Ph.D. Student - Earth System Science (UC Irvine), B.Sc. - Atmospheric Sciences (Cornell University)
By: Zachary Labe , 6:27 PM GMT on October 28, 2011
"Current Surface Map and Weekly History of Jet Stream Position"
"12hr Estimated Precipitation"
"Forecast Max Temperatures"
"Forecast Min Temperature"
"Forecast Weather at 2pm"
"Current Storm Reports"
"Forecast Model Links"
-NAM model 12z...Link
-GFS model 12z...Link
-NMM model 12z...Link
-SREF model 9z...Link
This is a pretty quick intro on computer model forecasts along with beneficial links. Computer models use the complex calculus algorithms to print out the forecasts. Despite our complaning with the models, without them much of us would be lost except in the short time. Some of the earlier computer models consisted of the ETA, NGM, and AVN, which forecasted generally less than 84hrs. They were highly inaccurate, but provided a basis. The ETA was actually the computer model that helped meteorologists predicted the "storm of the century" in March of 1993 so well in advance. But now more than ever we have a myriad of computer models available to the general public with many mesoscale models only available to NOAA. Lets start with the general. All current computer models are based off on the Zulu time. Zulu time is also known as UTC or Greenwich time . Generally to forecast timing of storms my special BUFKIT data transitions UTC time to EST time, which is helpful. But BUFKIT is a special download, which I won't get into, but the transition is helpful as for some reason I never am able to understand time zones too well, hahaha. Anyways let me start with each computer model...
-GFS (Mentioned most of all as it is a global model (Global Forecasting System))
~Available in 0z (initiates at 10:30pm), 6z (initiates at 4:30am), 12z (10:30am), 18z (4:30pm)
~Forecasts out to 384hrs
Cold bias on long range on 18z run
6z and 18z slightly unreliable
Northwest trend on lows within 84hrs of event
-ECMWF (This is another global model run by an international organization (European Centre for Medium-Range Weather Forecasts))
~Available in 0z (initiates at 1:30am), 12z (initiates at 1:30pm)
~Forecasts out to 240hrs
Overphases lows in 168hr range
Holds too much energy in southwest
Known as extremely accurate within 140hrs
-NAM (Mesoscale short range model)
~Available in 0z (Initiates at 9:30pm), 6z (Initiates at 3:30am), 12z (initiates at 9:30am), 18z (Initiates at 3:30pm)
~Forecasts out to 84hrs
Highly inaccurate towards the 84hrs
Handles coastal storms very well and southwest overrunning events
Tendency for way too much QPF
Those three above are the most common models for a beginner in computer models, but there are many more. All of the global models consistent of ensemble models also, such as the GFS has a myriad of ensemble (small models) that create a mean solution known as GEFS. They typically are too cold and southeast with low pressures, but some reason the NWS seems to enjoy using them. There are also more global models than the GFS and ECMWF... The UKMET is run by an internation organization and forecasts out to 144hrs. This model typically comes out an hr before the ECMWF and usually is pretty similar to the ECMWF. The ECMWF may also be known as the EURO by the way. There is also a Canadian model known as the GGEM/CMC, which again contains ensemble models. All of the internation models only run 0z and 12z runs. This is probably for the best as all models only receive new upper air data in 0z and 12z runs, so this is why the American model runs of the 6z and 18z are usually worthless. There is also a high resolution Canadian model known as the RGEM, which is very similar to the American NAM. High resolution (mesoscale models) are important as they usually are able to pick up on fine details such as frontogenesis, advection, adebiatic cooling, convection, etc. Some of this high resolution models include the WRF, HIRES NMM, RUC, ARW. They all are usually very accurate, but the WRF and HIRES NMM usually have wet bias.
As mentioned above there are ensemble models which come up with a mean solution instead of using one computer model's algorithms like the global models use. These ensemble mean solution are known as the MREF and SREF with MREF being in the medium range and SREF in the short range. SREF is usually pretty accurate and forecasts within 87hrs of an event. There are also other computer models used for hurricane forecasting, but I will not get into them. For instance one is the GFDL, which you may have heard of. I find I use mainly the GFS, NAM, and WRF/NMM in the summer, but use all of the models in the winter.
A few models to avoid...
Here is a list of links for explanations on how to interperate the models...
-PennState E Wall, which runs all of the models
-PennState E Wall tutorial on computer models (Check it out)
-Severe Weather parameters used on models, explanation
-Forecasting winter weather
-NCEP; used to find American Computer models
-Severe Weather Models
-Winter Weather Models
The last link is listed above as model soundings which takes all of the data to print out all sorts of information including precipitation type along with dynamics such as Omega. This is very complex and takes some time getting used to. Also you may see this data instead of in charts, it is sometimes used in SKEWT T charts.
I hope all of this information helped out... Keep in mind precipitation amounts is QPF, with 500mb aloft being the jet stream, 700mb aloft measuring relative humidity, 850mb aloft measuring 5,000ft aloft temperatures, 925-1000mb measuring surface temperatures. Generally I would look at the GFS and NAM first to get a hang of it along with reading the tutorial links. Use the 850mb and 2m charts for the GFS especially as they are pretty self explanatory and color coded. You will find some maps for international models are confusing and black and white.
Lower Susquehanna Valley Doppler...
(Courtesy of WGAL)
Follow my 24hr forecasts on Twitter... Link and Facebook... Link.
Please post storm reports in this blog from across the Northeast during the winter storm and please post location of observation in each report...
This blog is in progress. Check back soon...
Winter Forecast 2011-2012... Link
Historic October Nor'easter...
Higher cirrus will begin to flood the northern Middle Atlantic later in the day Friday after high temperatures in the mid 40s with partly cloudy skies. Clouds will begin to thicken and lower Friday night with precipitation moving northward as a 1000hPa coastal low pressure moves up the coast. The low pressure will follow an ideal track up and along the coast and if it were winter... this would be a major snowstorm for all of the Northeast including towards the coast. But given climatological means during late October, snow will only fall in certain locations well to the northwest. Precipitation will move northward as rain into Virginia and Maryland Friday night and move up into southern Pennsylvania also as rain. Evaporational cooling will begin as 2m temperatures fall to near dewpoint values late Friday night changing over to snow for southern Pennsylvania particularily for elevations above 800ft with a rain/snow mix in the valleys.
As the low begins to intensify further, a deformation axis will likely form well to the west of the center of circulation. The placement of this axis is key to the snow accumulation region as it will bring cold air down the surface in the form of dynamic cooling. Given the low track this axis is likely to form in eastern Pennsylvania up through central New England. Saturday will feature well below normal highs withs highs not reaching 40 as far south as Washington DC. Precipitation will intensify during the day. Ground temperatures initially in the 40s and marginal boundary layer temperatures will keep the highest accumulations in the higher elevations with several inches possible above 200ft (perhaps up to 12in as far south as South Mountain, Pennsylvania). Elsewhere valley locations can also expect some light snow accumulation as far south as northern Maryland. For the I-95 corridor, a mix of rain and snow is likely with no accumulation. Snow accumulations up through New England are likely 4-7in for the valleys and 8-14in for the higher elevations. I could even envision higher amounts given the strength of the deformation axis. Total QPF printout from guidance such as the ECMWF and GEFS suggest 1.0in-2.0in across the entire Northeast.
Most of the snow will fall during the day Saturday. If timing were a bit more favorable such as Saturday night, this would have been a potentially record breaking, rare October snow event.
Given the high leaf foliage in some areas, the combination of wet snow will cause major tree damage in some locations especially in eastern Pennsylvania. Power outages are possible. Snow ratios will be near 6/7:1 with rates up to 0.5in/hr at times. Given the heavier snow rates at times, even areas with boundary layer temperatures at 34-35F may see a quick accumulation.
I want to raise a few critical points...
1) Current ground temperatures really have little to do with snow accumulation when rates fall above 0.5in/hr. Snow can accumulate even on roads. The sun angle in late October is like mid February and has little impact. Combine this with marginal temperatures, cloud cover, and heavy snow rates this will cause snow to accumulate very easily after the initial lighter snow.
2) The biggest concern is the trees which I have been hyping for the last 24hrs. Snow accumulations possibly over 10in in some areas of south-central Pennsylvania up through New England could bring devastation to trees. The heaviness of the wet snow on foliage is a disaster. And looking a records, this has never happened before in many areas!
3) I-95 is a tough forecast and is dependent on exact low track. But given the dynamics of this storm which may cause even thundersnow, the big cities should all change to snow at some point.
4) Given the time of year, water temperatures are very warm and the temperature gradient and the natural baroclinicity along the coast will cause the storm to rapidly undergo bombogenesis with very unusual mesoscale features. This will cause some areas to see rapidly high QPF totals.
Selected City Accumulations for the Northeast...
Hagerstown, MD- Rain/snow then heavy snow. Accumulations 5-12in
Baltimore, MD- Rain changing to snow. Accumulations up to 4in especially the higher hills.
Salisbury, MD- Mostly rain with a trace of snow
Pittsburgh, PA- Rain/snow then snow. Accumulations 2-4in
State College PA- Moderate snow. 5-7in
Williamsport, PA- Heavy snow. Accumulations 5-10in
Altoona, PA- Moderate snow. Accumulations 5-7in
Harrisburg, PA- Heavy snow. Snow accumulations 6-14in
Lancaster, PA- Heavy snow. Snow accumulations 6-13in
Philadelphia, PA- Rain/snow changing to snow. Accumulations 1-3in+
Allentown, PA- Heavy snow. Snow accumulations 6-14in
Scranton, PA- Heavy snow. Snow accumulations 6-12in
Washington, DC- Rain changing to rain/snow. Spotty 1-3in accumulations are possible
Wilmington, DE- Rain changing to rain/snow. Spotty 1-4in accumulations are possible
Dover, DE- Rain changing to rain/snow with trace accumulations
Trenton, NJ- Rain/snow changing to snow. Accumulations 2-4in
New York City, NY- Rain/snow changing to snow. Accumulations 1-6in depending on location.
Poughkeepsie, NY- Heavy snow. Accumulations 5-10in
Binghamton, NY- Light snow. Accumulations 3-6in
Ithaca, NY- Light snow. Accumulations 2-4in
Albany, NY-Moderate snow. Accumulations 4-8in
Hartford, CT- Rain/snow changing to heavy snow. Accumulations 4-7in
Concord, NH- Heavy snow. Accumulations 5-10in
Providence, RI- Rain changing to rain/snow. Spotty 1-3in accumulations
Worcester, MA- Heavy snow. Snow accumulations 6-13in
Boston, MA- Rain/snow with spotty accumulations 1-3in
Nantucket, MA- Heavy rain and high winds up to 60mph
Hyannis, MA- Heavy rain and gale force winds in excess of 60mph
Portland, ME- Heavy snow. Accumulations 4-9in
Bangor, ME- Heavy snow. Accumulations 6-12in
"Subject to Change"
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.