Tropical cyclones (TC) under different climate conditions in the Northern Hemisphere have been investigated with the Max Planck Institute (MPI) coupled (ECHAM5/MPI-OM) and atmosphere (ECHAM5) climate models. The intensity and size of the TC depend crucially on resolution with higher wind speed and smaller scales at the higher resolutions. The typical size of the TC is reduced by a factor of 2.3 from T63 to T319 using the distance of the maximum wind speed from the centre of the storm as a measure. The full three-dimensional structure of the storms becomes increasingly more realistic as the resolution is increased.

For the T63 resolution, three ensemble runs are explored for the period 1860 until 2100 using the IPCC SRES scenario A1B and evaluated for three 30 yr periods at the end of the 19th, 20th and 21st century, respectively. While there is no significant change between the 19th and the 20th century, there is a considerable reduction in the number of the TC by some 20% in the 21st century, but no change in the number of the more intense storms. Reduction in the number of storms occurs in all regions. A single additional experiment at T213 resolution was run for the two latter 30-yr periods. The T213 is an atmospheric only experiment using the transient sea surface temperatures (SST) of the T63 resolution experiment. Also in this case, there is a reduction by some 10% in the number of simulated TC in the 21st century compared to the 20th century but a marked increase in the number of intense storms. The number of storms with maximum wind speeds greater than 50 m s−1 increases by a third. Most of the intensification takes place in the Eastern Pacific and in the Atlantic where also the number of storms more or less stays the same.

We identify two competing processes effecting TC in a warmer climate. First, the increase in the static stability and the reduced vertical circulation is suggested to contribute to the reduction in the number of storms. Second, the increase in temperature and water vapour provide more energy for the storms so that when favourable conditions occur, the higher SST and higher specific humidity will contribute to more intense storms. As the maximum intensity depends crucially on resolution, this will require higher resolution to have its full effect. The distribution of storms between different regions does not, at first approximation, depend on the temperature itself but on the distribution of the SST anomalies and their influence on the atmospheric circulation.

Two additional transient experiments at T319 resolution where run for 20 yr at the end of the 20th and 21st century, respectively, using the same conditions as in the T213 experiments. The results are consistent with the T213 study. The total number of TC were similar to the T213 experiment but were generally more intense. The change from the 20th to the 21st century was also similar with fewer TC in total but with more intense cyclones.