Insights & News

MetSwift’s Outlook for US Peak Tornado Season 2021

22nd January 2021

Early each year, Americans across large parts of the USA brace themselves for a reliable threat: tornadoes. According to official records held by the National Oceanic and Atmosphere Administration (NOAA), tornadoes have been responsible for the following yearly averages during 2000-2018 (1960-1999):

  • 1026 reported injuries (1588)
  • 79 fatalities (78)
  • $592 million in property loss
  • $163 million in crop loss

Reports were not so reliable in the earlier years but even so, we can see that improvements to warning systems have reduced the typical number of injuries by about a third when compared to 1960-1999.

Fatalities have stayed about the same though and even accounting for reporting reliability, that’s alarming. It’s testament to the tendency for tornadoes to develop suddenly and move erratically.

Of all weather types, tornadoes may well have the highest capacity to cause injury or death relative to frequency and scale (bearing in mind earthquakes are geology, not weather).

Twister Climatology

These violent vortexes have a distinct ‘peak season’ in which they occur most often.

For the contiguous US generally, this spans March-April, but it tends to extend through July in the Midwest region.

Seeking the Signs for Peak Season 2021

As I write this, the tropical Pacific waters are dominated by an event known as La Niña. This is the cold phase to a phenomenon known as the El Niño Southern Oscillation (ENSO).

It’s widely accepted in literature that during a La Niña, ‘peak tornado season’ tends to be more active than usual. So, this seems a good place to start looking for signals.

The stats in the graph below support the generally above average tendency for all contiguous US regions – but they also show that the difference to the long-term average can be small. The only standouts are the Southeast in March and May, and the Midwest in April.

To find stronger signals for a season, we need to consider other key driving forces as well. The next graph shows what we see when accounting for ocean surface temperature patterns in both the North Pacific and North Atlantic.

Now, there are several striking differences. Most of all, for the Southeast in April. That huge tornado count exists despite there being 5 matching years contributing. That’s because of one extremely anomalous month being matched: April 2011. Across the contiguous US, the ‘Super Outbreak’ of that month put down a staggering 106 F1+ tornadoes!

Elsewhere, we see a strong signal for above-average activity in the Midwest April-June. By contrast, below-normal activity is suggested for May-June in the Southwest and for June in the West.

At this stage, though, we’re still using a basic filtering technique. The MetSwift analogues algorithm goes well beyond this, considering more driving forces and their relative strength of influence on weather patterns in order to select the ‘best fit’ years to use as guidance.

Let’s study the results using the MetSwift analogues issued January 2021.

MetSwift Suggests: Expect More Tornadoes Than Usual

What really jumps out here is the predominance, across the months, of above-normal tornado counts for all regions. The only noteworthy exceptions are for the Midwest in May-June, and the West and Southwest in June.

Generally, there’s a suggestion that tornado activity will be focused further east than usual and be most anomalously high during March-April.

We can visualise that more clearly using a by-state breakdown:

Some of the largest changes in typical tornado count when looking at the MetSwift analogues versus long-term average:

Alabama (12 to 30) | Kansas (19 to 31) | Tennessee (9 to 18) | Arkansas (12 to 21)

Texas (40 to 24) | Florida (8 to 2) | Colorado (8 to 3)


There’s a lot of variability between years, though, so we shouldn’t take these figures too literally. It’s the overall pattern of suggested difference to the long-term average that’s worth taking home.

Take care everyone – especially if you’re in one of the more tornado-prone regions!

James Peacock MSc
Head Meteorologist at MetSwift

Cover Photo by Unknown Author is licensed under CC BY

Speak to one of our experts and discover how our analytic platform can revolutionise your weather & NatCat insight. Request a meeting, a chat & demo, or a free trial. Developed via consultation & collaboration.