www.roanoke.com

In Friday's Weather Journal column, looking at Thursday's winds but primarily reviewing the Feb. 10, 2008, windstorm, I only got to use a very small portion of my Q&A with Stephen Keighton and Phillip Manuel of the National Weather Service in Blacksburg. If you are interested, I have posted the entire e-mail conversation we had regarding the Feb. 10, 2008, windstorm in the extended entry below (click on "Read More" if you are in full blog mode).

Myatt: (1) Describe the causes of the Feb. 10 wind storm ... particulary why this specific event exceeded other high wind events that follow cold fronts?

Keighton: In terms of the meteorology, this particular case was extreme largely because all the pieces happened to come together just right. We frequently get gusty  and sometimes isolated damaging winds behind many cold fronts every single year, but the ones that produce widespread damage across the mountains and sometimes extending into the Piedmont (like the Feb 10 2008 event did) require a special set of circumstances. First, you need a very strong gradient producing anomalously strong winds a few thousand feet above the surface, and in this case for quite some distance aloft (as much as 20,000 feet they were much stronger than normal).  I've attached an image (from a Penn State web site) that shows in the shaded colors how many standard deviations from normal the wind magnitudes were at various levels (in this case the standard pressure levels), and you can see that is not so much at 850mb (~3,000 MSL) but more at 700 and 500 mb (10-20kft).  Next, it helps to have the wind direction just above the ridges perpendicular to those ridges, helping to create wave action. But it really helps to also have a stable layer or inversion dropping to just above mtn top level, which can often produce breaking or crashing waves. The stable layer will drop if there is strong sinking (subsidence) behind the front, so that is another key ingredient that was present on this day. If you then match up the timing of the strongest winds aloft with daytime heating, you can promote more effective mixing of these stronger winds down toward the ground as wind gusts. In the case of Feb 10, we had all these pieces come together at one time or another. The daytime mixing and the extreme wind speeds aloft were probably the most important factors in causing the widespread wind damage all the way out into the Piedmont. We've had other similar events extend damaging winds in some parts of then Piedmont, but this is the first event we have documented since the office opened in 1994 where every single county in our 40-county forecast area had wind damage or measured wind gusts over 50 kts (roughly 60 mph).
  

Manuel: First keep in mind the antecedent conditions needed for extreme fire behavior...dry fuels, low humidity, strong winds, and warm temperatures. If these 4 elements occur at the same time (the swiss cheese effect, all the holes line up, thresholds are met) then you can drive extreme fire behavior, a fire that cannot be controlled.  Wind is just part of the equation, a big part, but if the ground is already wet from rain/snow that occurred with the frontal passage then the wind can blow all it wants to and will not result in the extreme fire behavior.  In the case of February 10th, it was a dry frontal passage.  So what do we look for in determining whether a particular day may be conducive for extreme fire behavior?

Thresholds that we like to look at...

1. Sustained winds at or above 20 mph or frequent gusts above 25 mph.
2. Relative humidity below 30 percent. 
3. Temperatures above normal (preheats the fuels and dries them out more)
4. Dry fuels (fuels are dead grass, leaves, brush).  Occurs when you go 10 days or more without significant (quarter inch or more) of precipitation.  Occurs most often during a drought.

Line all of these things up and you have the proverbial "powder keg."  Once you have the powder keg, you need an initiation source... no spark no fire.  In the western United States this usually comes from lightning.  Here in the east, it is usually human.  In the case of February 10th, it was downed power lines...something that is usually not achieved until you exceed 35 mph sustained and gusts strong enough to down trees... gusts of say 58 mph or more (our severe thunderstorm criteria). 

For February 10th, we had a case of mass initiation starts over a large geographic area, Virginia and North Carolina.  Antecedent conditions were such that the powder keg was ready to blow.   No rain (dry fuels, in the midst of a drought), low RH (17 to 30 percent), and unusually strong wind for a long duration (sustained 30 to 47 mph, gusts between 60 and 70 mph for a 6-8 hour period).

 For a typical frontal passage in the Mid Atlantic during the winter and spring, the front is usually accompanied by some form of precipitation.  Winds may gust to 40 or 50 mph if the front is strong, but like I said before, the wind can blow all it wants too if the ground is wet.    The front that came through on February 10th was dry.  It was an unusually strong dry cold front and the dry air poured into the area after its passage and coinciding with the peak heating of the day.   The low level jet was about 80 knots, and since this passed over the area during the day, mixing brought these stronger winds to the surface, and not just for the Roanoke valley, it occurred over central and eastern VA/NC. Also noteworthy, the relative humidity never recovered the following night.  The Roanoke Regional Airport remained below 30 percent for 32 consecutive hours.
 

Myatt: (2) Any statistics/data that you may have that would give a picture of how severe, widespread or unusual this event was?

Keighton: I think part of this question was answered above with stats about all 40 counties impacted and the anomaly chart of upper level winds. Measured winds of 81 mph at the West Jefferson NC airport, 75 mph in Fairfield (Rockbridge County), 74 mph at the Hot Springs airport, and 67 mph at the Roanoke airport, are among the highest recorded winds from non-thunderstorms we are aware of in our area, but we don't keep accurate climatological records of wind gusts (I think it's the 3rd highest measured wind by the ROA ASOS, but I can't confirm that).

Manuel: This event affected a large geographic region... from the Blue Ridge to Coastal VA/NC.  According to the Virginia Department of Forestry, the fire occurrence on February 10^th, 2008, set a state record for the number of fires ever recorded across the state of Virginia in one day (not sure of the previous record, you would have to contact the DOF).  For the date, 351 wildfires were reported, consuming approximately 16,000 acres.  This one day event accounted for 28 percent of the total number of fires for the year.  For the year 1,241 fires were reported, burning 26,323 acres.  Annual normals are 1200 fires for 15,000 acres.  If you subtract the fires that occurred on February 10^th from the seasonal total, the 2008 fire season would have been below normal.

Keighton: Also, we were told by a number of county 911 centers that this was their busiest day ever (mainly because of all the fires).

Myatt: (3) What has the NWS-Blacksburg done to study this event, learn from it, and apply that to its future forecasting/warning?

Keighton: We have actually done quite a bit with this kind of event in general already, and so we have not done an in depth study of this particular one, but have reviewed it to see what was unique about it and specifically why the damaging winds extended so far east. We have a pretty good handle on when to issue high wind warnings for post-frontal widespread wind storms in the mountains and foothills, but the most difficult question is often when to extend these warnings east of the Blue Ridge into the Piedmont areas.  After reviewing this event, this scenario appears to require a unique combination of all the factors covered above, such as afternoon timing of when the strongest gradient in upper level winds and sinking air passes over the Blue Ridge, as well as the lowering of the inversion to just above mountain top level to create the breaking wave effect. Of course, knowledge of potential fire danger given fuel moistures, drought conditions, and expected humidities is also crucial toward determining the ultimate impacts from such an extreme event, and the Feb 10, 2008 event reminded us that even though fire danger season may not be officially underway, it does not mean that extreme fire danger situations cannot arise. Our office did have a Fire Weather Watch out by early in the morning the day before the event for the entire area, and then a Red Flag Warning out by the afternoon the day before the event. For the general public, we also had a High Wind Watch out two days before this event, and a warning out early in the morning the day before the event, but mainly for the favored areas in the mountains. As the event approached and then got underway, we continually expanded warning areas out into the Piedmont, so we were a bit behind on how widespread this event was. Our review of this event has helped us to understand just how precisely all these factors have to come together to produce such an extreme event. We also have begun a high wind climatology, which considers a number of different general patterns, including post cold front events (such as this one) and deepening coastal low events (which may appear similar to the public in terms of strong NW winds and cold advection, but aren't really associated with an obvious cold frontal passage since the system is really developing along the coast.  I'm going to try and attach a power pt presentation which has some charts showing that the cold frontal events are our most common, and that overall high wind events are quite common in February with a maximum start time for the events during the day (when the mixing down of higher momentum air is most efficient).   April 16, 2007 fell into the category of a deepening coastal low, and I say more below about the similarities and differences in answer to your next question.

 Myatt: (4) Is there any other event historically that somewhat compares to Feb. 10, 2008?

Keighton: There have been several over the last 5 to 10 years that shared some similarities in terms of either the large scale pattern or resultant widespread damage with Feb 10, 2008, but none that we have documented quite match just how widespread the damage was and how strong some of the winds were.Nov 13, 2003 had a similar pattern with a deep layer of strong gradient, but not quite as extreme in terms of the anomalies to climatology and resultant winds were not as widespread (mainly confined to mountains and foothills)

 April 16, 2007 was similar in terms of long duration and widespread damage out into the Piedmont (but maybe not every single county), but was a bit different in terms of large scale pattern (strongest gradient was in lower levels around the bottom of a deep upper level low).

Manuel: From a fire stand point, the Virginia Department of Forestry, might have your answer.  I don't.
  

Myatt: (5) The question I get on a regular basis: Have high wind events like these increased in frequency/severity in recent years? I regularly get people who have lived in the area for decades who say these kind of high wind episodes didn't used to occur as often as they do now.

Keighton: I'm sure there have been others many years ago, but we do not have the detailed documentation on anything prior to 1994 to really be able to compare them. With the rapid increase in spotters and other reporting stations and the opening of this office in 1994, we have certainly issued more high wind warnings and solicited more reports, so the official Storm Data publications will suggest more frequent events since this time, but I'm not convinced these events are any more frequent in reality than prior to that.  General population increases in mountainous areas will lead to that perception as well, but I can't say with any confidence that weather patterns have changed to make these kinds of events more frequent in the last decade.  I think Phil's answer below may have some better quotes for you than mine!

 Manuel: And I can provide my own war stories.  I grew up in Franklin County and was on the Snow Creek Volunteer Fire department for several years as a teenager.  I vaguely remember an event (late 70s, early 80s and in February?) when we had a very strong dry front come through in the middle of the night. It downed trees and power lines, and I remember responding to Smith Mountain Lake (along with other departments) trying to extinguish a fire that ended up burning a dozen or so mobile homes along the lake shore.  Aside from this 40 year knowledge base, I cannot recollect any other event that would compare.I have heard the North Carolina fire managers say, beware of the February cold fronts.  It is during a time of year when the atmosphere is very dynamic and the sun angle is increasing, providing for warmer afternoons.  Our normal fire season runs from February through the end of April (or until greenup,  grass turns green, leaves grow trees).  There is a secondary season during the fall, late October (leaves come off the trees) through November, or until you get a good soaking fall rain or winter storm.  During drought years, fire season can persist year round.

 Wind episodes will always be common here.  We live at the base of a dam... the Appalachian Mountains.  When water flows over a dam...the turbulent flow occurs at the base.   Since our prevailing wind flow is from the west...our mountains will act as a dam...the wind accelerating as it crosses the higher peaks and crashing into the foothills just east of the mountains... simple fluid dynamics.  Frequency of these events will fluctuate.  I think it would be safe to say, an active northern branch of the jet stream would likely generate more wind episodes for our geographic region.  Correlations could possibly be made to having more wind episodes during a La Nina Year as opposed to an El Nino... but I have not sat down and done any research on this. 

As for whether a fire season is going to be bad or not, this correlates with the rainfall patterns... drought primarily.