Every spring, the sky over the US Great Plains starts behaving like a giant steam boiler. Warm, moist air from the Gulf of Mexico collides with dry air from the west, and in that mix, thunderstorms are born that can rip the roof off a barn in minutes. Meteorologists have long tracked the 'fuel' for such storms, but so far they have mostly looked at average values. The new study measures how often that fuel reaches dangerous concentrations.

The authors used NARR reanalysis data from 1979 to 2015: for each grid point and each month, they calculated the maximum value of CAPE, SRH, and PROD. Then they fitted an extreme value distribution to these maxima, allowing its location to change over time or with El Niño influence. This approach separates random fluctuations from a real trend.

The result is seasonally uneven. In April, May, and August, PROD maxima increase significantly, with the trend visible in 313 grid points in April. This means that in regions where thunderstorms are already frequent, 'fuel' days become even more fuel-rich. For CAPE, the increase is seen in April, May, and June. The maximum annual increase in PROD is about 3% of the 1979 level.

February tells a different story. There, PROD and SRH extremes are tied more strongly to the El Niño phase than to time. When the tropical Pacific is cooler than usual (La Niña), the risk of extreme values is higher. The largest coefficient corresponds to a 10% decrease in PROD for each degree Celsius of the El Niño index.

Previously, it was unclear whether extreme values of the thunderstorm environment change over time. Averages might remain stable while the tails of the distribution creep upward. Now we see that in spring, the tails are indeed growing, and this is the first observational confirmation of what climate models predicted for CAPE. In practice, this means that insurance companies and forecast services should reassess springtime risks, especially in already vulnerable areas. The storm-sky thermometer shows a mild but persistent fever.