Beyond hurricanes: El Niño, La Niña, and Severe Convective Storms

As the 2026 El Niño rapidly develops, Inigo and collaborators have published a series of articles exploring how seasonal climate signals can inform decision-making, and how climate variability can be incorporated practically into risk transfer markets. Discussions of El Niño and La Niña in the (re)insurance industry are often framed around the Atlantic hurricane season, and for good reason. Hurricanes represent one of the industry’s most significant tail risks, where the difference between an ordinary year and one with major losses can be determined by the landfall of just one or two storms. However, the influence of the El Niño-Southern Oscillation (ENSO) extends far beyond these low-frequency, high-severity events. Through atmospheric teleconnections, El Niño and La Niña can tip the scales of probability for frequency perils such as US Severe Convective Storm (SCS), altering the environments that favour hail, tornado and damaging wind events.

In recent years, SCS has become one of the costliest natural catastrophe perils for the insurance industry. Unlike hurricanes, which generate losses through a handful of major events, SCS losses accumulate through numerous hail, tornado and damaging wind events occurring throughout the year. With insured losses exceeding $50 billion annually in the US from 2023 to 2025, understanding how climate patterns such as ENSO influence severe weather activity has become increasingly important for assessing seasonal fluctuations in risk.

How does ENSO influence US Convective Storms?

At a high level, the main mechanism linking ENSO and US severe convective storms is relatively simple. Changes in tropical Pacific sea surface temperatures alter the position of the jet stream over North America, which affects the supply of warm, moist Gulf air needed to fuel severe thunderstorms. El Niño generally suppresses these ingredients, while La Niña tends to enhance them, increasing the likelihood of hail, tornado and damaging wind events. This influence is most apparent during spring and is best linked to the ENSO state of the preceding winter and early spring, when these atmospheric patterns become established.

What does the data say?

To quantify how hail risk (the main loss driver of the SCS sub perils) varies between ENSO phases, we draw on INFER-Hail. This is Inigo’s AI-driven hail model which uses large-scale weather data interpolate between and correct for biases in historical hail observations. Figure 1 compares average hail-day frequency during historical El Niño and La Niña years, classified according to winter and early-spring ENSO conditions and expressed as anomalies relative to the long-term average.

A clear pattern emerges during La Niña years, hail activity increases across a broad corridor extending from Texas and Oklahoma through the Midwest and into the Tennessee Valley. In some locations, hail-day frequency is more than 15% above average following a La Niña winter. An opposing signal emerges during El Niño years, where hail is suppressed across many of the regions most exposed to severe convective storms. Cities such as Oklahoma City, St. Louis, Nashville and Chicago, for example, experience fewer severe hail day per year following El Niño winters.

These findings are consistent with previous research. Allen and Tippett (2015) identified a similar relationship between ENSO and severe thunderstorm environments, showing that La Niña conditions are generally associated with more favourable environments for hail and tornado occurrence across the central and southern United States. As with any climate signal, ENSO can’t predict how many damaging hailstorms occur in a given year. Instead, it alters the background odds, shifting the likelihood of particularly active or inactive seasons.

What can we learn from history?

Some of the most active severe weather seasons on record occurred following La Niña winters. The devastating 1974 Super Outbreak, which produced 149 tornadoes across the central and eastern United States, developed in April following a strong winter La Niña. The record-breaking April 2011 Super Outbreak also followed strong winter/spring La Niña conditions, with over 360 tornadoes including four EF5s (the most severe category). More broadly, several studies have found that many of the most extreme US tornado outbreaks have occurred during atmospheric patterns commonly associated with winter La Niña conditions that persist into spring.

As with any peril, translating changes in hazard into changes in loss is not straightforward. ENSO may influence the frequency of hailstorms, tornadoes and damaging winds, but insured losses also depend on where those events occur, how much property lies in their path, and broader factors such as inflation, insurance structures, exposure growth and reporting practices. Despite these complicating factors, industry-wide loss data reveals that since 1985, El Niño years have seen US SCS losses around 10% below the annual average, once adjusted for inflation and exposure growth (Figure 2).

Conclusion

Severe Convective Storms are a prime example that the seasonal forecast skill offered by ENSO extends well beyond the Atlantic hurricane season. Wildfire, flood, and a range of perils across regions including Japan, Australia and South America are all impacted by ENSO, suggesting that the opportunity for (re)insurance markets is broader than traditionally recognised.