During the early boreal winter (northeast) monsoon (November–December), cold air frequently bursts out from intense Siberian highs toward the Chinese coast in response to the development and movement of a 500-hPa trough. The resultant strong low-level northwesterlies turn into northeasterlies across the South China Sea as “cold surges.” On interacting with the near-equatorial trough, mesoscale convective systems form north of the trough, normally giving rise to heavy downpours and severe flooding, mainly along the coastal stretch in the east coast states of Peninsular Malaysia. In November 2014, a 1-week-long episode of heavy downpours, producing more than 800 mm of rain, occurred along the coastal stretch of northeastern Peninsular Malaysia. However, during December 2014, two episodes of extreme rainfall occurred mostly over inland and mountainous areas of the east coast of Peninsular Malaysia, in particular across its northern sector. These two unusual events, which lasted a total of 11 days and delivered more than 1100 mm of precipitation, resulted in extreme and widespread flooding, as well as extensive damage, in many inland areas. Analysis shows that the stronger wind surges from the South China Sea due to very intense cold-air outbreaks of the Siberian high developed under ENSO-neutral conditions. In addition, the mesoscale convective systems that developed across the northeastern Indian Ocean (near northern Sumatra) in response to the propagation of a 500-hPa short-wave trough across the Indian subcontinent toward China were the combined factors for these unusual extreme rainfall and flooding events along the east coast of Peninsular Malaysia.
Heavy rain with its associated flooding is an annual occurrence along the east coast of Peninsular Malaysia (Fig. 1) during the northeast monsoon. In November 2014, an episode of heavy rainfall and ensuing flooding occurred along the coastal stretch of northeastern Peninsular Malaysia. However, from mid-December onward, two episodes of extreme rainfall caused widespread flooding in Kelantan, Terengganu, and Pahang on the east coast of Peninsular Malaysia. The unusual situation was that these long-lasting extreme rainfall events were concentrated over the catchment areas in the upper reaches of the Kelantan and Pahang River basins instead of near the lower reaches, as normal. A mountainous hydrological station (Gunung Gagau) in Pahang recorded 624 mm of torrential rain from 15 to 19 December and 1395 mm from 20 to 24 December 2014. These rainfall amounts were claimed to have return periods of 40 and 1000 yr, respectively (Sun Daily, 20 January 2015). The extreme rainfall situation was aggravated by the low water-holding capacity of the lower reaches, excessive logging and land clearing, as well as sediment buildup in river areas. In fact, excessive land clearing caused serious erosion and swift drainage of water, leading to widespread flooding. River levels during these two episodes far exceeded those of the 1967 and 2004 floods. At the height of the flood, more than 66 500 people were evacuated, at least 21 people died, and 250 000 lost their homes. Estimated damages cost about 1 billion Malaysian ringgits (RM), almost U.S. $300 million (source: Malaysian National Security Council).
The available records show that the above flood-associated extreme meteorological events during December 2014 are the first to occur in Peninsular Malaysia. Thus, understanding the features and processes of these rare episodes is essential in advancing the scientific knowledge for better flood mitigation to minimize the socioeconomic impacts in the future (Valipour 2012, 2016). This is the core objective and approach of this paper. Several recent studies carried out for the Malaysia region show that there has been an increase in the frequency and intensity of extreme rainfall events (Suhaila et al. 2010; Zin et al. 2010). In addition, a few previous case studies of extreme rainfall over Peninsular Malaysia by Tangang et al. (2008) and Juneng et al. (2007) were mainly focused on the extreme precipitation events caused by the interaction of northeasterly cold surges with the Borneo vortex, as well as the impact on the events by phenomena such as the Madden–Julian oscillation. The relationship between the rainfall variability over the Maritime Continent region and ENSO has been studied by Hendon (2003), McBride et al. (2003), and Tangang and Juneng (2004). These studies reveal that the winter monsoon is weak during El Niño years and is likely to be strong during La Niña years. Recently, Wu and Leung (2009), as well as Tubi and Dayan (2012), studied the impact of ENSO on the Siberian high, the East Asian winter monsoon, and the modulation of rainfall over southern China and Hong Kong. Unlike the results from the above-mentioned previous studies, the cases examined in this paper, as revealed in section 3, show that the December 2014 episodes of unusual cold surge–induced extreme rainfall occurred mostly over inland and mountainous areas in the northern sector of the east coast of Peninsular Malaysia and were due to the influences of the intense Siberian high during an ENSO-neutral year along with developing mesoscale convective systems in the northeastern Indian Ocean.
2. Data and methods
Surface and upper-air reanalysis data at 0000 UTC from the European Centre for Medium-Range Weather Forecasts (ECMWF) are used. These so-called ERA-Interim (Dee et al. 2011) data have a spatial resolution of 0.75° × 0.75° in latitude and longitude. We complement the reanalysis data with a wide range of satellite observations (Rahimi et al. 2015). We use monthly, daily, and 3-hourly Tropical Rainfall Measuring Mission (TRMM) precipitation at 0.25° resolution; the data were obtained from the Goddard Earth Sciences Data and Information Services Center (GES-DISC) Interactive Online Visualization and Analysis Infrastructure (Giovanni). TRMM data (Bookhagen 2010) are merely used to complement the precipitation analysis in data-sparse land and sea areas throughout the region. Furthermore, selected daily/hourly gridded Multifunctional Transport Satellite-2 (MTSAT-2) infrared channel 1 (IR1) data, which are available from the Center for Environmental Remote Sensing (CEReS), Chiba University, Chiba, Japan, are also utilized. These data, which have a resolution of 0.04°, cover an area from 60°S to 60°N and from 85° to 205° longitude. The measured blackbody temperatures (TBBs) of these infrared images are color enhanced by suppressing those TBBs warmer than −20°C based on the convection index, as described by Chang et al. (2005). The coldest cloud tops are associated with colors ranging from blue to orange to pink. The colder the cloud tops are, the higher the clouds and, most likely, the heavier the thunderstorms. For example, near the strongest IR gradients at the leading edge of an enhanced image is the location of heavy rain with cumulonimbus clusters. TBBs from −70° to −80°C indicate that the cloud top has penetrated the tropopause level (above 100 hPa or around 53 000 ft) in this region.
December features a strong northeast monsoon during the early boreal winter and the above-stated extreme meteorological events in 2014 also occurred in this month. ERA-Interim mean December data in terms of mean sea level pressure and 925-hPa winds from 1981 to 2014 are hence used to assess the influence of ENSO on the strength of the Siberian high, which triggers the incursion of cold surges toward the equatorial South China Sea (SCS). Composited anomalies are obtained based on the latest standard climatological period of 1981–2010. The oceanic Niño index (ONI), which is available online (http://www.cpc.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml), is used to identify El Niño, La Niña, and neutral events in the tropical Pacific, as tabulated in Table 1.
To identify heavy/extreme rainfall periods and their distributions, precipitation data from meteorological stations on the east coast of Peninsular Malaysia (Fig. 1) are used: Kota Bahru [6°10′N, 102°17′E, 4.6 m above mean sea level (MSL)], Kuala Krai (5°32′N 102°12′E, 68.3 m MSL), Kuala Trengganu (5°23′N 103°06′E, 5 m MSL), Kuantan (3°47′N 103°13′E, 15 m MSL), and Mersing (2°27′N 103°50′E, 43.6 m MSL). Only the Kuala Krai station (some missing data during December 2014 are noted) is representative of the inland hilly area, while the others are coastal stations. The climatological base period for the reanalysis data is chosen as 1981–2010 while that for TRMM is from 1998 to 2013.
To detect all of the surge-induced events due to the cold-air outburst from the Siberian high moving toward the equatorial South China Sea, we adopt the cold surge index from Chang et al. (2005), which is chosen as the averaged 925-hPa meridional wind between 110° and 117.5°E along 15°N (Fig. 5a, described in more detail below). By adapting the index definition from Chang et al. (2005), we further define the easterly surge index [zonal wind surge due to strengthening or equatorward movement in the subtropical ridge of the northwestern Pacific (Raman et al. 1978) as a result of a Siberian high outbreak] as the averaged 925-hPa zonal wind between 7.5° and 15°N along 120°E. A surge event occurs when either one of these indices exceeds 8 m s−1. The surge intensity is classified into weak, moderate, and strong categories for a surge index between 8 and 10, 10 and 12, and greater than 12 m s−1 [as adopted from Chang et al. (2005)], respectively. Cold and/or easterly surges are necessary but not a sufficient condition to ensure the impingement and sustenance of surge-induced heavy rainfall events toward the east coast of Peninsular Malaysia.
3. Analysis and discussion
a. Basic climatological features
During November, the surface Siberian high over the northern Asiatic landmass (Fig. 2a) becomes a stable feature after its gradual buildup since September (Tick and Abu Samah 2004). The overlying 500-hPa trough aligns along the coast. Development of this trough favors the intensification of the Siberian high near Lake Baikal, causing a couple of cold-air outbreaks that advance toward the coast. Generally, the resultant low-level strong northwesterlies flow toward the East China coast, turn into northeasterlies as “cold surges,” and penetrate seaward across the South China Sea and toward equatorial Southeast Asia. This is made evident by the “cold tongue” of sea surface temperature (SST) aligned parallel to the coast of Vietnam with an isoline of 28°C stretching as far south as 6°N. Higher SSTs are consequently found off the coast of the Malaysian states of Sarawak and Sabah. The seaward portion of the near-equatorial trough is over the warmer SST region, which is an important force in the generation of low-level convergence (Lindzen and Nigam 1987). The near-equatorial trough in the equatorial SCS is oriented almost zonally around 5°N. Interactions among the strong northeasterlies with the trough give rise to strong cyclonic shears and the consequent development of mesoscale or synoptic-scale convective systems north of the trough (Ooi et al. 2011). High mean monthly rainfall totals of about 600 mm (Fig. 2b) are found along the coastal stretch of northeastern Peninsular Malaysia.
By December, the 500-hPa trough has become more pronounced (Fig. 2c). Mean sea level pressures over central Asia and China have increased from November, leading to the intensified north–south and east–west pressure gradients. As shown in Fig. 3a, a strong Siberian high or a split of this high moves southward and/or eastward (Chan 2005), giving rise to more pulses of northeast/east wind surges. Over the South China Sea, low-level northeasterly flow is stronger and persistent. Colder SSTs with an isoline of 27°C are located off the southeastern coast of Vietnam. The near-equatorial trough is now anchored around 3°N. Rainfall has increased in the SCS over the area of strong low-level northeasterly winds. Higher mean monthly rainfall of about 800 mm extends southward along the east coast of Peninsular Malaysia (Fig. 2d).
El Niño–Southern Oscillation (ENSO) has an influence on the intensity of the Siberian high (Wu and Leung 2009; Tubi and Dayan 2012) and thus modulates the cold surges and the associated monsoonal rainfall in the equatorial South China Sea region (Ooi 1999). This is clearly reflected by the anomalous mean sea level pressure and 925-hPa wind patterns during the strong early winter monsoon month of December (Figs. 3b–d). As shown in Fig. 3b, a generally negative mean sea level pressure anomaly west of Lake Baikal within the domain of 40°–60°N and 80°–120°E (as indicated by the rectangular box) implies that the Siberian high is weak during El Niño conditions. Correspondingly, the cold surges triggered by the Siberian high toward the equatorial SCS are weak, as shown by the anomalous south-southwesterlies across the SCS. During La Niña conditions (Fig. 3c), the Siberain high is strong, particularly in the vicinity and east of Lake Baikal. Hence, cold surges are strong, as revealed by the light anomalous southerlies across the SCS. Under ENSO-neutral conditions (Fig. 3d), even though the Siberian high is intense, the significant positive mean sea level pressure anomaly west of Lake Baikal is noted to align more zonally. In other words, the cold surges are weaker than those during La Niña conditions but are stronger than those during El Niño conditions, as supported by the presence of some weak anomalous southwesterlies across the SCS.
The Arctic Oscillation (AO) is also known to be related to the Siberian high intensity (Tubi and Dayan 2012). To evaluate the statistical significance of this and the above-mentioned observed anomalies with respect to the December rainfall anomalies in Peninsular Malaysia, multiple linear regression is performed on the area-averaged mean sea level pressure anomalies within the Siberian high domain (SHA; obtained from the rectangular box in Fig. 3) using ONI and the Arctic Oscillation index (AOI; available online at http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao_index.html) as proxies for two independent variables for ENSO and AO, respectively. The small p values of 0.33 and 0.42 for ONI and AOI, respectively, of the linear regression SHA = −0.42 × ONI − 0.24 × AOI + 0.25, show that the influence of ENSO is statistically more significant than that of AO even though the regression represents only about 22% of the variance, as indicated by the multiple R (Fig. 4a). Figure 4a also reveals that SHA is generally negative during El Niño years and positive during La Niña years. In addition, the regressed area-averaged rainfall anomalies (PMA) within the Peninsular Malaysia domain (defined as 2.5°–7.5°N, 102.5°–105.0°E, and obtained from the CMAP dataset) with SHA and SCA (area average of the mean sea level pressure within the southern China domain, i.e., 20°–30°N, 100°–120°E) as proxies give rise to the following relationship:
Apart from the increased variance of 32% (multiple R = 0.32) from this regression, the small p value of 0.17 from SHA as compared to that of SCA (i.e., 0.42) strongly indicates that PMA is greatly influenced by SHA (Fig. 4b). This is clearly visible in both the charts shown in Fig. 4.
b. Identification of heavy/extreme rainfall periods
A moderate-to-strong easterly surge occurred from 17 to 19 November 2014 when it initiated its gradual climb from 11 November (Fig. 5a) with pulses of weak cold surges on 13 and 19 November. Further examination of daily rainfall totals (Fig. 5b) and hydrographs at Kuala Kerai and Kota Bahru (Fig. 6) led to the identification of the first episode of surge-induced heavy rainfall along the east coast of Peninsular Malaysia to be from 13 to 20 November 2014. The flooding peaked at both locations around the alert level close to 19 November, indicating that the rainfall was normal. Rainfall data from a few meteorological stations that are widely spaced can only be used to assess the duration and areal extent but not the intensity of this episode.
By December 2014, four moderate-to-strong cold surges occurred intermittently together with five strong easterly surges preceding and/or trailing them (Fig. 5a). Using the same approach as above (Figs. 5b and 6), the complex mixed surges occurred between 11 and 19 December, as well as between 20 and 23 December 2014, and helped to identify the respective periods of 14–19 December and 20–24 December 2014 as the second and third episodes of the surge-induced extreme rainfall along the east coast of Peninsular Malaysia. Two significant flood peaks above the danger level at both Kuala Kerai and Kota Bahru (Fig. 6) around 18 and 24 December, respectively, not only reflect the extreme intensity of the rainfall but also its prolonged duration.
c. Case studies
1) Episode 1: 13–20 November 2014
A pronounced 500-hPa trough from Lake Baikal with advecting cold air had reached 35°N off the China coast on 13 November 2014 (Fig. 7a). The trough then moved east-northeastward with gradual weakening around 14 November. A cutoff low of mild cold air from northern China approached 40°N along the coast on 16 November and then disappeared shortly on its eastward track.
Underneath the upwind side of the 500-hPa trough is the region favored for surface anticyclonic development. A large Siberian high with a center pressure of about 1040 hPa was detected southwest of Lake Baikal on 12 November (not shown). The pressure of the center decreased to around 1030 hPa subsequently. In response to the progression of the 500-hPa trough, the 1016- and 1012-hPa isobars show an alternating pattern of a southward push and a northward retreat along 110°E from 13 to 20 November 2014 (Figs. 7 and 8). Strong northeasterly wind bands of 20–30 kt (where 1 kt = 0.51 m s−1) at 925 hPa persisted south of 20°N until 5°N. In general, the strong wind band was aligned from northeast to southwest. The near-equatorial trough was initially around 3°N and later dipped to 1°N on 19 November (Fig. 8c) as a result of a mild cold-air outburst along the Chinese coast on 16 November (Fig. 7d). Along 95°E (over the Indian Ocean) and north of the equator, a 15–25-kt easterly/east-northeasterly wind band appeared between 5° and 10°N from 17 to 20 November, indicating transient convective development there (Figs. 9e–h). This arose from the tightening of the pressure gradient in the Bay of Bengal on 18 and 19 November in response to the mild development of a 500-hPa short-wave trough across the Indian subcontinent (Figs. 8b,c).
Figure 10a shows vertical profiles of 0000 UTC specific humidities during November 2014 at 15°N 115°E (close to the source of cold surge), 10°N 110°E (close to southern Vietnam) and 5°N 105°E (close to Peninsular Malaysia), respectively. Specific humidity is used as a tracer of moisture and its differing time–height sections here can reflect the abundance, depth, and advection of moisture with respect to time at the selected locations. Significant moisture advection is an important forcing for not only convective initiation but also for generating deep moist convection (Doswell 1982). During the period of this heavy rain episode (indicated by the rectangular box in the figure), high moisture content (greater than 16 g kg−1 below 900 hPa) increased from the north toward the south, and its depth of availability also increased from 700 to above 400 hPa. Time series of the mean sea level pressure close and parallel to the rainband (Fig. 10b) shows that the pressure was lowest on 16~17 November, and the rainfall was noted to be heavier (Fig. 5b). This implies that the release of latent heat, which warms the atmosphere, was the principal factor controlling the mean sea level pressure. As shown in Fig. 10c, cyclonic vorticity penetrated above 300 hPa during the initial period of the episode but dropped to below 600 hPa after 17 November. In addition, significant cyclonic vorticity with core values of more than 2 × 10−5 s−1 occurred below 600 hPa initially and was reduced to about 1 × 10−5 s−1 after 17 November. Fig. 10c also indicates clearly that upper divergence, as represented by the weak anticyclonic shear above 300 hPa, is not a crucial factor in the development of mesoscale convective systems.
Figure 11 shows the daily TRMM precipitation data from 13 to 20 November 2014. The major rainbands were generally lodged along and off the coastal areas. The trough was initially located around 3°N from 13 to 15 November (Figs. 7a–c). The interaction of the cyclonic shear north of the trough with the surface discontinuities along the coast led to the coastal alignment of the convective cloud band. The rainband intensified and broadened on 18 November 2014 in response to a mild cold-air outbreak from the southern China coast on 16 November 2014. It became zonal and penetrated farther inland on 19 and 20 November 2014. Apart from the weakening and splitting of the 925-hPa winds along 105°E into northeasterly and northerly directions north of the trough after 17 November (Fig. 8a), the transient convective development between 5° and 10°N in the neighboring Indian Ocean (Figs. 9f–h) provided the momentum to cause the rainband in Peninsular Malaysia to be advected inland.
2) Episode 2: 14–19 December 2014
At 500 hPa, a pronounced northeast–southwest-oriented cold core trough moved southwestward, passed Lake Baikal on 15 December (Fig. 12b), deepened, and advected farther south to near 30°N (Fig. 12c) on 16 December; it then moved out from along the Chinese coast and weakened gradually on 17 December (Fig. 12d) before its subsequent east-northeastward track toward the sea.
A large intense Siberian high with center pressure of more than 1040 hPa was noted southwest of Lake Baikal on 14 December and moved southeastward toward southern China on 16 December with a subsequent reduction in core intensity and size. In response to the progression of the 500-hPa trough and the resultant outburst of intense cold air from the Siberian high toward the SCS and western North Pacific, the tight pressure gradient between the 1020- and 1012-hPa isobars shows pulsating fluctuations during this period (Figs. 12 and 13a,b). A strong northeasterly/east-northeasterly wind band of 20–45 kt at 925 hPa appeared, particularly across the SCS, and the 20-kt winds reached as far south as 5°N. Owing to the initial strong momentum, the near-equatorial trough was initially around 1°N and later shifted slightly northward to 2°N on 19 December. A 500-hPa cutoff low was visible over the northwestern part of the Indian subcontinent on 14 December (Fig. 12a). The prominent short-wave trough thus formed and then moved east-southeastward before weakening eventually over Indochina on 19 December (Fig. 13b). The ridge behind the trough intensified the surface high over the northern reaches of the Indian subcontinent, leading to the tightening of the pressure gradient (as reflected by the white 1016- and 1018-hPa contour bands) and strengthening of the 925-hPa northeasterly/east-northeasterly winds to 20–30 kt over the Bay of Bengal. The resultant increased cyclonic wind shear close to the northern near-equatorial trough across the Indian Ocean caused the intensification of convective activity off of northern Sumatra (Fig. 14).
In comparison with the November rainstorm (Fig. 10a), the time–height section of specific humidity during the period of this episode shows that moisture was already abundant and available throughout the troposphere from the South China Sea toward Peninsular Malaysia (Fig. 15a). As shown in Fig. 15b, the mean sea level pressure was the lowest on 16 December near 6°N, implying that extreme rainfall was confined mainly toward the northern part of Peninsular Malaysia’s east coast. Though the cyclonic vorticity penetrated up to 300 hPa (Fig. 15c), its core value is only 1.0 × 10−5 s−1 in the lower troposphere.
During the period from 14 to 19 December 2014, the broad rainbands (Fig. 16) were noted to penetrate more inland and southerly as compared with the November episode. This was partly due to the more southward location of the near-equatorial trough. In addition, apart from the impact of the convective activities off of northern Sumatra on 17 December 2014, the consecutive strong cold and easterly surges (Fig. 5a) provided sufficiently strong northeasterly wind momentum to overcome the effects of coastal surface discontinuity and advected the rainstorm farther inland, with the maximum precipitation core lying along the coast and its immediate sea area.
3) Episode 3: 20–24 December 2014
After the second episode, another prominent trough at 500 hPa with intense cold air advected from Lake Baikal moved toward the Chinese coast, reaching slightly south of 30°N on 21 December 2014 (Fig. 13d), before its subsequent east-northeastward track toward the sea.
A large intense Siberian high with a center pressure of more than 1040 hPa was detected southwest of Lake Baikal on 20 December (Fig. 13c). In response to the progression of the 500-hPa trough, this more intense Siberian high moved south-southeastward and burst out toward the SCS before its subsequent eastward movement to the western North Pacific on 22 December (Fig. 17a). In this episode, 20–45-kt northeasterly/easterly winds at 925 hPa persisted not only in the western North Pacific but also in the SCS. Winds of 30–35 kt reached as far south as 5°N in the SCS while the near-equatorial trough hovered around 2°N. A significant feature to note is that the intense Siberian high also burst out from the south-southwestward simultaneously, heading toward the Andaman Sea on 22 December and strengthening the northeasterlies there. The increased cyclonic shear in the southern Andaman Sea thus caused the already intensified convective activity off northern Sumatra during episode 2 to develop into a cyclonic disturbance and move slowly westward (Fig. 18).
As in episode 2, the time–height section of specific humidity during the period of this episode shows that moisture was already abundant and available throughout the troposphere from the South China Sea toward Peninsular Malaysia (Fig. 15a). In Fig. 15b, this episode shows a deeper drop in the mean sea level pressure on 22 and 23 December. The lowest pressure occurred over the central and southern parts of the east coast of Peninsular Malaysia, implying that heaviest rainfall amounts were confined there. Two strong and deep centers of cyclonic vorticity (on the order of more than 2.0 × 10−5 s−1) existed and, in particular, penetrated above 100 hPa on 22–23 December (Fig. 15c) in association with the occurrence of very deep and intense convective activity.
Figure 16 also shows the daily TRMM precipitation data from 20 to 24 December 2014. The rainbands were broader, penetrated farther inland, extended more southward, and lasted from 21 to 23 December, as compared with the second episode. During this episode, the consecutively stronger cold and easterly surges (Fig. 5a) provided the greater wind momentum necessary to overcome the effects of coastal surface discontinuity to advect the rainstorm farther inland, where it became lodged against the mountain ranges located near the west coast of Peninsular Malaysia. Meanwhile, the developing and westward-moving cyclonic disturbance off of northern Sumatra provided the additional momentum to cause the rainstorm to remain quasi-stationary over the inland area. Consistent with the lowest mean sea level pressure and deep and strong cyclonic vorticity, extensive and maximum precipitation occurred on 23 December 2014. Three-hourly enhanced MTSAT-2 TBB satellite images (Fig. 19, left) and the associated 3-hourly accumulated TRMM precipitation (Fig. 19, right) revealed the complex cloud intensity changes and distribution pattern variations. The distributions of varying intense precipitation cores are noted to be irregular in space and time while the enhanced MTSAT-2 TBB satellite imagery shows increased convection at night due to the nocturnal radiational cooling of cloud tops (McBride and Gray 1980), reaching a maximum during the early morning.
4) Precipitation during the episodes
During November 2014, high monthly precipitation totals of more than 800 mm were noted along the northeastern coast and its neighboring sea area of Peninsular Malaysia’s east coast (Fig. 20a, left). The anomaly for the November rainstorm (Fig. 20a, right) shows a significant increase of more than 500 mm. During December 2014, very high monthly precipitation totals of more than 1400 mm were noted inland as well as along the coastal region and its neighboring sea area (Fig. 20b, left), extending as far south as 2°N. The anomaly of two December rainstorms (Fig. 20b, right) shows a very significant increase in precipitation of more than 800 mm.
4. Summary and concluding remarks
The extreme rainstorms investigated in this study occurred during 2014, which was an exceptional ENSO-neutral year. The anomaly of the Siberian high showed that the high was very intense during December 2014. The cold air burst out strongly south-southeastward across the SCS and headed southeastward toward the western North Pacific. The anomalous light southerlies across the SCS indicate that the corresponding cold surges were very strong.
The composite average mean sea level pressure (Fig. 21) reveals that the core intensity of the Siberian high during the December rainstorm period was about 1040 hPa as compared with that of 1030 hPa in November. The south-southeastward outburst of the intense cold air from the Chinese coast gave rise to tighter pressure gradients both in the SCS and the western North Pacific. The impact of a 500-hPa short-wave trough over the Indian subcontinent also resulted in a tighter pressure gradient over the Bay of Bengal. This and the south-southwestward outburst of the Siberian high toward the Andaman Sea of the Indian Ocean led to stronger northeasterlies and an eventual development of a cyclonic circulation southeast of Sri Lanka. Hence, as shown in Fig. 22, the prevalence of very strong northeasterlies due to very strong and consecutive cold and easterly surges not only provided the necessary momentum to overcome the coastal discontinuity effect but also strong to very strong cyclonic shear north of the near-equatorial trough, for the development and sustenance of intense and deep mesoscale convective rainstorms over the inland areas of the east coast of Peninsular Malaysia. The developing and westward-moving cyclonic disturbance off of northern Sumatra provided the impetus for the intensification of the rainstorms to be located farther inland, contributing to the rapid and high river levels due to the more confined drainage basin. These results are clearly supported by the anomalous southeasterlies across Peninsular Malaysia and toward the cyclonic circulation southeast of Sri Lanka.
Malaysia and the neighboring equatorial South China Sea are located in the western part of the Maritime Continent. The distribution of the deep mesoscale convective systems in this monsoonal region is strongly influenced by its unique topographic orientation and thus can differ significantly from those of other regions. This is a data-sparse sea area, and the lack of long and comprehensive land data records pose severe limitations to our understanding of the key physical processes and the ability to provide reliable monsoon and rainfall forecasts. This is further compounded by the short or instantaneous time scale of the strong cold-air outbursts from China toward the equatorial SCS, which are attributed to the “inertia–gravity wave” (Lim and Chang 1981). Hence, there is a need to depend strongly and jointly on satellite data analysis, the continual compilation of case studies of similar or differing as well as unusual patterns, simple monsoon initiation indices, monitoring of the development and movement of 500-hPa troughs across surface features based on global predictive models or reanalysis data, and assessing the exceptional impacts of climate phenomena such as ENSO-neutral conditions. Only then can we understand better the basic features and processes generating some of the significant meteorological events in this region as a way to disseminate new and helpful information related to improved qualitative forecasting, warnings, and awareness.
ERA-Interim reanalysis data were provided by the ECMWF while TRMM and MTSAT-2 IR1 data were obtained from Giovanni and CERES, respectively. The data analysis supporting this project was carried out using the Grid Analysis and Display System (GrADS) software from OpenGrADS. We thank Dr. Peter Braesicke of Karlsruhe Institute of Technology, Karlsruhe, Germany, for his constructive feedback. This research study has been funded by the Fundamental Research Grant Scheme (FRGS) FP049-2013 and Trans Disciplinary Research Grant (TRGS) Project 417 TR001A-2015, as well as Institute of Ocean and Earth Sciences Research Grant (IOES) IOES-2014B of the Ministry of Higher Education, Malaysia. This work has also been strongly supported by the Vice Chancellor of the University of Malaya and the Director General of the Malaysian Meteorological Department.