Abstract

Observational analyses and numerical simulations are used to investigate a Puget Sound convergence zone (PSCZ) event that occurred in the lee of the Olympic Mountains of Washington State. The PSCZ, which develops when low-level airstreams are forced to converge over Puget Sound by the regional orography, is frequently associated with a mesoscale swath of clouds and precipitation across the central Puget Sound that stretches eastward over the western side of the Cascade Mountains. It was found that latent heat release enhances the PSCZ circulation and associated precipitation. Both the Olympic and the Cascade Mountains are important in the formation of the PSCZ. The Olympics deflect the low-level onshore flow into two branches, one along the Strait of Juan de Fuca and another around the southern flank of the Olympics; in addition, a lee trough, which develops to the east of this barrier, induces convergence over the central Puget Sound. The Cascades deflect low-level flow over northern Puget Sound into a more northerly direction and generate a windward ridge over the southern sound that contributes to a northward deflection of the flow. In a series of sensitivity experiments, the Froude number (Fr) was varied by changing the height of the Olympic Mountains. When the height of the Olympics is reduced by half (Fr ∼ 1), the lee trough is weaker than in the control run (Fr = 0.4) and the low-level flow tends to move over the Olympics with less deflection and converges farther downstream. When the height of the Olympics was doubled (Fr ∼ 0.19), most of low-level air is forced to flow around the mountains, producing weakened convergence over the south-central sound. In the control simulation, a pair of vortices appear to the north and the south of the convergence zone; in the double Olympics run, the vortices occur farther downstream than in the control. Lee vortices do not appear in the half Olympics simulation, and in a simulation without the Cascades, the vortices are weakened and more symmetrical.

1. Introduction

The meteorology of the northwest coast of the United States is greatly influenced by mesoscale circulations that result from the interaction between the synoptic-scale flow and the coastal orography. Specifically, low-level flow off the Pacific Ocean interacts with the coastal mountains of Washington, Oregon, and northern California, producing a variety of mesoscale phenomena, including coastal ridging, ageostrophic alongshore southerlies, convergence zones, lee troughs, and gap winds. In the first paper of this study (Chien et al. 1997, hereafter P1), the authors demonstrated the importance of the complex terrain of the western United States in the formation of coastal pressure ridging, the onshore push of marine air, and alongshore southerlies prior to frontal passage on 25–26 May 1992. Subsequent to frontal landfall, a Puget Sound convergence zone (PSCZ) formed over the Puget Sound of western Washington; the evolution and dynamics of this event are discussed in this paper.

Puget Sound lies between the isolated, bell-shaped Olympic Mountains (approximately 80 km in diameter, some peaks exceeding 2000 m) to the west and the extensive north–south-oriented Cascade Mountains (crest height of 1500–2250 m) to the east (Fig. 1, also refer to Fig. 1 of P1). The Strait of Juan de Fuca and the Chehalis Gap are low-level channels to the north and south of the Olympics.

Fig. 1.

Terrain height (domain 3) for the control experiment with a contour interval of 100 m. See Fig. 8 of Part I for domain configuration. Lines A–D indicate the positions of cross sections that are presented in section 4.

Fig. 1.

Terrain height (domain 3) for the control experiment with a contour interval of 100 m. See Fig. 8 of Part I for domain configuration. Lines A–D indicate the positions of cross sections that are presented in section 4.

The PSCZ is probably the most important topographically induced mesoscale phenomenon affecting the Puget Sound lowlands. It typically occurs dozens of times per year and significantly modifies the precipitation climatology of the region (Mass 1981). Mass (1981) completed a detailed study of 25 PSCZ events, finding that the PSCZ occurs when low-level westerly/northwesterly winds from off the Pacific Ocean are deflected around the Olympics and converge over Puget Sound. Convergence zones are most frequent during the late spring and early summer months, when the coastal flow is most often from the west. Typically, convergence zones form immediately after frontal passage, when the coastal winds shift from a southerly or southwesterly direction into the northwest. They usually produce a band of cloudiness and precipitation in northern and central Puget Sound, with nearly cloud-free skies to the north and south.

Other studies have documented how the interaction between the Olympics and the synoptic-scale flow produces a variety of mesoscale features. For example, Reed (1980) found that strong southwesterly synoptic flow impinging on the Olympics could generate a lee low associated with extremely large pressure gradient and strong winds (reaching 50 m s−1 or more) over portions of northern Puget Sound. Smith (1981) used linear mountain wave theory to explain the above pressure gradient. Mass and Dempsey (1985a,b) showed that the Olympics can produce lee troughing and low-level convergence for both easterly and westerly flows. Mass and Ferber (1990) and Ferber and Mass (1990), using a network of microbarographs, demonstrated that a weak windward ridge and a strong lee trough often form to the west and the east of the Olympics, respectively, during strong south-southwesterly synoptic-scale flow. Walter and Overland (1982) suggested that stronger lee troughing was associated with larger Froude numbers. More recently, Colle and Mass (1996), using aircraft data collected during the COAST1 (Coastal Observations and Simulations with Topography) field experiment and high-resolution mesoscale model simulations, documented the three-dimensional flow structures around the Olympics under strong southwesterly flow conditions. Steenburgh and Mass (1996) described the transient mesoscale response produced by the Olympics as the center of an intense cyclone passed north of the barrier.

A considerable amount of observational research has examined the flow around other mesoscale three-dimensional obstacles. For example, Smith (1982) reviewed observations and theory regarding the wind and pressure patterns produced by several mountain ranges such as the European and New Zealand Alps and the Andes. Aanensen (1965) described windward ridging and lee troughing on the Pennine Mountain Range of northern England. Smolarkiewicz et al. (1988) discussed the mesoscale wind field generated by the island of Hawaii, and Wilczak and Glendening (1988) described a mesoscale circulation, the Denver “cyclone,” that often occurs in the lee of the Palmer Ridge of central Colorado.

Theoretical and laboratory work suggests that low–Froude number2 flow (Fr < 1) tends to pass around rather than over three-dimensional obstacles, producing wake effects and convergence in the lee of barriers (Drazin 1961; Smith 1979; Baines 1979). Linear theory does not agree well with observations for low Froude numbers since nonlinear effects are dominant (Smolarkiewicz and Rotunno 1989; Colle and Mass 1996). For large–Froude number flow (Fr > 1), linear theory becomes more realistic. Pressure patterns for the latter flow regime typically include windward ridging and leeward troughing of similar amplitude, resembling the configuration found in Smith’s linear model (Smith 1980, 1982, 1988).

Compared with other mesoscale obstacles around the world, the isolated and approximately bell-shaped Olympics are particularly suitable for the study of airflow around a three-dimensional barrier. Unfortunately, the observational network of the region, especially above the surface, is relatively coarse compared with the mesoscale circulations generated by the terrain. To produce a mesoscale dataset suitable for study, the Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model Version 5 (MM5) was used to simulate the three-dimensional structure of the PSCZ that occurred after the onshore push event described in Chien et al. (1997). Key questions that are addressed in this paper include the following. What is the role of troughing to the lee of the Olympics in the formation of the PSCZ? What is the importance of windward ridging on the Cascades in deflecting the flow? Why are there typically clear zones to the north and south of the PSCZ cloud band? What role does latent heating play in modulating the strength and structure of the PSCZ circulation? Are there lee vortex structures associated with the PSCZ? If so, what flow regime is favorable for the occurrence of such vortices?

Section 2 presents an overview of the event. In section 3, the model and experimental design are described briefly. The simulation results are presented in section 4. Section 5 discusses sensitivity studies designed to assess the importance of various topographic features and physical mechanisms in the formation of the PSCZ. A summary is presented in the final section.

2. Case description

The PSCZ event presented here occurred over the Puget Sound region between 0300 UTC and 1800 UTC 26 May 1992, as the weak synoptic front/trough described in P1 passed to the east of the region. At 1200 UTC 25 May (see Fig. 3c of P1), prior to the occurrence of the event, the synoptic front/trough was about 300 km offshore and winds along the coast were generally from the southwest. Twelve hours later at 0000 UTC 26 May (see Fig. 3e of P1), the front was making landfall along the Pacific Northwest coast and the coastal winds had begun to veer to a more westerly direction.

Fig. 3.

Visible satellite imagery at 1501 UTC 26 May 1992.

Fig. 3.

Visible satellite imagery at 1501 UTC 26 May 1992.

For a detailed description of the event, mesoscale analyses of sea level pressure, plotted station models, and 3-h rainfall for the period in which the PSCZ developed are presented in Fig. 2. At 0300 UTC 26 May (Fig. 2a), the synoptic front had made landfall, coastal winds had shifted to the west and northwest, and weak convergence associated with the front was evident over Puget Sound. By 0600 UTC 26 May, the front had passed out of the domain (Fig. 2b). With stronger westerly/northwesterly flow behind the front, weak lee troughing started to develop over the western Puget Sound area in the lee of the Olympic Mountains. Precipitation was observed both to the north and south of Puget Sound. At 0900 UTC 26 May (Fig. 2c), precipitation had moved eastward to the western foothills of the Cascades, associated with enhanced upslope (westerly) motion following frontal passage. At this time, the winds over the sound ranged from weak southerlies to the south to calm winds and weak northerlies to the north.

Fig. 2.

Mesoscale analyses at (a) 0300, (b) 0600, (c) 0900, (d) 1200, (e) 1500, and (f) 1800 UTC 26 May 1992. Sea level pressure interval is 1 mb. Three-hour precipitation (mm) ending at the observation time is found to the right of the small squares. Precipitation data is obtained from Hourly Precipitation Data of Washington State and Oregon, May 1992 (NOAA 1992).

Fig. 2.

Mesoscale analyses at (a) 0300, (b) 0600, (c) 0900, (d) 1200, (e) 1500, and (f) 1800 UTC 26 May 1992. Sea level pressure interval is 1 mb. Three-hour precipitation (mm) ending at the observation time is found to the right of the small squares. Precipitation data is obtained from Hourly Precipitation Data of Washington State and Oregon, May 1992 (NOAA 1992).

At 1200 UTC 26 May (Fig. 2d), lee troughing and converging airstreams were evident over Puget Sound; the character of the rainfall had begun to alter during the preceding 3-h period as the heaviest rain (reaching 5 mm over the period) was now limited to an area extending from central Puget Sound into the Cascades. While overcast skies were evident over central Puget Sound, to the south and north partly cloudy sky conditions were observed. Three hours later at 1500 UTC 26 May (Fig. 2e), the convergence over central Puget Sound had substantially intensified as a result of strengthening of both the northerly and southerly airstreams, and precipitation extended from the Cascades to northern Puget Sound. Another convergence zone appeared to have formed between northwesterly flow in the Strait of Georgia and southerly flow moving northward from the eastern Strait of Juan de Fuca. Low-level convergence and enhanced precipitation continued and weakened over the central sound through 1800 UTC 26 May (Fig. 2f).

Figure 3 presents a visible GOES (Geostationary Operational Environmental Satellite) image at 1501 UTC 26 May 1992. At this time the synoptic front was exiting Washington State to the northeast and a cloud band associated with the convergence zone extended from central Puget Sound to the western slopes of the Cascades. Broken to partly cloudy skies were observed to the north and south of the band. In contrast, skies were nearly cloud-free over the eastern (lee) slopes of the Olympics and in the lee of the Cascades and the mountains of Vancouver Island.

3. Model and experimental design

a. Model description

The PSU–NCAR MM5 was used in this study. It is a nonhydrostatic, multinested, primitive equation mesoscale model described in Grell et al. (1994). In the model simulations, the multilayer Blackadar (1979) parameterization was used to represent planetary boundary layer processes including surface fluxes of heat, moisture, and momentum. The drag enhancement in the Blackadar PBL scheme of Chien and Mass (1994), which parameterizes subgrid-scale drag over complex terrain, was also applied.3 The hydrological cycle includes the subgrid-scale convective parameterization of Grell (1993) and a grid-resolvable explicit moisture scheme, including prognostic equations for cloud water, rainwater, and ice (Dudhia 1993). In addition, the simulations also applied the upper radiation boundary condition of Klemp and Durran (1983), as well as relaxation lateral boundary conditions to nudge the model-predicted variables toward a large-scale analysis.

b. Experimental design

The domain configuration of the model simulations is shown in Fig. 8a of P1. Nested grids were used with horizontal resolutions of 45, 15, and 5 km (domains 1–3, respectively). The domain and terrain of the 5-km inner nest is shown in Fig. 1. The simulations for domains 1 and 2 were initialized at 1200 UTC 24 May and ended at 1800 UTC 26 May 1992. The model output from domain 2 was then used to create initial conditions for the inner domain at 0000 UTC 26 May and hourly boundary data for the next 18 h. Twenty-seven sigma4 levels were used in the vertical and the time step was 15 s. Six experiments, listed in Table 1, were conducted for domain 3. The control experiment applied all the physics discussed in the previous section and used realistic terrain. Latent heat release was turned off and the realistic terrain was used in the NLH (no latent heat release) run. To answer questions regarding the role of the lee troughing and the nature of lee vortices, the height of the Olympics was set to zero, reduced by half, and doubled in the NO (no Olympics), HO (half Olympics), and DO (double Olympics) runs, respectively, with full physics being applied. Finally, the Cascade Mountains were removed in the NC (no Cascades) run, with full physics being applied, in an attempt to examine the role of windward ridging on the Cascades in the formation of the PSCZ. The atmospheric volumes that were formerly occupied by mountains were initialized by interpolating the pressure-level model output of the outer domain to appropriate sigma levels in the inner nest.

Fig. 8.

East–west cross sections along line B of Fig. 1 at (a) 0600 and (b) 1200 UTC 26 May, and along line D of Fig. 1 at (c) 0600 and (d) 1200 UTC 26 May 1992. Thick solid lines are isentropes at 2-K interval. Wind vectors represent alongsection winds, with scales shown at the upper-right corner of each plot (horizontal wind: meters per second; vertical velocity: microbars per second). Shaded areas with thin lines (interval: 2 m s−1) denote the northerly [for (a) and (b)] and southerly [for (c) and (d)] wind components.

Fig. 8.

East–west cross sections along line B of Fig. 1 at (a) 0600 and (b) 1200 UTC 26 May, and along line D of Fig. 1 at (c) 0600 and (d) 1200 UTC 26 May 1992. Thick solid lines are isentropes at 2-K interval. Wind vectors represent alongsection winds, with scales shown at the upper-right corner of each plot (horizontal wind: meters per second; vertical velocity: microbars per second). Shaded areas with thin lines (interval: 2 m s−1) denote the northerly [for (a) and (b)] and southerly [for (c) and (d)] wind components.

Table 1.

The experimental design of the MM5 simulations.

The experimental design of the MM5 simulations.
The experimental design of the MM5 simulations.

4. Evolution of the simulated event

In this section, the output of the control experiment for the 5-km inner domain is used to describe and diagnose the three-dimensional structural evolution of the PSCZ event.

a. Surface evolution

Figure 4 presents simulated sea level pressure, surface winds, and 3-h accumulated precipitation for the control experiment. At hour 3 (0300 UTC 26 May; Fig. 4a), the front/trough and the associated rain band are moving inland over Puget Sound and the Strait of Georgia. Higher pressure and westerly/northwesterly winds are developing over the ocean, a lee trough is forming to the east of the Olympics, and strong westerly flow is surging eastward into the Strait of Juan de Fuca. At this time, there is weak low-level convergence over Puget Sound. Three hours later at 0600 UTC 26 May (Fig. 4b), the synoptic front/trough has moved farther inland, accompanied by substantial rainfall over Puget Sound and the western slopes of the Cascades.

Fig. 4.

Simulated sea level pressure (thick solid lines at 1-mb interval, with an additional dashed line at a 0.5-mb interval), winds (m s−1) at the lowest σ level (∼40 m AGL), and 3-h accumulated precipitation (shaded area with thin contour lines starting at 0.5 mm with a 1.5-mm interval) ending at plotted times for (a) 0300, (b) 0600, (c) 0900, (d) 1200, (e) 1500, and (f) 1800 UTC 26 May 1992.

Fig. 4.

Simulated sea level pressure (thick solid lines at 1-mb interval, with an additional dashed line at a 0.5-mb interval), winds (m s−1) at the lowest σ level (∼40 m AGL), and 3-h accumulated precipitation (shaded area with thin contour lines starting at 0.5 mm with a 1.5-mm interval) ending at plotted times for (a) 0300, (b) 0600, (c) 0900, (d) 1200, (e) 1500, and (f) 1800 UTC 26 May 1992.

By 0900 UTC 26 May (Fig. 4c), the winds over the Strait of Juan de Fuca have strengthened, with a maximum wind speed exceeding 10 m s−1 near the eastern terminus of the strait. The low-level incoming flow diverges over the northern sound: one branch moves toward the Strait of Georgia, while the other heads southward toward Puget Sound. Moderate precipitation continues over the western slopes of the Cascades, while drying is apparent to the west. Windward ridging and lee troughing are developing upstream and downstream of the Olympics, respectively. Compared with the observations (cf. Fig. 2c), the model realistically simulates the wind and pressure fields and the distribution of precipitation.

At 1200 UTC 26 May (Fig. 4d), the precipitation over the Cascades has decreased substantially as the front moved eastward out of the domain. The trough in the lee of the Olympics has shifted southward and there is a suggestion of mesoscale pressure ridging to the east of the sound. These pressure changes contribute to an increase of the southwesterly flow over the southern Puget Sound and northerly/northeasterly flow to the north. At this time there is low-level convergence over the central sound that is associated with a band of light rainfall extending from the central sound to the Cascades.

By 1500 UTC 26 May (Fig. 4e), the low-level wind convergence has strengthened significantly, with a distinct precipitation band over the central sound that extends to the east. Considerable precipitation accumulated within this band during the previous 3 h, with a maximum exceeding 5 mm over the western slopes of the Cascades. The simulated precipitation pattern agrees well with observations (cf. Fig. 2e). As shown later in Fig. 11, the band was associated with a zone of enhanced upward motion, reaching 0.33 m s−1 at 850 mb.

Fig. 11.

Simulated cloud water (shaded area is greater than 5 × 10−5 kg kg−1) and vertical motion (contour interval is 0.05 m s−1; thick solid lines: upward motion; dashed lines: downward motion; zero lines are not plotted) at (a) 950, (b) 850, and (c) 750 mb for the control simulation at hour 15 (1500 UTC 25 May 1992).

Fig. 11.

Simulated cloud water (shaded area is greater than 5 × 10−5 kg kg−1) and vertical motion (contour interval is 0.05 m s−1; thick solid lines: upward motion; dashed lines: downward motion; zero lines are not plotted) at (a) 950, (b) 850, and (c) 750 mb for the control simulation at hour 15 (1500 UTC 25 May 1992).

By 1800 UTC 26 May, the onshore flow over the ocean has weakened considerably, resulting in an attenuated windward pressure ridge (Fig. 4f). The lee low appears to have strengthened and moved westward; this change is at least partially the result of an increase of near-surface temperature over high terrain during the late morning, and the resulting effects on sea level pressure reduction. The rainband associated with the convergence zone is still evident, although attenuated in amplitude.

b. Three-dimensional wind structure of the convergence zone

In this section, the three-dimensional wind structure to the lee of the Olympics is presented by showing the horizontal wind vectors at every grid point and streamlines at hour 15 (1500 UTC 26 May; Fig. 5), when the PSCZ was most intense. At both 150 (not shown) and 300 m (Fig. 5a), well-defined convergent airstreams are found over Puget Sound: westerly flow through the Strait of Juan de Fuca turns southward into the sound, while westerly flow moving around the southern flanks of the Olympics backs into the southwest. Easterly flow is found over the central sound, resulting in the appearance of a well-defined cyclonic vortex to the south and an incomplete anticyclonic vortex to the north. Progressing through 600 (not shown) to 1200 m (Fig. 5b), the area of confluence is displaced eastward and two vortices of similar shape and amplitude, one cyclonic and the other anticyclonic, are found over Puget Sound. The pattern changes considerably by 2000 m (Fig. 5c), with the loss of both the easterly flow over the central sound and the adjacent two vortices; complex wind perturbations induced by mountain waves are evident to the southeast of the Olympics, and an area of divergence is located to the east of Puget Sound immediately above the region of strongest low-level convergence and convection. This divergent wind field was still apparent at 2500 m and faded by 3000 m (not shown). At 3500 m, there is little evidence of either orographic or convective effects (not shown).

Fig. 5.

Simulated wind vectors and streamlines near Puget Sound for the control run at (a) 300, (b) 1200, and (c) 2000 m at hour 15 (1500 UTC 26 May 1992). Wind speed is in meters per second. The directions of the streamlines are omitted.

Fig. 5.

Simulated wind vectors and streamlines near Puget Sound for the control run at (a) 300, (b) 1200, and (c) 2000 m at hour 15 (1500 UTC 26 May 1992). Wind speed is in meters per second. The directions of the streamlines are omitted.

It is of interest to compare the flow to the lee of the Olympics to that found downwind of the island of Hawaii, which encompasses a mesoscale orographic barrier similar in horizontal extent to that of the Olympics. As noted in Smolarkiewicz et al. (1988) and Smolarkiewicz and Rotunno (1989), a pair of vortices often appear in the lee of Hawaii during the typical low–Froude number conditions (∼0.2–0.3) of the area. These Hawaiian vortices are similar in both scale and location to those produced by the model over Puget Sound, although one should note that the Froude number of the flow upstream of the Olympics is approximately 0.4 for the above control simulation. Later in this paper, the relative contributions of the Olympics and Cascades to the generation of the Puget Sound vortices and the alteration of these circulations as Froude number varies will be examined.

c. Cross sections

To examine the three-dimensional structural evolution of the PSCZ, vertical cross sections were constructed along and across Puget Sound during the event (Figs. 6 and 7). These sections show isentropes, wind vectors, and cloud water mixing ratio.

Fig. 6.

North–south cross sections along line A in Fig. 1 at (a) 0900, (b) 1200, (c) 1500, and (d) 1800 UTC 26 May 1992 for the control simulation. Thick solid lines are isentropes at 2-K interval. Wind vectors represent flow within the cross section. Wind vector scales are shown at the upper-right corner of each plot (horizontal wind: meters per second; vertical velocity: microbars per second). Shaded areas denote cloud water mixing ratio greater than 5 × 10−5 kg kg−1, with an interval of 1 × 10−4 kg kg−1. Points E–H and M–P in (c) indicate the ending locations of trajectories in Fig. 10.

Fig. 6.

North–south cross sections along line A in Fig. 1 at (a) 0900, (b) 1200, (c) 1500, and (d) 1800 UTC 26 May 1992 for the control simulation. Thick solid lines are isentropes at 2-K interval. Wind vectors represent flow within the cross section. Wind vector scales are shown at the upper-right corner of each plot (horizontal wind: meters per second; vertical velocity: microbars per second). Shaded areas denote cloud water mixing ratio greater than 5 × 10−5 kg kg−1, with an interval of 1 × 10−4 kg kg−1. Points E–H and M–P in (c) indicate the ending locations of trajectories in Fig. 10.

Fig. 7.

East–west cross sections along line C in Fig. 1 at (a) 0900, (b) 1200, (c) 1500, and (d) 1800 UTC 26 May 1992. Presented fields and contour conventions are the same as for Fig. 6.

Fig. 7.

East–west cross sections along line C in Fig. 1 at (a) 0900, (b) 1200, (c) 1500, and (d) 1800 UTC 26 May 1992. Presented fields and contour conventions are the same as for Fig. 6.

By hour 9 (0900 UTC 26 May; Fig. 6a), the convergence zone has begun to form near the 125-km mark of the north–south section (along line A of Fig. 1), resulting in weak upward motion and enhanced cloud water to approximately 2400 m. To the north of the convergence zone, there is a region of clear skies associated with subsiding motion. Three hours later at 1200 UTC 26 May (Fig. 6b), both the low-level southerly and northerly wind components over the southern and northern sound, respectively, have increased substantially, resulting in enhanced low-level convergence, increased upward motion, and enhanced cloud water; this area of enhanced upward motion and cloud water slopes upward toward the north. Between 850 and 700 mb, the flow diverges and subsides to the north and south, producing cloud-free regions. At the northern and southern ends of the cross section, cloud water mixing ratio is slightly larger due to the lifting effect of the Cascades on the westerly flow.

Between 1200 and 1500 UTC 26 May (Fig. 6c), the convergence zone remains nearly stationary as it reaches its strongest stage. The upward motion is stronger, deeper, and less tilted at this time. To the north and south, compensating outflow is found mainly between approximately 850 and 700 mb, with stronger subsidence to the north. At points M and P, where weak upward motion is found, the absence of cloud is due to the dry midtropospheric origin of the air parcels reaching these locations (as will be shown later in the trajectory section). By 1800 UTC 26 May (Fig. 6d), the PSCZ circulation has weakened and moved northward, with a concomitant reduction in vertical motion and clouds.

Figure 7 shows a series of east–west cross sections through the Olympics along line C in Fig. 1. At 0900 UTC 26 May (Fig. 7a), when the PSCZ is developing, westerly flow moves over the crest of the Olympics and subsides on the barrier’s eastern slopes, producing clear skies and a weak lee trough (cf. Fig. 4c). Upslope flow and clouds are apparent on the windward (western) slopes of the Olympics and Cascades, with a lobe of enhanced cloudiness over Puget Sound associated with the incipient PSCZ.

Three hours later at 1200 UTC 26 May (Fig. 7b), the PSCZ and its associated upward motion have intensified, resulting in enhanced cloud water over the sound. Subsidence to the lee of the Olympics has attenuated, while to the east the upslope flow on the Cascades and associated cloud water have lessened. At 1500 UTC 26 May (Fig. 7c), the PSCZ produces strong vertical motion over the sound, with associated clouds extending to 700 mb. In the lee of the Olympics, clear skies are associated with subsidence. By 1800 UTC 26 May (Fig. 7d), the PSCZ has substantially weakened, and only scattered clouds are present over the Puget Sound area. Strong daytime heating has resulted in the development of a dry-adiabatic layer over the Olympics.

To examine the eastward progression of the westerly flow and its interaction with the Cascades, east–west cross sections were made along the Strait of Juan de Fuca and the Chehalis Gap (lines B and D of Fig. 1) north and south of the Olympics, respectively. Shading indicates northerly winds in cross section B (Figs. 8a,b) and southerly winds in cross section D (Figs. 8c,d). Along the Strait (section B) at hour 6 (0600 UTC 26 May; Fig. 8a), the leading edge of the strong westerly flow and associated upward motion reflects the position of the synoptic front. To the east of this transition, low-level airflow turns southward as it approaches the Cascades. Six hours later at 1200 UTC 26 May (Fig. 8b), the westerly flow has reached the foothills of the Cascades and the low-level northerly component has increased within about 75 km of the barrier. There are at least two reasons why the flow turns northerly when approaching the Cascades: first, the lee trough that forms to the east of the Olympics generates a southward pressure gradient force over the northern sound, as will be discussed in the next section. Second, the northwest–southeast orientation of the Cascades in this area blocks the airflow, producing a weak windward pressure ridge that contributes to a southward deflection.

Cross section D at hour 6 (0600 UTC 26 May; Fig. 8c) shows that the low-level air that flows around the southern flank of the Olympics turns southwesterly within 75 km of the Cascades. At this time, strong upward motion associated with frontal passage is found aloft and there is the suggestion of damming of cooler air west of the Cascades. Six hours later at 1200 UTC 26 May (Fig. 8d), weak damming is still evident in the cross section, resulting in a weak windward surface pressure ridge (cf. Fig. 4d) and enhanced low-level southerly flow that strengthens the PSCZ.

d. Force balances

To document the changing force balances experienced by air parcels as they flow around the Olympics during a convergence zone event, the magnitudes of the terms in the momentum equation were diagnosed. The nonhydrostatic equations used in the diagnosis are the same as Eqs. (2.2.1) and (2.2.2) of Grell et al. (1994) except that p* is decoupled.5 For simplicity, the momentum equation is summarized as

 
formula

where d/dt is the total derivative, which includes local time derivative and 3D advection terms; PGF represents the pressure gradient force; and the residual term includes diffusion, friction, vertical entrainment, and numerical errors. The time differential (or local change) of wind is approximated with a centered time difference (Δt = 30 min).

Figure 9 presents the force balances of the above terms at the model’s lowest sigma level (∼40 m AGL) at several grid points surrounding the Olympic Mountains at 1200 UTC 26 May. At this time, strong westerly to northwesterly flow, roughly in geotriptic balance (between Coriolis, pressure gradient, and friction forces), is present at nearshore points A and G. With an eastward bulge of sea level pressure along the central Strait of Juan de Fuca and troughing to the lee (east) of the Olympics, pressure gradient forces at points B and C are directed to the southeast and south, respectively. Since the Coriolis force is also directed toward the south at these points, there are large southward and southwestward flow accelerations at B and C, respectively. At point D, the pressure gradient force is directed nearly along the wind vector to the south-southeast, resulting in a strong northerly downgradient wind. At points H and I, the pressure gradient force is larger than the Coriolis force and the residual terms, resulting in a northward acceleration toward the lee trough. At point F, the northward pressure gradient force resulting from both troughing in the lee of the Olympics and windward ridging on the Cascades produces a veering of the surface wind to the northeast (southwesterly wind).

Fig. 9.

Simulated sea level pressure (solid lines, 0.5-mb interval) and force balances at the lowest σ level (σ = 0.995) for grid points surrounding the Olympics at 1200 UTC 26 May 1992. The positions of the grid points are shown in the inset at the lower-left corner. Vectors (1–5) at each grid point represent the wind vector as well as the total derivative, residual, Coriolis, and pressure gradient terms in the momentum equation, respectively. Their magnitudes and the corresponding lengths are shown at the bottom. Shading denotes the area where the model terrain is higher than 300 m.

Fig. 9.

Simulated sea level pressure (solid lines, 0.5-mb interval) and force balances at the lowest σ level (σ = 0.995) for grid points surrounding the Olympics at 1200 UTC 26 May 1992. The positions of the grid points are shown in the inset at the lower-left corner. Vectors (1–5) at each grid point represent the wind vector as well as the total derivative, residual, Coriolis, and pressure gradient terms in the momentum equation, respectively. Their magnitudes and the corresponding lengths are shown at the bottom. Shading denotes the area where the model terrain is higher than 300 m.

e. Trajectories, vertical motion, and the distribution of clouds

To provide a description of the three-dimensional air flow associated with the convergence zone event, and to gain insight into the associated complex mesoscale distributions of clouds and precipitation, backward trajectories were released at low levels (σ = 0.99 and 0.86) over the Puget Sound area at hour 15. The trajectories were integrated backward toward the model initialization time using the three-dimensional pressure and wind output from the control simulation. Then, the trajectories are compared to the vertical motion and cloud water fields at several levels.

Figures 10a and 10b show trajectories that end near the surface at σ = 0.99 (∼70 m AGL) to the east of the Olympics. Trajectory A, which ended to the northeast of the Olympics, began over the ocean at approximately 900 mb, approached and was deflected north of the Olympics between hours 6 and 12, and then descended over the northeast corner of the Olympic Peninsula. Trajectory B started near the surface over the ocean, ascended rapidly over the northwest side of the Olympics while being deflected to the north, descended over the northern side of the barrier, and then veered to the south while rising and sinking over the northeast side of the Olympics. Trajectory C originated near 800 mb southwest of the Olympics and then descended rapidly to near the surface over the eastern slopes of the barrier. Its cyclonic turning during descent is a result of being caught up in the southern vortex to the lee of the Olympics (cf. Fig. 5). In contrast, trajectory D rose considerably on the windward side of the Olympics, later descending to the surface on the southeast lee slopes. Trajectories E–H, released at the same level but farther to the east, experienced a considerably different history. In general, these air parcels remained at low levels as they were deflected around the Olympics and later converged over Puget Sound. A minor exception to this characterization applies to trajectory E, which started near 900 mb and descended to the surface while being deflected to the north of the Olympics.

Fig. 10.

Backward trajectories released at σ = 0.99 [∼70 m AGL; (a) and (b)], and σ = 0.86 [∼1250 m AGL; (c) and (d)], ending at hour 15 (1500 UTC 26 May 1992) and starting at the initial time (0000 UTC 26 May 1992). The width of the trajectories indicates pressure height according to the legend. The trajectory positions every 3 h are indicated by short lines. Shading represents the area where model terrain is higher than 200 m, with a 200-m interval.

Fig. 10.

Backward trajectories released at σ = 0.99 [∼70 m AGL; (a) and (b)], and σ = 0.86 [∼1250 m AGL; (c) and (d)], ending at hour 15 (1500 UTC 26 May 1992) and starting at the initial time (0000 UTC 26 May 1992). The width of the trajectories indicates pressure height according to the legend. The trajectory positions every 3 h are indicated by short lines. Shading represents the area where model terrain is higher than 200 m, with a 200-m interval.

Trajectories launched at σ = 0.86 (∼1250 m AGL) are presented in Figs. 10c and 10d. Trajectory I approached northern Puget Sound from the northwest and experienced modest rising and sinking as it crossed the mountains of Vancouver Island. Trajectories J–L evinced little vertical displacement as they approached and crossed the Olympics, and only modest subsidence on the lee side. Parcels N and O, which were nearly saturated as they began at low levels northwest of the Olympics (see Table 2), rose slowly while being deflected around the Olympics, and then converged and rose rapidly over the central sound. In contrast, the adjacent parcels (M and P), which originated at high levels (∼800 mb) within relatively dry air, experienced only modest deflection in either the horizontal or vertical. The different origins of the above trajectories explain the observed cloud pattern at this level (cf. Fig. 6c), with clouds associated with the rapidly rising trajectories N and O, and cloud-free conditions at M and P related to their drier origins.

Table 2.

Relative humidity of air parcels along trajectories M–P of Fig. 10d. At hours 0 and 3, trajectories M and P are out of the domain.

Relative humidity of air parcels along trajectories M–P of Fig. 10d. At hours 0 and 3, trajectories M and P are out of the domain.
Relative humidity of air parcels along trajectories M–P of Fig. 10d. At hours 0 and 3, trajectories M and P are out of the domain.

Vertical velocity and cloud water fields at 950 mb at 1500 UTC 26 May (Fig. 11a) reflect the subsidence on the northeastern and southeastern sides of the Olympics shown in the trajectories above, as well as the strong upward motion and large cloud water near the central sound associated with the low-level converging airstreams. In addition, low marine clouds extend over the coastal zone.

The strong updraft near the central sound, which is the result of low-level wind convergence, strengthens considerably at 850 mb and stretches eastward over the western slopes of the Cascades (Fig. 11b). To the lee (east) of the Olympics and the Cascades, there is generally downward motion at 850 mb; these subsiding areas closely correspond to the regions of clear or nearly clear skies in the simulated cloud water field and in the satellite imagery (cf. Fig. 3). The shallow nature of the convergence zone circulations is evident at 750 mb, where the PSCZ updraft is attenuated and split (Fig. 11c). There is the suggestion of a wavelike feature over the eastern slopes of the Olympics at this level, with rising motion over the lee slopes.

5. Sensitivity studies

Five sensitivity experiments were conducted to examine the importance of latent heat release and the topography surrounding Puget Sound (the Olympics and the Cascades) in the formation of the PSCZ. Because the PSCZ event was strongest between 1200 and 1500 UTC 26 May, only the simulation results at hour 15 (1500 UTC 26 May) for each experiment are presented in the following discussions.

a. Impact of latent heat release

To examine the impact of latent heat release in the formation of the PSCZ, the no latent heat (NLH) simulation was completed. Figures 12a (sea level pressure, 3-h precipitation, and 40-m winds at 1500 UTC 26 May) and 12b (the difference in 40-m wind and sea level pressure between the control and NLH experiments) show that without latent heating the lee trough to the east of the Olympics is much weaker, flow across the barrier is decreased, and low-level wind convergence near the central sound is attenuated compared to the control run (cf. Fig. 4e); in addition, there is lighter and more widespread precipitation over Puget Sound and the western slopes of the Cascades. Sea level pressure over the Cascades is generally lower in the control run than in the NLH simulation since latent heating produces hydrostatic pressure falls over the windward slopes of the barrier.

Fig. 12.

Model results of the NLH (no latent heat release) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) difference fields (control minus NLH) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; (c) north–south cross section A; and (d) east–west cross section C. The fields in (a), (c), and (d) are the same as in the corresponding figures of the control run.

Fig. 12.

Model results of the NLH (no latent heat release) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) difference fields (control minus NLH) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; (c) north–south cross section A; and (d) east–west cross section C. The fields in (a), (c), and (d) are the same as in the corresponding figures of the control run.

Figure 12c shows the north–south cross section for the NLH run at 1500 UTC 26 May. In this section, the low-level northerly winds over the northern sound are weaker and shallower than in the control run (cf. Fig. 6c), and there is greatly attenuated vertical motion above the convergence zone. Without latent heat release, low-level temperatures between approximately 750 and 950 mb are considerably cooler than in the control run, resulting in a more stable layer between 750 and 850 mb that constrains the vertical development of the PSCZ. Also note that the two clear zones to the north and south are gone because of both a weakening of compensating subsidence around the attenuated convection and a decrease in orographic subsidence in the lee of the barrier (suggested in cross section C) as more air flows around rather than over. The east–west cross section C of the NLH run (Fig. 12d) also demonstrates that low-level upward motion near the convergence zone is considerably shallower than the control experiment (cf. Fig. 7c). On the windward side of the Olympics, stability between approximately 920 and 850 mb is larger than in the control run. As a result, little flow crosses the crest at low levels and subsidence to the lee of the Olympics is very weak in the NLH run. With attenuated subsidence, the lee trough is weaker and clouds extend to the eastern slopes of the Olympics. These differences indicate that latent heat release is very important 1) in generating strong updrafts in the PSCZ, especially around 850 mb, and 2) in producing clear zones to the north and south of the PSCZ by enhancing downslope on the Olympics and compensating subsidence around the updraft.

b. Impact of the Olympic Mountains

To examine the importance of the Olympics in the formation of the PSCZ and the associated flow structure under different flow regimes, the height of the Olympic Mountains was varied from zero (NO) and half (HO) to double (DO) the value used in the control run.

1) No Olympics

Without the Olympics at hour 15, there is no lee trough and the flow slowly backs from northwesterly on the coast to westerly over western Washington (Fig. 13a). The convergence zone and associated rainband are no longer present; over the southern sound only a weak southerly component is apparent, probably representing the blocking effects of the Cascades. Over the western slopes of the Cascades, moderate precipitation has accumulated during the previous 3 h, mainly because of upslope flow on the Cascades. The differences between the control and NO runs at 1500 UTC 26 May for both sea level pressure and surface wind (Fig. 13b) show that the Olympics generate a windward ridge and a lee trough to the northwest and the southeast of the barrier, respectively. The trough induces strong northerly flow into the northern sound, easterly flow over the central sound, and weak southerly flow over the southern sound.

Fig. 13.

Model results of the NO (no Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) difference fields (control minus NO) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; (c) north–south cross section A; and (d) east–west cross section C. The fields in (a), (c), and (d) are the same as in the corresponding figures of the control run.

Fig. 13.

Model results of the NO (no Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) difference fields (control minus NO) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; (c) north–south cross section A; and (d) east–west cross section C. The fields in (a), (c), and (d) are the same as in the corresponding figures of the control run.

The north–south cross section A along Puget Sound for the NO run at hour 15 (Fig. 13c) shows that without the Olympics low-level northerly flow is absent over the northern sound, and, instead, only weak southerly flow is present at low levels along the entire section. In the east–west cross section across the central sound (Fig. 13d), strong upward motion over the western slopes of the Cascades produces deep clouds and moderate rainfall. Without low-level convergence as found in the control experiment (cf. Fig. 7c), deep clouds over the Puget Sound region are absent.

2) Half Olympics

In the HO experiment the height of the Olympics is halved, resulting in an increase of the Froude number to approximately 1 (it was about 0.4 in the control run). The results of HO at hour 15 (Fig. 14a) show less deflection of the low-level airflow over the windward side of the Olympics than in the control experiment (cf. Fig. 4e), and both the windward ridge and the lee trough of the Olympics are attenuated. Because of the weaker lee trough, the northerly and southerly wind components over the northern and southern sound are significantly reduced in magnitude. The wind converges farther downstream, resulting in a weaker convergence zone over the central sound and the absence of precipitation immediately east of the Olympics. Over the western slopes of the Cascades, 3-h precipitation and winds are quite similar to that of the control experiment.

Fig. 14.

Model results of the HO (half Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) north–south cross section A; and (d) east–west cross section C. The fields in each panel are the same as in the corresponding figures of the control run.

Fig. 14.

Model results of the HO (half Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) north–south cross section A; and (d) east–west cross section C. The fields in each panel are the same as in the corresponding figures of the control run.

Figure 14b shows that low-level winds at 300 m ASL (above sea level) over the northern (southern) sound of the HO run have less northerly (southerly) components than in the control run (cf. Fig. 5a). The convergence is weak over the central sound, and there is no easterly flow. Stronger convergence occurs farther downstream over the western slopes of the Cascades than in the control run. Because of the higher Froude number in this simulation, there are no vortices to the lee of the Olympics.

The north–south cross section A for the half-Olympics simulation (Fig. 14c) shows a very weak and shallow convergence zone over the sound at 1500 UTC 26 May. In the east–west cross section C (Fig. 14d), considerably more air flows over the Olympics because of the larger Froude number. The areas of strongest upward motion and deepest clouds have been displaced eastward toward the Cascades in this simulation compared to the control run (cf. Fig. 7c). Also note that the amplitude of the mountain waves in the HO run appears to be larger than in the control simulation.

3) Double Olympics

When the height of the Olympic Mountains is doubled, the Froude number of the incoming flow drops to approximately 0.20, roughly the same value as observed during many of the Hawaiian studies. The upstream wind speed and static stability are nearly the same as in the control run. Figure 15a shows that low pressure is centered over the crest of the Olympics (which may be at least partially spurious due to the extra terrain height in the sea level pressure reduction calculation), precipitation is attenuated and is more amorphous than that found in the control run (Fig. 4e), and a well-defined convergence zone is apparent in the low-level wind field (also see Fig 15b). As in the control run (see Fig. 5), easterly flow is found over the central sound and a dual vortex structure is seen to the east of the Olympics (Fig. 15b). However, the southerly wind component over the southern sound in the DO run is weaker than in the control run, resulting in a southward displacement of both the convergence zone and the easterly flow. The vortices become better defined at 1200 m ASL (Fig. 15c) and extend farther downstream than in the control run.

Fig. 15.

Model results of the DO (double Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) 1200-m winds; (d) north–south cross section A; and (e) east–west cross section C. The fields in each panel are the same as in the corresponding figures of the control run.

Fig. 15.

Model results of the DO (double Olympics) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) 1200-m winds; (d) north–south cross section A; and (e) east–west cross section C. The fields in each panel are the same as in the corresponding figures of the control run.

The north–south cross section for the DO run at 1500 UTC 26 May (Fig. 15d) shows that the low-level southerly wind component over the southern sound is somewhat weaker than in the control run and the low-level convergence is displaced slightly toward the south. Convection in the convergence zone is weaker and its associated clouds are not as narrow and deep as in the control run. The east–west cross section C at hour 15 (1500 UTC 26 May) indicates that with the double-height Olympics there is less subsidence in the lee of the Olympics since most of the air flows around the mountains rather than over (Fig. 15e). As a result, weak updrafts and large cloud water extend farther westward toward the eastern foothills of the Olympics. Over the western slopes of the Cascades, the upslope wind components are significantly weaker than in the control run, resulting in less low-level clouds and precipitation over this region (also see Fig. 15a).

In the above HO and DO simulations, the height of the mountains was varied to model different flow regimes. In terms of Froude number, they are equivalent to simulations with twice (HO) and half (DO) upstream wind speed of the control run, if the mountains were the same and the stability did not vary. To determine whether this range (0.2–1) of Froude number occurs in nature, Froude numbers were computed from the soundings at Quillayute (UIL in Fig. 1) for nearly an entire year. It was found that a wide range of Froude numbers (0.1–2) did occur for westerly and northwesterly flow regimes, and some of the low–Froude number events (Fr ∼ 0.2) were associated with weak PSCZ.

c. Impact of the Cascade Mountains

The realistic Olympics were retained and the Cascades were removed in the NC run. The Froude number of this experiment increased to about 0.45 because of a slight increase in upstream wind speed compared to the control run. Without the Cascades, the mesoscale pressure ridging to the east of Puget Sound disappeared at hour 15 (Figs. 16a and 4e). Similar to the control experiment, the Olympics deflect the low-level incoming flow into two branches, but, in contrast, the convergence over the sound is weakened. Without the Cascades, southerly flow over the southern sound is no longer evident, the northerly flow over the northern sound has been attenuated, and precipitation is absent both within the convergence zone and on the windward slopes of the Cascades (Figs. 16a,b). Without the Cascades, the two vortices, and particularly the southern one, have considerably weakened at 300 m and have become more similar in amplitude (Figs. 16b and 5a). It is interesting to note that a run with no Cascades and a doubled Olympics, which had an associated upstream Froude number of about 0.25, did have two strong symmetrical vortices to the lee of the Olympics (not shown).

Fig. 16.

Model results of the NC (no Cascades) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) difference fields (control minus NC) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; and (d) north–south cross section A. The fields in (a), (b), and (d) are the same as in the corresponding figures of the control run.

Fig. 16.

Model results of the NC (no Cascades) run at hour 15 (1500 UTC 26 May 1992). (a) Surface winds, sea level pressure, and 3-h precipitation; (b) 300-m winds; (c) difference fields (control minus NC) of sea level pressure (0.25-mb contour interval, with negative contours dashed) and wind vectors (m s−1) at the lowest sigma level; and (d) north–south cross section A. The fields in (a), (b), and (d) are the same as in the corresponding figures of the control run.

The influence of the Cascades is highlighted in Fig. 16c, which presents the sea level pressure and surface wind differences between the control and NC runs. The Cascades enhance sea level pressure on their windward slopes, especially downwind of the major paths to the Pacific north and south of the Olympics. This orographic pressure ridging enhances southerly flow over the southern sound and easterly flow over the central and northern sound.

The north–south cross section A for the NC run (Fig. 16d) shows that northerly flow is generated over the northern sound, accelerating toward the trough in the lee of the Olympics. However, without the Cascades, there is no southerly flow over the southern sound and the PSCZ is significantly weaker. In fact, only speed convergence is present in low levels and convection and the associated clouds are absent.

These results suggest that the Cascade Mountains are important in the formation of the PSCZ because they produce windward ridging that contributes to 1) the easterly component that turns northwesterly winds into northerly or northeasterly flow over the northern sound and 2) a southerly component over the southern sound. Blocking by the Cascades also clearly contributes to the enhancement of precipitation within the PSCZ.

6. Discussion and conclusions

This paper describes an observational and numerical study of a Puget Sound convergence zone (PSCZ) event that formed over the Puget Sound of Washington State after a synoptic front passed to the east. Other regional effects associated with this frontal passage, such as the prefrontal onshore surge of marine air, thermal troughing, and coastal southerlies were documented in an earlier paper (Chien et al. 1997). The PSU–NCAR MM5 was used to examine the three-dimensional structure and origin of the PSCZ. In addition to a control experiment, several sensitivity experiments were conducted to determine the importance of latent heat release, the Olympic Mountains, and the Cascade Mountains in the formation of the PSCZ. It was found that the MM5 could successfully simulate this PSCZ event, duplicating most aspects of the observed wind and precipitation fields.

As the coastal winds veered from a southerly to a westerly/northwesterly direction following frontal passage, the flow was deflected by the Olympics into two branches: one passing through the Strait of Juan de Fuca and the other around the southern flank of the Olympics. At the same time, troughing developed over the eastern slopes of the Olympics. This troughing, coupled with pressure ridging on the windward slopes of the Cascades, resulted in the southward and westward acceleration of the airstream that had come through the Strait of Juan de Fuca. Over the southern sound, windward pressure ridging on the Cascades, coupled with the Olympic lee trough, accelerated the flow coming around the southern flank of the Olympics toward the northeast. The convergence of the two airflows over the central sound resulted in the PSCZ, with clouds and rainfall over the central Puget Sound area and zones of clear/partly cloudy skies to the south and north. The above points are supported by sensitivity experiments in which either the Olympics or the Cascades are removed, and it is concluded that the Olympics and the Cascades are both important in the formation of the PSCZ.

It is also found that latent heat release is important in the formation of the PSCZ. The low-level convergence associated with the PSCZ produces upward motion, condensation, and latent heat release in the lower troposphere. Because of this latent heat release, the low-level atmosphere becomes less stable, providing a more favorable environment for further strengthening of convection, low-level convergence, and compensating subsidence to the north and the south. In addition, orographically forced subsidence to the lee of the Olympics and the mountains of Vancouver Island contribute to the development of clear zones to the north and south of the PSCZ. Without latent heating, the air on the windward side of the Olympics is more stable and thus experiences greater deflection around the Olympics, resulting in a weakening of the subsidence and troughing to the lee of the barrier.

Detailed low-level wind analyses for the simulated PSCZ event show that a pair of vortices appear in the lee of the Olympics up to approximately 1200 m. Between the vortices, well-defined easterly flow—predominantly forced by the Olympics lee trough—is found over the central sound. Above 1200 m, the vortices are no longer evident and divergent winds develop above the zone of low-level convergence. Comparisons of this vortex pattern with those observed or modeled downwind of the island of Hawaii (Smolarkiewicz et al. 1988; Smolarkiewicz and Rotunno 1989), a mesoscale orographic barrier similar in horizontal extent to that of the Olympics, show similar structures in both scale and location. This similarity is evident even though the Froude number of the flow upstream of the Olympics is approximately 0.4 for the above control simulation, compared to the 0.2–0.3 that is typical for the flow approaching the island of Hawaii. As noted below, this is consistent with Hunt and Snyder (1980), who found that a Froude number of 0.4 is still within the lee vortex regime.

For the experiment in which the Olympics were reduced by half, the Froude number becomes larger (∼1). With the reduced topography, the trough in the lee of the Olympics is weaker than in the control run, there is less horizontal deflection, more of the low-level flow moves over the Olympic Mountains, and flow convergence occurs farther downstream. Because of the large Froude number, there is no vortex formation in the lee of the Olympics for this experiment. In another experiment in which the height of the Olympics was doubled (Fr ∼ 0.19), most air tends to move around the mountains and converges in the immediate lee of the barrier. The convergence is weaker than in the control run, and the vortices occur farther downstream of the Olympics.

In a sensitivity experiment in which the Cascades are removed and the realistic Olympics are retained, the Olympics act much like an isolated three-dimensional barrier such as the island of Hawaii. As expected, the simulated vorticity structure appears to be more symmetric than in the control run, but the two vortices are considerably weakened because of the slightly larger Froude number (∼0.45) of this run. Decreasing the Froude number to approximately 0.25 by doubling the height of the Olympics restores the two vortices, each similar in amplitude and shape. These vortex patterns are quite consistent with the laboratory experiments of Hunt and Snyder (1980), in which vortices do not appear in the lee of the barrier when Fr = 1. For smaller Froude numbers such as Fr = 0.4 and 0.2, lee vortices were observed. Our study suggests that a necessary condition for the appearance of two lee vortices to the east of Olympic Mountains is Fr < 0.45.

The authors have observed many weak PSCZ events in which strong westerly flow through the Strait of Juan de Fuca does not occur. As suggested in Fig. 13b (which displayed the differences between the control and the no Olympics runs), troughing in the lee of the Olympics alone can induce modest convergence in the central sound, resulting in a weak PSCZ event, especially if the lower troposphere is moist and relatively unstable. Strong westerly flow in the Strait of Juan de Fuca contributes to the intensification of the PSCZ since the strongest events are usually associated with such onshore flow. In fact, many PSCZ events appear to occur in two steps. First, as the large-scale winds turn westerly/northwesterly, generally during or immediately after frontal passage, lee troughing develops in the lee of the Olympics and weak convergence is found over the central sound, with weak easterlies to the north and southwesterlies to the south. Convergence-zone-related precipitation is generally weak at this stage. Second, as strong flow moves through the strait and is deflected southward into the preexisting convergence zone, the convergence and associated precipitation are greatly strengthened. To a considerable degree this evolution was evident during the May 1992 event presented above.

This paper has not explored convective effects such as propagating gust fronts emanating from the orographically induced convection that occurs during PSCZ events. Recent cases viewed with Doppler radar imagery (WSR-88D) have revealed wind shifts associated with such gust fronts moving southward into the sound. These southerly to northerly wind shifts are sometimes taken as convergence zone passage and can initiate weak to moderate convection if the flow is sufficiently unstable.

Fig. 2.

(Continued)

Fig. 2.

(Continued)

Fig. 4.

(Continued)

Fig. 4.

(Continued)

Fig. 15.

(Continued)

Fig. 15.

(Continued)

Acknowledgments

This research was supported by the National Science Foundation under Grants ATM-9111011 and ATM 94-16866 and the ONR Coastal Meteorology Accelerated Research Initiative (Grant NH45543-4454-44). Use of the MM5 was made possible by the Microscale and Mesoscale Meteorological Division of the National Center for Atmospheric Research. The mesoscale model was run at the Scientific Computing Division of NCAR. The authors would like to express their gratitude to Ying-Hwa (Bill) Kuo for support and helpful discussions, and three anonymous reviewers for useful recommendations.

REFERENCES

REFERENCES
Aanensen, C. J., 1965: Gales in Yorkshire in February 1962. Geophys. Mem., 108, 1–44.
Baines, P. G., 1979: Observations of stratified flow past three-dimensional barriers. J. Geophys. Res., 84, 7834–7838.
Blackadar, A. K., 1979: High resolution models of the planetary boundary layer. Advances in Environmental Science and Engineering, Vol. 1, Pfafflin and Ziegler, Eds., Gordon and Breach, 50–85.
Chien, F.-C., and C. F. Mass, 1994: A numerical study of the interaction between frontal systems and coastal mountains. Proc. Fourth PSU/NCAR Mesoscale Model Users’ Workshop, Boulder, CO, NCAR, 78–82.
——, ——, and Y.-H. Kuo, 1997: Interaction of a warm-season frontal system with the coastal mountains of the western United States. Part I: Prefrontal onshore push, coastal ridging, and alongshore southerlies. Mon. Wea. Rev., 125, 1705–1729.
Colle, B. A., and C. F. Mass, 1996: An observational and modeling study of the interaction of low-level southwesterly flow with the Olympic Mountains during COAST IOP 4. Mon. Wea. Rev., 124, 2152–2175.
Drazin, P. G., 1961: On the steady flow of fluid of variable density past an obstacle. Tellus, 13, 239–251.
Dudhia, J., 1993: A nonhydrostatic version of the Penn State/NCAR mesoscale model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev., 121, 1493–1513.
Ferber, G. K., and C. F. Mass, 1990: Surface pressure perturbations produced by an isolated mesoscale topographic barrier. Part II: Influence on regional circulation. Mon. Wea. Rev., 118, 2597–2606.
Grell, G. A., 1993: Prognostic evaluation of assumptions used by cumulus parameterizations. Mon. Wea. Rev., 121, 764–787.
——, J. Dudhia, and D. R. Stauffer, 1994: A description of the fifth-generation Penn State/NCAR mesoscale model (MM5). NCAR/TN-398+STR, 138 pp. [Available from MMM Division, NCAR, P.O. Box 3000, Boulder, CO 80307.]
.
Hunt, J. C. R., and W. H. Snyder, 1980: Experiments on stably and neutrally stratified flow over a model three-dimensional hill. J. Fluid Mech., 96, 671–704.
Klemp, J. B, and D. R. Durran, 1983: An upper boundary condition permitting internal gravity wave radiation in numerical mesoscale models. Mon. Wea. Rev., 111, 430–444.
Mass, C. F., 1981: Topographically forced convergence in western Washington State. Mon. Wea. Rev., 109, 1335–1347.
——, and D. P. Dempesy, 1985a: A topographically forced convergence line in the lee of the Olympic Mountains. Mon. Wea. Rev., 113, 659–663.
——, and ——, 1985b: A one-level, mesoscale model for diagnosing surface winds in mountainous and coastal regions. Mon. Wea. Rev., 113, 1211–1227.
——, and G. K. Ferber, 1990: Surface pressure perturbations produced by an isolated mesoscale topographic barrier. Part I: General characteristics and dynamics. Mon. Wea. Rev., 118, 2579–2596.
NOAA, 1992: Hourly precipitation data, Washington and Oregon, Vol. 42 (5).
Reed, R. J., 1980: Destructive winds caused by an orographically induced mesoscale cyclone. Bull. Amer. Meteor. Soc., 61, 1346–1355.
Smith, R. B., 1979: The influence of mountains on the atmosphere. Advances in Geophysics, Vol. 21, Academic Press, 187–230.
——, 1980: Linear theory of a stratified hydrostatic flow past an isolated mountain. Tellus, 32, 348–364.
——, 1981: An alternative explanation for the destruction of the Hood Canal Bridge. Bull. Amer. Meteor. Soc., 62, 1319–1320.
——, 1982: Synoptic observations and theory of orographically disturbed wind and pressure. J. Atmos. Sci., 39, 60–70.
——, 1988: Linear theory of stratified flow past an isolated mountain in isosteric coordinates. J. Atmos. Sci., 45, 3889–3896.
Smolarkiewicz, P. R., and R. Rotunno, 1989: Low Froude number flow past three-dimensional obstacles. Part I: Baroclinic generated lee vorticities. J. Atmos. Sci., 46, 1154–1164.
——, R. M. Rasmussen, and T. L. Clark, 1988: On the dynamics of Hawaiian cloud bands: Island forcing. J. Atmos. Sci., 45, 1872–1905.
Steenburgh, W. J., and C. F. Mass, 1996: The interaction of an intense midlatitude cyclone with coastal orography. Mon. Wea. Rev., 124, 1329–1352.
Walter, B. A., and J. E. Overland, 1982: Response of stratified flow in the lee of the Olympic Mountains. Mon. Wea. Rev., 110, 1458–1473.
Wilczak, J. M., and J. W. Glendening, 1988: Observations and mixed-layer modeling of a terrain-induced mesoscale gyre: The Denver cyclone. Mon. Wea. Rev., 116, 1599–1622.

Footnotes

Corresponding author address: Dr. Clifford Mass, Department of Atmospheric Sciences, University of Washington, P.O. Box 351640, Seattle, WA 98195-1640.

1

COAST is a field program that was held along the northwestern coast of the United States during November and December of 1993.

2

The Froude number: Fr = U(hmN)−1, where U is the wind component approaching the barrier, hm is the height of the obstacle, and N is the buoyancy frequency [(g/θ̄)∂θ/∂z]1/2.

3

Specifically, higher resolution terrain data was used to calculate the subgrid-scale terrain roughness surrounding the grid points. Then, the friction velocity at each grid point was increased by a factor proportional to the subgrid-scale roughness.

4

The vertical coordinate σ is defined as (ppt) (pspt)−1, where p is pressure, ps is surface pressure, and pt is a constant pressure at the top of the model (50 mb). Here, σ = 0.995, 0.985, 0.97, 0.945, 0.91, 0.87, 0.83, 0.79, 0.75, 0.71, 0.67, 0.63, 0.59, 0.55, 0.51, 0.47, 0.43, 0.39, 0.35, 0.31, 0.27, 0.23, 0.19, 0.14, 0.11, 0.07, and 0.025.

5

Here, p∗ = pspt, where ps is surface pressure and pt is pressure at the top of the model.