Average 20 year old weather nerd. Plymouth State University Meteorology, Class of 2018. NOAA Hollings scholar. Summer 2016 intern at NWS Boston.
By: MAweatherboy1, 11:51 PM GMT on May 17, 2015
Good evening. Today's blog is a bit of a new one for me, a review of yesterday's severe weather event. In their initial Day 1 outlook, the SPC issued a large moderate risk area, spanning through southern Nebraska, central Kansas, and central Oklahoma, with adjacent enhanced risk areas.
Figure 1: SPC Day 1 Convective outlook for 5/16, valid at 12z.
The entire moderate risk was associated with 15% hatched tornado probabilities, a usual indicator of a significant tornado outbreak. As with most severe weather events, there was considerable uncertainty regarding how exactly the day's threat would unfold. Many factors must come together to allow for a tornado outbreak to occur; storms need large amounts of instability, or CAPE (Convective Available Potential Energy), copious atmospheric moisture, and strong wind shear, with the veering winds allowing storms to spin. Yesterday had most of these ingredients, with a fairly strong surface low pressure moving through the Midwest. A large, moisture rich warm sector became well established in association with this system, and upper level wind profiles appeared very favorable for strong, long lived supercells capable of producing strong tornadoes, large hail, and damaging wind. So what was the result of yesterday's event? Below is the SPC's 1630z convective outlook, which was similar to the 12z and 13z outlooks. Overlayed are all the preliminary storm reports relayed to the SPC for the day, with a legend in the bottom right of the image.
Figure 2: SPC Preliminary 1630z Day 1 outlook verification.
A few things stand out in the above image. Clearly, there was a good deal of severe weather yesterday. 211 storm reports is not an average day. Interestingly, however, nearly all of the reports occurred outside the Moderate Risk area. You can draw your own conclusions, however I would say that the SPC forecast ultimately performed rather poorly. Or at the least, it was not up to the usual standard of what we see on these verification reports. I mention this not as a knock on the SPC. They are top notch forecasters, working tirelessly to improve public safety through their accurate predictions. But precisely forecasting every weather event is an impossible task. The atmosphere is too complex and we do not understand nearly enough about it. I am sure most of the SPC meteorologists would not consider this a great forecast. There are some upsides to a forecast like this, however. No two days of weather, no two storms are ever identical, but similar setups can show up again year after year, and so called "analog" forecasts can be used to help get an idea of how a day's event will go. Recognizing mistakes in one forecast can help improve another. So what went wrong yesterday?
Around mid-morning yesterday, I made a comment on the main blog indicating I thought we were in for another "bust" severe weather event. Indeed, this year and last have seen several potential significant severe weather outbreaks turn out much less impressive than what they were forecast or what some aspects of the atmosphere would have allowed for. Bust is a somewhat subjective term, and you can argue that yesterday wasn't a bust, or at least a total bust. But it did not live up to its potential, nor were the majority of the storm reports in the areas they were expected to be in. From what I saw watching the radar for the better part of yesterday, and what is known about how thunderstorm development and maintenance happens, I can pick out two main reasons for yesterday's forecast problems, with each one impacting a different area. Let's look at the northern portion of the Moderate Risk first, from the bottom of Kansas northward. There's no question, this was a complete bust. Hardly a single storm report came in for that region of Moderate risk in the 1630z outlook. This bust is actually pretty easy to explain. The key has to do with what happens before the main show, in the morning hours. Often times, preceding an afternoon severe weather event, a region will receive morning convection- showers and thunderstorms, which can be strong and even severe, but which rarely produce a widespread, significant severe weather outbreak. These storms can actually serve a hindering purpose for a potential outbreak. When convection and/or low cloudiness lingers too long into the late morning or even early afternoon, it stabilizes the atmosphere, and can prevent additional thunderstorms from developing or strengthening in the mid-late afternoon, which is the most diurnally favorable time of day for storm formation given strong solar heating, which destabilizes the atmosphere. This lack of "recovery" of the atmosphere can prevent an outbreak even when favorable upper dynamics exist. Not all days are like this; sometimes there is no morning convection. And sometimes there is morning convection, but it moves out early enough to allow for plenty of hours of sunshine to once again destabilize the atmosphere. Indeed, many of the greatest outbreaks occur on days like these. But yesterday falls into the first category for the northern portion of the Moderate risk. Significant convection and cloud cover was in place for the entirety of the morning and into the afternoon across northern Oklahoma, Kansas, and southern Nebraska. In the 24 hour radar loop from the Wichita, KS linked below, notice how convection was nearly constant during the day over the region.
The lack of any significant break in convection meant that the northern portion of the moderate risk saw plenty of rain and storms, some strong to severe, but very little in the way of supercells that could produce tornadoes. You can see a line of storms strengthening as it pushes off to the east at the end of the loop, and this produced some damaging wind and a couple possible tornadoes in the slight and enhanced risk areas east of the moderate risk. By and large, however, the ultimate result of most of the storms in this area was the prevention of stronger storms, and no apparent tornadoes touched down in Kansas yesterday, a major forecast bust. The SPC did eventually recognize this. The 20z outlook update significantly reduced the size of the moderate risk, cutting it down to only the southern portions of the previous area. This was more of a "nowcast" than a forecast, with the SPC noting the following in their discussion, citing stabilizing morning convection and additional disruptive convection to the south:
"PERHAPS THE MOST NOTABLE ADJUSTMENTS HAVE BEEN
MADE ACROSS PORTIONS OF THE CENTRAL PLAINS...WHERE THE ENVIRONMENT
HAS BEEN IMPACTED BY CONSIDERABLE EARLY PERIOD CONVECTION. HIGHEST
SEVERE STORM PROBABILITIES SEEM LIKELY TO BE FOCUSED ALONG A
CONVECTIVELY GENERATED SURFACE BOUNDARY NOW /ROUGHLY/ EXTENDING
SOUTH OF WICHITA KS THROUGH THE ENID AND CLINTON-SHERMAN OKLAHOMA
AREAS. THE BOUNDARY LAYER ALONG THE DRYLINE ACROSS WESTERN KANSAS
IS RECEIVING CONSIDERABLE INSOLATION...BUT MOISTURE RETURN HAS BEEN
DISRUPTED BY CONVECTION TO SOUTH."
Below is the 20z categorical SPC outlook, with storm reports for 20z and beyond overlayed. Note the downgrade of the previous norther portion of the moderate risk to "enhanced" risk, which still proved to be too high of a risk for what evolved.
Figure 3: SPC Preliminary 2000z Day 1 outlook verification.
So in large part, excessive morning convection is what caused a major severe weather bust for much of Kansas, Nebraska, and northern Oklahoma. Now onto to areas farther south. This portion of the forecast was better than what transpired farther north, but it still left plenty of room for improvement. Looking at the verification map above, you can see much of what was left of the moderate risk at 20z still received little to nothing in the way of severe weather. Numerous severe storm reports, including tornadoes, are clustered in the southwest portion of the moderate risk area, as well as adjacent portions of the enhanced risk area. Why did this happen? The best way that I can put it is that the atmosphere was almost too favorable for storm development. In these areas farther south and west, morning convection moved out in plenty of time for later, stronger storms to form. In most tornado outbreaks, storm formation takes time. Diurnal instability must build up enough for storms to finally overcome a so-called "cap" that can be located above the surface level. Bob Henson discussed this in a recent entry on the main blog, saying, "During some years, spring thunderstorms are strongly “capped” across the Southern Plains by a recurrent layer of very warm air one to two miles above the surface that moves across the region from New Mexico and west Texas. This spring, a strong low-latitude jet stream (likely aided by El Niño) has led to cool and wet conditions across NM and west TX and cut back on the strength of the cap. This allows showers and storms to develop more readily over the Plains, but it also hinders the buildup of explosive instability that can enhance the severity of late-day storms." You can read the full entry here. Caps are often thought of as the enemy of severe thunderstorms, a warm, stable layer that those big cumulus clouds just can't seem to bust through. The truth is, though, caps are, at least in my view, another of those key ingredients in severe thunderstorm formation. I think this is something that can be undervalued at times in forecasts. Yesterday had plenty of instability build across northern Texas and into southwest Oklahoma, but there was no resistance for the storms to form. Linked below is a 24 hour radar loop from the Lubbock, TX radar site. The speed of the loop makes it a bit tough to see, but you can observe supercell thunderstorms develop by very early afternoon, which is quite early for long lived tornadic supercells to be developing. Additional supercells developed farther south later on, and did produce numerous severe hail and wind reports, but were removed from the most favorable environment for tornadoes.
These supercells were indeed the big severe weather producers for the day, as seen in the earlier storm report maps. But they did not produce a major outbreak, for a couple of reasons. For one, the best upper level support was farther northeast, in the moderate risk area, as expected. If storms had waited longer to initiate, they would have had a better opportunity to take advantage of this favorable setup. In addition, the sheer number of storms, some supercells and some line segments, that developed precluded any chance of a major outbreak. By the time these storms reached the most favorable area, they had congealed into a mess of mostly linear storms, with a low threat for tornadoes, since most tornadoes, and nearly all intense ones, form from supercells, not linear structured storms. These linear forms do carry a threat of damaging straightline winds, and indeed parts of northeast Oklahoma and western Missouri reported numerous severe wind incidents associated with lines of thunderstorms.
We did get one clear case of what yesterday was capable of. One very strong tornado touched down near Elmer, Oklahoma, associated with a strong supercell that had formed further southwest. This was a long tracked tornado, with some of the more impressive radar velocities observed in recent years, and a large wedge tornado was well observed by several storm chasers in the area. This tornado has not yet been given a rating. Luckily, the areas impacted were sparsely populated, and damage was minimal. This may also contribute to a tornado rating lower than what the actual intensity was, given a lack of hard evidence (damage) to rate the tornado stronger than an EF2 or EF3, despite radar velocities arguing for an EF4 or EF5. This tornado was only a preview, however, of a movie we never got to see. This supercell was only starting to move into the most favorable region for storms before the excessive amount of convection around it began to choke it off. Indeed, the worst case scenario for yesterday, and perhaps one of the more likely scenarios, was for half a dozen storms like that to form and tear through the moderate risk area, with several strong tornadoes possible. If there had been more capping, if storms had initiated a little later, a little farther east, it could've been a whole different story. Below is one more linked radar loop, from the Frederick, OK radar yesterday. Given the high speed and one hour increments, you can't quite make out the previously mentioned violent tornado producing supercell, but what you can see is several supercells emerging from the west and merging into a line, a nasty line of storms but not one likely to produce strong tornadoes.
Figure 4: A large wedge tornado, with a bolt of lighting, near Elmer, OK yesterday.
Weather is just so interesting, you could do a case study like this on pretty much any day, and in much more detail than what I've gone into. However, given that I stuck my neck out a little yesterday and was ultimately very accurate in my thoughts, I thought I'd go a little deeper on what exactly I was thinking yesterday, and why the day turned out as it did. I hope you've enjoyed the entry, and feel free to leave comments, questions, and your own thoughts below. I hope to make some more entries in the coming months now that I'm off school for the summer, on tropical and/or severe weather. That's all for today though; have a great week!
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.