GLOBAL CLIMATE-CHANGE AND TROPICAL CYCLONES

This paper offers an overview of the authors' studies during a specialized international symposium (Mexico, 22 November-1 December 1993) where they aimed at making an objective assessment of whether climate changes, consequent on an expected doubling of atmospheric CO2 in the next six or seven decades, are likely to increase significantly the frequency or intensity of tropical cyclones (TC). Out of three methodologies available for addressing the question they employ two, discarding the third for reasons set out in the appendix. In the first methodology, the authors enumerate reasons why, in tropical oceans, the increase in sea surface temperature (SST) suggested by climate change models might be expected to affect either (i) TC frequency, because a well-established set of six conditions for TC formation include a condition that SST should exceed 26 degrees C, or (ii) TC intensity, because this is indicated by thermodynamic analysis to depend critically on the temperature at which energy transfer to air near the sea surface takes place. But careful study of both suggestions indicates that the expected effects of increased SST would be largely self-limiting (i) because the other five conditions strictly control how far the band of latitudes for TC formation can be further widened, and (ii) because intense winds at the sea surface may receive their energy input at a temperature significantly depressed by evaporation of spray, and possibly through sea surface cooling. In the second methodology, the authors study available historical records that have very large year-to-year variability in TC statistics. They find practically no consistent statistical relationships with temperature anomalies; also, a thorough analysis of how the El Nino-Southern Oscillation cycle influences the frequency acid distribution of TCs shows any direct effects of local SST changes to be negligible. The authors conclude that, even though the possibility of some minor indirect effects of global warming on TC frequency and intensity cannot be excluded, they must effectively be ''swamped'' by large natural variability.