2000 EST 27 FEB 2012….. STARDATE 201202.27
Exactly one week ago — FEB 19– eastern Kentucky the southern half a West Virginia and Virginia experienced a significant snowstorm. I have had a week now to think about this event and to compile all of my forecasts… statements and present a thorough review. Thus this web page entry.
In addition I will explain why I argued so strongly that this was going to be a KY WVA VA snowstorm and that the system had little chance of turning up the coast into the big cities of the northeast. As we go through this analysis I will stress several key points.
- The pattern strongly favored this Low tracking ENE off the NC coast and not coming up the coast. Unfortunately many Meteorologists and weather Hobbyists along the East coast were ignorant of this and blindly clung to the old saying
” the north trend “.
- The GFS Model performance sucked moose. The Model performance was appallingly bad in the period from day 10/ FEB 10 to Day2/FEB 17 with frequent wild swings from solution to solution.
- The close adherence of this seriously flawed Model by weather hobbyist and professional meteorologist caused unnecessary Hedging and uncertainty in the forecast.
The first statement was issued by me based upon the Day 10 12z European model February 10 run…. which showed a significant Low pressure area along the just off the southeast coast of North Carolina. It turns out that the day 10 European model from FEB 10 was amazingly accurate and as close to a direct hit with respect to the actual surface on February 19 for a ten day forecast as you have. Of course the FEB 10 European model was off by one day but still the day 10 European model depiction was pretty darn good.
FRIDAY FEB 10 235 PM I posted this:
* ALERT DAY 10 EUROPEAN MODEL HAS MAJOR EAST COAST SNOWSTORM … even for most of NC and all of MD and VA on FEB 20
Of course several times this winter we seen false alerts from weather models especially at Day 8 Day 10 and Day 12. However in this case something was different with respect to the large scale or synoptic pattern. The PRIMARY reason WHY I believe this event had a better chance to become a reality and a significant event … even at Day 10… was the collapse that was forecasted by all the models of the super intense and large Polar Vortex (PV) over Eastern Canada.
I have said this many times before and I am sure I will say again. The overwhelming amount of data shows that having a large powerful intense PV over eastern Canada— over Hudson’s Bay or central or Northern Quebec / or Labrador–is an extremely negative or Hostile synoptic feature to development of major East Coast Winter storms. Having the PV below 500 dm in that geographical position is a manifestation of the + AO pattern. That being said … if one looks at the 500 mb maps in January and early February you would have seen a large intense PV situated consistently over northern Hudson’s bay or Northern Quebec or Baffin island. Sometimes this feature has been as low as 474 dm and has been so large that it has 6 or 8 lines — height lines –around it. That represents a lot of wind and energy and becomes extremely hard to get the Polar Jet (PJ) to dive south and merge or “phase” with the subtropical jet (STJ) over the Eastern CONUS. And without the phasing… you don’t get big winter East Coast storms.
In addition on the 10th I argued that another key aspect to this POSSIBLE event happening was the cold front and significant Low pressure area tracking through the Great Lakes and February 17. In order for this possible winter storm scenario to develop the Great Lakes Low had to move into southeastern Canada and become a 50 /50 Low. This in turn would force the cold air south into the Northeast US and prevent the next system and tracking up into the Ohio Valley.
The 12z FEB 10 european had some partial support the next morning from the 6z GFS but the model truncation after day 8/192 hour clearly caused the flawed GFS Model to “lose ” the southern system (the short wave) at the 500 MB.
The 12z FEB 11 operational European model had the southeast Low but it was further south. The 12z euro ensemble had the Low 200 miles further North and this trend continued with the 0z FEB 12 run of the Operational European… and I posted this on the Facebook (FB) page at 0155 hrs 12FEB
*** ALERT *** ABOUT FEB 19-20 THREAT ****
The NEW 0Z — early Sunday AM – run of the ECMWF ( the european) shows a CLASSIC VA -NC -MD SNOWSTORM … but only IF one looks at UPPER AIR maps . But the surface map does NOT look all that good. At 500mb on day 8 – 7PM SAT FEB 18 the 0z euro has closed 500 Low centered over central ILL -IND…. the surface Low for some reason is waaaay to the south which makes NO sense . The Model has the surface Low on the SC coast which is a tad further North than the 12z Saturday run. BY the Morning of FEB 19 the closed 500 low is right over southern VA … and there is some snow falling over central and eastern VA…. but again the Surface Low is waaaaaaay to the east which makes NO sense.
On the other hand the 0z GFS took the Low into the Great Lakes and developed a strong SE ridge.
However the 12z GFS FEB 12 run showed a Major change or “flip flop” in that the Model took the Low from MI to the GA coast and showed the southern energy — the short wave– weakening rapidly as it propagated east along the Gulf coast.
The 12z Euro run on 12 FEB remained very consistent. The model detected 2 pieces of energy — 2 short waves - over TX and KS – that merged or Phased into a stronger southern short wave that developed a Negative tilt. This in turn argued that the surface Low should be closer to the coast than what the Model was depicting with its surface map for FEB 19.
A SIGNIFICANT development occurred on the 0z FEB 13 Model runs when the CMC or Canadian model which had been showing no development of the southern shortwave for the last several model runs… turned it towards the European solution by developing a significant area of Low pressure along the Southeast US coast that bought significant snow into the southern half of Virginia. Such a development of course began speculation about the new north trend as it almost always does but as we will see in this case that was not to be for number of different reasons. On MONDAY FEB 13 at 0121 hrs I posted this …
0Z CMC MAPS FOR ALL TO SEE.. SIGNIFICANT SNOWFALL for southern VA standards NOT for say NYC or PHL or BOS standards into southern 25% of VA. This is BIG leap in what THIS Model was showing when compared to earlier runs…
The trend continued with the 0z 13 FEB run of the European model which showed it even bigger and more developed southeast US coastal Low tracking the low over or close to Cape Hatteras. However the model began to experience problems with the northern and the precipitation shield.
Later that day the the operational run of the 12z FEB 13 run of GFS Model came out with a dramatically different solution which caused a significant reaction with some of the more poorly informed meteorologists. The GFS showed a major jump northward in its position of the southeast low so that it bought significant snow into the big cities of the northeast. Upon seeing this run I posted the following on the FB page
* COMMENTS ON 12Z MONDAY 13 FEB GFS run — REGARDING FEB 19-20 ** Ignore this run of the GFS ..it has huge error in it that Just forces the model to go off the reservation into idiot land.
This was followed by the assessment from HPC/ NWS the declared the 12z 13 FEB operational GFS to be seriously flawed bcause of corrupted data the getting into the model .
Of course that drew a very strong reaction from several sources both from whether hobbyists and from professional meteorologist who should have known better. The statement from HPC was pretty strong and I don’t see how anybody could have possibly ignored their argument.
Leading the charge that the GFS solution was correct after all was of course Joe Bastardi. He made the argument that this dropping the operational GFS and its north trend was the same sort of mistake that HPC made with regard to the East Coast historic snowstorm 25-26 DEC 2010. If you recall the particulars of that event … it brought 10 to 20 inches of snow from eastern NC through Norfolk across the lower Maryland eastern shore into the big cities of Philadelphia New Jersey and New York City and then into Southern New England.
Bastardi argument was of course absurd and had little basis in reality but he counts on weather hobbyists not having a firm grasp of what the actual facts were in a particular situation. The truth is that in the days leading up to the 25-26 DEC 2010 historic snowstorm ALL of the models had shown the same solution for several days. This was not a case of one model showing ABC while another model showed XYZ and another showed HIJ. But more importantly on the day before the actual 25 DEC 2010 Major East Coast Snowstorm … all the model data — the European … the Canadian ….the British AND the GFS ….and ALL of the short range models for that matter as well — all took the potential coastal snowstorm way out to sea . It was NOT the case of a ONE model coming up with a bizarre solution and being dropped by HPC.
In addition his analogy also fails because just as all the models both the global and shore range models suddenly took the coastal storm out to sea on the December 24 model runs… that Night on Christmas Eve …ALL of the models suddenly reversed course and bought and be major snowstorm for the immediate coastal areas from Virginia to Boston.
Never let facts get in the way of good Bullshit.
Fortunately even while the debate regarding the obviously flawed 12z GFS Model was raging on midday on the 13th of February the 12z Ukmet came out and showed a Major se US Low tracking over Cape Hatteras and moving out to sea bringing a significant snowstorm to much of western NC …VA … and southern Maryland.
1239PM FEB 13 I made this post on the FB page
**ALERT *** ALERT *** 12Z UKMET GOES BOOM.. shows major east coast snowstorm FEB 19…. for entire Middle Atlantic even VA and MD into NYC LI and southern CT.
This was followed shortly by the 12z run of the European model which developed the southeast low a little deeper and took it further to the north ride along the South Carolina North Carolina coastal areas with the track running from Charleston to Wilmington to Cape Hatteras. This cause me to make the following post on the FB page.
FEB 13 132 PM *** ALERT *** ALERT ***ALERT …. this is No drill … 12Z EURO brings in MUCH stronger SE US Low with HEAVY SNOW to all of central and eastern VA –over 6 inchesand much of Northern NC …. POSSIBLE ACCUMULATING SNOW late on the 18th into the 19th… WOOF WOOF BARK WOOF…
At this Point I tried to make the argument that the overall pattern clearly did not support a north trend and that the system was not a threat for the big cities the northeast. Despite the strongly worded statement from HPC and the fact that the 12z European Canadian and British models all strongly disagreed with the flawed 12z 13 FEB GFS run… there continue to be a lot of desperate hope and the strongly worded assertions that the 12z GFS was somehow … that HPC was wrong .. had to be at least take into consideration .
However there was another way to of solve this dilemma as to which model solution was valid and which one was not. That would be to look at the historical analogs and finding similar events as well as using your brain to understand the potential scenarios that a forecast of may consider.
There are specific reasons or mechanisms in the atmosphere that determine WHY Low pressure areas in the winter months that form in the Gulf will turn up the coast and why some do Lows are forced to turn out to sea near or to the south of Cape Hatteras. This is a science and it is not supposed to be based on whimsical Nonsense. These 3 particular instances or analogs of significant snowstorms for the lower Middle Atlantic region highlight the reasons and the mechanisms as to WHY Low pressure along the southeast US coast might NOT come up the coast.
However if you are unfamiliar with these events and you do not understand the significance of the players on the field especially at 500 MB then you would mis-read the Models and end at making a bad forecast period
3 CLASSIC LOWER MIDDLE ATLANTIC — NC VA southern MD – SNOWSTORMS OF THE PAST.
case #1 30 JAN 1995…
This first event from 30 JAN 1995… is a very close match to the event of February 19, 2012. This first image shows the upper air map on the 30 for January, 1995. As you can see there is an ULL — Upper Level Low — over the Tennessee Valley in a manner that was very similar to the event of 19 FEB 2012. But in addition there was a strong PV – polar vortex- located over Eastern Canada. The PV in the Middle of JAN 1995 was a very intense large PV… in a manner of very similar to what we saw in the beginning of February 2012.
The SURFACE Map shows the se US Low tracking along with just off the Southeast U.S. Coast very close to Cape Hatteras. These
precipitation maps from 29-30 JAN 1995… and 30-31 JAN 1995 showed that significant snow did fall over much of Eastern Kentucky the southern portions of West Virginia and much of southwestern and Central Virginia. Even more significant is that in this situation just like in 19 FEB 2012…… the heaviest snows fell over sw VA and the southern Piedmont.
Notice here that he northern jet stream does NOT drop or phase into the southern jet stream. Keep in mind that it is the phasing of the two jet streams between the Mississippi River and the East Coast which is what develops the coastal Low. If the two streams do not phase there is no MECHANISM for the southern short wave or Low to come up the coast.
CASE #2 2-3 JAN 2002…
This is also a very close match to the situation of 19 FEB 2012…. Because the winter of 2001-02 has been very similar in terms of
the overall temperatures and lack of snow over the eastern half of the CONUS…. And because the winter of 2001-02 was forecasted to be a fairly cold and stormy winter over the eastern CONUS just like 2011-12.
Here is the SURFACE MAP from 2 JAN 2002 and it seems to be a near ideal setup for significant snowstorm into the big cities of the northeast US.
As you can see from these upper air maps JAN 1-2-3 , 2002… there is a large and very pronounced ULL that was centered over Iowa and Missouri on January 1. This feature dropped into the Deep South and then rotated off the Southeast U.S. Coast by January 3. Note again that there was a very deep and very intense PV — Polar Vortex – at 480 dm or lower located over Quebec Canada. Also … note the very strong Pacific trough crashing into the West Coast JAN 1-2, 2002… which again bears a striking resemblance to the 500 MB FEB 17-19, 2012. The analog here is striking we close to what we saw last weekend over TN KY NV and VA.
Notice here that AGAIN the northern jet stream does NOT drop or phase into the southern jet stream. Keep in mind that it is the phasing of the two jet streams between the Mississippi River and the East Coast which is what develops the coastal Low. If the two
streams do not phase there is no MECHANISM for the southern short wave or Low to come up the coast.
Its Just that simple.
And by the way …there was widespread speculation and forecast hoopla that this Low on the se US Coast 2 JAN 2002 was also going to turn up the coast. And finally here is the snowfall accumulation map. In this instance the best snows were NOT over sw VA and the
southern Piedmont but over central and southeastern Virginia. However the principle and the upper maps remain quite relevant and analogous to 19 FEB 2012.
CASE # 3… 25-27 MARCH 1971…
Again the surface maps appeared to be quite favorable for significant East Coast snowstorm. There is a large arctic HIGH centered over the Great Lakes covering all of the eastern CONUS and a stationary front over Florida and the gulf coast into Southern Texas with significant precipitation developing on the front. However by 17 March 1971 the significant surface low pressure area moves off the North Carolina Coast and out to sea by passing the big cities of the Northeast. WHY did this Low also ”miss” the big cities of the NE and why heavy snow ONLY fall over eastern KY southern WVA northern NC and the southern half of Virginia?
The 500 MB maps clearly show what happened. In this case we have a very nice 50/50 Low centered over Newfoundland Canada and this feature is a strong signal for potentially East Coast snowstorm . We also have a very pronounced -NAO as well . And the PV is not very strong and is displaced over North Central Canada. On 25 MARCH 1971 – the 1st map– we see a Upper Low Located over KS and the PJ –polar Jet – is located along or just to the north of the was Canada border and drops into New England… while the Southern jet stream can be found running from California to Texas and Oklahoma then into the southeastern states.
On 26 MARCH 1971… the weak disturbance in Kansas with the Upper Low has developed significantly and is now showing in negative tilt in the short wave trough — SEE BLACK LINE– running from eastern KY to the Florida panhandle. Notice here that AGAIN the northern jet stream does not drop or phase into the southern jet stream. By 27 MARCH 1971 the southern short wave or Low has moved off the Southeast U.S. Coast into the southwest portions of the north Atlantic.
And finally the snowfall map for this event in late March 1971 showed a large area of 8 to 14″ of snow across much southwest … Central … South central and into Eastern Virginia. Less than 1 inch of snow fell in Washington, DC and 4″ of snow fell in Norfolk and Virginia Beach and 7″ fell in Williamsburg.
Significantly after the 12z 13 FEB operational GFS screw up… the 12z GFS ensemble on 13FEB was much closer towards the European Canadian and British models scenarios . Indeed the GFS ensemble mean was also focusing on the rapid propagation or eastward movement of an next incoming trough into the West Coast and towards the Rockies.
The 0z 14 FEB Global model runs became a significant turning point in with respect to the probability of a significant snowstorm for the lower Middle Atlantic region . The operational GFS showed a much more organized area of low pressure coming further north along the southeast US coast and had snow/ rain up to the North Carolina Virginia State line for the first time. This was followed by the 0z run of CMC ( the Canadian) which showed a major snowstorm for the southern half of Virginia and southern West Virginia. The 0z European Model run showed perhaps the strongest surface Low yet with significant snows far north as Philadelphia and Southern New Jersey as the 0z European Model has the Low coming as far North as 37 degrees N latitude before being turned out to sea.
However the 12z 14 FEB European model shifted the Low to the south again. At this Point I began to focus more heavily as to whether not there was any chance for phasing to occur between the two jet streams.
All winter long the pattern has featured a very strong and progressive or fast flowing Pacific jet which has not really relaxed long enough to allow for any sort of serious cold air into reasons coming south from Canada. Very fast progress of flows are resistant to phasing between the southern and the polar jet. I stated as such on the FB page
FEB 14 144PM..
I have for the last few days of favored the flatter solution as a general overall outlook be cause of this particular winter pattern. The last two winters …if you farted too loud you get a snowstorm the East Coast because the pattern allowed phasing… in fact it forced a lot of phasing or merging between the two jet streams. This result in a lot of big East Coast winter storms.
This winter as we all know has been a very mild winter because the mean pattern or flow has been coming in from the Pacific not from Canada and the arctic regions. Thus with THIS possible event … I suspect that constant Pacific Jet stream energy will keep coming in from the West Coast pushing the Southern System on an easterly or eastern northeasterly track and not allowing it to turn up the coast.
The critical issue is whether or not the strong piece of energy in the southern jet stream over Alabama Mississippi mergers or phases with the northern jet stream. If it does…IF… that would pull the system further north and Northern Virginia Maryland into fairly and New Jersey would see significant snow. IF it does NOT phase then this would favor a flatter solution increase the snow threat for central and sw VA into Northern NC
ON 14 FEB I Posted a NOTE on the FB page “WHERE WE STAND ON THIS POSSIBLE THREAT FEB 19-20”.
Withe respect to the debate as to whether not the operational run of the 12z 14FEB GFS was valid or BS… recall this run of the GFS has the Low coming up the coast and dropping heavy snow from DCA to BOS … was clearly proven by the 0z 15 FEB operational GFS which did a another flip flop and took the Low back to the south… along the Southeast U.S. Coast over Cape Hatteras and out to sea.
This new flip flop by the operational GFS model cause me to make the following commentin disgust on the FB page…
This is why I HATE the GFS past 72 hours when it comes to East Coast winter storms. The 0z GFS model run was a massive …severe unprecedented shift… taking the coastal Low from the Georgia Coast … suddenly 300 Miles to the nw over ATL… From passing south and east of Cape Hatteras to passing west of Norfolk … from showing NO precipitation of any kind north Washington, DC
…to driving heavy snow and rain into New York City Albany and will Willaimsport PA. The fact that you may be living in the northeast big cities and the 0Z GFS is the picture perfect and one that you want to see for your location does it mean that the solution is a viable.
The proof ? …all the OTHER weather models… ALL OF THEM…. did not show ANY significant change in their track or forecast and the updated 6z GFS has gone back to the southeast U.S. coastal solution.
At this point in my statements to my clients …I again stated that for Central Virginia and the Richmond metro area this was probably going to be a rain event for at least a portion of the event but I could not categorically rule out the threat of snow if the trend for colder temperatures continued. And for southwest Virginia up through the Shenandoah this look like a pretty significant snowstorm . One of these statements was read on air during the morning news show on WRVA1140 AM by Jimmy Barrett. This prompted one of the on air meteorologists in the Richmond metro area to comment that ”despite some rumors of snow you may have heard about for Sunday the precipitation will be all rain”
The 0z 15 FEB run of the GFS… Amazingly enough had a another massive model flip flop and reversal… and now took the southeast coastal Low well inland of New York City and Philly which of course forced the model to produce widespread significant rain. ON the FB page I Posted the following
FEB 15 117AM …. ** ALERT **8 ALERT ** IGNORE THE 12Z GFS RUN … AGAIN ** its way inland andhas rain into NYC…. c onsider the GFS last run 4 runs… 12z TUES flat Low moves off NC coast out to sea… 18z TUESDAY a tad further North… 0z WAY NORTH low tracks over ATL then west of Norfolk… 6z Flat wave off the NC coast …12z way inland and north!.
After seeing the 12Z runs of the European …British and ….Canadian models I finally committed to this publicly on the FB page at 126pm 15 FEB
GOOD SIZE SNOWSTORM LOOKING MORE and MORE LIKELY NORTHERN VA … NW VA Northern half of SHENANDOAH VALLEY MD DE south NJ
This commitment to forecasting a snowstorm for the lower Middle Atlantic was rewarded by it more changes in some of the models late on the 15th of FEB… the 0z 16 FEB Model runs. As I have said many …many times… the operational GFS is a awful model when itcomes to forecasting East Coast winter storms. It has a lot of trouble handling Southern short waves… it has a lot of trouble Phasing the 2 jet streams just to name a few. But once you move within 84 to 72 hours of the event the GFS was a much better model.
Sure enough at exactly 84 hours before the start of the event on12z 19 FEB…. the new 0z 15 FEB GFS Model showed a dramatic turn towards the European solution.
at 1108pm FEB 15 i made the following post on the FB page
**ALERT *** 0Z GFS COMING IN VERY DIFFERENT FROM THE 12Z GFS RUN and much more like the ECMWF / UKMET
This was supported by the Canadian model 0z FEB 15 run and like the European and the new GFS and Ukmet the Canadian model showed no significant snow north of the Washington, DC area.
The model trend continue to show the southeast U.S. Low moving over or just south of Cape Hatteras and out to sea and each one was coming in slightly colder than the previous run on all the models. After posting the 12z FEB 16 GFS model forecast SNOW map on the FB page I wrote the following on the FB page ( which should of been under NEW NOTE or posted here on the web site )
16FEB 1205PM … ** 12Z MODEL DISCUSSION 16 FEB MODEL…. VALID FOR FEB 19 EVENT*** and in this thread I stated that so far No model was bringing the Low north or heavy snow into any of the big cities of the northeast. In addition it looked to me that this was going to be a sw VA… eastern KY southern WVA big snow… And that the rain could turn to snow for several hours in Richmond but the lack of cold air was going to be a problem east of the Blue Ridge mountains.
Shortly after I wrote that the new 12z 16 FEB European model came in dramatically colder even in central and Eastern Virginia. The European model snow map showed a large area of 40 inch snows even through the Richmond metro area.
131PM FEB 16 ** ALERT *** ALERT ** 12Z EURO BRING IN COLDER AIR TO RIC CHANGES RAIN TO SIGNIFICANT SNOWFALL SUNDAY AFTER 1PM then moderate to heavy snow until late sunday evening … MODEL says 4 to 8 inches of snow Roanoke Charlottesville Lynchburg Richmond Fredericksburg BUT … BUT that is assuming a 10:1 ratio…. if this happens the snow ratio will wont be 10:1 around RIC
At this Point I decide to as time the issue a 1ST GUESS snowfall forecast map at 610pm 16 FEB .
As we moved into the short rain model timeframe snow mounts on many the models began to increase especially east of the Shenandoah Valley into the Piedmont and Central Virginia as the data turn colder and colder with each one. This was especially true on the 0z 6z and 12z Model runs of the NAM and GFS. It was at this point that the short range models began to show significant snow developing Sunday afternoon across Central Virginia and continuing to 1:00 AM on the morning of the 20th.
Significant rain events changing over snow with significant accumulations are not likely in Central Virginia but has been known to happen. But more importantly the reason for the colder temperature profiles on the shore range models was because the precipitation and the system coming out of the gulf was moving slower . This in turn allowed for colder air to move southward lay on the 18th into the morning of the 19th.
On Friday morning February 17 I did a live on air into view with Jimmy Barret 1140am WRVA. My appearance followed immediately after the morning news broadcast which featured the some clown from the Weather Channel named Ray Stagick. I have no idea if this guy is a meteorologist are not but if he is… is he is a pretty awful one. TWC committed to inch of snow or less at the at the end of the event for the Richmond metro area. As you can hear from the broadcast..I went for 2-4 from the city East and 4 to 8 west and sw and nw of the city and 10″ in sw VA. Still I suppose I should not complain about the weather channel too muchsince they make me look really good.
During the morning hours on 15 FEB all of the Local TV stations the their forecasted snow amounts for Central Virginia in the Richmond metro area to 1-2” while TV 12 continue to show less than 1 inch. All the TV stations to agreed that western and southwestern Virginia would see several inches of snow and the media stations in the Charlottesville and southwest Virginia / Roanoke markets played up the threat of significant accumulating snow. Unbelievably even as NWSFO AKQ began to mention the possibility of accumulating snow TWC –The Weather Channel- still refuse to discuss any other outcome except for a rain event with a brief period as snow at the end across central Virginia.
FIRST CALL MAP was issued late in the day on the 17th.
It was based on the short range data which showed a longer delay with the arrival of the precipitation. Most of the other forecasts and model data were premised on the idea that the precipitation arriving Saturday night before the cold air arrived. However the shore range models continue to emphasize a delay with each passing run so that most the precipitation would not arrive until Sunday dawned over the Virginia North Carolina State line and rapidly expanding northward towards Maryland.
However this delay was significant because it allowed the reinforcing shot of cold air to sweep into PA MD WVA and VA. It was not that the air mass was dramatically colder but the due points were significantly drier. The temperature profiles began to show that the precipitation might start as rain or as an Exe but a larger and larger portion of this event over Central Virginia was going to fall as snow and NOT rain. And over southwest Virginia the shore range models are really began to crank out some impressive snow totals.
Winter storm watches and warnings were issued by NWS for all of southwest Virginia the southern third of West Virginia and Eastern Kentucky.
LAST CALL MAP issued Saturday 18 FEB 7PM EST.
This was where I made my first serious mistake. During the last 2 runs of the European model the model snow algorithm showed significant snowfall — 8-12” in west to east Band stretching from southwest Virginia/ Roanoke east across the southern Piedmont stopping just to the west of Richmond. Even though most of the other short range models on Saturday did not show Richmond getting over 8 inches of snow I decided to gamble as a war and extend the 8 to 12 inch band of snow into the southern Piedmont and fairly close to the western side of the Richmond metro area.
This was a mistake. I thought that the model is increasing the snow fall accumulations because the model has detected the colder drier air and the later start time which meant that more the precipitation would actually fall was snow and not mix precipitation or rain changing to snow.
During the afternoon winter storm watch is were issued for all of Central Virginia for 3 to 6 inch snow amounts.
Lastly… Surely after I issued the LAST CALL SNOW FORECAST map the last row of the shore range models came out and a dramatically reduce the northward extent of the precipitation and the amount of significant precipitation over central… southwest Virginia and Southern Virginia. This causes a reaction from several uneducated weather hobbyists – called weather weenies—that of course posted the radar trends did not support significant snow getting into central and southwestern Virginia .
AS we now know the 0z FEB 19 shore range model runs were all completely mistaken. However the shift was so complete on all the shore range models that I decided to issue a reduced snowfall map. Even if the 0z FEB 19 shore range models were wrong it was clear that was not going to be a band of 8-12” snows from Roanoke extending into going through the southern Piedmont almost to the western side of the Richmond metro area.
I believe that I quickly deduce the potential for this event earlier than anybody else based upon my knowledge and study of the differences between East Coast snowstorms and lower Middle Atlantic snowstorms.
I recognize the progress nature of the pattern which would force the system to slide off the North Carolina Coast.
My FIRST CALL MAP was best and my forecast on WRVA 1140 was spot on.
One other note… I did we see several emails from folks complaining that the actual website had not been updated and appeared to be completely unaware that the updates… were being posted directly at the FB page. I will have to make efforts to ensure that annuity is aware of this feature.