Wednesday, May 19, 2010

post storm analysis

Well my gut telling me that south central OK would be the best play wasn't quite correct.

The second initiation of convection to the south near Lawton just took forever to congeal into a supercell after many many mergers of smaller storms. The hint is that the cap was too strong, or the shear was too strong for the minimal instability being released. This would suggest that the stronger dynamical forcing to the north was a key factor in the rapid transition from small cells into bigger supercells.

The shear though was also questionable at FWD , despite 200 SRH, only the lowest km had significant hodograph curvature. aloft the hodograph was a straight line. At OUN, within some sort of convection, the hodograph was much larger and curvature much more pronounced.

The observed 500 hPa frontal zone was pretty much in contact with the main supercell of the day and had a pretty good connection to the initiation in southern OK. V2 soundings might shed light on this .

Overall I am not sure what to think of an event like this. The soundings to the north in OK appeared to be uncapped and the shear great. Tornado production was similar to the day before ... apparently present but most short lived. Moisture was present, SRH was high, BRNSHR was in the 10-80 range, and the hodographs all had good curvature.

To the south, there was a cap, perhaps weaker forcing until much later, hodographs were curved by trended to be more straightline (window of opportunity was shorter to the south perhaps), but the instability was greater. I did notice that the southern supercells, had outflow attached to them. So perhaps the mid level dryness enhanced cold pool strength. This case has lots of interesting questions.

dynamical devil

The dryline bulge is in western Oklahoma somewhat removed from the 500 hPa frontal zone as seen in water vapor imagery. However the circulation about this front has resulted (or speculated to have resulted in a narrow moist tongue over the proximate location of the dryline in OK.

I expect CI to be imminent, and perusal of vis sat imagery confirms the presence of a cumulus field in western OK. Not sure if the remnant outflow boundary is really a front or if the nocturnal MCS simply cooled and clouded over the northern portion of the front. Thus, not sure if this front is very active....or if it will truly be a focus for CI.

Vis imagery wave clouds would indicate that it is still relatively stable in Central and southern OK.

First echos on the dryline at 18:56 strengthened by 19:14 UTC. Looks like convection is right on time.

The devil is in the dynamics

A nice little potential outbreak today for OK and North TX. Complicated situation as the night time MCS moves away from OKC, with no outflow boundary clearly visible. The low level cloud field is eerily similar to previous days. The complications are that the elevated mixed layer is well into the area, and the 500 hPa frontal zone is already through AMA. Surface convergence will be there to the north but it may be the dynamics that sets this thing off.

The cap is the big player and the models favor this opinion. NCAR ARW run from last night says multiple periods of initiation to the north, remniscent of May 10, again occurring earlier in the 18-19 UTC time frame. Then later on initiation occurs along a dryline boundary (around 22 UTC) before growing upscale. NSSL ARW run is a lot different probably because of the poor precipitation forecast it produced to the west of the MCS. So too much precipitation to the west virtually ensures that a subsequent trigger for new convection occurs earlier.

Both models agree that initial development will be scattered but not isolated. Thus storm interactions and outflow development may hinder chances at discreet, cyclic tornadic supercells. Today may be a tail-end Charlie kind of day.

The models are not very different with the dryline, but it appears the 500 hPa front is much stronger in NCARs run...though i have no proof since the NSSL run does not make available fields aloft...however the CI that occurs in NCARs run appears associated with the 500 hPa frontal circulation. This is inconclusive however because the 850 RH field is near saturated as well. Breaking the cap is easier when the PBL top becomes saturated.

If I were chasing today, I would try to be south of OKC, probably in Norman. Wait to see how the cap holds to the north and if early storms stay isolated and discreet go for them. Then haul ass south later in the day as the next wave of CI occurs along the 500 hPa front. Otherwise south and east of Norman. The show should start by 19-20 UTC either way...3 hours from now.

Not sure of the tornado potential. Yesterday appeared to be ripe for tornados, yet the pictures and reports indicate short lived but strong tornados. Nice supercells on radar, though. wonder if the shear profiles were not ideal enough. The shear profiles from NCARs model at OUN at 21-23 UTC appear good enough, so this may be a right time right place scenario...brief window is implied. To the north it is less certain since storms may be on the warm front. I am still hedging my bets southward of OKC ...

Wednesday, May 12, 2010

warm front MCS

So I fell asleep last night and missed out on real time watching of the Mesoscale Convective Systems that formed. May 11-12 2010 .

I am always fascinated by the different evolutionary modes of convective initiation. Yesterday's mode was scattered convection under a moderate shear-high instability day. But it was effectively capped, given that multiple elevated supercells formed and quickly dissipated much to the chagrin of VORTEX2 participants. It was a classic Iowa warm front event, with little to no forcing but shear and instability abound.

at 2200 UTC, the first supercell was in the process of initiating in SW OK and after 1 hour falls apart. It skirts an area to the north just south of woodward I believe, and another set of CI takes place between 2345 and 0015 UTC. Again short lived supercells. However, the anvil aloft is cold and the cloud does not dissipate. So it looks like a storm is there, but on radar it was a useless blob of 30 dbz echo.

By 0145 the warm front in SE KS gets hot and explodes into the primary MCS. This initial convection gets going and after perhaps 45 minutes to an hour backbuilds to the west along the front while that old cloud shield drifts in by 0402. Not much longer later the old cloud shield is rapidly dissipated around the MCS as it swells to maximum size virtually covering MO by 1100 UTC.

at 0345 UTC SW KS sees another MCS initiate and suddenly from KS into NE convection is firing up. This MCS looks asymmetric almost half circle like but looks very similar to an MCC. Once these two systems get close enough to each other the eastern MCS appears to dissipate, at least its western end. The radar reflectivity pattern was chaotic in nature. This is a different breed of cloud system most notably MCC like. It doesnt come with much severe weather oddly enough given its environment. But the cloud shields are huge.

After looping tonights water vapor imagery .... the eastern MCS traveled from IL through VA out over the Atlantic. I noticed a pretty compelling anticyclonic gyre aloft. very classic long lived system.

Also note that CI occurred on both sides of the water vapor dry slot ... to the east failed and to the west got going.

The surface data doesn't look that compelling .. where the warm front convergence was occuring in SE KS is where CI took place along the front. beyond that moisture was slowly working north (implied moisture convergence). Both systems were on the anticyclonic shear side of the upper level jet which was just by position only since the wind field aloft was pretty homogenous.

Soundings from LMN indicated the cap had weakened boasting CAPE of between 2-3000 J kg-1 at 2100 UTC. The wind sbacked at 0000 UTC and the low levels became saturated as the cap layer seemed to strengthen a bit. though perhaps the lapse rate steepened under slight cooling aloft. 6 hours later the layer below 850 had a nasty 50 kt LLJ and warmed and moistened appreciably. moisture depth to the east was high, as the low levels appeared to be saturated.

visible imagery indicated some sort of cloud boundaries, like elevated outflow boundaries that might be implicated here. But I dont have access to the OK mesonet data to see if anything showed up there. In my experience it doesnt show up on the sfc winds, but it might be in the pressure field.

Interesting day. Maybe V2 will have some photogrammetry of these supercell initiations.

Monday, May 10, 2010

Post storm wrap up

Wild day. 37 reported tornados. maximum hail size of 4.25". Max reported wind gust to 100 mph.
Looks like a number of fatalities have occurred as well.

Well forecast event by SPC. The uncertainty was well forecast at 1200 and 1300 Day 1 outlooks:
1. Initial storms would take time to organize thus Initiation area would be removed from maximum tornado risk.
2. moisture return would not be an issue.

BUT, the 1630 UTC update suggested that 21 UTC would be the new initiation time. Doubt crept in. Even to go as far as to say that central and southern OK would be capped and that models were over-aggressive. The 2000 UTC update re-instated the aggressive coverage southward based on the satellite data: cumulus cloud field developed in advance of and in proximity to the dryline.

What was not well forecast was the timing of CI. SPC was calling for a 18-19 UTC CI in SW KS, with another bout further south into OK by 21-22 UTC then further south into central and southern OK. The focus was on the cap and favorable moisture return.

Indeed the actual event sequence began like SPC said: SW KS near DDC for 2nd and 3rd storms in a bunch at 18 UTC. The storms appeared further SE into OK to the northwest of WWR by 1850 UTC and another east southeast of Gage and a few weak storms further south. These storms never really got going though they were players later when they merged with the northern storm ... that storm became tornadic after those mergers though it isnt clear they were the cause.

Now heres where the forecast goes awry. The dryline has not been dogged by storms. It was still going strong. At 2000 UTC a weak storm complex was near JWG. It almost dies, then out of nowhere re-intensifies into a nasty linear looking supercell. Much like the Nov 2005 Ames tornadic supercell or really much like high shear supercells.

By 2100 UTC a new batch of weak storms in advance of the dryline form and these quickly become supercellular in a cluster. The dryline is far removed from these storms. Much further south more storms get going, more likely previously involved storms with the dryline down there. Between these 2 areas another loan supercell formed and merged with the OUN cluster. This rogue supercell seemingly merged into the southern side of the northern supercells and that was in Norman or close by. differential storm motion for some reason on that storm.

New storms form off the dryline by JWG and move northeast to form a more linear looking cluster. By 2142 more storms form on the dryline to the west of OUN. This cluster appeared to form a line before devolving or maybe evolving into nasty hook echo supercells. Quite spectacular.

The supercells came in waves, seperated by time and/or space in clusters. They were at times poorly spaced but productive, and well spaced and spectacular. We saw the whole range of high shear supercells. More importantly we saw that the model forecast paradigm broke down. The soundings launched by vortex2 will be invaluable to assessing the environment on this day. This will be one of the most studied days in VORTEX2.

Big cities under the gun

Wichita and Oklahoma city are under tornado warnings and basically in between. A number of large tornado reports have come out, amidst the very brief 30s to 3 minute variety. The supercells have shown all kinds of high shear characteristics today.

The most crazy is a left mover split that eventually produced tornado warnings in kansas off the woodward storm. It crept in behind another storm, and then that one went tornadic as well. The woodward or likely a variant of it, because it merged with at least two other storms (of linear shape and very weak reflectivity speaking).

and the reports are coming out fast. You actually have to wonder what the impact of having a great weather dept. at OU has done for safety in Oklahoma. If there are storms you can bet like 30-60 chasers will be out combing the plains for tornados, and some of them will be relaying information back to the NWS. This doesnt prevent damage, but it should make people more aware of tornados and there warnings.

and the tornado warnings come out

Couple of reported tornados in what looks like a set of massive circular blobs which was the most continuous feature associated with the shortwave trough.

DDC initial storm just got its act together. Possible hook and strong rotation noted on radar. It appeared the evolution of such storms were very typical of high shear environments ... stretched out along the mean wind until repeated bouts of deep convection upshear culminated in protecting the storm.

This is also the case for the woodward storm though it got going later and split with the left mover not moving very far off yet. The main tornadic portion of the cell appeared to be linear! it now has a kidney bean shape with reflectivity over 65.

Situation is still somewhat chaotic between these two nice storms. no warnings and there appears to be no organization with them which is strange.

Dew points up to 68-70 in OK just west and south of OKC. Despite that the storms south of the woddward storm just cant get going. They are struggling. would love to see the sounding from OUN, as there might be a stronger cap in place which would protect OUN till later tonight.

FWD sounding was strongly capped. so was DDC though I will have to review the radar data at 1715 to see which air mass was being sampled, environment of low level storm.

Nice outflow in CO.
1st big storm went up by DDC in advance of the dryline along a cloud street. ci took 20-25 minutes. appears to have a second storm on its flank over the same initiation area. This area is where the winds have stayed from the southeast with dew points near 61 in DDC.

Storms that were ongoing associated with the low actually became tor warned first. DDC storms and initiation zippered along the dryline south still not doing much.

Cumulus cloud field sprang up in OK in apparent advance of the dryline where temperatures have gotten to 90 F with dew points into 20's. The structure of the cloud fields kind of resembles the model bulges in dryline position especially the NCAR ARW run from 00 UTC last night valid at 18 UTC. Indeed the OKMESONET says the dryline is in OK...what a surge that was...but probably not given the station spacing down there.

Northeast storm motion on these initial cells especially down into the woodward region. I am starting to think that a round 2 may be likely as the drying evident in water vapor imagery over the western TX panhandle increases and indicates that is the base of the shortwave trough. in advance of this feature is a secondary line of moisture evident on vis imagery which I believed to be part of the main dryline, but it may behind the surface dryline and may overtake the dryline in a few hours. This paradigm indicates a 2nd bout of CI as the forcing coincides with the dryline boundary...I am out on a limb here.

afternoon update

Indeed the severe weather risk areas migrate towards SW KS in advance of the surface low. Per my thinking in the previous post, storm initiation is indeed being forecast earlier. The dryline is really beefing up with dew points dropping down to near 10 F and surface winds sustained to nearly 35-40 mph with gusts to 50-60.

Current sat trends indicate multiple "cloud streets" though not of the fine scale variety behind the dry line. This indicates to me that low precip supercells may be the initial mode, however the speed at which the shortwave is progressing suggests that by the time CI does occur, they may be right in the moisture plume which argues against the low precip mode.

However, the linear banding of the cloud features at this point points to a more linear development mode. One wonders if this early convection, if it gets going, will be round 1 followed by round 2 of supercells later in the afternoon. Still more cloud lines popping up now near rthe dryline at 1715 UTC.

Waiting for 18 UTC sounding for DDC...maybe another hour or so before it appears on line if they launched at 1710 UTC like I hope they did. I also expect one from AMA at 18 or 21 UTC and one at OUN at 20 or 21 UTC. Hopefully the ARM site at LMN will also launch a few.

Dryline just went by Liberal KS, 82 over 45 this hour, was 75 over 59. Interesting that the dryline appears to be making north easterly gains into KS but not strong easterly surges in the panhandle ... recent trends indicate it is moving but not as fast further to the south. Will have to watch the progression.

Max dewpoints in OK up to 68 in the south, still holding strong at 65 in the plume to the north.

CI iminent along a line from the TX-OK border North to DDC in the next hour and a half over the OK panhandle.

The morning update

Much to my surprise, the moisture is returning with dew points up to 60 in OKC at 15 UTC with a 65 just south into TX in the narrow plume of moisture to the west of FWD. It looks like the moisture is deep at FWD on the sounding from this morning. The shear is already large and there is no reason to suspect it will diminish. The hodographs are large and semi-circular from 0-3km.

Currently though there is a deep low cloud deck with indications of wave clouds. This corresponds nicely with OUN's sounding and low level stable layer. The water vapor imagery continues to indicate a dry intrusion aloft is setting up from MAF to OUN to Tulsa. The dryline itself is already in place at AMA, and should be moving rather quickly with the arrival of the shortwave in the next few hours. Convective initiation should occur according to the data prior to 21 UTC and possibly as early as 19 UTC if I have done my wave phase speed analysis correctly.

Presumed 700 hPa cold advection is already occurring into the TX panhandle. BIG question is when will the cloud cover break up over Oklahoma, and where it will recede long enough for temperatures to climb. If the clouds hold strong, which is in doubt, then where they do break up will determine the locations, at least partly, for CI. Sounds like May 3rd all over again.

East moving storms with right movers to the south-southeast. Gonna be a long day for some folks in Oklahoma. And SPC and I are of the same opinion ... maximum threat over a limited window of time. As per my discussion last night, the rightfully acknowledge the limited span of the best thermodynamics and rightfully acknowledge that the shear may be too strong such that storms may need time to fully organize. Thus the CI forecast location is somewhat removed from the maximum tornado risk area. I still wonder though if the environment is really conducive to long track tornados.

I think the answer is yes, but I wouldnt be surprised if a few storms go up, put down a tornado, then proceed to do nothing more that put down very large hail.

Also the nature of the straightline hodograph over the CI area might be conducive to splitting supercells initially. At least a portion of the northern end of the CI area may rapidly transition to an MCS.

ARW run from NCAR indicates that CI will occur by 20 UTC, and I would guess is probably an hour or two late given the use of the MYJ PBL scheme. But it does indicate that the threat area for more isolated storms along the dryline will be further south into Oklahoma. But once storms like these get going in the model there is no telling how correct it is given the bad mass fluxes at this scale (3 km). Suffice it to say the model is trending towards my line of thinking.

Clouds starting to thin over SW OK ... game on. First batch of CI I forecast for east of Liberal KS. If I were chaing today I would be in Oklahoma City east side waiting for data until 1pm.

Potentially dangerous situation tornado watchboxes should be the norm today.

May 10 the forecast for tornadoes?

It isn't rare to have a day 2 moderate risk for the central plains including OK and KS. It is rare to see a moderate risk during VORTEX2 operations ;) . The prevailing wisdom is that if you put 100 vehicles and 400 scientists and 8 radar trucks in the Plains ...ain't nuthin' going to happen.

So I thought I would get in on the forecast action.

Broad scale specifics from the NAM model: a sfc low races out of CO tonight across southern KS and moves into northeast KS by tomorrow night. This feature is due to a pretty sharp shortwave trough at 500 hPa (about half way up in the atmosphere in terms of mass) which gives the appearance of a negative tilt into tomorrow. Note that the flow around this trough (the jet streak too) is cyclonically curved. The jet streak itself is actually located across OK and TX panhandle. The jet streak at 300 hPa is located further north and there is a hint that the flow aloft will be anticyclonic in nature across southern Oklahoma.

Players in Convective initiation: The warm from will be a breeding ground for all kinds of convection tonight into tomorrow evening. The dryline will be a major player as it punches into Oklahoma City. BUT the shear appears to be oriented perpendicular to the boundaries ... a sign that storms should start and possibly remain discreet.

The odd things I see:
The motion is fast for the shortwave aloft, and thus the moisture return may be in question ... moisture takes time to be transported in great quantities. 0500 UTC data in Texas show 60's creeping into the TX panhandle but dew points only 50 at OKC and DFW. Hi-res NSSL model is consistent through 6 hours with observed data. Models are indicating this to be the case ... that only low to mid 60's with high temperatures near 83 will be possible, though behind the dryline it may hit 90.

NSSL model also indicating the low may stall. Operational models may be aggressive in bringing this system out. Complicating factor is the system to the NW that drops in behind, and system to the east which seems to not move too much. Hard to know since the jet aloft to the rear of the shortwave is not fully ashore.

Day 1 outlook updated to reflect a high risk. Can't argue against that. Models are all producing favorable parameters for Oklahoma tornados. This is easy to forecast maximum risk where all parameters come together nicely. However, as a chaser, where would you want to be knowing:
1. fast storm motions
2. seemingly moisture limited to the east (from the limited model runs I have seen)
3. complicated cloud forecast
4. complicated surface low and dryline evolution and thus placement.

It appears SPC is playing up the moisture (to 70 F) with temps in the mid 80's, and heavily favoring the track of the midlevel jet (rightfully so).

The BIG question is what kind of supercells will be the norm? low precip, high precip, or classic? If low precip come off the dryline how will they evolve in the face of ridiculous shear (40 meters per second) and a somewhat strong cap which is sure to come out given the steady fetch from ABQ?

I have to wonder if there will be early storms in KS that get into OK and further complicate this situation with outflow boundaries. There doesn't appear to be a whole lot of limiting factors the models are picking up on, so the devil is in the details ... morning models will be key as well as the radar and moisture progress on the soundings.

High risk is warranted but there is more going on here. May 3rd 1999 comes to mind in terms of an OKC to Tulsa outbreak. Needless to say I will have the radar looping tomorrow.