Now that a few days have passed, and I have had time to collect my thoughts, I wanted to explain my feelings about this day and thoughts about its implications (My chase log will appear separately).
So here goes:
I first met Tim at a derelict gas station in Nebraska, and was struck by how easy going, approachable and honest he was. We had a very interesting chat about the storm that had failed to produce a tornado on the day, and talked shop about storm environments - like so many of us do when we are out on the plains. But Tim was different - when someone asked him a question they were never made to feel below him, in fact they left the conversation with a sense of awe and excitement. That is a rare gift to have, and it was far from his only talent - his wizardry for gadgetry and the design of probes for in situ tornado measurements was simply astounding. This would not be the last time I spoke to him either, from having a chat at a Denny's eating breakfast at 1am after a successful chase, and other times that were as simple as a wave or a polite hello. The last time I saw Tim was on the dry-line last May watching 2012 produce yet another disappointment (which was later followed by the fail-recho). Tim was one of those who gave a good name to chasing - a gentleman chaser if ever there was. The way he modestly handled the public spotlight, and gave his all to promoting the science while maintaining the safety of his team had a deep impact on me and how I thought about chasing. One only need to watch an episode of storm chasers or a National Geographic special to see that he always put safety first, was level headed and a master of his craft. So why have I told you all this? I don't claim to be a close friend, or someone who knew Tim well. However, I considered Tim a role model, one of those that I could aspire to be like in the way that he never sought the spotlight, never put those he chased with in danger, and someone that could tie the experience of chasing back into the scientific and public domains. His ability to communicate to people at any level of knowledge is something that I think so many of us can learn from - talking to children in schools, the general public in libraries - the next generation of researchers, chasers and the future of public safety depends on this pursuit. I can honestly say that the world of research storm chasing has lost a champion who will be sorely missed.
I never had the pleasure of meeting Paul, but from every account he was a lovely person with a funny side, and from every photo that I have seen I think its safe to say an incredibly talented photographer. My thoughts and heartfelt sorrow go out to the Samaras family - to lose one is a terrible heartache, to lose two generations simultaneously is a sorrow that I can not even bear to imagine.
I only ever met Carl the once (well at least chatting to him directly), on April 8th this year at breakfast at a hotel in Great Bend. It was an interesting discussion on the day before, but I certainly would not say I had enough time to get to know him. Again, his reputation in many circles speaks volumes about the man, and he too will be sorely missed by chasers and scientists alike.
Now I would like to change tack, and look at the implications for chasing.
(Note for those who want to skip a lengthy discussion of the environmental parameters skip to the the next bold underlined text).
(Note 2: This is an expression of my interpretation, and others who are more familiar with significant or violent tornado environments may come to different conclusions.)
So what happened on Friday - how is it possible that four storm chasers were killed, after no deaths for what has been 40 to 50 years of people being out there? This troubles me, as Tim was so mindful of his surroundings and remarkably cool under the pressure of approaching tornadoes having been in the situation so often before. The environmental parameters that we saw last Friday were simply frightening and remarkable; The OUN (Norman, Oklahoma) sounding for 00Z shows Convective Available Potential Energy (CAPE) likely above 4000 J/kg before storm modification, storm relative helicity (SRH) in the lowest atmospheric layer (0-1km) in the 300 J/kg range, lifted condensation levels (LCLs) that were below 750m prior to storm formation. In this environment the threat for the rapid development of a tornado is incredibly high, and given the low cloud bases large tornadoes become more likely.
And yet this likely doesn't just encapsulate how volatile the environment was - so why is this the case? Well, surface temperatures that I observed at the lake in west El Reno prior to initiation were closer to 90F, while surface dewpoints in the Chickasha area prior to initiation were showing 77F on the mesonet. This yields a higher CAPE than the observed value at OUN 00Z - probably on the order of 5000 J/kg or greater effective (and this can be seen in the most unstable parcel for the 18Z sounding). Next, consider the influence of the storms motion - the storm during its early parts and initial tornado production was moving southeast - from an environmental helicity point of view this lengthens the hodograph substantially. Now, combine this with the east-southeast surface flow that could be found near the boundary of the triple point and we have SREH (Storm relative effective helicity) well into the "what the heck does that even mean" territory (300-500 or even more). Even in the profile above, the significant tornado parameter (STP) is in excess of 7, and the supercell composite parameter (SCP) is near 29 - NWS mesoanalysis paints these values at the time of both these values in vicinity of northern Oklahoma city and further westward towards the dryline/boundary intersection - possibly these were even higher by the time the storm began modifying the environment. While the 00Z 0-3 kilometer CAPE (a favored tornado metric) does not appear to be very large (45 J/kg) forecast values were suggesting this would move into the realms of 100 J/kg - certainly more than sufficient for significant tornadoes given the other parameters. Finally I would note that the deep layer shear is not very strong for the large amount of moisture, and given the relatively modest flow at 500-300mb, this suggested HP modes would be the end result.
Adding this crazy soup of parameters together indicates a few things - but lets compare it to the climatology. If you consider the distribution of significant tornadoes in the recent paper by Grams et al. (2012), surface dewpoints were above the 90th percentile for significant tornado events, and for surface temperatures - the resultant LCL height reveals that neither of these factors was a negative contribution, and only served to increase the vertical gradient and hence thermodynamic energy. LCL heights were low - certainly below the 75th percentile of the Grams climatology for the southern plains, while mixed-layer CAPE was above the 90th percentile - more like an environment in the Northern plains during the summer. 0-1km shear from the OUN sounding was near to the mean for the significant tornado sample, but it is likely that this is somewhat under-representative in the effective environment. Lets look relative to the local climatology for significant tornadoes:
|Source: Screenshot of the NWS Mesoanalysis tornado segments for the vicinity of El Reno, OK, this is the product of work by Bryan Smith, Rich Thompson, Roger Edwards and others which can be found in a 3-part paper in Weather and Forecasting, 2012.|
Long story short, its pretty easy to see that we are looking at a type of environment that doesn't occur very often. We have very high available CAPE, low LCLs and significant wind shear and helicity in the near surface layers. Most of us haven't chased many environments like this - so lets think about potential implications to our storm mode and tornadic behavior:
1. Tornadoes in this environment are likely to develop quickly, be strong and have multiple vortex characteristics before wedging out. Here's a couple of events that spring to mind where the parameter space was pushed to the upper right of the distribution in terms of say helicity or CAPE: Joplin MO, Plainview IL. In each case you ended up with violent tornadoes with unstable or quickly changing vortex structures that develop rapidly. I'm not saying these are great analogues, but more just a couple of cases where I can think of similar tornado behavior.
2. With LCLs so low, you have the potential for the mesocyclone or tornado cyclone to potentially situate itself very close to the ground as a response to the rapid up-motion - with high values of environmental helicity meaning that spinups and strong vortices are likely, especially with such a strong mesocyclone. Consider as an example the Greensburg KS 2007 storm and its behaviour after dusk, or Hallam NE. There are probably other examples but I will leave those to people who are more familiar. The moisture that produces these low LCLs and the weak mid level shear tends to also push the mode towards the hybrid HP/classic structure, where the mesocyclone is still located in the SW quadrant of the storm rather than the east.
But what does this mean for those who chase?
Storm motion from this profile will likely be pretty much to the east-southeast for right moving supercells - and I think many of us worked this out fairly quickly. Occlusion of the mesocyclone will be to the left of path - again pretty stock standard and expected. So where did we go wrong ? (I say we as even though I kept a respectful distance, I still found myself running rapidly east during the expansion phase to avoid the southeastward rear-flank downdraft surge and the tornadic circulation, though I was fortunately far enough south to stay clear of the later).
This opinion may not be the most popular, but playing the notch is like playing russian roulette - there is eventually going to be a time when the bullet gets you. I've held this opinion for some time (after nearly coming unstuck one day in Nebraska in a sparse road network), and in HP modes this tends to keep me at the very least further east, and more often southeast. In the past few years, we've seen time and again storm chasers who came close to being unstuck - a banged ego here, a shattered windshield there. But what happens on the day that you don't see the monster coming? People are no longer even chasing just the notch, but in under the mesocyclonic circulation - and in this case the tornado-cyclone. So in a low LCL environment, with a bucketload of helicity and no shortage of CAPE in the lowest layer - is that the right place to be? I'm not trying to tell people how to chase, just raise the issues and implications that we have to deal with after this event.
Now - in the case of the TWISTEX team, I don't have a single problem with them being there - this was what they did, and did so effectively for many years. I think (and this is a personal opinion only) some series of unfortunate events or timing (for example deploying probes, which is when they are at their least mobile) during the rapid expansion of the tornado-cyclone to the surface and turning of the tornado with acceleration northeast likely caught them out - with subvortices that had forward motions of 150 mph as documented by the XPol radar, they may not have even seen it coming, or were unable to do anything to get away. It is a testament to their mission and skill that they likely managed to deploy probes in the path of this tornado - a tragic but hopefully lasting legacy to tornado science that befits three of the great chasers.
For other chasers - maybe we need to reassess how we approach storms that are way out there on the scale of environmental parameters and question whether you want that particular tornado to be your last. The huge width of this tornado was definitely a surprise to us all as was the rapid movement, and reflected the combination of an incredibly strong mesocyclone and the available environmental energy and helicity. As Jon Davies pointed out, and from my viewpoint was also observable on the day - the tornado did not appear to be at the center of the mesocyclone, but rather closer to the periphery - which perhaps explains the increase in forward motion and change in direction to NE at 40mph when combined with the RFD surge to the southeast.
In a different set of circumstances - this could have been much, much worse - the secondary mesocyclonic signature could have also produced a strong tornado tracking southeast into the panicked masses - leaving storm-chasers with no escape. Or perhaps the primary mesocyclone could have continued moving southeast at increased speed overtaking more chasers as it expanded, fortunately the RFD surge probably prevented this. My girlfriend/navigator was in the right hand seat on the day, and showed me the horrific image of spotter-network icons that were either in the path or very close to the very strong mesocyclone - and I was terrified that we would lose many more than we did.
So where does this leave me - I don't think it will stop me chasing, as many proved that this circulation could be safely navigated and documented with incredible images and data coming to light. However, I think that I will certainly think twice about what the environment means on a particular day before chasing - a lesson that I think would serve many aspiring chasers to the field well. I extend my deepest sympathies to the families of all the victims, both non-chasers and chasers for your loss, they will be missed by all. But as a final remark, here is a plea to the chasing community:
'If you value Tim, Paul and Carl's memories, then think more about the implications of the environment on how you chase, how it reflects on each of us when all is done and dusted - Tim, Paul and Carl were always very conscious of chasing safe. Please don't sacrifice your life in pursuit of that one tornado shot - it is never worth it, the chaser community needs no more funerals.'
This post was written by John Allen, and all the opinions and speculation are his own and do not reflect those of others involved in the Hunters of Thunder.
Weather and Forecasting 27:1, 106-123, , , , , . (2012) A Climatology and Comparison of Parameters for Significant Tornado Events in the United States.
Smith, Bryan T., Richard L. Thompson, Jeremy S. Grams, Chris Broyles, Harold E. Brooks, 2012: Convective Modes for Significant Severe Thunderstorms in the Contiguous United States. Part I: Storm Classification and Climatology. Wea. Forecasting, 27, 1114–1135. doi: http://dx.doi.org/10.1175/WAF-D-11-00115.1
Weather and Forecasting 27:5, 1136-1154 , , , , . (2012) Convective Modes for Significant Severe Thunderstorms in the Contiguous United States. Part II: Supercell and QLCS Tornado Environments.