Is Observatory up to the mark?
Some newspapers carried a report earlier this month that in the past three years (2007-2009), the Observatory had issued Tropical Cyclone Signals No. 3 or No. 8 on 21 occasions. Of them, ten turned out to be accurate while for the other 11 occasions wind conditions were not as severe as what had been forecast. In other words, the Observatory is spot on in its tropical cyclone warning about half of the time.
I would like to discuss this further here.
When we say the issuing of a Tropical Cyclone (TC) signal is accurate, we mean that during the occasion strong wind (for TC No. 3) or gale force wind (for TC No. 8) has blown at half or more of the 8 reference weather stations in Hong Kong.
While in one of my previous blogs (15 May 2009) I pointed out that tropical cyclone forecasting has consistently improved in recent decades, difficulties still remain in the timely issuance of TC signals.
First, tropical cyclones are known to have rather erratic movement and development. This is especially so when they approach the coast, where a transition from a marine environment to a continental one may have a profound influence on their behaviour. Among other factors, the transition may include the cut-off in the supply of moisture over land which will limit the storm's development, and a change in surface characteristics, for instance an increase in friction due to land surface's roughness or an encounter with hilly terrain, both of which tend to throw a storm off its expected track.
The passage of Severe Tropical Storm Goni in August 2009 is an illustrative case of tropical cyclones' erratic movement as they make landfall. Figure 1 shows the post-analysis of Goni's haphazard journey as it crossed the South China Sea some 100 km from Hong Kong.
Second, some time has to be allowed for schoolchildren and the working public to go home, for the public transport to be geared up for the flux of people homeward, and for the local communication infrastructure to be well prepared for the sudden surge in demand as people set about winding down their work and contacting their families and friends. Without this lead time, the society may be thrown into chaos. Thus, waiting until the criteria are met or about to be met before issuing the signal, thereby ensuring a high chance the warning criteria will be met, is not a viable option. This would put people into jeopardy.
Take the case of Typhoon Hagupit in September 2008. It passed within 200 km of Hong Kong, and the combined effect of its storm surge and the high tides resulted in a maximum sea level rise of 3.5 metres. As a result, severe flooding occurred in low-lying areas such as Tai O, Tuen Mun and Lei Yue Mun. Our computations showed that had the storm moved a further 100 km closer, it would bring an additional rise of 1 to 2 metres. Such a scenario would have been catastrophic had we not given advance warning to members of the public.
Figure 2 Typhoon Hagupit brought severe flooding to Tai O in September 2008
(photo courtesy of TVB)
I am saying all these not to shirk our responsibilities, but to underpin the importance of operating the tropical cyclone warning service in such a way that public safety is safeguarded. We will constantly review and seek improvement to our services. Our team of professional meteorologists will continue to do their best to further improve the skills of storm forecasting, to better protect the public.