• Asai, T., 1988: Meso-scale features of heavy snowfalls in Japan Sea coastal regions of Japan (in Japanese). Tenki,35, 156–161.

  • Atlas, D., S.-H. Chou, and W. P. Byerly, 1983: The influence of coastal shape on winter mesoscale air–sea interaction. Mon. Wea. Rev.,111, 245–252.

  • Bond, N. A., and M. A. Shapiro, 1991: Polar lows over the Gulf of Alaska in conditions of reverse shear. Mon. Wea. Rev.,119, 551–572.

  • Crook, N. A., T. L. Clark, and M. W. Moncrieff, 1990: The Denver cyclone. Part I: Generation in low Froude number flow. J. Atmos. Sci.,47, 2725–2742.

  • Douglas, M. W., L. S. Fedor, and M. A. Shapiro, 1991: Polar low structure over the northern Gulf of Alaska based on research aircraft observations. Mon. Wea. Rev.,119, 32–54.

  • Forbes, G. S., and W. D. Lottes, 1985: Classification of mesoscale vortices in polar airstreams and the influence of the large-scale environment on their evolutions. Tellus,37A, 132–155.

  • Lee, T.-Y., and Y.-Y. Park, 1996: Formation of a mesoscale trough over the Korean peninsula during an excursion of Siberian high. J. Meteor. Soc. Japan,74, 299–323.

  • Lin, Y.-L., and D. J. Perkey, 1989: Numerical modeling of a process of lee cyclogenesis. J. Atmos. Sci.,46, 3685–3697.

  • ——, N.-H. Lin, and R. P. Weglarz, 1992: Numerical modeling studies of lee mesolows, mesovortices and mesocyclones with application to the formation of Taiwan mesolows. Meteor. Atmos. Phys.,49, 43–67.

  • Mansfield, D. A., 1974: Polar lows: The development of baroclinic disturbances in cold air outbreaks. Quart. J. Roy. Meteor. Soc.,100, 541–554.

  • Nagata, M., 1991: Further numerical study on the formation of the convergent cloud band over the Japan Sea in winter. J. Meteor. Soc. Japan,69, 419–427.

  • ——, 1993: Meso-β-scale vortices developing along the Japan-Sea Polar-Airmass Convergence Zone (JPCZ) cloud band: Numerical simulation. J. Meteor. Soc. Japan,71, 43–57.

  • ——, M. Ikawa, S. Yoshizumi, and T. Yoshida, 1986: On the formation of a convergent cloud band over the Japan Sea in winter:Numerical experiments. J. Meteor. Soc. Japan,64, 841–855.

  • Ninomiya, K., 1989: Polar/comma-cloud lows over the Japan Sea and the northwestern Pacific in winter. J. Meteor. Soc. Japan,67, 83–97.

  • ——, 1991: Polar low development over the east coast of the Asian continent on 9–11 December 1985. J. Meteor. Soc. Japan,69, 669–685.

  • Rasmussen, E., 1979: The polar low as an extratropical CISK disturbance. Quart. J. Roy. Meteor. Soc.,105, 531–549.

  • Smith, R. B., 1984: A theory of lee cyclogenesis. J. Atmos. Sci.,41, 1159–1168.

  • ——, 1986: Further development of a theory of lee cyclogenesis. J. Atmos. Sci.,43, 1582–1602.

  • ——, 1989: Comment on the “Low Froude number flow past three-dimensional obstacles. Part I: Baroclinically generated lee vortices.” J. Atmos. Sci.,46, 3611–3613.

  • Smolarkiewicz, P. K., and R. Rotunno, 1989a: Low Froude number flow past three-dimensional obstacles. Part I: Baroclinically generated lee vortices. J. Atmos. Sci.,46, 1154–1164.

  • ——, and ——, 1989b: Reply. J. Atmos. Sci.,46, 3614–3617.

  • ——, R. Rasmussen, and T. L. Clark, 1988: On the dynamics of Hawaiian cloud bands: Island forcing. J. Atmos. Sci.,45, 1872–1905.

  • Sun, W.-Y., and J.-D. Chern, 1993: Diurnal variation of lee vortices in Taiwan and the surrounding area. J. Atmos. Sci.,50, 3404–3430.

  • ——, ——, C.-C. Wu, and W.-R. Hsu, 1991: Numerical simulation of mesoscale circulation in Taiwan and surrounding area. Mon. Wea. Rev.,119, 2558–2573.

  • Tremback, C. J., 1990: Numerical simulation of a mesoscale convective complex: Model development and numerical results. Department of Atmospheric Science Paper 465, Colorado State University, 247 pp.

  • ——, and R. Kessler, 1985: A surface temperature and moisture parameterization for use in mesoscale numerical models. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 355–358.

  • ——, G. J. Tripoli, and W. R. Cotton, 1985: A regional scale atmospheric numerical model including explicit moist physics and a hydrostatic time-split scheme. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 433–434.

  • Tripoli, G. J., and W. R. Cotton, 1982: The Colorado State University three-dimensional cloud/mesoscale model—1982. Part I: General theoretical framework and sensitivity experiments. J. Rech. Atmos.,16, 185–220.

  • Tsuboki, K., and G. Wakahama, 1992: Mesoscale cyclogenesis in winter monsoon air streams: Quasi-geostrophic baroclinic instability as a mechanism of the cyclogenesis off the west coast of Hokkaido Island, Japan. J. Meteor. Soc. Japan,70, 77–93.

  • Yagi, S., T. Muramatsu, T. Uchiyama, and N. Kurokawa, 1986: “Convergent band cloud” and “Cu-Cb line” over Japan Sea affected by topographic features in the coast of the Asian continent (in Japanese). Tenki,33, 453–465.

  • View in gallery

    Geographic map and smoothed topography in the mesoscale model domain. Contour interval is 200 m. The KMC represents the northern Korean mountain complex, located in the northern part of the Korean peninsula.

  • View in gallery

    Sea level pressure (hPa) for (a) 1200 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

  • View in gallery

    GMS satellite IR imagery for (a) 2100 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995.

  • View in gallery

    Height (m, solid) and temperature (°C, dashed) fields at 500 hPa: (a) 1200 UTC 28 and (b) 1200 UTC 29 January 1995. At 850 hPa: (c) 1200 UTC 28 and (d) 1200 UTC 29 January 1995.

  • View in gallery

    Difference of potential temperature between the 500- and 850-hPa levels (Δθ, K) at (a) 1200 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

  • View in gallery

    (a) Geopotential height (m, solid line) and absolute vorticity (10−5 s−1, dashed line) at the 500-hPa level, and (b) the Q vector and its divergence (·Q, 10−16 km−2 s−1) at the 850-hPa level at 1200 UTC 29 January 1995. Contour intervals for absolute vorticity and Q-vector divergence are 2 × 10−5 s−1 and 4 × 10−16 km−2 s−1, respectively. (b) Solid and dashed lines indicate positive and negative values, respectively.

  • View in gallery

    Simulated sea level pressure (hPa) at (a) 1800 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995. Terrain heights of 100, 500, 1000, and 1500 m are shaded with increasing darkness, respectively.

  • View in gallery

    Simulated streamlines at z* = 95 m for (a) 1200 UTC 28, (b) 1800 UTC 28, (c) 0000 UTC 29, and (d) 1200 UTC 29 January 1995.

  • View in gallery

    Same as Fig. 8, except at 850 hPa.

  • View in gallery

    The y–z cross sections of (a) equivalent potential temperature (K, solid) and cloud water mixing ratio (rc, dashed), and (b) υw wind vector fields at x = −100 km for 0000 UTC 29 January 1995. The isoline of rc starts from 0.1 g kg−1 with an interval of 0.1 g kg−1.

  • View in gallery

    Simulated cloud water mixing ratio (10−2 g kg−1) where z* equals (a) 2402 m and (b) 4019 m for 1200 UTC 29 January 1995. Contour interval is 0.2 g kg−1.

  • View in gallery

    Streamlines at 850 hPa obtained using the wind analysis data from the Japan Meteorological Agency for (a) 1800 UTC 28 and (b) 1200 UTC 29 January 1995.

  • View in gallery

    (a) Simulated streamlines at z* = 1315 m, (b) vertical velocity (cm s−1) at z* = 1548 m, and (c) cloud water mixing ratio (10−2 g kg−1) at z* = 2402 m, respectively. Solid and dashed lines in (b) indicate positive and negative values, respectively, with contour interval of 5 cm s−1.

  • View in gallery

    Changes in (a) sea level pressure and (b) average potential temperature in the lowest 444-m layer during the 3-h period from 1800 to 2100 UTC 28 January 1995. Solid and dashed lines indicate positive and negative values, respectively. Contour interval is 0.4 hPa (3h)−1 in panel (a) and 0.5 K (3h)−1 in panel (b).

  • View in gallery

    Simulated fields of (a) potential temperature (K) and (b) horizontal wind vector at z* = 585 m.

  • View in gallery

    Relative vorticity, stretching term, and tilting term are shown in (a), (c), and (e), respectively, for z* = 585 m, and in (b), (d), and (f), respectively, for z* = 1315 m. Contour interval is 20 × 10−6 s−1 in (a) and (b), and 50 × 10−10 s−2 in (c) through (f).

  • View in gallery

    Streamlines at 1315 m from expts (a) I1, (b) I2, and (c) I3.

  • View in gallery

    Sea level pressure perturbations (hPa, upper panels) and potential temperatures (K, lower panels) at 585 m from expts I1 (a,b); I2 (c,d); and I3 (e,f).

  • View in gallery

    Same as Fig. 17 except for I5 at three different hours.

  • View in gallery

    Simulated sea level pressure (hPa) from expt E1, in which the mountains over the northern Korean peninsula are removed, for (a) 1800 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995.

  • View in gallery

    Simulated streamlines at 850 hPa from E1 for (a) 1800 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

  • View in gallery

    Differences between RE and E1 (RE − E1) in (a) sea level pressure (hPa) and (b) wind vectors at 95 m over the ocean at 1800 UTC 28 January 1995. Contour interval in (a) is 0.5 hPa.

  • View in gallery

    Changes in (a) sea level pressure [hPa (3 h)−1] and (b) temperature [K (3 h)−1] during the 3-h period from 1200 to 1500 UTC 28 January 1995. Contour intervals in (a) and (b) are 0.5 hPa (3 h)−1 and 0.5 K (3 h)−1, respectively.

  • View in gallery

    Simulated sea level pressure fields from expt E2, in which the latent and sensible heat fluxes at the surface are assumed to be 0, at (a) 1200 UTC 28 and (b) 1200 UTC 29 January 1995.

  • View in gallery

    Difference between expts RE and E2 (RE − E2) in (a) sea level pressure (hPa) and (b) wind vector at 95 m at 0000 UTC 29 January 1995.

  • View in gallery

    (a) Simulated sea level pressure from E3, in which the condensational heating is not considered, and (b) the difference between the experiments RE and E3 (RE − E3) in the potential temperature difference (Δθ, K) for 1200 UTC 29 January 1995.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 43 39 4
PDF Downloads 13 11 0

A Numerical Modeling Study of Mesoscale Cyclogenesis to the East of the Korean Peninsula

View More View Less
  • 1 Department of Atmospheric Sciences, Yonsei University, Seoul, Korea
  • | 2 Department of Marine, Earth and Atmospheric Sciences, North Carolina State University, Raleigh, North Carolina
© Get Permissions
Full access

Abstract

Numerical simulations and the analysis of observational data are employed to understand the mesoscale cyclogenesis in a polar airstream that occurred over the sea to the east of the Korean peninsula on 28–29 January 1995. The observational analysis shows that a mesoscale low develops over the southeastern East Sea (Japan Sea) on 29 January 1995. Satellite imagery also indicates that a meso-β-scale vortex forms on the lee side of the northern Korean mountain complex (KMC), which is located in the northern Korean peninsula, and that a meso-α-scale cyclone develops over the southeastern East Sea at a later time. The mesoscale cyclone forms in the lower troposphere with strong baroclinicity and cyclonic circulation under the influence of an upper-level synoptic-scale cold vortex.

Numerical simulation has captured major features of the observed cyclogenesis very well. The cyclogenesis occurs in a progressive manner. Basically, four distinctive stages of the cyclogenesis are identified. 1) First, a surface pressure trough forms on the lee side of the KMC under a northwesterly synoptic-scale flow that is deflected anticyclonically over the KMC. 2) Second, the lee trough deepens further into a strong convergence zone and a meso-β-scale vortex. 3) Next, the meso-β-scale vortex develops into a meso-α-scale vortex as the vortex and the trough begin to move southeastward from the lee of the KMC. 4) Finally, the surface trough deepens into a closed low and the meso-α-scale vortex becomes collocated with this deepening surface low to form a meso-α-scale cyclone over the southeastern East Sea.

Several sensitivity experiments are performed to isolate the effects of a topography, warmer sea surface, diurnal thermal forcing, and latent heat release. During stages 1 and 2, it is found that the KMC and low-level baroclinicity are responsible for generating the strong lee trough and vortex. During stage 3, the development of the meso-α-scale vortex is brought on by the tilting of horizontal vorticity and vertical stretching in a synoptic-scale cyclonic circulation. In the final stage, the condensational heating plays the key role for the development of the meso-α-scale cyclone under the influence of an upper-level synoptic-scale cold vortex. The presence of the warm sea surface is found to be a necessary condition for the development of a polar air convergence zone and the mesoscale cyclone. It is also found that the low-level baroclinicity is essential for the present case of mesoscale cyclogenesis.

Corresponding author address: Dr. Tae-Young Lee, Department of Atmospheric Sciences, Yonsei University, Seoul 120-749, Korea.

Email: lty@atmos.yonsei.ac.kr

Abstract

Numerical simulations and the analysis of observational data are employed to understand the mesoscale cyclogenesis in a polar airstream that occurred over the sea to the east of the Korean peninsula on 28–29 January 1995. The observational analysis shows that a mesoscale low develops over the southeastern East Sea (Japan Sea) on 29 January 1995. Satellite imagery also indicates that a meso-β-scale vortex forms on the lee side of the northern Korean mountain complex (KMC), which is located in the northern Korean peninsula, and that a meso-α-scale cyclone develops over the southeastern East Sea at a later time. The mesoscale cyclone forms in the lower troposphere with strong baroclinicity and cyclonic circulation under the influence of an upper-level synoptic-scale cold vortex.

Numerical simulation has captured major features of the observed cyclogenesis very well. The cyclogenesis occurs in a progressive manner. Basically, four distinctive stages of the cyclogenesis are identified. 1) First, a surface pressure trough forms on the lee side of the KMC under a northwesterly synoptic-scale flow that is deflected anticyclonically over the KMC. 2) Second, the lee trough deepens further into a strong convergence zone and a meso-β-scale vortex. 3) Next, the meso-β-scale vortex develops into a meso-α-scale vortex as the vortex and the trough begin to move southeastward from the lee of the KMC. 4) Finally, the surface trough deepens into a closed low and the meso-α-scale vortex becomes collocated with this deepening surface low to form a meso-α-scale cyclone over the southeastern East Sea.

Several sensitivity experiments are performed to isolate the effects of a topography, warmer sea surface, diurnal thermal forcing, and latent heat release. During stages 1 and 2, it is found that the KMC and low-level baroclinicity are responsible for generating the strong lee trough and vortex. During stage 3, the development of the meso-α-scale vortex is brought on by the tilting of horizontal vorticity and vertical stretching in a synoptic-scale cyclonic circulation. In the final stage, the condensational heating plays the key role for the development of the meso-α-scale cyclone under the influence of an upper-level synoptic-scale cold vortex. The presence of the warm sea surface is found to be a necessary condition for the development of a polar air convergence zone and the mesoscale cyclone. It is also found that the low-level baroclinicity is essential for the present case of mesoscale cyclogenesis.

Corresponding author address: Dr. Tae-Young Lee, Department of Atmospheric Sciences, Yonsei University, Seoul 120-749, Korea.

Email: lty@atmos.yonsei.ac.kr

1. Introduction

Satellite imagery often shows the development of meso-α-scale lows over the ocean to the east of the Asian continent during winter outbreaks of polar air (Ninomiya 1989; Tsuboki and Wakahama 1992). Some of these lows are found over the East Sea (Japan Sea) between the northeastern coast of the Korean peninsula and the west coast of Japan. Once these lows form, some of them keep developing when they are moving northeastward along the west coast of Japan. Although the development of the mesoscale lows over the ocean around northern Japan have been studied extensively, the genesis mechanism of these mesoscale lows to the east of the Korean peninsula is still not well understood and deserves further study.

Ninomiya (1989) found that polar lows were observed to form about 500–1000 km north of major polar frontal zones, where strong low-level baroclinicity is maintained by sea surface fluxes in the polar air mass between the continent and the relatively warm ocean. He defined the polar low in this region as a meso-α-scale low accompanied by spiral or comma cloud system with a scale of 200–700 km. He also found that these mesoscale lows rarely appear over the Asian continent, Yellow Sea, and East China Sea.

As the polar air streams out over the ocean from the Asian continent, surface weather maps often show a sharp trough that starts from immediately to the lee side of the mountains in the northern Korean peninsula, that is, the northern Korean mountain complex (referred to as KMC hereafter) (Fig. 1), and extends toward the ocean. Satellite imagery sometimes shows mesoscale vortices that form to the southeast of the KMC and propagate southeastward. These characteristics suggest that the topography of the Korean peninsula may play some significant role in the mesoscale cyclogenesis to the east of the peninsula.

Previous studies have revealed the importance of the KMC on the mesoscale disturbances around the peninsula. Yagi et al. (1986) indicated that the low-level polar air convergence zone to the east of Korea may result from the dynamic effect of the KMC. Nagata (1991) suggested that the blocking effect of the KMC, the land–sea thermal contrast, and the characteristic SST distribution equally contribute to the formation of the convergence zone. Asai (1988) showed that zones of frequent occurrence of mesoscale vortices were found over the sea to the east of the Korean peninsula and to the west of Hokkaido Island, Japan, and indicated that these zones were collocated with the polar air convergence zone. Nagata (1993) proposed that barotropic shear instability was the dominant development mechanism of meso-β-scale vortices along this zone. Direct effects of the KMC on the formation of mesoscale cyclones over the East Sea, however, are rarely discussed in the previous studies.

The KMC consists of various peaks with heights greater than 2 km, and the length and width of the area above 1-km height are about 320 and 100–160 km, respectively. The shape of the KMC is asymmetric with a steeper slope to the east coast. The Froude number for a typical flow during night around the KMC is about 0.3–0.7, which is sufficiently low for the mountains to affect the flow significantly according to the previous studies [e.g., Smolarkiewicz and Rotunno (1989a,b); Lin et al. (1992)]. Lee and Park (1996) indicated that the KMC may be directly responsible for the formation of some mesoscale disturbances around the Korean peninsula.

In simulating an inviscid flow over the Hawaiian mountains, Smolarkiewicz et al. (1988) found that a pair of vortices formed on the lee side of the island, which then shed downstream at later times. Smolarkiewicz and Rotunno (1989a,b) found that this type of lee vortex may occur when a low Froude number, inviscid, nonrotating stratified flow passes over an isolated mountain. According to Smolarkiewicz and Rotunno, the Froude number, defined as U/Nh, where U is the basic wind speed, N the Väisälä frequency, and h the mountain height, needs to be less than 0.5 in order to produce the lee vortices. Higher Froude number flows have also been found to be able to produce lee vortices (Lin et al. 1992). The lee vortex is formed by either the baroclinically generated vorticity (Smolarkiewicz and Rotunno 1989a,b) or the generation of potential vorticity (Smith 1989). These types of lee vortices have also been simulated in numerical experiments of the Denver cyclone (Crook et al. 1990) and the Taiwan mesolow (Sun et al. 1991; Lin et al. 1992). In the Taiwan case, Lin et al. (1992) also found that the cyclonic vortex collocates with the mesolow in a rotating fluid flow system and may therefore be classified as a mesocyclone.

Smith (1984, 1986) proposed a theory which views lee cyclogenesis as the formation of the first trough of a standing baroclinic wave. The theory requires that the basic wind reverses its direction at a certain level. This wind reversal height (i.e., critical level in a steady-state flow) satisfies the general conditions observed to accompany lee cyclogenesis in the Alps, since lee cyclogenesis is often associated with the passage of a cold front there. This theory has also been applied to explain cyclogenesis in the lee of the Appalachians (Smith 1986). This type of lee cyclogenesis is found to be more effective in a nonlinear flow (Lin and Perkey 1989). The splitting of low-level flow is more pronounced for a low Froude number flow. The ageostrophic advection of cold air is able to strengthen the mountain-induced high and the lee cyclone. In addition, lee cyclogenesis is strengthened by both the low-level sensible heating and the turning of the wind associated with boundary layer processes. Therefore, the presence of low-level baroclinicity across the east coast of the Korean peninsula and the boundary layer forcing associated with the warm sea surface may play a similar role in the formation of the lee cyclone over the KMC.

Mechanisms for the formation and development of meso-α-scale lows have been a subject of considerable interest, especially for the polar lows over the North Atlantic Ocean, the Norwegian Sea, the Barent Sea, and the Gulf of Alaska, as well as the lows around Japan. Mansfield (1974) suggested that polar lows were shallow baroclinic disturbances, whereas Rasmussen (1979) viewed the polar low as an extratropical disturbance driven by conditional instability of the second kind (CISK). Later, Forbes and Lottes (1985) suggested that both the baroclinic and CISK mechanisms were important for polar low development. Based on observations, Bond and Shapiro (1991) found that mesoscale cyclogenesis exists in the large-scale parent low over the Gulf of Alaska and suggested frontogenesis at low levels as a polar low genesis mechanism. Douglas et al. (1991) suggested that the observed evolution of the polar low over the Gulf of Alaska may be significantly influenced by 1) flow modification by the high mountains ringing the Gulf of Alaska, 2) the varying synoptic-scale flow over the Gulf of Alaska, and 3) heat and moisture fluxes from the underlying ocean surface.

Ninomiya (1991) suggested that a similar mechanism, as proposed by Bond and Shapiro (1991), was responsible for the polar low genesis to the east of the Asian continent. He suggested that the mesoscale low over the east coast of Asia formed, under the influence of a cold vortex aloft, in the west–east-oriented trough within the northwestern quadrant of the synoptic-scale low that developed over the northwestern Pacific. Tsuboki and Wakahama (1992) suggested that the meso-α-scale cyclones off the west coast of Hokkaido Island, Japan, were due to baroclinic instability associated with a particular baroclinic flow. It appears that the formation mechanism of the meso-α-scale cylone or polar low in this region is extremely complicated and deserves further study. In this study, we will focus on the earlier stages of meso-α-scale cyclone formation to the east of the Korean peninsula.

The meso-α-scale cyclone presented in this study occurred on 29 January 1995 over the southeastern East Sea in the vicinity of the polar air convergence zone. In this study, we will analyze the mesoscale cyclogenesis event and discuss its mechanism. In section 2, we describe the formation and development of a mesoscale vortex to the east of the peninsula based on synoptic analyses and satellite imagery. Several numerical experiments are performed to investigate the observed mesoscale cyclogenesis. The description of the numerical experiments is given in section 3. The results from the control experiment and the comparison of its results with observations are presented in section 4. In section 5, results from five idealized experiments and three sensitivity experiments with real data are discussed. Concluding remarks can be found in section 6.

2. Observational analysis

The mesoscale cyclogenesis of the present case is observed on 28–29 January 1995 across the sea between the KMC (northern Korean mountain complex) and the west coast of Japan. Figure 2 shows the sea level pressure (SLP) patterns for the period of 1200 UTC 28–1200 UTC 29 January 1995. At 1200 UTC 28 January, two mesoscale troughs are found, one over the Sakhalin Islands and the other over the northeastern coast of the Korean peninsula. A mesoscale low is found off the midwest coast of Japan. The pressure gradient is weak throughout the eastern Asian continent and also over the area to the west of the Korean peninsula. A significant trough has developed over Japan by 0000 UTC 29 January. At 1200 UTC 29 January, another mesoscale low has developed off the midwest coast of Japan (near 39°N, 138°E), at the location similar to that of the mesoscale low at 1200 UTC 28 January. The present study is interested in the formation of this mesoscale low. The pressure drops about 8 hPa during the period of 0000–1200 UTC 29 January over the area of this low. With the development of the mesoscale low, the zonal gradient of surface pressure has also significantly increased over the sea between the Korean peninsula and the mesoscale low. Detailed surface pressure analysis over the northern Korean peninsula shows that a low pressure area is persistently found over the northeastern coast of the peninsula, although the synoptic charts do not show this feature well.

Satellite imagery shows that a mesoscale vortex develops during 1500–2100 UTC 28 January to the southeast of the KMC (Fig. 3a). The line of convective clouds to the southeast of the vortex indicate the existence of an elongated low-level convergence in a northwest–southeast direction (Figs. 3a and 3b). The relationship between the band of convective cloud and the lower-level convergence has been investigated by several studies [e.g., Yagi et al. (1986); Nagata et al. (1986)]. Hereafter, this cloud band will be called the convergent cloud band following Nagata et al. (1986). During this early stage, the lee vortex moves slowly southeastward (Fig. 3b).

The imagery shows the movement of the convergent cloud band. At 0600 UTC 29 January, the convergent cloud band is found between the northeastern coast of the peninsula and the west coast of Japan. A vertex point of the band is located near 38.2°N, 130.8°E at 0600 UTC. It keeps moving southeastward, and reaches the point near 36.8°N, 133.7°E at 1200 UTC 29 January. The speed of movement is faster during 0300–1200 UTC 29 January than before 0300 UTC. High-level clouds are found at 0600 UTC around 40°N, 135°E far to the north of the cloud band. These high clouds have been advected from the area to the southeast of Vladivostok (43.1°N, 131.9°E) and from the southwest. The clouds are then advected northeastward.

A well-developed meso-α-scale cyclonic circulation is found over the southeastern part of the East Sea at 1200 UTC 29 January (Fig. 3d). This cyclonic circulation is over the convergence zone, and is located to the southwest of the mesoscale low shown in the corresponding surface map (Fig. 2c). The relatively large cloud mass to the northeast of the cyclone consists of high-level thin clouds advected from the west and the clouds associated with the mesoscale low. The cloud-top temperature over the cyclone area (southern part of the cloud mass) ranges from −33° to −37°C. This range of cloud-top temperatures indicates that the clouds associated with the meso-α-scale cyclone are below the 500-hPa level, where the air temperature over the cyclone area ranges from −35° to −40°C.

Satellite imagery can be useful for understanding the movement of the mesoscale pressure system over the sea. The satellite observations described above indicate that the mesoscale cyclone is associated with the polar air convergence zone that extends with a V-shaped pattern from the lee of the KMC to the mesoscale cyclone. The vertex point of the V-shaped convergent cloud band has moved southeastward from a point near the vortex formation area in the lee of the KMC. These satellite observations may indicate that the development of the mesoscale trough to the west of central Japan at 1200 UTC 29 January (Fig. 2c) is associated with the southeastward-moving band of polar air convergence.

The 500-hPa chart shows the presence of a cold-core cutoff low over eastern Manchuria at 1200 UTC 28 January (Fig. 4). This low is collocated vertically with the low at 850 hPa. A synoptic-scale ridge is found to the east of the low. The low at 500 hPa moves slowly southeastward and becomes more symmetric by 1200 UTC 29 January, when the ridge to the east becomes stronger. The low at 850 hPa has also moved southeastward and shows two mesoscale lows around Hokkaido Island, Japan, at 1200 UTC 29 January. During this 24-h period, the 850-hPa airflow around the northern Korean peninsula has changed from weak northwesterlies to stronger northerlies.

The 850-hPa thermal trough develops toward the peninsula as the 500-hPa cold-core low moves southeastward during the 24-h period. At the 850-hPa level, temperature decreases are found over the northern peninsula and most of the East Sea, except for the northeastern part of the sea where a temperature increase is found. On the other hand, the temperature decrease at the 500-hPa level is relatively large (4–5 K) over the middle part of the sea due to the southeastward movement of the 500-hPa cold-core low.

These changes in temperature at the 850- and 500-hPa levels result in the decrease of atmospheric static stability over the middle and northeastern part of the East Sea. Figure 5 shows the potential temperature difference (Δθ) between the 500- and 850-hPa levels obtained using the 2.5° × 2.5° analysis data from the Japan Meteorological Agency (JMA). At 1200 UTC 28 January, the minimum difference is found to the north of the surface mesoscale low off the midwest coast of Japan. The difference increases moderately over the eastern part of the sea during the next 12 h. Then a significant decrease occurs during 0000–1200 UTC 29 January over the middle and eastern parts of the sea. The relatively small difference at 1200 UTC 29 January over the northeastern part may be due to the combined effect of the southeastward movement of the upper-level cold vortex and lower tropospheric heating associated with the mesoscale low. The contribution of the latter part will be discussed further in section 5d.

Figure 6 shows the 500-hPa absolute vorticity and geopotential height and the 850-hPa divergence of Q vector for 1200 UTC 29 January. A vorticity maximum is found over the northern part of the Korean peninsula. The figure indicates that some positive vorticity advection exists near the mesoscale low off the midwest coast of Japan, but it is not significant (Fig. 6a). Upper-level vorticity advection is generally weak throughout the present case (not shown). The Q vector and its divergence at 850 hPa indicate an upward motion around the mesoscale low (Fig. 6b).

The mesocyclone shown in this section forms over the area of polar air convergence over the East Sea. The development of the meso-α-scale cyclone occurs in an environment with strong baroclinicity, synoptic-scale cyclonic circulation, and a cold vortex aloft whose center is located to the north of the surface cyclone. Satellite data suggest that the depth of the meso-α-scale cyclone is limited to be below the 500-hPa level. Observation also indicates that the KMC may play some roles in the mesoscale cyclogenesis, since the surface pressure trough is found to the lee side of the KMC and the satellite imagery shows that a mesoscale vortex develops to the lee side of the KMC and then travels southeastward.

3. Numerical experiments

In the following sections, we will use a mesoscale model to investigate the present case of mesoscale cyclogenesis. This study employs the Colorado State University RAMS mesoscale numerical model (Tripoli and Cotton 1982; Tremback et al. 1985). The model assumes a hydrostatic and anelastic atmosphere in terrain-following sigma-z (z*) coordinates. A Smagorinsky-type eddy coefficient is used for eddy diffusion and a simplified Kuo-type parameterization is used for cumulus convection (Tremback 1990). The ground surface temperature is predicted by the soil model of Tremback and Kessler (1985).

Several numerical experiments have been performed to investigate the mesoscale cyclogenesis, and major features of the experiments are summarized in Table 1. Experiment RE is designed to simulate the observed flow fields and to obtain deeper insights into the mesoscale cyclogenesis. Experiments I1 through I5 study the effects of the mountain on the airflow in an environment characterized by a significant low-level baroclinicity and warm sea surface to the downstream side of an idealized terrain. Experiments E1 through E3 study the roles of physical factors in the observed mesoscale cyclogenesis.

All experiments with real data (RE, E1–E3) are carried out in a domain that contains 107 × 90 × 17 grid points with a constant horizontal grid interval of 20 km. The model domain and smoothed topography are shown in Fig. 1. These experiments are carried out for 36 h from 0000 UTC 28 to 1200 UTC 29 January 1995. Initial meteorological fields and synoptic-scale tendencies at lateral boundaries are prepared using the 2.5° × 2.5° analysis data provided by JMA. The initial ground surface temperature is obtained by extrapolating the initial air temperature above the ground. Climatological data is used for the sea surface temperature. The roughness length of the land surface is assumed to be 0.05 m. The experimental design for the idealized experiments (I1–I5) will be described in section 5a.

4. The control experiment and comparisons with observations

The simulated SLP is shown in Fig. 7 at 6-h intervals. A mesoscale trough forms on the lee (southeast) side of the KMC during the daytime and persists through the evening (not shown). At 1800 UTC 28 January, the trough becomes sharper (Fig. 7a). It extends over the sea to the west coast of Japan. Near southern Hokkaido, Japan, a mesoscale low has developed off the midwest coast of Japan and moved to its present location during the past 12 h. The lee trough departs from the KMC after 2100 UTC 28 January and travels southeastward. The trough deepens faster after 0000 UTC 29 January at the rate of about 2 hPa (3 h)−1. It then develops into a closed low on its east side at 0900 UTC and deepens further with its central pressure below 1014 hPa at 1200 UTC 29 January near the midwest coast of Japan (Fig. 7d). The trough extends from the East Korea Bay (around x = −400 km, y = −50 km) to this low with a V-shaped pattern.

Streamlines at z* = 95 m and 850-hPa level are shown in Figs. 8 and 9, respectively. At 1200 UTC 28 January, a low-level convergence zone is found between the lee side of the KMC and the mesoscale low near Japan (Fig. 8a). Over the KMC, the incoming northwesterly flow splits and the northern branch undergoes an anticyclonic circulation. This anticyclonic circulation is associated with the mountain-induced high pressure over an isolated mesoscale mountain. A sharp cyclonic turning is found immediately to the lee side. This flow pattern is similar to the baroclinic flow over an isolated mountain found in linear theory (Smith 1984, 1986) and numerical simulations (Lin and Perkey 1989), even though no wind reversal level exists in the present case. The convergence zone on the lee side of the KMC becomes stronger at 1800 UTC (Fig. 8b). Airflow from the north near Vladivostok and the Sikhote-Alin mountain range also blows into this convergence zone. This convergence zone departs from the lee side of the KMC after 2100 UTC 28 January and then moves southeastward. The vortex center, however, moves eastward (Figs. 8c and 8d).

At 850 hPa, a cyclonic mesovortex appears over the sea immediately to the lee side of the KMC at 1800 UTC 28 January (Fig. 9b). This vortex has formed at about 1500 UTC and grows in size without a noticeable movement until 2100 UTC and then moves southeastward growing into a meso-α-scale vortex. However, its size remains quasi-steady after 0300 UTC 29 January. The 850-hPa vortex is located within the mixed layer above the convergence zone and elongates in a west–east direction at earlier times and a southwest–northeast direction at later times.

Figure 10 shows the y–z cross sections of equivalent potential temperature, cloud water mixing ratio, and v–w wind vector fields at x = −100 km for 0000 UTC 29 January. The depth of the mixed layer over the ocean is less than 2 km in general, except over the area to the north of the convergence zone where clouds develop up to a 4-km height. The wind vector field shows that the convergence associated with the northerly incoming flow is found up to the 1315-m level (Fig. 10b). This indicates that the 850-hPa vortex is developing at the top of the layer with the polar air convergence. Figure 11 shows the cloud water mixing ratio fields for 1200 UTC 29 January at two levels, z* = 2402 and 4019 m. The southern boundary of clouds at 2402 m coincides with the location of the low-level convergence zone. The clouds at 4019 m are found to the north of the mesoscale low off the midwest coast of Japan.

Simulated heat fluxes from the sea surface show significant spatial variation with a local minimum over the trough region where the wind speeds are low (not shown). Sensible and latent heat fluxes are relatively large off the east coast of the Korean peninsula with maximum value larger than 250 and 400 W m−2 near the east coast, respectively.

The simulated SLP patterns (Fig. 7) agree reasonably well with those of observations. But the simulated SLP near the center of the mesoscale low off the midwest coast of Japan at 1200 UTC 29 January is somewhat higher than the observed SLP. The model results show more significant mesoscale features over the East Sea than the synoptic weather charts, especially the movement of the trough. The synoptic SLP chart (Fig. 2) does not show the mesoscale features over the sea due to sparse observations and may not be adequate for the verification of the model results. The simulated streamlines at 850 hPa (Fig. 9) agree very well with the analyzed streamlines (Fig. 12), except that the simulation shows more mesoscale features in the lee of the KMC. Both the simulated and observed streamlines show the transition of flow in the upstream of the KMC from northwesterlies at 1800 UTC 28 January to general northerlies at 1200 UTC 29 January. The location of the simulated mesoscale cyclonic vortex over the southeastern East Sea at 1200 UTC 29 January agrees very well with that of the analyzed vortex (Figs. 9d and 12b). The simulation is also in good agreement with the satellite observations. The development of the simulated 850-hPa vortex to the lee side of the KMC (Fig. 9b) matches well with the cloud pattern in the satellite imagery during 1800–2100 UTC 28 January (Fig. 3a). The simulated cyclonic circulation and its location also match reasonably with those deduced from the cloud imagery for 1200 UTC 29 January. A good agreement can be found between the location of simulated clouds over the low-level convergence zone and that of the convergent cloud band found in the satellite imagery during the later stage (e.g., Figs. 3d and 11a). The relatively clear area near southern Hokkaido also matches reasonably well with the simulated mesoscale ridge area (Figs. 3d and 7d).

It is difficult to find the meso-α-scale vortex in the satellite imagery for 0600 UTC 29 January (Fig. 3c). According to the simulated results, the vortex is embedded in the mixed layer where clouds form primarily by turbulent mixing or by ascent of air over the convergence zone. A relatively strong upward motion associated with the lower-level convergence is found in the northern and western parts of the vortex where relatively deep clouds develop, while a weak downward motion is found over the southern part where no cloud is found above the mixed layer (Fig. 13). It appears that the northern and western parts of the vortex are covered by the relatively deep clouds that developed over the convergence zone. The southern boundary of the simulated cloud at z* = 2402 m (Fig. 13c) matches very well with that of the convergent cloud band in satellite imagery (Fig. 3c). These results together suggest that the simulated clouds may not show the meso-α-scale vortex pattern. This may also imply that there can be more meso-α-scale vortices over the sea than what is shown in the satellite imagery. In addition, Figs. 10 and 13 indicate that the cyclone is shallow and basically is a boundary layer disturbance since the major upward motion is below 3 km. This is in agreement with the satellite observations that indicate that the clouds associated with the meso-α-scale cyclone are limited below 500 hPa.

The present results have shown that a mesoscale trough forms over the sea on the lee side of the KMC and then develops into a strong convergence zone and a vortex in a relatively complicated environment: a mixed layer over the sea with strong low-level baroclinicity. These results are further investigated to determine the processes that are responsible for the development of the leeside disturbance.

The development of the leeside trough during the daytime appears to be associated mainly with diurnal heating and thermal advection. The solar heating of the elevated plateau (i.e., the KMC) and differential thermal advection create a thermal ridge over the mountaintop and downslope area, consequently producing a leeside trough. The impact of diurnal heating will be discussed later using idealized experiments. Differential thermal advection is important throughout the event and will be described in detail below.

Figure 14 shows the changes in SLP and average temperature in the lowest 444-m layer during the 3-h period from 1800 to 2100 UTC 28 January (0300–0600 LST 29 January). SLP has increased over the continent and over the sea to the west of the peninsula, while it has decreased over most of the area to the east of the continent and the peninsula. A decrease in pressure is found in the lee side of the KMC, where a temperature increase is found. The KMC contributes both directly and indirectly to the leeside temperature change. As it splits and deflects the approaching airstream, air above it descends over the leeside resulting in a warmer region over the downslope area through adiabatic warming. In the mean time, the deflected airstream to the northeast of the KMC advects colder air to the area surrounding the lee trough, while thermal advection is weakened over the downslope area and in the lee trough (Fig. 15). In the trough over the sea, heating of the mixed layer by the warm sea surface is generally stronger than the advective cooling, while the opposite is true in the area surrounding the trough. These results suggest that the differential thermal advection and the adiabatic descent are the key processes that contribute to the development of the strong leeside trough. Analysis also indicates that the differential heating of air in between and surrounding the trough is important for the deepening of the trough over the ocean throughout the period.

The mesoscale vortex, which appears on the lee side of the KMC at 1800 UTC (Fig. 9b), develops over the strong convergence zone. In order to explain the development of this vortex, the vorticity budget has been investigated. Figure 16 shows the relative vorticity, the stretching, and tilting terms that are found to be mainly responsible for the vorticity generation. It can be found that the locations of maximum vorticity at 585- and 1315-m levels coincide with that of the maximum vertical stretching at 585 m (Figs. 16a–c). In addition, the values of the maximum vorticity at the two levels are similar (about 2.4 × 10−4 s−1), even though the combined stretching and tilting effects at 1315 m are weaker than the stretching at 585 m. This is due to the turbulent mixing of vorticity in the mixed layer. This indicates that the vertical stretching is the major source of vorticity near the area of vorticity maximum. The tilting is important at higher levels (i.e., 1315 m) in the eastern part of the vortex where its importance is comparable to that of the stretching at lower levels. It also contributes to the anticyclonic vorticity to the north of the vortex. Note that another vorticity source may be derived from upstream of the KMC either through potential vorticity generation (Smith 1989) or through the tilting of horizontal vorticity as proposed by Smolarkiewicz and Rotunno (1989a,b).

In this vortex area where vorticity does not significantly vary with height in the layer below 1315 m (Figs. 16a and 16b), the flow maintains a convergence pattern at lower levels (Fig. 8b), although a vortex forms near the 850-hPa level. This variation of flow pattern with the height can be explained by the anticyclonic vorticity to the north of the large positive vorticity. As found in Figs. 16a and 16b, the anticyclonic vorticity to the north of the positive vorticity area is much stronger at 1315 m than that at 585 m. The vortex can be better defined with the significant contrast of positive and negative vorticity. As mentioned previously, the tilting is important for the anticyclonic vorticity to the north of the vortex. Another way to look at this feature is that the flow at 850 hPa behaves more like an inviscid, low Froude number flow over a mountain [e.g., Smolarkiewicz and Rotunno (1989a,b)], while the convergence associated with the flow near the surface is influenced by friction in the mixed layer.

The departure of the leeside trough from the KMC after 2100 UTC 28 (Fig. 7b) is mainly caused by the cold air advection in the lee coastal area. It occurs gradually. The air deflected by the KMC to the left-hand side (facing downstream) turns anticyclonically and flows into the lee side trough, bringing colder air into the area (Fig. 15b). This cold air advection pushes the thermal ridge on the lee coastal area southwestward and weakens the middle portion of the extended thermal ridge. In the meantime, the lee trough over the sea steadily intensifies and a significant increase of wind speed is found in the northwestern part of the trough (not shown). The enhanced airflow in the lee coastal area pushes the trough farther southwestward, and causes the trough over the sea to become detached from the KMC. After the detachment, the strong northerly in the northwestern part of the trough forces the western part of the trough to move southward fast.

This simulation has reproduced the observed features of the present mesocyclogenesis fairly well and has revealed some valuable insights into the event. The cyclogenesis is occurring in a progressive manner. It appears that four distinctive stages of the cyclogenesis can be identified. 1) First, a surface pressure trough forms on the lee side of the KMC under a northwesterly synoptic-scale flow which is deflected anticyclonically over the KMC. 2) Second, the lee trough deepens further into a strong convergence zone and a meso-β-scale vortex. 3) Then the meso-β-scale vortex develops into a meso-α-scale vortex as the trough and the vortex begin to move southeastward from the lee of the KMC. 4) Finally, the surface trough deepens into a closed low and the meso-α-scale vortex becomes collocated with this deepening surface low to form a meso-α-scale cyclone over the southeastern East Sea where static stability of the 500–850-hPa layer is relatively weak. The simulated results suggest the importance of the KMC for the genesis of the trough and the convergence zone. The results also indicate the importance of the warm sea surface and low-level baroclinicity for the mesoscale cyclogenesis considered here.

5. Roles of physical factors in the meso-α-scale cyclogenesis

Explanation of the mesoscale cyclogenesis shown in the previous sections requires further understanding of the roles of some physical factors, such as the KMC, warm sea surface, condensational heating, etc., in the formation of the lee trough–vortex and the transition from vortex into cyclone. This section describes the roles of these physical factors in the mesoscale cyclogenesis based on the results of numerical sensitivity experiments.

a. Roles of the KMC (northern Korean mountain complex)

Five idealized experiments have been performed to study the effects of the KMC on the airflow in an environment characterized by significant low-level baroclinicity and a warm sea surface to the lee side (Table 1). All experiments, except for experiment I1, include the warm sea surface downstream of the idealized mountain. Experiments I1 through I3 do not consider the diurnal heating of the land surface. Experiment I1 is designed to show the mountain effect without the sea. Experiments I3 and I2 are to show the effects of the mountain with and without the east–west baroclinicity, respectively. Experiments I5 and I4 are to show the effects of diurnal heating with and without the horizontal temperature gradient, respectively. These experiments have used an elongated bell-shaped mountain which possesses the characteristic size and shape of the KMC. The model domain has 86 × 71 × 17 grid points with a constant horizontal grid size of 20 km.

The initial fields for I3 and I5 are obtained by running the model for 4 h during which time cooling of the atmosphere is gradually imposed over the land as a function of height and distance from the shoreline so that the initial air temperature near the surface is about 8 K lower at 800 km inland from the shoreline than that at the shoreline. The cooling rate decreases to 0 at z = 5 km. Initial temperature and humidity profiles for this 4-h run are taken from the JMA analysis data for 1200 UTC 28 January at a location upstream of the KMC. A uniform westerly wind of 7.5 m s−1 is assumed initially. The sea surface temperature is 274 K at the coast and increases toward the sea at the rate of 1 K (100 km)−1. Other model input data are the same as those used in experiment RE. Initial fields for all other experiments are also obtained from the 4-h run, but with no prescribed artificial cooling. This is necessary for the experiments with the sea to have the same atmospheric boundary layers over the sea, although they are not exactly the same due to the difference in upstream condition. The wind fields obtained from the 4-h run with horizontal temperature gradient are not yet in thermal wind balance. The vertical wind shear, however, becomes more consistent with the thermal wind as the integration proceeds. The Froude numbers (U/Nh) for the initial upstream flows with and without the horizontal temperature gradient are about 0.30 and 0.35, respectively, if we assume U = 7.5 m s−1.

Figure 17 shows the streamlines at z* = 1315 m from experiments I1, I2, and I3. When there exists a flat land surface instead of sea on the downstream side of the mountain and diurnal heating is not considered (Fig. 17a), a weak trough appears on the lee side. The leeside disturbance is stronger with the warm sea surface than without it (Fig. 17b). At lower levels, however, the disturbance is slightly stronger in I1 (not shown). The difference between I1 and I2 becomes more noticeable at later times. This may be explained by the following two impacts of a warm sea surface. The perturbation induced by the mountain is vertically well mixed over the sea and the horizontal temperature gradient produced by the land–sea thermal contrast enhances the development of disturbance. When the horizontal temperature gradient is introduced upstream in I3, the leeside disturbance becomes stronger and a pair of cyclonic and anticyclonic vortices appear on the downstream side (Fig. 17c). The temporal variation of flow pattern in I3 is not shown here, but it is similar to that of I5, which will be shown later.

Figure 18 shows the SLP perturbation and the potential temperature fields at z* = 585 m. The pressure perturbation in I1 shows a circular pattern on the downstream side, while the perturbation in I3 shows a trough pattern. Due to the addition of the temperature gradient, the strongest contrast in temperature is found in I3 between the warmer leeside area and its surroundings. Note that the temperature gradient near the coastline is enhanced by the land–sea thermal contrast (Figs. 18d and 18f). The stronger development of the leeside disturbance in I3 is mainly due to the differential thermal advection, as discussed in the previous section. While the subsidence produced by the mountain induces warming over the downslope area, the differential thermal advection produces further thermal contrast between the lee trough and its surrounding area, consequently producing a stronger trough or vortex. This differential advection also helps the development of anticyclonic vortex in I3. The anticyclonically deflected air flows into the trough that extends from the lee slope to the sea. The air flowing toward the lee slope forms an anticyclonic circulation to the north of the trough with the contribution of anticyclonic vorticity generation through the column shrinking. Results of I2 and I3 are similar to the lee cyclogenesis simulated for an idealized baroclinic flow over a bell-shaped mountain by Lin and Perkey (1989).

The diurnal heating of the land surface without the initial temperature gradient brings perturbations to the airflow whose pattern and magnitude are similar to those in I2 (not shown). But the introduction of the temperature gradient and the diurnal heating together produces significant perturbations to the airflow (Fig. 19). At 1600 LST, the left branch (facing downstream) of airflow passing over the mountain turns anticyclonically toward the leeside trough, forming a strong convergence zone on the right-hand side of the downstream area (Fig. 19a). By 1900 LST, a pair of mesoscale vortices develop in the downstream area (Fig. 19b). The disturbance, however, becomes weaker later as the convergence at low levels over the mountain weakens (Fig. 19c). This may be mainly due to the dissipation of the warm core over the mountain. A detailed comparison indicates that the overall development of the lee vortices at various levels is stronger in I5 than that in I3. The importance of diurnal surface heating is also found by Sun and Chern (1993) in a study of the lee vortices in Taiwan. In their case, however, they found that an existing vortex over the sea eventually vanishes during the daytime due to the influence of the strong solar heating of ground surface. In the present wintertime case, the heating of elevated ground surface contributes to the development of the lee vortices in the presence of a significant horizontal temperature gradient.

The results of these idealized experiments indicate that the idealized mountain with the characteristics of the KMC can produce a mesoscale trough and vortex on its lee side. The low-level baroclinicity, enhanced by the land–sea thermal contrast, provides an important source for the leeside disturbance development. Diurnal heating of an elevated plateau (i.e., the KMC) is also found to be important for the development of the leeside disturbance.

In order to investigate the effect of the KMC further in a real setting, the KMC is removed in experiment E1. A small low appears in an area of weak pressure gradient to the east of the middle Korean peninsula at 1800 UTC 28 (Fig. 20a). At 0000 UTC 29 January, the low (1021 hPa) is embedded in the trough over the middle of the sea (Fig. 20b). This low keeps moving eastward and reaches near the midwest coast of Japan at 1200 UTC 29 January (Fig. 20d). The low near the midwest coast of Japan is similar to that for RE (Fig. 7d) in its location and strength. However, the SLP pattern in Fig. 20d does not show the well-defined trough found in RE to extend from the lee of the KMC to the west coast of Japan in a V-shaped pattern (Fig. 7d). According to the streamlines at 850 hPa (Fig. 21), the lee vortex is not found, but a cyclonic vortex develops over the surface low at about 0600 UTC 29 January. The cyclonic vortex at 1200 UTC 29 in E1 is weaker and less organized than that found in RE.

The differences between RE and E1 in SLP and the surface winds are shown in Fig. 22 for 1800 UTC 28 January. Positive values of pressure difference are found over the KMC (Fig. 22a), and they may be interpreted mainly as the result of pressure increase due to radiative cooling of the elevated mountainous area in experiment RE. Negative values in the lee side of the KMC can be related to the effects of both the KMC and warm sea surface. The wind vector difference clearly demonstrates the importance of the KMC in the formation of the convergence zone (Fig. 22b). However, the formation of the convergence zone will be discussed later, since it is also related to the effect of the warm sea surface.

Results of experiment E1 indicate that mesoscale cyclogenesis can also occur in the absence of the KMC. This cyclogenesis appears to be initiated by the influences of preexisting mesoscale cyclone off the midwest coast of Japan and the land–sea thermal contrast along the east coast of the Korean peninsula during the nighttime. This will be discussed in the next section.

b. Roles of the preexisting meso-α-scale cyclone and land–sea thermal contrast

Results of experiment E1 are analyzed further to investigate the formation of the mesoscale low off the mideast coast of the Korean peninsula in E1. Figure 23 shows the changes in SLP and temperature in the lowest 444-m layer during the 3-h period from 1200 to 1500 UTC 28 January when a small low forms in the area of weak pressure gradient near (39°N, 130°E).

A pressure decrease is found off the east coast of the Korean peninsula (Fig. 23a). The figure indicates that the pressure decrease is related to the differential heating of the air (Fig. 23b). A heat budget analysis suggests that the temperature increase over the area of low formation is due to the heating by the underlying warm sea surface in the presence of relatively weak cold air advection. To the east of this area, warming of the mixed layer by the sea surface is offset or dominated by the cold air advection in the western part of the mesocyclone off the west coast of Japan. The temperature decrease over the peninsula is due to the radiative cooling of the land surface and cold air advection. This differential heating of air results in the pressure decrease and, consequently, the formation of a low off the east coast of the peninsula.

The results shown in Fig. 23 and the tendency of isobars to be parallel to the coastline during the nighttime (Figs. 20a,b) suggest the importance of coastal shape for the low formation. These effects of the land–sea thermal contrast and coastal shape are also important in the results of RE, as will be shown later. But in RE, the KMC plays a more pronounced role in the mesocyclogenesis by producing a trough on its lee side. The preexisting mesoscale cyclone should play the same role in the mesocyclogenesis in experiment RE as that found in the results of experiment E1.

c. Roles of warm sea surface

The effect of the ocean is investigated by deactivating the sensible and latent heat fluxes (experiment E2). The SLP pattern from E2 shows that a surface trough appears to the south of the KMC at 1200 UTC 28 January (Fig. 24a). The pattern at later times, however, shows no significant disturbance to the lee side of the KMC (not shown). It can be seen that the mesoscale cyclone and the trough associated with it do not develop without the effects of a warm sea surface (Fig. 24b). According to the streamline analysis (not shown), the anticyclonic deflection appears over the KMC and a low-level convergence zone develops on the lee side, but the convergence zone quickly moves southwestward to the south of the KMC. A weak shortwave appears at higher levels over the KMC, but it does not develop into a vortex.

Figure 25 shows the differences between experiments RE and E2 in SLP and wind vectors at z* = 95 m for 0000 UTC 29 January. The pattern of pressure difference over the East Sea somewhat resembles that of SLP in RE. The SLP is about 2–7 hPa lower in RE than that in E2 over most of the area to the east of the continent and the peninsula (Fig. 25a). Relatively large differences in SLP and wind vector are located over the trough area of RE. The wind vector difference shows a strong cyclonic convergence pattern along the trough (Fig. 25b).

The wind vector difference and the relatively strong gradient of pressure difference along the east coast of the Korean peninsula indicate the importance of land–sea thermal contrast and the coastline shape for the development of a cyclonic circulation and convergence zone to the east of the peninsula. The wind vector difference over the coastal area may be interpreted as a mechanism analogous to the land breeze (cf. Figs. 22b and 25b). Due to the coastal shape, this land-breeze-type wind may contribute to the development of the cyclonic circulation over the sea around the KMC. In the ocean area near Vladivostok, where the synoptic scale pressure gradient is relatively weak in the lower troposphere (see Fig. 4c), a significant portion of the prevailing northerly airflow can be explained by the ocean effect (i.e., Fig. 25b). This northerly flow is important for the development of the convergence zone. These also indicate that the synoptic-scale pressure pattern in the lower troposphere is important for the formation of the convergence zone. Atlas et al. (1983) also found the importance of the coastal shape for the formation of a convergence line off the coast of Long Island, New York, during a cold air outbreak. They suggested that the major difference in the air on either side of the convergence line was due to the difference in the temperature of underlying sea surface and the path length of overwater travel, which depends on the coastal shape.

Comparison of experiments RE and E2 indicates that the East Sea produces a favorable atmospheric environment for the mesoscale cyclogenesis through the heating of the lower troposphere, which results in a region of lower pressure and low-level cyclonic circulations to the east of the Korean peninsula. The KMC produces a lee trough, but the warm sea surface plays the crucial role for the development of the convergence zone. The impact of the ocean may vary with the case depending on the temperature of the airflow moving out off the continent. As implied by the result of E2, the impact would be weaker for an airflow with higher air temperature due to a smaller heat supply from the sea.

d. Role of condensational heating

This section discusses the results of E3, in which the condensational heating is neglected. The patterns of SLP and streamlines from E3 are not significantly different from those of RE until the trough reaches the area of cyclone development, although the trough and convergence is weaker without the condensational heating. The effect of condensational heating is noticeable at 1200 UTC 29 January (Fig. 26). The cyclone does not develop and the SLP pattern remains as a trough, suggesting that a major effect of condensational heating is to enhance the deepening of the trough and produce a well-defined mesoscale cyclone. The pressure decrease associated with the condensational heating is about 2.5 hPa over the mesoscale trough and cyclone. The difference between RE and E3 in Δθ [i.e., Δθ(RE) − Δθ(E3)] over the trough and mesoscale cyclone is up to 3–4 K (Fig. 26b). This indicates that the condensational heating contributes most of the decrease of Δθ over the trough in RE during the 6-h period from 0600 to 1200 UTC 29 January. The results of E3 also indicate that the meso-α-scale cyclone does not develop by the influence of upper-level forcing alone.

6. Concluding remarks

Numerical simulations and the analysis of observational data are employed to understand the meso-α-scale cyclogenesis in a polar airstream which occurred over the sea to the east of the Korean peninsula on 28–29 January 1995. The observational analysis shows that a mesoscale low develops over the southeastern East Sea (Japan Sea) on 29 January 1995. Satellite imagery also indicates that a meso-β-scale vortex forms on the lee side of the northern Korean mountain complex (KMC), which is located in the northern Korean peninsula, and that a meso-α-scale cyclone develops over the southeastern East Sea at later times. The mesoscale cyclone forms in the lower troposphere characterized by strong baroclinicity and cyclonic circulation under the influence of an upper-level cold vortex.

Numerical simulations have captured major features of the observed cyclogenesis very well and revealed some valuable insights into the cyclogenesis process discussed here. The cyclogenesis occurs in a progressive manner. Basically, four distinctive stages of the cyclogenesis are identified. 1) First, a surface pressure trough forms on the lee side of the KMC under a northwesterly synoptic-scale flow which is deflected anticyclonically over the KMC. 2) Second, the lee trough deepens further into a meso-β-scale vortex and a strong convergence zone, to which the cold air originating from the region between Vladivostok and the Sikhote-Alin range blows. 3) Then the meso-β-scale vortex develops into a meso-α-scale vortex as the vortex and the trough begin to move southeastward from the lee of the KMC. 4) Finally, the surface trough deepens into a closed low and the meso-α-scale vortex becomes collocated with this deepening surface low to form a meso-α-scale cyclone over the southeastern East Sea.

Several numerical sensitivity experiments are performed to isolate the effects of topography, a warmer sea surface, diurnal thermal forcing, and latent heat release. During stages 1 and 2, it is found that the KMC and low-level baroclinicity are responsible for generating the lee trough. The presence of a warm sea surface is found to be necessary for the development of the polar air convergence zone. The present study has shown that both dynamic and thermal effects associated with the KMC are important. The solar heating and differential cold air advection may help to build a warm ridge over the lee slope of the KMC during the daytime. As a result, a trough and convergence zone form over the lee side of the KMC in the afternoon. It is found that the KMC contributes both directly and indirectly to the strengthening of the leeside trough during the nighttime. While the subsidence produced by the mountain induces warming over the downslope area, the differential thermal advection caused by the deflected flow produces further thermal contrast between the lee trough and its surrounding area, consequently strengthening the trough. A mesoscale vortex forms over the strong convergence zone and it is best defined at approximately the 850-hPa level, which is near the top of the layer of polar air convergence. The major source of vorticity in the area of vorticity maximum is the vertical stretching caused by the lower-level convergence, while the tilting of the horizontal vorticity is also found to be an important source of vorticity in the eastern part of the vortex. Both the tilting and vertical stretching in a synoptic-scale cyclonic circulation are found to be important for the development of the meso-α-scale vortex during stage 3. In the final stage, condensational heating plays the key role for the development of the meso-α-scale cyclone.

This study suggests that the effects of KMC are important for the initial development of the mesoscale cyclone, and that the genesis and development of the cyclone require the influences of both a warm sea surface and condensational heating. Differential cold air advection is found to be an important process throughout all the stages of evolution, which allows for stronger development of the lee trough and vortex and also the deepening of the trough over the ocean. In other words, the low-level baroclinicity is essential for the present case of mesoscale cyclogenesis. The southeastward movement of the upper-level cold vortex and the condensational heating of the lower troposphere contribute to the development of the meso-α-scale cyclone by decreasing the atmospheric stability. Sensitivity tests also indicate that the present meso-α-scale cyclone does not develop by the influence of upper-level forcing alone. The possible contribution from the CISK is still not well understood and needs to be investigated in the future.

This study also indicates that, under the atmospheric conditions given in this case, a mesoscale cyclone is still able to form without the influence of the KMC. However, this mesoscale cyclone is initiated over the sea, instead of over the lee side of the KMC. In that case, the influences of the preexisting mesoscale cyclone near the midwest coast of Japan and the land–sea thermal contrast contribute to the formation of the low off the mideast coast of the Korean peninsula. The basic wind direction around the KMC and the KMC orientation are found to be important for the development of trough to the east of the KMC. The KMC may not be able to produce a significant impact on the airflow over the sea, when a strong synoptic-scale northerly flow prevails around the KMC and over the sea.

According to this numerical study, more meso-α-scale vortices may form over the convergence zone than what is shown in the satellite imagery. The formation and organization of these meso-α-scale vortices deserve a further study, since they may organize into a meso-α-scale cyclone when they reach a location of weak static stability, such as the area under the upper-level cold vortex. The importance of the effects of the KMC on the initial development of the mesoscale cyclone is often missing in forecasting and from previous studies due to the lack of data over the sea. This study implies that a mesoscale numerical model with a fine grid resolution can be used to improve the prediction of mesoscale disturbances that develop to the east of the Korean peninsula.

Acknowledgments

This work is supported by the Korea Science and Engineering Foundation under Grant 94-0703-02-01-3. The authors wish to thank the anonymous reviewers for their valuable comments and Dr. R. P. Weglarz for his proofreading and comments. They would like to thank Drs. Pielke and Cotton at Colorado State University for allowing them to use the CSU-RAMS model. They also thank the Korea Meteorological Administration (KMA) for providing the Japan Meteorological Agency analysis data. Thanks are also extended to Mr. Y.-H. Kim of KMA for processing the satellite data. Computations were performed on the CRAYC90 supercomputer of Systems and Engineering Research Institute, Taejon, Korea.

REFERENCES

  • Asai, T., 1988: Meso-scale features of heavy snowfalls in Japan Sea coastal regions of Japan (in Japanese). Tenki,35, 156–161.

  • Atlas, D., S.-H. Chou, and W. P. Byerly, 1983: The influence of coastal shape on winter mesoscale air–sea interaction. Mon. Wea. Rev.,111, 245–252.

  • Bond, N. A., and M. A. Shapiro, 1991: Polar lows over the Gulf of Alaska in conditions of reverse shear. Mon. Wea. Rev.,119, 551–572.

  • Crook, N. A., T. L. Clark, and M. W. Moncrieff, 1990: The Denver cyclone. Part I: Generation in low Froude number flow. J. Atmos. Sci.,47, 2725–2742.

  • Douglas, M. W., L. S. Fedor, and M. A. Shapiro, 1991: Polar low structure over the northern Gulf of Alaska based on research aircraft observations. Mon. Wea. Rev.,119, 32–54.

  • Forbes, G. S., and W. D. Lottes, 1985: Classification of mesoscale vortices in polar airstreams and the influence of the large-scale environment on their evolutions. Tellus,37A, 132–155.

  • Lee, T.-Y., and Y.-Y. Park, 1996: Formation of a mesoscale trough over the Korean peninsula during an excursion of Siberian high. J. Meteor. Soc. Japan,74, 299–323.

  • Lin, Y.-L., and D. J. Perkey, 1989: Numerical modeling of a process of lee cyclogenesis. J. Atmos. Sci.,46, 3685–3697.

  • ——, N.-H. Lin, and R. P. Weglarz, 1992: Numerical modeling studies of lee mesolows, mesovortices and mesocyclones with application to the formation of Taiwan mesolows. Meteor. Atmos. Phys.,49, 43–67.

  • Mansfield, D. A., 1974: Polar lows: The development of baroclinic disturbances in cold air outbreaks. Quart. J. Roy. Meteor. Soc.,100, 541–554.

  • Nagata, M., 1991: Further numerical study on the formation of the convergent cloud band over the Japan Sea in winter. J. Meteor. Soc. Japan,69, 419–427.

  • ——, 1993: Meso-β-scale vortices developing along the Japan-Sea Polar-Airmass Convergence Zone (JPCZ) cloud band: Numerical simulation. J. Meteor. Soc. Japan,71, 43–57.

  • ——, M. Ikawa, S. Yoshizumi, and T. Yoshida, 1986: On the formation of a convergent cloud band over the Japan Sea in winter:Numerical experiments. J. Meteor. Soc. Japan,64, 841–855.

  • Ninomiya, K., 1989: Polar/comma-cloud lows over the Japan Sea and the northwestern Pacific in winter. J. Meteor. Soc. Japan,67, 83–97.

  • ——, 1991: Polar low development over the east coast of the Asian continent on 9–11 December 1985. J. Meteor. Soc. Japan,69, 669–685.

  • Rasmussen, E., 1979: The polar low as an extratropical CISK disturbance. Quart. J. Roy. Meteor. Soc.,105, 531–549.

  • Smith, R. B., 1984: A theory of lee cyclogenesis. J. Atmos. Sci.,41, 1159–1168.

  • ——, 1986: Further development of a theory of lee cyclogenesis. J. Atmos. Sci.,43, 1582–1602.

  • ——, 1989: Comment on the “Low Froude number flow past three-dimensional obstacles. Part I: Baroclinically generated lee vortices.” J. Atmos. Sci.,46, 3611–3613.

  • Smolarkiewicz, P. K., and R. Rotunno, 1989a: Low Froude number flow past three-dimensional obstacles. Part I: Baroclinically generated lee vortices. J. Atmos. Sci.,46, 1154–1164.

  • ——, and ——, 1989b: Reply. J. Atmos. Sci.,46, 3614–3617.

  • ——, R. Rasmussen, and T. L. Clark, 1988: On the dynamics of Hawaiian cloud bands: Island forcing. J. Atmos. Sci.,45, 1872–1905.

  • Sun, W.-Y., and J.-D. Chern, 1993: Diurnal variation of lee vortices in Taiwan and the surrounding area. J. Atmos. Sci.,50, 3404–3430.

  • ——, ——, C.-C. Wu, and W.-R. Hsu, 1991: Numerical simulation of mesoscale circulation in Taiwan and surrounding area. Mon. Wea. Rev.,119, 2558–2573.

  • Tremback, C. J., 1990: Numerical simulation of a mesoscale convective complex: Model development and numerical results. Department of Atmospheric Science Paper 465, Colorado State University, 247 pp.

  • ——, and R. Kessler, 1985: A surface temperature and moisture parameterization for use in mesoscale numerical models. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 355–358.

  • ——, G. J. Tripoli, and W. R. Cotton, 1985: A regional scale atmospheric numerical model including explicit moist physics and a hydrostatic time-split scheme. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 433–434.

  • Tripoli, G. J., and W. R. Cotton, 1982: The Colorado State University three-dimensional cloud/mesoscale model—1982. Part I: General theoretical framework and sensitivity experiments. J. Rech. Atmos.,16, 185–220.

  • Tsuboki, K., and G. Wakahama, 1992: Mesoscale cyclogenesis in winter monsoon air streams: Quasi-geostrophic baroclinic instability as a mechanism of the cyclogenesis off the west coast of Hokkaido Island, Japan. J. Meteor. Soc. Japan,70, 77–93.

  • Yagi, S., T. Muramatsu, T. Uchiyama, and N. Kurokawa, 1986: “Convergent band cloud” and “Cu-Cb line” over Japan Sea affected by topographic features in the coast of the Asian continent (in Japanese). Tenki,33, 453–465.

Fig. 1.
Fig. 1.

Geographic map and smoothed topography in the mesoscale model domain. Contour interval is 200 m. The KMC represents the northern Korean mountain complex, located in the northern part of the Korean peninsula.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 2.
Fig. 2.

Sea level pressure (hPa) for (a) 1200 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 3.
Fig. 3.

GMS satellite IR imagery for (a) 2100 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 4.
Fig. 4.

Height (m, solid) and temperature (°C, dashed) fields at 500 hPa: (a) 1200 UTC 28 and (b) 1200 UTC 29 January 1995. At 850 hPa: (c) 1200 UTC 28 and (d) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 5.
Fig. 5.

Difference of potential temperature between the 500- and 850-hPa levels (Δθ, K) at (a) 1200 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 6.
Fig. 6.

(a) Geopotential height (m, solid line) and absolute vorticity (10−5 s−1, dashed line) at the 500-hPa level, and (b) the Q vector and its divergence (·Q, 10−16 km−2 s−1) at the 850-hPa level at 1200 UTC 29 January 1995. Contour intervals for absolute vorticity and Q-vector divergence are 2 × 10−5 s−1 and 4 × 10−16 km−2 s−1, respectively. (b) Solid and dashed lines indicate positive and negative values, respectively.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 7.
Fig. 7.

Simulated sea level pressure (hPa) at (a) 1800 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995. Terrain heights of 100, 500, 1000, and 1500 m are shaded with increasing darkness, respectively.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 8.
Fig. 8.

Simulated streamlines at z* = 95 m for (a) 1200 UTC 28, (b) 1800 UTC 28, (c) 0000 UTC 29, and (d) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 9.
Fig. 9.

Same as Fig. 8, except at 850 hPa.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 10.
Fig. 10.

The y–z cross sections of (a) equivalent potential temperature (K, solid) and cloud water mixing ratio (rc, dashed), and (b) υw wind vector fields at x = −100 km for 0000 UTC 29 January 1995. The isoline of rc starts from 0.1 g kg−1 with an interval of 0.1 g kg−1.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 11.
Fig. 11.

Simulated cloud water mixing ratio (10−2 g kg−1) where z* equals (a) 2402 m and (b) 4019 m for 1200 UTC 29 January 1995. Contour interval is 0.2 g kg−1.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 12.
Fig. 12.

Streamlines at 850 hPa obtained using the wind analysis data from the Japan Meteorological Agency for (a) 1800 UTC 28 and (b) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 13.
Fig. 13.

(a) Simulated streamlines at z* = 1315 m, (b) vertical velocity (cm s−1) at z* = 1548 m, and (c) cloud water mixing ratio (10−2 g kg−1) at z* = 2402 m, respectively. Solid and dashed lines in (b) indicate positive and negative values, respectively, with contour interval of 5 cm s−1.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 14.
Fig. 14.

Changes in (a) sea level pressure and (b) average potential temperature in the lowest 444-m layer during the 3-h period from 1800 to 2100 UTC 28 January 1995. Solid and dashed lines indicate positive and negative values, respectively. Contour interval is 0.4 hPa (3h)−1 in panel (a) and 0.5 K (3h)−1 in panel (b).

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 15.
Fig. 15.

Simulated fields of (a) potential temperature (K) and (b) horizontal wind vector at z* = 585 m.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 16.
Fig. 16.

Relative vorticity, stretching term, and tilting term are shown in (a), (c), and (e), respectively, for z* = 585 m, and in (b), (d), and (f), respectively, for z* = 1315 m. Contour interval is 20 × 10−6 s−1 in (a) and (b), and 50 × 10−10 s−2 in (c) through (f).

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 17.
Fig. 17.

Streamlines at 1315 m from expts (a) I1, (b) I2, and (c) I3.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 18.
Fig. 18.

Sea level pressure perturbations (hPa, upper panels) and potential temperatures (K, lower panels) at 585 m from expts I1 (a,b); I2 (c,d); and I3 (e,f).

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 19.
Fig. 19.

Same as Fig. 17 except for I5 at three different hours.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 20.
Fig. 20.

Simulated sea level pressure (hPa) from expt E1, in which the mountains over the northern Korean peninsula are removed, for (a) 1800 UTC 28, (b) 0000 UTC 29, (c) 0600 UTC 29, and (d) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 21.
Fig. 21.

Simulated streamlines at 850 hPa from E1 for (a) 1800 UTC 28, (b) 0000 UTC 29, and (c) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 22.
Fig. 22.

Differences between RE and E1 (RE − E1) in (a) sea level pressure (hPa) and (b) wind vectors at 95 m over the ocean at 1800 UTC 28 January 1995. Contour interval in (a) is 0.5 hPa.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 23.
Fig. 23.

Changes in (a) sea level pressure [hPa (3 h)−1] and (b) temperature [K (3 h)−1] during the 3-h period from 1200 to 1500 UTC 28 January 1995. Contour intervals in (a) and (b) are 0.5 hPa (3 h)−1 and 0.5 K (3 h)−1, respectively.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 24.
Fig. 24.

Simulated sea level pressure fields from expt E2, in which the latent and sensible heat fluxes at the surface are assumed to be 0, at (a) 1200 UTC 28 and (b) 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 25.
Fig. 25.

Difference between expts RE and E2 (RE − E2) in (a) sea level pressure (hPa) and (b) wind vector at 95 m at 0000 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Fig. 26.
Fig. 26.

(a) Simulated sea level pressure from E3, in which the condensational heating is not considered, and (b) the difference between the experiments RE and E3 (RE − E3) in the potential temperature difference (Δθ, K) for 1200 UTC 29 January 1995.

Citation: Monthly Weather Review 126, 9; 10.1175/1520-0493(1998)126<2305:ANMSOM>2.0.CO;2

Table 1.

Summary of numerical experiments.

Table 1.
Save