Storm Retrospective: Kinsley/Rozel Tornadoes, May 18, 2013

John Monteverdi

Thom Trimble 

All images copyright © John Monteverdi and Thom Trimble

The storm that eventually produced at least two major tornadoes  near Rozel, KS formed southwest of Kinsley, KS at around 6:45 PM CDT.  

The storm developed, we believe, in a narrow slot of inflow air that had about 3F higher dew points and slightly lower temperatures than did the surrounding air being ingested by storms further north and south.  I've drawn an arrow on the 2100 UTC surface chart to show the flow of air into the storm genesis area.  Note that the dew points on the upstream portion of the arrow are in the 70s.

I wish I could say I had foreseen that narrow jet of air.   But I hadn't.   It's that kind of mesoscale detail that cannot be forecast in advance.  And the luck here was that we were in the exact position at which that kind of feature could really cause explosive development of a storm.   It certainly did in this case.

The storm immediately had a bell shaped lowering (see image to bottom center). Within in a few minutes it had a very good visual appearance with a rain core out to its northeast and something like an RFD cut developing too.

Arrow indicates approximate path of what we think
was a moist inflow jet that the storm intercepted.

We stayed with this storm for an hour or so and  during that time were treated to an absolutely fabulous example of a storm proceeding through the supercell cascade.  While the storm was about 5 miles west of us, the southeast winds really kidked up a notch....probably 20-30 mph.  We believe that the  inflow jet, as explained above, was being intercepted by the storm.   I got a dewpoint of 70.8F on my Kestrel and  a temperature of 84F.  We commented  that it was hard to figure why the storm had such a high base and no wall cloud.     

Bell shaped lowering on rapidly developing storm
west of Kinsley KS.  

Shortly after the development of that inflow jet,  though, the storm base lowered and the RFD slot sharpened.  The updraft area showed signs of bvecoming an RFD occlusion...and, then a nub formed on the north side of the horeshoe base (see picture below)  That rapidly descended visually and became a large tornado, somewhere between a wedge and cone.

Horeshoe base, showing rear flank downdraft cut, and developing cloud base circulation
(the nub above the centered tree).   

About 2 minutes after previous image.   Tornado circulation now developing at ground level.

This is the instant of initial contact of the tornado with the ground.  It's zoomed, but you can see the same trees and farm buildings that are to the left of the distant forming funnel in the picture above.

The good news is that we got great digital stills of the whole thing, and also video.  The bad news is that things happened so quickly that I never had an opportunity to setup the good video camera on a tripod.  Nevertheless, I have video documentaiton of the first tornado development on the lowered base.  

The radar below shows the storm when the tornado cycled up to its strongest.  Our position is shown as the yellow circle.   The names plotted there are the other chasers and storm spotters who were reporting their positions.  The purple triangle is a Tornado Vortex Signature.

GRLevel3 Radar Plot of Kinsley/Rozel storm at time of tornado genesis.   Our location shown as yellow circle.  The magenta triangle is not the tornado but a radar signature called Tornado Vortex Signature. (Time shown is PDT)

Tornado damage (blue and yellow..estimated by
ground survey;  red estimated by DOWS)
(Image courtesy of NWS)

The tornado moved northeastward, as shown on the graphic on the right, until the process known as tornado or RFD occlusion completed.  The tornado turned to the left, cyclonically, around the mesocyclone.  The red symbol at right is the only location where the tornado was sampled by the mobile doppler radar (Doppler on Wheels--DOWS).  There the winds were measured at 165-185 mph and the circulation width at the ground at 3300 feet.  On this basis the tornado was rated an EF4.

At the time the tornado was most intense, the radar signature was extremely impressive.  I've included at right the radar plot from the DOWS at that time below.

DOWS radar plot corresponding to red
symbol above.  Upper left, reflectivity;  Upper right, storm
relative velocity.  (Image Courtesy of NWS)

The tornado was extremely impressive visually for nearly its entire life cycle.   Its visual appearance varied from a cone to a wedge.  It was quite obvious that the funnel was extremely large and had violent circulation around it.

We drove north towards Rozel and the tornado began to narrow...and at first we thought it had begun the roping out process.   But shortly after that the trunk redeveloped with lofted debris north of town.   I told Thom I thought another tornado might develop back along the flanking line, which by that time was east of us...as we were in roaring RFD (warm) air.  

These three images were taken over a five minute period and correspond to when tornado #1 was at its most intense.  We were located about 2 miles south of Rozel at the time and the pictures are shot looking northwest.

Tornado begins to narrow at its base just west of Rozel.

New tornado formed east of Rozel rapidly as first tornado was in
the process of dissipating.  Setting sun provided wonderful
color to all cloud features.

We turned east and moved through Rozel.   We wondered why the police had roadblocked several roads east of the first tornado. Out of the murk, we could then see another fully developed cone...nearly a wedge briefly back along the flanking line.  We barely got pictures of it.

This also was a very substantial tornado, as you can see in the picture at left.  The setting sun began to produced unusual orange and purple highlights on this thing.  It really was gorgeous, towering above us to the east.

The DOWS did not manage to redeploy so quickly to sample the rotation with this tornado.  But it probably would have been at least an EF2 from the looks of it, though that is very difficult to judge from visual appearance alone.

Meanwhile, back at the original occlusion point, the original tornado was still pulsing, but in a downward trend.  At one point it redeveloped a massive cone.


While we were watching the second tornado that formed further southeast along the flanking line, yet another tornado
seemed to appear to the north.

We're debating whether that was a new tornado, or the remnant of the original tornado as the main updraft area probably had crossed the road at that point. That tornado refused to give up the ghost, so to speak.   It took nearly 30 minutes for it to finally rope out to nothing.

To the right  are some pictures of the rope out phase of tornado 1 (or perhaps, as explained above,  it was tornado 3).   We were getting confused in the darkness and the episodes of good and poor visibility  And while it was doing that, it still had a substantial ground level circulation. Finally, even the tiny remnant of the funnel seemingly floating disconnected was associated with ground level debris.

Thom and I agree that tornado 1 was the strongest, and largest, tornado we've seen in our chase careers.  And now that we've had several days to consider what we saw and experience,  we agree that it was the most phenomenal chase of our chase experiences together.

From the decision to get us there, to the good fortune that made us stay in place in Kinsley instead of going north or south, to the road we chose with absolutely perfect vantage points, all was just about perfect.

Last remnants of the rope out phase of Tornado 1 (or it may have been a third tornado).  The last portion of the
funnel seemed to float free, yet it was still associated with circulation at ground level and debris.

May 22, 2013: Down Days Coming and Early Exit

Retrospective

Thom and I are in Wichita Falls after a day in north-central Texas.  Although there looks to be a supercell-favoring pattern tomorrow (May 23) in northwest TX and the southern TX Panhandle, I've decided to return home with Thom, who has get back due to family issues (good ones, though).

This will be the last active log, and I will now concentrate on trying to get some pictures up for our productive days, mostly May 18 and 19.

I am impressed with the numercial models, particularly the ensembles, which suggested a pattern favorable for supercells in the central Plains almost 10 days in advance.

It turns out that several research projects on tornado formation and evolution were able to get excellent data sets during the last five days.  The Moore, Oklahoma tragedy casts a pall on this.  Nonetheless, we are going to learn much about tornadogenesis because of these couple of projects.  They were not on the Oklahoma City supercell, however, as they were south where most storm spotters and chasers were that day.

I just learned that the tornado we documented near Kinsley, KS has been rated an EF4.   There was no damage (as far as I know) to assess to determine this.  But the Doppler on Wheels (DOWs) were there, and they were able to estimate the wind speed associated with the funnel.  The map above shows the track for the first tornado.   The colors correspond to damage or DOW evidence for a rating.  The red tag shows the position of the tornado when the winds were 165-185 mph.  This corresponds to an EF4 rating.  The radar plot below right shows the DOW plotted radar images at the time the tornado was at maximum strength.   We have pictures and video corresponding to most of the 7 mile damage path for this tornado.

The damaging tornado that struck Moore is still making the news.   I was interviewed via cell phone by KTVU-Channel 2 in Oakland...but I declined other interviews (one from CNN, and directed them to my colleagues who are actually in the Oklahoma City area and who have done refereed research on the previous tornadoes that struck that area.

One reason I don't want to answer specific questions about the Moore tornado is that Thom and I were not there, nor have we assessed the damage.  Several of my colleagues have already been involved in that, and their results already are belying the hyperbole about that tornado (for example, it was not two miles wide, nor did it have a two mile wide damage path.  So far, I've only heard of documentation of damage consistent with an EF4 rating.   But I only get that information hours after the surveys take place.

I'm personally curious what meteorological parameters were in play in the Oklahoma City area that afternoon, that were not in play further south, where most researchers, spotters and chasers were placed.  It's not obvious from the weather data and charts I have for that hour.

Yesterday's Chase

The early runs of the models suggested a few isolated supercells forming around 1PM between Fort Worth and Abilene.  By the time we got into that area, the models were showing none of that, but more of a discontinuous line.  And that's what happened.  We played hopscotch with the rapidly progressing front, trying to stay in the rich moisture ahead of it.

When we got to Waco, several of the cells on the south end of the line became somewhat discrete.  They showed some signs of becoming supercells, but never did.  So, at around 6PM, we called it a chase, and drifted back to Wichita Falls.  Along the way, we changed our plane reservations from Sunday afternoon to tomorrow, Thursday, May 23.

This will be my earliest exit from a storm chase in all my years of chasing.  However, apart from the first day, we've had pretty intense chasing every day of the trip.  Of course, the best and most productive days for documenting tornadoes occurred on May 18 and 19 for us.

Once I have more time to examine our photography and videos for May 18, I will begin posting photo documentation of that day.

May 21, 2013: North Central TX

Another Round of Supercells Expected Today:  Risk of Strong Tornadoes Again

Not much time to post as we expect early initiation.  We are currently near Gainesville TX and then are heading southwest to the area just east of Abilene.

The morning weather map shows the low pressure area having drifted southward into central TX overnight.   The setup is a bit different today as 500 mb winds (see graphic) have slackened just a bit.

However, there is still the tap of high dew point air coming into central TX.   The surface flow is a bit weaker than yesterday.   Yet forecast hodographs indicate wind shear favorable for supercells, although the window for tornadoes appears short.

An additional factor is an outflow boundary that is shifting southward out of the thunderstorms we just left in Ardmore.  That also might provide a focus if storms forming to the west cross it later today.  Unfortunately, the Dallas-Fort Worth metroplex may be under the gun later today.

Moore Tornado

Just an awful coincidence that the tornado that devastated Oklahoma City's southern suburbs took basically the same path as the May 3, 1999 tornado, and very similar path to the tornado that struck the same area a couple of years later.

Most chasers were on the storms further south.  Thom and I passed that storm when it was developing and figured that more substantial storms would be further south...so we went on by, in fact around it, to get to the tornadic storm further south.  We did not get to any storm today that had a tornado.

Just as well that we didn't leave even later, as we came from Ponca City in the north....and would have probably chased that storm as our only option later in the day.

It looks like all my chase colleagues are safe, and most who are down here are choosing not to contribute to the confusion in Moore....as I35 passes right through that beleagured town.

The graphic is designed by Meredith Botnick to show support for the people of Oklahoma who have endured tragedy today and yesterday.

There is a plea for $10 donations, via a text number.

Text 32333
Subject:  FOOD

That will provide a $10 donation for victim relief.

May 20, 2013: S-Central Oklahoma

More Tornadic Supercells Today in Oklahoma:  Preliminary Target for Storm Intiation---Lawton, OK

(More complete discussion with illustrations later.  Also, still have not had time to fill in log with pictures.   Looks like we may have a down day soon, on which I can catch up).

We overnighted in Ponca City, Oklahoma.   Just returned from a nice run around Lake Ponca.

The morning weather map shows a surface low in sw Oklahoma, with a dry line extending southwestward into north Texas.  A cold front extends from that dry line, with the intersection near Altus OK, back northwestward.

A warm front extends from that intersection across Oklahoma to se KS.  That three way intersection is often referred to as a "triple point" by severe weather meteorologists.

Ahead of the dry line, an air mass rich in moisture is found to the warm front.  That air mass is projected to have surface based CAPE values (a measure of how warm ascending air parcels will be relative to their surroundings through the deep troposphere) of 4000-5000 J/kg (see graphic to right).   These really are values at the upper end of the sort we usually see with severe weather, and indicate (Metr 201 and 415/715 students will know this) updrafts that should support extremely large hail.  On the map above, the light green represents areas with 60F+ dew point, and the darker green 70+.

Once again a very strong jet stream  (see graphic at upper right) will produce wind shear values favorable for supercell convection and, in parts of Oklahoma, low level wind shear favorable for these supercells to develop tornadoes.

The issues center on two aspects of  the forecast storm motions.  First, the storms will be moving at an angle that's too close to the dry line, instead of at right angles.   This could mean that each storm will be running over a previous storm's cold pool....and we can end up with a big line of storms, as we did yesterday in northern Oklahoma.  If that happens, the time window for tornado formation will be small...before each storm runs into that cold pool.

Second, unless the surface winds cooperate, storm motions could be 40-50 mph, as they were yesterday, making for exceedingly difficult chasing....it's almost impossible to set in one location and take pictures, for example, when the storm rushes by like that.  Fortunately, the surface winds are forecast to be more southerly or south-southeasterly, rather than hard southeasterly.  In that case, we're likely to have storm motions of around 30 mph.

In any case, we are off to our target area, where the visible image above shows a low level cumulus field already bubbling.

Brief Elephant Trunk Tornado near South Haven KS

Tornado near Sourh Haven KS developed right over our heads.

May 19, 2013 -- Provisional Target for Storm Initiation North-Central Oklahoma

Another Round of Tornadic Supercells Expected:  Provisional Target--northernl OK/se KS

(Complete log of yesterday's magnificent cyclic tornadic supercell will be posted during the day today)

Our first look at data this morning shows a triple point (join between a warm front, dry line and cold front) somewhere jn northwest of Oklahoma City today, drifting much through the day, and ending up near I35 east of Enid (see graphic to the right).  There is an outflow boundary from last night's convection over southeast KS and northeastern OK.

Dew points east of the triple point are in the 60's with 70s surging northward in southeasterly flow over the eastern two thirds of Oklahoma into southeast KS.  On the chart at right, green areas reprsent dew points >60F.  Deepest green indicates dwpoints >70F.

The 500 mb wind flow (~18000 ft) will be the strongest of the trip so far, with widespread 40-50 knots in southwest flow over the threat area.

Hodographs (diagrams that show the variation of wind direction and speed with height)  plotted for this afternoon just east of the dry line show a pattern favorable both for supercells and for a wind profile in the lowest km or so supportive of tornadoes.  Some of those tornadoes might be strong and violent.  Other measures of updraft rotation, such as 0-3 km Storm Relative Helicity (see diagram at right) are consistent with the impressive hodographs (the green and yellow areas on the chart at right are an estimate of updraft rotation).

The 12 UTC run of the WRF-NAM is not yet in.   So Thom and I will go out on a run, and we'll see then what our actual intiial target will be.  Our target is shown in the red box above.   My thought is that the outflow boundary will act as a storm initiation mechanism.  There is also enhanced shear along that.  That shear can be ingested into the updraft of thunderstorms and augment their rotation.

Meanwhile, tomorrow's pattern looks to be in about the same area, or slightly further southeast...but with conditions favoring another outbreak of supercells, some with strong and violent tornadoes.

Tornado 1 Ropes Out

Tornado in process of formation, between Kinsley and Rozel, KS...May 18, 2013

Funnel clouds near Kinsley KS

Dryline initiation

Convergence of Chasers

Here's an image showing the positions of storm spotters/chasers at 11:49 AM CDT.    You can see that there will be quite a few of them out today.  Let's hope that all engage in safe behavior.  Apparently in the Dallas area last Wednesday there were chasers parked in the middle of highways, for goodness sake.

May 18, 2013: Target Areas Range from sw Nebraska to scentral KS

Significant Risk for Tornadic Supercells over Broad Area Today

Thom and I are just heading south from North Platte.   

While significant CAPE exists from I80 south surmounted by more than adequate storm relative shear for supercells, I don't much like the storm motions relative to the dry line and the mid tropospheric winds further north for sustained discrete convection and storms that don't collide with other storms' cold outflows.

The RAP and WRF-NAM are really confused about progged dew point and CAPE/CIN fields.   But both are in agreement about the strong surface pressure falls in sw KS with the low pressure area developing eastward into that area  and a dry line bulge coming out of OK into that area.   The forecast hodographs are impressive for the area from Wichita west to Dodge CIty and up to I70.   So right now we are targeting that area.

However, we are going to think harder about that target when we get to McCook or so.  The HRRR is consistent too in showing cells in SW Nebraska with a lot of updraft helicity.  But later runs also show really imprssive looking cells in south-central KS.

Below I've included the WRF-NAM forecast for Pratt KS at around 7PM this evening.  The buoyancy is really upper end, with 5000 J/kg and no CIN.  Metr 201 and 698 students will remember that the forecast for maximum upward velocity (assuming no entrainment and no turbulent mixing) is the square root of 2 X CAPE.  In reality, maximum updraft at the EL will be about 1/3 or lsss of that....but that's till nearly 100 mph....enough to justify the forecast of cantaloupe sized hail.

KLNX - SuperRes Reflectivity Tilt 1 2:25 PM

KLNX - SuperRes Reflectivity Tilt 1 2:25 PM

Friday May 17 Target---NW Nebraska

The "Road Less Traveled".   Panorama shot on Highway 61 in the Sand Hills, somewhere north of Hyannis

Initial Target--Through the Sand Hills

The morning weather map shows that yesterday's warm front has drifted north to the South Dakota border.

Southerly winds have brought mid 50 dew points into most of Nebraska east of a dry line that extends from the center of a cyclone in ne CO southward.

As the main trough in the middle and upper troposphere (see 500 mb chart posted) continues to evolve, pressures should fall in that low, drawing southeasterly winds across most of Nebraska, and bringing even 60+ dew points into the state.


 Given the winds aloft, southeasterly surface winds are very favorable for the kind of deep layer shear (see hodograph) that encourages supercellular convection.

Right now, the best combination of this deep layer shear with the buoyancy most favorable for deep convection (see sounding) is projected to be in the red box drawn on the surface chart.  There are also several aspects of the deep layer shear in combination with the projected low level shear suggestive of relatively long lived  rotating storms, although they may tend towards the HP mode.


The High Resolution Rapid Refresh Model (HRRR) depicts storm initiation in the eastern Nebraska Panhandle around 20 UTC (3 PM CDT).  I've included that graphic and also what the HRRR depicts that echo doing by 0000 UTC (7PM CDT).  The model creates a very intense mostly isolated storm that appears to be a supercell, having moved northeastward to just south of the South Dakota border.

So our target is not far today.  Also, it does look like the highest risk for tornadic supercells tomorrow will be along the KS-Nebraska border, so whatever happens today, we will be in good position for tomorrow's storms.





The Chase

Panorama of Alliance, Nebraska elevated supercell, view towards the northwest..   Detail of the updraft base is given below.  Inflow bands at midlevel and a large low level "tail" inflow can be seen on the right side of the image.

The cell we targeted was east of Alliance.  It had the familiar
hook and flying eagle presentation of supercells.

We drifted west to Ogallala as the first round of initiation began to occur in eastern Wyoming.  We drifted north on state highway 61 through the Sand Hills and into Cherry County.

Cherry County has only a few roads, so it is a target area dreaded by chasers.  The early initiation was pretty much as the HRRR depicted (as shown in the last section.  The early storms congealed into a small line, and we attempted to get on "tail end" charlie.

That we did.



That portion of the line was broken up into somewhat discrete cells.  These were rotating and presented as small supercells on radar.  We targeted the south cell that had somewhat of a couplet on storm relative velocity plots.

The storm also had a moderate couplet on it, as shown on the
North Platte Doppler radar storm relative velocity profile
(The light red (outbound) next to the dark green (inbound).

By the way, it's worth it to point out that our decision not to activate Mobile ThreatNet was a good one.   ATT has excellent coverage through most of the Plains, and good enough coverage even in the traditional "holes" (like the Sand Hills) for RadarScope to get files (even though at exactly the same location GRLevel was not able to receive files.

In addition Thom had an iPhone using Verizon, and we were often able to alternate phones when GRLevel was not able to get files, and one of the other, Verizon or ATT was not in range.





As we dropped southeast of that storm, we could see it was elevated.  It was clearly a supercell, but the the surface dew points were too low and the condensation was developing four or five thousand feet above the ground.  There we could see the familiar horsehoe base, with a Rear Flank Downdraft cut.


The panorama of this storm (shown above)  shows that it had midlevel cloud bands extending into it and a large inflow band at lower midlevels.  It was indeed a supercell, and somewhat picturesque.  However, it did not have a chance of becoming tornadic.

Final view of "tail end Charlie" showing detail of horseshoe base, indicative of a mesocyclone in the lower midlevels of the storm.   The rear flank downdraft cut is also visible at the center of the horseshoe and lined up well with the radar presentation (examples above).