Thursday, April 30, 2009

April 29 review: mesoscale magic in the TX South Plains

A potent supercell tracked between the Plainview and Matador TX areas yesterday, producing a couple substantial/photogenic tornadoes northeast of the town of Floydada. Surprisingly, not terribly many local chasers were on it... perhaps due to its relatively short tornadic period before going OFD... or due to other sucker storms existing farther E. SPC handled the forecast well IMO. Model signals were pretty subtle for a significant event... though knowing that OFBs might come into play later, coupled with the amount of BL moisture lurking to the southeast, was a sign that some mesoscale magic could definitely occur.

During the morning hours, an elongated shortwave trough tracked eastward thru KS-OK, supporting scattered thunderstorm activity. The airmass feeding into these storms was characterized by anomalously strong precipitable water values and effective parcels already rooted near the surface. A strong MCS was thus gradually born over the Red River valley, on the southern flank of the convective activity. The far western segment of the MCS outflow appeared to settle southwestward, and by early afternoon curved E-NE-N along a Crowell-Turkey-Wayside-Amarillo TX line (see 1945Z visible satellite image below). A second surface boundary was also evident on satellite, extending E-W from Crowell-Matador-Plainview and intersecting the dryline near Springlake. The latter E-W boundary was not apparent via the TX mesonet, but gradually appeared as a fineline of reflectivity on Lubbock's radar... its signature probably becoming blended with an increasingly well-defined RFD gustfront off the supercell-to-be. Also of note, a convergence line/surface trough was present in the moist sector a few counties east of the dryline, generally straddling the I-27 corridor.

A trailing shortwave trough, low-amplitude but well-defined in nature, approached west TX by mid-afternoon. Persistent isallobaric forcing, likely a result of both an upswing in QG ascent and increased heating of the outflow air, produced a well-defined low-level mass response along and east of the dryline between 21Z and 23Z. (
See sfc obs below... no frontal analysis done as my usual analysis software doesn't have the TX mesonet data.) Enhanced downward mixing of westerly momentum forced the dryline eastward, where it merged with the convergence line near I-27; meanwhile, surface winds in the moist sector increased and backed substantially to the ESE. The supercell in question developed near the intersection of the E-W boundary and the effective dryline by 21Z, and tracked E to ESE along or just north of that boundary.

The Jayton profiler, located about 50 miles SE of where the storm became tornadic near Floydada, captured the mass response quite dramatically. Note in the 22Z and 00Z Jayton profiler hodographs below, the 500-1000 m flow backing and increasing to near 30 kt. Accordingly, given such a strong sr-inflow component and a perfectly kinked and curving hodo below 1 km... strong to very strong 0-1 km SRH resulted. Using the Plainview storm motion, 0-1 km SRH on the 00Z Jayton profiler was > 330 m2/s2 (with continued strong curvature through 3 km contributing to > 550 m2/s2 of 0-3 km SRH).
Also note, 0-6 km bulk shear was moderate to strong, in excess of 50 kt. In comparison, the 00Z Amarillo RAOB hodograph (third hodograph below) showed 0-1 km SRH of ~200 m2/s2... though it's possible the strongest low-level mass response hadn't yet occurred by the time of balloon launch (~23Z) and/or occurred farther south only. With neither OFB evident in surface data and likely becoming increasingly diffuse given increased mixing/heating and mass response... it is tough to know whether or not the boundary enhanced horizontal vorticity available to the storm. Regardless, I would think the shear sampled by the 00Z Jayton profiler is a reasonable estimate for the Floydada area tornadoes, with any changes in the shear due to increased elevation on the caprock escarpment probably marginal.

The thermodynamic environment available to this storm was likewise very impressive, due to the influx of upper 60s surface dewpoints up onto the caprock combined with fairly modest capping aloft. In order to achieve the correct elevation/pressure surface for Floydada (which is a good 500 feet lower than AMA), I used a base NAM-WRF model sounding with a 900mb pressure surface (ideal given relative pressure surfaces in the region that evening) and applied the AMA RAOB data to it, blending the BL using TX mesonet data. The resulting dewpoint lapse rate is dry adiabatic (similar to that observed at AMA)... but this is known to happen when anomalously rich near-surface moisture exists. Bottom line, note the strong instability (MLCAPE > 3100 J/kg), very strong 0-3 km MLCAPE (~150 J/kg), and sufficiently low MLLCLs (1200 m). The second sounding below is the raw 00Z AMA RAOB, which likely got "munched" by anvil-level convection just above 200 mb.

In reviewing this data, it is probably a good thing for TX residents that the storm quickly became outflow dominant (with the gust front apparently undercutting the updraft and encouraging evolution to a small bow)... because the mesoscale environment was truly primed for strong to violent tornadoes. If I had been free to chase on this day, I hope that I woulda taken the chance on it.

Sunday, April 26, 2009

April 25: underwhelming OK storms, and tornado near KC metro

Observed a couple near- and after-dark supercells yesterday south of Clinton, OK between 830 and 930 PM. The structure of the lead supercell (first pic below) was pretty trashy... high-based and dominated by HP processes. The trailing supercell initially had a stacked plates look to it... and briefly developed an increasingly low, compact, and rounded base beneath its vigorous/hard updraft (second pic below)... before raising its base again and looking pretty underwhelming after dark. (In hindsight, I probably should have exited my vehicle to ascertain whether outflow air from the lead supercell had contaminated the environment of the trailing supercell-- which actually tracked farther NW). Heard some radio reports of wall clouds with moderate rotation from these supercells, and was surprised... I didn't see anything of that sort... in general they were high-based and non-impressive.

Similar to what happened the day prior (April 24), afternoon boundary layer moisture mixed out a good bit yesterday across far southwestern OK (see 23Z surface map below) in vicinity of the storms... with surface dewpoints falling to the 58-61F range. The 12Z NAM suggested this might be a problem... and BL trajectories emanating from the strongly/deeply mixed moist sector airmass around Childress/Wichita Falls didn't help any. While it appeared that more substantial moisture survived mixing and was lurking immediately to the east, that was not really the case... with the richest moisture being rather shallow on the 00Z OKC sounding (surface dews 64F but 100mb ML dews 60F)... not much improvement from the afternoon before. This was a far cry from what was forecast in previous model runs, and the "real-world" result was not surprising: high-based supercells amidst modest SRH before dark... transitioning to more low-based but increasingly elevated trash with diurnal cooling. Bottom line, by the time diurnal cooling/decoupling dramatically improved LCLs and SRH... the BL was probably too stable.

Interestingly, it appears that richer BL moisture pooled ahead of the quasistationary front from northcentral OK to northeastern KS, possibly due to less vigorous vertical mixing. This, combined with cooler thermal profiles through the low- and mid-troposphere, resulted in strongly surface-based environments but with far lower daytime LCLs than in SW OK. This may have been a factor in allowing a supercell to initiate on the front near Lawrence KS, move east a bit, and produce a beautiful 15-minute duration tornado in Leavenworth county around 630 PM. Another tornadic supercell moving off the front affected the Enid area of northcentral OK shortly after dark as well.

The 00Z Topeka RAOB (sounding #1 below) was launched just to the cool side of the quasistationary front and ~40 miles west of the Leavenworth co. storm. Given the shallow slope of the frontal system (frontal inversion only ~350 meters deep) topped by a moist layer/cap aloft almost identical to that observed in the warm sector well SSE on the SGF RAOB... it looked perfectly feasible to modify the Topeka RAOB to represent the Leavenworth co. tornado (sounding #2 below). Given relatively rich moisture just above the inversion, again the assumption is that somewhat richer moisture (ML dewpoints around 62F) survived mixing ahead of the front in comparison to in southwestern/southcentral OK. The modified Topeka sounding cranks out moderate instability with MLCAPE of 1689 J/kg; a strongly surface-based environment given MLCIN of 23 J/kg and 0-3 km MLCAPE of 59 J/kg; and favorably low MLLCL at 973 meters.

Hodograph #1 below is that observed on the 00Z Topeka RAOB; while hodograph #2 below shows the 22Z Lathrop MO wind profiler (in the warm sector ~60 miles ENE of the tornado). While near-surface winds backed slightly 22-00Z just ahead of the front, moreso than at Lathrop... the Lathrop profiler also showed more curvature through 1 km (~1 km flow more veered) than the warm sector RAOBs at SGF and OKC, which looked suspect. So it may be a wash... would estimate 0-1 km SRH of around 140 m2/s2 with 0-6 km bulk shear of 35 kt for the Leavenworth co. tornado. (Mid- and high-level sr flow was obviously pretty poor as well, though that didn't seem a major factor).

This was probably the 7th or 8th time I've been chasing well away from KC, only to have a semi-unexpected substantial tornado event occur close to home. Interestingly, it seems that many of these tricky far northeast KS-northwest MO "missed events" occur amidst shear environments that are not particularly impressive, with tornadoes probably occurring equally if not moreso due to processes associated with strong near-ground vertical stretching rather than low-level shear; the May 29 2004 Camden Point MO F3, June 4 2005 Hiawatha KS F2, and June 8 2005 Oskaloosa KS F0 come to mind, all of these being gorgeous (and in the former two cases, very strong) tornadoes. Mother nature likes to keep it interesting I guess, and I don't see that ever changing. Looking back at the model data from yesterday morning... I probably should have given the frontal boundary in that area a "closer look" with regard to chasing... esp since I was due back at work today.

Congrats to Dick and Darin for catching a gorgeous stovepipe tornado today not far from yesterday's action area... Roger Mills county OK. Sfc dews had no problem holding in the mid 60s today in the warm sector.

Friday, April 24, 2009

April 25 forecast

Gonna be as brief as possible here... in between getting off of work at 11 PM and charging my camera batteries while seeing if I can rid myself of a migraine-like headache.

Tomorrow's severe weather forecast is a bit complex, but at least should be centered over a fairly localized area over the southern High Plains. 12Z RAOBs this morning at BRO/CRP/DFW showed that the deeper BL over the Gulf of Mexico had finally retreated inland, with impressive moisture quality (100 mb ML dewpoints ranging from 61F near the Red River to 68F along the gulf coast). This moisture mixed out a bit this afternoon while spreading northward over north TX and OK... but remained at an impressive depth of 150-175mb on the 00Z RAOBs.

Model concensus is for the cold front to stall across the northern TX panhandle, far northwestern OK and southeastern KS by early afternoon tomorrow... with the effective dryline becoming pronounced near the TX panhandle-western OK border via diurnal mixing to its west and rich moisture flooding northwestward to its east. Previous model forecasts of ML dewpoints in the 63-65F range in the target area now look fairly realistic... far more impressive moisture than we've seen lately... have to go way back to the Feb 10 event along the Red River to find moisture like that.

Timing and location of initiation is uncertain. Upper air analysis and NAM/RUC initalizations suggest a possible low-amplitude impulse over the CO river valley this evening that could potentially overspread the southern High Plains tomorrow aftn/eve... but think it will be of little consequence, and that the vertical motion field on the NAM may be more a function of the convection explosion forecast to begin around 00Z. FWIW, the flow at jet level is forecast to become increasingly difluent and divergent through the afternoon and evening, which may offer some support for upward vertical motion. On that note, I expect a fair amount of cirrus overspreading the area tomorrow... anyone's guess what impact it will have on mixing/temps/initiation.

The triple point eastnortheastward along the cold front is the most obvious place to expect initiation by around peak heating, where convergence will be strong. Despite the front likely lifting back north after dark due to considerable pressure falls overspreading the cool sector... still not sure I'm crazy about the triple point target... will have to evaluate that more later.

The dryline itself is forecast to be weakly convergent by mid- to late-afternoon. Interesting to note the NAM is forecasting a very deeply mixed moist sector tomorrow afternoon (sfc temps 85-90F or more) as far north as far southwestern OK and northwestern TX. It's possible this could enhance low-level baroclinicity/cyclogenesis and more vigorously mix a dryline bulge into southwestern OK during the afternoon. Equally as possible if not moreso... diurnal backing of the low-level flow and revving of the LLJ could result in a more strongly convergent
retreating dryline after 00Z... model signals suggest it's quite possible this could be the mechanism for initiation. This evolution might actually be "better" given SRH will be weak to nil prior to 00Z.

It's tough to trust anything in the NAM in the 00Z-06Z time frame given the convective explosion that is forecast. However, given the anticipated quality/depth of BL moisture, capping aloft that shouldn't be out of control (~7H temps not exceeding 8C), the added benefit of elevated terrain in the High Plains, and LLJ expected to accelerate to 40-50 kt after 00Z... it's realistic to expect a window of strong surface-based instability (MLCAPE 2000-3000 J/kg with very weak CINH) for a few hours after dark, coupled with 0-1 km SRH easily exceeding 300 m2/s2. Thus, if storms fire, potential exists for tornadic supercells and a strong tornado or two prior to 03-04Z. Mid- and high-level SR-flow is quite modest, but I can't help but be reminded of the loose analog case of 5-12-04... which featured modest flow aloft as well but had no trouble producing a dry-side-of-classic tornadic supercell.

Preliminary target = Allison, TX in far eastcentral portions of the panhandle. Haven't chased TX in three years!

Saturday, April 18, 2009

cold core tornado event occurring over SW KS!

The usual cold core goldmine of SW KS is producing again... with 6+ tor reports in the DDC area including reports of two on the ground at once. I finished a string of five midnight shifts at 7 A.M. and wish I had looked closer at this potential target and considered trying to chase after work, rather than driving home and crashing for some much needed sleep... but that's fatigue for ya.

The set-up fits the classic Jon Davies pattern pretty well, though the low-level pressure/height field (and thus low-level shear) is very weak. Cells fired first near the intersection of the slow-moving occluded front and a weakly defined west-to-east boundary... then built southsoutheastward along the occluded front. Only the cells near the subtle boundary intersection and/or just to the cool side of the west-to-east boundary have been tornadic. While ambient horizontal and vertical vorticity associated with the boundary intersection and mid-level low are probably playing a role, low-level stretching--associated with steep low- thru mid-level lapse rates and very large low-level instability--are probably dominating w.r.t. tornadogenesis processes. The 12Z DDC sounding this morning sampled 700mb and 500mb temps of -5°C and -25°C respectively... more typical of a late winter cold core set-up (though low-level theta-e is modest, an offsetting factor). Had I seen that 12Z sounding before going to sleep, I may have reconsidered whether to chase.

Anyway, big kudos for John Hart & SPC for going with a tor box on this one in spite of very marginal low-level flow/shear!

Saturday, April 11, 2009

April 9 eastern KS-OK border chase

Dick, Darin, and I agreed on a target in the Coffeyville KS-Bartlesville OK area on Thursday. Given relatively underwhelming moisture return across the board and comparatively poor roads/terrain in the obvious warm sector target, it was definitely a day to give close consideration to the southcentral KS cold core potential ahead of the mid-level low center... but we shared too many concerns about that target being productive.

We moved as far west as Sedan KS 3-4 PM, letting a few marginally impressive severe storms trail by well to our west. We gradually moved east, staying near Highway 166 to keep our options open. With nothing really becoming dominant and with storms farther south largely taking a linear configuration, things were looking a bit grim as we reached Edna KS around 545 PM. Suddenly, a semi-discrete and intensifying storm along the OK turnpike caught our eye. Moments later it received a tor-warning, and that sealed it--we headed farther east on 166 for an intercept. Given the torrential rain and getting caught behind a semi... it took us longer to get there than it should have.

Eerily, we intercepted the storm in the town of Picher OK (struck by a killer EF-4 tornado in May 2008), with tornado sirens blazing as cityfolk looked warily off to the SW to try to catch a glimpse of the storm.
The supercell was in the process of rapidly contracting, and completely dissolved not 15 minutes later. We got a good look at the compact rainfree base (and a couple nub funnels... see pix below, view looking south at 6:37 PM) before it wrapped in rain. Lots of dime hail, with maybe a few quarters, in the core... impressive considering the storm was going downhill so quickly. Interesting that an EM reported a tornado near Miami only a couple minutes prior to our close intercept of the base, but concensus from numerous chaser/public videos is that the funnel cloud in question did not touch down.

Significant tornadoes managed to occur farther south in the Mena AR and Shreveport LA areas. It's possible that the environment was more strongly surface-based in a wider corridor in those areas due to stronger moisture return. Clearly, had our dewpoints been 5°F higher across the board, tornado potential would have increased considerably. There also may have been a few problems with the local shear profile in northeast OK/far southeast KS (e.g. initially modest 0-1 km SRH, or a veering-backing pattern in the vertical), but wind profiler data wasn't much help in conclusively gauging that one way or another.

As for the cold core target along the trailing surface trough/psuedo-occluded front, surface dewpoints appeared to verify about 2°F lower than forecast from the 12Z NAM and as much as 4°F lower than from the 12Z RUC. Given the related LCL heights and 0-3 km CAPE had been forecast to be only modestly favorable to the cool side of the front, I can only assume that the parameters took a further hit given the poorer-than-forecast moisture. Pressure rises in the wake of the primary surface low, with the wind shift pushing southward as a cold front, may have also acted to minimize the width and/or duration of the unstable boundary layer along and N of the front... if not undercutting updrafts altogether. The 00Z/12Z NAM and 09Z/12Z RUC did a good job with forecasting the evolution of this surface boundary, despite my having worried that the southward push from the NAM was a result of modeled cold pooling associated with very vigorous convective initiation forecast along the boundary. The progressive easterly motion of the mid-level low not far behind the surface low, also noted by Dick, probably had a role in the isallobaric forcing and frontal evolution. In the loose analog case of 10-26-06, the cyclone did not appear to be as vertically stacked, probably favoring the more stationary nature of the occluded boundary.

Wednesday, April 8, 2009

April 9 forecast

The 00Z models show a slight southward shift from previous runs, which is a real downer given the poor chase territory over much of northeast OK. I think the potential for a pseudo cold core play looks better in some regards on the 00Z NAM, as previous runs had 35-55 kt 500mb flow (southeast of the mid-level low) overspreading the surface low and adjacent warm front. The 25-35 kt 500mb flow forecast off the 00Z run looks better for keeping storm motions slower (lessening the risk of them either weakening or quickly becoming elevated), and is in the range consistent with a lot of other cold core tornado events (particularly those tied directly to the cool side of a baroclinic boundary). I'm a little leary of sticking close to the surface low tomorrow though, given the progressive nature of the low and strong low-level cold advection forecast in its wake. I'm also not sure how "tornado friendly" the cool side of the effective warm front will be, given the rather late arrival of the substantial low-level moisture (if it in fact arrives as forecast) and 700-500mb temperatures that are not especially impressive for an early April cold core event. My initial hunch is to play the northern end of the warm sector, east of the low a bit (say, the far eastern KS-OK border), and hope something goes up on the northern sections of the dryline and gets its act together in far northeast OK or far southeast KS before outrunning the narrow instability axis. It is certainly possible that a better cold core play might evolve farther west, tied to the surface low/warm front intersection, and surface obs and visible satellite later on tomorrow will help with figuring that out. If we can't get upper 50s dews up toward the KS-OK border before dark, the sfc low target (w/ the coldest temps aloft and strong PVA) may be the only viable option.

The just-in 00Z NAM-WRF-NMM 4km WRF from Matt Pyle has no trouble initiating an arc of storms across northeast OK between 5 and 7 P.M. local time, which is interesting.

This would be a much tougher forecast (from a chase strategy standpiont) if it was occurring smack dab in the middle of KS. As it is... I refuse to mess with "Ozark" country (especially around, if not after, sunset), where parameters are more typically "stacked up" ... i.e., stronger deep layer shear, stronger moisture/CAPE, and thus bigger EHI/STP values than points farther northwest. Thus, the cold core or pseudo cold core (i.e. open warm sector ESE of the sfc low) looks like the most practical target at this time.