Typhoon is one of the major natural hazards on the earth. Current research on typhoon is mainly focused on its intensity, i.e., its maximum wind speed. However, the devastating power of typhoon is not only related to its maximum wind, but also related to its coverage or radius of gale force wind. For example, hurricane Sandy is only a category one hurricane (maximum wind of 35 m/s) when it landfalled on the eastern coast of the US. However, it brought a damage equivalent to ~ 50 billion US dollars and became the sixth in terms of the total loss caused by hurricane in the US history. One main reason is the amazing size of Sandy, whose radius of gale force wind (> 17 m/s) reaches 1610 km. this induced wide spread of storm surges up to 3 meters high along the vulnerable east coast of the US. Although we have a good understanding of the genesis and maintenance of typhoon now, what influence the size of typhoon remains largely unknown. The intensity and track forecast of typhoon has progressed gradually, but the size remains unpredictable most of the times.

For the first time, by using large amount of satellite data and global atmospheric model simulations, we found that rainfall area of tropical cyclones (another indicator of tropical cyclone size) is primarily controlled by sea surface temperature in the local tropical cyclone environment relative to the tropical mean sea surface temperature, or the relative sea surface temperature, instead of the conventionally thought absolute SST. the main reason is that the mid-troposphere is more humid over area with higher relative SST over the tropics due to prevalent convective heating. As a tropical cyclone moving into an area with high relative SST, its rainfall area will expand and so does the wind field. Vice versa, a tropical cyclone entering a low relative SST area will tend to contract. This partly explains why tropical cyclones in the western north Pacific is generally larger than those in the eastern Pacific ocean.

The work implies that tropical cyclone rainfall area in a warmer climate is expected to change little since relative SST will remains largely unchanged in a uniform warming climate. As a result, the increased amount of tropical cyclone rainfall will be mainly contributed by the increased intensity and rainfall rate instead of the increased rainfall area.