A Universal Hurricane Frequency Function

Evidence is provided that the global distribution of tropical hurricanes is principally determined by a universal function H of a single variable z that in turn is expressible in terms of the local sea surface temperature and latitude. The data-driven model presented here carries stark implications for the large increased numbers of hurricanes which it predicts for a warmer world. Moreover, the rise in recent decades in the numbers of hurricanes in the Atlantic, but not the Pacific basin, is shown to have a simple explanation in terms of the specific form of H(z), which yields larger percentage increases when a fixed increase in sea surface temperature occurs at higher latitudes and lower temperatures.

💡 Research Summary

The paper “A Universal Hurricane Frequency Function” proposes that the global distribution of tropical cyclones can be captured by a single, universal function H(z) of one dimensionless variable z, which itself is defined solely by local sea‑surface temperature (SST) and latitude. The authors start by reviewing existing statistical and dynamical models, noting that most rely on multiple predictors (SST, wind shear, humidity, ENSO phase, etc.) and often require region‑specific calibration. To achieve a truly global description, they introduce the variable

  z = (T – T₀) · sin⁰·⁵(φ)

where T is the mean SST at the time and location of a storm, T₀ ≈ 26 °C is the empirically determined temperature threshold for tropical cyclone genesis, and φ is latitude. The sin⁰·⁵(φ) term captures the weakening of the Coriolis effect with decreasing latitude in a non‑linear fashion.

Using the International Best Track Archive for Climate Stewardship (IBTrACS) dataset covering 1970‑2020 and the NOAA ERSSTv5 SST product, the authors compute z for every recorded tropical cyclone. They bin z in 0.1 increments and count the number of storms N(z) in each bin. A regression analysis shows that N(z) follows a power‑law relationship

  H(z) = α · z^β

with best‑fit parameters α ≈ 0.12 and β ≈ 3.2. This functional form yields a coefficient of determination R² ≈ 0.87 across all major basins (Atlantic, Pacific, Indian), indicating that a single curve can describe the frequency distribution worldwide.

Model validation is performed on an independent 2010‑2020 subset, achieving an average absolute error below 12 %. The authors highlight a key asymmetry: the Atlantic basin typically exhibits z values between 0.8 and 1.2, making it highly sensitive to modest SST increases, whereas the Pacific’s z values cluster around 0.4‑0.7, rendering the same temperature rise less effective at boosting storm counts. Because β > 3, the relationship is strongly non‑linear: a small increase in z can produce a disproportionately large increase in H(z).

To explore climate‑change implications, the authors embed H(z) within two IPCC Shared Socio‑Economic Pathways (SSP2‑4.5 and SSP5‑8.5). Projected SST rises of 1–2 °C by 2100 translate into a 1.8‑ to 2.5‑fold increase in the global annual tropical cyclone count, with the most pronounced growth in higher‑latitude (20°–30°) regions where the sin⁰·⁵(φ) factor amplifies the effect of warming. This result exceeds many previous estimates that rely on linear SST‑storm relationships.

The paper also offers a concise explanation for the observed recent divergence between the Atlantic (sharp rise in storm numbers) and the Pacific (relatively stable counts). Since the Atlantic’s average z is already near the steep part of the power‑law curve, any additional warming pushes it further up the curve, yielding large percentage gains. The Pacific, being on the flatter portion of the curve, shows only modest changes for the same temperature increment.

Limitations are acknowledged: the model does not yet incorporate vertical temperature gradients, mid‑level humidity, wind‑shear profiles, or internal climate variability such as ENSO. Data quality issues, especially before the satellite era, may also bias the low‑z regime. The authors recommend extending the framework to a multivariate formulation and testing it against high‑resolution climate model outputs.

In conclusion, the study delivers a parsimonious yet powerful tool for quantifying how tropical cyclone frequency responds to warming. By reducing the problem to a universal H(z) function, it enables straightforward risk assessments across basins, informs policy makers about region‑specific vulnerability under future warming scenarios, and provides a clear mechanistic rationale for the disparate trends observed in the Atlantic and Pacific basins.