Skip to main content
Hong Kong Observatory Brand Hong Kong - Asia's world city
GovHK Traditional Chinese Simplified Chinese Search
Search Site Map Contact Us
red dot
Print Version
Print Version

Wednesday, 12th August 2009

Dragon Boat Festival climate

An insightful reader queried about the blog released on 25 May 2009. In essence, he questions that since Tuen Ng (Dragon Boat day) tends to vary from year to year, whether a warming trend could indeed be identified for the period around Tuen Ng.

The Chinese calendar is a lunisolar calendar, not a purely solar one like the Gregorian calendar we commonly use nowadays. The date of a Chinese festival like Tuen Ng tends to change from year to year with the addition of a leap month every two to three years.

        Figure 1

Figure 1

Figure 1 shows the dates of Tuen Ng over the years. One can see that even though the day of the festival varies from year to year, there is a degree of regularity to the pattern. If one computes the mean date over 10-year periods (as used in the 25 May blog) and its variation over the past 60 years, one finds the following:

        Table 1


Mean day of year for Tuen Ng
(from 1 January)

deviation (days)

1949-1958 161.8 9.1
1959-1968 162.2 8.9
1969-1978 162.7 8.6
1979-1988 163.0 9.2
1989-1998 160.6 9.1
1999-2008 164.1 8.3

It can be seen from Table 1 that the mean day of Tuen Ng in each of the 10-year periods changes quite little --- between day 160 and 164. Its variation (standard deviation in statistical term) is about 9 days.

Now back to the discussion of the climate for Tuen Ng. In the 25 May blog, we counted the number of 'comfortable nights' (defined in the blog as those having a minimum of 25 degrees Celsius or below at the Hong Kong Observatory) over the 7-day period ending on Tuen Ng and identified a decreasing trend in the number of such nights over the three 10-year periods from 1979 to 2008. Thus, we said the local weather has become warmer and warmer.

However, with the benefit of hindsight the reader correctly points out that because of cooler weather, the Tuen Ng in 2009 had six comfortable nights out of the 7-day period. Would that upset the warming trend? Here are the results after shifting all 10-year periods forward by one year to cover 2009:

        Table 2

Year 7-day periods ending on Tuen Ng

Number of days with minimum of 25.0 o C or below


1950-1959 3854%
1960-1969 34 49%
1970-1979 25 36%
1980-1989 3246%
1990-1999 22 31%
2000-2009 26 37%

Indeed, from Table 2, the number of comfortable nights during 2000-2009 did not fall and as a matter of fact, rose a little compared with the previous decade. Does this mean a reversal of the warning trend? The answer is no. The numbers in the table still indicate a general declining trend, even though it is not a monotonic fall. And things are not going to return to those good old days where half or over half of the number of nights were 25 degrees or less, for the reasons given below.

Separate calculations indicate that local temperatures in recent decades have been rising at a rate of 0.2 degrees per decade. Our projection based on results of the Intergovernmental Panel on Climate Change (IPCC) is that such a warming trend will continue unabated if there is no basic change to our current lifestyle.

In terms of the minimum temperature for the 7-day period ending on Tuen Ng, the graph below plots 10-year running mean of the minimum temperature, from 1949 to 2009. The use of running mean is to remove short-term fluctuations to bring out long-term changes. From Figure 2, the rising trend in the average minimum temperature is evident: 1.4 degrees over the past 60 years. [In statistical terms, the increase is at a rate of 0.24 degrees per decade, at 5% significance.]

       Figure 2
Figure 2

B.Y. Lee

Last revision date: <17 Jan 2013>