Forecasting tornado risk three weeks ahead is becoming possible

Early results from SWERVE show that forecasts in weeks two and three can sometimes detect meaningful signals of heightened severe weather activity. The signals, produced through consensus forecasts that blend multiple models, perform better than climatology alone. Instead of simply repeating seasonal averages, they highlight specific windows when the atmosphere appears more favorable than normal for tornadoes and damaging storms.

For decades, forecasters assumed that beyond about 10 days the atmosphere became too chaotic to predict. SWERVE suggests that, under certain large-scale conditions, useful guidance may extend much further than previously thought.

Severe weather warnings have become more accurate and timely, but they remain limited in how far in advance they can prepare people. Tornado warnings currently average 8–13 minutes, while SPC convective outlooks extend to 8 days. Beyond that, there has been no official product.

If reliable windows of elevated risk can be identified two or three weeks ahead, the benefits could be substantial. Emergency managers could adjust staffing well before outbreaks. Utility companies could stage repair crews and materials, farmers could adapt planting or harvesting schedules, and large events could plan around potential disruptions. For communities, earlier risk communication reduces the surprise factor and gives families more time to prepare.

The absence of official guidance for weeks two to four has long been identified by NOAA as a critical societal challenge. Severe storms often occur during this gap, too far out for short-term forecasts but too close to be captured in seasonal outlooks.

SWERVE directly targets this gap. By testing both operational numerical models and experimental products, the program is evaluating how well existing tools can identify severe weather risks across the contiguous United States at 14–21 days.

SWERVE is a cross-disciplinary collaboration involving NOAA laboratories, cooperative institutes, NWS forecasters, SPC meteorologists, emergency managers, and academic researchers. Its aim is not only to assess scientific skill but also to understand how extended-range forecasts could be communicated and used in practice.

By blending outputs from multiple models, SWERVE reduces dependence on a single forecast and instead highlights periods when different systems align on heightened risk.

The atmosphere is inherently chaotic, and small errors in initial conditions grow rapidly over time, limiting deterministic forecast skill beyond 7–10 days. Yet certain large-scale patterns create temporary windows of predictability.

One key influence is the Madden–Julian Oscillation (MJO), a tropical rainfall system that shifts eastward around the globe every 30–60 days and alters jet stream behavior. Other teleconnections, including shifts in the polar jet, can shape storm environments weeks in advance.

By tracking these large-scale drivers, forecasters can sometimes anticipate when conditions across the continental United States may become more favorable for severe weather, even if they cannot specify where tornadoes will form.

SWERVE remains in its pilot phase, and researchers emphasize that these products are still experimental. Forecasts in this range cannot predict the exact timing or location of tornadoes. Instead, they provide probabilistic guidance that signals when risks are higher than normal.

Turning these forecasts into operational products will require years of validation, user engagement, and training. NOAA’s history with the Hazardous Weather Testbed shows that moving from experiment to practice is a gradual process. Even so, SWERVE marks the first structured attempt to close the weeks 2–4 forecasting gap, and it is already demonstrating that useful guidance may be possible.

References:

¹ SWERVE: Predicting severe weather weeks in advance – NSSL News – September 30, 2025

Reet Kaur

I’m a science journalist and researcher at The Watchers, contributing to the Epicenter edition, where I cover peer-reviewed scientific research and emerging discoveries across Earth and space sciences. With a background in astronomy and a passion for environmental science, I’ve worked in shark and coral conservation in Fiji, conducting reef and shark-behavior research, contributing to mangrove restoration, and earning PADI Open Water and Coral Reef Certifications. I bring a blend of scientific rigor and storytelling to illuminate the discoveries shaping our planet and beyond.

This author does not have any more posts.