Achtergrond

Floods in Manila August 2012

On the three days of 7–9 August 2012 unusually strong rains poured almost a meter of rain on Manila, the capital of the Philippines, causing widespread flooding in the city.

The rains on 7–9 August were very strong monsoon rains that lasted for three days. The intensity was increased by two tropical storms passing to the North of the Philippines. Comparing with historical data over 1947–2010, the amount of rain on a single day was the third largest recorded, implying a return time of around 20 year. However, as the rains lasted three days, the total amount has a return time of more than 600 years. In general, greenhouse warming will produce more intense short-duration rain events, but the contribution of global warming in intensifying this specific event is still an open question.

Observations

There is an international network for the exchange of weather information used for weather forecasts called GTS. The observations at the station “Science Garden” in Manila for 7, 8 and 9 August were 323.4, 363.7 and 292.6 mm/day respectively. However, data on GTS is notoriously unreliable. One of the most common errors is decimal point displacement, so that a rainfall event of 800.1 mm (Siliguri, India, 14 July 1951) was in fact only 80.01 mm. To check this we verified that the first two values were also mentioned in a newspaper interview with the Philippine Meteorological Service, PAGASA. Secondly, the values are consistent with 3-hourly values also recorded on GTS. Thirdly, the total amount is similar to the amount observed with the TRMM satellite instrument, see Figure 1. TRMM is most accurate over the ocean and shows amounts over 900 mm on the ocean in front of Manila. To put these amounts in perspective, the 3-day sum is more than the mean annual rainfall in the Netherlands.

Figure 1. Left: GTS observations at Manila “Science Garden” meteorological station. Right: TRMM rainfall estimates of 25 July until 8 August 2012, most of which fell in the last three days.

Meteorological situation

On 6 August 323.4 mm of precipitation had already poured down over the Metro Manila. This huge amount is more than half the average rainfall of 504 mm for the entire month of August. This torrential rain resulted from an unusual atmospheric situation of the southwest monsoon (called 'habagat' locally), which was amplified by the tropical storm (TS) Haiku located 200 km away in the North of Philippine (see Figure 2).

Figure 2. IR satellite imagery on August 6, 2012 at 18:00 UTC. Areas of red are high cloud tops and are expected to have the heaviest rainfall (credit : http://wounderground.com).

From the synoptic analyses as seen in Figure 3, the intensification speed of the wind in the outer area of cyclonic flow of the storm and a divergence pattern over the West Pacific east of Manila pulled more moist air mass from the west over the Philippines. A similar situation occurred two days later with the remnants of typhoon Saola.

The monsoon wind, which is part of the Asian summer monsoon circulation, converged over the South China Sea and was lifted up, resulting in heavy thunderstorms that moved toward Manila. In addition, the geographical situation of Manila and surrounding region intensified the event. The monsoon winds have to pass over the Sierra Madre mountains which makes the rainfall more intense on the western side where Manila is located rather than in the east over the West Pacific ocean, as seen in the satellite image (Fig. 2) or in TRMM data (Fig. 1).

Figure 3. Left: The 850mb wind situation overlay to Sea Surface Temperature. Right: The anomaly of 925mb Relative Humidity (around 800 meter height).

Moreover, during the days when the rain poured down excessively, there was about 4%–10% increase of the air relative humidity that indicated the abundant moisture over the southern part of South China Sea where the origin of the wind was that blew toward the Philippines as “habagat”, the South West Monsoon. 

How extreme was the event?

The SACA&D database of daily climate data from South-East Asia contains many series from the Philippine, among them data for the same Science Garden station. The `blended series’ combines data from different series and covers the period from 1946–2010, with a few gaps. In general these data show a wet season peaking in August, with days alternating between dry and very wet, almost every year includes days with more than 100 mm of rainfall.

The wettest day on record was caused by Typhoon Ondoy on 27 September 2009, when the Science Garden station recorded 454.9 mm of rainfall. A useful way to present the data is to plot the wettest day of the year against the position of the year in the record list. The 454.9 of 2009 occurred once in 64 years. An event as large or larger than the second one (2007, 391.9 mm, not associated with a typhoon) occurred twice in 64 years, or once every 32 years. The third wettest day was 2012 (which suggests that there has been an upward trend; statistically the trend is significant at p<0.1). The return time of 2012 in the observations up to 2010 is therefore around 20 years, not very extreme. This is shown by the blue line in the left panel of Fig. 4. The green lines give a theoretical fit assuming a constant climate and that all extremes are of the same type. 

Figure 4. Left: extreme value distribution of single-day precipitation. The two highest crosses denote 2009 (typhoon Ondoy) and 2007, the blue line denotes the wettest day so far in 2012.

The extreme part of the event on 6–8 August was the long duration of the heavy rainfall. In the right panel of Fig. 4 we show the same plot for three-day sums. The 2012 value is much higher than the previous highest value, 562 mm in 1970. As the value for 2012 is so much higher than previously recorded, the return time is quite high, the 95% confidence interval starts at about 1000 years, again under the assumption of a stationary climate and similar types. Including 2012 there is again a barely significant increase in the extreme events at p<0.1. It is therefore well possible that these kinds of events are more likely now than they were in the past, a simple linear analysis suggests that this would decrease the lower bound on the return time to 700 years.

Questions

Although the amplitude of this event was clearly dominated by a coincidence of natural factors, we could question whether climate changed also played are role in increasing the odds for such an extreme. It is generally accepted that with the global temperature rise the amount of moisture in the air will rise. Over the oceans, and near the surface, this is expected to occur with a rate close to 6 % per degree of warming. More moisture is very likely causing more intense rainfall events, everything else being equal. For short duration events, this could even lead to a dependency close to 14 % per degree. For the tropics, global climate models also generally predict positive changes in extremes, yet with large uncertainty ranges. 

For this event it would be very interesting to investigate to what extent it was influenced by global warming already occurring. However, this is far from certain as, contrary to most of the rest of the world ocean, the upstream South China Sea has been cooling over the last 15 years after a steady increase of one degree since 1910. Temperatures in July 2012 were about half a degree lower than in the late 1990s. On the other hand, the moisture source could originate from other regions further away. Therefore, a more in depth analysis of this case is needed.

Other mechanisms could also have played a role, such as the increase in air pollution that suppresses light precipitation but increases heavy downpours. The urban heat island of Manila could also have intensified the rainfall. A careful study that examines all these possibilities must be performed to be able to draw conclusions.

If anthropogenic factors such as global warming, air pollution or urban heat island effects are confirmed to have intensified the rainfall in this case, it would mean that the probabilities of such heavy downpours in megacities in the tropics are higher now than an analysis of past data would show. This would necessitate adaptation to the new situation.

Niet gevonden wat u zocht? Zoek meer achtergrond artikelen