1 Introduction

The Asian summer monsoon comprises two major systems: the East and South Asian summer monsoons. The domain of the South Asian monsoon, including the Southeast Asian monsoon, is confined to the tropics with complicated topography (LinHo and Wang 2002; Ding and Chan 2005). The northern Indian Ocean is separated into the Arabian Sea, Bay of Bengal, and South China Sea (SCS) by the Indian Subcontinent and the Indochina Peninsula, generating a large difference in the solar heating of the surface between land and sea. The mountains of these two land masses are primarily north–south-oriented, and they intersect with the prevailing lower-level westerly and southwesterly winds at high angles. The east–west-oriented Himalayas have been considered as a blocking of the monsoon from expanding northward of the mountains (Ramisch et al. 2016; Tada et al. 2016). The cross-equatorial current that is culminated in the Somali jet strongly influences the monsoon. Thus, the South Asian monsoon has an exclusively tropical meteorology with high regional variability. With continued solar heating in late May, the South Asian monsoon expands from Indochina northwestward along the eastern Bay of Bengal. The inception of the Indian monsoon is moderately late; it occurs in early June and is confined to the southwestern coastal region of India. By late June, the monsoon has expanded to the foothills of the Himalayas and eastern India. It continues to expand westward in the Gangetic plain in July. Recent renewed emphasis has been placed on the blocking of the Himalayas and the topographically-induced condensational heating at the southern slope of the mountains, which is commonly referred to as candle-like heating (Boos and Kuang 2010; Wu et al. 2012; Chen et al. 2014; He et al. 2015).

The domain of the East Asian summer monsoon covers a large range of latitudes. Prior to August, the monsoon rain is primarily produced in the frontal system located on the northwestern flank of the Western North Pacific (WNP) high and in the tropical convergence zone south of the high. As the WNP high progresses northward, these two rainy bands follow, with a dry band in between. Thus, evolutions of the monsoon in subregions of the East Asia–WNP are highly correlated. According to the location of the WNP high and the precipitation in Taiwan, Chou et al. (2011) divided the Asian summer monsoon into three phases. In Phase 1 (May 15–June 15), the western stretch of the WNP high is located south of Taiwan. This is the early Meiyu–Okinawa Baiu season (Okada and Yamazaki 2012; Wang et al. 2016). In Phase 2 (June 15–July 24), the western stretch of the WNP high and the frontal system that extends from the Yangtze valley northward and northeastward to Japan and Korea are located near Taiwan and are nearly stagnant. This is the core Meiyu–Baiu rainy season (Sampe and Xie 2010; Chen and Bordoni 2014). In Phase 3 (after late July), the WNP high jumps to southern Japan, terminating the Baiu in Japan (Ueda et al. 1995; Suzuki and Hoskins 2009).

Because the locations of the frontal system are coupled to the WNP high, one of the main topics of the East Asia–WNP summer monsoon studies is identifying the factors that control the northward movement of the WNP high. As shown later in this paper, the speed of the northward movement of the WNP high in Phase 2 (i.e., the frontal system in the core Meiyu–Baiu season) becomes nearly stagnant and is only half the speed of that in Phase 1. The purpose of this study was to identify the mechanisms that cause the WNP high to move northward at different speeds in Phases 1 and 2, specifically the mechanisms that connect the WNP high to the upper tropospheric circulation, as in the transition from Phase 2 to 3, which has been addressed by previous studies (Lin and Lu 2008; Wu and Chou 2012).

The remaining paper is organized as follows. Section 2 introduces the data and model, and Sects. 3 and 4 highlight the phase connection between South and East Asian monsoons, with the focus on the influence of the upper tropospheric circulation. Section 5 investigates the impact of the Himalayas on the northward extent of the Asian summer monsoon in a series of numerical experiments. Concluding remarks are provided in Sect. 6.

2 Data and model

The data used in this study were as follows: (a) Taiwanese Central Weather Bureau (CWB) precipitation records (i.e. daily data from eight CWB stations at low plains distributed evenly over the island) and (b) geopotential height, pressure, winds, shortwave and longwave radiation at the top of the atmosphere and at the surface, and latent and sensible heat fluxes at the surface from the National Centers for Environmental Prediction (NCEP) reanalysis (Kalnay et al. 1996). It covers a 33-year period from 1978 to 2010 and has a spatial resolution of 2.5° latitude–longitude and a temporal resolution of 1 day.

Community Atmosphere Model (CAM) version 5.1 of the National Center for Atmospheric Research (NCAR) was also used to simulate the atmospheric response to the artificially prescribed mountains (by replacing topography data in the model). The CAM simulations were integrated using the finite volume dynamical core with 30 vertical levels at a horizontal resolution of approximately 1°. Parameterizations in the CAM include cloud macrophysics and microphysics as well as radiation, moist processes, surface models, and turbulent mixing (Neale et al. 2013). The boundary and initial conditions of the simulations were derived from the Community Earth System Model preindustrial control experiment (Vertenstein et al. 2010). The integration length was 25 years for each simulation, and the outputs for 6–25 years were analyzed. Experimental design will be provided in Sect. 5.1.

3 South Asian summer monsoon and upper tropospheric circulation

In this section we summarize and introduce the evolution of monsoonal heating in South Asia and its topographical dependence, and then examine the influence of the upper tropospheric circulation by the heating of the atmosphere and the surface.

3.1 Evolution of the South Asian summer monsoon

A crucial factor affecting the evolution of the South Asian summer monsoon is strengthening of the westerlies in the Indian Ocean, Arabian Sea, and Bay of Bengal in May and June, which is largely influenced by the contrast of the large-scale heating over land and sea and the cross-equatorial flow from the Southern Hemisphere. The complicated orography of the Indian Subcontinent, Indochina Peninsula, and Himalayas adds to the monsoon evolution. The mostly north–south-oriented mountains of the Indochina Peninsula and Indian Subcontinent interact with the westerlies of the Arabian Sea and Bay of Bengal at nearly normal angles, inducing uplifting and precipitation, whereas the east–west-oriented Himalayas block the monsoon from expanding further north. Following the strengthening of the westerlies in the Arabian Sea and Bay of Bengal, the monsoon expands northward along the west coasts of India and Indochina. Figure 1 displays the westerly wind at 850 hPa over the northern Indian Ocean (0°–10°N; 50–100°E), Arabian Sea (10–20°N; 55–75°E), and Bay of Bengal (10–20°N; 85–95°E). In May, the westerly wind is stronger in the south (Indian Ocean) than in the north (Arabian Sea and Bay of Bengal), and in the east (Bay of Bengal) than in the west (Arabian Sea). The westerly wind in these regions increases continuously until mid-June (Indian Ocean and Bay of Bengal) and late June (Arabian Sea). It attains a maximum strength with small variations in July and weakens afterward.

Fig. 1
figure 1

Westerly wind speed at 850 hPa over the northern Indian Ocean (0°–10°N; 50–100°E), Arabian Sea (10–20°N; 55–75°E), and Bay of Bengal (10–20°N; 80–100°E). The x-axis is time ranging from May 1 to August 31

Full size image

Closely related to the evolution of the westerlies, the monsoon onset dates and the characteristics of the monsoon evolution differ substantially between Indochina and India. In Indochina, the monsoon quickly expands over a large portion of the peninsula in mid-May when a sudden jump in the lower-level westerlies (Fig. 1) and southerlies (not shown) occurs. This monsoon onset over a large domain is displayed in Fig. 2a, b. The north–south orientation of most mountain ranges in Indochina, which enables humidity to easily be transported northward by surface winds, might be the major cause of this large-domain monsoon onset; the mesoscale mountainous effects on the downstream East Asia–WNP region have been further indicated (Wu et al. 2014; Wu and Hsu 2016).

Fig. 2
figure 2

Ten-day mean vertically integrated (atmosphere + surface) heating (W m−2) and geopotential height (m) at 700 hPa

Full size image

Evolution of the monsoon in time and space can be most effectively studied using precipitation information. However, accurate estimations of precipitation over land are difficult whether from atmospheric model simulations or from satellite retrievals. Because precipitation is the most critical atmospheric heat source, we used the column-integrated (including the surface) heating as a surrogate for precipitation to infer the strength of the monsoon. Using the NCEP reanalysis data, we computed the column-integrated heating, H, from \(H=({S_t} \downarrow - {S_t} \uparrow - {R_t} \uparrow )+(P\, - \,E),\) where S t and R t , respectively denote the solar and infrared radiation at the top of the atmosphere, P denotes latent heat release due to the precipitation, and E denotes the surface evaporation in units of heat flux (W m−2). The arrows denote upward or downward radiation.

Seasonal evolution of the column heating as well as geopotential heights at 700 hPa in the domain of the Asian summer monsoon is displayed in Fig. 2. Each panel in the figure represents a 10-day mean. The heating accelerates in early May and especially in mid-May (Fig. 2b) when strong heating occurs in southwestern Indochina and the southeastern Bay of Bengal. Thus, mid-May is commonly identified as the time of the summer monsoon onset not only in Indochina but also in all of East and South Asia (Wu and Hsu 2016).

After the monsoon onset in Indochina, the Asian summer monsoon continues to evolve with enhanced solar heating as summer proceeds. By late May, the WNP high retreats eastward, and deep convections expand to the SCS (Fig. 2c), which is associated with the heating of the warm pool as well as the eastward retreat and northward shift of the WNP high. This is the onset of the summer monsoon in the SCS (Wu 2002). In the Bay of Bengal, the low-level westerly continues to strengthen, generating heavy precipitation in southern Myanmar and western Thailand. The westerlies in the Arabian Sea are still weak in late May (Fig. 1), and the summer monsoon in the Indian Subcontinent has not yet commenced.

June is when the South Asian summer monsoon circulation rapidly intensifies (LinHo and Wang 2002; Minoura et al. 2003). This is mostly reflected in the strengthening of the low-level westerlies in the Arabian Sea and Bay of Bengal (Fig. 1) and the South Asian high (Fig. 3). By mid-June, heavy precipitation expands to the head of the Bay of Bengal and covers the eastern half of the bay (Fig. 2e). The northern front of rainfall (or heating) in South and Southeast Asia is pushed to the foothills of the Himalayas by late June (Fig. 2f). On the other side of India, the westerlies increase rapidly in early June in the southern Arabian Sea (Fig. 1). When the westerlies encounter the north–south-oriented Western Ghats in southwestern India, it causes heavy rainfall and triggers the Indian summer monsoon (Fig. 2d). The monsoon continues to strengthen and expand after the onset but is confined to southwestern India (Fig. 2e). Most of India remains dry throughout the entire month of June because it is influenced by the descending air from the dry regions of Saudi Arabia, Iran, and Afghanistan. With the continued strengthening of the Somali jet in early July, the westerly is pushed further north deep into the Arabian Sea (not shown in figures). Simultaneously, the descending dry air retreats northward. With the weakened subsidence, the monsoon in northeastern India expands westward to the Gangetic Plain. The westward expansion of the monsoon continues until late July (Fig. 2g–i). Heating (or precipitation) in southeastern India and the southwestern Bay of Bengal is weak throughout the summer.

Fig. 3
figure 3

Ten-day mean streamlines and geopotential height (m) at 200 hPa. The black circle identifies the location of the maximum geopotential height (i.e. center of the South Asian high), whereas the star is the center of the previous 10-day period. For earlier 10-day periods, the center is identified by solid circles

Full size image

3.2 Upper tropospheric circulation

The upper tropospheric circulation in Asia is influenced largely by the overall heating below. Figure 3 shows evolutions of circulation and the geopotential height at 200 hPa. Each panel represents 10-day mean conditions. Before May, the Maritime Continent and the Intertropical Convergence Zone southeast of the Philippines are two regions with strong heating because of deep convections. At the onset of the Asian summer monsoon in mid-May, the center moves rapidly from the SCS to Indochina (Fig. 3b). This is also the center of a large-scale anticyclonic circulation with westerlies to the north and easterlies to the south. The center continues to move northwestward through July and starts to retreat in mid-August. The 200-hPa geopotential height increases as the center proceeds northwestward. The strength of the upper tropospheric high reflects the overall heating of the troposphere, implying that the Asian summer monsoon intensifies as it expands from Indochina westward to India.

The northwestward shift of the center is more clearly displayed in Fig. 4a, b. The speed of the northward shift of the center before late June is twice as fast as that in July and early August. The average speed is 0.25° latitude/day before late June and 0.125° latitude/day in July and early August. We attribute this slow northward shift of the high center after late June to the blocking of the monsoon by the Himalayas from expanding northward. Figure 4c indicates that the heating in the Himalayan region (25–30°N; 80–100°E) increases steadily from late May to late June and reaches a maximum through July. It weakens in early August, implying a weakening of the South Asian summer monsoon. The reduced speed of the northward shift of the high center after mid-June, as displayed in Fig. 4a, indicates that the monsoon heating is blocked by the Himalayas from expanding further north.

Fig. 4
figure 4

Latitude (a) and longitude (b) of the center of the South Asian high at 200 hPa. The two vertical lines separate the three monsoon phases. c Vertically integrated total heating (W m−2) in the Himalayan region (25–30°N; 80–100°E). The x-axis is time ranging from May 1 to August 31

Full size image

3.3 Three phases of the South Asian summer monsoon

Because the westerlies in the Arabian Sea and Bay of Bengal strongly influence the Indian summer monsoon, we indicate in Fig. 5 the longitudinal distributions of the 10-day mean zonal wind in the latitude band 10–20°N encompassing the Arabian Sea, India, the Bay of Bengal, Indochina, and the SCS. The locations of India and Indochina are marked by solid bars at the bottom of the figure. We divided the 10-day mean westerlies into three groups with distinct evolution speeds. Across the 10–20°N latitude band, the westerlies strengthen rapidly from mid-May to mid-June (Fig. 5a), reach a maximum with little change in strength from mid-June to late July (Fig. 5b), and start to weaken in late July or early August (Fig. 5c). As will be shown later, these three phases of the South Asian westerlies coincide exactly with the three phases of the East Asia–WNP precipitation. Webster and Yang (1992) proposed using an index for measuring the strength of the Asian summer monsoon circulation. This index is defined as the vertical wind shear between the 200- and 850-hPa levels averaged over the tropical northern Indian Ocean (0°–20°N; 40–110°E). It indicates that the strength of the Asian summer monsoon circulation increases rapidly from mid-May to late June, reaches a mature state with little variation from late June to early August, and decreases afterwards (not shown in the figures). Thus, the entire East and South Asian summer monsoons can be grouped into three phases: developing, mature, and retreating.

Fig. 5
figure 5

Longitudinal variations of the zonal wind at 850 hPa (m s−1) averaged over the latitude band 10–20°N

Full size image

In the domain of the South Asian summer monsoon, the westerlies are stronger in the south (Indian Ocean) than in the north (Arabian Sea and Bay of Bengal) and in the east (Bay of Bengal) than in the west (Arabian Sea) from May through early June (Fig. 1). This order is reversed after mid-June, implying that the strengthened westerlies shift northwestward. The northwestward shifts of the South Asian high center correspond to the evolution of the westerlies (Fig. 4). In Phase 3, the South Asian high center retreats southward, but only slowly. The South Asian high in Phase 3 is not only controlled by the heating of the South Asian summer monsoon but is also influenced in August by the heating of the interior Asian dry land across the Middle East, Central Asia, western China, and Mongolia (Wu and Chou 2012).

4 Impact of upper tropospheric circulation on the East Asian summer monsoon

4.1 Three phases of Taiwan precipitation and the Western North Pacific high

Precipitation in the East Asia–WNP is closely related to the strength and position of the WNP high. Precipitation is suppressed near the high ridge. On the southern flank of the high, a lower-level convergence of tropical air with vigorous convections is present. On the northern flank of the high, a convergence of air from the tropics and midlatitudes is present, which forms frontal weather systems and produces heavy precipitation. Taiwan is located on the western margin of the WNP high, and its climate is highly sensitive to the location and strength of this high. Therefore, precipitation in Taiwan can be used as a measure for the overall climate in the East Asia–WNP (Chou et al. 2011).

Figure 6 shows the annual variation of precipitation in Taiwan. The data were obtained from eight CWB stations at low plains distributed evenly over the island. Two precipitation maxima occur between mid-May and mid-June and between late July and September. A dry period occurs between these maxima, from mid-June to late July. These three precipitation phases are closely related to the location of the WNP high. Figure 7a–c display the 700-hPa streamlines and geopotential heights of the three phases. The WNP high shifts from the tropics in Phase 1 to the subtropics in Phase 2 and to midlatitudes in Phase 3. In Phase 1, the western stretch of the high extends to the northern Philippines and the SCS. To the northwest, a southwesterly flow occurs from the Bay of Bengal to Indochina, southern China, and Taiwan. This southwesterly flow is associated with the rain-producing frontal system that was intensively studied by Chang et al. (2000). In Phase 2, the WNP high ridge shifts northward to a location near Taiwan, and the frontal system shifts to mid-eastern China, Japan, and Korea. The wet phase of Taiwan ends, while the Meiyu–Baiu–Changma rainy season begins (Kato 1989; Chen et al. 2004). In Phase 3, the WNP high shifts further northward to the south of Japan. The Meiyu–Baiu–Changma rainy season ends; most of the East Asia–WNP is under the influence of the flow from the tropical WNP.

Fig. 6
figure 6

Mean annual variation (1978–2010) of the precipitation (mm day−1) in Taiwan at eight low plain weather stations. The x-axis is time ranging from January 1 to November 31

Full size image
Fig. 7
figure 7

Streamlines and the geopotential height (m) at a–c 700 hPa and d–f 200 hPa of the three phases of the East Asian summer monsoon: Phase 1 (May 15–June 15), Phase 2 (June 15–July 24), and Phase 3 (July 24–August 31)

Full size image

4.2 South Asian high and the East Asian summer monsoon

The overall evolution of the Asian summer monsoon (i.e. the northward propagation of the East Asian summer monsoon and the northwestward propagation of the South Asian summer monsoon) is related to the seasonal procession of the solar heating. These observations alone do not provide information on the connection between the two (or three when including the WNP summer monsoon) monsoon systems. They also do not explain why both the East and South Asian summer monsoons, which are distant geographically, can be grouped into three identical evolution phases.

In Phases 1 and 2 of the East Asian summer monsoon, the rain band corresponds to the frontal system, which is located on the northwestern flank of the WNP high ridge. Thus, in the East Asia–WNP region, the upper tropospheric high ridge, which is an eastern extension of the South Asian high ridge, is located to the north of the WNP high ridge indicating a vertical coupling of the two high-pressure systems. This is demonstrated in Fig. 7 that, over the 120–140°E longitude band, the eastern stretch of the South Asian high (at 200 hPa) overlaps the western stretch of the WNP high (at 700 hPa) with the former located north of the latter. In Phase 2 (mid-June to late July), as constrained by the South Asian high, the WNP high has slow northward movement (Fig. 4a) and, thus, a stagnant frontal system in the Meiyu–Baiu–Changma rainy season.

In Phase 3, the upper tropospheric circulation in the midlatitudes significantly influences the East Asia–WNP climate (Wu and Chou 2012). Before late July, the sea surface temperature in the subtropical WNP (15–25°N, 130–150°E) is high, but the atmosphere is stable because it is under the influence of the WNP high (Wu et al. 2009; Wu and Chou 2012). By late July, the continued solar heating of the interior Asia and weakening of the East Asian jet stream induces intrusion of the upper-level high potential vorticity from midlatitudes into the subtropical WNP, incites deep convections, and forces the WNP high to shift rapidly to the south of Japan (Wu and Chou 2012). This corresponds to the end of the Baiu rainy season. Because of the blocking effect of the Himalayas, the center of the South Asian high is constrained to approximately 30°N. Thus, the WNP high is located to the north of the South Asian high, the position of which is reversed from Phase 2. This is clearly displayed in Fig. 7c, f. With the WNP high shifting to the midlatitudes, the East Asia–WNP is controlled by the tropical meteorology in the WNP. The strongest typhoon activities occur in August. The topographically associated regulation of the connection between the South and East Asian monsoons is weak in Phase 3.

Because the South Asian high is strongly controlled by the South Asian summer monsoonal heating, which has a geographical limit in latitude advancement, we postulate that these two monsoon systems are remotely connected through the upper tropospheric circulation. In addition to the South Asian high, there are other important factors affecting the season evolution of the East Asian monsoon. For example, eastern extension of the South Asian high has been known to be also influenced by the so-called Silk Road pattern, particular in August (Enomoto et al. 2003). Besides, interactions between the East Asian summer monsoonal heating and the South Asian high in interannual timescale have been found (Wei et al. 2015; Zhang et al. 2016).

5 Sensitivity to the meridional position of the Himalayas

5.1 Experimental design

We have diagnosed using reanalysis data in the previous sections that the Himalayas block the South Asian summer monsoon from rapidly expanding northward after mid-June and constrain the center of the upper-level South Asian high to near 30°N before the end of Phase 2 of the summer monsoon in late July. It is reasonable to expect that the candle-like effect of the blocking becomes weaker when the Himalayas are located further north, i.e. farther away from the source region of moisture. In the extreme case where the Himalayas are removed, the monsoon would penetrate further north but become substantially weaker.

Previous studies have indicated that the seasonality and regional characteristics of the Asian summer monsoon can be effectively simulated in the CAM (Boos and Kuang 2010; Chiang et al. 2015; Wu et al. 2016). To further explore the effect of the Himalayas on constraining the center of the upper-level South Asian high to approximately 30°N in Phase 2, we conducted in this study four sets of CAM simulations with differences in topography (at 1° horizontal resolution). The sea surface temperature in the simulations is fixed at climatological monthly mean values; thus, circulation changes can only be attributed to the dynamic and thermodynamic effects of the topography. In addition to the control simulation (T1, Fig. 8a), which has an actual topography of 1° resolution, we first reduce the height of a large region (45–125°E, 20–60°N) by 90% in the second set of simulations (T0, Fig. 8b). The Himalayas are then restored over the flat Asian continent in the third set of simulations (T0H, Fig. 8c). The final set of simulations (T0NH, Fig. 8d) is conducted by including the same topography as T0H, except shifting the Himalayas northward by 5° latitudes.

Fig. 8
figure 8

Topography (km) in the experimental simulations: a T1, b T0, c T0H, and d T0NH. e Seasonal variation of the location of the South Asian high at 200 hPa. In e, the x-axis is time ranging from May 15 to July 31 and 7-day running mean has been applied. The interval between points is 5 days

Full size image

What we can learn from the comparison among T0, T0H, and T0NH is the impact of the Himalayas and their latitudinal position, on basis of a paleoclimate-style geography where the Asian continent is extremely flat. The results provide a simple insight for understanding the blocking and candle heating of the Himalayas. It is noticed that the altitude of the Tibetan Plateau is about 5 km, and the monsoon would still be blocked by the plateau even devoid of the Himalayas. Therefore, the T0 experiment is chosen to eliminate the blocking effect by all mountains and high lands, especially the Tibetan Plateau, so that the blocking effect due to the Himalayas can be cleanly isolated. We have also compared extra experiments with the similar idea of testing the impact of the latitudinal position of the Himalayas, by reducing the height of the Himalayas and expanding the Himalayas southward to 20°N (can be identified as T1NH and T1SH). The contrast of the blocking effect of the Himalayas between T1SH and T1NH is similar to that between T0H and T0NH (not shown in the figure). Nevertheless, the thermal and mechanical effects of the massive Tibetan Plateau–Himalayas make the explanation much more complex. That is one of the reasons for choosing T0 in this study.

5.2 Sensitivity to the Himalayas on a flat Asian continent

Figure 8e displays the seasonal variations of the South Asian high center in the four sets of simulations. Comparing the case of a flat Asian continent (T0) with that of the actual topography (T1) illustrates that the impact of the topography is large. Notably, the northward propagation of the high center is substantially delayed and can only reach lower latitudes throughout the entire summer when all mountains are removed. When the Himalayas are inserted into the flat Asian continent (T0H), the northward propagation of the high center accelerates (T0H versus T0). The high center reaches 28.5°N in early July and remains nearly stagnant afterward, which is half month earlier than T0. This latitude coincides with the southern edge of the Himalayas (Fig. 8c), indicating that the high center is blocked from moving northward by the mountain range. To further demonstrate the impact of the Himalayas on the evolution of the upper-level high center, the curves of T0NH and T0H show that the high center can reach higher latitudes when the mountain range is moved northward. The displacement of the high center by 5 °N (33.5 versus 28.5°N) in late July is also consistent with the displacement of the Himalayas.

The three curves of T0, T0H, and T0NH also indicate that without the Himalayas and, hence, devoid of candle heating, the northward expansion of the monsoon is slow and cannot penetrate far north (red curve). When the Himalayas are present (T0H and T0NH), the monsoon expedites its northward expansion and can reach higher latitudes. Comparing T0H and T0NH, the monsoon expands northward more rapidly, particularly in the Phase 1, when the Himalayas are located at lower latitudes because of a stronger blocking effect. This stronger blocking effect is demonstrated in Fig. 9 which displays the meridional circulation along 90°E in Phase 1 (upper row) and Phase 2 (lower row). Compared with T1 which has the actual topography, the candle heating, as indicated by the strong upward motion at the southern slope of the Himalayas in Fig. 9a, e, disappears in T0 which has a flat Asian continent (Fig. 9b, f). The candle heating is restored in T0H associated with the enhancement of the low-level southerly wind (Fig. 9c, g). When the Himalayas is located further north by 5°, the southerly wind (green shadings in Fig. 9d, h) can penetrate further northward, but the candle heating is weaker as indicated by the weaker vertical motion. This weaker candle effect is caused by the fact that the Himalayas is now located further away from the heat source in the tropics.

Fig. 9
figure 9

Meridional wind speeds (shadings, m s−1) and vertical motion (contour lines, Pa min−1) along 90°E: a–d Phase 1 of T1, T0, T0H, and T0NH; e–h Phase 2 of T1, T0, T0H, and T0NH. Black bars denote topography, and positive contour lines for vertical motion are not displayed

Full size image

We used the meridional wind shear between 200 and 850 hPa in (70–110°E, 10–30°N) as suggested by (Goswami et al. 1999) to quantify the strength of the Indian monsoon Hadley cell. In Phase 1, the Hadley cell index is approximately 1.70 m s−1 in T1, 0.03 m s−1 in T0, 0.24 m s−1 in T0H, and −0.07 m s−1 in T0NH. In Phase 2, the Hadley cell is significantly stronger. It is 4.33 m s−1 in T1, 2.21 m s−1 in T0, 3.85 m s−1 in T0H, and 3.00 m s−1 in T0NH. Compared with T0, the Hadley cell in Phase 2 intensifies by 74% in T0H and 36% in T0NH. Accordingly, the Hadley cell is stronger with (T0H) than without (T0) the Himalayas; also, the Hadley cell becomes weaker as the Himalayas are moved northward of its present location (T0NH).

Figure 10 displays the spatial distributions of the mean vertically integrated total heating, moisture fluxes at 925 hPa, and the geopotential heights at 200 and 700 hPa in the simulations. In Phase 1 (left column), the Asian summer monsoon is weak without the Tibetan Plateau heating and the Himalayas blocking, and the South Asian and WNP highs are also weak (Fig. 10a, T1 figure is not shown). Heating is weak except in the Indian Ocean and the Bay of Bengal. When the Himalayas are restored, the monsoonal heating associated with the lower tropospheric moisture convergence increases over the Bay of Bengal and the foothills of the Himalayas (Fig. 10b). Meanwhile, the northward transport of moisture is blocked by the mountains. The consequent strengthening of the WNP high favors more heating (or precipitation) in the regions surrounding Taiwan. When the Himalayas are moved to a position north of its current position (T0NH), the monsoon heating and South Asian high are weakened, indicating a weakened monsoon when the regions of the candle heating are moved north and away from the source of moisture (Fig. 10c). The strengthened westward extension of the WNP high enhances heating between Taiwan and Japan, which is located at the northwestern flank of the WNP high. Notably, a northward shift of the Himalayas enhances moisture transport (southwesterly) in southern Japan (T0NH, Fig. 10c).

Fig. 10
figure 10

a Phase 1 (May 15–June 15) mean vertically integrated total heating (shadings, W m−2), moisture fluxes at 925 hPa (vectors, g kg−1 m s−1), and geopotential heights (contour lines, m) at 200 hPa (orange) and 700 hPa (purple) in the experimental simulation of T0. b Phase 1 mean changes of heating (shadings) and moisture fluxes from T0 to T0H and the geopotential heights in T0H. c Same as b except for the T0NH. d–f Same as a–c except for Phase 2 (June 15–July 24)

Full size image

In Phase 2 (right column of Fig. 10), the Meiyu–Baiu band is visible between Taiwan and Japan (Fig. 10d). The restoration of the Himalayas strengthens the western strength of the WNP high. The consequently enhanced southerly flow increases heating over East Asia (west of Taiwan, Fig. 10e). When the Himalayas are moved northward (T0NH), the South Asian high can penetrate further north with the ridge over north of Taiwan (Fig. 10f versus e), and regions of enhanced lower level southwesterly and heating exhibit over East Asia (Fig. 10f). Interestingly, compared with T0H, the western stretch of the WNP high in T0NH does not shift northward with the Himalayas, which may be attributable to a weaker condensational heating in South Asia. Apparently in the T0NH with the Himalayas’ blocking effect taking place at a location north of its present location, the enhanced East Asian heating is primarily attributable to the northward shift of the center of the upper-level high, indicating a weakened control of the Meiyu–Baiu by the WNP high.

6 Concluding remarks

According to the evolution of the lower-level pressure in the East Asia–WNP and the westerlies in South Asia, both the East and South Asian summer monsoons can be divided into strengthening (May 15–June 15), mature (June 15–July 24), and retreating (after 24 July) phases. These phases of the two monsoons are identical in duration. Because South Asia and the East Asia–WNP are distinct in geographical location and topography, it is unlikely that the three identical phases are caused by direct connections between these monsoons, such as the connection between the South Asian westerlies and the East Asian southerlies. Thus, we argue that the two monsoon systems are indirectly connected through the evolution of the upper tropospheric circulation.

6.1 Upper tropospheric circulation as a bridge connecting the South and East Asian monsoons

After the onset of the South Asian summer monsoon in mid-May, the lower-level westerlies and southerlies in the Bay of Bengal continue to strengthen in June. The South Asian monsoon expands northward along the west coast of Indochina and eastern sector of the Bay of Bengal from early June and up to the head of the Bay of Bengal and the foothills of the Himalayas in late June. During the same period, the southerlies and westerlies in the Arabian Sea also intensify, but the Indian Subcontinent is under the influence of subsiding air above a shallow surface layer of a thermal low. Until early July, the air that is subsiding from regions of Iran–Afghanistan–Pakistan retreats northward, and the monsoon expands westward from the head of the Bay of Bengal into the Gangetic Plain. The monsoon reaches a maximum strength in late July and retreats slowly in August. In response to the evolution of the South Asian summer monsoonal heating, the center of the upper tropospheric high pressure system also propagates northwestward from Indochina in mid-May to the Himalayas–southern Tibetan Plateau in late June. With the northward expansion of the monsoon blocked by the Himalayas, the center of the South Asian high is constrained to latitude of approximately 27–30°N in late June through early August.

In the East Asia–WNP, the monsoon rain is associated with the frontal system located on the northwestern flank of the WNP high. In Phase 1 (the strengthening phase) of the summer monsoon, the WNP high extends westward from the Philippine Sea to the northern SCS, and the frontal system located on the northwestern flank of the WNP high ridge covers southeastern China and Taiwan. This is the early Meiyu season in southeastern China and Taiwan. In Phase 2 (the mature phase) of the summer monsoon, the WNP high moves to a location near Taiwan, and the rainy band extends from central eastern China to Japan and Korea. This is the core Meiyu–Baiu season. Because strong tropospheric heating induces high pressure in the upper troposphere, the high upper tropospheric pressure in the East Asia–WNP is located to the north of the WNP high pressure ridge in both Phases 1 and 2 of the monsoon.

In response to the northwestward expansion of the South Asian summer monsoon, the center of the South Asian high shifts quickly from Indochina to the foothills of the Himalayas in monsoon Phase 1 and remains nearly stagnant at approximately 27–30°N in Phase 2. Because the ridge of the upper tropospheric pressure in the East Asia–WNP is an eastern extension of the South Asian high, it also shifts northward quickly in Phase 1 and slowly in Phase 2. The ridge of the upper tropospheric pressure is coupled to and located north of the ridge of the lower tropospheric pressure in the East Asia–WNP. Thus, the evolution of the WNP high is constrained by the evolution of the South Asian high, and the East Asian summer monsoon is connected to the South Asian summer monsoon through the impact of the South Asian monsoon heating on the upper tropospheric circulation.

6.2 Numerical results

Analyses of the simulated data sets indicate that one of the crucial factors that causes the slow northward motion of the Meiyu–Baiu–Changma rain band in Phase 2 is the blocking of the South Asian summer monsoon from expanding northward by the Himalayas. Experiments using the CAM of NCAR further indicate that the near-stationary nature of the East Asian frontal system in Phase 2 is remotely related to the blocking of the South Asian summer monsoon by the Himalayas. The model simulations are based on a paleoclimate-style experiment in which the influence of the Himalayas and its latitudinal position in a flat Asian continent (T0) on the monsoon is investigated. In the extreme case where the Himalayas are removed, the monsoon can penetrate further north but is substantially weakened. When the Himalayas are inserted into a flat Asian continent (T0H), the northward propagation of the South Asian high accelerates until reaching at the southern edge of the Himalayas, indicating that the high center is blocked from moving northward by the mountain range. When the Himalayas are located further north by 5° (T0NH), the southerly wind can penetrate further northward, but the candle heating is weaker as indicated by the weaker vertical motion. This weaker candle effect is caused by the fact that the Himalayas are located further away from the heat source in the tropics. The enhanced southwesterly and heating in East Asia penetrate further northward in T0NH than in T0H, which is consistent with the finding that the upper tropospheric high ridge also shifts further northward with the Himalayas. With a weaker condensational heating in South Asia, the T0NH has a weaker WNP high than T0H. This is an indication of a weakened control of Meiyu–Baiu by the WNP high, which, in turn, is closely related to the eastern stretch of the South Asian high.