17 July 2010 Severe Weather Outbreak

 

This was an extremely challenging severe weather event. SPC had a "See text" with the early afternoon update. The forecast guidance has been of little help this summer, and SPC, our office here at ALB and surrounding offices have had questionable success predicting the mode and extent of convection, with some non-event watches being issued, and then at other times, like Saturday 7/17, convection was downplayed, yet a watch could have been useful before convection fired.  Keep in mind, that on the previous day, Friday 7/16, we were under a slight risk from SPC and we were geared up for a long day/night of severe weather operations with the approach of a cold front.  Very little severe weather happened on Friday.  Yesterday, we were only under a "See text" from SPC, and even when a mesoscale discussion was issued when the first storms began to fire, SPC stated no watch was expected.  A later mesoscale discussion mentioned between 2-4 pm a Watch was possible, and a portion of upstate NY and New England would be added to a SLIGHT RISK. At this time, the convection was isolated to scattered mainly in the BGM forecast area.

 

The convection initiated over central NY was associated with a weak surface trough (or possibly even a lake breeze boundary). The main surface cold front remained north of the Saint Lawrence River Valley across southern Ontario and Quebec. A 500 hPa closed upper low was situated just S/SE of James Bay. Cyclonic flow persisted aloft with a number of short waves passing across the region. The forecast thinking was for some scattered showers and thunderstorms with a surface cold frontal passage overnight. An isolated severe event may be possible, and the afternoon forecast had small hail and gusty winds in the grids/zones.

 

The limiting factor for potential convection was a mid level cap in the 600-700 hPa layer. The 12Z observational sounding at KALB showed this cap, and the model soundings. It appeared convective activity would be limited or isolated at best. The initial convection in the early to mid afternoon eroded this cap, and a couple clusters of convection moved across the eastern half of BGM's CWA into the western Mohawk Valley and northern Catskills. One cell produced wind damage in Jefferson in Schoharie county and 1" hail in Ashland in western Greene County. This was an impressive cell, as it even produced a funnel cloud in Delaware County. This cell weakened, as it moved across the Catskills, but reintensified as it moved across the Hudson River Valley into the Taconics of southern Columbia County. On the radar, these initial cluster of cells put out several outflow boundaries that new convection developed off. The structure of the southern Columbia, northern Litchfield, and southern Berkshire cell was impressive in the vertical.

 

The forecast staff thought the convection would wane as we approached 23Z. SPC mentioned no box was likely due to limited coverage (small area), and isolated severe weather threat. They felt the limited severe threat would diminish around 02Z. Cell regeneration kept occurring near the western Mohawk Valley, with southern Herkimer, and western Fulton counties getting pounded by training cells. Significant wind damage occurred in southern Herkimer and western Fulton counties. STP radar estimates were 3-4+ inches north of I-90, and south of Hinckley reservoir. The FFG values were only 1.5 inches for 6 hours in southern Herkimer. This prompted a FFW. Later we learned a beaver dam broke flooding portions of state RT 29 in Oppenheim in western Fulton County.

The convection increased in coverage ahead of the slowly descending cold front. The RUC40 showed a mid level short wave in advance of the surface cold front. An impressive cluster of convection moved across the Capital Region between 8-9 pm and caused widespread wind damage in eastern Rensselaer County shortly after 9 pm. There were also large hail reports of greater than 1". The cell bowed out as it went towards VT and and western MA.

 

The convection tended to train a bit with new cells regenerating along an outflow boundary, and the short wave aloft. The cells extended from the northern Catskills, eastward across the Capital Region, Taconics and into the Berkshires between 10 pm and 2 am. The frequency of the cloud to ground lightning was impressive. The lightning stroke density was very high, with up to 2000+ strikes in an hour in and near the ALY CWA.

 

Another impressive bow echo formed west of Schoharie and Greene counties. Damage reports were minimal likely due to the sparsely population of the area. Larger polygons were used to cover the lines, since multiple areas showed elevated cores, higher VIL's and velocity maxima.

 

We had some clues to the development of strong to potentially severe convection, but the outbreak certainly exceeded expectations, at least here in the ALB forecast area.  The high reflectivity cores and wind signatures were as good as any outbreak we have ever seen.  When the convection formed into a complex during the evening and tracked through the ALB area, there was a lightning display that was probably one of the most impressive lightning displays in many years.

 

Graphics – Model, Ensemble and MOS guidance

 

 

 

Above left:  ECMWF 500 hPa heights and vorticity loop initialized 00Z 17 July and ending 12Z 17 July. 

Above right:  Conceptualized warm season upper cut-off low climatology from CSTAR study.  Note the similarities, however, the ECMWF shows a progressive closed low, trending to an open wave, with an ill-defined vorticity maximum on the southern periphery of the upper low.

 

  

 

Above left:  Four panel of 500 hPa heights and vorticity from the NAM80, GFS, ECMWF and GGEM initialized 12Z 15 July.  There are some indications in all the guidance that the main northern stream upper impulse tracking through southern/southeastern Canada tracks closest to our region Saturday, along with the associated upper dynamics.

Above center:  Same model sources as left except MSLP.  Note the very ill-defined surface features on Friday and Saturday.  However, there is some indication of a surface trough Friday, and a slightly better defined boundary on Saturday, but not in all four models.

Above right:  Same model sources as others, except QPF.  The QPF fields were quite well-defined for Friday afternoon, when very little developed.  There was also an indication of another round of precipitation Saturday afternoon and evening.

 

  

 

Above left:  Four panel of 850 hPa winds and isotachs from the NAM80, GFS, ECMWF and GGEM initialized 12Z 15 July.  All the models showed increasing low-level jet energy, but no well-defined jet core for Friday.  However, the wind fields increased significantly over the region in the NAM and GFS for Saturday afternoon.

Above center:  Same as left, except low level theta-e.  The 925 hPa theta-e showed one ridge and gradient tracking through on Friday, and a lesser-defined theta-e ridge and gradient tracking through Saturday.  The NAM showed the Saturday feature best.

Above right:  Same as others and center except 700-500 hPa lapse rate.  None of the models showed much more than a lapse rate of -6C at any given time, so there were no clear signals for enhanced updrafts in the midlevel lapse rate displays.

 

   

 

Above left:  Four panel of 500 hPa heights and vorticity from the NAM80, GFS, ECMWF and GGEM initialized 00Z 16 July. 

Above left center:  Same as left except MSLP. 

Above right center:  Same as others, except QPF. 

Above right:  Same as others except 700 hPa relative humidity and vertical velocity.

This next run of the guidance did not show much difference from the previous version.  However, there were still hints at more of a surface trough passage Friday, and more of the meaningful cold frontal passage Saturday.  The timing of the surface boundary was slightly different in all the guidance.  The 700 hPa vertical motion and relative humidity did show a more definitive feature Friday afternoon in the GFS, and a hint of an axis of moisture and vertical motion on Saturday afternoon/evening.

 

  

 

Above left:  Four panel of 850 hPa winds and isotachs from the NAM80, GFS, ECMWF and GGEM initialized 00Z 16 July. 

Above center:  Same as left, except low level theta-e. 

Above right:  Same as others except 700-500 hPa lapse rate.

The Saturday afternoon strengthening low-level jet segment became slightly better defined in all the guidance at this point.  The 2 separate theta-e ridges for Friday afternoon/evening and Saturday afternoon/evening, and associated gradients were also slight better defined.  The midlevel lapse rates showed no clear signals.

 

  

 

Above left:  Four panel of 500 hPa heights and vorticity from the NAM80, GFS, ECMWF and GGEM initialized 12Z 16 July. 

Above center:  Same as left except MSLP. 

Above right:  Same as others except QPF.

There is still not much difference in this set of guidance, with all guidance still suggesting a well-defined feature tracking through the region Friday afternoon.  Again, not much of anything happened Friday.  All the models are now suggesting some precipitation Saturday afternoon/evening, though.  However, only the GFS suggests a definitive feature tracking through NY and New England, compared to the other models that suggest the precipitation ends before entering New England.

 

  

 

Above left:  Four panel of 700 hPa relative humidity and vertical velocity from the NAM80, GFS, ECMWF and GGEM initialized 12Z 16 July. 

Above center:  Same as left except 850 hPa winds. 

Above right:  Same as others except low level theta-e.

The 700 hPa relative humidity and vertical motion is less-defined, however, the GFS seems to resolve the Saturday afternoon feature best.  There is a little more model consensus for the strengthening low-level jet energy.  The low-level theta-e ridges and gradients are still a bit nebulous, but can be tracked in each set of guidance, but the Saturday afternoon feature is of varying strengths and timing in each model.

 

  

 

Above left:  Four panel of 500 hPa heights and vorticity from the NAM80, GFS, ECMWF and GGEM initialized 00Z 17 July. 

Above center:  Same as left except MSLP. 

Above right:  Same as others except QPF.

This run of the guidance still did not show clearly-defined features at 500 hPa and the surface.  Although, the mean upper trough axis was still indicated as crossing through the region later Saturday into Sunday, and a more definitive surface wind shift tracking through, with varying opinions on timing of the passage of the wind shift.  The GFS still showed a feature tracking through NY and New England in the QPF field.

 

  

 

Above left:  Four panel of 700 hPa relative humidity and vertical velocity from the NAM80, GFS, ECMWF and GGEM initialized 00Z 17 July. 

Above center:  Same as left except 850 hPa winds. 

Above right:  Same as others except low level theta-e.

The GFS continued to depict the feature at 700 hPa best, although, it seems all models showed a better-defined feature in the 700 relative humidity fields.  All models showed a low-level jet segment tracking through the region, but with various timing, and various southward extent.  The low-level theta-e ridges and gradients seemed less defined on this particular model run, which is confusing because this run of the models was closest to the onset.

 

    

 

Above left:  GFS MOS guidance for Albany, NY initialized 06Z 17 July and 12Z 17 July. 

Above left center:  Same as left except for Glens Falls, NY.

Above center:  same as others except for Bennington, VT.

Above right center:  Same as others except for Poughkeepsie, NY.

Above right:  Same as others except for Rome, NY.

The 6-hour probabilities at all sites were less than 25% at ALB and DDH, which is the threshold for scattered showers and thunderstorms, therefore, MOS guidance was suggesting isolated activity at best, but the timing was centered at 00Z-06Z, which is when the convection occurred in reality.  The 6-hour probabilities were higher at GFL and POU, and were timed differently, so even though the MOS was showing some skill, some of the output sent conflicting messages.

 

   

 

Above left:  GFS MOS guidance for Albany, NY initialized 18Z 17 July and 00Z 18 July. 

Above left center:  Same as left except for Glens fall, NY.

Above right center:  Same as others except for Bennington, VT.

Above right:  Same as others except for Poughkeepsie, NY.

The MOS guidance showed much more skill in the 18Z run, but of course the feature was being resolved better in the data, and was generally underway in central/western NY. 

 

    

 

Above left:  NAM MOS guidance for Albany, NY initialized 00Z 17 July and 12Z 17 July. 

Above left center:  Same as left except for Glens Falls, NY.

Above center:  same as others except for Bennington, VT.

Above right center:  Same as others except for Poughkeepsie, NY.

Above right:  Same as others except for Rome, NY.

Note the 00Z 17 July MOS showed slightly enhanced 6-hour probabilities around 06Z 18 July, but mainly isolated convection.  The 12Z July guidance was much more bullish about probabilities for convection around 06Z, showing more skill.  However, keeping in mind the plan view guidance above all the MOS displays, there seemed to be some conflicts between the ETA MOS and the plan view guidance, with the plan view displays less emphatic about a clearly defined feature tracking from NY through New England.

 

  

 

Above left:  Short Range Ensemble Forecast (SREF) probability for 0.05 inches in 6 hours, initialized 21Z 15 July, valid 18Z 16 July through 00Z 17 July.

Above center:  Same as left except initialized 03Z 16 July, valid 18Z 16 July through 00Z 17 July.

Above right:  Global Ensemble Forecast System (GEFS) initialized 00Z 16 July, valid 18Z 16 July through 00Z 17 July.

Note the ensemble guidance was indicating likely to categorical chances for measurable precipitation on Friday, when very little happened.

 

    

 

Above left:  Short Range Ensemble Forecast (SREF) probability loop for 0.05 inches in 6 hours, initialized 21Z 16 July, valid 18Z 17 July through 06Z 18 July.

Above center:  Same as left except initialized 09Z 17 July, valid 18Z 17 July through 06Z 18 July.

Above right:  Same as other except, initialized 15Z 17 July, valid 18Z 17 July through 06Z 18 July.

Notice the SREF was not indicating much of a chance of 0.05’ of precipitation during any 6-hour period over eastern NY or New England until the 15Z 17 July run, which became available after the event began, and was consequently of little to no help.

 

  

Above left:  Global Ensemble Forecast System (GEFS) probability loop for 0.05 inches in 6 hours, initialized 12Z 16 July, valid 18Z 17 July through 06Z 18 July.

Above center:  Same as left except initialized 00Z 17 July, valid 18Z 17 July through 06Z 18 July.

Above right:  Same as other except, initialized 12Z 17 July, valid 18Z 17 July through 06Z 18 July.

Notice the GEFS consistently showed high probabilities for 0.05” of precipitation over central and eastern NY, perhaps just nudging into western New England.  This guidance showed more skill than the SREF, and much of the MOS guidance.

 

Graphics – Data analysis just prior to convective initiation and during the severe weather event

 

 

 

Above left:  Water Vapor satellite imagery loop from 1130Z through 2300Z 17 July.

Above right:  Visible satellite imagery loop from 1730Z to 2315Z 17 July.

Note the closed low north of the Great Lakes tracking east, and the well-defined upper impulse tracking out of the Great Lakes and Oh Valley.  The convection fired in proximity to this upper impulse. This upper impulse was not resolved well in the NWP models shown previously.

 

 

Above:  Storm Prediction Center severe weather outlook for day 1 issued on Friday 16 July at 1230Z.

Note the slight risk for Friday 16 July, and the reference to relatively well-defined features along a cold front.  Also note (at the bottom) that very little severe weather occurred Friday 16 July.

 

 

Above:  Storm Prediction Center severe weather outlook for day 1 issued at 1619Z (top) and 2000Z (bottom).

The Storm Prediction Center was having difficulty resolving the salient features, as were the NWS Offices across the northeastern U.S., as NY and New England were not added to the slight risk until after the event began.

 

 

 

Above left:  Probabilities for severe hail from the Storm Prediction Center issued at 2000Z.

Above right:  Same as left except for severe winds.

The probabilities for hail and severe winds were difficult to define.  As will be seen later, large hail and damaging winds were observed during this event, even though there were no probabilities for hail indicated.  Temperatures were relatively warm aloft, supporting very limited hail, but still, severe hail was observed.

 

   

 

Above left:  Skew T sounding from Albany, NY at 12Z 17 July. 

Above left center:  Same as left except for Buffalo, NY.

Above center:  same as others except for Pittsburgh, PA.

Above right center:  Same as others except for Detroit, MI.

Above right:  Same as others except for Maniwaki, ON.

Note that the ALB sounding showed a relatively stable atmosphere, and relatively dry.  However, upstream, there was low-level thermal and moisture advection, as well as cooling aloft.  So based on the vertical wind profiles, by evening, central and eastern NY could be expected to be considerably more moist and unstable, supporting any convection that could develop upstream.

 

 

 

 

Above:  Storm Prediction Center mesoscale discussions from 1905Z 17 July and 2243Z 17 July.

Mesoscale discussion were mainly reactionary, based on convection that had already developed.  The convection was expected to occasionally become severe, but coverage was supposed to remain relatively limited, even as it moved into western New England.  Again, the Storm Prediction Center and NWS Offices across the northeastern U.S. could do little more than react to developing conditions, since surface and upper features supporting the convection were so subtle.

 

 

 

Above left:  LAPS surface based CAPE loop from 16Z 17 July through 12Z 18 July.

Above right:  MSAS MSLP and wind barbs loop from 16Z 17 July through 12Z 18 July.

There was sufficient CAPE for strong convection prior to 00Z 17 July, generally >1000 J/KG, but note that after 00Z 18 July there was a widespread area of <500 J/KG over central and eastern NY, where the convection exploded and tracked through the Capital District through about 05Z 18 July.  The low-level focusing mechanism must have been rooted above the surface.  The wind barbs and MSLP did not show much surface convergence, or a well-defined surface boundary.

 

  

 

Above left:  Skew T sounding from Albany, NY at 00Z 18 July. 

Above center:  Same as left except for Buffalo, NY.

Above right:  same as others except for Pittsburgh, PA.

The 00Z ALB sounding showed considerably more moisture and instability than at 12Z, and also showed some semblance of an “Inverted V” sounding, a characteristic found in some research to support damaging winds.  Instability and moisture also remained upstream.  The ALB sounding is considered proximate to the severe convection, since it developed just upstream between 22Z-24Z and tracked through our region during the next 4-5 hours.

 

 

 

Above left:  Storm Prediction Center moisture convergence valid 00Z 18 July.

Above right:  Storm Prediction Center maximum unstable CAPE (MUCAPE) valid 00Z 18 July.

Note the moisture convergence in western NY/PA, and the 2000+ J/KG CAPE over eastern NY and western New England.  The instability was certainly present but the surface forcing was a bit more ambiguous by 00Z, when severe convection was already being observed in central NY.

 

  

 

Above left:  Storm Prediction Center 850 hPa analysis valid 00Z 18 July. 

Above center:  Same as left except 500 hPa.

Above right:  same as others except for 250 hPa.

There was a wind shift at 850 hPa around the Great Lakes, with a low-level jet segment tracking through western NY.  There was also a broad wind shift at 700 hPa within a broad upper trough around the Great Lakes.  Our region was just north of a 250 hPa jet segment, but it did not look like we were in the favorable left entrance region of the upper jet.

 

 

 

Above left:  Storm Prediction Center 850 hPa analysis valid 01Z 18 July. 

Above right:  Same as left except 05Z.

These 850 hPa analyses 4 hour apart also show the subtle and ill-defined nature to the features that were affecting the development, mode and maintenance of the convection that was occurring over central and eastern NY.

 

  

 

Above left:  Storm Prediction Center mid level 700-500 hPa analysis valid 00Z 18 July. 

Above right:  Same as left except 02Z.

Midlevel lapse rates were steepest in western and central PA, into western NY.  However, there was severe convection with significant updrafts in central and eastern NY.

 

 

 

Above left:  Positive and negative lightning strikes for the 5 minute period from 0200Z-0205Z 18 July.

Above right:  Same as above left except for the Capital District of NY, and for the time period 0230Z-0235Z 18 July.

The convection that tracked through the Capital District had one of the most impressive lightning displays in recent memory.  The band of convection between Cooperstown and Schnectady in the image to the left showed 214 cloud-to-ground strikes, not including cloud-to-cloud lightning, in 5 minutes.  That is one cloud-to-ground lightning flash roughly every second.  When the convection tracked across Albany, there were 75 cloud-to-ground lightning flashes, not including cloud-to-cloud flashes, in 5 minutes.  That is one flash roughly every 4 seconds, again not including cloud-to-cloud.

 

 

Above:  Loop of hourly positive and negative lightning strikes from 1800Z 17 July through 1500Z 18 July.

This loop shows how the cluster of severe convection develop and tracked over the course of multiple hours.

 

    

 

Above left:  Radar base reflectivity loop from KENX from 1950Z-2100Z 17 July with severe weather polygons overlayed.

Above left center:  Same as left except 2100Z-2300Z 17 July.

Above center:  Same as left except 2300Z 17 July through 0100Z 18 July.

Above right center:  Same as left except 0100Z-0300Z 18 July.

Above right:  Same as left except 0300Z-0500Z 18 July.

Note the development and movement of the convection, and the progression of severe thunderstorm warnings. 

 

 

 

Above:  Severe weather reports for 16 July from the Storm Prediction Center (left) and from NWS Taunton, MA (right, from Iowa State Environmental Mesonet website).

 

    

 

Above left:  Severe weather reports for 17 July from the Storm Prediction Center.

Above left center:  Severe weather reports from NWS Buffalo, NY (from Iowa State Environmental Mesonet website).

Above center:  Severe weather reports from NWS Binghamton, NY (from Iowa State Environmental Mesonet website).

Above right center:  Severe weather reports from NWS Albany, NY (from Iowa State Environmental Mesonet website).

Above right:  Severe weather reports from NWS Taunton, MA (from Iowa State Environmental Mesonet website).

 

Note the large number of severe weather reports from the 17 July outbreak that was very much under forecasted.  The only conclusion that can be drawn from looking at the guidance and data from this event is that there are still many small scale and convective scale processes that cannot be resolved in the data network and the guidance.  Sometimes we forecasters have to react to changing conditions in the data, and learn to recognize trends in data to improve 6-12 hour predictions of convection.  Guidance is improving, but we forecasters have to learn what subtle signals in model/ensemble output on which we need to focus 12 hours or more prior to a potential event, and there were some signals in the guidance 12-36 hours before convective initiation that provided clues to the Saturday 17 July outbreak.  It is difficult for humans to overcome biases such as when a consistently advertised potential severe weather day associated with a cold front the day before, on Friday 16 July, does not materialize, but we have to learn to take one day and one atmospheric feature/system at a time, and give each one equal attention.

 

Any comments, questions or suggestions, please mail to:Neil.Stuart@noaa.gov