1. Introduction
The majority of the rainfall over tropical Southeast Asia equatorward of 10°N occurs in the winter (Cheang 1977), while over the northern part of Southeast Asia and East Asia, the rainfall occurs in the summer (Ramage 1952; Chen et al. 2004). Stretching from 8° to 23.5°N (Tropic of Cancer), Vietnam is located between these two monsoon rainfall regimes. The rainfall season varies from the summer [June–September (JJAS)] regime in the north, the fall [September–November (SON)] regime in the central, and a combination of these two (May–November) regimes in the south (Chen et al. 2012). According to the Dartmouth Flood Observatory (DFO; DFO 2010) flood archive, heavy rainfall events leading to great floods in central Vietnam usually occur in late fall. As shown by the y–t diagram of the station rainfall superimposed with wind vectors along the coast of Vietnam (Fig. 1a), rainfall during October–November 2008 occurred primarily in central and southern Vietnam. Because late fall is the dry season in northern Vietnam, the southeastern Asian community was surprised by the heavy rainfall event on 30–31 October 2008, which occurred near Hanoi, Vietnam. The peak rainfall during this event reached 433.3 mm day−1 on 31 October 2008 (Fig. 1b), leading to the first great flood in the past 25 years in this city. The high-impact flood event caused 92 fatalities, damaged 55 000 houses, left 40 000 people homeless, and resulted in approximately $500 million (U.S. dollars) worth of property damage (Son 2008).
(a) The y–t diagram of station rainfall superimposed with wind vectors (interpolated from the gridded NCEP GFS surface wind) and isotach of surface winds; stations are marked by red dots along the eastern coasts of China, Vietnam, and Malaysia, and vortices crossing the coast of Vietnam during October–November 2008 are indicated by red arrows. (top left) Terrain height represented by the color scale on the top. The magnitude of the wind vector and the isotach contour |V(925 hPa)| are shown on the top right of (a), while the axis for precipitation P is shown in the bottom right. (b) The precipitation time series (blue bars) at the Hanoi surface station (WMO 48820) for October–November 2008 along with a climatological precipitation time series (red bars); underneath the time series, blue crosses mark the dates the vortices crossed the coast of Vietnam and the red cross indicates the date a “bomb” formed southeast of the southern tip of the Kamchatka Peninsula. (c) The JMA surface analysis chart at 0000 UTC 30 Oct 2008; the thick red cross indicates the location of the midlatitude cyclone, which would develop into an explosive cyclone; the thin blue cross indicates the vortex cyclone, which would become an HRF cyclone. Surface observations can be found in the original JMA chart (http://www.jma.go.jp/en/g3/). (d) The 850-hPa streamline chart superimposed with TRMM precipitation on 30 Oct 2008; the propagation path of the easterly disturbance is added with the locations of the easterly disturbance’s trough line (short dark red line), cold surge vortex–SCS cyclone (red crosses), and the HRF cyclone on 30 Oct 2008. Blue arrows indicate the direction of the cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The y–t diagram of station rainfall superimposed with wind vectors (interpolated from the gridded NCEP GFS surface wind) and isotach of surface winds; stations are marked by red dots along the eastern coasts of China, Vietnam, and Malaysia, and vortices crossing the coast of Vietnam during October–November 2008 are indicated by red arrows. (top left) Terrain height represented by the color scale on the top. The magnitude of the wind vector and the isotach contour |V(925 hPa)| are shown on the top right of (a), while the axis for precipitation P is shown in the bottom right. (b) The precipitation time series (blue bars) at the Hanoi surface station (WMO 48820) for October–November 2008 along with a climatological precipitation time series (red bars); underneath the time series, blue crosses mark the dates the vortices crossed the coast of Vietnam and the red cross indicates the date a “bomb” formed southeast of the southern tip of the Kamchatka Peninsula. (c) The JMA surface analysis chart at 0000 UTC 30 Oct 2008; the thick red cross indicates the location of the midlatitude cyclone, which would develop into an explosive cyclone; the thin blue cross indicates the vortex cyclone, which would become an HRF cyclone. Surface observations can be found in the original JMA chart (http://www.jma.go.jp/en/g3/). (d) The 850-hPa streamline chart superimposed with TRMM precipitation on 30 Oct 2008; the propagation path of the easterly disturbance is added with the locations of the easterly disturbance’s trough line (short dark red line), cold surge vortex–SCS cyclone (red crosses), and the HRF cyclone on 30 Oct 2008. Blue arrows indicate the direction of the cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The y–t diagram of station rainfall superimposed with wind vectors (interpolated from the gridded NCEP GFS surface wind) and isotach of surface winds; stations are marked by red dots along the eastern coasts of China, Vietnam, and Malaysia, and vortices crossing the coast of Vietnam during October–November 2008 are indicated by red arrows. (top left) Terrain height represented by the color scale on the top. The magnitude of the wind vector and the isotach contour |V(925 hPa)| are shown on the top right of (a), while the axis for precipitation P is shown in the bottom right. (b) The precipitation time series (blue bars) at the Hanoi surface station (WMO 48820) for October–November 2008 along with a climatological precipitation time series (red bars); underneath the time series, blue crosses mark the dates the vortices crossed the coast of Vietnam and the red cross indicates the date a “bomb” formed southeast of the southern tip of the Kamchatka Peninsula. (c) The JMA surface analysis chart at 0000 UTC 30 Oct 2008; the thick red cross indicates the location of the midlatitude cyclone, which would develop into an explosive cyclone; the thin blue cross indicates the vortex cyclone, which would become an HRF cyclone. Surface observations can be found in the original JMA chart (http://www.jma.go.jp/en/g3/). (d) The 850-hPa streamline chart superimposed with TRMM precipitation on 30 Oct 2008; the propagation path of the easterly disturbance is added with the locations of the easterly disturbance’s trough line (short dark red line), cold surge vortex–SCS cyclone (red crosses), and the HRF cyclone on 30 Oct 2008. Blue arrows indicate the direction of the cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Cheang (1977) observed the winter rainfall in tropical southeastern Asia is primarily produced by intensified equatorial vortices in the presence of cold surge flows. Using modern National Centers for Environmental Prediction (NCEP-II) reanalyses, Tsay (2004) extended and documented Cheang’s observation to cover the period of 1979–2004. Based on Tsay’s documentation, the reexamination of Cheang’s intensification mechanism of heavy rainfall vortices and the climatology of rainfall produced by these vortices was proposed in the Science Plan of the East Asian Monsoon Experiment (EAMEX; Chen 2007) in collaboration with the Monsoon Asian Hydro-Atmosphere Scientific Research and Prediction Initiative (MAHASRI; MAHASRI 2006). Later, Yokoi and Matsumoto (2008) examined the environment favorable for the occurrence of heavy rainfall vortices in central Vietnam and claimed the coexistence of a tropical depression–like vortex and a cold surge flow is required. Truong et al. (2009) suggested the topographic–environmental flow interaction may also facilitate the development of heavy rainfall–flood (HRF) events in central Vietnam.
Despite recent efforts to explore the cause of heavy rainfall in tropical southeastern Asia suggested by Cheang (1977), several basic aspects of this cause have not been addressed. They include the following:
The equatorial vortex described by Cheang is basically the easterly wave. How does this type of wave evolve into a vortex?
This vortex, named as cold surge vortex1 hereafter because of its formation mechanism, may intensify across the South China Sea (SCS) by its interaction with a cold surge flow. What is the dynamic mechanism to intensify this vortex to become an HRF cyclone?
As a result of the intensification of a cold surge vortex being a local dynamic process, what weather systems in both latitudinal regions are involved with the midlatitude–tropical interaction, observed from a perspective of large-scale circulation?
How is the HRF event maintained hydrologically?
In the midlatitudes, the cold surge is formed by the northerly or northwesterly flow existing between a surface low–high couplet and attendant upper-level trough–ridge couplet. This couplet takes about a week to move across the northeastern Asian coast and another week for the next trough–ridge couplet, linked to a new cold surge, to cross this same coast. However, these two trough–ridge couplets may not always be connected. Thus, the time interval between occurrences of two major cold surges may be a couple of weeks (e.g., Petterssen 1956; Palmén and Newton 1969). Figures 1a,b show that the three vortices reached the eastern coast of Vietnam during October–November 2008. The time interval between vortices is 18 days. For the occurrence of a strong cold surge across East Asia, the upper short-wave disturbance coupled with this cold surge can be tracked upstream to the western North Atlantic (Joung and Hitchman 1982). A low pressure system coupled with the cold surge may also lead to the development of explosive midlatitude cyclones (Sanders and Gyakum 1980; Chen et al. 1992) downstream in the western North Pacific.
In a climatological sense, the East Asian cold surges may occur 3–6 times during October–November, when there is coupling with the eastward-propagating short-wave trough–ridge system. The surface lows associated with these cold surges lead to the formation of northwestern Pacific cyclogenesis. As a result of a life cycle of 5 days for easterly waves, it is possible for more than 10 easterly waves to propagate across the Philippine Sea and farther westward during this time. Thus, several cold surge vortices are likely to form. According to the DFO HRF event archive, only one to two HRF events occur during the late fall in central Vietnam. The differences in the population of these three types of tropical disturbances during this time suggest that an additional mechanism, beyond the formation of the cold surge vortex, regulates the synoptic development of the HRF event. As indicated by the convective activity in the western tropical Pacific, the intraseasonal 30–60-day mode [i.e., Madden–Julian oscillation (MJO; Madden and Julian 1971, 1972)] is active over this region (e.g., Lau and Chan 1985). The intraseasonal mode modulates not only the southeastern Asian monsoon, but also the easterly wave activity in the western tropical Pacific (e.g., Chen and Weng 1996). In view of the possible coexistence of the easterly wave, cold surge flow, and the intraseasonal 30–60-day oscillation, the 2008 Hanoi HRF event may develop from an interaction among these monsoon modes.
The 2008 Hanoi HRF event occurred farther north than the other Vietnam events in the fall. The interannual variation of the atmospheric circulation in the western tropical Pacific during fall 2008 should have some effect on the track of the HRF cyclone and the water vapor transport toward the northern part of Southeast Asia. The interannual circulation variation is generally considered as the atmospheric response to sea surface temperature (SST) change. For instance, the response of the summer circulation to SST anomalies in the western Pacific is depicted by the Japan–Pacific Oscillation (Nitta 1987), while the response of the winter circulation is depicted by a short-wave train along the North Pacific rim (Chen 2002). The dynamics of a large-scale circulation in the midlatitudes belong to the Rossby regime, but those in the tropics belong to the Sverdrup regime (Chen 2005, 2010). To relate the interannual variation of the tropical circulation in fall 2008 over the tropical southeastern Asia–western Pacific region to the rainfall change in this region, the relationship between diabatic heating generated by rainfall and divergent and rotational circulations through the Sverdrup dynamics will be used to infer the maintenance of the 2008 anomalous circulation and its effect on the track of the Hanoi HRF cyclone and water vapor transport.
An effort is made to explore several issues concerning the formation of the Hanoi HRF event on 30–31 October 2008 and the maintenance of its rainfall: 1) what is the synoptic environment favorable for the formation of this event in the context of the midlatitude–tropical interactions, 2) how is this event dynamically formed and hydrologically maintained through multiple-scale processes of several potential monsoon modes, and 3) how is this event affected by the interannual variation of the eastern Asian circulation that occurs in northern Vietnam. This effort will be pursued with the following analyses outlined below. The data used in this study are presented in section 2. The synoptic evolution of this HRF event is illustrated using two sets of streamline charts in section 3. One set emphasizes the eastward propagation of the midlatitude upper trough–ridge system related to the eastern Asian cold surge and the explosive cyclone in the western North Pacific. The other shows the westward propagation of easterly disturbances and the formation of a cold surge vortex south of the Philippines. The other also shows the vortex’s development into a sizable cyclone linked to the occurrence of the Hanoi HRF event. The intensification of this cold surge vortex is illustrated in section 4 in terms of a spectral analysis of the zonal winds of the HRF cyclone. The streamfunction tendency created by significant monsoon modes identified by the spectral analysis of the zonal wind is also shown in section 4. The maintenance of the Hanoi HRF event by multiple scale processes is presented in section 5, while the possible impact of the anomalous circulation during fall 2008 on the water vapor supply to maintain this HRF event is discussed in section 6. Conclusions, including summary and suggestions for future study, are provided in section 7.
2. Data
To illustrate the development of the Hanoi HRF cyclone from a subsynoptic perturbation, high spatial and temporal resolution data are needed. This resolution requirement is met by initial analyses of the NCEP GFS (Kanamitsu et al. 1991; Yang et al. 2006) with a 0.5° longitude × 0.5° latitude horizontal resolution operationally issued every 6 h. The synoptic development of the HRF cyclone is presented in terms of streamline charts superimposed with rainfall. These streamline analyses are verified with surface analyses by three operational centers. The surface high and low centers in the subtropics and midlatitudes of the Northern Hemisphere are identified with the National Climatic Data Center (NCDC) Service Records Retention System (SRRS) analysis and forecast charts archived at the National Oceanic and Atmospheric Administration (NOAA) Operational Model Archive Distribution System (NOMADS; Rutledge et al. 2006; data are available online at http://nomads.ncdc.noaa.gov/), while the upper ridges and troughs are verified by the SRRS 500-hPa geopotential height maps. The surface analysis maps issued by the Thai Meteorological Department (TMD; TMD 2008) are used to replace the SRRS surface analysis maps to match the synoptic development of the weather systems in the tropics with those in the midlatitudes. The surface analysis maps issued by Japan Meteorological Agency (JMA; JMA 2008) are also used to verify the HRF cyclone depicted in the midlatitudes and the tropics.
The development and maintenance of the 2008 Hanoi HRF event are illustrated by means of the water vapor budget analysis, which requires rainfall data with a spatial and temporal resolution comparable to the NCEP GFS initial analyses. The only rainfall data that can satisfy this requirement are derived rainfall measurements from the Tropical Rainfall Measurement Mission (TRMM; Simpson et al. 1996), which covers a global domain between 50°S and 50°N with a 0.25° longitude × 0.25° latitude resolution available every 3 h. These satellite-derived measurements are supplemented with rainfall measurements from World Meteorological Organization (WMO) surface stations every 3 h. In addition, the flood disaster events during October–November in Vietnam collected by the DFO since 1979 are also used as a reference in the discussion.
3. Synoptic development of the Hanoi heavy rainfall event
It has been suggested that the 2008 Hanoi HRF event was caused by the intensification of an easterly disturbance through the midlatitude–tropical interactions. To infer possible processes involved in the development of this event from the perspective of the large-scale environment in eastern and southeastern Asia, a synoptic analysis of the weather systems in both the midlatitudes and the tropics is presented.
a. Midlatitudes
Observing severe polar outbreaks (i.e., strong eastern Asian cold surges) in northeast Asia, Joung and Hitchman (1982) tracked the location of the upper troughs coupled with cold surges upstream to the western North Atlantic. Their tracking approach uses the 2-day difference of upper geopotential height. One basic criterion of tracking upper troughs is the life span of any trough used in this approach should be at least 2 days or longer. If the decaying process of troughs or small-scale waves is shorter than 2 days, the tracks of these weak perturbations may not be accurately provided by this approach. Therefore, it is supplemented by the track of surface cyclones coupled with these troughs, using NCEP surface analysis charts issued at 0000 UTC. The upper trough, coupled with the surface cyclone linked to the cold surge involved with the development of the Hanoi HRF event, may be traced westward to the western North Atlantic. The details of the daily synoptic evolution of the upper trough–ridge system and the coupled surface high–low system, which led to this cold surge event, were traced and archived by EAMEX (2010).
On 20 October, the Arctic storm at the northern end of the central Asian trough line (southeast of the Kara Sea) and the newly formed surface storm east of the southern tail of this trough moved eastward with this trough (Fig. 2b). This storm was steered by the upper trough to pass through the valley between the Altai Mountains and Tianshan, but blocked by the Tibetan Plateau (Fig. 2a). Developing into a large-scale cutoff low, the upper trough propagated east-northeast over the next couple of days (Fig. 2d). The propagation of this trough allowed shortwave disturbances moving around it and repeated surface development along the northeastern Asian coastline and the adjacent ocean. On 23 October, the surface low moved across the Mongolia plateau and a cold surge flow formed southeast of Lake Baikal in Kazakhstan between this surface low to its northeast and the surface high pressure over the Tibetan Plateau to its southwest (Fig. 2c).2 During the following week, the eastward-propagating upper trough slowed down and deepened more, from when it moved across the strong baroclinic zone over the eastern seaboard of northeast Asia, as shown from the locations of the trough in Figs. 2f,h,j. The deepening of this trough has several important implications related to the intensification of the tropical easterly disturbance linked to the synoptic development of the Hanoi HRF event and potential use in medium-range forecasting of the HRF events along the eastern coast of Vietnam:
The intensification of the surface low northeast of the cold surge flow coupled with the deepening upper trough led to the formation of an explosive cyclone (i.e., “bomb” coined by Sanders and Gyakum 1980) southeast of the Kamchatka Peninsula (indicated by a thick red cross in Fig. 1d) on 31 October. However, the synchronous occurrence of the Hanoi HRF event with this explosive cyclone may offer a means to forecast such events.
In Fig. 2c, a newly formed cold surge flow appeared along the coast of northeastern Asia and provided a means for a midlatitude–tropical interaction. It is also shown in Figs. 2e,g,i that the Pacific surface high pressure associated with the cold surge helped enhance the tropical easterlies after 20 October. The details of these interactions related to the development of the cold surge vortex will be discussed in section 4.
The equatorward cold surge flow across the northern part of the South China Sea and the western tropical Pacific trade winds merged in establishing strong easterlies. The southward deepening of an upper trough across northern India (Fig. 2d) facilitated the development of a surface cyclone over the western Bay of Bengal (Fig. 2c). This cyclone development will be further illustrated in section 3b. The tropical westerly flow associated with this cyclone and the South China Sea easterlies form a strong cyclonic shear environment favorable for the development of a synoptic vortex over the tropical South China Sea and Indochina.
Joung and Hitchman (1982) estimated the average time for an upper trough to propagate from the western North Atlantic to the coast of East Asia and to develop a strong cold surge event is about 6–7 days. As indicated by the trajectory (thick red lines) of surface cyclones (red dots) added with the corresponding upper trough lines (thin red solid lines) in Fig. 3, the time needed for their movements from the initiation of the surface cyclone north of Tanshian to the western North Pacific is about 10 days (~20–30 October 2008) longer than measured by Joung and Hitchman (1982).
Fig. 2. (left) NCEP SRRS analysis charts and (right) GFS 500-hPa streamlines superimposed with zonal wind speed. The dark red solid line, red cross, and red thick, solid line are trajectories of cyclones, locations of cyclones, and trough lines at 0000 UTC on the specified date, respectively. (left) The date for each chart is on the top-left corner of the chart. Symbols, H and L, are surface highs and lows, respectively, identified by the NCEP SRRS analysis charts. Locations of several important geographic features (e.g., Kara Sea, Lake Baikal, Altai Mountains, and Tianshan) are referred to in Fig. 3. The areas colored green and red represent high terrain and high pressure, respectively. The mean sea level pressure thickness is depicted by black contours and red dashed lines, respectively with contour intervals of 4 hPa for MSLP and 40 m for thickness (MSLP).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Fig. 2. (left) NCEP SRRS analysis charts and (right) GFS 500-hPa streamlines superimposed with zonal wind speed. The dark red solid line, red cross, and red thick, solid line are trajectories of cyclones, locations of cyclones, and trough lines at 0000 UTC on the specified date, respectively. (left) The date for each chart is on the top-left corner of the chart. Symbols, H and L, are surface highs and lows, respectively, identified by the NCEP SRRS analysis charts. Locations of several important geographic features (e.g., Kara Sea, Lake Baikal, Altai Mountains, and Tianshan) are referred to in Fig. 3. The areas colored green and red represent high terrain and high pressure, respectively. The mean sea level pressure thickness is depicted by black contours and red dashed lines, respectively with contour intervals of 4 hPa for MSLP and 40 m for thickness (MSLP).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Fig. 2. (left) NCEP SRRS analysis charts and (right) GFS 500-hPa streamlines superimposed with zonal wind speed. The dark red solid line, red cross, and red thick, solid line are trajectories of cyclones, locations of cyclones, and trough lines at 0000 UTC on the specified date, respectively. (left) The date for each chart is on the top-left corner of the chart. Symbols, H and L, are surface highs and lows, respectively, identified by the NCEP SRRS analysis charts. Locations of several important geographic features (e.g., Kara Sea, Lake Baikal, Altai Mountains, and Tianshan) are referred to in Fig. 3. The areas colored green and red represent high terrain and high pressure, respectively. The mean sea level pressure thickness is depicted by black contours and red dashed lines, respectively with contour intervals of 4 hPa for MSLP and 40 m for thickness (MSLP).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Fig. 3. The propagation path for the concerned cyclone coupled with the upper trough related to the late-October 2008 cold surge and the Hanoi HRF event. Red dots and thin solid red lines are the daily locations of cyclones and the 500-hPa troughs.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Fig. 3. The propagation path for the concerned cyclone coupled with the upper trough related to the late-October 2008 cold surge and the Hanoi HRF event. Red dots and thin solid red lines are the daily locations of cyclones and the 500-hPa troughs.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Fig. 3. The propagation path for the concerned cyclone coupled with the upper trough related to the late-October 2008 cold surge and the Hanoi HRF event. Red dots and thin solid red lines are the daily locations of cyclones and the 500-hPa troughs.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
b. Tropics
The interactions between tropical easterly disturbances and eastern Asian cold surge flows may lead to the development of cold surge vortices, which bring winter rainfall to tropical southeastern Asia (Cheang 1977, 1987). McBride (1995) pointed out that the interaction of tropical easterly disturbances with the monsoon trough is a tropical cyclogenesis mechanism in the western tropical Pacific. Can the interaction of the easterly disturbance with the cyclonic shear flow into the South China Sea during the fall of 2008 lead to the formation of a vortex? Are other factors affecting the evolution of this vortex into the HRF cyclone that caused the 2008 Hanoi heavy rainfall event and its subsequent dissipation?
Existing in the shallow atmospheric layer above the surface, the cold surge flow can be well depicted in terms of 925-hPa streamline line charts prepared with the NCEP GFS analysis (the middle column of Fig. 4). The surface synoptic condition in southern and southeastern Asia portrayed by the 925-hPa streamline charts is further illustrated with the TMD surface analysis charts (the left column of Fig. 4). On 23 October 2008, a tropical easterly disturbance (a yellow circular spot in Fig. 4a and a short thick solid red line in Fig. 4b) propagated to the western tropical Pacific. At the same time, a low-level cyclone appeared (red circular spot) over the western Bay of Bengal adjacent to India. Two days later (25 October), the easterly disturbance moved closer to the southern Philippines. At this time, a small shallow vortex close to the surface formed in the Celebes Sea (Fig. 4e). This vortex was embedded in the tropical shear flow established by cold surge flows from northeastern China and the cold surge flow of the remnant oceanic anticyclone in the North Pacific. On 26 October, the easterly disturbance merged with this shallow vortex (Figs. 4g,h) and formed a cold surge vortex, but did not reach 500 hPa (Fig. 4i). Two days later, 28 October, this vortex grew into a cyclone across the South China Sea, covering Indochina and the tropical SCS (Figs. 4j,k), and extended vertically beyond 500 hPa (Fig. 4l). The formation of a cold surge vortex on 26 October and the development of this vortex into an HRF cyclone3 on 28 October through the interaction of the eastern Asian cold surge flow and the easterly disturbance are discernible in Figs. 4g,h,i. Then, the HRF cyclone and the Bengal cyclone were encircled by a zonally oriented trough (Figs. 4m,n). The encounter of these two cyclones eventually merged into one centered at the west coast of Burma on 2 November (Figs. 4p,q). As indicated by thin red lines, the eastward migration of the Bengal cyclone and the westward propagation of the HRF cyclone are well verified by the TMD surface analysis chart in the left column of Fig. 4. The westward propagation and growth of the easterly disturbance into the HRF cyclone took place along the ITCZ and the southeastern Asian cyclonic shear flow.
Analyses for (top to bottom) (left page) 0000 UTC 23 Oct to 1200 UTC 26 Oct and (right page) 0000 UTC 28 Oct to 0000 UTC 2 Nov. (left) The surface analysis maps of the TMD, (middle) the 925-hPa streamline charts superimposed with TRMM precipitation, and (right) the 500-hPa streamline charts superimposed with zonal wind speed. The propagation paths (red dashed lines) of the Bengal cyclone and an easterly disturbance–cold surge vortex–HRF cyclone: locations for these two types of disturbances are marked by red and yellow circular spots, respectively. Symbols H and L are also added to indicate surface high and low identified by the NCEP SRRS analysis charts. The location of Lake Baikal is referred to in Fig. 3.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Analyses for (top to bottom) (left page) 0000 UTC 23 Oct to 1200 UTC 26 Oct and (right page) 0000 UTC 28 Oct to 0000 UTC 2 Nov. (left) The surface analysis maps of the TMD, (middle) the 925-hPa streamline charts superimposed with TRMM precipitation, and (right) the 500-hPa streamline charts superimposed with zonal wind speed. The propagation paths (red dashed lines) of the Bengal cyclone and an easterly disturbance–cold surge vortex–HRF cyclone: locations for these two types of disturbances are marked by red and yellow circular spots, respectively. Symbols H and L are also added to indicate surface high and low identified by the NCEP SRRS analysis charts. The location of Lake Baikal is referred to in Fig. 3.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Analyses for (top to bottom) (left page) 0000 UTC 23 Oct to 1200 UTC 26 Oct and (right page) 0000 UTC 28 Oct to 0000 UTC 2 Nov. (left) The surface analysis maps of the TMD, (middle) the 925-hPa streamline charts superimposed with TRMM precipitation, and (right) the 500-hPa streamline charts superimposed with zonal wind speed. The propagation paths (red dashed lines) of the Bengal cyclone and an easterly disturbance–cold surge vortex–HRF cyclone: locations for these two types of disturbances are marked by red and yellow circular spots, respectively. Symbols H and L are also added to indicate surface high and low identified by the NCEP SRRS analysis charts. The location of Lake Baikal is referred to in Fig. 3.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
On 23 October, a 500-hPa cutoff low appeared at the southern end of the trough across the Indian subcontinent (Fig. 4c). Underneath this cutoff low was a low-level cyclone centered at the eastern Arabian Sea, southwest of India (Figs. 4a,b). At this time, the eastward extension of the surface cyclone formed another cyclonic center over the western Bay of Bengal. In the next couple of days, the upper-level cutoff low decayed (Fig. 4f) and the corresponding surface cyclone disappeared, but the upper trough deepened southeastward to reach the northwestern part of the Bay of Bengal and the newly formed eastern cyclonic center developed into the Bengal cyclone (Figs. 4g, h). On 26 October, the upper trough deepened further (Fig. 4i) and the Bengal cyclone intensified more (indicated by surface pressure in Fig. 4g). In the next two days, the cold surge vortex developed into an HRF cyclone (Figs. 4j,k). Its vertical extent reached beyond 500 hPa on 30 October, the Bengal cyclone already dissipated (Fig. 4m) and the Bengal trough started to fill (Fig. 4o). The vertical extent of the HRF cyclone connected to the filled Bengal trough like a cutoff low (Fig. 4o). This development facilitated the merger of the Bengal and HRF cyclones. The filled Bengal trough was eventually separated from the upper part of the HRF cyclone (Fig. 4r). This separation led to the decay of this cyclone and the dissipation of the merged Bengal–HRF cyclone (Figs. 4p,q).
It was noted in section 3a that the trough located over the Okhotsk Sea on 30 October moved eastward to the western part of the Bering Sea. This synoptic development resulted in the formation of an explosive cyclone (Sanders and Gyakum 1980) centered over Karaginsky Island, east of the northeastern Kamchatka Peninsula (Fig. 5a). The synchronous occurrence of the Hanoi HRF event and the formation of this explosive cyclone have a synoptically important implication. The HRF cyclone intensified not only through the interaction of the tropical easterly disturbance with the newly formed East Asian cold surge flow, but also through the midlatitude–tropical interaction of the tropical trade easterlies in the western tropical Pacific with the remnant oceanic cold surge flow (Fig. 5a). The y–t diagrams of u(850 hPa) cutting through the center of the HRF cyclone at 109°E, the center of the explosive cyclone at 160°E south of 25°N, and the central part of the trade easterlies at 167°E north of 25°N are shown in Figs. 5b,c, respectively. The maximum westerlies and easterlies of the HRF cyclone and the North Pacific anticyclone are marked by blue and red arrows, respectively, in these two y–t diagrams; the maximum intensities of the HRF cyclone and the North Pacific anticyclone happen simultaneously. The enhanced southwesterly flow into Hanoi coincides with the strongest flow associated with these two weather systems. This synchronization occurs not only for this case, but is common to all 31 HRF events in Vietnam during October–December of 1979–2008, except for two events on 4 October 1992 and 13 October 1995. These two events occurred 2 days after these northwest Pacific explosive cyclones formed.
(a) The 850-hPa streamline chart superimposed with isotachs |V(850 hPa)| The y–t diagrams of u(850 hPa) at (b)109°E and (c) 25°N and 160°E south of 25°N. The color scales of u(850 hPa) are shown on the top right of (b),(c). The cold surge flows in (a) are marked by blue shafts. The longitudes of the (b),(c) u(850 hPa) y–t diagrams are also marked by red lines in (a). The red lines in (b),(c) indicate 31 Oct. The maximum easterlies and westerlies of both the Indochina cyclone and the North Pacific anticyclone are indicated by blue and red arrows, respectively. Open blue arrows are added in (a) to indicate cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The 850-hPa streamline chart superimposed with isotachs |V(850 hPa)| The y–t diagrams of u(850 hPa) at (b)109°E and (c) 25°N and 160°E south of 25°N. The color scales of u(850 hPa) are shown on the top right of (b),(c). The cold surge flows in (a) are marked by blue shafts. The longitudes of the (b),(c) u(850 hPa) y–t diagrams are also marked by red lines in (a). The red lines in (b),(c) indicate 31 Oct. The maximum easterlies and westerlies of both the Indochina cyclone and the North Pacific anticyclone are indicated by blue and red arrows, respectively. Open blue arrows are added in (a) to indicate cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The 850-hPa streamline chart superimposed with isotachs |V(850 hPa)| The y–t diagrams of u(850 hPa) at (b)109°E and (c) 25°N and 160°E south of 25°N. The color scales of u(850 hPa) are shown on the top right of (b),(c). The cold surge flows in (a) are marked by blue shafts. The longitudes of the (b),(c) u(850 hPa) y–t diagrams are also marked by red lines in (a). The red lines in (b),(c) indicate 31 Oct. The maximum easterlies and westerlies of both the Indochina cyclone and the North Pacific anticyclone are indicated by blue and red arrows, respectively. Open blue arrows are added in (a) to indicate cold surge flows.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
4. Formation and development mechanism of the HRF cyclone
a. Spectral analysis of zonal wind
Of three cold surge vortices that approached the Vietnam coast during October–November 2008 (Fig. 1a), one of these vortices intensified and developed into an HRF cyclone. This development is shown by the time series u(850 hPa) at 18° and 4°N (thick black lines in Fig. 6a). The peak values of easterly and westerly zonal winds, respectively, at these two latitudes appear on 31 October; a powerful intensification a week before and a quick weakening 4 days after the HRF cyclone reached its maximum intensity. The monsoon modes involved with this development are revealed from a spectral analysis of a 6-month u(850 hPa) time series (1 August 2008–31 January 2009) at two latitudinal locations, 18° and 4°N, along 109°E (Figs. 6b,c) performed with Madden and Julian’s (1971) scheme. The three monsoon mode signals, 30–60, 12–24, and 5 days, denoted as 
(a) The time series for 
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The time series for
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The time series for
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The contributions from the three monsoon modes to this HRF cyclone intensification are illustrated by the time series for the filtered zonal winds (
b. Streamfunction tendency
Even without the geostrophic constraint, the tropical circulation is well depicted by the streamfunction ψ, which is the inverse Laplace transform of vorticity ζ. As expressed by the vorticity equation, the vorticity tendency ζt is the result of all dynamic processes, including horizontal advection of vorticity, vortex stretching, tilting, vertical advection, and friction, but the last three processes are not significant in our analysis. The advantage of using the streamfunction tendency ψt to illustrate the development of the HRF cyclone is that it provides a simple approach to separate contributions to the development of this cyclone by the three monsoon modes. The cold surge flow exists primarily in the lowest layer of troposphere. To illustrate the role played by the cold surge flow in the formation of a cold surge vortex and its ensuing development, a streamfunction tendency analysis is performed at 925 hPa.
On 25 October, a negative ψt(925 hPa) center appeared over the already existing small shallow vortex in the Celebes Sea (Fig. 7a) ahead of the trough line of an easterly disturbance depicted by a short thick, solid red line. This negative ψt(925 hPa) center not only strengthened this vortex, but also enabled it to propagate westward. In less than a day on 26 October, this easterly disturbance merged with the shallow vortex in the Celebes Sea (Fig. 7b). A lateral interaction of this cold surge vortex with the southward intrusion of the cold surge flow across the South China Sea is shown by the connection between these two synoptic elements through 925-hPa streamlines and the negative ψt(925 hPa) center. Propagating northwestward across the South China Sea in the next three days, the size of this vortex was amplified and enlarged to cover the entire tropical South China Sea (Fig. 4k) and extended vertically beyond the midtroposphere (Fig. 4l). On 29 October, the center of this cyclonic circulation was located at the coast of central Vietnam (Fig. 7c). After this stage, the merger of the HRF cyclone with the India–Bengal trough formed a strong cyclonic shear flow between the eastern Asian cold surge flow in the north and tropical westerlies of the India–Bengal trough in the south, consistent with that of the 925-hPa climatological flow for October–November (not shown). This shear flow provided the environment with increased cyclonic shear vorticity. It also facilitated the interaction between this cyclone and the cold surge flow, and the intensification of the HRF cyclone.
The 925-hPa streamline charts superimposed with total streamfunction tendency [ψt (925 hPa)] for specified dates. The propagation path of the concerned easterly disturbance–cold surge vortex–Indochina cyclone (red dashed line) and locations of these disturbances on a specified date are superimposed on the streamline charts. H and L are surface high and low, respectively, identified by the NCEP SRRS analysis charts.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The 925-hPa streamline charts superimposed with total streamfunction tendency [ψt (925 hPa)] for specified dates. The propagation path of the concerned easterly disturbance–cold surge vortex–Indochina cyclone (red dashed line) and locations of these disturbances on a specified date are superimposed on the streamline charts. H and L are surface high and low, respectively, identified by the NCEP SRRS analysis charts.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The 925-hPa streamline charts superimposed with total streamfunction tendency [ψt (925 hPa)] for specified dates. The propagation path of the concerned easterly disturbance–cold surge vortex–Indochina cyclone (red dashed line) and locations of these disturbances on a specified date are superimposed on the streamline charts. H and L are surface high and low, respectively, identified by the NCEP SRRS analysis charts.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
It has been suggested that the HRF cyclone develops from a cold surge vortex in the Celebes Sea by an in-phase constructive interference between the three monsoon modes. To substantiate this mechanism, compare the flow patterns of the HRF cyclone in Fig. 7c and the streamline chart for (
The 925-hPa streamline charts 1200 UTC 29 Oct 2008: (a) wind anomalies of the 12–24- and 5-day modes combined, (b) wind anomalies of 30–60-day mode, (c) wind departures from climatology, and (d) sum of (a) and (b). All streamline charts are superimposed with the corresponding streamfunction tendencies. The red dashed line in (a)–(d) is the propagation path of the easterly disturbance, cold surge vortex, and tropical cyclone. The color scale for streamfunction tendencies in different modes is shown on the right.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The 925-hPa streamline charts 1200 UTC 29 Oct 2008: (a) wind anomalies of the 12–24- and 5-day modes combined, (b) wind anomalies of 30–60-day mode, (c) wind departures from climatology, and (d) sum of (a) and (b). All streamline charts are superimposed with the corresponding streamfunction tendencies. The red dashed line in (a)–(d) is the propagation path of the easterly disturbance, cold surge vortex, and tropical cyclone. The color scale for streamfunction tendencies in different modes is shown on the right.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The 925-hPa streamline charts 1200 UTC 29 Oct 2008: (a) wind anomalies of the 12–24- and 5-day modes combined, (b) wind anomalies of 30–60-day mode, (c) wind departures from climatology, and (d) sum of (a) and (b). All streamline charts are superimposed with the corresponding streamfunction tendencies. The red dashed line in (a)–(d) is the propagation path of the easterly disturbance, cold surge vortex, and tropical cyclone. The color scale for streamfunction tendencies in different modes is shown on the right.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
More evidence of the role played by the three monsoon modes in the development of the cold surge vortex into the HRF cyclone is shown by the streamline charts combined with the streamfunction tendency of the three modes 
5. Maintenance of the Hanoi heavy rainfall event
The major rainfall in the global tropics occurs over the Asian monsoon region, tropical Africa, Central America, and the ITCZ (e.g., Chen 1985). The (ψQ, P) chart of October–November 2008 over Asia and the Pacific (Fig. 9a) exhibits the rainfall center over the Asian monsoon region, ITCZ, and the North and South Pacific convergence zones, consistent with the long-term climatology of (ψQ, P). Water vapor is transported by the low-level circulation, as depicted by the 2-month-mean value of ψQ in Fig. 9a, but this does not shed much light on how the precipitation is maintained. According to Eq. (3), the maintenance of rainfall by the convergence of water vapor flux is shown by (χQ, QD, P) in Fig. 9b, averaged for October–November 2008. It was shown in Fig. 1b that the rainfall associated with the 2008 Hanoi HRF event increased rapidly on 30 October and reached its maximum on 31 October. Despite the maximum rainfall, the hydrological conditions on 31 October resemble those for 30 October. To understand the cause of the dramatic rainfall increase to reach the maximum rainfall, note the flow field presented in Fig. 4n, showing a strong well-organized southeast–northwest flow formed by the HRF cyclone and the Taiwan anticyclone on 30 October transporting warm moist air to Hanoi. This is confirmed by the ψQ structure (Fig. 9c) with a strong southeast–northwest water vapor flux (open red arrow) established by the cyclonic ψQ cell over Indochina and the anticyclonic ψQ cell east of Taiwan. Although the rainfall center was located at Hanoi downstream of the southeast–northwest water vapor flux, it is not immediately clear how this center is maintained. This concern is clarified by the (χQ, QD, P) distribution (Fig. 9d)—the southern and southeastern Asian monsoon region was covered by a planetary-scale (χQ, QD) convergent center during October–November (Fig. 9b). Four convergent centers of (χQ, QD) are embedded within the Asian monsoon χQ center, while one of them is located at Hanoi. Our next concern is the formation–maintenance of this rainfall center by the three monsoon modes.
The (a) (ψQ, P) and (b) (χQ, QD, P) for October–November 2008 mean and (c) (ψQ, P) and (d) (χQ, QD, P) on 30 Oct 2008. The TRMM precipitation (in blue) is superimposed in (a)–(d). The contour interval of ψQ and χQ is 107 kg s−1 shown in the top-right corner of each chart. Scales for P and QD amplitude are provided on the right side of (a)–(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The (a) (ψQ, P) and (b) (χQ, QD, P) for October–November 2008 mean and (c) (ψQ, P) and (d) (χQ, QD, P) on 30 Oct 2008. The TRMM precipitation (in blue) is superimposed in (a)–(d). The contour interval of ψQ and χQ is 107 kg s−1 shown in the top-right corner of each chart. Scales for P and QD amplitude are provided on the right side of (a)–(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The (a) (ψQ, P) and (b) (χQ, QD, P) for October–November 2008 mean and (c) (ψQ, P) and (d) (χQ, QD, P) on 30 Oct 2008. The TRMM precipitation (in blue) is superimposed in (a)–(d). The contour interval of ψQ and χQ is 107 kg s−1 shown in the top-right corner of each chart. Scales for P and QD amplitude are provided on the right side of (a)–(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Measurement of the Hanoi daily rainfall is drawn from two different rainfall data sources (WMO station 48820 and TRMM). The 2008 Hanoi HRF event reported by station observations can be complementary to the TRMM dataset and aid in verification. The time series for rainfalls and power spectra of these daily rainfall totals are shown in Fig. 10. The two different rainfall data sources exhibit similar signals consistent for the daily variations at Hanoi. The distinctive signals that emerge from the rainfall power spectra (the 30 ~ 60-, 12 ~ 24-, and 5-day mode) are the same as the spectra for the zonal winds in Fig. 6.
(a) The precipitation time series of the Hanoi station (blue strip) and TRMM (red solid line). (b) Power spectra of every time series in (a). The thin lines are the 99% confidence level for each dataset, while the gray solid line is the mean 99% confidence level. The frequency bands for 30–60-, 12–24-, and 5-day modes are marked by three light blue strips in (b).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The precipitation time series of the Hanoi station (blue strip) and TRMM (red solid line). (b) Power spectra of every time series in (a). The thin lines are the 99% confidence level for each dataset, while the gray solid line is the mean 99% confidence level. The frequency bands for 30–60-, 12–24-, and 5-day modes are marked by three light blue strips in (b).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
(a) The precipitation time series of the Hanoi station (blue strip) and TRMM (red solid line). (b) Power spectra of every time series in (a). The thin lines are the 99% confidence level for each dataset, while the gray solid line is the mean 99% confidence level. The frequency bands for 30–60-, 12–24-, and 5-day modes are marked by three light blue strips in (b).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
To gain a perspective of how the Hanoi HRF event is maintained by these monsoon modes, the x–t diagrams of precipitation P and its three mode components, 
The x–t diagrams of (a) total precipitation P, (b) 
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The x–t diagrams of (a) total precipitation P, (b)
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The x–t diagrams of (a) total precipitation P, (b)
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The convergence of water vapor flux from the three monsoon modes that maintained the Hanoi HRF event is shown in Fig. 12. The 30–60-day mode, 
Potential function of water vapor flux superimposed with precipitation on 30 Oct 2008: (a) 30–60-, (b) 12–24-, (c) 5-day modes, (d) sum of (a)–(c), and (e) daily departure of (χQ, QD, P) from the October–November 2008 mean (χQ, QD, P) shown in Fig. 9b. The contour intervals of potential function of water vapor flux in (a)–(c) and (d),(e) are 2 × 106 kg s−1 and 5 × 106 kg s−1, respectively (bottom-right corner of each chart). Precipitation is scaled by color and shown at the top right of (a),(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Potential function of water vapor flux superimposed with precipitation on 30 Oct 2008: (a) 30–60-, (b) 12–24-, (c) 5-day modes, (d) sum of (a)–(c), and (e) daily departure of (χQ, QD, P) from the October–November 2008 mean (χQ, QD, P) shown in Fig. 9b. The contour intervals of potential function of water vapor flux in (a)–(c) and (d),(e) are 2 × 106 kg s−1 and 5 × 106 kg s−1, respectively (bottom-right corner of each chart). Precipitation is scaled by color and shown at the top right of (a),(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Potential function of water vapor flux superimposed with precipitation on 30 Oct 2008: (a) 30–60-, (b) 12–24-, (c) 5-day modes, (d) sum of (a)–(c), and (e) daily departure of (χQ, QD, P) from the October–November 2008 mean (χQ, QD, P) shown in Fig. 9b. The contour intervals of potential function of water vapor flux in (a)–(c) and (d),(e) are 2 × 106 kg s−1 and 5 × 106 kg s−1, respectively (bottom-right corner of each chart). Precipitation is scaled by color and shown at the top right of (a),(d).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
6. Discussion
Most locations of the heavy rainfall–flood events in Vietnam during October–November appeared over central Vietnam south of 19°N (Fig. 13) in the past three decades. The 2008 Hanoi case occurred exceptionally farther north than other years because of the unusual track of the cold surge vortex–HRF cyclone (Fig. 4), which was steered northwestward from the Celebes Sea to northern Vietnam. Was there any interannual change in the atmospheric circulation over eastern and southeastern Asia during October–November 2008, enabling the cold surge vortex to form in the Celebes Sea and move northwestward, instead of westward, as most cold surge vortices related to the central Vietnam HRF events occur?
Locations of Vietnam flood events during October–December over the past 30 years archived by the Dartmouth Flood Observatory. The location of the 2008 Hanoi flood is marked by a red cross, while the other events are marked by blue dots.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Locations of Vietnam flood events during October–December over the past 30 years archived by the Dartmouth Flood Observatory. The location of the 2008 Hanoi flood is marked by a red cross, while the other events are marked by blue dots.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Locations of Vietnam flood events during October–December over the past 30 years archived by the Dartmouth Flood Observatory. The location of the 2008 Hanoi flood is marked by a red cross, while the other events are marked by blue dots.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Because rainfall is maintained by convergence of water vapor flux, excessive rainfall at Hanoi should be reflected by the distribution of Δ(χQ, QD, P) departures during October–November 2008 (Fig. 14a) from their climatological mean values. A large-scale positive Δ(χQ, QD) center appeared over tropical southern and southeastern Asia, coincident with the (χQ, QD,) center over the Asian monsoon region (Fig. 9b). Two negative Δ(χQ, QD, P) centers overlap with dry regions between the North and South Pacific convergence zones, and the ITCZ. A strong east–west differentiation of ΔP anomalies existed north of the equator between the positive ΔP centers over the Asian monsoon region and the negative ΔP centers in the western tropical Pacific. South of the equator, a weak east–west differentiation of ΔP anomalies appeared between the Asian monsoon region and the western Indian Ocean. In addition to these east–west differentiations of ΔP anomalies, there are also distinct north–south differentiations of ΔP between positive ΔP anomalies along the ITCZ and either the North or South Pacific convergence zone. As indicated by ΔQD, water vapor is divergent from the negative ΔP anomaly areas and convergent toward the positive ΔP anomaly areas. A response of the atmospheric circulation to the diabatic heating generated by ΔP is the divergent circulation with its lower-tropospheric pattern resembling ΔQD; it includes the east–west circulations north and south of the equator and the local Hadley circulation.
Departures of (a)Δ(χQ, QD, P), (b)Δ(ψQ, P), and (c) Δ[ψ(850 hPa), SST] from their corresponding variables averaged over October–November for the period 1979–2009. Contour intervals of ΔχQ, ΔψQ, and Δψ(850 hPa) are shown in the top-right corner of (a)–(c), while ΔP and ΔSST are scaled by colors shown in the bottom-right corner of (a)–(c).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Departures of (a)Δ(χQ, QD, P), (b)Δ(ψQ, P), and (c) Δ[ψ(850 hPa), SST] from their corresponding variables averaged over October–November for the period 1979–2009. Contour intervals of ΔχQ, ΔψQ, and Δψ(850 hPa) are shown in the top-right corner of (a)–(c), while ΔP and ΔSST are scaled by colors shown in the bottom-right corner of (a)–(c).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
Departures of (a)Δ(χQ, QD, P), (b)Δ(ψQ, P), and (c) Δ[ψ(850 hPa), SST] from their corresponding variables averaged over October–November for the period 1979–2009. Contour intervals of ΔχQ, ΔψQ, and Δψ(850 hPa) are shown in the top-right corner of (a)–(c), while ΔP and ΔSST are scaled by colors shown in the bottom-right corner of (a)–(c).
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
As illustrated in appendix D, the large-scale circulation in the tropics is well depicted by the Sverdrup relationship [i.e., a balance between the meridional advection of planetary vorticity (i.e., β term) and vortex stretching in the vorticity equation; Chen 2005, 2010]. A significant diabatic heating in the tropics is produced by the release of latent heat through convection and rainfall. As inferred from the ΔP distribution east and west of 120°E, the meridional flow of the North Pacific anticyclone can be modulated by the east–west differential heating generated by the ΔP anomalies between the Asian monsoon region and the western Pacific–the western Indian Ocean. The local Hadley circulation over the western Pacific can also be modulated by the meridional differential heating, generated by the ΔP anomalies between the North–South Pacific convergence zones, the ITCZ, and the dry region between these convergence zones.
Well accepted by the meteorological community, global circulation may be changed by tropical SST variations, such as those shown in Fig. 14c. This circulation change may be illustrated from a perspective of the hydrological cycle. In response to the diabatic heating generated by ΔP anomalies through vortex stretching, an interaction between the divergent and rotational circulation, the Δψ(850 hPa) anomalies (Fig. 14c) exhibit a four-leaf structure radiating from the Asia monsoon ΔP center. The ΔψQ anomalies (Fig. 14b) exhibit the same structure. In response to Δ(χQ, QD, P) (Fig. 14a), the change of the Asian monsoon circulation is reflected by the Δψ (850 hPa) anomalies, which supply water vapor to maintain the Hanoi HRF event, as indicated by the ΔψQ anomalies.
The atmospheric circulation change in response to diabatic heating generated by ΔP can extend above 500 hPa. The anomalous southeast–northwest flow across the South China Sea indicated by the Δψ structure (Fig. 14c) is not only to supply water vapor and maintain the Hanoi HRF event, but also to steer the cold surge vortex anomalously northwestward to northern Vietnam.
7. Conclusions
The synoptic development mechanism and forecast of HRF events during late fall in Vietnam have been long-standing issues in the regional weather–climate system. Occurrence of the unusual 2008 Hanoi event offered an excellent opportunity to explore the aforementioned issue. It was observed by Cheang (1977) that heavy winter rainfall events in tropical Southeast Asia were produced by cold surge vortices evolved from the interactions between easterly disturbances and cold surge flows. Although these vortices are generated by a local mechanism, cold surge flows originate from the midlatitudes. Thus, the development and formation of the Hanoi HRF event should be explored from a large-scale perspective in the context of the midlatitude–tropical interaction. The major findings of this effort can be summarized as follows:
The upper short-wave trough associated with the severe cold surge in East Asia can be backtracked to the western North Atlantic (Joung and Hitchman 1982). The time needed for this disturbance to reach the East Asian coast is about 6 ~ 7 days. The surface high–low couplet of the cold surge related to the cold surge vortex (which evolved into the 2008 Hanoi HRF event) was linked to an upper trough–ridge system and could backtrack its genesis location in the valley between Tianshan and the Altai Mountains. The time for this couplet and the associated cold surge flow to reach the East Asian coast is about 10 days, longer than Joung and Hitchman’s observation. An explosive cyclone, which reached the intensification rate of a “bomb” (Sanders and Gyakum 1980), was also formed downstream from the surface low associated with the cold surge simultaneously, with the peak of the Hanoi HRF event being reached on 31 October 2008.
The synoptic analysis revealed that the cold surge vortex, which eventually developed into the HRF cyclone, formed when an easterly disturbance propagated through the Celebes Sea. This disturbance merged with a surface vortex originating within an east–west-oriented tropical cyclonic shear flow established by the East Asian cold surge flow and the tropical westerlies from the Indian Ocean.
Three monsoon modes are distinct from the spectral analysis of the zonal winds north and south of the HRF cyclone. The easterly disturbance belongs to the 5-day mode, while the cold surge flow exhibits a strong 12–24 day signal. The cold surge vortex developed in concert with these two monsoon modes. The 30–60-day monsoon mode, a regional response to the global MJO, develops a cyclonic shear flow in the tropical southern and southeastern Asia, which provides an environment favorable for the development of this vortex into the HRF cyclone. This cyclone peaked in intensity when a constructive in-phase interference of the three monsoon modes occurred.
The x–t diagrams of rainfall anomalies for three monsoon modes showed the 2008 Hanoi heavy rainfall on 30–31 October occurred when the rainfall anomalies for these three monsoon modes had an in-phase constructive interference. A northwestward water vapor flux, as depicted by the streamfunctions for the three monsoon modes, was well organized to transport moisture from the South China Sea to northern Vietnam. The divergent circulations of the three monsoon modes helped to converge the moisture flux, which helped maintain the heavy rainfall in Hanoi.
Over the past three decades, only one of the 31 heavy rainfall–flood events during October–December in Vietnam occurred far north in Hanoi. In response to the anomalous warm sea surface temperature anomalies over the western tropical Pacific in October–November 2008, an anomalous anticyclonic circulation appeared over this region. The southeasterly flow associated with this anticyclone across the northern part of the South China Sea not only facilitated the northwestward propagation of the HRF cyclone, but also enhanced the northwestward water vapor transport into Hanoi to maintain the 2008 HRF event.
Diagnostic analyses performed in this study indicate the formation mechanism of the cold surge vortex leading to the occurrence of the 2008 Hanoi heavy rainfall flood event includes factors beyond the interaction between the easterly disturbance and the cold surge flow. The development of a large-scale environment favorable for the cold surge vortex formation and the ensuing intensification of this vortex by the intraseasonal monsoon mode have not been considered. The supply mechanism of water vapor flux to maintain the heavy rainfall event was also neglected. New findings highlighted previously may enable us to formulate some parameters useful in operational forecasts for the Vietnam heavy rainfall–flood events in late fall. These include the following:
The cold surge flow developed from the newly formed surface high–low couplet in the vicinity of the Altai Mountains can be observed and monitored several days ahead of the occurrence of the cold surge on the eastern seaboard of eastern and northeastern Asia. Accompanying this cold surge, the formation of an explosive cyclone from the surface low associated with the cold surge synchronized with the occurrence of the Vietnam HRF event. Using the NCEP GFS forecast model, it is feasible to predict several days ahead for Vietnam flood events. Thus, tracking the upper short-wave trough–ridge system and the coupled surface high–low system can provide an early warning.
The 2008 Hanoi event occurred when the low-level convergent center–the upper-level divergent center of the global MJO was located in the Asian monsoon region coincident with a constructive interference of rainfall anomalies of the three monsoon modes at Hanoi. These factors can be used as excellent monitoring tools for the occurrence of the Vietnam HRF events.
The interannual variation of the western Pacific subtropical anticyclone in response to warm SST anomalies in the western tropical Pacific is an informative factor for an operational center to predict the potential latitudinal location of the Vietnam HRF events.
In view of the potential application of these parameters presented for operational forecasts of the Vietnam HRF events, a companion study is currently ongoing: 1) to explore whether new findings of this study are common to those events which occurred in the past three decades and 2) to develop some simple forecast guidelines using these parameters for operational forecasts. Results from this effort will be reported in the future.
Acknowledgments
This study is part of the East Asian Monsoon Experiment (EAMEX) in collaboration with the Monsoon Asian Hydro-Atmosphere Scientific Research and Prediction Initiative (MAHASRI) and the National Hydro-Meteorological Services of Vietnam, and supported in part by the Cheney Research Fund and NSF Grant ATM-0836220. Ming-Cheng Yen’s effort is supported by NSC Grant NSC99-2111-M-008-012. We thank professor Jun Matsumoto of Tokyo Metropolitan University for motivating us to explore the cause of the Vietnam heavy rainfall–flood events during late fall. Comments and suggestions offered by three reviewers were very helpful in clarifying some basic issues presented in this study.
APPENDIX A
Definition of the Easterly Disturbance, Cold Surge Vortex, and HRF Cyclone
The easterly disturbance, the cold surge vortex, and the HRF cyclone are defined by the maximum 850-hPa speed of their cyclonic circulations, as marked in Fig. A1:
easterly disturbance: |V(850 hPa)|max ≲ 5 m s−1;
cold surge vortex: 5 m s−1 ≲ |V(850 hPa)|max < 8 m s−1;
cyclone: 8 m s−1 ≲ |V(850 hPa)|max.
The time series for the maximum isotach |V(850 hPa)| of easterly disturbance, cold surge vortex, and HRF cyclone leading to the development of the 2008 Hanoi HRF event.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for the maximum isotach |V(850 hPa)| of easterly disturbance, cold surge vortex, and HRF cyclone leading to the development of the 2008 Hanoi HRF event.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for the maximum isotach |V(850 hPa)| of easterly disturbance, cold surge vortex, and HRF cyclone leading to the development of the 2008 Hanoi HRF event.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
APPENDIX B
Indication of the Northeastern Asian Cold Surge at Seoul, South Korea, during Late October 2008
As shown by the 850-hPa streamline chart on 30 October 2008 in Fig. 1d, the northeastern Asian cold surge related to the 2008 Hanoi HRF event moved across the Korean Peninsula in late October 2008. In contrast with this figure, it is revealed from the y–t diagram for 925-hPa wind vectors with red dots along the eastern coast of Asia (Fig. 1a) that a wind surge appeared at 35°N on 28 October 2008. The contrast between thickness (500–1000 hPa, red dashed line) in Fig. 2i and the 925-hPa streamline chart in Fig. 4n shows a strong cold-air advection across South Korea. Associated with this cold-air advection, the wind surge shown in Fig. 1a is apparently a cold surge. To further confirm our inference, the time series for pS and TS at Seoul, South Korea, during late October 2008 is shown in Fig. B1. The first cold surge initiated on 23 October 2008 did not reach Hanoi (not shown), but the second cold surge did reach Hanoi.
The time series for surface pressure ps and temperature TS at Seoul, South Korea, during late October 2008.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for surface pressure ps and temperature TS at Seoul, South Korea, during late October 2008.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for surface pressure ps and temperature TS at Seoul, South Korea, during late October 2008.
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
APPENDIX C
Illustration of the Difference between the Time Series for u(850 hPa) and 
The difference between the time series for u(850 hPa) and 
The time series for u(thick black line), 
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for u(thick black line),
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
The time series for u(thick black line),
Citation: Monthly Weather Review 140, 4; 10.1175/MWR-D-11-00111.1
APPENDIX D
Circulation Change in Response to Differential Heating
REFERENCES
Cheang, B. K., 1977: Synoptic feature and structure of some equatorial vortices over the South China Sea in the Malaysian region during the winter monsoon of December 1973. Pure Appl. Geophys., 115, 1303–1333.
Cheang, B. K., 1987: Short and long-range monsoon prediction in Southeast Asia. Monsoons, J. S. Fein and P. L. Stephens, Eds., Wiley-Interscience, 579–606.
Chen, S.-J., Y.-H. Kuo, P.-Z. Zhang, and Q.-F. Bai, 1992: Climatology of explosive cyclones off the East Asian coast. Mon. Wea. Rev., 120, 3029–3035.
Chen, T.-C., 1985: Global water vapor flux and maintenance during FGGE. Mon. Wea. Rev., 113, 1801–1819.
Chen, T.-C., 2002: A North Pacific short-wave train during the extreme phases of ENSO. J. Climate, 15, 2359–2376.
Chen, T.-C., 2005: The structure and maintenance of stationary waves in the winter Northern Hemisphere. J. Atmos. Sci., 62, 3637–3660.
Chen, T.-C., 2007: East Asian Monsoon Field Experiment (EAMEX): Participation of the MAHASRI (post-GAME) International Field Experiment. 50 pp. [Available online at http://eamex.iastate.edu/Download/eamex_science_plan.pdf.]
Chen, T.-C., 2010: Characteristics of summer stationary waves in the Northern Hemisphere. J. Climate, 23, 4489–4507.
Chen, T.-C., and W. E. Baker, 1986: Global diabatic heating during FGGE SOP-1 and SOP-2. Mon. Wea. Rev., 114, 2578–2589.
Chen, T.-C., and S.-P. Weng, 1996: Some effects of the intraseasonal oscillation on the equatorial waves over the western tropical Pacific–South China Sea region during the northern summer. Mon. Wea. Rev., 124, 751–756.
Chen, T.-C., S.-Y. Wang, W.-R. Huang, and M.-C. Yen, 2004: Variation of the East Asian summer monsoon rainfall. J. Climate, 17, 744–762.
Chen, T.-C., J.-D. Tsay, M.-C. Yen, and J. Matsumoto, 2012: Interannual variation of the late fall rainfall in central Vietnam. J. Climate, 25, 392–413.
Compo, G. P., G. N. Kiladis, and P. J. Webster, 1999: The horizontal and vertical structure of East Asian winter monsoon pressure surges. Quart. J. Roy. Meteor. Soc., 125, 29–54.
DFO, cited 2010: Global active archive of large flood events. Dartmouth Flood Observatory. [Available online at http://floodobservatory.colorado.edu/Archives/index.html.]
EAMEX, cited 2010: East Asian Monsoon Experiment. [Available online at http://eamex.iastate.edu/synoptic_development_of_08_Hanoi_HRF_event.pdf.]
JMA, cited 2008: Weather maps. Japan Meteorological Agency. [Available online at http://www.jmbsc.or.jp/english/index-e.html.]
Joung, C. H., and M. H. Hitchman, 1982: On the role of successive downstream development in East Asian polar air outbreaks. Mon. Wea. Rev., 110, 1224–1237.
Kanamitsu, M., and Coauthors, 1991: Recent changes implemented into the global forecast system at NMC. Wea. Forecasting, 6, 425–435.
Lau, K.-M., and P. H. Chan, 1985: Aspects of the 40–50 day oscillation during the northern winter as inferred from outgoing longwave radiation. Mon. Wea. Rev., 113, 1889–1909.
Madden, R. A., and P. R. Julian, 1971: Detection of a 40–50 days oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28, 702–708.
Madden, R. A., and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29, 1109–1123.
MAHASRI, 2006: Monsoon Asian Hydro-Atmosphere Scientific Research and Prediction Initiative, version 4.1. International drafting committee of the post-GAME planning working group, science plan proposal, 58 pp. [Available online at http://www.gewex.org/PAN-GEWEX-MTG/MAHASRI_SciencePlan_v4.1.pdf.]
McBride, J. L., 1995: Tropical cyclone formation. Global Perspectives on Tropical Cyclones, Tech. Doc. WMO/TD-693, R. L. Elsberry, Ed., World Meteorological Organization, 63--105.
Mitchell, J. M., B. Dzerdzeevskii, H. Flohn, W. L. Hofmeyr, H. H. Lamb, K. N. Rao, and C. C. Walléen, 1966: Climate change. WMO Tech. Note 79, World Meteorological Organization, Geneva, Switzerland, 79 pp.
Murakami, M., 1976: Analysis of summer monsoon fluctuations over India. J. Meteor. Soc. Japan, 54, 15–31.
Murakami, M., 1979: Large-scale aspects of deep convective activity over the GATE area. Mon. Wea. Rev., 107, 994–1013.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65, 373–390.
Palmén, E., and C. W. Newton, 1969: Atmospheric Circulation Systems. International Geophysics Series, Vol. 13, Academic Press, 603 pp.
Petterssen, S., 1956: Motion and Motion Systems. Vol. 2, Weather Analysis and Forecasting, McGraw–Hill, 428 pp.
Ramage, C. S., 1952: Variation of rainfall over south China through the wet season. Bull. Amer. Meteor. Soc., 33, 308–311.
Rutledge, G. K., J. Alpert, and W. Ebisuzaki, 2006: NOMADS: A climate and weather model archive at the National Oceanic and Atmospheric Administration. Bull. Amer. Meteor. Soc., 87, 327–341.
Sanders, F., and J. R. Gyakum, 1980: Synoptic–dynamic climatology of the “bomb.” Mon. Wea. Rev., 108, 1589–1606.
Simpson, J., C. Kummerow, W. K. Tao, and R. F. Adler, 1996: On the tropical rainfall measuring mission (TRMM). Meteor. Atmos. Phys., 60, 19–36.
Son, T., 2008: Vietnam aims for quick full recovery from historic floods. Thanh Nien News, 15 November 2008. [Available online at http://www.thanhniennews.com/2008/Pages/20081115134845043760.aspx.]
TMD, cited 2008: Weather charts. Thai Meteorological Department. [Available online at http://www.tmd.go.th/en/weather_map.php.]
Truong, N. M., T. T. Tien, R. A. Pielke, C. L. Castro, and G. Leoncini, 2009: A modified Kain–Fritsch scheme and its application for the simulation of an extreme precipitation event in Vietnam. Mon. Wea. Rev., 137, 766–789.
Tsay, J. D., 2004: Water vapor budget of cold surge vortices. M. S. thesis, Iowa State University, 120 pp. [Available online at http://eamex.iastate.edu/Download/MS_vortex.pdf.]
Yanai, M., T. Maruyama, T. Nitta, and Y. Hayashi, 1968: Power spectra of large-scale disturbances over the tropical Pacific. J. Meteor. Soc. Japan, 46, 308–323.
Yang, F., H. L. Pan, S. K. Krueger, S. Moorthi, and S. J. Lord, 2006: Evaluation of the NCEP Global Forecast System at the ARM SGP Site. Mon. Wea. Rev., 134, 3668–3690.
Yokoi, S., and J. Matsumoto, 2008: Collaborative effects of cold surge and tropical depression-type disturbance on heavy rainfall in central Vietnam. Mon. Wea. Rev., 136, 3275–3287.
The cold surge vortex is a vortex, which reaches a speed of 8 m s−1 shown in Fig. A1 (appendix A), formed either by the direct interaction of the eastern Asian cold surge flow with a land–mountain or by the interaction of an easterly disturbance with the cold surge flow.
The temporal evolution of the northeastern Asian cold surge during late October 2008 is illustrated in terms of the time series of surface pressure and temperature at Seoul, Korea, in appendix B.
The easterly disturbance, cold surge vortex, and HRF cyclone are defined by their wind speed in Fig. A1 of appendix A.
A discrepancy between the time series for u(850 hPa) and 
