Predict/Guess the Number of Tornadoes and the First High Risk of 2016

[Stebo]

Precipitation over the S Plains has been almost nil since Christmastime, and you can see a few splotches of D0 showing up over TX/OK on the drought monitor. It certainly isn't a major concern yet, but it means we have less leeway over the next 30-45 days. If for some reason we remain fairly dry through mid-late March, then I would expect the usual over-mixing issues over the S Plains to commence during storm season until/unless a couple big, wet systems come to the rescue (which they did last year, after a dry winter).

I'm not worried at all yet. Persistent west coast ridging during mid-late winter isn't what I'd like to see if I had the choice, but until March it's not all that relevant (other than starting to run up a short-term precip deficit). It looks like we'll average a slightly -AO for DJF, which in my chase season rankings is a bearish indicator for Plains spring activity, as is the +PDO. These aren't particularly strong correlations, of course.

One minor quibble: I would argue 1988 is a decent analog, as painful as it is to admit. There are probably 4-5 others that are equally strong or better, but still.

1988 isn't a good analog, by this time the drought was already taking shape over the plains and outside a few minor D0 spots in Texas, the plains aren't in a drought. Furthermore the based upon the MJO being progressive and forecast to move into phase 7 at a high amplitude by early March and the evidence in the Euro weeklies of a pattern change which would match the MJO progression as well, I don't think the southern plains will remain dry for much longer.

[Stebo]

Also I would argue that the AO will end up slightly positive for the DJF months especially if the projection for the rest of the month ends up being correct.

 

Dec was very positive, January very negative until the 20th then positive for the last 10 days, this month has only had about 5 days of -AO.

[thewxmann]

I'm sorry if I came on hard but, I just see what you say on twitter and on here and it gets irksome. You don't always have to be conservative especially when things are looking up. The problem is, you always use the worst case scenario for a negative outcome, just like bringing up 87-88, that isn't even like the current year. I mean it just doesn't make sense to constantly have a negative or conservative theme on everything.

 

For the record, I only brought up 1987-88 as a response to the PNA statement, that's it... I wasn't using it as an analog to this year at all. I apologize if it came off that way, I actually have no idea what the best analogs are this year (aside from the Nino years).

 

As for what's on Twitter, those are my unfiltered and uncensored thoughts, most of which I would never post on a weather forum. Of course I'm going to be hating on this pattern. CA vs. the +PNA ridge is like Cam Newton being pummeled by the Denver pass rush in the Super Bowl, of course he's not going to be having happy thoughts.

[brettjrob]

1988 isn't a good analog, by this time the drought was already taking shape over the plains and outside a few minor D0 spots in Texas, the plains aren't in a drought. Furthermore the based upon the MJO being progressive and forecast to move into phase 7 at a high amplitude by early March and the evidence in the Euro weeklies of a pattern change which would match the MJO progression as well, I don't think the southern plains will remain dry for much longer.

 

Despite being by far the "worst" spring on record post-1950, it looks like 1988 actually wasn't that dry, at least until late spring. The reason I say it's a decent analog is that it was a second-year waning strong Nino, had a +PDO that dipped early in the winter and resurged a bit by Feb/Mar, and its DJF H5 height anomalies have some similarities to this winter's. Given that the ONI exceeded 2.0 this winter, I think any spring coming out of a strong Nino has to be given at least a little weight this year.

 

As for the AO: the January mean (-1.45) was the exact negative of the December mean (+1.45), so this month's mean will essentially determine the sign of the DJF value. Looks like the latest ensemble forecasts are a little more optimistic that we'll tilt towards the + side than when I looked a few days ago, but it's kind of a moot point, as the DJF mean should be within the +/-0.25 range for sure. So barring any large errors in the current 7-14 day forecast, any AO signal for the upcoming spring should be quite weak.

[Quincy]

The longest stretch between high risk days (since 1990) was 671 days from 6/5/1999 to 4/6/2001. If there is no high risk issued by April 4th of this year, we'll exceed that stretch.

[Quincy]

We're off to a fast start. Probably the fastest start since 2008. Long-term guidance is also encouraging.

[PGM]

I'll go with 1350 tornadoes and first high risk April 15th. We're off to a good start with over 150 confirmed tornadoes at the end of February!

[Torchageddon]

Number of tornadoes in the US: 1324

First High Risk: April 1

[nrgjeff]

Going 1325 tornadoes with an Upper Midwest surge. First high risk May 4

[Quincy]

Going 1325 tornadoes with an Upper Midwest surge. First high risk May 4

The data I'm looking at also seems to suggest an enhanced potential for the Upper Midwest, perhaps northern Plains too. Hope it's a fun spring and I'm cautiously optimistic that it will be!

[bdgwx]

It looks like we may be approaching one of the longest stretches between high risk convective outlooks. Assuming this week's system does not escalate into a high risk then we may surpass the 671 day stretch from 6/5/1999 to 4/6/2001 next week.

[andyhb]

It looks like we may be approaching one of the longest stretches between high risk convective outlooks. Assuming this week's system does not escalate into a high risk then we may surpass the 671 day stretch from 6/5/1999 to 4/6/2001 next week.

 

The odd thing being that last year was the most active in terms of overall tornado numbers since 2011. Lack of high end setups (or certain things mitigating potentially high end setup, see 5/9 last year) has been really noticeable.

[Quincy]

The odd thing being that last year was the most active in terms of overall tornado numbers since 2011. Lack of high end setups (or certain things mitigating potentially high end setup, see 5/9 last year) has been really noticeable.

The first half of 2014 had three high risk outlooks, but the second half of the year followed close to climo with only a few modest fall events. Not that much different from 2012, but a far cry from 2013, that had a major 11/17 outbreak.

 

2015 sticks out like a sore thumb for significantly above average precipitation across the Plains. It seemed like HP events were strongly favored over LP, with several setups being contaminated by morning convection and/or a messy storm evolution.

 

We need more of a moisture balance if we're going to see a return to higher-end setups. 

[tornadohunter]

The first half of 2014 had three high risk outlooks, but the second half of the year followed close to climo with only a few modest fall events. Not that much different from 2012, but a far cry from 2013, that had a major 11/17 outbreak.

 

2015 sticks out like a sore thumb for significantly above average precipitation across the Plains. It seemed like HP events were strongly favored over LP, with several setups being contaminated by morning convection and/or a messy storm evolution.

 

We need more of a moisture balance if we're going to see a return to higher-end setups. 

 

The odd thing being that last year was the most active in terms of overall tornado numbers since 2011. Lack of high end setups (or certain things mitigating potentially high end setup, see 5/9 last year) has been really noticeable.

 

Who remembers may 16th, 2015? That day could have been a legit high risk had it not been for morning convection

[RCNYILWX]

All the aforementioned issues in reducing the potential ceiling of events is valid, but I think the addition of ENH to the SPC outlooks has also had an impact. In fact, Bill Bunting gave a remote presentation to my WFO (Chicago) and when asked a question along those lines, he basically admitted as much. I think that MDT risks are more likely to be issued in place of what might have been HIGH in the past, except for exceptionally high end setups and some ENH days may have been MDT in the past.

One example off the top of my head is 5/6/15, which received an ENH upgrade from Slight, but in hindsight would have verified a MDT or even a small HIGH from KS up into NE.

Sent from my SM-G935V using Tapatalk

[Quincy]

All the aforementioned issues in reducing the potential ceiling of events is valid, but I think the addition of ENH to the SPC outlooks has also had an impact. In fact, Bill Bunting gave a remote presentation to my WFO (Chicago) and when asked a question along those lines, he basically admitted as much. I think that MDT risks are more likely to be issued in place of what might have been HIGH in the past, except for exceptionally high end setups and some ENH days may have been MDT in the past.

One example off the top of my head is 5/6/15, which received an ENH upgrade from Slight, but in hindsight would have verified a MDT or even a small HIGH from KS up into NE.

Sent from my SM-G935V using Tapatalk

I think the most apparent reason for the lack of HIGH risk outlooks may have to do with forecaster confidence, and that higher-tier outlooks are less likely to be made unless there is strong confidence. With the increasing resolution of mesoscale models, it has become much easier to predict storm mode and the severity of severe weather events in the short term.

 

We've also seen a general decrease in the number of higher-end outlook "busts" over the past few years. For example, there hasn't been a HIGH risk outlook verified with 10 or less tornadoes since 2007 (8 years), however in the previous 8 years, there were six such "busts." It should be noted that a HIGH risk can be issued for wind-only, including derecho events.

 

I'm not sure if there really is much of a difference with the introduction of ENH, since the probability tables have barely changed. The only explanation would be if SPC is hedging more on terms/words than following a strict scientific approach with probabilities. I wonder though... what is the method for creating an outlook? Do forecasters think of probabilities first and then refine, or do they say "this feels like an ENH risk" or "I expect a MDT risk verification."?

These refinements would only impact a couple of Day 1 tornado and severe wind outlooks during any year. -SPC

Maybe a placebo effect? The only cases that a previously MDT risk would now be a ENH are:

• 15% tornado with no hatched/significant tornado risk
• 45% wind with no hatched/significant risk

I looked through all of the ENH risks since the change and couldn't find one example of that. (unless it happened on a day that was upgraded to MDT at some point)

 

Or maybe it all comes back to the prior point about forecaster confidence. Has SPC been more "conservative" with outlooks lately? With the benefit of better technology, maybe it's just a function of improved forecaster skill.

 

Verification is an interesting point. According to the Practically Perfect Forecast analysis technique, it's fairly common for events to "verify" as high risk, especially if there are a cluster of reports in one small area, and/or if severe storms roll through a densely populated location. In hindsight, a lot of these "HIGH risks" could be next to impossible to pin down, more than a couple of hours in advance.

 

The March 18th event was SLGT, but verified HIGH via wind according to PPF.

http://www.eas.slu.edu/CIPS/RTIME/PPF_realtime.php

 

It has been pointed out by some that PPF may be flawed due to its tendency to seemingly inflate the probabilities in tightly clustered severe events.

[bdgwx]

Those are some really good points Quincy and RCNYILWX. Specifically that the new enhanced option may be skewing or at least redefining what it means to be a moderate or high risk and that there are generally less high risk busts today than there were 10+ years ago due to improved skill. I also wonder if the 4/27/11 outbreak has somewhat desensitized some of us somewhat. I know I sometimes, albeit incorrectly, look at a particular setup and think "oh...well that doesn't look that bad compared to 4/27/11". But I also can't help thinking there really has been a reduction in the environmental factors favoring big tornado outbreaks recently. So maybe it's a combination of a lot things. Anyway, yeah, interesting points.

[Smokey_mountain_vols]

Number of tornadoes 1314

First high risk March 27th.

Stebo with the homerun hit. Just off by 2 days.

Sent from my HTC6525LVW using Tapatalk

[andyhb]

All the aforementioned issues in reducing the potential ceiling of events is valid, but I think the addition of ENH to the SPC outlooks has also had an impact. In fact, Bill Bunting gave a remote presentation to my WFO (Chicago) and when asked a question along those lines, he basically admitted as much. I think that MDT risks are more likely to be issued in place of what might have been HIGH in the past, except for exceptionally high end setups and some ENH days may have been MDT in the past.

One example off the top of my head is 5/6/15, which received an ENH upgrade from Slight, but in hindsight would have verified a MDT or even a small HIGH from KS up into NE.

Sent from my SM-G935V using Tapatalk

 

One way I've found to judge the magnitude of threats we've had is to look at activity on this forum. There has not been a severe event where we have used the main forum for discussion here since April 27th-29th, 2014. Span of almost 2 years (and that was also the last high risk for tornadoes specifically).

 

5/9 last year definitely could've been on the higher end (and we still saw some more significant tornadoes with it) if there weren't so many problems with antecedent convection. 5/16 could've as well, but that had issues with water loading, antecedent convection north of Oklahoma and also VBV in the wind profiles. I do agree on the enhanced addition also being a factor, although I really don't think there has been a single setup worthy of a high risk since the last one in June, 2014. 12/23 last year was probably the closest, but that had problems with warmth in the mid levels south of the long lived tornadic supercell in N MS.

[rolltide_130]

5/9 last year definitely could've been on the higher end (and we still saw some more significant tornadoes with it) if there weren't so many problems with antecedent convection. 5/16 could've as well, but that had issues with water loading, antecedent convection north of Oklahoma and also VBV in the wind profiles. I do agree on the enhanced addition also being a factor, although I really don't think there has been a single setup worthy of a high risk since the last one in June, 2014. 12/23 last year was probably the closest, but that had problems with warmth in the mid levels south of the long lived tornadic supercell in N MS.

 

FWIW, had the SPC put a high risk roughly around the area of my crappy drawing, it would've verified. Although with that said the northern extent of the MDT failed to verify on 12/23 IIRC.

 

[brettjrob]

Just remembered that today marks four years since the last Plains HIGH risk, unless I'm forgetting something.

 

Before that, going back to at least 2006, we never went much more than a year without one.

 

2006-04-06
2007-05-05
2008-05-22
2008-05-29
2008-06-05
2009-04-26 (bad bust, to be fair)
2010-05-10
2011-05-24
2012-04-14
...

[Quincy]

We're near average year-to-date with tornadoes (below if you use the 10-year average/LSRs, but above if you use a 25 or 30-year sample) and unless tomorrow overperforms, we risk falling below average for the first time since early in February. I'm working on a project with Ian for USTornadoes that will get into more details on our progress year-to-date. Look for this in a few days.

 

Although April is likely to finish below average as well (barring some significant event(s) late in the month), we're finally rounding a corner here where severe threats should steadily pick up by next weekend. The "long-awaited" troughing across the western third of the country is likely to return and the CFS has spat out multiple runs in a row showing basically a string of threat days starting around the 24th and more or less continuing well into May.

 

I won't speculate too much on May yet, although I am encouraged in general, but I think I have good confidence in calling for at least a near average final week of April, if not above average, if the puzzle pieces were to all come together.

[rolltide_130]

1155 and April 26th.  Hard to go wrong with the last 2 weeks in April.

 

While it's still a week out, you might wind up nailing this one if things don't trend away from what's being shown for next Tuesday. 

[ValpoVike]

1155 and April 26th.  Hard to go wrong with the last 2 weeks in April.

 

I may have a reasonable shot to nail my high risk prediction date.  

[Eskimo Joe]

Less than 900, no high risk.

[jojo762]

Less than 900, no high risk.

wow. Pretty conservative number. And especially conservative on the High Risk given that we could get one in the near future...

[hlcater]

wow. Pretty conservative number. And especially conservative on the High Risk given that we could get one in the near future...

I say moderate, the way recent runs have looked, with the exception of the bullish 12z GFS, point more to a moderate risk event more than a high risk one. However if the 12z GFS has it right, that would probably be a high risk, in my rather inexperienced eyes at least.

[weatherextreme]

I say moderate, the way recent runs have looked, with the exception of the bullish 12z GFS, point more to a moderate risk event more than a high risk one. However if the 12z GFS has it right, that would probably be a high risk, in my rather inexperienced eyes at least.

 

 

Still 4 days out, but it looks like the hype-train has settled down a little, for now.

[andyhb]

Well with tomorrow likely not going to hit the high risk threshold due to numerous caveats (unless something significant changes between now and then) and the pattern afterwards (especially following this weekend) looking pretty hostile to severe potential, time is starting to tick on getting that pink categorical outlook to show up this spring.

[thewxmann]

Well with tomorrow likely not going to hit the high risk threshold due to numerous caveats (unless something significant changes between now and then) and the pattern afterwards (especially following this weekend) looking pretty hostile to severe potential, time is starting to tick on getting that pink categorical outlook to show up this spring.

 

My 6/10 prediction looking solid still...