1. Introduction
Borneo, located in the Maritime Continent in Southeast Asia, is the third largest island in the world. Its climate is strongly affected by the interannual variability of El Niño–Southern Oscillation (ENSO) (Ropelewski and Halpert 1987, 1996). However, the ENSO impact on Borneo rainfall is seasonally dependent: during an El Niño year anomalous dry conditions characterize the austral spring [September–November (SON)], giving way to a dipolar structure of wet/southwest versus dry/northeast during the austral summer [December–February (DJF)] (Aldrian and Susanto 2003; Juneng and Tangang 2005). Qian (2008) found that rainfall in this region is mostly concentrated over islands, especially over mountains, indicating the important role of diurnal cycle of land–sea breezes and mountain–valley winds in precipitation processes. The local precipitation and its movement also depend on wind advection and propagation, as found in the analysis of Tropical Rainfall Measuring Mission (TRMM) precipitation radar data by Ichikawa and Yasunari (2006). The windward terrain lifting and the rain shadow effect may also play important roles in this monsoonal region (Chang et al. 2005b). On the other hand, the intensity of the Asian–Australian monsoon and the prevalent weather regimes are subject to the influence of the Pacific–Indian Ocean basin-scale interannual variability of ENSO (Aldrian and Susanto 2003; Juneng and Tangang 2005; Giannini et al. 2007; Moron et al. 2009, 2010; Robertson et al. 2010). Multiscale interactions among ENSO, monsoons, and diurnal cycle are found to be key to understanding dipolar patterns of ENSO rainfall variability over the nearby Java Island. Moron et al. (2009) and Qian et al. (2010) found that quiescent weather types are more prevalent during the austral summer of El Niño years, leading to an enhancement of the diurnal cycle and positive seasonal rainfall anomalies over the inland mountains, surrounded by dry anomalies over the lowlands. We hypothesize that analogous multiscale interactions can explain the dipolar anomalies over Borneo during DJF of ENSO years as well, and the aim of the current study is to explore this hypothesis using a weather-typing analysis similar to that used in Moron et al. (2009) and Qian et al. (2010).
Figure 1 shows the topography of the island of Borneo, which spans the equator from about 4°S to 7°N, and 108° to 118°E, composed of the states of Sarawak and Sabah of Malaysia and the nation of Brunei in the north, and the Indonesian provinces of West, Central, South, and East Kalimantan in the south. Relatively speaking, the island is narrower in the north and wider in the south, with mountains in the central area and lowland near the coast. Specifically, a central mountain range runs from northeast to southwest along the Malaysia/Indonesia border in the north with some high peaks over 2000 m above the sea level, and then extends south to Central Kalimantan with relatively low heights in the Kalimantan territory. Two lower and smaller branch mountain ranges exist at both sides of the larger central mountain range, one stretching toward the west along the border of Sarawak and West Kalimantan and the other extending toward the southeast all the way into the southern tip of South Kalimantan. Northwest of Borneo is the South China Sea, the Malay Peninsula, and Sumatra Island. The Java Sea and Java Island are located in the south, while Makasar Strait and Sulawesi Island are to the east. As will be seen later, the geography and topography are very important for rainfall distribution and variability over Borneo.
Terrain heights (m) over Borneo Island and surrounding regions based on the USGS 2-min observation. Black contours show terrain heights of 500 m.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Terrain heights (m) over Borneo Island and surrounding regions based on the USGS 2-min observation. Black contours show terrain heights of 500 m.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Terrain heights (m) over Borneo Island and surrounding regions based on the USGS 2-min observation. Black contours show terrain heights of 500 m.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
The spatial and temporal rainfall variability over Borneo will be analyzed using observational data. Section 2 lists the data used for the analysis. Results are given in section 3. We first describe observations which confirm the ENSO-related dipolar pattern of rainfall anomalies over Borneo shown by Aldrian and Susanto (2003) and Juneng and Tangang (2005). The spatial distribution of rainfall is then analyzed from the perspective of weather-typing analysis based on low-level wind circulation patterns over the region, using the classification of Moron et al. (2009). Then, the progressive diurnal cycle of rainfall formation and development is analyzed in detail for each of these weather types. In light of the difference in the spatial distribution of rainfall among the weather types, the spatial pattern of seasonal rainfall variability over Borneo is elucidated by considering the ENSO influences on the frequency of occurrence of the weather types. Conclusions are drawn in section 4.
2. Observed data
Satellite estimated precipitation data cover both land and ocean, which is important for analyzing rainfall patterns across the coastline of Borneo. We used a satellite-estimated rainfall dataset with a quarter-degree grid resolution: the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC) Morphing Technique (CMORPH; Janowiak et al. 2005). The 3-hourly high-frequency CMORPH data (available starting from December 2002) can be used to study the diurnal cycle of rainfall associated with land–sea breezes and mountain–valley winds (Qian 2008). Although there are not enough years of CMORPH data available (only about 10 years or so) for the ENSO composite analysis of interannual variability, but hundreds to thousands of days included in the CMORPH record are sufficient for a statistically meaningful analysis of daily weather types and the diurnal cycle of precipitation.
The monthly Global Precipitation Climatology Centre (GPCC) precipitation data (0.5° × 0.5°; available online at http://iridl.ldeo.columbia.edu/SOURCES/.WCRP/.GCOS/.GPCC/) in the period of 1901–2007 have been used for a composite analysis of ENSO impact on the Borneo rainfall. This dataset may smooth out the “true” variability because there are fewer rain gauges in Borneo than the number of 0.5° × 0.5° grids, especially in the center of the island.
The daily 850-hPa winds of National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) Reanalysis I Project (1979–present, referred to as NRP herein; Kalnay et al. 1996; Kanamitsu et al. 2002) were also used in the composite analysis of seasonal wind anomalies and in the clustering analysis of daily weather types.
The Quick Scatterometer (QuikSCAT) sea wind data (around 10 m above ocean only, twice daily with morning and evening passes, 0.25° × 0.25°, available starting from 1999; Liu 2002) have been used to roughly illustrate the land–sea breezes around Borneo.
The Extended Record Sea Surface Temperature (ERSST; 1981–2010, 1° × 1°) (Smith and Reynolds 2005) global dataset was used for ENSO composite of regional SSTs.
Based on the availability of the datasets, there are two sorts of analyses in the paper: one is of seasonal means, which involves multidecades of data (such as GPCC, NCEP reanalysis, and ERSST) for ENSO composite analysis, and the other is of daily or 3-hourly data (such as CMORPH and QuikSCAT), which are good enough for daily weather typing analysis and diurnal cycle analysis but the record is too short for the ENSO composite analysis of interannual variability.
3. Results
a. Opposite signs of rainfall variability over southwest and northeast Borneo associated with ENSO
Figures 2a and 2b show the climatology of the GPCC precipitation and low-level winds (at 850 hPa) of the NRP and their composite (or averaged) anomalies typical of El Niño years (middle panels) and La Niña years (lower panels) in September–November (left panels) and December–February (right panels). In the ENSO anomaly panels (Figs. 2c–f), those significant above a two-sided 90% level according to a Student’s t test are shown by dark colors (while those statistically insignificant are shown by the lightest brown or green colors). To enhance confidence in the statistical analysis, we use data of as many years as possible in the GPCC record in the composite analysis. The ENSO years used in the composite are the strongest 21 warm (25 cold) ENSO years during the period of 1901 to 2007 with the maximum 3-monthly SST anomaly in DJF in the Niño-3.4 region (5°S–5°N, 120°–170°W) being greater than 1°C (less than −1°C).
(top) Climatology, (middle) El Niño − climatology, and (bottom) La Niña − climatology of the GPCC (1901–2007) precipitation (mm day−1; shaded), and NRP winds (vector) at 850 hPa, for (left) SON and (right) DJF. ENSO developing years are denoted by (0). Rainfall anomalies significant (insignificant) above the 90% level of a Student’s t test are shown by the darker (lighter) colors.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
(top) Climatology, (middle) El Niño − climatology, and (bottom) La Niña − climatology of the GPCC (1901–2007) precipitation (mm day−1; shaded), and NRP winds (vector) at 850 hPa, for (left) SON and (right) DJF. ENSO developing years are denoted by (0). Rainfall anomalies significant (insignificant) above the 90% level of a Student’s t test are shown by the darker (lighter) colors.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
(top) Climatology, (middle) El Niño − climatology, and (bottom) La Niña − climatology of the GPCC (1901–2007) precipitation (mm day−1; shaded), and NRP winds (vector) at 850 hPa, for (left) SON and (right) DJF. ENSO developing years are denoted by (0). Rainfall anomalies significant (insignificant) above the 90% level of a Student’s t test are shown by the darker (lighter) colors.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
In the average in SON (Fig. 2a), rainfall is heaviest over northwestern Borneo. In DJF (Fig. 2b), heavy rainfall (>8 mm day−1) extends to southern Borneo covering the West, Central, and South Kalimantan provinces of Indonesia, along with the seasonal southward migration of the intertropical convergence zone (ITCZ) when the Asian monsoon transfers to the Australian monsoon. The heavy rainfall near the northwest coast of Borneo has been attributed to gravity wave activity (Johnson and Priegnitz 1981; Houze et al. 1981; Mapes et al. 2003) and the topographic rainfall over the coastal mountain slopes caused by terrain-lifting (Chang et al. 2005a; Ichikawa and Yasunari 2006). In general, the East Kalimantan coast receives less rainfall than West Kalimantan because of gravity waves originating from the high mountains in Sulawesi, as argued by Wu et al. (2009). The gravity waves from Sulawesi propagate westward and upward, arriving at the east coast of Borneo in afternoon hours, generating warm temperature aloft over East Kalimantan, thus stabilizing the atmosphere (more stable vertical profile of temperature with warmer and lighter air in upper layers of the atmosphere) and prohibiting precipitation over eastern Borneo. Besides the impacts of the island-scale gravity waves, smaller-scale features associated with the shapes of coastal lines (i.e., headlands or bays) have also left their marks on the finescale spatial structures of rainfall distribution, which will be discussed later by using CMORPH estimates of rainfall.
It is generally thought that rainfall over Indonesia is less than normal in El Niño years, especially in the premonsoon transitioning season (Haylock and McBride 2001; Hendon 2003; Moron et al. 2009, 2010). But as also pointed out by others, rainfall pattern is very complex over the vast region of the Indonesian islands because of the very complex topography of islands and mountains [293 different climate areas are classified in Indonesia according to the Badan Meteorologi Klimatologi dan Geofisika of Indonesia (BMKG) in terms of long-term climatology]. Similarly, the interannual variability of rainfall over Indonesia is also by no means spatially homogeneous or temporally coherent, especially in the peak rainy season (Haylock and McBride 2001; Qian et al. 2010).
Over Borneo, the regional-scale rainfall anomalies in the peak rainy season are opposite in sign between southwestern and northeastern Borneo, as shown in the observational studies of Juneng and Tangang (2005) and Aldrian and Susanto (2003). By using the GPCC precipitation data (1901–2007), we also found that the spatially homogeneous (or temporally coherent when considering the statistical significant part) rainfall anomalies over Borneo in SON with negative (positive) anomalies in El Niño (La Niña) years (Figs. 2c,e). In DJF, however, opposite signs of rainfall anomalies are found over a smaller area in southwestern Borneo and a larger area from central to northeastern Borneo (Figs. 2d,f), consistent with other ENSO impact studies (Aldrian and Susanto 2003; Juneng and Tangang 2005). Similar southwest–northeast dipolar patterns of rainfall anomalies were obtained based on GPCC in 1979–2007 (not shown). The anomalies in El Niño years are generally more pronounced than those in La Niña years; this asymmetry between the warm and cold ENSO events is also in line with the regional climate variability study over Java (Qian et al. 2010) and other large-scale impact studies (Hoerling et al. 1997; Kumar and Hoerling 1998).
Interesting differences are noted between the narrower north Borneo and the wider south Borneo in the ENSO composites of rainfall anomalies (Figs. 2c–f). In SON, the larger and statistically significant anomalies are in south Borneo. However, in DJF, the larger and statistically significant anomalies are in north Borneo, whereas in south Borneo there is an east–west dipolar pattern (which is statistically less significant in La Niña years than in El Niño years). In terms of the 850-hPa monsoonal winds, the zonal winds change sign north and south of about the equator (or between the equator and 2.5°N; for convenience, the areas north and south of 1°N in Borneo are called north and south Borneo, respectively, in this paper) (Figs. 2a,b). For example, in DJF, easterlies (westerlies) are north (south) of 1°N. In El Niño years, the wind anomalies are easterlies (Fig. 2d, their magnitudes are larger in the south than in the north). Therefore, in El Niño years, the monsoonal wind speed is larger (smaller) than normal in north (south) Borneo.
The inverse relationship between monsoonal wind speed and local diurnal cycle of land–sea and mountain–valley breezes (i.e., the stronger the monsoon, the weaker the diurnal cycle; therefore, we call it the “monsoonal damping effect” for convenience), which was proposed in Qian et al. (2010) to explain a dipolar rainfall variability over Java Island, can be used to explain the rainfall anomalies in north Borneo (which, like Java, is relatively narrow). For instance, in DJF, the strengthened (weakened) mean monsoonal wind speed in north Borneo in El Niño (La Niña) years will exert more (less) disruption to the local diurnal cycle, so therefore the sea- and valley-breeze convergence will be weaker (stronger) than normal and thus produce less (more) rainfall in north Borneo in El Niño (La Niña) years, as can be seen in Figs. 2d and 2f, respectively.
Besides the monsoonal damping effect, there is anomalous large-scale sinking (rising) of the Walker circulation over the Maritime Continent in El Niño (La Niña) years [see Bjerknes (1969) and Giannini et al. (2007); for brevity, we call it the “Walker circulation effect” hereafter]. In north Borneo, the monsoonal damping effect and the Walker circulation effect are of the same sign and thus reinforce each other in influencing local rainfall in DJF. Therefore, the rainfall anomalies in northern Borneo in DJF are large in magnitude and statistically significant.
The principle of the above discussion also applies for north Borneo in SON (Figs. 2c,e). The direction of the mean seasonal winds at 850 hPa in SON (Fig. 2a) is opposite to that in DJF, blowing from east to west in south Borneo and then veering toward the northeast in north Borneo. Therefore, in SON the wind anomalies in El Niño years (Fig. 2c) act to reduce mean wind speed in north Borneo, thus strengthening the diurnal cycle and sea-breeze convergence and increasing rainfall (by the monsoonal damping effect), which is opposite to the large-scale sinking and drying effect over the Maritime Continent by the El Niño (i.e., the Walker circulation effect). Hence, the large-scale Walker circulation effect and the local island-scale monsoonal damping effect counteract each other (the former is slightly stronger), and thus the negative rainfall anomalies are small in magnitude and statistically insignificant in north Borneo (Fig. 2c). The La Niña year anomalies (Fig. 2e) in north Borneo can in some extent be explained similarly.
The rainfall anomalies in SON in south Borneo are statistically significant (especially in El Niño year anomalies shown in Figs. 2c). Note that the winds are strengthened (weakened) in El Niño (La Niña) years there. Therefore, the sea-breeze convergence is weakened (strengthened) by the monsoonal damping effect and rainfall is reduced (enhanced) in south Borneo, especially near the coast, in El Niño (La Niña) years. Here, the Walker circulation effect and monsoonal damping effect are of the same sign, reinforcing each other.
Having explained the interannual rainfall variability over north Borneo in SON and DJF and over south Borneo in SON, now let us turn to our major task: to understand the puzzling dipolar pattern of rainfall anomalies associated with ENSO in south Borneo in DJF (Figs. 2d,f). In general, wind anomalies are northeasterly in El Niño years and southwesterly in La Niña years. Therefore, considering Borneo Island as a whole, it is interesting to note that positive rainfall anomalies are located downwind (i.e., in the wake area) of the 850-hPa wind anomalies for both the El Niño and La Niña year composites. In contrast, negative rainfall anomalies are on the windward side of the island in DJF. This suggests that coastal processes associated with land–sea breezes may play an important role in this spatial pattern of rainfall anomalies.
b. Weather types and their frequencies and rainfall distribution
The reason for this dipolelike pattern of rainfall anomalies in DJF in south Borneo, however, has not been clearly understood yet. We will try to explain this phenomenon by analyzing the rainfall formation and development process in different weather regimes while considering the ENSO impact on the variability of frequency of the weather regimes. The weather regime analysis was proved very useful in the study of rainfall variability over Java Island (Qian et al. 2010). In the current study, the ENSO impact on monsoon intensity and the favor of particular weather regimes under certain climate conditions will be shown to be a key to understand the spatial heterogeneity of rainfall variability over Borneo.
Daily weather types during the premonsoon and monsoon season (September–February) based on 28-yr (1979–2007) daily wind reanalysis data of NRP were identified over the Maritime Continent (12°S–6°N, 90°–130°E) by Moron et al. (2010), and are used here as the basis of the current study. A k-means cluster analysis was applied to unfiltered 850-hPa daily winds in the subspace of the nine leading principal components, which account for 75% of the total variance. Initially, a prescribed number k of clusters are specified and daily observations are agglomerated around centroids chosen from random seeds. Then, iterative clustering calculations are carried out to minimize the sum, over all clusters, of the within-cluster spread, to finally reach the clusters that localize relatively high concentrations in data distribution in the atmospheric phase space. The multiyear daily time sequence of weather types describes a systematic monsoonal evolution, as well as the variability at seasonal and interannual time scales (Moron et al. 2010; Qian et al. 2010).
The above clustering analysis based on 850-hPa winds produced five weather types (WT1–WT5) typical in the Maritime Continent region from September to February. These weather types are the same as those depicted in Moron et al. (2010) and Qian et al. (2010). However, the spatial pattern of rainfall and winds will be further studied by using new datasets such as the QuikSCAT sea winds as well as the CMORPH precipitation data (Fig. 4).
The time evolution of the frequencies of the five WTs is shown in Fig. 3 (from 1 September to 28 February). A weather type (e.g., WT3) could be scattered in any month in this period; that is, it could occur in both SON and DJF. On average (Fig. 3a), the dry WT1 mostly occurs in August–September and WT2 frequently takes place in the premonsoon season in September–November, and the three wet WTs gradually dominate in turn from late December to February. In the all-year average, the progressing of frequency from WT1 to WT5 is rather gradual along the time (Fig. 3a).
Seasonal evolution of frequencies of weather types WT1 to WT5, from September to February, for (a) all years, (b) El Niño years, and (c) La Niña years. Frequencies larger than 10%, 30%, 50%, and 70% are shaded from light to heavy. The date of 1 December is marked by a horizontal dashed line.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Seasonal evolution of frequencies of weather types WT1 to WT5, from September to February, for (a) all years, (b) El Niño years, and (c) La Niña years. Frequencies larger than 10%, 30%, 50%, and 70% are shaded from light to heavy. The date of 1 December is marked by a horizontal dashed line.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Seasonal evolution of frequencies of weather types WT1 to WT5, from September to February, for (a) all years, (b) El Niño years, and (c) La Niña years. Frequencies larger than 10%, 30%, 50%, and 70% are shaded from light to heavy. The date of 1 December is marked by a horizontal dashed line.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
However, the frequency of occurrence of the WTs changes dramatically in El Niño years (Fig. 3b) and La Niña years (Fig. 3c). In El Niño (La Niña) years, many more (fewer) days are of weather types WT2 and WT4 but fewer (more) days of WT3 and WT5—see also Fig. 9 of Qian et al. (2010). Specifically, in DJF, 64% of days are of the quiescent weather type WT4 in El Niño years as compared to 40% of days of WT4 in the all-year average. In contrast, very few days (9%) are of WT3 in the DJF of El Niño years, while the all-year averaged frequency of WT3 is 24%. What are the impacts on rainfall of these differences in the frequencies of WTs? Let us first compare the spatial pattern of the rainfall and winds of each WT in turn over Borneo.
Figure 4 shows the spatial distribution of daily-average rainfall and 850-hPa winds (red arrows), averaged over the number of days for each of the five WTs based on the CMORPH satellite estimates and the NRP reanalysis winds. The high-resolution CMORPH data are used to show the heterogeneous spatial structures of rainfall, which are due to the complex topography over Borneo and the diurnal cycle of winds near the coast (land–sea breezes) and mountains (mountain–valley winds), and their interaction with large-scale monsoonal winds, as will be shown in detail later. The twice-daily (morning and evening passes) QuikSCAT satellite observed sea surface winds are used to roughly illustrate the direction and magnitude of sea breezes (black arrows) around Borneo (evening pass minus the daily mean).
Averaged CMORPH precipitation (mm day−1, shaded), NRP 850-hPa winds (m s−1, red arrows), and QuikSCAT sea breezes at 10 m (black arrows, local evening press minus daily mean) for the five weather types (WTs). The thin black curves illustrate terrain heights of 250 m using the USGS 10-min data.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Averaged CMORPH precipitation (mm day−1, shaded), NRP 850-hPa winds (m s−1, red arrows), and QuikSCAT sea breezes at 10 m (black arrows, local evening press minus daily mean) for the five weather types (WTs). The thin black curves illustrate terrain heights of 250 m using the USGS 10-min data.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
Averaged CMORPH precipitation (mm day−1, shaded), NRP 850-hPa winds (m s−1, red arrows), and QuikSCAT sea breezes at 10 m (black arrows, local evening press minus daily mean) for the five weather types (WTs). The thin black curves illustrate terrain heights of 250 m using the USGS 10-min data.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
The sea breezes (averaged among the number of days in each WT) can be as strong as 2 m s−1 in some coastal areas, such as on the mountain slope in the northwest coast facing the South China Sea where rainfall is also heavy. Figure 4 indicates that the sea breezes tend to be stronger on the lee side (i.e., wake) of the island compared to the windward side, with respect to the direction of the daily mean 850-hPa winds in each WT. They are also stronger in some bay areas (with concave coastline), where the local thermally driven sea breezes are sheltered from the interference of the large-scale winds by local topography. Maximum rainfall is found over some of those regions (wake and bay areas), for example, near 3°N, 116°E in the northern East Kalimantan province (in WT1) or in the Malaysia coast of the state of Sabah facing northeast (near 5°N, 117°E in WT1 and WT3). Some headland regions (with convex coastline), such as the west corner of Borneo on the Indonesia–Malaysia border (about 2°N, 110°E) and the Sabah coast in northeast Malaysia, garner large amount of rainfall due to the converging angle of sea breezes. [The mechanism of sea breeze convergence is discussed in detail in Qian (2008).] The above evidence suggests that the diurnal cycle of sea breezes, coupled with the ambient large-scale daily winds, is important in explaining cross-island contrasts in the coastal rainfall over Borneo.
Comparison of the location of maximum rainfall area and the direction of the averaged NRP 850-hPa winds in Fig. 4 also shows that the maximum daily-average rainfall is located downwind with respect to the direction of the 850-hPa winds, on the leeside slope of the mountain range, which holds true for all five WTs. In WT1, large-scale winds are southeasterlies south of the equator and southwesterlies north of the equator and the maximum rainfall is over the northeast part of Borneo. In WT2, and WT4 as well, heavy rainfall is downwind on the western and southern slopes of the central mountain range; in the northwest in WT2 with southeasterly winds at 850 hPa, and in the southwest part in WT4 with northeasterly winds at 850 hPa. In contrast, in WT3 and WT5, heavy rainfall is over the eastern slope of the central mountain range; in the Central, South, and East Kalimantan provinces of Indonesia and the northeastern Sabah state of Malaysia in WT3 when westerly and southwesterly winds prevail at 850 hPa over most of Borneo; and on the southeastern slope of the central mountain range in WT5 with northwesterly winds at 850 hPa over the Kalimantan Provinces of Indonesia. The contrast is particularly strong between WT3 and WT4: more rainfall occurs in northeast Borneo in WT3 but in southwest Borneo in WT4. Therefore, more (less) frequent occurrence of WT4 (WT3) in El Niño years (Fig. 3) causes the positive and negative rainfall anomalies in southwest and northeast Borneo, respectively (Fig. 2d). Next, the physical processes that cause the heterogeneity of rainfall distribution will be analyzed in detail.
c. Mechanistic analysis of the northeast–southwest dipole: ENSO, weather regimes, and diurnal cycle and their role on the rainfall distribution and variability over Borneo
The analysis in the previous section indicates that both the diurnal cycle and the direction of horizontal wind might be important factors in shaping the spatial rainfall pattern in the five WTs. In this section, we will examine the diurnal cycle of the 3-hourly CMORPH data for each WT. Special attention will be paid on the contrast between land and sea and between mountain and valley.
Figures 5–9 show the diurnal cycle of the 3-hourly CMORPH rainfall, from that right after sunrise [0700–1000 local time (LT)] to that in the early morning (0400–0700 LT), in panels (a)–(h), while the daily averaged anomalies are shown in panel (i), for the five WTs, respectively. Averaged NRP large-scale winds at 850 hPa are also shown in panels (a)–(h). Terrain heights of 250 m are also given to show the location of mountains. Land breezes in the morning (as shown by the QuikSCAT morning-pass sea winds minus the daily mean) are illustrated in the 0700–1000 LT panels. Opposite wind directions (computed as the evening pass minus the daily mean) illustrating sea breezes in the afternoon are shown in the 1600–1900 LT panels. It is worth noting that the spatial extent of the land–sea breezes is not only limited in the vicinity of the coast (the local diurnal cycle), but they could reach far areas in the open sea (near 3°N, 108°E, which is about 500 km away from the northwest coast of Borneo; Figs. 5a,d)—similar to the large-scale diurnal variation of winds to the west of the Philippines that is associated with the delayed surface heating/cooling over the Asian continent and Indo-China Peninsula in June 2004 (Park et al. 2011). This kind of diurnal cycle of winds over the South China Sea seems stronger in WT1 (which occurs more frequently in the fall in early SON, thus with stronger surface heating over the Asian continent; Fig. 5) than those in other WTs (which happen more frequently in late fall and winter, thus with weaker surface heating over the Asian Continent; Figs. 6–9). A strong diurnal cycle of rainfall, with small (large) amount of rainfall in the morning (afternoon) over Borneo is common in all five weather types. However, the diurnal cycle of rainfall and local winds interacts with large-scale winds, resulting in dramatic differences in rainfall distribution between the weather types, as elaborated one by one as follows.
(a)–(h) WT1 diurnal cycle of CMORPH rain in the weather type 1. LT denotes the local time in Borneo. QuikSCAT land–sea breezes are illustrated by the twice-daily morning and evening passes in the 0700–1000 and 1600–1900 LT panels (thin black arrows). (i) Anomalous CMORPH precipitation (mm day−1) and anomalous NRP 850-hPa winds (m s−1) for the weather type WT1, where the WT frequency-weighted averaged climatologies have been subtracted to show the character of the weather type.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
(a)–(h) WT1 diurnal cycle of CMORPH rain in the weather type 1. LT denotes the local time in Borneo. QuikSCAT land–sea breezes are illustrated by the twice-daily morning and evening passes in the 0700–1000 and 1600–1900 LT panels (thin black arrows). (i) Anomalous CMORPH precipitation (mm day−1) and anomalous NRP 850-hPa winds (m s−1) for the weather type WT1, where the WT frequency-weighted averaged climatologies have been subtracted to show the character of the weather type.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
(a)–(h) WT1 diurnal cycle of CMORPH rain in the weather type 1. LT denotes the local time in Borneo. QuikSCAT land–sea breezes are illustrated by the twice-daily morning and evening passes in the 0700–1000 and 1600–1900 LT panels (thin black arrows). (i) Anomalous CMORPH precipitation (mm day−1) and anomalous NRP 850-hPa winds (m s−1) for the weather type WT1, where the WT frequency-weighted averaged climatologies have been subtracted to show the character of the weather type.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
In WT1 (Fig. 5), a relatively dry weather type occurring mostly before the austral monsoon season, heavy rainfall is in the South China Sea, offshore of the northwest coast in the morning (Fig. 5a), contributed by land breezes down from the coastal mountains in northwest Borneo (note in Fig. 1 that mountains are quite close to the northwest coast). Light rainfall starts to appear at noon on the mountain slopes near the northwest coast and northeast coast of Borneo (Fig. 5b). In the afternoon (1600–1900 LT; Fig. 5d), rainfall reaches maximum in the headland area in the northwest corner of West Kalimantan and Sarawak and in the relatively narrow northern portion of Borneo (the Sabah state of Malaysia and the northern East Kalimantan province of Indonesia). The convergence of sea breezes toward the headlands (with an converging angle of onshore winds) or toward the narrow portions of the island (with converging winds from opposite coasts) lifts up moist air and triggers convection, which can be further amplified by the cumulus-merger processes (Qian 2008; Simpson et al. 1980, 1993). The cumulus-merger process continues to enhance the moist convection by condensational heating [i.e., by the mechanism of the conditional instability of the second kind (CISK)] and prolongs rainfall to midnight, mostly over northeast Borneo (Figs. 5e,f). Meanwhile, the large-scale southwesterly winds at 850 hPa over northern Borneo appear to advect the maximum rainfall area downwind toward northeast, to the valley and bay area of northern East Kalimantan. Considering the characteristic of short life span (30 min or so) of the heavy downpours observed in this region (Bidin and Chappell 2006), the apparent downwind propagation of rainfall could be either caused by advection of clouds or by the regeneration of new convections contributed by moistened air transferred from the antecedent upwind rainy region. Another possible cause for the nocturnal maximum rainfall in the valley is the convergence of downslope mountain winds from the surrounding mountains on the north, west, and south sides toward the basin or bay area in the north East Kalimantan province.
It is interesting to note that the heaviest rainfall occurs on the lee side of the mountains and of the island (in the wake area in respect to the synoptic 850-hPa wind direction; i.e., in the northeast of Borneo, in Figs. 4a, 5d, and 5e), where sea breezes oppose the synoptic winds. For convenience, we will call this phenomenon the “wake effect.”
Figure 5i shows the daily-average rainfall and 850-hPa wind anomalies of WT1 with respect to a frequency-weighted climatology that is the weighted average of monthly climatology by using frequencies of the WT in each of the six months (September–February) as the weights (Qian et al. 2010; Moron et al. 2010). This panel [i.e., Fig. 5i; see also panel (i) of Figs. 6–9 for WT2 to WT5 as well] is to highlight the anomalous character of rainfall and low-level winds. For WT1, the negative and positive rainfall anomalies are located in the upwind and downwind directions of the regional winds anomalies respectively, indicating the wake effect on coastal rainfall.
In a dry and transitioning weather type WT2 (Fig. 6), with large-scale monsoon rain gradually shifted to the south, southern Borneo gets more rainfall than that in WT1 (Fig. 4b vs Fig. 4a); comparing to WT1, the monsoonal winds in WT2 are rather weak. Rainfall appears in headland of northeast coast in 1000–1300 LT (Fig. 6b) and reaches substantial amount in 1300–1600 LT (Fig. 6c), occurring 3 h earlier than that in WT1. From 1600 to 1900 LT, rainfall occurs along all the coastal regions around Borneo, especially along the southwest coast and on the western slopes of the mountain range near the northwest coasts (i.e., in the wake area in respect to the easterly winds at 850 hPa), with sea breezes converging to the coastal rainbelt, reinforced by valley breezes with almost the same phase of timing. Note that rising air in the rainy region may draw low-level convergence into it, which may further enhance the sea breezes (i.e., the CISK mechanism). It is also interesting to note that another rainbelt emerges at the eastern windward side of the central mountain range after 1900–2200 LT (Fig. 6e). The two rainbelts last until midnight and even after in the valleys south and west of the central mountain range. Again, the positive rainfall anomalies for WT2 is downwind of mean winds (and anomalous winds) at 850 hPa (Fig. 6i).
For a strong monsoon weather type WT3 (Fig. 7), the regional-scale 850-hPa winds are westerly over most of Borneo. This weather type prevails in the rainy season (Fig. 3). The diurnal cycle of rainfall is very strong, indicating the role of moisture in enhancing the diurnal cycle (through the positive feedback of the CISK mechanism). In the morning, land breezes prevail and rainfall is over the seas surrounding Borneo during 0700–1000 LT (Fig. 7a). In 1000–1300 LT, rainfall over the seas propagates away from Borneo, probably due to gravity waves, while dissipating (Yang and Slingo 2001; Mapes et al. 2003). Rainfall starts to appear along the southwest coast and the headland area in East Kalimantan (about 1.5°N, 118°E) in the early afternoon (1300–1600 LT; Fig. 7c). The rain rate keeps increasing over all the coastal areas around Borneo, by rather strong sea breezes converging to the coastal areas. Note that the sea breezes along the east coast of Borneo in WT3 (wake area or lee side of regional-scale winds, Fig. 7d, the wake effect) are stronger than those in WT1 and WT2 (windward side of regional-scale winds, Figs. 5d and 6d). The rainband in the sea-breeze front around the coast of the island penetrates inland in the evening and forms heavy rainfall over the eastern slope (i.e., lee side) of the central mountain range (1900–2200 LT; Fig. 7e). When approaching midnight (2200–0100 LT; Fig. 7f), the heavy rainfall regions sustain on the lee side of the mountain. After midnight (0100–0400 LT; Fig. 7g), rainfall starts to dissipate, and another rainband appears off the northwest shore in the South China Sea, formed by land breezes and gravity waves that propagate away from Borneo to the South China Sea (Houze et al. 1981; Johnson and Priegnitz 1981; Yang and Slingo 2001; Mapes et al. 2003; Chang et al. 2005b). In the early morning, the remnant of the precipitation center is still over East Kalimantan about 0°, 116°E (0400–0700 LT, Fig. 7g) and lasts after sunrise (0700–1000 LT; Fig. 7a). As a result, the daily averaged rainfall is maximized over eastern Borneo in WT3 (Fig. 4c). The wake effect appears particularly pronounced for WT3; that is, positive rainfall anomalies for WT3 are downwind, which is over eastern Borneo (Fig. 7i).
For the quiescent monsoon type WT4 (Fig. 8), regional-scale winds over Borneo are northeasterly at 850 hPa, opposite to the seasonal westerly winds in the austral summer monsoon season that are prevalent in WT3 and WT5. The 850-hPa winds over the south and southeast of Borneo are quite weak. The nocturnal rainbelt off the northwest shore is the strongest in WT4 among the five WTs, probably due to the following three factors: the more favorable regional-scale condition for rainfall in the wet season than that in WT1 and WT2, the smaller wind velocity in WT4 than that in WT3 and WT5 (the inverse relationship between land–sea breezes and the monsoonal wind speed), and the alongshore (but slightly tilted offshore) direction of the regional-scale 850-hPa winds, which thus exert little interference on the nocturnal offshore precipitation. The earliest place to get rainfall during a day (1000–1300 and 1300–1600 LT; Figs. 8b,c) is the south coast of Borneo, which is located in the wake side of the island in respect to the regional-scale winds and the winds over there are weak (the wake effect). The rainfall along the northwest coast (except the most northern part where rainfall emerges during 1000–1300 LT) also forms during 1300–1600 and 1600–1900 LT. In contrast to that in WT3, the rainfall in WT4 along the east coast is rather small, probably due to the slightly onshore regional-scale northeasterly winds that disrupt the atmosphere boundary layer and the sea-breeze front there. The rainfall gradually increases in the southwest portion of Borneo in the afternoon, when the sea breezes penetrate into inland areas. Two northeast–southwest-oriented rainbands emerge in the evening and at night (after 1900 LT), one along the western slope and the other along the eastern slope of the central mountain range. The northern part of the two rainbands drifts southwestward following the regional-scale 850-hPa winds. The two rainbands are connected at their southwestern ends with maximum rain rates occurring over the valleys there. A third rainband over the South China Sea off the northwest coast emerges after midnight (Fig. 8g) and lasts until the morning (Figs. 8a,h), forming a pattern of three parallel rainbands in northeast–southwest orientation. The consequence of this diurnal cycle of rainfall and land–sea breezes is that the area of maximum daily averaged rainfall is in southwest Borneo (Fig. 4d). For the quiescent monsoon WT4, the wake effect leads to the positive rainfall anomalies over southwestern Borneo (Fig. 8i), in contrast to that in WT3 (Fig. 7i).
About the nocturnal and morning peak of rainfall over the coastal sea northwest of Borneo [panels (a), (g), and (h) in Figs. 7, 8, and 9], previous studies indicate that it may be contributed by the land-breeze convergence (Qian 2008) and gravity wave propagation (Mapes et al. 2003); in addition, a cold pool generated from the afternoon convection over land may also be responsible for the diurnal cycle of the offshore rainfall, as seen in the mesoscale convective systems near the Australian coast in the Equatorial Mesoscale Experiment (Mapes and Houze 1992) and near Borneo in the South China Sea Monsoon Experiment (Ciesielski and Johnson 2006).
Finally, let us look at a strong monsoon weather type, WT5 (Fig. 9), in which strong monsoon flows from the northwest and then turn toward southeast in the equatorial region (Figs. 9a–h and 4e). Rainfall starts at the coasts during 1300–1600 LT and intensifies during 1600–1900 LT in WT5 (Figs. 9c,d). Three rainbelts appear on the northwest, southeast, and southwest slopes (Fig. 9d). The timing of coastal rain starts slightly later in WT5 than that in other WTs (except in the southeast coast), probably due to stronger interference of regional-scale winds on the sea breezes. Rainy areas converge toward inland in the evening (Fig. 9e), then merge in 2200–0100 LT (Fig. 9f). After midnight, rainfall decreases in the middle over the south branch of the central mountain range in the territory of Kalimantan Provinces of Indonesia and remains in the valleys. But the rainy area southeast of the central mountain range (about 1°S, 115°E, on the lee side of the mountain) lasts longer, until early morning and after sunrise (0400–0700 LT and 0700–1000 LT). So again, in daily average, more rainfall in WT5 is located in the downstream direction (or the lee side of the mountain) of the large-scale 850-hPa winds (Figs. 4e and 9i)—the wake effect, as in other WTs (Fig. 4, and panel (i) of Figs. 5–9).
Considering the impact of wind regimes on the spatial rainfall distribution (Figs. 5–9), and the ENSO modulation of the frequency of the WTs (Fig. 3), we can interpret the rainfall anomaly pattern of the east–west dipole over south Borneo, as is summarized below.
Panel (i) of Figs. 5–9 shows that, in all five WTs, above (below) normal rainfall is located downstream (upstream) of regional-scale 850-hPa wind anomalies. That is, more rainfall is in the lee side of the island with respect to the regional-scale winds—the wake effect. WT1, WT3, and WT5 have more rainfall over eastern Borneo, while WT2 and WT4 have more rainfall over western Borneo. The east–west dipole pattern in rainfall anomalies is more pronounced in WT2, WT3, and WT4. Because WT2 and WT4 are more frequent in El Niño years (Fig. 3b), more rainfall occurs over the southwest part of Borneo (Fig. 2d). In La Niña years, more days are of the WT3 type(Fig. 3c), and thus above-normal rainfall takes place over eastern Borneo (Figs. 7i and 2f).
d. Possible role of regional SST variability
In El Niño years, regional SSTs in the seas in the Maritime Continent region vary from negative anomalies in SON to positive anomalies in DJF as shown in the ERSST data (Fig. 10). The warming is caused by the reduction of surface evaporation due to below normal surface wind speed (Hendon 2003) and increased solar radiation to the surface due to fewer than normal cloudy and rainy days above the seas in DJF (Giannini et al. 2007). In Fig. 10a, SST anomalies are quite small around Borneo, with mostly negative anomalies except off shore of the northwest and north Borneo. Therefore, impacts of local SST variability on Borneo rainfall should be weak in SON. In DJF (Fig. 10b), positive SST anomalies west and southwest of Borneo are larger than those east of Borneo in the Makasar Strait (between Borneo and Sulawesi). Warm waters heat up and moisten lower atmosphere, thus decreasing atmospheric static stability and enhancing rainfall. Therefore, the warmer SST anomalies in the southern South China Sea could increase rainfall over the sea, which may partly drift to the nearby coastal lands to increase rainfall over west Borneo and also over southern Borneo. Therefore, in addition to the mechanism of the wake effect discussed in Figs. 5–9, the regional SST variability could also partly contribute to the coastal rainfall variability. The latter deserves further investigation, for example, by a coupled regional atmosphere–ocean model.
The El Niño year-composite sea surface temperature anomalies in (a) SON and (b) DJF.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
The El Niño year-composite sea surface temperature anomalies in (a) SON and (b) DJF.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
The El Niño year-composite sea surface temperature anomalies in (a) SON and (b) DJF.
Citation: Journal of Climate 26, 5; 10.1175/JCLI-D-12-00178.1
4. Conclusions and discussion
Multiscale climate processes have been analyzed to understand the interannual rainfall variability over Borneo. Similar to that in Java (Qian et al. 2010), we also found some correspondence among ENSO, weather types, and the diurnal cycle over Borneo. But because Borneo is much larger and wider (especially in the southern part) than Java and the seasonality of rainfall and winds over Borneo (especially in the central and northern part) is somewhat different from that in Java (Aldrian and Susanto 2003), the diurnal cycle of rainfall and winds over Borneo and its interaction with ENSO and synoptic weather types have some distinct features.
We tried to explain the rainfall variability in Borneo in the SON and DJF seasons by 1) the Walker circulation effect [i.e., anomalous sinking (rising) motion of the air over the Maritime Continent associated with Walker circulation and El Niño (La Niña); Bjerknes 1969], 2) the monsoonal damping effect (i.e., the inverse relationship between monsoonal wind speed and the strength of the diurnal cycle of land–sea and mountain–valley breezes; Qian et al. 2010), and 3) the wake effect proposed in this paper.
During SON of El Niño years (Fig. 2c), the 850-hPa wind anomalies are easterlies, acting to increase monsoonal wind speed in south Borneo (thus weakening the diurnal cycle due to the monsoonal damping effect) but decrease wind speed in north Borneo (thus strengthen diurnal cycle). Therefore, in south Borneo, the weakened sea-breeze convergence tends to reduce rainfall, reinforcing the drying Walker circulation effect, so the dry anomalies are large and statistically significant in south Borneo in SON.
During SON of La Niña years (Fig. 2e), the strengthened sea-breeze convergence (associated with weakened monsoonal winds) in south Borneo tends to increase rainfall on the southern coastal lowlands, reinforcing the wet anomaly caused by the Walker circulation effect in La Niña years; thus, the wet anomalies are large and statistically significant over south Borneo. In north Borneo, the weakened sea-breeze convergence (associated with strengthened 850-hPa winds) tends to reduce rainfall, counteracting the wet anomalies caused by the Walker circulation effect, so the wet anomalies in north Borneo are rather small and in some areas statistically insignificant in La Niña years.
The spatial distribution of rainfall over Borneo in the austral spring and summer season on the daily scale has been shown to depend on the wind direction of the synoptic-scale circulation regimes, through a so-called wake effect, in which the sea breeze is enhanced on the sheltered downwind (lee) side of the mountains and of the island. It is proposed that the enhanced sea-breeze circulation on the sheltered side of the island leads to a larger diurnal cycle rainfall amounts there, compared to the windward side where the diurnal cycle is weaker due to disruption by the on-shore synoptic flow. The diurnal evolution is summarized as follows: the first stage of daily rainfall formation over Borneo is through the sea-breeze front that emerges on the coastal plains around noon, especially on the lee side of the island, through the wake effect; rainfall then intensifies in the afternoon and evening with the rainbelts converging toward the center of the island, reaching a maximum rainfall in the evening. The second stage consists of a prolonged rainfall in the valleys on the lee side of the mountains with respect to the low-level monsoonal winds (at 850 hPa) while gradually dissipating from about midnight to early morning.
The east–west dipolar pattern of rainfall anomalies in south Borneo (in DJF) associated with ENSO is shown to be consistent with the impact of ENSO on the prevalence of contrasting circulation types: El Niño years are associated with high prevalence of WT2 and WT4 (easterlies), which have more rainfall over western Borneo, while La Niña tends to characterized by WT3 (westerlies) with above-normal rainfall over eastern Borneo. Regional SST anomalies in the southern South China Sea become positive in the DJF of El Niño years, which could also partly contribute to the positive rainfall anomalies over western Borneo.
The proposed wake effect over Borneo contrasts with the terrain-lifting mechanism of Chang et al. (2005a) in which more rainfall occurs on the windward side of islands or mountains; while both consist of wind–terrain interactions, they lead to dipolar rainfall anomalies of the opposite sense. The wake effect is likely to be more prevalent in the deep tropics such as over Borneo where the large-scale synoptic winds are usually weak (and the diurnal cycle is especially strong), while the terrain-lifting effect is strong when synoptic winds are relatively strong and the slope of the terrain is steep (such as the western Ghats of India and Indo-China Peninsula and the eastern coast of the Philippines).
In DJF of El Niño years, the Walker circulation effect and the monsoonal damping effect cause the negative rainfall anomalies over north Borneo, while the wake effect cause the wet west versus dry east pattern of rainfall anomalies in south Borneo. These in combination account for the dipolar pattern of rainfall anomalies in DJF—the wetter southwest versus drier central and northeast over Borneo Island associated with El Niño.
Acknowledgments
JHQ thanks the University of Massachusetts at Lowell for a faculty start-up fund. We also thank the editor and two anonymous reviewers for their constructive comments.
REFERENCES
Aldrian, E., and R. D. Susanto, 2003: Identification of three dominant rainfall regions within Indonesia and their relationship to sea surface temperature. Int. J. Climatol., 23, 1435–1452.
Bidin, K., and N. A. Chappell, 2006: Characteristics of rain events at an inland locality in northeast Borneo, Malaysia. Hydrol. Processes, 20, 3835–3850.
Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97, 163–172.
Chang, C. P., P. A. Harr, and H.-J. Chen, 2005a: Synoptic disturbances over the equatorial South China Sea and western Maritime Continent during boreal winter. Mon. Wea. Rev., 133, 489–503.
Chang, C. P., Z. Wang, J. McBride, and C.-H. Liu, 2005b: Annual cycle of Southeast Asia–Maritime Continent rainfall and asymmetric monsoon transition. J. Climate, 18, 287–301.
Ciesielski, P. E., and R. H. Johnson, 2006: Contrasting characteristics of convection over the northern and southern South China Sea during SCSMEX. Mon. Wea. Rev., 134, 1041–1062.
Giannini, A., A. W. Robertson, and J.-H. Qian, 2007: A role of tropical tropospheric temperature adjustment to El Niño–Southern Oscillation in the seasonality of monsoonal Indonesia precipitation predictability. J. Geophys. Res., 112, D16110, doi:10.1029/2007JD008519.
Haylock, M., and J. McBride, 2001: Spatial coherence and predictability of Indonesian wet season rainfall. J. Climate, 14, 3882–3887.
Hendon, H. H., 2003: Indonesian rainfall variability: Impacts of ENSO and local air–sea interaction. J. Climate, 16, 1775–1790.
Hoerling, M. P., A. Kumar, and M. Zhong, 1997: El Niño, La Niña, and the nonlinearity of their teleconnections. J. Climate, 10, 1769–1786.
Houze, R. A., Jr., S. Geotis, F. Marks, and A. West, 1981: Winter monsoon convection in the vicinity of North Borneo. Part I: Structure and time variation of the clouds and precipitation. Mon. Wea. Rev., 109, 1595–1614.
Ichikawa, H., and T. Yasunari, 2006: Time–space characteristics of diurnal rainfall over Borneo and surrounding oceans as observed by TRMM-PR. J. Climate, 19, 1238–1260.
Janowiak, J. E., V. E. Kousky, and R. J. Joyce, 2005: Diurnal cycle of precipitation determined from the CMORPH high spatial and temporal resolution global precipitation analyses. J. Geophys. Res., 110, D23105, doi:10.1029/2005JD006156.
Johnson, R. H., and D. L. Priegnitz, 1981: Winter monsoon convection in the vicinity of North Borneo. Part II: Effects on large-scale fields. Mon. Wea. Rev., 109, 1615–1628.
Juneng, L., and F. T. Tangang, 2005: Evolution of ENSO-related rainfall anomalies in Southeast Asia region and its relationship with atmosphere–ocean variations in Indo-Pacific sector. Climate Dyn., 25, 337–350.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnili, M. Fiorino, and G. L. Potter, 2002: NCEP-DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643.
Kumar, A., and M. P. Hoerling, 1998: Annual cycle of Pacific–North American seasonal predictability associated with different phases of ENSO. J. Climate, 11, 3295–3308.
Liu, W. T., 2002: Progress in scatterometer application. J. Oceanogr., 58, 121–136.
Mapes, B. E., and R. A. Houze, 1992: An integrated view of the 1987 Australian monsoon and its mesoscale convective systems. I: Horizontal structure. Quart. J. Roy. Meteor. Soc., 118, 927–963.
Mapes, B. E., T. T. Warner, M. Xu, and A. J. Negri, 2003: Diurnal patterns of rainfall in northwestern South America. Part I: Observations and context. Mon. Wea. Rev., 131, 799–812.
Moron, V., A. W. Robertson, and R. Boer, 2009: Spatial coherence and seasonal predictability of monsoon onset over Indonesia. J. Climate, 22, 840–850.
Moron, V., A. W. Robertson, and J.-H. Qian, 2010: Local versus regional-scale characteristics of monsoon onset and post-onset rainfall over Indonesia. Climate Dyn., 34, 281–299.
Park, M.-S., C.-H. Ho, J. Kim, and R. L. Elsberry, 2011: Diurnal circulations and their multi-scale interaction leading to rainfall over the South China Sea upstream of the Philippines during intraseasonal monsoon westerly wind bursts. Climate Dyn., 37, 1483–1499.
Qian, J.-H., 2008: Why precipitation is mostly concentrated over islands in the Maritime Continent. J. Atmos. Sci., 65, 1428–1441.
Qian, J.-H., A. W. Robertson, and V. Moron, 2010: Interactions among ENSO, the monsoon, and diurnal cycle in rainfall variability over Java, Indonesia. J. Atmos. Sci., 67, 3509–3524.
Robertson, A. W., V. Moron, J.-H. Qian, C.-P. Chang, F. Tangang, E. Aldrian, T. Y. Koh, and L. Juneng, 2010: The Maritime Continent monsoon. The Global Monsoon System: Research and Forecast, 2nd ed., C.-P. Chang et al., Eds., World Scientific, 85–98.
Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 1606–1626.
Ropelewski, C. F., and M. S. Halpert, 1996: Quantifying Southern Oscillation–precipitation relationships. J. Climate, 9, 1043–1059.
Simpson, J., N. E. Westcott, R. J. Clerman, and R. A. Pielke, 1980: On cumulus mergers. Arch. Meteor. Geophys. Bioklimatol., 29A, 1–40.
Simpson, J., T. D. Keenan, B. Ferrier, R. H. Simpson, and G. J. Holland, 1993: Cumulus mergers in the Maritime Continent region. Meteor. Atmos. Phys., 51, 73–99.
Smith, T. M., and R. W. Reynolds, 2005: A global merged land–air–sea surface temperature reconstruction based on historical observations (1880–1997). J. Climate, 18, 2021–2036.
Wu, P., J. I. Hamada, M. D. Yamanaka, J. Matsumoto, and M. Hara, 2009: The impact of orographically induced gravity waves on the diurnal cycle of rainfall over southeast Kalimantan Island. Atmos. Oceanic Sci. Lett., 2, 35–39.
Yang, G.-Y., and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129, 784–801.
