Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 11:00 AM ET, 03/22/2010

Watching the weather models: Precipitation

By Dan Stillman

You, too, can predict the weather

Example of output from the GFS weather model.

Weather is the ultimate news story: It's always changing; it impacts everyone and everything; you never know for sure how it'll turn out; and it happens 24/7/365. Weather is the New York City of sciences. The science that never sleeps.

To keep up with the shifting winds, fluctuating temperatures, and constant movement of clouds and precipitation, computer models around the world churn out a seemingly endless stream of data both day and night. Did you know that most of the model-generated maps and numbers that meteorologists use to predict the weather are available to anyone and are just a few clicks away? They can be understood, even by the untrained weather enthusiast, with a little explanation.

Keep reading for where and when to access weather models, and an introduction to interpreting them...

The two main models used to forecast weather in the United States are known among meteorologists and weather junkies by their acronyms -- NAM and GFS -- more so than their full names (the North American Mesoscale Model and the Global Forecast System). Both models are operated by the National Weather Service's National Centers for Environmental Prediction (NCEP).

Output from these models and others are available at various government and university Web sites, as well as from a handful of privately funded sites. For this tutorial, we'll go straight to the source -- the NCEP models page at

Screenshot of NCEP models page.

Along the left side of the table on the page you'll see the selection of models available. In this post, we'll focus on the first two, NAM and GFS. Both models are run four times a day using current weather observations from around the world as input. The model run times are listed across the top of the chart:

00 UTC (7 p.m. EST/8 p.m. EDT the previous day);
06 UTC (1 a.m. EST/2 a.m. EDT);
12 UTC (7 a.m. EST/8 a.m. EDT);
18 UTC (1 p.m. EST/1p.m. EDT).

(See our previous post, "How to Tell Time Like a Meteorologist," for more on UTC, the standard time used to track and report meteorological information.)

The output from each NAM run is usually available about 2-3 hours after the model run time, while GFS takes closer to 4 hours. The date shown at the bottom of a given model/run box in the table indicates whether output from the present day's run is available.

Screenshot from NCEP's GFS model page.

Let's check out the latest output from the GFS. Graphics are provided in three resolutions -- coarse, medium and fine. For starters, I'd recommend clicking on the "medium" link for the most recent GFS run available. What you'll arrive at is a page with a different chart, this one with more columns and numbers than the last. We'll keep things simple here, aiming only to understand the series of maps linked to in the first column (labeled "10m-Wnd 06hr Pcpn").

Each number links to a map that will show predicted conditions for that many hours after the model run time (The 000 map shows what the observed conditions were at the model run time.) Let's take a look at the 24-hour forecast map. (Note: The maps are made available in chronological order as they are generated. If the 24-hour map isn't available when you happen to the visit the page, feel free to pick one that is.)

In this post, we're going to ignore everything except for the colored areas, which represent how many inches of liquid precipitation (legend is on the left) are predicted to fall during the 6 hours prior to the forecast time, which is shown below the map. Just remember, models are almost never perfect with their placement and amount of precipitation. So, an area showing no precipitation could still get some, especially if it's close to a precipitation area. And an area showing precipitation could end up staying dry, especially if it's near the outer edge of a precipitation area.

Now, let's compare the GFS model's 24-hour forecast map with that from the NAM model. To get to the desired NAM map, go back to the main NCEP model page (I'd suggest doing so in a new window, so that you'll be able to compare the maps side by side) and click the "medium" link for the most recent NAM run available. The column headers aren't quite the same on this page as on the GFS page. The maps in the second-to-last column will have what we want. Go ahead and click on the 24-hour forecast map in that column (labeled "MSLP 1000 - 500mb").

Again, ignoring everything except for the colored areas, does the NAM map look similar to the GFS map? Are there significant or slight differences in the amount and location of precipitation areas?

Part of being a good forecaster is getting to know which models are more accurate for different kinds of weather and weather patterns. NAM is a higher-resolution model than GFS, the result being that NAM tends to more precisely pinpoint gradients in precipitation and may identify small-scale areas of precipitation better than GFS. The price for NAM's higher resolution is that it only goes out to 84 hours, while GFS goes out to 384 hours. That said, each model has its own set of strengths and weaknesses -- the subject of a future post.

Now that we have an idea of where and how much precipitation is going to fall, the next question is what kind of precipitation will it be? Well, that depends on temperatures, which will be the topic of our next installment.

By Dan Stillman  | March 22, 2010; 11:00 AM ET
Categories:  You, Too, Can Predict the Weather  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Back to reality - wet and not as warm
Next: PM Update: First thunder of the season


Thank you so much for this post! It was extremely helpful. I can't wait for the next one!

Do you know any other good sites that explain weather models? Thanks!

Posted by: Yellowboy | March 22, 2010 1:03 PM | Report abuse

Thank you so much for this primer on weather forecasting, I think you should compile this series and get a book contract!

Posted by: Hoyas4Ever | March 22, 2010 1:50 PM | Report abuse

Hi Yellowboy - few other sites give good explanations. Thus here at CWG we're trying to summarize all the knowledge we learned in our educations. That was the only place I found a thorough lesson on weather models; however, if other CWG members have found sites online that I don't know about, I bet they will post more for all of us! :)

Posted by: Camden-CapitalWeatherGang | March 22, 2010 2:02 PM | Report abuse

One of the best sites for background on Numerical Weather Prediction (NWP), i.e., weather models is: You'll have to register, but there are no costs or other restrictions

This might be a bit to heavy for some, but not for those really interested.

See also Click on links to "A guide to interpreting the forecast images" and " A guide to interpreting the precipitation/temperature forecasts"

As I've offered before, I'll be glad to handle inidvidual questions off line. Email me at

Posted by: SteveT-CapitalWeatherGang | March 22, 2010 2:22 PM | Report abuse

Is that the first TS Warning of the year I see! Nasty looking cell - heading right up the 95 corridor quickly. Heads up DC and slightly east in about an hour or so! Big hail up to 1.25 indicated...

Posted by: DullesARC | March 22, 2010 2:46 PM | Report abuse


Posted by: Yellowboy | March 22, 2010 3:17 PM | Report abuse

Nice post and also thanks for SteveT's links.

Posted by: spgass1 | March 22, 2010 10:51 PM | Report abuse

Thanks muchly for this.

Posted by: nlcaldwell | March 23, 2010 1:02 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company