Volume 84 (2006) Issue 2 Pages 405-428
Influences of sea surface temperature (SST) spatial patterns and cumulus parameterizations on tropical cyclone (TC) frequency, in the context of global warming impacts, are investigated using an atmospheric general circulation model at T106 horizontal resolution. Simulated TCs in this high-resolution model are categorized into tropical storms (TSs) and tropical depressions (TDs). Model TSs are defined as TCs with maximum surface wind speed more than, or equal to 16 m s−1, for experiments with an Arakawa-Schubert cumulus parameterization. Another threshold of 14 m s−1 is used for those with a Kuo cumulus parameterization. Model TDs are defined as weaker TCs. Although the maximum wind speed, and the minimum central pressures of intense TCs are not realistically simulated in the model, geographical patterns of TS formation seem to be realistically simulated, with climatological and El Niño/La Niña SST conditions.
A series of experiments is conducted with doubled CO2 and with increased SSTs. A spatial pattern of SST, made by uniform 2 K warming, is used for experiments with both of the cumulus parameterizations. El Niño-like and La Niña-like warming patterns of SSTs, are used with the Arakawa-Schubert scheme. In these global warming experiments, frequency of TS formation decreases by 9.0-18.4% globally, and some of these changes are statistically significant. While no coherent changes in global frequency of relatively intense TCs (e.g., maximum surface wind ≥ 25 m s−1) are found in the warm-climate experiments, significant reduction in the total frequency of TSs and TDs resulted from all of these experiments. The results suggest that global frequency of relatively weak TCs may decrease in the future warm climate, but frequency of intense storms may either decrease or increase. Mean precipitation near TC centers is significantly heavier in the warming experiments than in the present-day experiments, as compared for TCs with the same maximum wind speed.