1. Introduction
During early summer, the East Asian monsoon is characterized by frequently and repeatedly occurring rainbands (Ding and Chan 2005; Xu et al. 2009). These rainbands are associated with a quasi-stationary front called the “mei-yu front” (Chen 1983; Ding 1992). The mei-yu front usually regulates a widespread precipitation rain belt over East Asia. However, heavy rainfall and rainfall maxima are often produced by mesoscale convective systems (MCSs) embedded in the mei-yu frontal cloud band (Kuo and Chen 1990; Zhang et al. 1990; Ninomiya 2000; Zhang et al. 2003). Heavy-rain-producing mesoscale or synoptic-scale disturbances along the mei-yu front may move eastward or stay stationary showing a wavelike distribution (Chen et al. 2008). Heavy rainfall is also associated with a low-level vortex developing on the east flank of the Tibetan Plateau and moving into or becoming a cyclone center on the mei-yu front (Tao and Ding 1981; Ninomiya and Akiyama 1992; Chang et al. 1998). The presence of a southerly or southwesterly low-level jet (LLJ) to the south of the disturbances significantly helps the development of heavy-rain-bearing systems (Chen and Yu 1988; G. T.-J. Chen et al. 2005, 2006). Locations where the warm and moist southwesterly flow directly and persistently interacts with cold air intrusion from the north are also favorable for heavy precipitation (Ding 1992). These and other synoptic settings favorable for the development of heavy rainfall during the monsoon season have been examined and summarized by many authors (Kuo and Chen 1990; Chen and Li 1995; C.-S. Chen et al. 2005).
Since total precipitation at any point is directly proportional to the rate and duration of rainfall, quasi-stationary moving MCSs are favorable for producing extreme rainfall over a region. MCSs, especially slowly moving or quasi-stationary MCSs, are major causes of extreme rainfall in the warm season—for example, in the United States (Maddox et al. 1979; Houze et al. 1990; Doswell et al. 1996; Schumacher and Johnson 2005, 2006; Moore et al. 2003). All these studies agree with the importance of a LLJ to those quasi-stationary heavily precipitating MCSs through persistent advection of warm moist air into the system. Another major component is the continuously redeveloping deep convection moving over the same region (Chappell 1986; Doswell et al. 1996; Davis 2001) triggered by surface fronts, outflow boundaries, mesoscale convective vortices (MCVs), or orographic lifting (Sanders 2000; Raymond and Jiang 1990; Pontrelli et al. 1999; Schumacher and Johnson 2008, 2009). “Training line” and “back-building” processes are two major patterns (Schumacher and Johnson 2005, 2006). The training line process occurs when a series of convective cells move in a line parallel to the storm direction, but propagate very little in the perpendicular direction. On the other hand, the back-building process is a situation when convective cells repeatedly develop upstream of previous deep convection and pass over the same region.
The complex terrain over southern China and Taiwan can help to produce heavy rainfall and flash flooding through the interaction between prevailing monsoon flow and orography-generated local circulations (Li et al. 1997; Yeh and Chen 2002; Li and Chen 1998; C.-S. Chen et al. 2005). Under the presence of a similar southwesterly or quasi-stationary subtropical front, specific locations of heavy precipitation may vary from time to time (Li et al. 1997; C.-S. Chen et al. 2005; Chen et al. 2007). Heavy rainfall occurs both over mountain, slopes, and coastal regions during the passage of mei-yu fronts (Chen et al. 2007). Mechanisms for orography-enhanced heavy precipitation can be summarized as follows: 1) lifting by the terrain, through forced ascent as the air moves up the slope, or more indirect, as the airflow is blocked by the mountain, or when mountains act as a high-level heat source (Chen et al. 1991; Chen and Feng 2001; Akaeda et al. 1995); 2) convergence between the prevailing monsoon flow and offshore flow during the night (Chen and Li 1995; Yeh and Chen 2002; C.-S. Chen et al. 2005); 3) convergence between monsoon southwesterly flow and a low-level barrier jet or between postfrontal northwesterly flow and a barrier jet (Chen and Li 1995; Li et al. 1997; Yeh and Chen 2002); and 4) stagnation of MCSs or squall lines blocked by terrain, leading to extreme accumulation of precipitation (Reeves and Lin 2007).
Theoretically, for a nonrotating conditionally unstable monsoon flow impinging on a mesoscale mountain ridge, the low-level flow may cross the mountain or be blocked and split depending on the Froude number (Chu and Lin 2000; Chen and Lin 2005a,b). For a large Froude number (Fr > 1) (Fr = U/Nh, where U is the zonal wind component normal to the barrier, h is the mesoscale mountain height, and N is the Brunt–Vaisala frequency), the flow will cross the mountain. For a small Froude number (Fr < 1), the low-level flow will be blocked and redirected around the mountain barrier (Sun et al. 1991; Chen and Li 1995; Li and Chen 1998). The northward-redirected southwesterly monsoon flows impinging on the high mountains of Taiwan is an example. During the Taiwan Area Mesoscale Experiment (TAMEX), southwesterly flows are frequently blocked and redirected by the Central Mountain Range in conditions of small Froude number (Chen and Li 1995; Li and Chen 1998). The northern branch of the splitting airflow could accelerate down the orographically induced pressure gradient and become a low-level barrier jet.
During the Terrain-influenced Monsoon Rainfall Experiment (TiMREX), one long-duration heavily precipitating mesoscale system develops offshore and drops about a half meter of rain on the coastal cities of southwest Taiwan. Inside this heavily precipitating system, new convection keeps developing offshore and feeding a large precipitation shield. In this paper we aim to answer this question: What is the initiation and maintenance mechanism for this extreme rainfall event? We note that for this coastal rainfall event, neither a nocturnal downslope–offshore flow nor an island barrier jet was observed to be responsible for the enhancement or maintenance. Our working hypothesis is that previous precipitation before the development of the long-duration system forms a cold pool over the island and adjacent ocean, and that boundary between the cold pool and upstream warm and moist unstable air focuses the ascent and initiates convection well offshore. This paper is motivated to examine the proposed hypothesis of this extreme rainfall event using detailed data from TiMREX. The experiment design of TiMREX, data, and analysis methods are described in section 2. Overviews of synoptic conditions, storm environments, and upstream conditions are presented in section 3, while storm initiation, evolution, and convective structures are analyzed in section 4 based on Doppler radar measurements. Finally, the possible initiation and maintenance mechanisms are summarized using all supporting evidence in section 5.
2. Data and methodology
TiMREX is a joint U.S.–Taiwan multiagency field experiment (Lee et al. 2009) providing high-resolution measurements of monsoon precipitation systems over southwest Taiwan and the upstream ocean in May and June of 2008. The goal of TiMREX is to improve the understanding and parameterization of the physical process of the heavy precipitation systems and associated environments within the context of active monsoon flow. Details of the experiment’s design, motivation, and observations are summarized in Lee et al. (2009). Locations of major facilities are shown in Fig. 1a.
(a) Map of the distribution of major facilities deployed during the TiMREX field phase. All the facilities are consistent with markers in the legend. Between 1200 UTC 14–16 Jun research ship 99810 was within the region outlined by the blue triangle. Solid green lines are locations of RHI scans of S-Pol shown below in Fig. 10. Dashed box is the region for calculating time series of rainfall average. (b) Time series of hourly rain gauge recorded rainfall averaged over southwest Taiwan [dashed box in (a)].
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
a. Large-scale reanalysis and rainfall data
The National Centers for Environmental Prediction (NCEP) Climate Forecast System Reanalysis (CFSR) data (Saha et al. 2010) are used to examine the large-scale flow and synoptic conditions. The CFSR is a global, high-resolution coupled atmosphere–ocean–land surface–sea ice reanalysis system. The CFSR atmospheric reanalysis has a horizontal spatial resolution of 0.5° × 0.5° and 6-hourly temporal resolution (at 0000, 0600, 1200, and 1800 UTC). The CFSR dataset used in this study has 37 vertical levels ranging from the surface to 1 hPa. Most of the levels are at 25-hPa intervals. The CFSR data are used to analyze fields of potential temperature and winds. Diagnostics of frontogenesis, differential vorticity advection, and thermal advection are also conducted using CFSR. The 700-hPa thermal advection is determined by the advection of mean temperature determined by the thickness between 750 and 650 hPa.
The Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) 3B42 rain product (Huffman et al. 2007) is used for an overview of the daily large-scale rainfall distribution pattern associated with the mei-yu front. The TMPA 3B42 dataset has 3-hourly temporal resolution and 0.25° × 0.25° spatial resolution. This rainfall algorithm uses passive microwave measurements from low-Earth-orbit satellites and infrared radiance measurements from geostationary satellites, as well as land surface precipitation rain gauge analyses when available.
b. Surface and upper-air network
During TiMREX, the operational surface network consists of 25 conventional weather stations, 418 automatic rain gauges, 57 GPS integrated water vapor sensors, and a lightning location system (LLS). The hourly rain data from the automatic rain gauges are used in this study.
TiMREX includes frequent island, ship, and aircraft sondes to collect information on storm environment and upstream conditions (Fig. 1a). During the subject heavy precipitation event during intensive observing period 8 (IOP8) (14–16 June), the soundings are launched every 3 h from Taiwan Island and every 6 h from the upstream small islands and research ship. Note that the research ship (99810 in Fig. 1a) is not stationary, but from 1200 UTC 14–16 June is within the blue triangle; exact locations are given in Davis and Lee (2012). Four different sonde types (Vaisala RS80, Vaisala RS92, Meisei, and Graw) are used. Quality control procedures on the sonde data are conducted to correct for a tropospheric deep dry bias and for contamination of thermodynamic variables in the lower troposphere by the ship itself (Ciesielski et al. 2010). The corrected sonde data show good agreement with independent estimates.
c. Radar network and analysis
1) Radar data and quality control
Doppler radars are deployed at the southwest coast of Taiwan during TiMREX (Fig. 1a). Measurements from one of the operational radars—radar code of Chi-Gu (RCCG) and the research radar [National Center for Atmospheric Research (NCAR)’s S-band polarimetric (S-Pol) radar] are used in this study. Both of these S-band radars scan simultaneously a full volume in about 7.5 min. The S-Pol radar is strategically placed ~70 km from the RCCG radar to form a dual-Doppler radar pair. The S-Pol also collects polarimetric data that is used to remove the ground clutter and other signal noise. The NCAR S-Pol radar data quality control and calibration process is executed by the NCAR TiMREX data archive office (http://sowmex.cwb.gov.tw).
The RCCG radar is operated and calibrated by the Taiwan Central Weather Bureau and widely used in publications (Chang et al. 2009; Zhang et al. 2009; Liou and Chang 2009; Tsai 2010). In this study, ground-clutter-contaminated data and second trip echoes are discarded through the removal of stationary echoes and echoes below 1-km altitude, while aliased Doppler velocities are unfolded. After the quality control, radar volume data from both RCCG and S-Pol are interpolated from the radar spherical coordinates onto 1.0-km horizontal resolution and 1.0-km vertical resolution Cartesian coordinates using the NCAR REORDER software package (Oye and Case 1995). A closest point weighting function with effective radius of 1.5 km is used in the interpolation process. The output Cartesian-based radar volume covers 400 km × 400 km horizontally and 20 km vertically every 7.5 min.
2) Definition of radar precipitation features (RPFs)
Based on the concept of TRMM precipitation features (Nesbitt et al. 2000), contiguous raining pixels at the altitude of 2 km of a radar volume scan from the RCCG radar are defined as radar precipitation features (RPFs). RPFs smaller than 50 km2 are removed from the dataset. The size of an RPF is defined by the total area of raining pixels at 2 km. Convective proxies are defined for the RPF, including maximum height of specific radar reflectivities (e.g., maximum height of the 20-dBZ echo) and convective core size (area of pixels >40 dBZ). Large, medium, and small convective clusters are also defined. A large convective cluster is defined as an RPF with convective core larger than 400 km2, while a small convective cluster is defined as an RPF with convective core smaller than 100 km2. Accordingly, a medium convective cluster is in the size range between large and small clusters. The location of an RPF or convective cluster is determined by the center of the longest axis of the feature.
3) Dual-Doppler wind analysis
Dual-Doppler wind analysis is conducted between S-Pol and RCCG (Fig. 1a). The wind synthesis and subsequent computations in the dual-Doppler analysis are performed in the NCAR Custom Editing and Display of Reduced Information in Cartesian Space (CEDRIC; Mohr et al. 1986) system. Before processing the dual-Doppler analysis, radar volumes collected at nearly the same time (at most 2 min apart) from both radars are interpolated into the same-resolution Cartesian coordinate with S-Pol as the original point. An average storm movement vector is calculated for each event to perform a differential advection during the dual-Doppler synthesis. The storm motions are determined by tracking the low-level radar reflectivity of the convective lines.
The horizontal wind components (U and V) are calculated directly without setting any boundary conditions. The vertical velocity is derived from integrating the mass continuity equation and considering the hydrometeor fall speed. The mass continuity equation is integrated upward from the lowest grid height of 1.0 km with 1.0-km vertical resolution. The retrieved wind fields from dual-Doppler analysis contain both random and nonrandom errors. Errors can occur because of incorrect storm advection assumptions, boundary conditions, poorly sampled low-level and upper-level divergence, and temporal sampling difference between the two radars (Doviak et al. 1976; Wilson et al. 1984).
3. Weather conditions and storm environments
During the period of 14–16 June, a quasi-stationary mei-yu wind shear line approaches and stays over the southeast China coast and the Taiwan Strait. Associated with the mei-yu shear line, robust convection and heavy precipitation systems develop over and southwest of Taiwan on 14 June. Rainy convective systems pass southwest Taiwan by 15 June (stratiform precipitation and heavy clouds are dominating during daylight of 15 June), but develop again on the evening of 15 June and evolve further on 16 June (Fig. 1b). Though rainy systems on 14 and 16 June both have a long duration, their storm type and evolution is quite different. Storms on 14 June evolve more like squall lines and propagate from northwest to southeast, while raining systems on 16 June remain nearly stationary. This section examines the rainfall distribution, synoptic conditions, and storm environments for the extreme rainfall event of 16 June.
a. Rainfall distributions
The quasi-stationary raining system of 16 June drops the heaviest rain over the upstream ocean and the coastal cities of southwest Taiwan (Figs. 2–3). Rainfall maxima over the southwest coast exceed 300 mm and decrease toward the mountain ranges, with only about 100 mm produced over the slope or mountain regions (Fig. 3). Very little precipitation crosses the mountain, with less than 50 mm rain on the eastern side of the mountain ridge. We can show that this rainfall distribution pattern is explained by the movement of precipitating clouds that continuously develop over the upstream ocean to the coast of the island, instead of developing or becoming enhanced over the island. This long-duration precipitation system is quite stationary and heavy rainfall is limited to the southwest part of the island and upstream ocean (heavy rain region less than 150 km in diameter). We should also note that for yet-unknown reasons, both radar (not shown) and satellite estimates (Fig. 2) underestimate rainfall on 16 June (Fig. 3), but are closer to rain gauges on 14 June (not shown). This may be an important issue for the future study.
Three-hourly rainfall based on TRMM 3B42 rain analysis over regions centered at Taiwan from 2100 UTC 15 Jun to 1200 UTC 16 Jun.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Distribution of 4-hourly rainfall based on rain gauges during (a) 2000–2400 UTC 15 Jun, (b) 0000–0400 UTC 16 Jun, (c) 0400–0800 UTC 16 Jun, and (d) 0800–1200 UTC 16 Jun.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Figure 3 also reveals that during the 16-h period of precipitation over Taiwan, the heaviest rain is always confined to the coast of southwest Taiwan with very slight changes of location. The heavy rain approaches the coast at about 2000 UTC on 15 June (Fig. 3a) after the mei-yu front–related rainband moves to the south of the island in the morning of 15 June. The heaviest rain peaks during 0000–0400 UTC or 0800–1200 LST (LST = UTC + 8 h) on 16 June, producing about 60–80 mm of rain on the coastal plain during this 4-h time period (Fig. 3b). During this period, heavy precipitation (about 10 mm h−1) reaches the foothills of the south–north-oriented mountain range. After 0400 UTC on 16 June, heaviest precipitation moves slightly to the north, but still remains on the southwest coast of the island (Figs. 3c,d). Precipitation almost ceases after 1200 UTC on 16 June.
b. Synoptic overview and diagnosis
The heavy rainfall event over southwest Taiwan on 16 June occurs under the large-scale setting of a quasi-stationary mei-yu front over south China. The development of the mei-yu front can be traced back to 12 June when a southwest vortex and Plateau leeside trough (Ding and Chan 2005) develop. By 13 June, the southwest vortex (or 850-hPa cyclone) associated with the 500-hPa leeside trough becomes very well defined over southwest China at the foothills of the Yun-Gui Plateau and propagates toward the northeast (Figs. 4–5). The 500-hPa trough and 850-hPa cyclone produces strong low- to middle-level southwesterly flow. As the 500-hPa trough accompanied by the 850-hPa cyclone lifts out to the northeast, the flow over Taiwan weakens. But the initial 850-hPa cyclone helps to place the mei-yu boundary near Taiwan by 14 June (Fig. 4b). By 16 June, both the midlevel trough and 850-hPa cyclone move off the Chinese continent and become unrecognizable (Figs. 4d and 5d). In addition, winds over southeast China and Taiwan become nearly calm. We should note that there is a strong wind shear reversal between the strong southwesterly 850-hPa flow concurrent with weak 500-hPa southeasterly flow over southwest Taiwan. Strong shear reversal has been documented in back-building MCS events over the United States by Schumacher and Johnson (2008). Also on 16 June, another low pressure system develops in the lower troposphere (Fig. 5d) over southwestern China. But this low pressure system is not as strong as the one on 13 June, and the cyclonic winds around this system are much weaker. Since this low pressure system is on a relatively small scale (200-km diameter) and still far from Taiwan (500 km), the influence of this system on the heavy precipitation system over South China Sea and southwest Taiwan is uncertain.
The 850-hPa subjective analysis over East Asia at 0000 UTC on (a) 13 Jun, (b) 14 Jun, (c) 15 Jun, and (d) 16 Jun 2008. Geopotential heights are contoured black solid lines, while temperatures are in red dashed contours.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
As in Fig. 4, but for analysis at 500 hPa.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Figures 6 and 7 show the maps of 850-hPa wind superimposed with potential temperature and thermal advection at 700 hPa, respectively. These diagnostics show that during 16 June heavy rain case, the main temperature gradients associated with the mei-yu front at the 850-hPa level (or the baroclinic zone) is still over southern China (Fig. 6). The mei-yu boundary (or front) or low-level wind shear line never approaches Taiwan during the 13–16 June wet spell. During the heavy rain periods on 16 June, the ocean upstream of southwest Taiwan is dominated by warm moist southwesterly flow. Warm advection at 700 hPa is evident over the upstream ocean indicating that the large-scale environment is favorable for rising motion or convection (Fig. 7). But the large-scale ascent is not the major factor controlling the timing, location, or amount of the heavy precipitation on 16 June, although of course slow ascent is important in preconditioning the environment and helping to enhance convective instability. If the vertical motion is estimated from the thermal advection according to quasi-geostrophic (QG) theory, the rising motion produced by the large scale is on the order of 1 cm s−1. However, the rising motion required by the heavy precipitation (e.g., averaging 10 mm h−1) is on the order of 1 m s−1 over a 50–100-km region, even assuming precipitation efficiency is 100% and mixing ratio is 10 g kg−1. Therefore, large-scale rising motion, while important, is far too weak to account for the observed distribution of heavy rainfall.
Map of winds (arrows) and potential temperature (contours) at 850 hPa at different times: (a) 0000 UTC 14 Jun, (b) 1800 UTC 14 Jun, (c) 0000 UTC 15 Jun, (d) 1800 UTC 15 Jun, (e) 0000 UTC 16 Jun, and (f) 0600 UTC 16 Jun. Thermal advection at 700 hPa is coupled by color shading.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Map of thermal advection (shaded; only warm advection is plotted) and winds (arrows) at 700 hPa during different times: (a) 0000 UTC 14 Jun, (b) 1800 UTC 14 Jun, (c) 0000 UTC 15 Jun, (d) 1800 UTC 15 Jun, (e) 0000 UTC 16 Jun, and (f) 0600 UTC 16 Jun. Unit: m s−1 K (km)−1. Potential temperature at 850 hPa is superimposed by contours.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
c. Upstream low-level jets
Figure 8 shows the time series of winds over the upstream ocean (station 99810) and southwest coast of Taiwan (station 46750). Compared to the rainfall time series (Fig. 1a), strong upstream southwesterly flows are quite consistent with the heavy precipitation over southwest Taiwan during the whole period of 14–16 June (Fig. 8a). However, low-level flows over southwest Taiwan are much weaker throughout the whole rainy period (Fig. 8b). We will show that this is due to a combination of an extensive rain-produced cold pool combined with orographic blocking effects. After the low-level flows over the coast turn to strong southerly (similar to the upstream flows) by 1200 UTC of 16 June, the heavy precipitation system ends. There are no dropsondes available for this case until 0900 UTC on 16 June when the storm is already at the very late and decaying stage. Davis and Lee (2012) show that by 0900 UTC, the low-level flow has already switched to strong southerly and split over southwest Taiwan.
Time–height cross section of winds ( barbs) and potential temperature (color shaded) over (a) upstream ocean and (b) coast sounding site. Note that the upstream ship was close to Taiwan before 1200 UTC on 14 Jun, and within the blue triangle (Fig. 1a) for the next 48 h; see Davis and Lee (2012) for exact track.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Before the development of the quasi-stationary heavy precipitation system (16 June) over southwest Taiwan and the upstream ocean, low-level southwest flows start to restore warm and unstable conditions from the relatively stable conditions formed by previous precipitation on 14–15 June (Fig. 8). By 0000 UTC on 16 June, the upstream flows off southwest Taiwan coast reach the intensity of an LLJ (e.g., 850-hPa wind > 12.5 m s−1; Bonner 1968; Chen and Yu 1988) as shown in Figs. 8 and 9. The LLJ is limited to the upstream ocean away from southwest Taiwan, while low-level winds over the coast of southeast China and southern Taiwan are much weaker (Fig. 9). The LLJ is evident only up to 850–800 hPa and disappears at 700 hPa. The LLJ is also recorded in the radar radial velocity from the southwest direction (Fig. 10). The low-level jet extends from the upstream ocean to within about 50–60 km from the southwest Taiwan coast. There is strong convergence between the upstream low-level jet and weak winds within 40–50 km of the island.
Observations of wind field (barbs) and potential temperatures (K) from sounding over Taiwan and adjacent area at 0000 UTC on 16 Jun 2008. Wind field is shown at different levels: (a) 925, (b) 850, (c) 800, and (d) 700 hPa.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Radial velocity derived from S-Pol radar RHI scans at different directions (green lines in Fig. 1a) near 0200 UTC 16 Jun: (a) 222°, (b) 226°, and (c) 228°.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
During the heavy precipitation period of 0000–0600 UTC on 16 June, the southwesterly LLJ blows more than 15 m s−1 between 950 and 900 hPa. But over the coast, winds have average speed of about 5 m s−1. The wind direction is southeasterly in the lowest hundred meters and turns anticyclonically with elevation. By comparing winds upstream and near the coast, it can be concluded that significant convergence exists between the upstream ocean and the southwest coast of Taiwan. Not coincidentally, this region happens to be the exact region where convection and heavy precipitation continuously develops for about 15 h on this day.
d. Upstream conditions and storm environments
Figures 11 and 12 show soundings during the first few hours of the precipitation system over the upstream and coastal site, separately. There is a clear difference between upstream conditions and storm environment over the coast. After the storm system of 14 June passes Taiwan, the low-level environment over southwest Taiwan becomes relatively colder and stable. However, the upstream atmosphere becomes more unstable with large CAPE (> 2000 J kg−1) by 1800 UTC on 15 June. It is likely that previous precipitation formed a cold pool covering the island and adjacent ocean, but sea–air flux restored the upstream environment faster than the atmosphere over and close to the island. As shown in Davis and Lee (2012), the sea surface temperature south-southwest of Taiwan exceeds 28°C, and our Fig. 2 shows a nearly rain-free area about 300 km in diameter that would be favorable for developing a warm moist boundary layer. The unstable air mass and LLJ over the upstream ocean combined to initiate new convection and provide a continuing moisture source for the long-lived raining system. Though the moist neutral air over southwest Taiwan can be easy to lift and form precipitation, it cannot support strong convection. In this case, convection triggered upstream and fed by warm and moist air from upstream is important to this long-duration mesoscale convective system.
Sounding profiles over the ocean upstream of southwest Taiwan (99810) at (a) 0600 UTC 15 Jun and (b) 0000 UTC 16 Jun.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
As in Fig. 11, but for site at the southwest Taiwan coast (46750).
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
e. Precipitation-induced cold pool
The precipitation on 14 June probably creates and leaves behind a cold pool over the island and adjacent upstream ocean. The cold pool over the ocean may be restored by air–sea fluxes with a fetch of 300 km upstream rain-free region (Fig. 2) that is sufficient for any rain-cooled air to be modified to near equilibrium with a 28°–29°C ocean. The cold pool over the island is enhanced and maintained by the continuous precipitation and heavy clouds during the daylight of 15 June (Fig. 8). For example, there is about 4–5-K difference of potential temperature between the upstream ocean and the island. The potential temperature near the surface drops more than 5 K after the heavy precipitation over southwest Taiwan on 14 June and the low theta condition persists as precipitation continues (Fig. 8b). This lower-temperature layer is 500–600 m deep and lasts from the middle of 14–16 June. By 0000 UTC of 16 June, there is a 3°–4°C difference between the upstream ocean and the southwest part of the island in the lowest 200-m levels (Figs. 13a,b). There is still a 2°C depression of the temperature over the island from the upstream ocean at 925 hPa (Figs. 13c,d). Thus, a precipitation-formed cold pool, though moderate, exists over the southwest part of the island.
Low-level observations from soundings over Taiwan and adjacent area at 0000 UTC 16 Jun 2008. Wind field (barbs) and temperature (°C) is shown at different levels: (a) 1000, (b) 975, (c) 950, and (d) 925 hPa.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
The high terrain of Taiwan may help to block the southwesterly monsoon flow in this case and maintain the cold pool sitting over southwest Taiwan. In theory, low-level flow blocking or splitting will occur in the condition of small Froude number [Fr = U/(Nh)]. During the rainy event (e.g., 0000–0600 UTC 16 June) the U component at the terrain height h (2 km) is about 7 m s−1, while N is ~0.015 s−1. In this case, the Froude number of the low-level atmosphere over the upstream ocean is ~0.3. Under this condition the low-level flow is very likely to be redirected or decelerated. This is probably the case since the low-level (e.g., 850 hPa) winds over southwest Taiwan are quite calm (<3 m s−1) compared to the strong winds (>12 m s−1) over the upstream ocean.
In spite of having a temperature boundary of only about 2°C between the upstream warm and moist LLJ, and the relatively cool and calm air mass, in this case there was significant convergence between these air masses. The upstream LLJ could be lifted over the shallow cold pool extending from the island as shown by the example in Fig. 10. However, over time, the LLJ may also help to shrink the cold pool by horizontal mixing along its southwest flank and/or vertical mixing from above.
4. Evolution of heavily raining systems
This section examines the evolution of the heavy precipitation system in terms of large-scale view from satellite, mesoscale evolution from radar images, finer-scale evolution of convection and precipitation upstream and near the coast of southwest Taiwan, and statistics of convective cells.
a. Snapshots of satellite and radar measurements
As has been mentioned above, vigorous convection and precipitation developed along the rainband of mei-yu front over the coast of southeast China and the Taiwan Strait starting from 14 June. During the day on 15 June, convection propagates quite far south of Taiwan Island, with only scattered precipitation over the island. However, deep convection, indicated by the very cold IR brightness temperature (BT), starts to develop over the adjacent upstream ocean and coast of southwest Taiwan by 2000 UTC 15 June (Fig. 14). Two hours later, convection coverage increases and grows into a very well-defined MCS with low BT covering the whole southwest Taiwan and adjacent ocean area. This IR-viewed MCS persists more than 12 h over exactly the same region, although decreasing slightly in area with time. The coldest region of the IR images is located slightly away from southwest Taiwan, indicating the development of deep convection upstream offshore (Figs. 14d–f).
Brightness temperature (K) from IR channel of the Multifunctional Transport Satellite (MTSAT) over and around Taiwan at (a) 2257 UTC 15 Jun and (b)–(f) thereafter every 2 h.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Snapshots of radar images shown in Fig. 15 reveal more details about the convective cores related to the large precipitation region. In summary, convective cells or lines (e.g., echoes > 40 dBZ) keep developing over the upstream ocean and move very slowly to the southwest coast, but decay after reaching the island. The large heavy precipitation shield appears stationary, supported by the continuous development of new convection. Before 0600 UTC, convective lines stay quasi stationary over the upstream ocean with heavy precipitation reaching southwest Taiwan. Smaller regions of new convection continue to develop southwest of the island and merge into these convective lines.
Radar images with reflectivity > 20 dBZ from RCCG over southwest Taiwan at (a) 2301 UTC 15 Jun and (b)–(f) thereafter every 2 h. The thick white dashed line in (b) denotes the approximate edge of the cold pool.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
b. Statistics of precipitation and convection
The above summaries of the storm evolution are also confirmed by the occurrence frequency (statistics based on the RCCG radar) of both significant radar echoes and defined convective pixels during the period of heavy precipitation (Fig. 16). About 90% of the time during 2200 UTC 15 June–1200 UTC 16 June, southwest Taiwan and upstream ocean 100 km away experiences precipitation with radar reflectivity > 30 dBZ (Fig. 16a). The maximum area of 30-dBZ occurrence appears oval shaped from 119.0° to 120.5°E. The distribution of occurrence of convective radar echoes is very different from that of the 30-dBZ occurrence (Fig. 16b). A west–east-oriented band-shape maximum of convection frequency (35%) is located over the upstream ocean about 40 km away from the coast. Another convection occurrence maximum center (25%) is located just on the edge of the southwest coast. However, a very low frequency (<10%) of convection is observed over the heavy precipitation region of the plain of southwest Taiwan. In other words, the large amount of precipitation over southwest Taiwan is not mainly contributed by convective rain but by heavy stratiform precipitation evolving from the upstream convection. Therefore, continuous development of convection upstream is the main cause of the heavy precipitation over the island.
Occurrence frequency (%) of (a) radar echoes > 30 dBZ and (b) defined radar convective pixels during the period of 2200 UTC 15 Jun–1200 UTC 16 Jun. Thick white dashed line denotes the suspected location of the approximate cold pool edge at 0200 UTC 16 Jun. The X indicates the location of RCCG radar. S-Pol radar and its RHI scans from Fig. 10 are indicated by black filled circle and thick black solid line in (b).
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
c. Evolution of convective clusters
Figure 17 shows the distribution of defined convective clusters of different size every 3 h. During the first 6 h, numerous convective lines (medium to large size with diameter >10 or 20 km) are concentrated over the west–east band-shape region 40–50 km from the southwest coast (Figs. 17a,b). Some convective clusters, either large or small, move downstream to the coast, but very few survive. Before 0900 UTC, only a few small cells appear over the low land of southwest Taiwan (Figs. 17a–c). At the same time, many small convective clusters (e.g., with diameter <10 km) develop over the southwest edge of the heavy precipitation region. These small convective clusters deepen while propagating downstream and finally merge with those large convective lines (not shown). In summary, convective clusters keep developing upstream during the heavy precipitation period within the thermodynamically unstable environment and support those major convective cores by merging with them. Conditions are closer to moist neutral over the island, so the deep convection does not persist there, but evolves into continuous stratiform precipitation.
Distribution of defined convective cells of different size every 3 h on 16 Jun: (a) 0000–0300 UTC, (b) 0300–0600 UTC, (c) 0600–0900 UTC, and (d) 0900–1200 UTC. Meanings of different markers are indicated in the legend of (c). Thick black dashed line denotes suspected location of the approximate cold pool edge at 0200 UTC 16 Jun.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
d. Mesoscale flow structures based on dual-Doppler wind analysis
Figure 18 shows dual-Doppler-derived horizontal wind field and vertical motion at different levels at 0016 UTC on 16 June. The southwesterly LLJ is clearly defined at the 2-km level upstream, southwest of the convective line (40 dBZ), converging with weaker southwesterly winds downstream (Fig. 18a). To the north of the major convective core, westerly winds are observed. Evidently, the major convection region is in this convergence zone between the strong upstream southwesterly flows, weaker downstream southwesterlies, and westerlies to the north. Above the southwest LLJ, the flow turns to weaker westerlies with altitude (Figs. 18b–d). This is quite consistent with that observed by the ship soundings over the upstream ocean (Fig. 10). The derived upward motion is located very close to the deep convection and convergence zone and peaks in the midtroposphere (6–8 km). However, the vertical motion related to the convective cores is very possibly overestimated, especially above midtroposphere (with average above 10 m s−1). Therefore, the vertical cross section of the mesoscale flow is not shown.
Horizontal dual-Doppler wind analysis during 0016 UTC 16 Jun at (a) 2, (b) 4, (c) 6, and (d) 8 km. The scale of the wind arrow is located at the bottom. The shaded color indicates vertical motion, while the black contours represent regions with radar reflectivity > 40 dBZ.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
5. Hypothesis of initiation maintenance and supporting evidence
a. Hypothesis of initiation maintenance
Based on above-mentioned results and discussions, it can be summarized that the long-duration heavy precipitating system is maintained by convection developing continuously over the upstream ocean. Convective cells may initially grow as they move toward the island, but by the time they reach the island they are in the process of merging into the quasi-stationary stratiform precipitation mass on shore. This process is similar to the “echo-training” or back-building processes (Chappell 1986; Doswell et al. 1996; Davis 2001; Schumacher and Johnson 2005). The back-building system features new convection developing from the rear of the storm and propagating over the same region before the older convection completely disappears. This process requires some initiation factors to develop convection from the backside continuously. These include surface fronts, convectively generated outflow boundaries, orographic lifting, or interaction between MCVs and a LLJ.
Of all the conditions listed above, surface fronts and MCVs evidently do not exist in the case being studied. Our hypothesis is that previous precipitation before the development of the long-duration system forms a cold pool over the island and adjacent ocean, and the boundary between the cold pool and upstream moist air triggers new convection. This cold pool is weak and shallow, but acts as if to extend the island and its elevated terrain 40–50 km into the upstream ocean. Similar terrain-tied cold pool (e.g., cold air damming) has been reported in other parts of the world. For example, cold air damming east of the Appalachians over the eastern United States has been shown to effectively extend the Appalachians to the east coast in situations of coastal cyclone redevelopment in the cool season in particular (Richwein 1980; Bell and Bosart 1988). Also, for a warm season example, Srock and Bosart (2009) discuss how a heavy rainfall event ahead of tropical storm Marco (1990) is influenced by cold air damming east of the Appalachians.
The upstream LLJ starts being lifted well offshore by the cold pool rather than at the foothills of the mountain range over the island. The axis of the LLJ (shown by the radar RHIs in Fig. 10) slopes upward at the location of about 50 km from the southwest coast as it approaches the radar over the coast. Convection occurring most frequently over the LLJ upward-sloping zone is a reasonable explanation of convection initiation as the warm moist air mass rises at the edge of the remnant cold pool. After the cold pool weakens, new convection stops developing and the precipitation system slowly dies. The weakening of the cold pool is possibly related to the fact that southerly flow (Fig. 8) moves its remains northward where the convection can no longer be initiated in the same location. A schematic diagram of the hypothesis is shown in Fig. 19. We should also note that orographic blocking would result in flow deceleration upstream (Chen and Li 1995) with rising motion before reaching the foothills even without the presence of cold pool. However, with the presence of cold pool due to precipitation cooling and nocturnal cooling, blocking is more evident and persistently locked into the same location.
Schematic diagram of the hypothesis of initiation, evolution, and maintenance for the long-duration heavy precipitation system.
Citation: Monthly Weather Review 140, 8; 10.1175/MWR-D-11-00208.1
Schumacher and Johnson (2008) report that a reversal in the vertical shear above the LLJ is an important condition for backbuilding in particular cases. A long-lived MCS can be maintained by a nearly stationary low-level gravity wave without support by a convectively generated cold pool. A strong reversal of the wind shear with height is responsible for keeping the wave in this case nearly stationary. Though we propose the long-lived MCS in this study to be maintained by a precipitation-generated cold pool, the shear reversal organizing mechanism may be also applied to this back-building MCS event as in Schumacher and Johnson (2008). The feature of wind reversal above the LLJ is found in this study (Fig. 11), but further work is needed to find out whether a gravity wave may also exist in this case.
b. Summary of supporting evidence
Key components supporting the initiation and maintenance of the long-duration back-building system are shown in Fig. 19. They include 1) LLJ over the upstream ocean, 2) warm and moist unstable air over the upstream ocean, 3) cold pool (stable air) over the island and adjacent ocean, 4) convergent boundary between the upstream LLJ and calmer winds over the coast, and 5) convection developing continuously near this boundary for many hours.
Figures 8–10 clearly show LLJs over upstream ocean and properties of LLJs including depth, area, intensity, and time series. In addition, all the soundings made between 1800 UTC 15 June and 0000 UTC 16 June over the upstream ocean indicate the warm and moist unstable nature of the upstream air mass (Figs. 11 and 12). On the other hand, the cold pool over the island and adjacent ocean is revealed through the difference between near-surface temperature or potential temperature over the upstream ocean and the coast (Figs. 8 and 13) or the wind boundaries between the low-level jet and light wind within the cold pool (Fig. 10). Furthermore, the radial velocity of several RHI scans toward the southwest from S-Pol clearly shows the boundary between the LLJ and downstream lower troposphere (Fig. 10). Figure 10 also shows that the LLJ is slightly lifted upward when approaching the convergence boundary. The boundaries and convergence between the LLJ and relatively calm flow downstream is also indicated by the wind analysis from soundings (Fig. 8). All these environmental components are acting together to continuously initiate and organize convection from the backside of the long-lived multiple MCSs (Figs. 14–17).
6. Summary and conclusions
During the Terrain-influenced Monsoon Rainfall Experiment (TiMREX), one long-duration (about 15 h) extreme rainfall system develops offshore and drops about a half meter of rain on the coastal cities. This study first investigates the storm evolution on both large and mesoscale, convective processes, upstream and storm environments, and then develops and examines a hypothesis for the initiation and maintenance of the long-duration heavy precipitation process. Major summaries and conclusions on this study are listed in the following points.
Associated with the northeastward propagation of a southwest vortex and upper-level short-wave trough, a quasi-stationary mei-yu rainband develops over Taiwan Strait, South China Sea, and western Taiwan during 14–16 June.
Mei-yu front–related “squall lines” pass south Taiwan on 14 June, and convection over southwest Taiwan stops on 15 June (or stratiform precipitation and heavy clouds are dominating), but new convection redevelops southwest offshore by 1800 UTC 15 June and forms heavily raining MCSs over southwest Taiwan and adjacent ocean for about 18 h.
Inside the large MCSs, new convection keeps developing upstream offshore at the boundary between a precipitation-formed cold pool and LLJ, but decays or dies after moving into the island, dropping the heaviest rain over the upstream ocean and coastal regions.
Warm and moist unstable air carried by the LLJ from the upstream ocean keeps feeding the heavily precipitating system and convection inside the system, while the island is dominated by relatively cold (3°–4°C colder near surface) and calm conditions not favorable for the development of deep convection there.
The initiation and maintenance process of this long-lived rainy mesoscale system is similar to the “back-building–quasi-stationary” process described for extreme rainfall MCSs in the United States by others in the literature; what is unique in this case is that the cold pool is trapped by high terrains over Taiwan.
A hypothesis of “cold pool extending orographic effect” has been developed and examined: A cold pool from previous persistent precipitation forms a partial barrier to low-level moist southwest airflow, but LLJ and sea–air flux partially restores the warm, moist air upstream of Taiwan and destabilizes the lower atmosphere upstream. Convection is then triggered upstream of the cold pool boundary and the cold pool over the island prevents the convection from continuing to develop there, resulting in most of the heavy precipitation falling from the stratiform portion of the MCS.
Acknowledgments
This research was supported by NASA Precipitation Measurement Mission Grant NAG513628 under the direction of Dr. Ramesh Kakar. This study was also supported by NSF Grant AGS-1142558. We thank the SoWMEX/TiMREX science team and office sponsored by the United Sates National Science Foundation and Taiwan Central Weather Bureau and National Science Council for providing the observational and analysis data used in this paper. Special thanks go to P. Ciesielski (CSU) and R. Rilling (NCAR) for their assistance in accessing the sounding data and NCAR S-Pol data. We greatly appreciate and thank Dr. Russ Schumacher and one anonymous reviewer for careful and constructive reviews that resulted in significant improvements to the manuscript.
REFERENCES
Akaeda, K., J. Reisner, and D. Parsons, 1995: The role of mesoscale and topographically induced circulations initiating a flash flood observed during the TAMEX project. Mon. Wea. Rev., 123, 1720–1739.
Bell, G. D., and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev., 116, 137–161.
Bonner, W. D., 1968: Climatology of the low level jet. Mon. Wea. Rev., 96, 833–850.
Chang, C.-P., S.-C. Hou, H.-S. Kuo, and G.T.-J. Chen, 1998: The development of an intense East Asian summer monsoon disturbance with strong vertical coupling. Mon. Wea. Rev., 126, 2692–2712.
Chang, P.-L., P.-F. Lin, B. J.-D. Jou, and J. Zhang, 2009: An application of reflectivity climatology in constructing radar hybrid scans over complex terrain. J. Atmos. Oceanic Technol., 26, 1315–1327.
Chappell, C. F., 1986: Quasi-stationary convective events. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 289–309.
Chen, C.-S., W.-S. Chen, and Z.-S. Deng, 1991: A study of a mountain-generated precipitation system in northern Taiwan during TAMEX IOP 8. Mon. Wea. Rev., 119, 2574–2606.
Chen, C.-S., W.-S. Chen, Y.-L. Chen, P.-L. Lin, and H.-C. Lai, 2005: Investigation of orographic effects on two heavy rainfall events over southwestern Taiwan during the Mei-yu season. Atmos. Res., 73, 101–130.
Chen, C.-S., Y.-L. Chen, C.-L. Liu, P.-L. Lin, and W.-C. Chen, 2007: The statistics of heavy rainfall occurrences in Taiwan. Wea. Forecasting, 22, 981–1002.
Chen, G. T.-J., 1983: Observational aspects of the Mei-yu phenomena in subtropical China. J. Meteor. Soc. Japan, 61, 306–312.
Chen, G. T.-J., and C.-C. Yu, 1988: Study of low-level jet and extremely heavy rainfall over northern Taiwan in the mei-yu season. Mon. Wea. Rev., 116, 884–891.
Chen, G. T.-J., C.-C. Wang, and D. T.-W. Lin, 2005: Characteristics of low-level jets over northern Taiwan in mei-yu season and their relationship to heavy rain events. Mon. Wea. Rev., 133, 20–43.
Chen, G. T.-J., C.-C. Wang, and L.-F. Lin, 2006: A diagnostic study of a retreating mei-yu front and the accompanying low-level jet formation and intensification. Mon. Wea. Rev., 134, 874–896.
Chen, G. T.-J., C.-C. Wang, and S. W. Chang, 2008: A diagnostic case study of mei-yu frontogenesis and development of wavelike frontal disturbances in the subtropical environment. Mon. Wea. Rev., 136, 41–61.
Chen, S.-H., and Y.-L. Lin, 2005a: Effects of the basic wind speed and CAPE on flow regimes associated with a conditionally unstable flow over a mesoscale mountain. J. Atmos. Sci., 62, 331–350.
Chen, S.-H., and Y.-L. Lin, 2005b: Orographic effects on a conditionally unstable flow over an idealized three-dimensional mesoscale mountain. Meteor. Atmos. Phys., 88, 331–350.
Chen, Y.-L., and J. Li, 1995: Large-scale conditions favorable for the development of heavy rainfall during TAMEX IOP 3. Mon. Wea. Rev., 123, 2978–3002.
Chen, Y.-L., and J. Feng, 2001: Numerical simulations of airflow and cloud distributions over the windward side of the island of Hawaii. Part I: The effects of trade-wind inversion. Mon. Wea. Rev., 129, 1117–1134.
Chu, C.-M., and Y.-L. Lin, 2000: Effects of orography on the generation and propagation of mesoscale convective systems in a two-dimensional conditionally unstable flow. J. Atmos. Sci., 57, 3817–3837.
Ciesielski, P. E., and Coauthors, 2010: Quality-controlled upper-air sounding dataset for TiMREX/SoWMEX: Development and corrections. J. Atmos. Oceanic Technol., 27, 1802–1821.
Davis, C. A., and W.-C. Lee, 2012: Mesoscale analysis of heavy rainfall episodes from SoWMEX/TiMREX. J. Atmos. Sci., 69, 521–537.
Davis, R. S., 2001: Flash flood forecast and detection methods. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 481–525.
Ding, Y., 1992: Summer monsoon rainfalls in China. J. Meteor. Soc. Japan, 70, 373–396.
Ding, Y., and J. C.-L. Chan, 2005: The East Asian summer monsoon: An overview. Meteor. Atmos. Phys., 89, 117–142.
Doswell, C. A., III, H. E. Brooks, and R. A. Maddox, 1996: Flash flood forecasting: An ingredients-based methodology. Wea. Forecasting, 11, 560–581.
Doviak, R. J., P. S. Ray, R. G. Strauch, and L. J. Miller, 1976: Error estimation in wind fields derived from dual-Doppler radar measurement. J. Appl. Meteor., 15, 868–878.
Houze, R. A., Jr., B. F. Smull, and P. Dodge, 1990: Mesoscale organization of springtime rainstorms in Oklahoma. Mon. Wea. Rev., 118, 613–654.
Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 38–55.
Kuo, Y.-H., and G. T.-J. Chen, 1990: The Taiwan Area Mesoscale Experiment (TAMEX): An overview. Bull. Amer. Meteor. Soc., 71, 488–503.
Lee, W.-C., B. J. D. Jou, C.-R. Chen, and J.-A. Moore, 2009: Overview of SoWMEX/TiMREX. Preprints, 34th Conf. on Radar Meteorology, Williamsburg, VA, Amer. Meteor. Soc., 9B.2. [Available online at http://ams.confex.com/ams/34Radar/techprogram/paper_156254.htm.]
Li, J., and Y.-L. Chen, 1998: Barrier jets during TAMEX. Mon. Wea. Rev., 126, 959–971.
Li, J., Y.-L. Chen, and W.-C. Lee, 1997: Analysis of a heavy rainfall event during TAMEX. Mon. Wea. Rev., 125, 1060–1082.
Liou, Y.-C., and Y.-J. Chang, 2009: A variational multiple–Doppler radar three-dimensional wind synthesis method and its impacts on thermodynamic retrieval. Mon. Wea. Rev., 137, 3992–4010.
Maddox, R. A., C. F. Chappell, and L. R. Hoxit, 1979: Synoptic and meso-α scale aspects of flash flood events. Bull. Amer. Meteor. Soc., 60, 115–123.
Mohr, C. G., L. J. Miller, R. L. Vaughan, and H. W. Frank, 1986: The merger of mesoscale datasets into a common Cartesian format for efficient and systematic analyses. J. Atmos. Oceanic Technol., 3, 143–161.
Moore, J. T., F. H. Glass, C. E. Graves, S. M. Rochette, and M. J. Singer, 2003: The environment of warm-season elevated thunderstorms associated with heavy rainfall over the central United States. Wea. Forecasting, 18, 861–878.
Nesbitt, S. W., E. J. Zipser, and D. J. Cecil, 2000: A census of precipitation features in the tropics using TRMM: Radar, ice scattering, and lightning observations. J. Climate, 13, 4087–4106.
Ninomiya, K., 2000: Large and meso-α-scale characteristics of Meiyu/Baiu front associated with intense rainfalls in 1-10 July 1991. J. Meteor. Soc. Japan, 78, 141–157.
Ninomiya, K., and T. Akiyama, 1992: Multi-scale features of Baiu, the summer monsoon over Japan and East Asia. J. Meteor. Soc. Japan, 70, 467–495.
Oye, D., and M. Case, 1995: REORDER: A program for gridding radar data: Installation and use manual for the UNIX version. NCAR/ATD, 30 pp.
Pontrelli, M. D., G. Bryan, and J. M. Fritsch, 1999: The Madison County, Virginia, flash flood of 27 June 1995. Wea. Forecasting, 14, 384–404.
Raymond, D. J., and H. Jiang, 1990: A theory for long-lived mesoscale convective systems. J. Atmos. Sci., 47, 3067–3077.
Reeves, H. D., and Y.-L. Lin, 2007: The effects of a mountain on the propagation of a preexisting convective system for blocked and unblocked regimes. J. Atmos. Sci., 64, 2401–2421.
Richwein, B. A., 1980: The damming effect of the southern Appalachians. Natl. Wea. Dig., 5, 2–12.
Saha, S., and Coauthors, 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015–1057.
Sanders, F., 2000: Frontal focusing of a flooding rainstorm. Mon. Wea. Rev., 128, 4155–4159.
Schumacher, R. S., and R. H. Johnson, 2005: Organization and environmental properties of extreme-rain-producing mesoscale convective systems. Mon. Wea. Rev., 133, 961–976.
Schumacher, R. S., and R. H. Johnson, 2006: Characteristics of U.S. extreme rain events during 1999–2003. Wea. Forecasting, 21, 69–85.
Schumacher, R. S., and R. H. Johnson, 2008: Mesoscale processes contributing to extreme rainfall in a midlatitude warm-season flash flood. Mon. Wea. Rev., 136, 3964–3986.
Schumacher, R. S., and R. H. Johnson, 2009: Quasi-stationary, extreme-rain-producing convective systems associated with midlevel cyclonic circulations. Wea. Forecasting, 24, 555–574.
Srock, A. F., and L. F. Bosart, 2009: Heavy precipitation associated with southern Appalachian cold-air damming and Carolina coastal frontogenesis in advance of weak landfalling tropical storm Marco (1990). Mon. Wea. Rev., 137, 2448–2470.
Sun, W.-Y., C. C. Wu, and W. R. Hsu, 1991: Numerical simulation of mesoscale circulation in Taiwan and surrounding area. Mon. Wea. Rev., 119, 2558–2573.
Tao, S., and Y. Ding, 1981: Observational evidence of the influence of the Qinghai-Xizang (Tibet) Plateau on the occurrence of heavy rain and severe convective storms in China. Bull. Amer. Meteor. Soc., 62, 23–30.
Tsai, K.-J., 2010: Comparisons of reflectivity and rain rate derived from the TRMM precipitation radar and the CWB Cigu radar over the southwestern area of Taiwan in 2008. M.S. thesis, Graduate Institute of Atmospheric Physics, National Central University, 69 pp.
Wilson, J. W., R. D. Roberts, C. Kessinger, and J. McCarthy, 1984: Microburst wind structure and evaluation of Doppler radar for airport wind shear detection. J. Climate Appl. Meteor., 23, 898–915.
Xu, W., E. J. Zipser, and C. Liu, 2009: Rainfall characteristics and convective properties of mei-yu precipitation systems over south China, Taiwan, and the South China Sea. Part I: TRMM observations. Mon. Wea. Rev., 137, 4261–4275.
Yeh, H.-C., and Y.-L. Chen, 2002: The role of offshore convergence on coastal rainfall during TAMEX IOP 3. Mon. Wea. Rev., 130, 2709–2730.
Zhang, J., and Coauthors, 2009: High-resolution QPE system for Taiwan. Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications, S. K. Park and L. Xu, Eds., Springer, 147–162.
Zhang, Q., K.-H. Lau, Y.-H. Kuo, and S. J. Chen, 2003: A numerical study of a mesoscale convective system over the Taiwan Strait. Mon. Wea. Rev., 131, 1150–1170.