Because tropical cyclones derive their energy from the latent heat of condensation of water vapor put into the atmosphere through evaporation of ocean water, many scientists have hypothesized that an increase of ocean surface temperature as a result of global warming would lead to an increase in tropical cyclone activity as well as intensity. Our 2004 paper shows that at least for typhoons in the western North Pacific Ocean, this hypothesis is wrong. Years with above-normal temperatures in this ocean actually had below-normal number of tropical cyclones, and vice versa.

The main reason is that in this ocean basin, the water temperatures in the region where most tropical cyclones form are always greater than or equal to 28^oC, which is above the threshold for tropical cyclone formation (around 27^oC). Therefore, a slight increase in water temperature does not have any effect. The negative correlation we found is simply a reflection of the relationship between tropical cyclone activity and El Niño. In an El Niño year, water temperatures in the western North Pacific Ocean tend to be below normal, and hence the negative correlation with tropical cyclone activity.

The debate of whether global warming causes an increase in the number of more intense tropical cyclones has been heating up since 2005. In a recent Conference on Hurricanes and Tropical Meteorology of the American Meteorological Society (April 2006), we had a panel discussion on this issue, of which I was one of the four panelists. Two panelists continued to argue that global warming leads to an increase in ocean temperature and hence more intense tropical cyclones. Another panelist countered that their conclusion was based on sometimes very inaccurate estimates of tropical cyclone intensity especially in the pre-meteorological-satellite era, and that the recent increase in the number of very intense hurricanes in the Atlantic was partly due to improved satellite technologies in estimating tropical cyclone intensity. Further, if the past data were largely correct, the era of the 1960s had almost the same number of intense hurricanes, which means that such a number goes through cycles with periods of tens of years (that is, multi-decadal variations).

This was basically the same argument I made for typhoons in the western North Pacific Ocean. There were just as many intense typhoons in the 1960s and early 1970s as in the 1990s, and this result was published as a "Technical Comment" in Science in 2006 ("Comment on ‘Changes in tropical cyclone number, duration, and intensity in a warming environment,’" Science 311[5768]: 1713, 24 March 2006). That is, the recent increase in intense typhoon occurrence frequency is simply part of the multi-decadal variations in such occurrences over a long period of time. I therefore maintain that at least for the western North Pacific, the ocean temperature has very little effect on typhoon intensity and if global warming is to have an effect, it is through the modification of the atmospheric circulation on a global scale. I am preparing a paper at the moment to present this argument.

The 2006 paper deals with another very important topic of tropical cyclone, that of changes in the track of a tropical cyclone as it is about to hit land. Although our understanding of the physics of tropical cyclone movement and our ability to predict tropical cyclone track have improved significantly in recent years, the location at which a tropical cyclone will hit land is still very uncertain. This is because most of the previous studies did not address the specific issues that as a tropical cyclone approaches land, frictional effects at different parts of the tropical cyclone would be different (larger friction over land), and so would the supply of moisture (less coming from land).

Our 2006 paper shows that the effect of friction would cause a tropical cyclone to be "attracted" towards land, and making loops at the same time. We explained these results using the theory that I developed earlier (in "Relation between potential vorticity tendency and tropical cyclone motion," Journal of the Atmospheric Sciences 59[8]: 1317-36, April 2002). This result has important implications in the monitoring and prediction of a tropical cyclone track as it is about to hit land.

In the area of tropical cyclones (the other half of my research team works on monsoons and regional climate prediction), my current foci are: (1) changes in the tropical cyclone track and its associated wind and rainfall distributions before, during, and after it hits land, with our 2006 Journal of the Atmospheric Sciences paper forming the first part of the study; (2) the effect of global warming on tropical cyclone activity and intensity, which is a continuation of my past research as represented by my 2004 Journal of Climate paper and my 2006 "Technical Comment" in Science; (3) variations of tropical cyclone activity on different time scales, from intraseasonal to centennial, as an extension of my past research on this topic; and (4) seasonal predictions of tropical cyclone activity to improve on my previous prediction schemes.

With the effect of global warming continuing, a very important research direction in this field would be to identify a possible relationship between global warming and tropical cyclone number and intensity. This will involve the analyses of past data as well as numerical predictions using estimations of future concentrations of greenhouse gases. The analyses of past data do not have to be limited to actual observed number of cyclones.

Much work has been carried out to examine other geophysical proxies such as lake and cave deposits and coral reefs to determine tropical cyclone activity on much longer time scales (up to thousands of years). Thus, with much more research in this area, we should be in a much better position to say, in 5 to 10 years’ time, whether global warming does have an effect on tropical cyclone activity and intensity, and if so, through what mechanisms.

A related subject is the seasonal prediction of tropical cyclone activity. With the advance in computer power, it will be possible in the next few years to make numerical climate predictions of such activity with reasonable accuracy. Such research is useful for coastal communities to make adequate preparations for the upcoming tropical cyclone season.

Because all these studies hinge upon our understanding of how tropical cyclones form and intensify, much more basic and theoretical research is necessary on the physical processes responsible for the formation and intensity change of tropical cyclones. Despite much research in this area for a few decades and although the number of papers has been on the increase, a theoretical framework that can be embraced by most in the community has yet to emerge.

Another important problem is the effect of land on tropical cyclone track and its associated wind and rainfall distributions. In addition to being an interesting topic in geophysical fluid dynamics, an improved understanding of this problem has far-reaching consequences for coastal communities in terms of forecasting and disaster preparedness. It is actually quite surprising that research in this area has not been very active until recently. Presumably with the damage from Hurricane Katrina in 2005, such research will increase significantly.

Last, but not least, the field of tropical cyclone research will begin to include social scientists in the area of disaster preparedness and mitigation, and societal impacts. We will see these scientists working alongside with meteorologists to develop tropical cyclone warning systems and community education and awareness programmes. Much more interdisciplinary research will therefore emerge as a result.