May 2015 Tornado Forecast
Climo tornado probabilities for May 15th.
The April tornado forecast was a trial run and a lot has been learned since it was made. After some success with that forecast, I will be incorporating a lot more statistical analysis (analog data) into coming up with a forecast for the month of May. Click here to skip to the tornado forecast details.
Quickly recapping April: Overall, April was a fairly steady month for tornado reports across the U.S., with 21 out of 30 days reporting tornadoes. April 1-9 was very active, including a notable, early season EF-4 tornado in northern Illinois. April 10-18 was relatively inactive, but had tornadoes reported in all but two of those days. April 19-27 was rather active, but no single day featured a major or significant outbreak. Most of the tornadoes in the month were EF-1 or weaker, although the April 9th EF-4 was an exception. Preliminary tornado counts indicate that the month was near to slightly below average in tornado activity. The prior forecast issued on March 31st predicted slightly above average tornado counts. Once the numbers become official, I will get into more detail with comparing the forecast vs. actual.
NWS drought update for the Plains.
Before we get into May, there are some things to consider with the current state of the U.S. Although a long-term drought continues for much of the Plains, heavy rainfall throughout April reduced drought conditions considerably across portions of Oklahoma and Texas. This is critical, because relative tornado inactivity during the first half of spring 2014 across the Plains was at least partially attributed to a lack of available moisture. When we think about creating environments that support tornadoes, moisture is key and that moisture source can originate in Texas. Yes, the Gulf of Mexico is another player, but that moisture must also be able to track hundreds of miles inland, across the big state of Texas. Additionally, the overall atmospheric pattern has shifted somewhat, as although there has been continued troughiness across eastern Canada and the Northeast, we are seeing a trend toward more ridging developing. This will have many implications on how the tornado season evolves into May. With less blocking across northeast Canada toward the North Pole, that should tend to favor positive height tendencies as well. (Not to mention seasonality helps with that as well) For this outlook, the focus will be split into two portions. May 1-10 and May 11-31. I will still include breaking the month down into thirds at the end of the outlook, but keep in mind that specifics beyond May 10th lack the confidence needed to get into extreme detail in that time period. May is also a potentially volatile month, as one outbreak or even one day for tornadoes can effectively skew the big picture. It would be difficult to nearly impossible to predict any such outbreak more than 10 days in advance. The point of this outlook is to look at May as a whole.
[wpanchor id=”details”]May 1st through 10th: After a quiet finish to the tornado season in April, the first few days of May look to continue that theme. This should not last long, though, as both the operational and ensemble forecast models are in strong agreement with the expected pattern. On paper, it looks encouraging for severe weather prospects. An upper level trough and surface low should exit the New England area may May 3rd, leaving the Lower 48 with a relatively zonal pattern, at least for a short-time. Downstream ridging across the East should give way to a digging trough across the West Coast by May 4th to 5th. With some blocking across eastern Canada, we are looking at western U.S. trough that will persist for several days, with potentially multiple impulses rotating around from the Intermountain West into the Plains.
Model forecast 500mb heights and anomalies, centered on May 7th.
This could ultimately evolve in many ways. At this point, the most probable scenario would appear to be two to three, perhaps four, days that are conducive for minor to moderate severe weather events, including tornadoes. I would not want to write off the potential for a big event just yet, but it seems more likely that we will have a series of smaller events, kind of how April panned out. What will be working in the favor of severe weather will be the bonus of a better moisture source across the southern Plains and moderate to strong instability ahead of approaching shortwaves. It is the details that get somewhat muddled out, with the amount of clearing/destablization, storm mode, locally backed winds, etc. Essentially, the pattern looks favorable for severe weather and tornadoes, particularly between May 5th and 8th. As mentioned before, maybe only two of those days wind up decent for severe, but if the pattern lines up just right, there could be multiple tornadoes each day.
Verdict: Near average tornado activity between May 1st and 10th. The period should begin quiet, but then see multiple days with at least modest tornado activity. The period may close out with a couple of quiet days as the pattern trends more zonal, assuming the model projections are correct.
May 11st through 31st: I have put a lot of time into assessing analog data for this period and not so heavily relying upon forecast model ensembles, as was the case with the April tornado outlook. What has been most alarming lately is that the longer-range models have really struggled beyond day 10 over the past several weeks and have often showed little to no skill in the period of days 15 to 30. This means that the models have flip-flopped back and forth, giving little meaningful insight to what might lie ahead. Now, I am not totally discounting the models here. Sometimes they can struggle during a large scale pattern change, much like what we may be seeing into May. I put a fair amount of stock into the European ensembles/weeklies, but I think taking a look at the past will help shed some light on what might happen in the future.
Top 5 analogs for mid to late May.
To come up with analogs, I have looked at two things in particular. First, I assessed the short-term model analogs for May 1-2. Secondly, I looked at the day 6-10 ensemble predictions from the GEFS and those analogs. There was some overlap and plenty of similarities to note. In the broad scheme, the analogs began with the western troughing/eastern ridging that we are expecting to see into the first week of May. The analogs are then in strong agreement with positive height anomalies along and east of the Rockies from the 2nd week of May, pretty much through the entire month. After gaining visibility to the latest 00/30 Euro Weeklies, pretty much the same is the case here. That only increases confidence in this forecast. I pretty much disregarded the CFS entirely for this outlook, as its members are all over the place for May and as an ensemble system, has not had a good track record as of late.
Top 5 analog mean 500mb heights.
Going back to the ridge axis, that is the key thing to consider. The indications, in general, point toward a mean ridge axis setting up somewhere between the Missouri River and the Mississippi River. The latter, or a further east ridge axis, would be more supportive of tornadoes, allowing for troughs to dig a bit further south and west into the Plains. Even the former would be encouraging, as it would tend to favor High Plains activity. That region has not had a great chase season in a while. Based on the reports, it appears as if 2010 was the last solid year there and 2008 was decent too, particularly across western Kansas. Either way, since we are looking at a broad period here and kind of splitting hairs, there is nothing necessarily stopping a rogue, deep trough to swing through, although it would appear that if that was going to happen, it would tend to favor late May over mid-May. As a result, I am leaning toward a somewhat more active end to May, even though I do expect a fair amount of activity in the middle of the month as well. (When I say fair amount of activity, I expect several days with scattered severe reports, again, in some ways similar to what April featured)
Verdict: Near average tornado activity between May 11th and 31st. The middle of May holds some potential, but is forecast to feature slightly below average tornado reports. Into late May, there are some indications that tornado activity could rebound to near to slightly above average levels with more potential activity. When looking at some of the recent analogs (over the past 25 years), 3 out of 4 had slightly to moderately below average tornado counts in May. On the flip side, they tended to favor average to even very active months of June. Since May 2015 does not appear to be a particularly close match to any of those analogs, I would take that information with a grain of salt. If anything, it does give at least some additional confidence in the thinking that May could end more active and lead into another active June. I do not want to get ahead of myself here, so we will hold off on further discussion about June until the next monthly tornado forecast.
- May 1-10 forecast: Near average tornado counts in the U.S.
- May 11-20 forecast: Slightly below average tornado counts in the U.S.
- May 21-31 forecast: Near to slightly above average tornado counts in the U.S.
May 2015 tornado estimate: 260 tornadoes in the U.S. (near to slightly below average) For those wondering, the top 5 analogs are as follows, beginning with the given date and extending out 20 days: 4/28/1994, 5/12/1991, 4/28/1957, 5/10/1993 and 5/18/1977, in that order. The “recent” analogs mentioned earlier in the post were 1990, 1991, 1993 and 1994.
