PerspectiveClimate Change

Whither Hurricane Activity?

See allHide authors and affiliations

Science  31 Oct 2008:
Vol. 322, Issue 5902, pp. 687-689
DOI: 10.1126/science.1164396

A key question in the study of near-term climate change is whether there is a causal connection between warming tropical sea surface temperatures (SSTs) and Atlantic hurricane activity (1-3). Such a connection would imply that the marked increase in Atlantic hurricane activity since the early 1990s is a harbinger of larger changes to come and that part of that increase could be attributed to human actions (3). However, the increase could also be a result of the warming of the Atlantic relative to other ocean basins (4), which is not expected to continue in the long term (5). On current evidence, can we decide which interpretation is likely to be correct?

To appreciate the problem, consider the observed relation between hurricane activity [power dissipation index (PDI)] (6) and SST in the main development region of Atlantic hurricanes (hereafter “absolute SST”). Between 1946 and 2007, this relation can be defined by a simple linear regression between the two quantities (see Supporting Online Material). This observed relation can be extrapolated into the 21st century using absolute SSTs calculated from global climate model projections (see the figure, top panel) (7). By 2100, the model projections' lower bound on 5-year averaged Atlantic hurricane activity is comparable to the PDI level of 2005, when four major hurricanes (sustained winds of over 100 knots) struck the continental United States, causing more than $100 billion in damage. The upper bound of the projected 5-year average exceeds 2005 levels by more than a factor of two. This is a sobering outlook that, combined with rising sea levels, would have dramatic implications for residents of regions impacted by Atlantic hurricanes.

However, there is an alternate future, equally consistent with observed links between SST and Atlantic hurricane activity. Observational relationships (4), theories that provide an upper limit to hurricane intensity (5), and high-resolution model studies (8) suggest that it is the SST in the tropical Atlantic main development region relative to the tropical mean SST that controls fluctuations in Atlantic hurricane activity. Between 1946 and 2007, this “relative SST” (see the figure, bottom panel) is as well correlated with Atlantic hurricane activity as the absolute SST. However, relative SST does not experience a substantial trend in 21st-century projections. Hence, a future where relative SST controls Atlantic hurricane activity is a future similar to the recent past, with periods of higher and lower hurricane activity relative to present-day conditions due to natural climate variability, but with little long-term trend.

Past and extrapolated changes in Atlantic hurricane activity.

Observed PDI anomalies are regressed onto observed absolute and relative SST over the period from 1946 to 2007, and these regression models are used to build estimates of PDI from output of global climate models for historical and future conditions. Anomalies are shown relative to the 1981 to 2000 average (2.13 × 1011 m3 s−2). The green bar denotes the approximate range of PDI anomaly predicted by the statistical/dynamical calculations of (12). The other green symbols denote the approximate values suggested by high-resolution dynamical models: circle (8), star (13), and diamond (15). SST indices are computed over the region 70°W-20°W, 7.5°N-22.5°N, and the zero-line indicates the average over the period from 1981 to 2000. See Supporting Online Material for details.

From the perspective of correlation and inferred causality, this analysis suggests that we are presently at an impasse. Additional empirical studies are unlikely to resolve this conflict in the near future: Many years of data will be required to reject one hypothesis in favor of the other, and the climate model projections of hurricane activity using the two statistical models do not diverge completely until the mid-2020s. Thus, it is both necessary and desirable to appeal to nonempirical evidence to evaluate which future is more likely.

Physical arguments suggest that hurricane activity depends partly on atmospheric instability (2), which increases with local warming but is not determined by Atlantic SSTs alone (5). Warming of remote ocean basins warms the upper troposphere and stabilizes the atmosphere (5). Furthermore, relative Atlantic SST warming is associated with atmospheric circulation changes that make the environment more favorable to hurricane development and intensification (9-11).

Further evidence comes from high-resolution dynamical techniques that attempt to represent the finer spatial and temporal scales essential to hurricanes, which century-scale global climate models cannot capture due to computational constraints. High-resolution dynamical calculations under climate change scenarios (8, 12-14) (green symbols in the figure) are consistent with the dominance of relative SSTs as a control on hurricane activity. Even the dynamical simulation showing the most marked increase in Atlantic hurricane activity under climate change (13) is within the projected range for relative SST but outside the projected range for absolute SST.

Whether the physical connections between hurricane activity and SST are more accurately captured by absolute or relative SST also has fundamental implications for our interpretation of the past. If the correlation of activity with absolute SST represents a causal relation, then at least part of the recent increase in activity in the Atlantic can be connected to tropical Atlantic warming driven by human-induced increases in greenhouse gases and, possibly, recent reductions in Atlantic aerosol loading (3, 15, 16). In contrast, if relative SST contains the causal link, an attribution of the recent increase in hurricane activity to human activities is not appropriate, because the recent changes in relative SST in the Atlantic are not yet distinct from natural climate variability.

We stand on the cusp of potentially large changes to Atlantic hurricane activity. The issue is not whether SST is a predictor of this activity but how it is a predictor. Given the evidence suggesting that relative SST controls hurricane activity, efforts to link changes in hurricane activity to absolute SST must not be based solely on statistical relationships but must also offer alternative theories and models that can be used to test the physical arguments underlying this premise. In either case, continuing to move beyond empirical statistical relationships into a fuller, dynamically based understanding of the tropical atmosphere must be of the highest priority, including assessing and improving the quality of regional SST projections in global climate models.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S8


References and Notes

View Abstract

Navigate This Article