1. Introduction
Forecasting the maintenance of mesoscale convective systems (MCSs) is a well-known operational problem and has served as the focus of numerous studies (e.g., Gale et al. 2002; Coniglio et al. 2006, 2007). Such studies were often performed with the goal of better understanding the fundamental dynamics of MCSs as well as anticipating their potential hazards. For simplicity, few of these studies accounted for the impacts of topography, even though organized convective systems are known to encounter complex terrain (e.g., Chen and Chou 1993; Keighton et al. 2007). Explorations of topography and organized convection have increased in recent years, including both observationally based (e.g., Keighton et al. 2007; Parker and Ahijevych 2007) and modeling-based studies (e.g., Frame and Markowski 2006, hereafter FM06; Reeves and Lin 2007).
Letkewicz and Parker (2010, hereafter LP10) built upon the work of Keighton et al. (2007) by examining 40 squall lines observed to encounter the Appalachian Mountains. The squall lines were categorized based on whether or not they were able to continue to produce severe weather in the lee of the terrain. “Crossing” cases were associated with severe weather reports both upstream and downstream of the mountains, while “noncrossing” cases only produced severe weather upstream of the barrier, and often dissipated upon encountering the terrain (Keighton et al. 2007). Although the synoptic-scale upper-level flow patterns were similar between case types, analysis of representative inflow soundings for 20 crossing and 20 noncrossing cases (both upstream and downstream of the Appalachians) revealed that their downstream environments were significantly different (LP10). This finding stands somewhat in contrast to previous research (e.g., Reeves and Lin 2007) that primarily studied upstream effects (i.e., blocking of the cold pool) of terrain on a convective system. LP10 found that crossing cases were characterized by downstream thermodynamic environments containing higher instability and lower convective inhibition, consistent with previous research on MCS maintenance (e.g., Coniglio et al. 2007; Cohen et al. 2007). This finding was also consistent with the conclusion by Keighton et al. (2007) that “crossers” are most common during the day and “noncrossers” are more common at night. LP10 recommended that forecasters assess the most-unstable parcel’s convective available potential energy (MUCAPE) in the lee because it is the most general requirement for storms to develop and be maintained (including elevated storms) and, when paired with the most-unstable parcel’s convective inhibition (MUCIN), it correctly diagnosed a majority of the cases in the LP10 dataset.
Here, the authors take it as a fairly elementary conclusion that at least modest CAPE and small convective inhibition (CIN) are necessary for convective maintenance. However, in contrast to other MCS maintenance studies (e.g., Coniglio et al. 2007; Cohen et al. 2007), LP10 also found that the crossing cases moved into downstream environments typified by weaker vertical wind shear and a weaker mean wind. LP10 attempted to interpret this result in terms of the conceptual model of FM06, who described how topography impacts preexisting organized convection. Their idealized simulations showed that as squall lines traverse a barrier, they undergo orographic enhancement, followed by suppression, and then lee reinvigoration due to changes in the cold pool induced by the topography. As the system ascends the barrier, it undergoes orographic enhancement due to the upslope component of flow near the gust front. Then, as the system descends the barrier, the cold pool thins as the outflow becomes supercritical.1 This thinning of the cold pool adds to the lee suppression induced by ambient downslope flow and the sinking branch of an orographic gravity wave. When the cold pool reaches the base of the mountain in the lee, the outflow transitions back to the subcritical regime, suddenly slowing and deepening; this abrupt change in the depth of the fluid is known as a hydraulic jump, and its formation leads to subsequent restrengthening of the convection due to a stronger lift along the gust front.
In the context of the FM06 conceptual model, LP10 hypothesized that a weaker mean wind and less bulk shear downstream would promote MCS maintenance in a few ways. First, a weaker mean wind would induce weaker downslope flow in the lee, resulting in less lee suppression during the key stage when restrengthening must occur. The benefit of weaker bulk shear was interpreted in light of the squall-line maintenance theory of Rotunno et al. [1988, hereafter referred to as the Rotunno–Klemp–Weisman (RKW) theory]: less shear was hypothesized to provide a better balance with systems’ weakened cold pools in the lee, resulting in stronger lifting. Finally, LP10 conjectured that stronger lee shear promotes greater entrainment into developing updrafts, thus suppressing convection in the lee. This effect was theorized to be more profound in the smaller-CAPE noncrosser environments.
Unfortunately, a wide spread in the observed wind profiles in LP10 made it difficult to ascertain the true predictive capability of the wind profile. Furthermore, the inherent correlation of the mean wind and bulk shear made it unclear which ingredient was the most dynamically relevant in the observations. Another unanswered question was whether or not the wind profile alone is sufficient to discriminate between crossing and noncrossing MCSs (or whether its predictive capability is dependent upon the thermodynamic environment). The primary goal of this study was to ascertain the role of the wind profile (if any) in maintaining organized convection that encounters terrain. This was accomplished through idealized modeling tests that varied the mean wind and low-level shear independently, both with and without corresponding modifications to the thermodynamic environment. Section 2 describes the model setup and experiments while section 3 discusses the results of each sensitivity test along with the relevant processes. Finally, section 4 concludes with a summary of the key findings and implications of this work.
2. Data and methods
a. Model design
This study employed version 1.14 of the Bryan Cloud Model (CM1; Bryan and Fritsch 2002), a three-dimensional nonhydrostatic numerical model, to simulate idealized squall lines. The horizontal grid spacing was 500 m and the vertical grid spacing was stretched from 150 m at the model surface to 500 m aloft. This grid spacing is a bit coarser than what Bryan et al. (2003) recommended for convection, but it is comparable to that used by FM06 in the vertical and finer in the horizontal. The grid spacing provided a reasonable depiction of system structure (e.g., Weisman et al. 1997), while also being affordable enough to permit a large number of experiments. The vertical grid reached 20 km in height with the upper 6 km serving as a Rayleigh damping layer. The periodic y dimension was 60 km in extent; the across-line x dimension was nominally 600 km, although this was increased for the faster-moving squall-line simulations. Line-end effects were eliminated by the periodic y dimension; the focus here was on the key convective-scale processes at work during the lee redevelopment stages of the simulated systems.
To preserve comparability between the present study and FM06, convection was initiated in the same manner, using a line thermal with a potential temperature perturbation θ ′ of +4 K on which random temperature perturbations of up to 0.1 K were imposed, ensuring the development of 3D motions. The horizontally homogeneous base-state environment was identical to FM06, utilizing an idealized thermodynamic profile (Weisman and Klemp 1982) and wind profile (used in many numerical studies of squall lines (e.g., Rotunno et al. 1988; Weisman et al. 1988; see Fig. 1). Variations applied to the base state are discussed in section 2b. Additional simulations were attempted using the composite observed soundings from LP10; however, in the present idealized horizontally homogeneous framework, long-lived mature convective systems were unable to be produced in such environments due to the hindering effects of dry air and CIN (e.g., Figs. 4 and 6 of LP10). Such issues are known to be common in idealized simulations of deep moist convection (e.g., Munoz and Wilhelmson 1993; McCaul and Cohen 2004; Kirkpatrick et al. 2007).
Homogeneous base-state (a) thermodynamic and u-wind (mountain perpendicular) profiles testing the sensitivity to (b) the mean wind, (c) the low-level shear, and (d) the low-level shear with a constant mean wind. The thermodynamic profile in (a) is adapted from Weisman and Klemp (1982), with a constant boundary layer mixing ratio of 14 g kg−1.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
Precipitation microphysics were parameterized with the scheme of Thompson et al. (2004). The Coriolis force, surface physics, and radiation were neglected in order to isolate the key convective-scale dynamics. Real-world MCSs are potentially impacted by such processes, particularly as they pertain to the development of diurnally driven slope flows. However, before undertaking more complicated simulations, it was desirable to first isolate and understand the effects of varying the wind profile.
b. Experimental design
The sensitivity tests chosen for this study were motivated by the intriguing finding within LP10 that the wind profile is a statistically significant discriminator of crossers from noncrossers. Is the wind profile so important that it can be the deciding factor in environments with identical thermodynamic profiles? Clearly, if CAPE is zero then no convection is possible, regardless of the wind profile. However, it is unclear to what degree a noncrossing wind profile (i.e., a wind profile with stronger shear and mean wind) would suppress an MCS given the benefit of ample instability. To pursue this question, for the preponderance of the simulations the thermodynamic sounding was held fixed (Fig. 1a) while the wind profile was systematically varied (Figs. 1b–d). These experiments can be split into two categories: changes to the mean wind (Fig. 1b) and changes to the low-level shear (Fig. 1c). Since variations in shear inherently result in changes to the mean wind, further tests were performed that kept the mean wind in the lowest 2.5 km constant while changing the shear in that same layer (Fig. 1d). While previous studies have examined the behavior of precipitation systems encountering topography using changes to the base-state flow (e.g., Reeves and Lin 2007; Miglietta and Rotunno 2009), the current study used the FM06 conceptual model as a starting point and examined the sensitivities of this model to changes in the wind profile.
LP10 speculated that the wind profile would have a greater impact on the crossing process when the thermodynamic environment is less favorable. To quantify this, two types of experiments were performed. The first type introduced low-level cooling in the lee over the lowest 2 km AGL in the manner of Parker (2008), decreasing the surface temperature by 6 or 12 K. The lee cooling was achieved in each case by the time the squall line reached the peak of the mountain. The second type of experiment gradually modified the moisture profile in the lee by nudging it to a drier reference profile over time. As in the cooling simulations, the moisture modification was complete by the time the squall line reached the mountain peak. One set of tests decreased the boundary layer mixing ratio from 14 to 11 g kg−1 while another set of tests relaxed the 0–3-km moisture profile toward the averaged value of all LP10’s observed cases. [Interested readers can look ahead to Fig. 12 in order to see skew T–logp diagrams from these cooling and drying experiments (the results are discussed in section 3d).]
c. Definitions
Keighton et al. (2007) defined a crossing or noncrossing MCS based upon the production of severe weather reports in the lee of the mountains. Unfortunately, it is hard to produce a defendable metric for surface wind gusts or hail size in the present idealized model framework. Thus, a simple, objective definition of a crossing MCS was applied in this study. In the results and discussion that follows, a simulated squall line will be called a crosser if its along-line averaged simulated surface reflectivity is greater than 40 dBZ and its along-line averaged maximum vertical velocity is greater than 5 m s−1 in the lee of the barrier. Both criteria must be met at the same time for the MCS to be deemed a crosser. Otherwise, the squall line will be considered a noncrossing MCS. While these definitions probably do not overlap with those of Keighton et al. (2007), they at least serve to distinguish weaker (and likely nonsevere) convection from stronger (and potentially severe) convection.
3. Results
a. Control
To assure consistency with the work of FM06 and set a baseline for the experiments, control simulations were performed both with and without topography. Hovmöller diagrams of selected fields in Fig. 2 illustrate the evolution of the squall line over time, and Fig. 3 shows horizontal plan views of surface simulated reflectivity. The contrasts between the two control runs clearly demonstrate the impact of terrain on a convective system. Upstream (west) of the mountain peak, the base-state wind’s interaction with the terrain induces small amounts of rising motion, slightly enhancing ascent at the lowest model level (Fig. 2f). Downstream (east) of the mountain peak, there is an obvious weakening in the convection (Figs. 2d and 3f) due to ambient sinking motion as the system and its cold pool move downslope (Figs. 2e,f). This suppression is followed by a sudden reintensification near the base of the mountain (x = +70 km in Figs. 2d and 3g,h), consistent with the conceptual model of FM06. The strength of the cold pool over time also reflects this weakening and reinvigoration pattern in the presence of terrain (evident in Fig. 2e). In the immediate lee, the cold pool is warmer when topography is present as a result of partial blocking and downslope warming; however, when convection reintensifies near the base of the mountain in the lee, the cold pool still acquires a strength comparable to that simulated with no topography (Figs. 2b,e).
Hovmöller diagrams comparing (left) along-line averaged surface simulated reflectivity, (middle) θ ′ (K), and (right) lowest model level vertical velocity w (m s−1) for the (from top to bottom) no-terrain, control, increased mean wind +5 m s−1, and decreased mean wind −5 m s−1 simulations. The x axis in each represents the distance from the mountain peak (at x = 0) in km and the y axis displays simulation time in h. (f) The arrow points to a feature discussed in the text.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
Evolution of the surface simulated reflectivity for the control wind profile (a)–(d) without and (e)–(h) with terrain over time, as noted. The thick black contour represents −2-K θ ′ and denotes the position of the outflow boundary.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
As in FM06, this sudden reintensification of the convection in the lee is attributed to a hydraulic jump of the cold pool. To illustrate the evolution of the cold pool as it traverses the barrier, along-line averaged cross sections were computed relative to the objectively determined surface gust front location at every y point. This provided a clearer picture of the leading edge of the cold pool while correcting along-line phasing differences due to bowing (e.g., Figs. 3e,f). These cross sections illustrate the FM06 mechanism at work: vertical velocities become weaker and shallower as the cold pool traverses the mountain (Figs. 4a–c) and subsequently reintensify and deepen as the outflow pools in a hydraulic jump at the base of the mountain (Figs. 4d,e). While the outflow does not thin during descent as dramatically as in FM06, surface wind speeds near the head of the cold pool strengthen during this period and weaken after the outflow deepens, consistent with a transition from supercritical to subcritical flow (not shown). The less drastic thinning of the cold pool than in FM06 is mainly tied to the use of a broader barrier width, which results in more gradual system variations.
Along-line averaged vertical cross sections of vertical velocity w (m s−1; shaded) and θ ′ (K; contoured) over time (as noted), computed relative to the objectively determined surface gust front location at every y point for the control simulation with terrain. The x axis indicates the distance from the leading edge of the cold pool in km.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
In addition to this fundamental FM06 sequence, further deepening and cooling of the outflow occurs farther downstream after the cold pool’s hydraulic jump occurs (Fig. 4f). This further strengthening of the outflow is tied to the incorporation of new cells that develop approximately 5–10 km ahead of the outflow after it reaches the base of the mountain (Fig. 3g). While the development of these new cells is at least partly related to lifting along the outflow boundary, a dry simulation shows that the environmental flow over the barrier is associated with midlevel vertical motion near the location of these new cells (Fig. 5a). Broad, upwind-tilted orographic gravity waves are the initial response of the flow to the imposed terrain (Fig. 5b). However, as the prevailing flow continues to traverse the barrier and descend in the lee, an environmental hydraulic jump develops (Fig. 5c). Such a feature has been recognized to generate convective cells in the lee of terrain in previous studies of orographically generated convection (e.g., Chu and Lin 2000; Chen and Lin 2005; Miglietta and Rotunno 2009). Immediately downstream of the environmental hydraulic jump there exists a train of high-amplitude, short-wavelength gravity waves. Such a wave train has also been demonstrated in previous studies (e.g., Fig. 3c of Durran 1986). In the control run, the environmental hydraulic jump is visible as a thin line of upward velocities (see arrow at x = +50 km in Fig. 2f). Once the zone of lifting associated with the system’s outflow reaches this pocket of ascent, several additional convective cells develop and quickly merge with the parent MCS.
Illustration of the development of the hydraulic jump in the environmental flow leading to a wave train using (a) a Hovmöller diagram of surface u wind (m s−1; shaded) and 3-km vertical velocity w (m s−1; contoured) and (b),(c) vertical cross sections at t = 120 and t = 250 min, respectively, of vertical velocity w (m s−1; shaded) and θ ′ (K; contoured).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
The experiments performed by FM06 used rain-only microphysics, whereas the present simulations include ice. To determine whether the current simulations with ice fall into the same parameter space as FM06, an additional control simulation with an identical model configuration but with rain-only microphysics was performed (Kessler 1969). A comparison of the two control simulations in Fig. 6 illustrates a similar overall evolution; however, it is clear that the inclusion of ice microphysics allowed for a larger stratiform region (Figs. 6a,b) and lower θe in the surface outflow (Figs. 6c,d). The rear inflow is correspondingly stronger with ice, and it descends more substantially. Consequently, the amount of midlevel θe air brought down to the surface is also enhanced (Figs. 6e,f). Despite these differences, the upstream Froude numbers (i.e., the degree of blocking) within in the two cool pools were similar (not shown). The control simulation with ice microphysics contained a cooler and drier cold pool with higher stability, yet the descending rear inflow compensated for this by adding kinetic energy to the cold pool, producing a comparable outflow Froude number to the rain-only simulation. Thus, the current simulations with ice are expected to fall in the same part of the parameter space as those of FM06. The inclusion of ice microphysics, however, leads to more realistic system structures in the present study.
Comparison of control simulations with (left) ice microphysics and (right) rain-only microphysics, using (a),(b) Hovmöller diagrams of along-line averaged surface simulated reflectivity; (c),(d) plan views of θe (K, shaded) and simulated reflectivity (contoured every 10 dBZ); and (e),(f) vertical cross sections of along-line averaged θe (shaded) and u wind (m s−1).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
In summary, the control simulation largely upheld the FM06 conceptual model of enhancement, suppression, and reinvigoration. During lee reinvigoration, lifting following the hydraulic jump of the cold pool was locally supplemented by ascent s a result of a hydraulic jump of the large-scale flow. The tandem acted to initiate additional preline cells in the lee. Once the original system merged with this new convection, its updrafts became significantly stronger and its cold pool depth and temperature approached those in the no-terrain control run. As a result, this simulation is considered a crossing MCS and it serves as a baseline for the rest of the experiments.
b. Sensitivity to the mean wind
It was evident in the control simulation that slope flows can have both a beneficial and detrimental impact on organized convection encountering terrain. Since the observed crossing cases in LP10 were associated with a weaker mean wind than the noncrossing cases, the first set of sensitivity experiments tested the hypothesis that a stronger mean wind is less favorable for a crossing MCS due to stronger lee downslope flow, while a weaker mean wind is more favorable for a crossing MCS due to weaker lee downslope flow.
When the wind speed at every level of the wind profile was increased by 5 m s−1 (Fig. 1b), the overall evolution of the convective system still adhered to the pattern of orographic enhancement, suppression, and restrengthening (Fig. 2g). The stronger base-state wind speed–induced enhanced slope flows, resulting in more orographic enhancement and colder outflow on the upstream side of the barrier (Fig. 2h). The higher mean wind also produced stronger lee sinking motion (Fig. 2i), which acted to suppress the convection for a greater distance downstream from the mountain (i.e., strong convective redevelopment at x = +100 km downstream in Fig. 2g vs x = +70 km downstream in Fig. 2d) and weaken the updrafts in the lee (Fig. 7). However, the updrafts remained vigorous enough (and the simulated reflectivity remained high enough) that this system meets the definition of a crossing MCS (section 2c).
Hovmöller diagrams of the along-line average of the maximum vertical velocity w (m s−1) for the (a) control and (b) increased mean wind +5 m s−1 simulations. Smoothing has been applied to the field to account for insufficient time resolution in the output, allowing it to appear more continuous.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
In summary, a stronger mean wind did not prevent the MCS from successfully traversing the terrain, despite enhanced lee suppression of the system. Given the greater lee suppression, it is worth considering whether further increases to the mean wind would lead to dissipation of the MCS. An additional simulation that increased the mean wind by 10 m s−1 produced widespread pre-MCS convection over the terrain (due to intense low-level upslope flow), making it impossible to assess the actual MCS–terrain interaction. Based upon repeated experimentation, it appears that the primary effect of increasing the mean wind is simply to intensify the pattern of enhancement and suppression due to the stronger slope flows. At least in the favorable Weisman and Klemp (1982) thermodynamic sounding (with large CAPE and minimal CIN), the enhanced cross-barrier flow is not sufficient to suppress convective redevelopment in the lee of the terrain.
Decreasing the mean wind by 5 m s−1 (without changing the bulk shear) provided a different environment than the previous experiments due to a low-level flow reversal (easterly wind at the surface changing to westerly aloft; see Fig. 1b). This led to the development of a new convective system well ahead of the original squall line that was traversing the mountain; this new convection eventually replaced the original system (Fig. 2j). Other studies have found similar results in low wind speed experiments (e.g., Chu and Lin 2000; Chen and Lin 2005; Reeves and Lin 2007; Miglietta and Rotunno 2009). Anecdotally, the authors have observed similar behavior on a number of days near the Appalachian Mountains, so it is worthwhile to understand the origins of the new system. Two additional simulations were run to determine the source of the new MCS: one without moisture and one with moisture but without the initial line thermal used to generate the primary squall line. Examination of the dry simulation did not reveal any environmental mountain waves that could have contributed to the new convection on the east side of the barrier. However, the model run with moisture but without the primary (western) squall line still produced a new system to the east of the barrier in a similar location to the one in the original “mean wind −5 m s−1” simulation (not shown). Inspection of this simulation revealed that the convective system was a simple result of moist upslope flow impinging on the topography. Clouds were generated by the gradual ascent, and convection eventually developed as small-scale perturbations within the model lifted some parcels to their levels of free convection.
While the orographically generated squall line makes it difficult to interpret the evolution of the original MCS, it appears that the first MCS experienced a similar pattern of orographic interruption and enhancement as the other simulations (cf. Figs. 2d,g,j). West of the barrier, convection was suppressed before it reached the barrier peak, in large part due to the presence of ambient sinking motion resulting from the low-level easterly flow encountering the barrier (Fig. 2l). Despite this suppression, convective reinvigoration still occurred, notably closer to the peak (x = +30 km in Fig. 2j) as a result of the ambient rising motion to the east of the barrier. Due to its weakening west of the peak, the original MCS exhibited a warmer cold pool (Fig. 2k). The cold pool still remained strong enough to produce some lee reinvigoration and surface-based convection despite the presence of the second MCS. It should be noted that this simulation may not be directly comparable to the other experiments, as the formation of the second MCS likely altered the downstream environment. Furthermore, the second MCS that develops also makes it difficult to parse out the purported benefit of a weaker environmental mean wind; however, the original system was indeed able to overcome environmental subsidence west of the mountain peak, and was suppressed for a shorter distance and duration before being enhanced by ascent to the east of the barrier.
c. Sensitivity to the low-level shear
In addition to a weaker mean wind, LP10 found that crossing MCSs were associated with weaker vertical wind shear. Taking RKW theory into consideration, it was hypothesized that because the cold pool weakens and becomes partially blocked as it crosses the barrier, less environmental shear is required to balance the cold pool in order to create a vigorous, upright gust front updraft. Several simulations that varied the low-level shear tested this hypothesis.
The first test increased the bulk shear in the lowest 2.5 km by 5 m s−1 (Fig. 1c). As determined by the averaged reflectivity (Fig. 8a) and the averaged maximum vertical velocity (not shown), this increase in shear did not prevent a crossing MCS. A few differences between this simulation and the control were clear. First, the increased shear resulted in a warmer cold pool upstream (cf. Figs. 2e and 8b), which seemed to occur because the stronger shear initially produced an overturning updraft and a small amount of leading stratiform precipitation (e.g., as in Parker and Johnson 2004). Such a configuration entails less evaporation of precipitation into the trailing cold pool. The second difference was that the system restrengthened slightly farther downstream of the mountain (x = +100 km in Fig. 8a). Finally, the squall line contained more bowing segments as a result of the greater shear (not shown), which is consistent with the sensitivities reported by Weisman (1993). Despite these low-level differences, the terrain-induced evolution was similar to the other simulations; increasing the shear was not particularly detrimental to the maintenance of the squall line.
As in Fig. 2, but for the simulations varying the low-level shear.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
A comparison of the simulations in Fig. 9 demonstrates the influence of terrain on C, and consequently on C/ΔU. Without terrain, C and C/ΔU change very little over time in the control simulation and C/ΔU remains >1 (gray line with circles in Fig. 9). However, in the presence of terrain, C and C/ΔU are modified as the convective system encounters the barrier (black line with squares in Fig. 9). As discussed previously, the cold pool undergoes orographic enhancement, suppression, and reintensification, and these changes are directly reflected by C. In terms of RKW theory, in the control run with terrain, C/ΔU was still almost always >1, except at x = +50 km when C decreased in the lee and C/ΔU became ≈1 (Fig. 9b). In other words, according to RKW theory, the cold pool lifting in the control run was nearly optimal during the key stage of lee system redevelopment. Increasing the bulk shear by 5 m s−1 caused C to exhibit slightly larger fluctuations than in the control simulation, with C/ΔU falling a bit below 1 in the immediate lee (gray line with triangles, Fig. 9).
(a) Theoretical density current speed C and (b) C/ΔU computed at several locations upstream and downstream of the mountain peak (x = 0 km) for the simulations with varying low-level shear.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
Additional simulations tested further increases to the low-level shear in order to determine whether a noncrossing MCS would result. As the bulk shear in the lowest 2.5 km was increased by 10 and 15 m s−1 (over the control; Fig. 1c), C/ΔU correspondingly decreased downstream in the zone where system reformation occurs (x = 0 to x = +100 km; black line with circles and gray line with squares in Fig. 9b). However, despite the greater imbalance between the cold pool and shear, system redevelopment still occurred (Figs. 8d,g), with corresponding subsequent rebounds in C (Fig. 9a). Thus, within the present experiments’ parameter space, even if the shear is quite strong relative to the cold pool, it does not prevent system redevelopment and MCS maintenance across terrain.
A final sensitivity test decreased the bulk shear by 5 m s−1 in the lowest 2.5 km (Fig. 1c). Although C/ΔU was always >1 (black line with triangles in Fig. 9), the simulation nevertheless produced a crosser (Fig. 8j). This illustrates that a squall line can traverse the barrier in a lower-shear environment that is similar to what LP10 observed (although LP10 did not have a measure of C). The weaker shear caused the midlevel updraft to be displaced 10–20 km behind the gust front updraft throughout the simulation (e.g., midlevel updraft near x = −15 km and gust front updraft near x = 0 km in Fig. 10a). Even so, the lee reinvigoration process was sufficient to overcome the minimal CIN, renewing convection in the lee and producing a crossing MCS much like the control simulation.
As in Fig. 4, but for the simulation with low-level shear decreased by −5 m s−1.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
To recap, within a favorable thermodynamic environment, the cold pool–shear relationship does not appear to be a deciding factor for whether or not a simulated MCS will be a crosser. In the observations of LP10, shear and mean wind were highly correlated. Based upon the experiments described above, it appears that the mean wind has a greater dynamical impact upon the terrain-induced enhancement and suppression, while the shear has a secondary effect on the structure and intensity of the resulting convection. To fully isolate the effects of low-level shear from those of the mean wind, two additional sensitivity tests were performed that kept the mean wind constant in the lowest 2.5 km while either increasing or decreasing the bulk shear by 5 m s−1 (Fig. 1d). Under this configuration, the period of suppression in the immediate lee (x = 0 to x = +75 km) was much more pronounced in the case with weaker shear (Figs. 11a,b). The reason for this more notable suppression appears to be tied to the enhancement of the system’s cold pool upstream of the barrier, with increased stability of the cold pool entailing greater blocking by the terrain.
Hovmöller diagrams comparing (a),(b) along-line averaged surface simulated reflectivity and (c),(d) θ ′ (K) for the simulations with the same mean wind but varying low-level shear. The x axis in each represents the distance from the mountain peak (at x = 0) in km and the y axis displays simulated time in h.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
In the simulations presented herein, the mean wind contributes to the overall enhancement and suppression of the convection due to large-scale rising and sinking motion. Meanwhile, the shear influences the temperature and depth of the cold pool, and to some degree the ratio of C/ΔU in the lee. Despite these impacts, none of variations in the wind profile tested herein alone resulted in a noncrossing MCS. On the one hand, these experiments might overstate the role of the wind profile because it was modified over the entire domain (whereas the differences in LP10 were only significant in the lee). On the other hand, these simulations might understate the role of the wind profile because they used a very favorable thermodynamic environment.
d. Variations in the thermodynamic environment
The previous simulations clearly demonstrated that in a low CIN, high CAPE environment, modifications to the wind profile alone do not produce a noncrossing system, but do exert some influence on the evolution of an MCS traversing terrain. What is the influence of the wind profile in a lower instability, higher inhibition environment, such as what LP10 observed in their noncrossing cases? To answer this question, additional simulations were run in which the low levels were cooled or dried in the lee to achieve lower CAPE and higher CIN.
1) Lee low-level cooling
The first test cooled the surface in the lee of the terrain by 6 K, with maximum cooling achieved by the time the squall line reached the barrier peak, resulting in a surface-based CAPE (SBCAPE) of approximately 825 (2290) J kg−1 and a surface-based CIN (SBCIN) of −150 (0) J kg−1 in the lee (Fig. 12a). These were substantially lower SBCAPE and SBCIN values than in the control run (1790 and −20 J kg−1, respectively) whereas the MUCAPE and MUCIN values (based on a parcel located 1200 m above the surface) were unchanged. This cooling setup, while idealized, mimics the phenomenon of warm season cold air damming in the lee of the Appalachians (Bell and Bosart 1988). Despite the marked decrease in SBCAPE and increase in SBCIN, the MCS successfully crossed the terrain (Fig. 13a). While their severe weather potential is more limited, crossing cases have occurred when cold air damming is in place (Keighton et al. 2007). In the present simulations, the recurrence of strong lifting in the lee (Fig. 13e), in addition to available instability within and above the stable layer, maintained the squall line despite the enhanced surface-based inhibition.
Skew T–logp diagrams of the modified lee thermodynamic environments. Modifications include (a) cooling the surface 6 K (dashed) or 12 K (black), and (b) drying the low-levels to a constant 11 g kg−1 (black) or the average observed mixing ratio of all cases in LP10 (dashed).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
Comparison of the simulations cooling the surface in the lee by 6 and 12 K using Hovmöller diagrams of (a),(b) along-line averaged surface simulated reflectivity and (c),(d) θ ′ (K); and (e),(f) instantaneous along-line averaged cross sections of potential temperature (alternately shaded every 3 K from dark to light, starting at 291 K) and vertical velocity (m s−1, contoured).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
Since cooling the lee by 6 K still resulted in a small amount of SBCAPE, an additional test was performed that cooled the lee by 12 K. The resultant SBCAPE (MUCAPE) was 0 (1370) J kg−1 and SBCIN (MUCIN) was 0 (−5) J kg−1 (Fig. 12a). While the instability for both surface-based and elevated parcels was greatly reduced compared to the control, a crossing MCS (albeit somewhat less intense) was still produced (Fig. 13b). A key difference, however, was the development of a propagating internal bore on top of the stable layer as the system became elevated (leading edge near x = +220 km in Fig. 13f). This bore became the primary lifting mechanism that helped maintain the convection downstream of the terrain. This transition from a surface-based to elevated squall line appears to follow the same processes as described by Parker (2008).
Interestingly, the simulations with cooling actually had a narrower downstream zone of suppressed convection in the immediate lee of the barrier (x = 0 to x = +70 km; cf. Figs. 13a,b and 2d). In the control simulation, the system is cold pool driven, and lifting is suppressed until the cold pool’s hydraulic jump occurs. In contrast, cold pool-relative cross sections in Fig. 14 demonstrate that in the presence of cooling, the system becomes bore driven before the hydraulic jump in the cold pool occurs. As a result, strong updrafts are reestablished sooner (i.e., not as far downstream; Figs. 14a,b). As the amount of cooling increases, this transition to bore-driven lifting occurs more quickly. The vertical velocities weaken as the system moves farther downstream as a result of the lowered CAPE and more gradual slope in the isentropes (Fig. 14c). Thus, the shorter window of suppression does not equate to a stronger overall system.
Along-line averaged cross sections computed relative to the objectively determined surface gust front location (as in Fig. 4) comparing the (a) control, (b) surface cooling in the lee by 6 K, and (c) surface cooling in the lee by 12-K simulations. Vertical velocity w (m s−1) and CAPE are contoured (J kg−1; thick and thin contours, respectively), while θ ′ (K) is shaded. Each cross section is representative of the time at which the cold pool reached 60 km downstream of the peak of the barrier. The x axis indicates the distance from the leading edge of the cold pool in km.
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
The main implication of these cooling simulations is that MCS maintenance is strongly linked to the presence of sufficient downstream instability. Even in a lee environment with no surface-based instability, a crossing MCS was produced because elevated parcels in the lee with minimal inhibition were lifted to their levels of free convection. While this result is fairly unsurprising given the findings of LP10, additional cooling simulations with increases to the mean wind produced nearly identical results (not shown). This suggests that the cooling approach fell short of identifying a combination of nonzero MUCAPE and larger MUCIN that could critically test whether the wind profile plays an important role in a sufficiently unfavorable environment. Additional tests were employed that decreased MUCAPE and increased MUCIN via drying, as discussed next.
2) Lee low-level drying
Noncrossing cases have also been observed to move into drier environments (LP10), thus idealized changes to the moisture profile downstream were employed to weaken the instability and increase inhibition. Only the lowest 2–3 km were modified because the observed differences between case types in LP10 were concentrated in the low levels. As in the cooling simulations, drying was complete by the time the squall line reached the mountain peak. Decreasing the boundary layer mixing ratio in the lee to 11 g kg−1 (Fig. 12b) did not reduce the MUCAPE and increase the MUCIN enough to prevent system redevelopment in the lee or result in a noncrossing MCS, even when the mean wind was increased (not shown). Utilizing the average observed moisture profile for all soundings in LP10 over the lowest 3 km (Fig. 12b) resulted in SBCAPE and MUCAPE values of 600 J kg−1 and SBCIN and MUCIN of −20 J kg−1. This downstream environment produced a convective system that underwent a period of weak reintensification (from x = +50 to x = +100 km in Figs. 15a,c). This system is still classified as a crosser because of the lee updrafts of at least 5 m s−1 associated with simulated reflectivity of at least 40 dBZ. Increasing the mean wind in this environment, however, further hindered the system’s lee redevelopment (Figs. 15b,d), such that it became a noncrossing MCS according to the definition in this study. In other words, changes to the wind profile did play some role in determining MCS maintenance when the thermodynamic environment was sufficiently inhospitable for convection.
Comparison of the simulations drying the lowest 3 km in the lee to the average observed moisture profile of all cases in LP10 with (left) the control and (right) increased mean wind +5 m s−1 wind profiles. (a),(b) The along-line averaged surface simulated reflectivity and (c),(d) the along-line average of the maximum vertical velocity w (m s−1).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
An interesting question is whether or not this environment would support maintenance of a squall line over flat terrain. That is, would this environment lead to MCS dissipation regardless of the terrain? Additional simulations were run in which the “lee” was dried to the average observed moisture profile in LP10, but without terrain. When the terrain was removed, simulated reflectivities and vertical velocities remained above the crossing threshold well into the lee drying zone (Figs. 16a,e). Without terrain, the cold pool was not interrupted (Fig. 16c), nor was the system subject to downslope suppression. Instead, a more gradual weakening of the system occurred (Fig. 16a). Thus, it is clear that the impact of both terrain (with a stronger mean wind) and the less favorable lee thermodynamic environment both played a role in producing the simulated noncrossing MCS. Numerous other treatments of the downstream thermodynamic environment (not reported here) confirm MUCAPE as the most important parameter for simulated MCS maintenance over terrain, particularly when combined with MUCIN. The influence of the wind profile is generally smaller, but is not negligible when instability is small and inhibition is large.
Hovmöller diagrams comparing the simulations drying the lowest 3 km in the lee to the average observed moisture profile of all cases in LP10 with an increased mean wind +5 m s−1 both (left) without and (right) with terrain. (a),(b) The along-line averaged surface simulated reflectivity; (c),(d) the along-line averaged surface θ ′ (K); and (e),(f) the along-line average of the maximum vertical velocity w (m s−1).
Citation: Monthly Weather Review 139, 10; 10.1175/2011MWR3635.1
4. Discussion and conclusions
Through the use of idealized sensitivity tests, the first-order impacts of the wind profile upon MCSs traversing terrain have been examined. While LP10 showed that weaker shear and a weaker mean wind were associated with crossing MCSs, the present work finds that the wind profile alone does not play a primary role in determining idealized squall-line maintenance over a barrier similar to the Appalachians. However, the wind profile does influence the evolution of the MCS. Variations in the mean wind impacted the degree to which the system was orographically enhanced upstream and suppressed downstream, while changes in low-level shear were shown to modify the strength of the outflow, and to influence the conditions for gust front lifting in the lee. Despite these effects, a lee hydraulic jump in the outflow renewed lifting at the head of the cold pool (much like FM06 described) in each wind profile experiment. Tests that decreased the downstream instability through low-level cooling or drying still produced crossing MCSs with the control wind profile. A combination of substantial lee drying and increased mean wind did finally result in a terrain-induced noncrosser (as defined in this study).
In the datasets of Keighton et al. (2007) and LP10, most of the observed noncrossing MCSs failed to survive even as weak (nonsevere) convection in the lee. Likewise, Parker and Ahijevych (2007) estimated that only 10%–30% of convective systems survive crossing the Blue Ridge Mountains. The robustness of the squall lines in the present study is probably due to the extremely favorable initial sounding that was used (i.e., that of Weisman and Klemp 1982). The use of this sounding helps connect the present results to those of FM06, and it was found that the idealized model utilized herein had difficulty producing convection at all with the drier, more stable mean soundings from the LP10 study. Additional experiments (not shown) incorporating the LP10 observed soundings only into the lee of the idealized simulations were also unsatisfactory. Future studies with more inhospitable soundings would be useful, although a less idealized modeling framework (e.g., case study simulations) appears to be needed. A few other potentially important factors were also excluded. For example, the diurnal cycle was not included in the simulations, which affects the development of orographic slope flows and low-level jets. Furthermore, the present idealized study did not include cross-barrier changes in bulk shear and mean wind, which LP10 found to discriminate between crossing and noncrossing squall lines. This study also did not address situations in which the cold pool is completely blocked by the terrain, although in lower mountain chains such as the Appalachians, complete blocking of a 1–2-km-deep cold pool is unlikely. Finally, it has been shown by LP10 that both crossing and noncrossing cases tend to be associated with varying kinds of synoptic-scale fronts, thus it may be useful to explore the influence of larger-scale processes on squall-line longevity. Such avenues would provide additional understanding and help extend these idealized results toward tangible improvements in operational forecasting of these systems.
Acknowledgments
The authors thank members of the Convective Storms Group at North Carolina State University for their assistance and feedback throughout the course of this study. Comments from two anonymous reviewers provided valuable suggestions that improved the quality of this manuscript. This research was supported by the National Science Foundation under Grant ATM-0552154 and by National Ocean and Atmospheric Administration’s Collaborative Science, Technology and Applied Research (CSTAR) Program Award NA07NWS4680002.
REFERENCES
Bell, G. D., and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev., 116, 137–161.
Bryan, G. H., and M. J. Fritsch, 2002: A benchmark simulation for moist nonhydrostatic numerical models. Mon. Wea. Rev., 130, 2917–2928.
Bryan, G. H., J. C. Wyngaard, and M. J. Fritsch, 2003: Resolution requirements for the simulation of deep moist convection. Mon. Wea. Rev., 131, 2394–2416.
Bryan, G. H., J. C. Knievel, and M. D. Parker, 2006: A multimodel assessment of RKW theory’s revelance to squall-line characteristics. Mon. Wea. Rev., 134, 2772–2792.
Chen, G. J. T., and H. C. Chou, 1993: General characteristics of squall lines observed in TAMEX. Mon. Wea. Rev., 121, 726–733.
Chen, S.-H., and Y.-L. Lin, 2005: Effects of moist Froude number and CAPE on a conditionally unstable flow over a mesoscale mountain ridge. J. Atmos. Sci., 62, 331–350.
Chu, C.-M., and Y.-L. Lin, 2000: Effects of orography on the generation and propagation of mesoscale convective systems in a two-dimensionally conditionally unstable flow. J. Atmos. Sci., 57, 3817–3837.
Cohen, A. E., M. C. Coniglio, S. F. Corfidi, and S. J. Corfidi, 2007: Discrimination of mesoscale convective system environments using sounding observations. Wea. Forecasting, 22, 1045–1062.
Coniglio, M. C., D. J. Stensrud, and L. J. Wicker, 2006: Effects of upper-level shear on the structure and maintenance of strong quasi-linear mesoscale convective systems. J. Atmos. Sci., 63, 1231–1252.
Coniglio, M. C., H. E. Brooks, S. J. Weiss, and S. F. Corfidi, 2007: Forecasting the maintenance of quasi-linear mesoscale convective systems. Wea. Forecasting, 22, 556–570.
Durran, D. R., 1986: Another look at downslope windstorms. Part I: The development of analogs to supercritical flow in an infinitely deep, continuously stratified fluid. J. Atmos. Sci., 43, 2527–2543.
Frame, J., and P. Markowski, 2006: The interaction of simulated squall lines with idealized mountain ridges. Mon. Wea. Rev., 134, 1919–1941.
Gale, J. J., W. A. Gallus Jr., and K. A. Jungbluth, 2002: Toward improved prediction of mesoscale convective system dissipation. Wea. Forecasting, 17, 856–872.
Keighton, S., J. Jackson, J. Guyer, and J. Peters, 2007: A preliminary analysis of severe quasi-linear mesoscale convective systems crossing the Appalachians. Preprints, 22nd Conf. on Weather Analysis and Forecasting, Park City, UT, Amer. Meteor. Soc., P2.18.
Kessler, E., 1969: On the Distribution and Continuity of Water Substance in Atmospheric Circulations. Meteor. Monogr., No. 32, Amer. Meteor. Soc., 84 pp.
Kirkpatrick, J. C., E. W. McCaul Jr., and C. Cohen, 2007: The motion of simulated convective storms as a function of basic environmental parameters. Mon. Wea. Rev., 135, 3033–3051.
Letkewicz, C. E., and M. D. Parker, 2010: Forecasting the maintenance of mesoscale convective systems crossing the Appalachian Mountains. Wea. Forecasting, 25, 1179–1195.
McCaul, E. W., Jr., and C. Cohen, 2004: The initiation, longevity and morphology of simulated convective storms as a function of free tropospheric relative humidity. Preprints, 22nd Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., 8A.5.
Miglietta, M. M., and R. Rotunno, 2009: Numerical simulations of conditionally unstable flows over a mountain ridge. J. Atmos. Sci., 66, 1865–1885.
Munoz, L. A., and R. B. Wilhelmson, 1993: Structure and evolution of thunderstorms encountering temperature inversions. Preprints, 17th Conf. Severe Local Storms, St. Louis, MO, Amer. Meteor. Soc., 247–251.
Parker, M. D., 2008: Response of simulated convective squall lines to low-level cooling. J. Atmos. Sci., 65, 1323–1341.
Parker, M. D., and R. H. Johnson, 2004: Simulated convective lines with leading precipitation. Part I: Governing dynamics. J. Atmos. Sci., 61, 1637–1655.
Parker, M. D., and D. A. Ahijevych, 2007: Convective episodes in the east-central United States. Mon. Wea. Rev., 135, 3707–3727.
Reeves, H. D., and Y.-L. Lin, 2007: The effects of a mountain on the propagation of a preexisting convective system for blocked and unblocked flow regimes. J. Atmos. Sci., 64, 2401–2421.
Rotunno, R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45, 463–485.
Thompson, G., R. M. Rasmussen, and K. Manning, 2004: Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part I: Description and sensitivity analysis. Mon. Wea. Rev., 132, 519–542.
Weisman, M. L., 1993: The genesis of severe, long-lived bow echoes. J. Atmos. Sci., 50, 645–670.
Weisman, M. L., and J. B. Klemp, 1982: The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Mon. Wea. Rev., 110, 504–520.
Weisman, M. L., and R. Rotunno, 2004: “A theory for strong long-lived squall lines” revisited. J. Atmos. Sci., 61, 361–382.
Weisman, M. L., J. B. Klemp, and R. Rotunno, 1988: Structure and evolution of numerically simulated squall lines. J. Atmos. Sci., 45, 1990–2013.
Weisman, M. L., W. C. Skamarock, and J. B. Klemp, 1997: The resolution dependence of explicitly modeled convective systems. Mon. Wea. Rev., 125, 527–548.
Strictly speaking, the “supercritical” or “subcritical” terminology refers to wave propagation flow regimes in shallow-water theory. In this paper, these terms will be used to describe stratified flow that strongly resembles the regimes in shallow-water theory.
