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, 01/20/2011

NOAA-NWS begins Winter Weather Experiment

By Steve Tracton

Weather Service aims to improve winter forecasts

Radar image of "mesoscale" snow bands slamming New York City and New Jersey the evening of December 26. Yellow shades indicate very heavy snow. NOAA's Winter Weather experiment will attempt to better predict this kind of small scale weather feature. Source: Weather Underground

Despite major advances in scientific understanding and the development of computer-based weather models, there remains much room for improvement. Anyone who has followed this season's winter storm predictions has, no doubt, observed this. While there never will be perfect forecasts due to chaos (i.e. the butterfly effect), there is considerable opportunity for more accurate predictions and better estimates of associated levels of confidence in hazardous winter weather systems.

To address these issues, the National Center for Environmental Prediction's (NCEP) Hydrometeorological Prediction Center (HPC) - part of the National Weather Service (NWS) - recently embarked on what it refers to as the Winter Weather Experiment (WWE). The WWE, which will extend through February, is designed to identify promising techniques and tools most useful in addressing three especially challenging problems in Day 1 and 2 predictions:

1) Precipitation type and amount
2) Mesoscale (i.e. small scale) snowbands
3) Dealing with the ever present degrees of uncertainty in forecasts

Keep reading for details on this experiment...


The common challenge of items one and two is resolving the smaller-scale details ("mesoscale", horizontal dimensions from about 10 to several hundred miles) embedded in the larger-scale ("synoptic scale", extending from hundreds to around 3, 000 miles in horizontal extent) storm system which you'd typically see in routine analysis of surface and upper air weather maps.

As readily apparent from the storm earlier this week, the transition zone from snow to freezing rain can be very narrow and is dependent upon details in the vertical profile of temperature and moisture. Mesoscale snowbands can be as narrow as 5-10 miles in width and up to a few hundred miles in length. They reflect instabilities in the airflow over the sloping surfaces ("slantwise convection") of frontal systems aloft tied to the detailed structure of mesoscale disturbances. As in the recent experiences of the blizzards affecting the northeast (but not D.C., ugh!), their occurrence can significantly enhance snowfall rates and areal distribution of total amounts.

An illustration of where heavy precipitation bands tend to set up around low pressure areas. These small scale features can be very impactful but are difficult to model. Source: University Center for Atmospheric Research.

Enhancements in the predictability of precipitation type/amount and snowbands, of course, require accurate forecasts in the timing, track and intensity of the developing synoptic-scale (larger scale) storm. And, not incidentally, this larger-scale background can be affected by complex interactions with the mesoscale features in a feedback loop which tends to decrease the overall predictability on all scales. Did I say we're dealing with a formidable challenge?

A principal approach in the WWE is to further explore the use and value of the current modeling systems and procedures along with experimental high-resolution (4 kilometer) models and ensemble systems. Crucially important is forecaster interpretation of data sets and products derived from models in producing new products and displays.

Emphasis is upon quantifying and characterizing information on case dependent levels of uncertainty/confidence, which addresses a publically stated high-priority strategic goal of NOAA/NWS and NCEP: emphasizing improved decision support services.

The WWE is being executed within the context of the HPC Hydrometeorological Test Bed (HMT-HPC). A key aspect of the HMT is facilitating interactions among researchers and operational forecasters, including rotation of experts from within NCEP and NWS Weather Forecast Offices (WFOs).

Experimental WWE forecasts are generated once per day for Day 1 (12-36 hours) and Day 2 (36-60 hours) from 00Z (7 p.m.) and focus on a single storm system selected in advance. These forecasts are produced essentially independent of HPC's, so comparisons can be made between the corresponding operational and experimental predictions.

The potential benefits that might emerge from the WWE include determining the value and operational viability of experimental guidance that could be integrated into HPC operational procedures. Such complements to current operations might include more detailed and generally higher confident forecasts of hazardous winter weather.

But, success is certainly not guaranteed. Still, it is expected that knowledge and understanding of model capabilities/limitations and lessons learned from forecaster experience will lay a firm foundation for the design and conduct of prospective WWEs in years ahead. I'll be doing a follow-up post when an evaluation of the current WWE becomes available.

Important note: In view of the high impact of hazardous winter weather, NCEP has reallocated resources to spin up the Winter Weather Experiment this year and will continue to find ways to continue this work even in the face of the challenging budget environment.

NCEP does face limits in its capability to implement the new higher resolution models and ensembles being assessed in the WWE on the current operational computer. Current plans for acquiring the next generation computer are expected to enable running the higher-resolution models operationally by 2015. In the meantime, experimental runs will continue as NCEP partners with other research organizations to continue evaluating higher resolution modeling systems in prediction of extreme and high impact weather of all types.

By Steve Tracton  | January 20, 2011; 11:00 AM ET
Categories:  Government, Latest, Science, Tracton, Winter Storms  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Wintry cold by day, flakes by night
Next: Tonight's snow to be very light, again


A bunch of weather geeks are going to be doing the WWE? I really don't think they stand much of a chance in the ring. WHATCHA GONNA DO WHEN HULKAMANIA RUNS WILD ON YOU???

Posted by: rwalker66 | January 20, 2011 11:29 AM | Report abuse

Better forecasting of mesoscale snow bands should pave the way for thundersnow forecasting...something the NWS should have started doing years ago.

The current situation reminds me a bit of the pre-1950 reluctance to mention possible tornadoes when NWS' predecessor the Weather Bureau was forecasting severe thunderstorms. Apparently thundersnow is more frequent than earlier supposed; there seem to be at least two major preconditions for convective snow events. One is the presence of warmer water beneath cold air--the lake-effect situation; the other precondition involves deep or rapidly deepening synoptic low pressure. Occasionally both preconditions are present simultaneously in the Great Lakes and the Atlantic Coast. A third possible precondition for thundersnow may involve rapid orographic forcing in mountainous regions.

Posted by: Bombo47jea | January 20, 2011 11:38 AM | Report abuse

I find it very interesting that the NWS is focust on such a project when current models have trouble hammering out synoptic patterns.

Although maybe the two are mutually exclusive. I'm guessing they deem it more plausible to update mesoscale forecasting than synoptic forecasting. I guess that makes sense considering that chaos is probably less confounding in shorter chronology and smaller areas. But it would be nice to see the NWS bring out new computer models that had higher resolution forecasting further out from possible events.

Posted by: bbirnbau | January 20, 2011 12:28 PM | Report abuse

Yay for science.
The experiment has already started; nothing yet on how it's doing?

Posted by: FIREDRAGON47 | January 20, 2011 6:49 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company