• Alpers, W., , J-P. Chen, , I-I. Lin, , and C-C. Lien, 2007: Atmospheric fronts along the east coast of Taiwan studied by ERS synthetic aperture radar images. J. Atmos. Sci., 64 , 922937.

    • Search Google Scholar
    • Export Citation
  • Baker, R. D., , B. H. Lynn, , A. Boone, , W-K. Tao, , and J. Simpson, 2001: The influence of soil moisture, coastline curvature, and land-breeze circulations on sea-breeze-initiated precipitation. J. Hydrol., 2 , 193211.

    • Search Google Scholar
    • Export Citation
  • Blanchard, D. O., , and R. López, 1985: Spatial patterns of convection in south Florida. Mon. Wea. Rev., 113 , 12821299.

  • Carbone, R. E., , J. W. Wilson, , T. D. Keenan, , and J. M. Hacker, 2000: Tropical island convection in the absence of significant topography. Part I: Life cycle of diurnally forced convection. Mon. Wea. Rev., 128 , 34593480.

    • Search Google Scholar
    • Export Citation
  • Chen, G. T-J., 1983: Observational aspects of the mei-yu phenomenon in subtropical China. J. Meteor. Soc. Japan, 61 , 306312.

  • Chen, G. T-J., , and J-S. Yang, 1988: On the spatial and temporal patterns of heavy rainfall in Taiwan mei-yu season. (in Chinese with English abstract). Atmos. Sci., 16 , 151162.

    • Search Google Scholar
    • Export Citation
  • Chen, G. T-J., , H-C. Chou, , T-C. Chang, , and C-S. Liu, 2001: Frontal and non-frontal convection over northern Taiwan in mei-yu season. (in Chinese with English abstract). Atmos. Sci, 29 , 3752.

    • Search Google Scholar
    • Export Citation
  • Chen, Y-L., , and J. Feng, 1995: The influences of inversion height on precipitation and airflow over the island of Hawaii. Mon. Wea. Rev., 123 , 16601676.

    • Search Google Scholar
    • Export Citation
  • Chi, S-S., , and G. T-J. Chen, 1989: Case study of the MCSs and rainfall during TAMEX/Phase I. (in Chinese with English abstract). Atmos. Sci., 17 , 5975.

    • Search Google Scholar
    • Export Citation
  • Chien, C-J., , and P-L. Lin, 2004: Numerical simulations of local circulations over complex terrain in southeastern Taiwan (in Chinese with English abstract). Proc. Eighth Conf. of Atmospheric Sciences, TaoYuan, Taiwan, National Science Council of Taiwan and Central Weather Bureau of Taiwan, 50–56.

  • Feng, J., , and Y-L. Chen, 1998: Evolution of katabatic flow on the island of Hawaii on 10 August 1990. Mon. Wea. Rev., 126 , 21852199.

  • Frye, J. L., , and Y-L. Chen, 2001: Evolution of downslope flow under strong opposing trade winds and frequent trade-wind rainshowers over the island of Hawaii. Mon. Wea. Rev., 129 , 956977.

    • Search Google Scholar
    • Export Citation
  • Garrett, A. J., 1980: Orographic cloud over the eastern slopes of Mauna Loa volcano, Hawaii, related to insolation and wind. Mon. Wea. Rev., 108 , 931941.

    • Search Google Scholar
    • Export Citation
  • Grossman, R. L., , and D. R. Durran, 1984: Interaction of low-level flow with the western Ghat Mountains and offshore convection in the summer monsoon. Mon. Wea. Rev., 112 , 652672.

    • Search Google Scholar
    • Export Citation
  • Houze Jr., R. A., , S. G. Geotis, , F. D. Marks Jr., , and A. K. West, 1981: Winter monsoon convection in the vicinity of north Borneo. Part I: Structure and time variation of the clouds and precipitation. Mon. Wea. Rev., 109 , 15951614.

    • Search Google Scholar
    • Export Citation
  • Huang, W-M., 2007: Numerical simulations of the convective lines occurring off the southeastern coast of Taiwan on 14–15 May 1998 (in Chinese). M.S. thesis, Dept. of Atmospheric Sciences, Chinese Culture University, Taipei, Taiwan, 112 pp.

  • Johnson, R. H., , and J. F. Bresch, 1991: Diagnosed characteristics of precipitation systems over Taiwan during the May–June 1987 TAMEX. Mon. Wea. Rev., 119 , 25402557.

    • Search Google Scholar
    • Export Citation
  • Keenan, T. D., , and R. E. Carbone, 1992: A preliminary morphology of precipitation systems in tropical northern Australia. Quart. J. Roy. Meteor. Soc., 118 , 283326.

    • Search Google Scholar
    • Export Citation
  • Kingsmill, D. E., 1995: Convection initiation associated with a sea-breeze front, a gust front, and their collision. Mon. Wea. Rev., 123 , 29132933.

    • Search Google Scholar
    • Export Citation
  • Leu, Y-S., , and P-L. Lin, 2004: Characteristics of local circulation and boundary layer over southeastern Taiwan. Proc. Eighth Conf. of Atmospheric Sciences, TaoYuan, Taiwan, National Science Council of Taiwan and Central Weather Bureau of Taiwan, 43–49.

  • Liberti, G. L., , F. Cheruy, , and M. Desbois, 2001: Land effect on the diurnal cycle of clouds over the TOGA COARE area, as observed from GMS IR data. Mon. Wea. Rev., 129 , 15001517.

    • Search Google Scholar
    • Export Citation
  • Mahrer, Y., , and R. A. Pielke, 1977: The effects of topography on sea and land breezes in a two-dimensional numerical model. Mon. Wea. Rev., 105 , 11511162.

    • Search Google Scholar
    • Export Citation
  • Mapes, B. E., , T. T. Warner, , and M. Xu, 2003: Diurnal patterns of rainfall in northwestern South America. Part III: Diurnal gravity waves and nocturnal convection offshore. Mon. Wea. Rev., 131 , 830844.

    • Search Google Scholar
    • Export Citation
  • Murakami, M., 1983: Analysis of the deep convective activity over the western Pacific and Southeast Asia. Part I: Diurnal variation. J. Meteor. Soc. Japan, 61 , 6075.

    • Search Google Scholar
    • Export Citation
  • Nettleton, L., , S. Daud, , R. Neitzel, , C. Burghart, , W-C. Lee, , and P. Hildebrand, 1993: SOLO: A program to peruse and edit radar data. Preprints, 26th Conf. on Radar Meteorology, Norman, OK, Amer. Meteor. Soc., 338–339.

  • Pierrehumbert, R. T., , and B. Wyman, 1985: Upstream effects of mesoscale mountains. J. Atmos. Sci., 42 , 9771003.

  • Powell, M. D., 1982: The transition of the Hurricane Frederic boundary-layer wind field from the open Gulf of Mexico to landfall. Mon. Wea. Rev., 110 , 19121932.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., , and P. K. Smolarkiewicz, 1993: On the dynamics of Hawaiian cloud bands. Part III: Local aspects. J. Atmos. Sci., 50 , 15601572.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., , P. K. Smolarkiewicz, , and J. Warner, 1989: On the dynamics of Hawaiian cloud bands: Comparison of model results with observations and island climatology. J. Atmos. Sci., 46 , 15891608.

    • Search Google Scholar
    • Export Citation
  • Roeloffzen, J. C., , W. D. van den Berg, , and J. Oerlemans, 1986: Frictional convergence at coastlines. Tellus, 38A , 397411.

  • Rotunno, R., , J. B. Klemp, , and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463485.

  • Smith, R. B., 1979: The influence of mountains on the atmosphere. Advances in Geophysics, Vol. 21, Academic Press, 87–230.

  • Smolarkiewicz, P. K., , R. M. Rasmussen, , and T. L. Clark, 1988: On the dynamics of Hawaiian cloud bands: Island forcing. J. Atmos. Sci., 45 , 18721905.

    • Search Google Scholar
    • Export Citation
  • Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.

  • Sun, W. Y., , and J-D. Chern, 1993: Diurnal variation of lee vortices in Taiwan and the surrounding area. J. Atmos. Sci., 50 , 34043430.

    • Search Google Scholar
    • Export Citation
  • Wai, M. M-K., , P. T. Welsh, , and W-M. Ma, 1996: Interaction of secondary circulations with the summer monsoon and diurnal rainfall over Hong Kong. Bound.-Layer Meteor., 81 , 123146.

    • Search Google Scholar
    • Export Citation
  • Willoughby, H. E., , and P. G. Black, 1996: Hurricane Andrew in Florida: Dynamics of a disaster. Bull. Amer. Meteor. Soc., 77 , 543549.

  • Wilson, J. W., , and D. L. Megenhardt, 1997: Thunderstorm initiation, organization, and lifetime associated with Florida boundary layer convergence lines. Mon. Wea. Rev., 125 , 15071525.

    • Search Google Scholar
    • Export Citation
  • Yu, C-K., , and B. J-D. Jou, 2005: Radar observations of the diurnally forced offshore convective lines along the southeastern coast of Taiwan. Mon. Wea. Rev., 133 , 16131636.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Topography in southern Taiwan. Terrain height (m MSL) is indicated by shading. Location of the Doppler radar site at Green Island is denoted by the triangle. Locations of select surface observing stations (Green Island, GI; Chengkung, CK; Taitung, TT; Tawu, TW; and Lanyu, LY) and the NCEP–NCAR gridded data are denoted by open circles and square, respectively.

  • View in gallery

    Low-level PPI scans (∼1.6° elevation) of reflectivity (dBZ) from the GI radar from four selected cases of convective lines at (a) 0331 LST 10 Nov 2004, (b) 0331 LST 11 Jan 2000, (c) 0621 LST 17 Dec 2002, and (d) 0531 LST 9 Nov 2004.

  • View in gallery

    Maximum (upper curve) and mean (lower curve) reflectivities of each identified convective line obtained from the low-level PPI scan (1° ∼ 1.6° elevations) of the GI radar during their most organized stage. Horizontal thick lines indicate averaged values for the maximum and mean reflectivities. Vertical dashed lines mark the borders of different seasons.

  • View in gallery

    The spatial distribution of the formative frequency of convective lines (color shading) occurring off the southeastern coast of Taiwan, derived from the 1°-elevation PPI scans of radar reflectivity from the GI radar during 1998–2004. Topography in southern Taiwan is indicated with gray shading (key at left). Locations of select surface observing stations, as in Fig. 1, are also indicated for reference.

  • View in gallery

    (a) The distribution of correlation coefficients between the maximum coastal terrain height and the formative frequency of the convective lines as a function of inland distance along the southeastern coast of Taiwan. (b) Along-coast profile of maximum terrain height (shading) obtained within the 5-km-inland distance, and its corresponding profile of maximum formative frequency of convective lines (solid curve) found offshore. Relative locations of three coastal surface stations, as in Fig. 1, along this terrain profile are also indicated for reference.

  • View in gallery

    Time series of mean surface winds averaged during the formative days of the convective lines from three coastal surface stations and two offshore locations, as in Fig. 1. The offshore (onshore) flow component defined as the wind component perpendicular to the mean orientation of the coastline (∼25° clockwise from the north) is indicated by the solid (dashed) line with the positive value representing the wind toward the ocean. In each panel, full wind barbs correspond to 5 m s−1, half barbs to 2.5 m s−1.

  • View in gallery

    Number of cases of convective lines at different time intervals during a day for all 211 convective lines identified in this study.

  • View in gallery

    Number of cases of convective lines at their different durations.

  • View in gallery

    The mean environmental wind profile averaged from all identified cases of convective lines derived from the NCEP–NCAR gridded data located ∼130 km off the southeastern coast of Taiwan. Here, U (solid line) and V (dashed line) denote the cross-line and along-line wind components, respectively. Since convective lines are approximately parallel to the coast, the value of U is roughly identical to the offshore–onshore flow component, with the positive value representing the wind toward the ocean. Full wind barbs correspond to 5 m s−1, half barbs to 2.5 m s−1.

  • View in gallery

    (a) As in Fig. 7 but for the nearshore convective lines and (b) for the offshore convective lines.

  • View in gallery

    Distribution of statistical frequencies of the convective lines over different time intervals and offshore distances. Formative frequencies are indicated by shading and also contoured with a two-case interval. The thick dashed horizontal line marks the offshore distance at 40 km.

  • View in gallery

    (a) Locations of line segments used to extract reflectivity and radial velocity data from the low-level PPI scan (0.5°– 1.6° elevations) of the GI radar for the composite analysis shown in (b). Locations of TT, GI, and NCEP, as in Fig. 1, are also indicated. (b) Composite radial velocities (m s−1) within the nearshore (solid line) and offshore (dashed line) convective lines, with positive (negative) values representing the offshore (onshore) flow. The strongest reflectivity found within lines is located at X = 0 km.

  • View in gallery

    Distribution of Froude numbers calculated from each identified case of convective lines as a function of the offshore distance of their formative location. The events of the convective lines selected for the composite analysis shown in Fig. 12 are marked by solid triangles for reference. The thick solid line denotes the mean Froude number calculated with respect to a given offshore distance.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 13 13 6
PDF Downloads 15 15 4

Statistical Location and Timing of the Convective Lines off the Mountainous Coast of Southeastern Taiwan from Long-Term Radar Observations

View More View Less
  • 1 Department of Atmospheric Sciences, Chinese Culture University, Taipei, Taiwan
© Get Permissions
Full access

Abstract

In this study, observations from the C-band Doppler radar on Green Island, which is located off the southeastern coast of Taiwan, during 1998–2004 were analyzed to investigate the statistical characteristics of the convective lines occurring off the mountainous coast of southeastern Taiwan, with emphasis on their formative location and timing. A total of 211 cases of the lines were identified during the study period. It is shown that the lines were clearly a year-round phenomenon and five cases occurred per month on average. Statistical analyses for all identified cases reveal that the formative area of the lines was extended substantially from nearshore regions to at least ∼100 km offshore, with the region of the most frequent formation primarily confined to an elongated zone located ∼30 km off the coast. Along-coast variations of line formation were also evident and were shown to be closely related to the nearshore terrain features. The lines tended to form more frequently during the nighttime hours than the daytime hours, with a formative peak between 2000 and 2200 LST. Minimum formation was found near noon between 1200 and 1400 LST; however, considerable cases still could be found during the late morning and late afternoon hours. More than 70% of the lines (∼150 cases) had a duration of less than 4 h, and the mean duration for all lines was calculated to be ∼3.5 h. In addition, this study also documented statistical differences in the formative and flow characteristics between the nearshore and offshore lines, which are distinguished by offshore distances of formative location that are less or greater than 40 km, respectively. Of particular note, the analyses presented strongly suggest that the physical mechanisms contributing to the initiation of the nearshore and offshore lines are fundamentally different.

Corresponding author address: Prof. Cheng-Ku Yu, Dept. of Atmospheric Sciences, Chinese Culture University, 55, Hwa-Kang Rd., Yang-Ming-Shan, Taipei 111, Taiwan. Email: yuku@faculty.pccu.edu.tw

Abstract

In this study, observations from the C-band Doppler radar on Green Island, which is located off the southeastern coast of Taiwan, during 1998–2004 were analyzed to investigate the statistical characteristics of the convective lines occurring off the mountainous coast of southeastern Taiwan, with emphasis on their formative location and timing. A total of 211 cases of the lines were identified during the study period. It is shown that the lines were clearly a year-round phenomenon and five cases occurred per month on average. Statistical analyses for all identified cases reveal that the formative area of the lines was extended substantially from nearshore regions to at least ∼100 km offshore, with the region of the most frequent formation primarily confined to an elongated zone located ∼30 km off the coast. Along-coast variations of line formation were also evident and were shown to be closely related to the nearshore terrain features. The lines tended to form more frequently during the nighttime hours than the daytime hours, with a formative peak between 2000 and 2200 LST. Minimum formation was found near noon between 1200 and 1400 LST; however, considerable cases still could be found during the late morning and late afternoon hours. More than 70% of the lines (∼150 cases) had a duration of less than 4 h, and the mean duration for all lines was calculated to be ∼3.5 h. In addition, this study also documented statistical differences in the formative and flow characteristics between the nearshore and offshore lines, which are distinguished by offshore distances of formative location that are less or greater than 40 km, respectively. Of particular note, the analyses presented strongly suggest that the physical mechanisms contributing to the initiation of the nearshore and offshore lines are fundamentally different.

Corresponding author address: Prof. Cheng-Ku Yu, Dept. of Atmospheric Sciences, Chinese Culture University, 55, Hwa-Kang Rd., Yang-Ming-Shan, Taipei 111, Taiwan. Email: yuku@faculty.pccu.edu.tw

1. Introduction

Crucial for improving coastal weather forecasting is a better understanding of moist convection occurring in coastal regions. In nature, a variety of convective forcings probably occur near the coast. Two of the most well-known factors contributing to the formation of clouds and precipitation near the coast are the land–sea breezes (Chen 1983; Keenan and Carbone 1992; Kingsmill 1995; Wilson and Megenhardt 1997; Carbone et al. 2000) and the differential surface roughness between land and ocean (Powell 1982; Roeloffzen et al. 1986; Willoughby and Black 1996). For a coastal segment with significant topography, the development of coastal moist convection may also be related to the thermally induced katabatic–anabatic flow (Feng and Chen 1998; Frye and Chen 2001) and circulations produced dynamically by orographic blocking (Smolarkiewicz et al. 1988).

It is generally recognized that coastal forcings can result in the occurrence of clouds and precipitation not only over interior coastal regions but also over coastal water (i.e., the offshore convection), although the physical processes leading to the initiation of inland and offshore convection may be significantly distinct. A large number of previous investigations have proposed different mechanisms to explain the formation of moist convection observed off the coast. Despite the differences in geographical location and synoptic environment, boundary layer convergence produced as the offshore flow associated with the cool land breeze encounters synoptically prevailing flow was found to be a common forcing contributing to the initiation of offshore convective activities (e.g., Houze et al. 1981; Liberti et al. 2001). Similar to the scenario of this mechanism, if high mountains are present immediately adjacent to the coast, the characteristics of the offshore flow may also be influenced by thermally driven slope breezes (Mahrer and Pielke 1977; Garrett 1980; Wai et al. 1996) and/or by the mountain-induced return flow (Rasmussen et al. 1989; Rasmussen and Smolarkiewicz 1993; Alpers et al. 2007). Under these relatively complicated circumstances, the structure and intensity of the offshore flow would be determined by the mixed influence of land–sea thermal contrast and topographic effects, both of which are practically difficult to separate in real observations. In addition to the importance of coastal offshore flow on the initiation of clouds and precipitation, other convective forcings such as upstream blocking and diurnal gravity waves have also been described to interpret the occurrence of moist convection off the mountainous coasts (Grossman and Durran 1984; Mapes et al. 2003).

In contrast to the formative mechanisms of coastally generated moist convection, which have been largely discussed, another important issue dealing with their temporal and spatial variations has been relatively less explored. Processes leading to these variations are more challenging to investigate since they involve a wide variety of environmental factors (Blanchard and López 1985; Chen and Feng 1995; Baker et al. 2001). The presence of significant topography in coastal zones and the inherent interaction between convectively generated outflow and environmental kinematics and thermodynamics may also modulate coastal convection and further complicate its observed variations (Mahrer and Pielke 1977; Kingsmill 1995; Carbone et al. 2000).

Nevertheless, a dominant portion of these previous studies were confined primarily to the investigation of the distribution of clouds and precipitation over interior coasts. The specific physical mechanisms driving the temporal and spatial variabilities of oceanic convection near coastal regions are much less understood (e.g., Murakami 1983; Liberti et al. 2001). Although it is possible that the factors influencing the variations of inland, diurnally forced convection, as described above, would be similarly important in contributing to the variations of offshore convection, analyses of more observations collected over coastal water are required to fill the large gap in our knowledge regarding the remote influences of landmass on offshore convection.

Taiwan is a mountainous island, and in the absence of significant weather systems (e.g., fronts and typhoons) the land–sea breeze usually dominates boundary layer wind patterns and often plays an important role in the development of precipitation over Taiwan (Chen and Yang 1988; Chi and Chen 1989; Chen et al. 2001). Thermally induced mountain–valley winds were also frequently found to be significant in modulating the coastal land–sea breeze circulations and their associated moist convection (Johnson and Bresch 1991; Sun and Chern 1993; Leu and Lin 2004; Chien and Lin 2004). In distinct contrast to the western coast of Taiwan, the topography along eastern Taiwan is rather steep, with terrain rising abruptly to more than 1500 m MSL within 20 km of the shore (Fig. 1). Such a geographical configuration represents an excellent natural laboratory to explore our knowledge of coastal convection influenced by both land–sea thermal contrasts and topographic effects. In this particular geographical location, the cloud and/or precipitation lines occurring off the eastern coast of Taiwan are one of the most well-known and frequent mesoscale phenomena. These offshore lines are typically oriented parallel to the coast. They are frequently characterized by intense deep convection with multiple centers of heavy precipitation, which potentially threaten the safety of civil aviation and military activities over this local area. Some limited aspects of the lines, as well as their possible formative mechanisms, have been noted by a few recent observational studies.

With radar and surface observations collected from a specially chosen period (11–15 May 1998), when the convective lines were active off the southeastern coast of Taiwan, Yu and Jou (2005, hereafter referred to as YJ) documented that the lines were characterized by an elongated narrow zone of heavy precipitation and were generally located ∼10–30 km offshore, with a clear trend to initiate during nighttime hours but dissipating rapidly after sunrise. Their detailed analyses of an intense event on 14–15 May further showed that the low-level convergence, produced as the coastal cool offshore flow developing at night encountered the prevailing onshore flow, was a crucial convective forcing contributing to the formation of the offshore precipitation line. With synthetic aperture radar (SAR) images, Alpers et al. (2007, hereinafter referred to as ACLL) documented six cases of quasi-stationary atmospheric fronts off the eastern coast of Taiwan. Their results indicate that these coastal fronts could often result in the formation of coast-parallel cloud bands through the convergence between a weak easterly synoptic-scale wind and coastal offshore flow. They also argued that the presence of coastal offshore flow would probably result from the thermally driven land breeze–katabatic wind or from the recirculated airflow produced as the synoptic-scale onshore flow was blocked by coastal topography. However, owing to the lack of adequate observations along the eastern coast of Taiwan, ACLL could not explicitly address the origin of the coastal offshore flow. Moreover, the cloud bands documented in ACLL occurred during the daytime (mainly late morning) and were located farther offshore, ∼30–70 km away from the eastern coast of Taiwan. These observational aspects are distinctly different from those of the convective lines described in YJ, which were developing at night and located closer to shore.

In fact, these previous investigations of convective lines, focusing only on a few particular cases or months, prevent adequate understanding of their generality. In an effort to gain a more comprehensive view of the phenomenon, the present study uses the long-term observations from the C-band Doppler radar on Green Island collected during 1998–2004 to investigate their statistical characteristics, with particular emphasis on their observed location and timing. How these spatial and temporal variations relate to the ambient conditions (such as coastal winds, topographic features, and synoptic prevailing flow) is also discussed in this paper. This study is guided by several fundamental questions: Where was the place most preferred for formation of the convective lines? Did these lines develop more frequently during the nighttime or daytime hours? Were these observed characteristics related only to the diurnal forcings, or if not, to what extent were they the result of topographic effects or other as-yet-undocumented physical processes?

2. Overview of data and cases

The primary datasets used in this study were provided by the Weather Wing of the Chinese Air Force operational C-band (5.33 cm) Doppler radar on Green Island, off the eastern coast of Taiwan, and include volumetric distributions of reflectivity and radial velocity. Details about the Green Island radar characteristics can be found in YJ. The radar is situated ∼40 km off the southeastern coast of Taiwan (Fig. 1), providing broad coverage of precipitation information over the coastal region. In this study, the scanning volumes of radar data were collected from 1998 to 2004, with a temporal interval of 15 min–1 h between each volume. However, the Green Island radar was significantly damaged twice due to intense gust winds associated with Typhoons Otto (1998) and Bilis (2000). These two unfortunate events resulted in major observational gaps from the radar during 1998–2004. Additional gaps in radar observations were associated with some relatively short periods of breakdown and maintenance for the radar system. The monthly availability of radar observations during 1998–2004 is listed in Table 1. There were a total of 1253 days (∼3.5 yr) of radar observations available from 1998 to 2004. Other data sources used in this study are also indicated in Fig. 1, including routine surface and island observations within the coastal zone of southeastern Taiwan. A vertical sounding profile obtained from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis gridded data located ∼130 km off the southeastern coast of Taiwan (Fig. 1) is also used to represent the large-scale environmental conditions over the study domain.

In this study, the occurrence of a convective line off the southeastern coast was checked subjectively by screening a large set of low-level plan-position indicator (PPI) scans of reflectivity from the available Green Island radar data. Determination of a case for the present study was based on the following criteria. A visual line pattern of radar-observed reflectivity must persist for at least 1 h or longer. Also, the length of the convective line must be greater than 50 km during its most intense and/or organized period. These two criteria could effectively avoid including precipitation episodes with relatively transient features over coastal water. The ending time of an event was determined when the line pattern of the reflectivity became broken and much less organized and its associated precipitation elements started to dissipate persistently. Moreover, the large-scale environment associated with the observed convective lines needed to be characterized by weakly synoptically forced weather conditions. Hence, the diurnal forcings and the topographic effects could be maximized in our analyses, and the precipitation bands influenced or forced by significant weather systems (e.g., fronts, cyclones, or typhoons) were substantially precluded. Typical examples of low-level reflectivity patterns associated with the convective lines identified are shown in Fig. 2.

With comprehensive examination of available radar measurements, a total of 211 cases of convective lines satisfying the above criteria were identified. A list of the monthly number of cases for the identified lines and a complete listing of the dates of all identified cases are provided in Tables 1 and 2, respectively. It is shown that the convective lines could occur in different seasons and clearly were a year-round phenomenon. The most frequent occurrence was during springtime (74 cases), with much fewer cases during summer and the early autumn months (Table 1). On average, five convective lines occurred per month off the southeastern coast of Taiwan. Variations of the low-level strongest and mean radar reflectivities associated with each identified line during their most intense, organized stage are shown in Fig. 3. The mean reflectivity for each case was obtained by simply averaging all available reflectivity values along the line from the low-level PPI scan of the Green Island radar. The precipitation intensity of the convective lines appears to vary considerably case by case. The maximum reflectivity found within the lines was generally greater than 30 dBZ. A few extreme cases even reached above 50 dBZ. There was some suggestion of slightly stronger precipitation for late spring, summer, and late winter cases, but the overall trend of seasonal variations for the precipitation intensity was not very clear.

Since the Green Island radar is located over coastal water and is adjacent to the steep coastal barrier in southern Taiwan (Fig. 1), raw measurements from lower scanning elevations of the radar frequently contain considerable contamination due to ground, mountain, or sea clutter. These contaminated radar echoes and/or obviously spurious data were removed or corrected for all identified cases of convective lines using the NCAR SOLO software (Nettleton et al. 1993). This tedious editing work was crucial to ensure higher accuracy in the statistical analyses of these identified cases, which will be presented in sections 3 and 4.

3. Statistics of all cases

a. Spatial distribution

To investigate the characteristics of the spatial distribution of line formation, a representative 1°-elevation PPI scan of the radar reflectivity for each identified convective line was first selected during which the line initially appeared offshore. Since the reflectivity threshold of 15 dBZ, a value generally greater than the mean reflectivities of the observed lines (Fig. 3), was found to appropriately locate the convective line, the position of each line was then objectively identified in regions having radar reflectivities larger than the reflectivity threshold value. Figure 4 shows the spatial distribution of the line formation for the 211 identified cases. The area of the most frequent formation was primarily confined to an elongated, narrow zone located ∼30 km off the southeastern coast and approximately parallel to it. A local region of maximum formation could also be found very close to shore (∼5–10 km off the coast) near the coastal segment of the Chengkung station (CK in Fig. 4). Although the number of cases was obviously reduced off the elongated zone of maximum formation, the formative area of the convective lines was substantially extended from nearshore regions to at least ∼100 km offshore.

There was also some evidence of along-coast variations for line formation (Fig. 4). The peak of the formation (>24 cases) was located immediately adjacent to the Chengkung station. A secondary maximum of line formation (>18 cases) was located in a more southern coastal segment between the Taitung (TT) and Tawu (TW) stations. The distribution of formative frequency appears to have some connection with the coastal terrain features. To explore this relationship, a terrain analysis was performed along the southeastern coast of Taiwan. First, a sequence of line segments was drawn normal to the mean orientation of the coastline (∼25° clockwise from the north) for all positions along the coast. Along each line segment, the maximum terrain height within a specified inland distance and the maximum formative frequency of the convective lines found offshore were recorded. Afterward, a correlation coefficient calculated from all line segments within the analysis domain of Fig. 4 could be obtained at a specified inland distance.

Figure 5a shows the distribution of the correlation coefficients as a function of inland distance. The statistical significance of the correlation coefficient was evaluated by assuming that the statistic t has a Student’s distribution with 181 independent samples (179 degrees of freedom). This yielded threshold correlation coefficients of ∼0.1 and ∼0.2 at the 95% and 99% confidence levels, respectively. It appears that the most significant impact of terrain on line formation is found in the first 5 km inland, with a correlation coefficient of close to 0.6. The correlation coefficients are reduced persistently farther inland and are smaller than 0.4 beyond a 10-km-inland distance. This result suggests that the nearshore orographic features were more important in influencing the along-coast variations of line formation. The along-coast profile of the maximum terrain height obtained within the 5-km-inland distance and its corresponding maximum frequency found offshore, as shown in Fig. 5b, reveals a clear, consistent trend. Particularly, the main peaks of the line formation near Chengkung and between Taitung and Tawu correspond generally to the local highest mountains, where terrain heights can reach above 1000–1250 m (MSL).

Why the coastal area near Chengkung was most favored for line formation can be reasonably interpreted as a consequence of stronger and more frequent offshore flow in this local region due to the terrain influences. Previous observational and modeling studies of local circulations over southeastern Taiwan have suggested the general importance of both thermally driven slope breezes and the orographically induced return flow on the modification of coastal land–sea breezes (Leu and Lin 2004; Chien and Lin 2004; YJ; ACLL; Huang 2007). The impact of these topographic effects on the development of offshore flow would be most significant in coastal regions, particularly near Chengkung, since this site was located immediately adjacent to the highest and steepest mountain of the Coastal Range (cf. Fig. 4). It thus can be anticipated that the orographically modulated/driven offshore flow, as it encounters synoptically prevailing onshore flow, could produce a more pronounced and frequent boundary layer convergence that can help initiate offshore convection in this coastal area.

The impact of nearshore topography (particularly near the coastal segment of Chengkung) on coastal winds can be readily seen from the analyses and statistics of surface winds in the vicinity of southeastern Taiwan encompassing the study period. Figure 6 shows the time series of mean surface winds averaged during the formative days of the convective lines from three coastal stations (Chengkung, Taitung, and Tawu) and two offshore locations at Green Island and the NCEP–NCAR gridded data. Mean coastal winds exhibited a clear diurnal trend, with offshore flow prevailing at night and onshore flow during the day (Figs. 6a–c). In contrast, the hourly mean winds measured well offshore, as seen from Green Island (Fig. 6d) and the 1000-hPa NCEP–NCAR winds (Fig. 6e), indicated prevailing onshore flow and little evidence of diurnal variation. The time of onset of the mean offshore flow observed at Chengkung was around 1600 LST (LST = UTC + 8 h; Fig. 6a), ∼2–3 h earlier than those at Taitung and Tawu (Figs. 6b and 6c). Similar development of coastal offshore flow was also noted by YJ and they documented an earlier initiation of offshore flow and precipitation near Chengkung.

Moreover, as shown in Table 3, the mean magnitude of the offshore flow observed at Chengkung was calculated to be 1.2 m s−1, which was the strongest when compared with others elsewhere along the southeastern coast, such as at Taitung (0.8 m s−1) and Tawu (1.1 m s−1). Earlier onset and stronger intensity of mean offshore flow near Chengkung are consistent with a more direct influence of the thermally induced mountain winds in this particular coastal region. Table 3 also indicates the occurrence of offshore flow more frequently at Chengkung (70%) than other coastal stations. A considerable percentage of offshore flow (42%) was evident even during the daytime hours at Chengkung, in contrast to a relatively small percentage of daytime offshore flow observed at Taitung (26%) and Tawu (23%). It is obvious that the presence of daytime offshore flow at Chengkung would not likely be related to the diurnally driven forcings that favor only the development of coastal onshore flow during the daytime. Instead, it could be appropriately explained by the occurrence of orographically induced return flow as the prevailing onshore flow was blocked by extremely steep coastal mountains near Chengkung (cf. Fig. 4). As supported by a recent modeling study of diurnal circulations and convective lines over southeastern Taiwan by Huang (2007), coastal offshore flow frequently occurs during the daytime hours near Chengkung, and an enhanced blocking by the steep and high mountains along the Coastal Range was found to produce orographic return flow in this particular coastal segment.

b. Temporal distribution

The distribution of the formative time for all identified convective lines is shown in Fig. 7. It is clear that the convective lines tended to form more frequently during the nighttime hours than the daytime hours, with a formative peak between 2000 and 2200 LST. A relative maximum of line formation was found in the early morning between 0600 and 0800 LST. It is important to note that there was still a considerable number of cases (>40 cases) during late morning between 0800 and 1200 LST and during late afternoon between 1400 and 1800 LST. The formation of these daytime convective lines cannot be simply explained by the low-level convergence produced as the coastal offshore flow associated with the land breeze encounters the prevailing onshore flow as described in YJ. Instead, the orographic effects and/or other physical processes might be more relevant in this respect. This issue will be elaborated upon further in the next section.

The lifetime statistics for the convective lines are shown in Fig. 8. A dominant portion of the lines did not develop into a long-lived convective system, and only very few cases could persist longer than 6 h. More than 70% of the lines (∼150 cases) had a duration of less than 4 h. The mean duration for all lines was calculated to be ∼3.5 h. The relatively short-lived nature of the observed lines is basically consistent with the characteristics of the ambient vertical shear in the line’s environment.

As shown in Fig. 9, the mean wind profile calculated from all 211 identified cases indicates that the line environment was characterized by weak-to-moderate easterly onshore flow (∼5 m s−1) in the lowest 2 km (MSL), with prevailing westerly flow aloft. Particularly, an obvious westerly shear was evident above 1 km (MSL). Rather weak vertical wind shear and nearly constant potential temperature (not shown) found in the lowest few hundred meters may be indicative of typical characteristics of a well-mixed marine boundary layer (Stull 1988). Based on the consideration of vorticity balance dynamics, as proposed by Rotunno et al. (1988), the vertical tilt of moist convection is primarily determined by the relative vorticity magnitude of the convectively generated cold pool and the ambient vertical shear. Since the studied convective lines were typically characterized by not only a narrow zone of radar reflectivity but also a rather slow-moving speed, the spatial extent and strength of the cold pool produced by evaporative cooling of hydrometeors associated with the lines should be largely limited. In addition, the nighttime land breeze may represent another possible source of cool air feeding the convectively generated cold pool. Given a typically weak temperature contrast between the land and sea in this geographical location, this effect is also expected to be minor. With these characteristics, it appears reasonable to suggest a quite weak nature of the cold pool in the present study. The presence of a deep layer of pronounced westerly shear, as shown in Fig. 9, should be more dominant in determining the vertical tilt of moist convection, which substantially favors the line’s precipitation to being tilted eastward and to falling into the warmer inflow region1 (i.e., east of the line). This convective modification on the inflow side has been observed in YJ and was found to be fundamentally detrimental to the maintenance of the convective line through the reformation of moist convection ahead of the line.

4. Nearshore and offshore convective lines

a. Definition and statistical results

As shown in Fig. 4, in addition to a concentrated zone of line formation located close to the shore, there were still considerable cases of convective lines occurring well offshore. This result is consistent with observations from previous case studies of convective lines off the southeastern coast of Taiwan showing the diverse nature of the line location relative to the coastline, ranging from ∼10–30 km off the coast (e.g., YJ) to at least ∼30–70 km offshore (e.g., ACLL). Similar features could also be clearly seen in Fig. 2. However, an important but unclarified issue concerns whether there are some statistical differences in the formative time and flow characteristics between the lines formed nearshore and those lines observed well offshore. To explore these aspects, all convective lines identified in this study are initially classified into two categories, the nearshore and offshore lines, which are defined by whether the offshore distance of the formative location was less or greater than 40 km. The value of 40 km adopted herein was obtained from the typical propagation speed of the leading edge of the land breeze (∼1 m s−1, as observed in YJ) along the southeastern coast of Taiwan times the nighttime duration (∼12 h). As such, it could be considered to be an approximate length scale for the maximum seaward extent of the offshore flow associated with the nighttime land-breeze circulation. It can be generally expected that the formation of the nearshore lines would have had more of a chance to be influenced by the coastal diurnal forcings.

According to the classification described above, 148 (63) cases of the 211 identified convective lines are nearshore (offshore) lines. Results for the formative time of the nearshore and offshore lines are shown in Fig. 10. It is clear that a predominant portion of the nearshore lines was formed during the nighttime hours, with a formative peak between 2000 and 2200 LST (Fig. 10a). More than 83% of the nearshore lines (∼124 cases) were observed between 1800 and 0800 LST. A formative minimum could be found during the late morning and near noon (Fig. 10a). The formation of the offshore lines did not exhibit a clear, consistent trend of diurnal variation like the nearshore lines did (Fig. 10b). Although there was also some indication of more cases found during the nighttime hours, the temporal variations for the formation of the offshore lines were rather complicated and exhibited multiple local minima–maxima during the day. These features suggest a relatively looser link of the formation of the offshore lines to the diurnal forcings.

A more continuous, detailed aspect for the formative preference of the nearshore and offshore lines can be best seen from a time-offshore-distance section of statistical frequencies (Fig. 11). This analysis clearly confirms that the offshore distance of ∼40 km, as proposed earlier, is a physical borderline approximately separating two distinct characteristics of the formative distribution. There was clearly a bimodal distribution for the formation of the nearshore lines, with two major maxima located ∼10–30 km offshore between 1800 and 2200 LST and between 0400 and 0600 LST. The occurrence of the maximum during the early night at 1800–2200 LST was close to the onset of the nighttime coastal offshore flow (Fig. 6), which may initiate the nearshore convergence zone conducive to the triggering of moist convection off the coast. The second maximum in the early morning just before sunrise (0400–0600 LST) was probably related to the presence of the strongest coastal offshore flow observed during this particular period. As evident in Fig. 6, the mean coastal offshore flow was increased with time after sunset and reached a maximum of intensity around 0600 LST due to the persistent radiative cooling over land at night. Similar temporal variations of coastal offshore flow developing along the southeastern coast of Taiwan have been previously noted in YJ.

In contrast to a rather concentrated pattern for the formative preference of the nearshore lines, the formation of the offshore lines was distributed in a relatively uniform, random manner over a much wider zone of time and space. No obvious evidence of a localized maximum of formative areas could be seen. The contrasting features of the temporal and spatial distributions between the nearshore and offshore lines, as seen from Figs. 10 and 11, strongly imply that the physical mechanisms contributing to their initiation are fundamentally different.

b. Possible convective forcings

As described in section 3, the low-level winds over the study domain were primarily characterized by two distinct airflow patterns. One was the diurnally driven coastal winds along the southeastern coast of Taiwan with offshore (onshore) flow during the nighttime (daytime), and the other was the large-scale prevailing onshore flow observed over the offshore regions. Moreover, the heights of the level of free convection (LFC) seen from the NCEP–NCAR soundings (not shown) during the occurrence of convective lines were calculated to range from 0.6 to 2 km (MSL) and have a mean value of ∼1 km, implying a general need of low-level forcing for convective initiation. Given that the nearshore convective lines were largely confined to nighttime hours and close to shore, it is thus reasonable to speculate that the low-level convergence between the coastal offshore flow developing overnight and the prevailing onshore flow would probably be an important forcing for initiating these nearshore cases. As is evident in YJ, this low-level convective forcing was observed to be crucial in providing lifting to trigger moist convection for their observed convective lines occurring nearshore along the southeastern coast of Taiwan. Nevertheless, it is obvious that this kind of formative mechanism of convection cannot easily explain the observed characteristics of the offshore convective lines in terms of their formative times and locations.

In an attempt to clarify the uncertainties described above, the low-level airflow patterns within the nearshore and offshore lines observed from the low-elevation PPI scan of the Green Island radar were investigated. Owing to the inherent limitations of the observational geometry of the radar, the low-level onshore–offshore flow can be accurately retrieved only from radial velocity measurements with radar beams approximately normal to the coastline. This means that those particular cases of convective lines with relatively good coverage of radar echoes in the western or eastern vicinities of Green Island can be used to provide useful wind information. Given the high variability of the line location and the rather narrow widths of most of the observed convective lines, there were only 15 (14) cases of the nearshore (offshore) lines observed that met the requirements of the radar analysis. These selected cases are marked with font settings in Table 2. To obtain the representative features of the airflow in the vicinity of these two types of convective lines, a spatial composite of the low-level radial velocities in the cross-line direction was then performed for the nearshore and offshore cases, with respect to the position of the strongest reflectivity found within the lines (Fig. 12b). The positions of line segments used to extract radar measurements from each case are shown in Fig. 12a. Note that these line segments are chosen to be roughly perpendicular to the coastline and pass through regions near the radar site and, hence, the radial velocities obtained from these line segments are approximately identical to the magnitude of the onshore–offshore flow components.

Figure 12b indicates that the region of the nearshore line (solid line) was characterized by a pronounced wind transition from the onshore flow (negative radial velocities) east of the line to the offshore flow (positive radial velocities) on its west. Particularly, the location of maximum reflectivity within the line (i.e., X = 0 km in Fig. 12b) just coincided with the boundary of the two opposite flows with near-zero radial velocities. The onshore and offshore flow components had maximum magnitudes equal to ∼1.5 and ∼1 m s−1, respectively. The presence of the radar-derived offshore (onshore) flow observed on the western (eastern) side of the line is consistent with the mean ambient flow calculated from surface winds measured at Taitung and Green Island, as indicated in Table 4. These observed features further support the significance of the low-level convergence produced as the coastal offshore flow encounters prevailing onshore flow on the formation of the nearshore line.

Similar to the nearshore line, a zone of strong low-level convergence could be found within the offshore line (dashed line in Fig. 12b). However, it is clear the occurrence of this convergence was not relevant to the offshore flow, and instead it was caused by a pronounced decrease in the onshore flow (∼2 m s−1) from the east to the west of the line (i.e., toward the shore). No evidence of radar-derived offshore flow within the line was observed. Surface observations at Green Island and the 1000-hPa NCEP–NCAR winds (Table 4) also confirm the absence of ambient offshore flow on either side of the nearshore line as well as an ashore decrease in the onshore flow, with a stronger (much weaker) onshore flow of 6 m s−1 (1.3 m s−1) observed to the east (west) of the line.2 Note that the deceleration of the onshore flow for the offshore line, as evident in Table 4, was also accompanied by a counterclockwise wind shift from more easterly flow well offshore seen from the NCEP–NCAR data to approximately northeasterly flow nearshore observed at Green Island that was more parallel to the orientation of the coastal barrier shown in Fig. 12a. Given a weakly synoptically forced weather condition and prevailing oncoming easterly flow for the selected cases, the deflection and deceleration of winds observed off the mountainous coast of southeastern Taiwan were in a manner consistent with the influences of the upstream blocking by the coastal topography (e.g., Pierrehumbert and Wyman 1985).

The potential importance of upstream blocking effects caused by the interaction between the prevailing onshore flow and the coastal topography of southeastern Taiwan can be clearly seen in Fig. 13, which illustrates the distribution of the Froude number (Fr = U/NH, where U is the speed of the oncoming flow, N is the Brunt–Väisälä frequency, and H is the mountain height) calculated from each case of all identified convective lines as a function of the offshore distance of their formative locations. In this Froude number analysis, the upstream Brunt–Väisälä frequency and oncoming flow for each case were calculated below 1.5 km (MSL; i.e., the mean mountain height averaged along the coastal barrier in southern Taiwan) using the offshore sounding profile obtained from the NCEP–NCAR reanalysis gridded data. In Fig. 13, Froude numbers obtained from those particular events of convective lines selected for the composite analysis shown in Fig. 12 are marked by solid triangles for reference.

Figure 13 indicates Froude numbers calculated from each case of convective lines where all were less than unity and a dominant portion of the Froude numbers was actually below 0.5. In this relatively low Froude number flow regime, upstream blocking is expected to be a major atmospheric flow response to the presence of a mountain barrier (Smith 1979). This calculation may reasonably explain a clear trend of more (less) mountain-parallel winds observed at Green Island (NCEP–NCAR), as is evident in Fig. 6 and Table 4. Extremely low Froude numbers (≤0.1) could also be found in some limited cases of the nearshore and offshore convective lines. In this particular flow regime, the interaction of airflow with topography would probably be able to produce the offshore convergence zone, as the orographically induced return flow occurs near the coast and encounters prevailing oncoming flow (Smolarkiewicz et al. 1988). Note that a few nearshore cases belong to this flow regime (Fig. 13), and hence we cannot entirely rule out the terrain-induced return circulation as a potential factor that would contribute to the coastal offshore flow observed within the nearshore convective lines (cf. Fig. 12b).

Figure 13 also shows no significant variation of the mean Froude number (∼0.3) averaged with respect to the offshore distance (solid thick curve in Fig. 13). Orographic blocking would have approximately equal importance for both the nearshore and offshore cases of the convective lines, although the observed degree for the deceleration of the onshore flow as seen from Table 4 appears to be somewhat more pronounced for the offshore case. This result implies that the Froude number is not a good indicator capable of determining the preferred coastal distance for the line’s formation. Further detailed documentation of some other mesoscale aspects, such as the horizontal variations of the low-level kinematics and thermodynamics, the nature of the orographic blocking, and the diurnally forced circulations occurring within the coastal blocked zone, is crucial to clarify the relative contributions of the diurnal and orographic forcings to the initiation of the convective lines. Additional detailed observations collected over southeastern Taiwan that include both mountainous inland regions and offshore regions, plus high-resolution numerical simulations, will be required to explicitly address these scientific issues.

5. Conclusions

This study used long-term observations from the Green Island radar during 1998–2004 to document the statistical characteristics of the location and timing of the convective lines occurring off the mountainous coast of southeastern Taiwan under relatively undisturbed weather conditions. With a comprehensive examination of a large set of low-level PPI scans of reflectivity from the available radar measurements, a total of 211 cases of convective lines were identified during the study period. It is shown that the convective lines were clearly a year-round phenomenon with five cases occurring per month on average.

Investigation of the spatial distribution of the formation for the 211 identified lines revealed that the area of the most frequent formation was primarily confined to an elongated zone located ∼30 km off the southeastern coast of Taiwan. However, the formative area of the convective lines was observed to extend substantially from nearshore regions to at least ∼100 km offshore. There was also some evidence of along-coast variations for the line’s formation, with a peak located nearshore immediately adjacent to the steepest and highest mountain near the Chengkung station. These formative variations were shown to be closely related to the nearshore terrain features, and particularly the coastal area near Chengkung, the area most favored for line formation, because of its stronger and more frequent offshore flow due to terrain influences. The statistical results of formative times for all 211 convective lines indicate that they tended to form more frequently during nighttime hours than daytime hours, with a formative peak (31 cases) between 2000 and 2200 LST. Minimum formation was found near noon between 1200 and 1400 LST; however, considerable cases (>40) still could be found during the late morning and late afternoon. More than 70% of the lines (∼150 cases) had durations of less than 4 h, and the mean duration for all lines was calculated to be ∼3.5 h.

Statistical characteristics of the formative time and location for the nearshore (offshore) lines, which are defined by an offshore distance of formative location less (greater) than 40 km, were also documented. Analyses indicate that a dominant portion of the nearshore lines were found to form during nighttime hours and their formative locations were concentrated within a narrow zone of ∼10–30 km from the coast. Nevertheless, the formation of the offshore lines was distributed in a relatively uniform manner over a much wider zone of time and space and did not exhibit a clear, consistent trend in terms of diurnal variation like the nearshore lines did. Detailed composite analyses of low-level airflow patterns observed from the Green Island radar and surface stations within and adjacent to the lines further indicate that the formation of the nearshore lines was closely related to the low-level convergence produced as coastal offshore flow encountered the prevailing onshore flow. In distinct contrast, no evidence of low-level offshore flow was observed within the offshore line or its vicinity. The deceleration of the prevailing onshore flow due to upstream blocking appeared to be a major cause of line formation.

Although this study has revealed a significant role for both coastal offshore flow and orographic effects in the initiation and the spatial and temporal distributions of the convective lines occurring off the mountainous coast of southeastern Taiwan, several important scientific issues still deserve future investigation. For example, the relative importance of land breeze and thermally (and/or dynamically) driven slope circulations on the observed coastal offshore flow requires further clarification. Particularly given the low Froude numbers (a mean value of ∼0.3) characterizing the line environment (cf. Fig. 13), a dynamically produced windward-side downslope flow would probably occur over these flow regimes (e.g., Smolarkiewicz et al. 1988). Moreover, the preferred offshore distance for line formation did not seem to be easily determined by the relative magnitude of the Froude numbers calculated from each case of convective lines identified in this study. Comprehensive documentation of the nature of both the orographic blocking and the diurnally generated circulations, as well as their possible interaction occurring within the mountainous coastal zone of southeastern Taiwan under a variety of upstream conditions, will be crucial in further providing some clues for the processes controlling the formative location of the convective lines. Detailed case studies of individual offshore line events and the utility of modeling simulations should be helpful for explicitly addressing the relative contributions of the diurnal and orographic forcings to the initiation of the convective lines.

Acknowledgments

Green Island radar data used in this study were provided by the Weather Wing of the Chinese Air Force. The authors thank Mr. De-En Lin for assistance in gathering Green Island radar data and Mr. Ming-Jung Yang for assistance in providing surface and sounding data. We also thank three anonymous reviewers for their helpful comments on the manuscript. Research support was provided by the National Science Council of Taiwan under Grant NSC96-2111-M-034-001-MY3.

REFERENCES

  • Alpers, W., , J-P. Chen, , I-I. Lin, , and C-C. Lien, 2007: Atmospheric fronts along the east coast of Taiwan studied by ERS synthetic aperture radar images. J. Atmos. Sci., 64 , 922937.

    • Search Google Scholar
    • Export Citation
  • Baker, R. D., , B. H. Lynn, , A. Boone, , W-K. Tao, , and J. Simpson, 2001: The influence of soil moisture, coastline curvature, and land-breeze circulations on sea-breeze-initiated precipitation. J. Hydrol., 2 , 193211.

    • Search Google Scholar
    • Export Citation
  • Blanchard, D. O., , and R. López, 1985: Spatial patterns of convection in south Florida. Mon. Wea. Rev., 113 , 12821299.

  • Carbone, R. E., , J. W. Wilson, , T. D. Keenan, , and J. M. Hacker, 2000: Tropical island convection in the absence of significant topography. Part I: Life cycle of diurnally forced convection. Mon. Wea. Rev., 128 , 34593480.

    • Search Google Scholar
    • Export Citation
  • Chen, G. T-J., 1983: Observational aspects of the mei-yu phenomenon in subtropical China. J. Meteor. Soc. Japan, 61 , 306312.

  • Chen, G. T-J., , and J-S. Yang, 1988: On the spatial and temporal patterns of heavy rainfall in Taiwan mei-yu season. (in Chinese with English abstract). Atmos. Sci., 16 , 151162.

    • Search Google Scholar
    • Export Citation
  • Chen, G. T-J., , H-C. Chou, , T-C. Chang, , and C-S. Liu, 2001: Frontal and non-frontal convection over northern Taiwan in mei-yu season. (in Chinese with English abstract). Atmos. Sci, 29 , 3752.

    • Search Google Scholar
    • Export Citation
  • Chen, Y-L., , and J. Feng, 1995: The influences of inversion height on precipitation and airflow over the island of Hawaii. Mon. Wea. Rev., 123 , 16601676.

    • Search Google Scholar
    • Export Citation
  • Chi, S-S., , and G. T-J. Chen, 1989: Case study of the MCSs and rainfall during TAMEX/Phase I. (in Chinese with English abstract). Atmos. Sci., 17 , 5975.

    • Search Google Scholar
    • Export Citation
  • Chien, C-J., , and P-L. Lin, 2004: Numerical simulations of local circulations over complex terrain in southeastern Taiwan (in Chinese with English abstract). Proc. Eighth Conf. of Atmospheric Sciences, TaoYuan, Taiwan, National Science Council of Taiwan and Central Weather Bureau of Taiwan, 50–56.

  • Feng, J., , and Y-L. Chen, 1998: Evolution of katabatic flow on the island of Hawaii on 10 August 1990. Mon. Wea. Rev., 126 , 21852199.

  • Frye, J. L., , and Y-L. Chen, 2001: Evolution of downslope flow under strong opposing trade winds and frequent trade-wind rainshowers over the island of Hawaii. Mon. Wea. Rev., 129 , 956977.

    • Search Google Scholar
    • Export Citation
  • Garrett, A. J., 1980: Orographic cloud over the eastern slopes of Mauna Loa volcano, Hawaii, related to insolation and wind. Mon. Wea. Rev., 108 , 931941.

    • Search Google Scholar
    • Export Citation
  • Grossman, R. L., , and D. R. Durran, 1984: Interaction of low-level flow with the western Ghat Mountains and offshore convection in the summer monsoon. Mon. Wea. Rev., 112 , 652672.

    • Search Google Scholar
    • Export Citation
  • Houze Jr., R. A., , S. G. Geotis, , F. D. Marks Jr., , and A. K. West, 1981: Winter monsoon convection in the vicinity of north Borneo. Part I: Structure and time variation of the clouds and precipitation. Mon. Wea. Rev., 109 , 15951614.

    • Search Google Scholar
    • Export Citation
  • Huang, W-M., 2007: Numerical simulations of the convective lines occurring off the southeastern coast of Taiwan on 14–15 May 1998 (in Chinese). M.S. thesis, Dept. of Atmospheric Sciences, Chinese Culture University, Taipei, Taiwan, 112 pp.

  • Johnson, R. H., , and J. F. Bresch, 1991: Diagnosed characteristics of precipitation systems over Taiwan during the May–June 1987 TAMEX. Mon. Wea. Rev., 119 , 25402557.

    • Search Google Scholar
    • Export Citation
  • Keenan, T. D., , and R. E. Carbone, 1992: A preliminary morphology of precipitation systems in tropical northern Australia. Quart. J. Roy. Meteor. Soc., 118 , 283326.

    • Search Google Scholar
    • Export Citation
  • Kingsmill, D. E., 1995: Convection initiation associated with a sea-breeze front, a gust front, and their collision. Mon. Wea. Rev., 123 , 29132933.

    • Search Google Scholar
    • Export Citation
  • Leu, Y-S., , and P-L. Lin, 2004: Characteristics of local circulation and boundary layer over southeastern Taiwan. Proc. Eighth Conf. of Atmospheric Sciences, TaoYuan, Taiwan, National Science Council of Taiwan and Central Weather Bureau of Taiwan, 43–49.

  • Liberti, G. L., , F. Cheruy, , and M. Desbois, 2001: Land effect on the diurnal cycle of clouds over the TOGA COARE area, as observed from GMS IR data. Mon. Wea. Rev., 129 , 15001517.

    • Search Google Scholar
    • Export Citation
  • Mahrer, Y., , and R. A. Pielke, 1977: The effects of topography on sea and land breezes in a two-dimensional numerical model. Mon. Wea. Rev., 105 , 11511162.

    • Search Google Scholar
    • Export Citation
  • Mapes, B. E., , T. T. Warner, , and M. Xu, 2003: Diurnal patterns of rainfall in northwestern South America. Part III: Diurnal gravity waves and nocturnal convection offshore. Mon. Wea. Rev., 131 , 830844.

    • Search Google Scholar
    • Export Citation
  • Murakami, M., 1983: Analysis of the deep convective activity over the western Pacific and Southeast Asia. Part I: Diurnal variation. J. Meteor. Soc. Japan, 61 , 6075.

    • Search Google Scholar
    • Export Citation
  • Nettleton, L., , S. Daud, , R. Neitzel, , C. Burghart, , W-C. Lee, , and P. Hildebrand, 1993: SOLO: A program to peruse and edit radar data. Preprints, 26th Conf. on Radar Meteorology, Norman, OK, Amer. Meteor. Soc., 338–339.

  • Pierrehumbert, R. T., , and B. Wyman, 1985: Upstream effects of mesoscale mountains. J. Atmos. Sci., 42 , 9771003.

  • Powell, M. D., 1982: The transition of the Hurricane Frederic boundary-layer wind field from the open Gulf of Mexico to landfall. Mon. Wea. Rev., 110 , 19121932.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., , and P. K. Smolarkiewicz, 1993: On the dynamics of Hawaiian cloud bands. Part III: Local aspects. J. Atmos. Sci., 50 , 15601572.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., , P. K. Smolarkiewicz, , and J. Warner, 1989: On the dynamics of Hawaiian cloud bands: Comparison of model results with observations and island climatology. J. Atmos. Sci., 46 , 15891608.

    • Search Google Scholar
    • Export Citation
  • Roeloffzen, J. C., , W. D. van den Berg, , and J. Oerlemans, 1986: Frictional convergence at coastlines. Tellus, 38A , 397411.

  • Rotunno, R., , J. B. Klemp, , and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463485.

  • Smith, R. B., 1979: The influence of mountains on the atmosphere. Advances in Geophysics, Vol. 21, Academic Press, 87–230.

  • Smolarkiewicz, P. K., , R. M. Rasmussen, , and T. L. Clark, 1988: On the dynamics of Hawaiian cloud bands: Island forcing. J. Atmos. Sci., 45 , 18721905.

    • Search Google Scholar
    • Export Citation
  • Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.

  • Sun, W. Y., , and J-D. Chern, 1993: Diurnal variation of lee vortices in Taiwan and the surrounding area. J. Atmos. Sci., 50 , 34043430.

    • Search Google Scholar
    • Export Citation
  • Wai, M. M-K., , P. T. Welsh, , and W-M. Ma, 1996: Interaction of secondary circulations with the summer monsoon and diurnal rainfall over Hong Kong. Bound.-Layer Meteor., 81 , 123146.

    • Search Google Scholar
    • Export Citation
  • Willoughby, H. E., , and P. G. Black, 1996: Hurricane Andrew in Florida: Dynamics of a disaster. Bull. Amer. Meteor. Soc., 77 , 543549.

  • Wilson, J. W., , and D. L. Megenhardt, 1997: Thunderstorm initiation, organization, and lifetime associated with Florida boundary layer convergence lines. Mon. Wea. Rev., 125 , 15071525.

    • Search Google Scholar
    • Export Citation
  • Yu, C-K., , and B. J-D. Jou, 2005: Radar observations of the diurnally forced offshore convective lines along the southeastern coast of Taiwan. Mon. Wea. Rev., 133 , 16131636.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Topography in southern Taiwan. Terrain height (m MSL) is indicated by shading. Location of the Doppler radar site at Green Island is denoted by the triangle. Locations of select surface observing stations (Green Island, GI; Chengkung, CK; Taitung, TT; Tawu, TW; and Lanyu, LY) and the NCEP–NCAR gridded data are denoted by open circles and square, respectively.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 2.
Fig. 2.

Low-level PPI scans (∼1.6° elevation) of reflectivity (dBZ) from the GI radar from four selected cases of convective lines at (a) 0331 LST 10 Nov 2004, (b) 0331 LST 11 Jan 2000, (c) 0621 LST 17 Dec 2002, and (d) 0531 LST 9 Nov 2004.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 3.
Fig. 3.

Maximum (upper curve) and mean (lower curve) reflectivities of each identified convective line obtained from the low-level PPI scan (1° ∼ 1.6° elevations) of the GI radar during their most organized stage. Horizontal thick lines indicate averaged values for the maximum and mean reflectivities. Vertical dashed lines mark the borders of different seasons.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 4.
Fig. 4.

The spatial distribution of the formative frequency of convective lines (color shading) occurring off the southeastern coast of Taiwan, derived from the 1°-elevation PPI scans of radar reflectivity from the GI radar during 1998–2004. Topography in southern Taiwan is indicated with gray shading (key at left). Locations of select surface observing stations, as in Fig. 1, are also indicated for reference.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 5.
Fig. 5.

(a) The distribution of correlation coefficients between the maximum coastal terrain height and the formative frequency of the convective lines as a function of inland distance along the southeastern coast of Taiwan. (b) Along-coast profile of maximum terrain height (shading) obtained within the 5-km-inland distance, and its corresponding profile of maximum formative frequency of convective lines (solid curve) found offshore. Relative locations of three coastal surface stations, as in Fig. 1, along this terrain profile are also indicated for reference.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 6.
Fig. 6.

Time series of mean surface winds averaged during the formative days of the convective lines from three coastal surface stations and two offshore locations, as in Fig. 1. The offshore (onshore) flow component defined as the wind component perpendicular to the mean orientation of the coastline (∼25° clockwise from the north) is indicated by the solid (dashed) line with the positive value representing the wind toward the ocean. In each panel, full wind barbs correspond to 5 m s−1, half barbs to 2.5 m s−1.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 7.
Fig. 7.

Number of cases of convective lines at different time intervals during a day for all 211 convective lines identified in this study.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 8.
Fig. 8.

Number of cases of convective lines at their different durations.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 9.
Fig. 9.

The mean environmental wind profile averaged from all identified cases of convective lines derived from the NCEP–NCAR gridded data located ∼130 km off the southeastern coast of Taiwan. Here, U (solid line) and V (dashed line) denote the cross-line and along-line wind components, respectively. Since convective lines are approximately parallel to the coast, the value of U is roughly identical to the offshore–onshore flow component, with the positive value representing the wind toward the ocean. Full wind barbs correspond to 5 m s−1, half barbs to 2.5 m s−1.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 10.
Fig. 10.

(a) As in Fig. 7 but for the nearshore convective lines and (b) for the offshore convective lines.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 11.
Fig. 11.

Distribution of statistical frequencies of the convective lines over different time intervals and offshore distances. Formative frequencies are indicated by shading and also contoured with a two-case interval. The thick dashed horizontal line marks the offshore distance at 40 km.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 12.
Fig. 12.

(a) Locations of line segments used to extract reflectivity and radial velocity data from the low-level PPI scan (0.5°– 1.6° elevations) of the GI radar for the composite analysis shown in (b). Locations of TT, GI, and NCEP, as in Fig. 1, are also indicated. (b) Composite radial velocities (m s−1) within the nearshore (solid line) and offshore (dashed line) convective lines, with positive (negative) values representing the offshore (onshore) flow. The strongest reflectivity found within lines is located at X = 0 km.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Fig. 13.
Fig. 13.

Distribution of Froude numbers calculated from each identified case of convective lines as a function of the offshore distance of their formative location. The events of the convective lines selected for the composite analysis shown in Fig. 12 are marked by solid triangles for reference. The thick solid line denotes the mean Froude number calculated with respect to a given offshore distance.

Citation: Monthly Weather Review 136, 12; 10.1175/2008MWR2555.1

Table 1.

Monthly number of cases of convective lines identified in this study. Monthly available GI radar observations (days) during 1998–2004 are also indicated.

Table 1.
Table 2.

A complete list of formative dates for the convective lines occurring off the southeastern coast of Taiwan identified by the GI radar during the study period from 1998 to 2004. Note that in certain circumstances two separate events could occur on one day, and the date indicated by two successive days means that the duration of that event was across 0000 LST. Nearshore (offshore) cases selected for the composite analysis presented in section 4 are set in boldface (italics).

Table 2.
Table 3.

Percentages of the occurrences of offshore flow observed at three coastal surface stations calculated from the formative days of the convective lines. The mean magnitudes of the offshore flows calculated from three different percentages (i.e., during a full day, daytime, and nighttime) are also indicated.

Table 3.
Table 4.

Mean offshore (onshore) flows [positive (negative) value denoting offshore (onshore) flow] calculated from measured surface winds (locations in Fig. 12a) during the occurrence of the nearshore and offshore cases of convective lines selected for composite analysis of the radial velocities shown in Fig. 12b. For reference, the mean wind direction and speed are shown in parentheses.

Table 4.
1

Examination of the radar observations indicates that the convective lines identified in this study generally moved very slowly or were even quasi-stationary. Hence, it is reasonable that the inflow side refers to the east of the line, given the prevailing easterly onshore flow at low levels.

2

Given a similar altitude for the GI surface station and the 1000-hPa isobar surface, and a rather weak mean vertical cross-coast shear in the line’s environment in the lowest 1 km (cf. Fig. 9), comparisons of winds obtained from them should be able to provide a meaningful signal for the low-level flow alteration.

Save