?

By frnash on Tuesday, December 23, 2008 - 04:08 pm: Edit Post

"Im a little confused what a unit is represents …"

You & me both. I didn't feel much like researching it any further at 3:14AM, though. However my curiosity remains, so I'll try to dig up an answer sometime soon. :-)

By canon on Wednesday, December 24, 2008 - 09:29 am: Edit Post

A "unit" in super whopper Techno-jargon might
be some sort of conversion that represents the
speed or storage capacity of the system.. r.aspx

Those are all worthwhile links..


Have a Merry Christmas.

By coldbear on Friday, December 26, 2008 - 03:44 pm: Edit Post

Thanks for those great explanations.Thanks to frnash,canon,and admin. for your inputs. Now let me try to interpret this data. Happy New Year and stay alive.

Add a Message

This is a private posting area. Only registered users and moderators may post messages here.

Topics Last Day Last Week Tree View    Getting Started Formatting Troubleshooting Program Credits    New Messages Keyword Search Contact Moderators Edit Profile Administration