2004 had a couple of minor events in March and a pretty dull April in the Plains. I'm curious to see the longwave pattern looked like for that April to see if there are any similarities with this one. 2004 then went historic (in terms of chaser successes) after May 12. Before May 12 it was pretty dismal, but it really turned on a dime.
I pulled some of the upper air analyses from 2004 and there are some similarities, but differences as well. I'll share some of them below, but nothing is really conclusive in showing a strong pattern match.
Looking at March, 2004 and 2018 are quite different.
March of 2004 saw positive 500mb height anomalies across the continental U.S. with sharply negative height anomalies displaced to the Polar region, north of Canada. Surface temperatures were at or above average for just about all of the CONUS, but well above average over the central and western U.S. The month also saw an unseasonably far north shift in tornado activity, with clusters of tornadoes from North Texas into northwestern Oklahoma and central Kansas, likely due in part to the northward displacement (compared to climo) of the jet stream and resultant above average temperatures over the Plains.
On the other hand, 2018 has featured nearly constant troughing across the eastern U.S. with some troughing over the West Coast. Surface temperature anomalies show below average temperatures common across the northern and eastern tier of states, with above average temperatures over the southern Plains/southern Rockies.
Looking ahead, let's start with some of the similarities.
The early April progs for 2018 are similar to what was observed over April of 2004. If the forecast holds, troughing (negative height anomalies) will prevail over central/eastern Canada with ridging in the northeastern Pacific. Tornado activity for early April 2004 was unusually low, with just
29 tornadoes occurring from April 1-15.
For differences, the longer range data, namely the European weeklies, suggests large scale troughing prevails across much of Canada, with a tendency toward positive height anomalies (ridging) from the northern Pacific into the Bering Sea. In 2004, ridging was evident over the northeastern Pacific/western Canada, but troughing occurred southwest of Alaska. This pattern ultimately shifted eastward, allowing for a series of upper level troughs to dive across the western U.S., supporting increased tornado activity (well above average) in May across the Plains and Midwest.
An interesting note is that early spring 1997 is a common theme in the CPC day 6-10 analogs, accounting for 3/10 of the analogs. Years don't usually show up more than once or twice unless there is a strong correlation. Note that the northeastern U.S. saw a historic early spring snow storm from March 31-April 1, which fits the pattern of this year, as March has been much snowier than average in that part of the country.
Final thoughts:
I've dabbled a bit in analogs and long-range tornado prediction over the past few years and it's been tough. While there have been some successes, there have also been challenges as well. Since it only takes one or two big events to skew an entire tornado season, an otherwise quiet tornado season can be offset by a tornado outbreak sequence or even just a single historic event. Mesoscale details are important when considering tornadogenesis. A favorable looking pattern can go on to be a dud (April 2004 and 2013 were mentioned in this thread), while a relatively zonal flow can feature localized tornado outbreaks, if mesoscale features like a warm front or outflow boundary line up just right. May 18th, 2017 was an example of a tornado-driven high risk struggling to produce many photogenic tornadoes in the risk area, west of I-35, while most of the tornadoes of the event happened after dark in eastern Oklahoma and the Ozarks, displaced far from the high risk area.
The bottom line is that predicting tornado activity inside of two weeks is considerably easier than the period of 3-4 weeks out and beyond. You can't predict a mesoscale "accident" a month in advance, let alone inside of 24 hours, sometimes...
It's safe to say that April will most likely start with well below average tornado activity, but what happens beyond that is anybody's guess. I may dive a bit deeper into analyzing some analogs and such, but I'm not sure how worth it that is. Many of us are getting anxious for the upcoming peak tornado season, when it finally arrives. I'm somewhat more inclined to just leisurely look at weather maps and wait for May to surprise us with whatever nature has up its sleeve.
