Tuesday, January 22, 2008

Weather 101: How To Read Computer Model Graphics

One of the most common maps you see on Phillyweather.net are maps that are generated by NCEP (National Centers for Environmental Prediction), who generate free distribution maps for the GFS and NAM computer models, among others. Both of these computer models use the same graphics and run four times a day.

This edition of Weather 101 will take some of the basic points that are on a run of the NAM computer model and show you what it is saying and tell you what a possible forecast outcome will be based on this model. We are using one map to illustrate the points of what one map is showing. However, it's important to look at the whole context of what the model is showing and not just one map as multiple maps will show the direction a storm is moving, how fast it is moving, and changes in temperature associated with weather patterns. Additionally, one model's interpretation of upper air data may be different than another so looking at multiple computer models in analyzing what happens is important as well. In the case of the NAM it forecasts out to 84 hours from its start time, or 3 1/2 days away. The GFS, for reference purposes, forecasts out to 384 hours or 16 days away.

Feel free to click on the maps below to see them in a larger size.

Some of the most common questions that I run into are how to understand and develop a sense of reading these maps, what they entail, and what the map is showing in Philadelphia.
Based on this computer model run up above, the air 5000 feet above the surface is forecast by the NAM to be below -10 Celsius, which would equate to a temperature in the high 20's to low 30's in Philadelphia if this model were specifically correct regarding temperature. Since Philadelphia is north of the second blue line above the red line, this correlates to temperatures at 5000' (or the 850 millibar level of the atmosphere) of below -10 Celsius. A common equation to use with temperatures at this level when figuring out surface temperatures is to add between 5 and 15 degrees Celsius to the 850 temperature, with the lower numbers being used in winter and higher numbers used in the summer. A good 850 rule explanation is available at Wxrisk.com.

Why is the temperature at 850 mb important? Many times it is the indicator of what precipitation type falls from the sky weather it is liquid or frozen. There can be sleet if there is a warmer layer above the 850 level but it can be a good indicator of understanding where there is potential warm layers to deal with. In this case, since temperatures are as cold as they are the NAM would be indicating a likely snow scenario if it were correct for the time frame it is forecasting.

The low pressure system on this specific run of the NAM is broad (no distinct L in the Atlantic near the lowest pressure number on the graphic above) and high pressure is located over Illinois on the map (the black H up above). The closer the black lines are to each other and the closer low pressure is to high pressure the stronger winds will typically be. In this case, it may be breezy since high pressure is strong and there is some tightening of the wind gradient over Western Pennsylvania) but winds should be under 20 mph at the time of this forecast.
It's important to note how much precipitation falls as well. In this run of the NAM for Philadelphia we can see there will be at least .10" of liquid precipitation over a six hour period. Keep in mind that precipitation output on the NAM and GFS is always for the PRIOR six hours. In the case of this particular storm, the NAM itself is forecasting about 1/4" of liquid altogether but you can see that it is at least 1/10" of an inch as Philadelphia is in the second shade of green on the chart on the left. In South Jersey, the NAM is forecasting more precipitation in a six hour period as they are in a darker shade of green.
The time is important as well. In the case of NCEP model graphics, all times are in UTC or Greenwich Mean Time. The NAM map above indicates a forecast time of 7 PM EST for Thursday night. Model graphics will have times of 00 UTC (7 PM EST), 06 UTC (1 AM EST), 12 UTC (7 AM EST), and 18 UTC (1 PM EST). If we are in daylight savings time you want to add an hour to the times that are forecast.

On the bottom you can see two sets of times and dates. The 18 UTC is the run time of the model (meaning it was run at 1 PM)...that time is on the left next to the date of the model run (if you click on the map you can see the date is 01/22/08 -- that was the date the model forecast was made and the time the model ran was 18 UTC. On the bottom right is the time the forecast is valid for -- 01/25/08 at 00 UTC -- or 7 PM EST on Thursday night.

Now that we have collected all the pieces of data above, we can add each piece up to get an idea of what one computer model is suggesting for Philadelphia. The NAM in this case is thinking that about 1/4" of liquid precipitation will fall from 1 PM to 7 PM on Thursday afternoon, with temperatures that are likely below freezing, and breezy conditions. Without looking at all of the other layers of the atmosphere, we can deduce that we will likely see snow from this if this model is correct. Of course the computer models are not always correct but this is one computer's idea behind what will happen.

As weather conditions warrant over the coming weeks I will continue to do posts that will highlight different computer model maps and what they show as well as how to read upper level weather forecast maps that are put out by the GFS, NAM, and other computer models.