This work sits on the fringe of the knowledge map: its nearest neighbors are studies of the solar corona, solar wind, and fast radio bursts, not meteorology. Within meteorology, however, it continues a line of research on thunderstorm environments that previously analyzed mean values or annual aggregates. This article fills a gap by focusing on extremes and month-by-month dynamics.
The authors used NARR reanalysis data from 1979 to 2015 with a 3-hour time step and 1-degree resolution. For each grid point and each month, they computed the maxima of CAPE, SRH, and PROD (the square root of the product of CAPE and SRH). They then fitted a generalized extreme value (GEV) distribution to the monthly maxima, allowing the location parameter to depend on time or the El Niño index. Significance was assessed using the Benjamini–Hochberg procedure to control false discoveries.
Significant positive trends were found for PROD maxima in April, May, and August; for CAPE in April, May, and June; and for SRH in April and May. In April, the PROD trend is significant at 313 grid points at a 20% false discovery rate, implying at least 250 real trends. The maximum increase in PROD is about 3% per year relative to the 1979 level. For February, a strong negative relationship was found between PROD and SRH extremes and the El Niño index: during La Niña (negative phase), the risk is higher. The largest coefficient corresponds to a 10% decrease in PROD per degree Celsius of the index.
This work differs from its archive neighbors (solar physics) in being a completely different field, but within meteorology it refines previous studies of thunderstorm environments. Unlike studies that used annual maxima or mean values, this month-by-month analysis of extremes reveals seasonally dependent trends and a link to El Niño specifically in the tails of the distribution. This is the first study to show an observed increase in CAPE extremes over the US, which had previously only been predicted by climate models.
The topic could develop toward studying the spatial dependence of CAPE and SRH extremes, as well as their joint distribution. This would allow more accurate estimation of the probability of simultaneous extreme conditions at different locations, which is important for assessing the risk of large thunderstorm outbreaks. It would also be promising to link PROD extremes to the actual number of severe thunderstorms per day, to test how well the parameter reflects real impacts.
The results will affect severe thunderstorm risk assessment and insurance in the US, especially in spring. They are also useful for climate modeling and seasonal forecasting.
The next step is to analyze the spatial dependence of CAPE and SRH extremes. It would also be worthwhile to check whether days with PROD maxima correspond to days with the highest number of reported severe thunderstorms.
In meteorology, the question remains open as to how changes in thunderstorm environments translate into actual thunderstorm frequency and intensity. This work links extremes of environmental parameters to climate indices, but it does not answer the question of a direct causal link with tornado and hail reports. This is a key unresolved problem for assessing real risk.
The maximum annual increase in PROD in April is about 3% relative to the 1979 level. For comparison, that is like adding about 26 meters to the height of Mount Everest every year.