Why the differences between the ECMWF & GFS for the Nor-Easter next week? Has to do with phasing of the streams


The numerical models historically have had a very difficult time trying to predict cut-off lows across the Southwestern U.S. that eject and try to phase with northern stream energy.

This is the case now, in regards to next weeks, all important holiday week forecast. The GFS Operational has consistently forecast a strong Nor-Easter with inland snow and a mixture/rain near the coastal areas & major metros in the Northeast.

The GFS Ensembles have trended in favor of the GFS operational, recently. The ECMWF operational is moving a much weaker system more off-shore with less of an impact.  The differences between the two can be seen by examining the upper air charts (500mb) and see how the models handle the northern and southern streams (height/vorticity).

The ensembles will be the most useful guidance over the next day or so as we try to get a handle on the likelihood that the streams will phase and to what extent.  Until then, I would be very leery of anyone telling you they know which solution is correct or what exactly is going to happen. These patterns are very complex and historically very challenging for the models to accurately simulate at these time ranges (4-5 days out).



GFS Ensemble 500mb Height & Vorticity 12z Wed

GFS Ensemble  SFC Pressure & Precip 00Z Thu.






ECMWF 850 00Z THU                                                                                 GFS 850 00Z THU

Notice the large differences. The GFS develops a huge cyclone over the Northeast while the ECMWF does not, because the ECMWF does not completely phase the two streams.

-Mike Dross

11/22/13  -8:30PM


Another Tornado Debris Signature Caught on WSR-88D: Near Raleigh, NC

An apparent brief touch down of a tornado in Central North Carolina yesterday evening was detected by the Raleigh, NC WSR-88D Radar. The RDA was fairly close to the storm with the 0.5 degree beam altitude only around 1390 feet.

While the NWS survey has not yet been completed,  it appears this will very likely be an EF0 or EF1 tornado, based on the damage reported, photos and radar data.

Update 2:10 PM EDT: NWS just confirmed EF0 with 85mph winds.  Path length 1.25 miles / 150 yards wide

The 22:31:59Z volume scan had a small, but defined TVS with gate to gate shear of ~81kts (+57/-24) and a well defined hook echo. It also developed a significant correlation coefficient minimum which was co-located with a reflectivity maximum in the hook where the TVS was also centered.

All of this convinced me, that there was indeed a tornado on the ground and in progress.   I quickly put together a “tweet” and sent it out at 22:36Z indicating that we now had a tornado on the ground.


I also submitted something to the effect,  on RAH NWSChat,  about a possible debris ball developing on radar.

Shortly there after we started receiving verification of damage and sightings of the tornado.


This is another example of how useful the Dual Polarization data can be when looking for Tornado Debris Signatures. Even for locating relatively “weak” tornadoes. Looking at the dual polarization data, it appears the “debris” was lofted to an altitude of around 2600 feet. My thinking is that much of this “debris” is likely tree leaves & foliage from the wooded areas where the tornado crossed, as well as, some other materials.

Click images to enlarge

Next Volume Scan shows "Debris Ball" with reflectivity maxium and correlation coefficient minimum, but velocity couplet weakening indicating tornado dissipating.

Photos of the wall cloud ( Credit: WRAL)

Damage to a home in Youngsville, NC

Updated: Severe Weather Threat Ohio Valley Wednesday-Wednesday Night

Complicated forecast for the Ohio-Valley into West Virginia and Western Pennsylvania, Wednesday into Thursday Morning.  As indicated in last nights discussion, synoptic forcing is very strong and will create favorable conditions for a large area of severe weather as an intense low pressure for this time of year, moves across the Ohio Valley.  The primary forecast problems are the convective mode that the storms take once they develop and whether the downstream ambient environment is contaminated by earlier convection that will disrupt the inflow and stabilize the boundary layer.  This is exactly what happens with the WRF-NMM and WRF-ARW 4KM models, each a little bit differently.  The WRF-4KM NAM CONUS Nest does not, and creates very little convective contamination prior to the main system arriving and develops a very dangerous environment across IL/IN/OH/ Western PA that would support a widespread severe weather event with supercells. While not explicitly, it would likely yield some tornadoes given the intense parameters it is generating.

Below are some animated GIF’s from the WRF-NMM to illustrate the forecast issue of convective contamination of the environment prior to the main forcing/convection related to the surface low.  If this AM convection does not develop across WV/VA as indicated, the stable outflow may not be near as strong and may not provide the cool theta-e values and therefore not weaken the main convective system as much or at all when it arrives late Wednesday night.

WRF-NMM Radar Simulation. Note Convection that develops across WV/VA AM Hours Wed Morning.

WRF-NMM  Note the Theta-e minimum associated with the AM convection across WV/VA mountains that works westward and helps weakens the approaching convective system late Wednesday night as it arrives in OH.

Even as it is,  with some possible convective contamination. The mesocale models are generating some strong indications of a wide spread severe weather event.

Across the Ohio-Valley Wednesday into Wednesday night, the  0-1km helicity and the 0-6km shear will increase as the surface low deepens and the upper level shortwave approaches.  Kinematic forcing will favor the formation of supercells and possibly tornadoes over a fairly large area, but the most favorable location will be just south of the stationary front/warm front which will maximize the 0-1km helicity.

Here is just one of the Updraft Helicity products from the various mesoscale models. This is from the WRF-NMM. It’s developing a long lived Supercell from just west of Chicago and tracking it southeastward. Other supercells develop across Indiana and Ohio. This is one of the longest and most intense updraft helicity values I have seen since observing these parameters.


Forecast Sounding 40 Miles South of Chicago, IL  01Z Wednesday Evening. WRF-4KM NAM Nest

Forecast Sounding near Wooster, OH this evening at 00Z. Very Unstable atmosphere in place ahead of main forcing. Supercells and tornadoes are possible if convection can develop  ahead of main convective system. 

While the threat from widespread damaging wind is certainly high and linear mode convection may very well be the primary mode of convection for the duration of this event. I feel that there very well may be a period of time where there are either discrete supercells or embedded supercells with in a line. Given all the parameters, I believe the tornado risk is quite high along and just south of the stationary front from Central IL through IN into OH.  This is a highly conditional forecast, since convective contamination from earlier storms may affect the environment or if the system becomes a large bow-echo or derecho very early in the evolution, then the tornado risk is greatly reduced.

Here is a map outlining the risk area of tornadoes based on all the available guidance through 05Z.


Severe Weather Likely Wednesday Into Thursday Ohio Valley into the Mid-Atlantic/Far Southern New England

An anomalously strong late spring upper level trough will develop along the Eastern U.S. Coast by Friday.  Two shortwaves, one ejecting out of the southern stream over California, will at least partially phase with a second shortwave that drops southward into the Great Lakes  on Thursday.  These will combine to strengthen a surface low that will traverse from Iowa eastward to the New Jersey Coast.

A two-day event is expected as severe convection will likely generate near the surface low Wednesday afternoon across Illinois and Indiana where the best forcing will exist, however further east across Ohio and West Virginia additional severe storms may develop with the aid of warm air advection and a mid level wind max in the northwest flow, during the afternoon. If discrete convection develops across Ohio/Western West Virginia  on Wednesday afternoon there could be supercells and possibly strong tornadoes, given the favorable shear and other parameters. This is well ahead of the main forcing of the the surface low back to the west.

Wednesday Night
Intense convection is expected to develop across Illinois and Indiana and track eastward during the overnight hours. As the surface low deepens, warm air advection will continue to help destabilize the atmosphere near the warm front which is likely to be near Northern Ohio into Western Pennsylvania . A greater risk of significant tornadoes exist along this boundary. Just north of the warm front elevated convection may produce large hail and damaging winds.

Additional storms may develop further southward into Southern Missouri, Kentucky and Tennessee.

As the upper level shortwaves begin to phase, per the NAM.  The surface low is expected to slow somewhat and should allow for some destabilzation across the eastern portion of Virginia, Maryland, Pennsylvania, New Jersey during the afternoon.

Wind fields will support supercells and possible tornadoes from Central Pennsylvania to  New Jersey southward into the Carolinas .  The best parameters for significant tornadoes on Thursday will exist near the warm front across Pennsylvania, Maryland, Delaware & into New Jersey.

*All of this is highly dependent on surface heating and interaction with prior convective outflow boundaries, none of which the models can properly resolve at this time range. 

Below are some of the products used to create the outlook. 



NAM Forecast Sounding for Columbus, OH Wednseday Evening.

If convection can initiate and overcome the convective inhibition across eastern Ohio Wednesday afternoon, the environmental conditions are favorable for discrete supercells and tornadoes.

Forecast Sounding for Baltimore, MD for Thursday Afternoon at 2PM EDT
NAM Sounding would support severe storms and possible tornadoes.

Severe Weather Outbreak Likely Friday/Friday Night Across Southern Plains / Midwest

Severe Weather Outbreak Likely Friday Afternoon/Night Midwest – Southern Plains

As an upper level low begins to move eastward, the strong core of the mid and upper level winds will rotate around the upper low. An increasingly favorable kinematic environment for supercells and potentially strong tornadoes will develop from Central Illinois/Central Indiana through Southern Missouri into Eastern/Central Oklahoma and  Northern/Western Arkansas.

The area at greatest risk for strong tornadoes is from Central/Eastern Oklahoma through Southern Missouri into Central Illinois.  Although there will likely be AM storms, some severe, the storms that will develop Friday afternoon and move into Ohio-Valley during the overnight hours are the ones that pose the greatest risk for that area. Boundaries from earlier convection in the day will likely focus localized helicities values over 400 m^2^s which will enhance the tornado potential with storms that develop in non-linear modes. Bulk Richardson Shear values favor supercell storm modes across these areas, at least initially.

Other storms will likely form further south across Oklahoma where the instability will become extreme with surface capes over 5000 and shear values in excess of 40kts. CIN will drop below 25 j/kg by afternoon and supercells should form with the aid of an approaching shortwave. With LCL’s AOB 1000 meters, tornadoes are quite likely, especially across Northeastern OK.

The Maximum Updraft Helicity product from the 4km WRF/ARW has been consistently generating strong supercells across OK, MO, IL Friday Afternoon & Evening.

1 Hour Max Updraft Helicity. Indicate Rotating storms develop late.

Max 1 Hour Updraft Helicity from WRF-ARW

WRF-4KM Simulated Radar 23Z Friday

0-6 KM Shear Increase as Mid-Level Jet Works into the Region

High Surface Dewpoints work northward into IL/IN feeding storms.

Surface CAPE is quite high.

Surface Low over MN and Secondary Low over SW OK will help back Low Level  Winds.

Supercell Composite Parameter indicates likelihood of rotating storms from OK northeastward into IL.

Significant Tornado Parameter indicates risk of strong tornadoes.

0-1Km Helicity 00Z

0-1km Helicity 03Z

250mb RRQ of Jet will enhance Vertical Motion

Upper Level Low will move Eastward with cold pool and wind max

St. Louis, MO Forecast Sounding for Friday Evening.  Small Cap, but supportive of Supercells and possible tornadoes

GFS Forecast Sounding for Springfield, IL for 00Z. Again kinematics and thermodynamics supportive of supercells and tornadoes.

Extreme Instability at Oklahoma City per the GFS Forecast Sounding

-Mike Dross

Latest Track Model Guidance. Thoughts on Sandy Landfall Wind Speeds.

Latest 18Z track model guidance continues to remain tightly clustered around a landfall location in New Jersey. The 12Z ECMWF is very consistent with its previous runs and the GFS.  Below is the spread and the mean (consensus) forecast tracks from the 12Z dynamic models / 18Z statistical models.

The mesoscale models are in good agreement of a low level wind max from 900-800 millibars around 100-110 knts that will move inland as Sandy approaches the coast.  The NAM 4km nest is forecasting surface gusts in the 80 knot range over open water which seems reasonable.  Generally a 90% reduction is used in warm core  eyewall from 850 millibars, to estimate surface wind speeds (10 meters). But since this system will not have very much deep convection, that will likely be too generous. My feeling is that much of the wind at 900mb will remain aloft due to the lack of deep convection, but occasional gusts will make it to the surface in some of the heavier showers that will rotate around the center of the cyclone.  If the models verify with the 850mb wind forecasts then it seems almost certain that surface wind speeds will exceed 70 mph in gusts. Also, based on the NAM 4km it appears  very strong gusts may occur much further north, near Boston. On the order of 60-70 mph near the coast.

Below are two images. The first is the latest track forecast.  The second is the 4km WRF Wind Gust Forecast. Note the fetch of 70+ wind gust from Nantucket to the New Jersey Coast.

Click this link for the 36 hour animation of the surface wind gust forecast




Added 925 Millibar Winds

Added 925 Mb winds to the NAM, NAM CONUS Nest & HWRF Models. They are in the drop down menus. Should be available with the 10-28-12 12Z runs.

Below is the forecast 925mb winds (just above the surface) as Sandy makes landfall. Winds of 95 knots are forecast  by the HWRF near Long Island. Mixing of these winds near surface by heavy rain showers will likely transport gust to 70 knots at times as the strong gradient north of the center rotates through.


GFS Atlantic 850mb vorticity added

The GFS 850mb vorticity was added this past weekend to the GFS Tropical menus. The 850mb vorticity can be a potential early indicator of tropical cyclone genesis.


The 850mb vorticity parameter has already been part of the ECMWF model guidance for a number of years.