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 11:10 AM ET, 04/16/2010

Weather models: More than just pretty pictures

By Dan Stillman

You, too, can predict the weather

* Short sleeves today, jackets tomorrow: Full Forecast | NatCast *

Want to predict the weather? Or at least understand some of the pictures and numbers meteorologists look at when making their forecasts? Most of this information is available on the Internet to anyone, free of charge. The tricky part is finding and making sense of the sometimes colorful maps and often confusing streams of numerical data. That's where we'll try to help, with this periodically appearing series on how You, Too, Can Predict the Weather.

In our last installment we began to explore the colorful maps produced by weather models, which are especially helpful for predicting the location, timing and amount of precipitation. We'll delve even deeper into the maps and the information they contain in future posts, but first its time to switch gears...

The models also produce alphanumeric forecasts called Model Output Statistics, or MOS (pronounced "moss"). These are predictions of various weather parameters, mostly at three-hour intervals, for specific locations. MOS is probably the easiest and quickest way to get an idea of what the models are thinking in terms of temperatures, clouds, winds, and chance of precipitation for the next couple of days and beyond.

The "One-Stop MOS" page at the National Weather Service Meteorological Development Lab Web site is one of the more convenient places to access MOS forecasts. After choosing a state or territory on the left side of the screen, check off one or more locations on the right and scroll all the way to the bottom. Pressing the "Submit Query" button will bring up the latest MOS forecasts for the location(s) selected.

All those rows and numbers and letters might seem quite confusing. Let's first get comfortable with several of the key rows in the "GFS MOS" or "NAM MOS" output, both of which mostly follow the same format. (If you remember, GFS and NAM are the two main U.S. weather models that we learned about last time.) Below is a sample MOS output and a key describing selected elements.

MOS screenshot (highlighting added). Credit: National Weather Service.

Line 1: Forecast location, model name, and date/time the model was run (if you've forgotten, or need a refresher, see our post on how to convert from UTC to local time)
Lines 2 & 3: Column headers for the rest of the table
DT=forecast date
HR=forecast hour in UTC (convert UTC to local time)
X/N: max./min. temperatures (°F)
00 UTC/8 p.m. EDT value is daytime high (highest temp in previous 12 hours)
12 UTC/8 a.m. EDT value is morning low (lowest temp in previous 12 hours)
TMP: temperature (°F)
CLD: cloud cover (CL=clear; FW=few; SC=scattered; BK=broken; OV=overcast).
WDR: compass direction near-surface wind is to come from, in tens of degrees.
WSP: near-surface wind speed (knots; convert to mph).
P06: probability of precipitation (PoP) during 6-hour period ending at that time.
P12: PoP during 12-hour period ending at that time.
TYP: precipitation type (R=rain; S=snow; Z=frozen).

Using the highlighted areas in the sample MOS above as an example, at Reagan National Airport (KDCA), the GFS model predicted the following conditions on April 17 at 06 UTC (2 a.m. EDT): 63°F, dew point 51°F, overcast, winds from the west-northwest (from 290° on a compass) at 7 knots. The morning low for April 17 was predicted at 57°F and the daytime high at 64°F.

*See for a description of elements not included above and additional information.

*The longer-term GFSX MOS, which predicts weather conditions at 12-hour intervals out to 8 days, has a little different format and is explained here:

As we briefly touched on in the last post, models are almost never perfect, and are sometimes significantly off. That's why MOS forecasts are referred to as "guidance" -- they are meant to guide rather than dictate a meteorologist's forecast. It's the job of the forecaster to anticipate the strengths and shortfalls of the different models given the expected weather pattern and local influences, such as nearby mountains or bodies of water.

Here are a few insider's hints as you try your hand at forecasting:

*The models tend to underestimate the effect of cloud cover on temperature. So, when overcast skies are forecast, daytime temperatures will often be a couple to several degrees lower than MOS, and nighttime temperatures a couple to several degrees higher.

*They also tend to underestimate the impact of wind on nighttime temperatures -- a light to moderate breeze can be enough to stir the atmosphere and keep temperatures from dropping as low as they would if winds were closer to calm.

*In the D.C. area at least, GFS MOS temperatures are often warmer than NAM MOS, and often accurately so. This is especially true with winds from the west or southwest, which is considered a warm, "downsloping" direction (air sinks and warms as it's blown from higher to lower elevations).

*An exception to the previous and first bullets above can sometimes be with a strong, warm flow from the south or southwest as a cold front approaches from the west. In this situation, temperatures often spike upward (even with cloudy skies) before the cold front comes through. NAM MOS (thanks to its higher resolution) sometimes better anticipates such spikes than GFS MOS.

*Strong winds from the north or northwest sometimes result in cooler-than-predicted daytime temperatures.

*One pitfall of a location-specific forecast such as given by MOS is that you can't see what's going on in surrounding areas. This is a particularly important consideration when it comes to probability of precipitation (PoP) and cloud forecasts. For example, MOS may predict a low PoP for a given location. But if you look at the model maps (like those we discussed last time), it may turn out that an area of precipitation is precariously close by.

*Note: MOS wind forecasts are for sustained winds. Depending on the overall weather pattern, wind gusts can be nearly or more than double the sustained wind speeds.

So there you have it. Between this post and the previous one in this series, we've now covered the basics of predicting both temperature and precipitation. Ready to go apply for a meteorologist position at your local TV station or National Weather Service office?

Not so fast... there are many more details and nuances to forecasting temperatures, precipitation and various other pieces of the weather pie. Not to mention, we haven't even touched on severe or wintry weather. The fun has just begun, and will continue when we meet next.

Some material in this post was excerpted and adapted with permission from the author's article in Weatherwise magazine.

By Dan Stillman  | April 16, 2010; 11:10 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: Short sleeves today, jackets tomorrow
Next: PM Update: A risk of showers and storms


In general, model guidance should be taken with a grain or two of salt, especially the farther out the prediction is made. The models tend to converge on a common solution as one gets closer to the selected forecast period.

Nonetheless, model guidance can sometimes be accurate even at long lead times. Some of the model runs "called" our December and February snowstorms as far out as ten days or two weeks out. One of the storms was forecast as a "snow event" for our area as far out as 185 hours, as I recall.

Posted by: Bombo47jea | April 16, 2010 12:20 PM | Report abuse

I like (no affiliation) for visualizations of MOS data. For example, here's DCA

Great post, keep it up.

Posted by: psilosome | April 16, 2010 5:34 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company