1. Introduction
The genesis of large-amplitude mesoscale gravity waves (MGWs) has been a topic of active research for many years. In the United States, such waves are most common east of the Rocky Mountains (Koppel et al. 2000) and are normally accompanied by cloud bands and precipitation (Koch and O'Handley 1997). Singular, large-amplitude waves are infrequent but may be relatively long lived, with some observed for over 12 h (Bosart and Sanders 1986; Ramamurthy et al. 1993). Identification of MGWs is normally based on analysis of correlations between surface pressure and wave-relative wind fields, and automated methods for wave detection have been developed (Bradshaw et al. 1999; Koch and Saleeby 2001). Despite the longevity of large-amplitude MGWs, their relative infrequency and the sparse upper-air observing network has meant that comprehensive three-dimensional datasets with high spatial and temporal continuity are rare. Specialized datasets from field experiments and modeling studies have been the primary source of information for our understanding of these waves.
Uccellini and Koch (1987, hereafter UK87) developed a climatology of MGW events over the midwestern United States. Wave genesis was found to occur north of surface warm fronts and near the inflection axis in the 300-mb height field under the left exit region of approaching upper-level jet streaks. Based on this flow configuration, geostrophic adjustment (Bosart and Seimon 1988; Ramamurthy et al. 1993; Kaplan et al. 1997) and shearing instability (Stobie et al. 1983; Pecnick and Young 1984; Einaudi et al. 1987) were hypothesized to be responsible for wave formation. Many case studies suggest both processes may have contributed to wave formation (e.g., Ferretti et al. 1988; Bosart et al. 1998). Additional mechanisms hypothesized to contribute to wave genesis include frontogenesis and frontal collapse (Ley and Peltier 1978; Bosart and Sanders 1986; Snyder et al. 1993), convection (Clark et al. 1986; Lin and Goff 1988; Powers and Reed 1993), orography and the mountain–plains solenoid (Zhang and Koch 2000; Koch et al. 2001), and initiation by the transverse circulation about a jet streak acting upon a low-level potential vorticity reservoir (Pokrandt et al. 1996).
Studies relating wave genesis to geostrophic adjustment have addressed the unbalanced (versus quasigeostrophic) nature of the flow downstream from the jet core (Van Tuyl and Young 1982; Koch and Dorian 1988). Koch and Dorian (1988) considered Lagrangian Rossby number values exceeding 0.5 as indicative of significantly unbalanced flow. Geostrophic adjustment associated with small (e.g., sub-Rossby radius of deformation) jetlets (Kaplan et al. 1997, 1998; Hamilton et al. 1998) has also been implicated in wave genesis in some cases, sometimes with convection present (Koch et al. 1998).
Mesoscale gravity waves are often accompanied by precipitation. In a 25-yr climatology of hourly surface observations, Koppel et al. (2000) identified 1038 inertia–gravity wave signatures, of which 46% were associated with convection. Convection has been found to generate or modulate gravity waves (Balachandran 1980; Miller and Sanders 1980), and gravity waves have also been shown to initiate convection (Koch 1979; Ferretti et al. 1988). Koch and Golus (1988) and Powers and Reed (1993) concluded that gravity waves and convective motions were interdependent in the cases examined and appeared coupled at times in a wave–CISK (conditional instability of the second kind) relationship.
Numerical modeling has been used to study wave genesis and the close relationship between mesoscale gravity waves and convection. Zhang and Fritsch (1988) found that gravity waves determined the structure, orientation, and propagation of their simulated squall line. Wave genesis was attributed to geostrophic adjustment associated with a low-level jet. In two-dimensional simulations with shear, Schmidt and Cotton (1990) found that gravity waves determined the squall-line speed and helped lift inflow air to saturation. In addition, the presence of an elevated mixed layer helped trap and sustain gravity waves. Cram et al. (1992) modeled a squall line in three dimensions and found line movement tied to an internal gravity wave, with discrete propagation by gust front processes also important. Convection was triggered by a cold front, and wave genesis was attributed to parameterized upper-level heating accompanying the convection, implying wave–CISK (Lindzen 1974; Raymond 1975; Davies 1979; Xu and Clark 1984) was active. Wave–CISK provides one framework for explaining the interaction between and maintenance of convection and gravity waves, though it does not address convective initiation (Koch et al. 1988) nor is it applicable to later stages of wave–convection interaction (Raymond 1976).
Powers and Reed (1993) revealed the importance of thermodynamical processes in MGW genesis through experiments with the Pennsylvania State University–National Center for Atmospheric Research (Penn State–NCAR) Mesoscale Model, version 4 (MM4). In their 30-km simulation, good agreement with observations was noted, including areas of wave formation and decay. Use of lower horizontal and vertical resolution resulted in fewer, weaker waves of larger wavelength. Significantly, they found that wave formation did not occur with latent heating omitted and concluded that convection, and possibly shearing instability, were responsible for wave genesis. The large amplitude (up to 7 mb) and longevity of the waves were attributed to ducting and wave–CISK forcing. In further modeling of this case, Powers (1997) determined that latent heating was more important than evaporative cooling and melting in modulating wave formation.
The aforementioned observational and numerical studies have revealed a strong relationship between precipitation and gravity waves. This was true for both warm-(Feretti et al. 1988) and cool-season (Bosart and Seimon 1988) events. While the continuing association between precipitation and mature gravity waves has been well documented, the origins of this relationship are often less clear. In addition, because the environment identified by UK87 is dynamically vigorous, most of the documented large-amplitude MGW events in the literature have multiple wave genesis mechanisms cited (e.g., Koch et al. 1993 discuss three). Even if only one mechanism were responsible, isolating it from the others would be difficult.
A combination of modeling and analysis of high-resolution observational datasets seems best suited to understanding wave genesis. Detailed three-dimensional observations are useful but always spatially or temporally incomplete, while modeling may aid in identifying the key physics provided the results can be verified with observations on comparable scales. Observational studies of an MGW event were recently published by Rauber et al. (2001, hereafter Part I) and Yang et al. (2001, hereafter Part II) for an MGW occurring during the Storm-scale Operational and Research Meteorology-Fronts Experiment Systems Test (STORM-FEST; Szoke et al. 1994).
On 14 February 1992, a long-lived MGW developed in association with a weak precipitation band over southwest Kansas. The wave was tracked for 14 h through Kansas, Missouri, and Illinois. Profiler, surface mesonet, 3-hourly rawinsonde data, and dual-Doppler measurements were used to provide a detailed 3D analysis of the wave, precipitation, and wind field as the wave crossed the STORM-FEST network.
The earliest surface pressure signatures of wave motion were observed near 1600 UTC as a dry air mass, originating as downslope flow in the lee of the Rocky Mountains, ascended a warm front east of the lee cyclone (Part I). A weak rainband developed simultaneously with the wave of depression at the leading edge of the dry air mass. The orientation of the wave front, determined from isochrone analyses of minimum pressure occurrence, and the orientation of the rainband, determined from radar analyses, corresponded closely to the leading edge of the dry air mass. The gravity wave and rainband remained tied to the leading edge of the dry air mass for the first 8–10 h of evolution. As the wave entered central Missouri, the convection and wave decoupled from the leading edge of the dry air mass. At about the same time, surface barograms suggested a change to a wave of elevation as the wave propagated into eastern Missouri and Illinois.
At 2100 UTC, the rainband, mature gravity wave, and leading edge of the dry air mass passed over the dual-Doppler network in northeast Kansas. Three-dimensional kinematic fields were synthesized from the dual-Doppler data and thermodynamic retrieval techniques were applied to derive perturbation pressure and virtual temperature analyses (Part II). A downdraft at the rear of the rainband was attributed to evaporative cooling of precipitation above the warm front and descent accompanying the deceleration of air at the leading edge of the dry air mass (see Part II, Fig. 15). This downdraft acted to depress the height of the low-level stable layer, creating the wave signature in the surface barograms. However, the environment present during wave genesis, 7 h earlier in southwest Kansas, was not well observed. Evaporation of precipitation with the newly formed rainband was hypothesized to have led to subsidence warming and depression of the inversion, creating the first indications of a solitary wave of depression, but no evidence was available to test this hypothesis.
Other researchers have examined the same event. Koch and O'Handley (1997) investigated wave genesis using MM4 with 20-km grid spacing. Gravity waves formed in a region of unbalanced flow aloft and a stable duct layer north of the surface warm front. MM4 exhibited some skill in reproducing the early wave structure, but was less accurate after precipitation developed along the wave. Koch and Siedlarz (1999, hereafter KS99) identified three incipient gravity waves across Kansas by 1830 UTC, the most significant wave located in the southwestern part of the state. They ruled out convection as a genesis mechanism for any of the incipient waves, but stated convection had a role in wave maintenance. Trexler and Koch (2000) studied profiler observations and, utilizing kinematic methods, derived time–height analyses of vertical velocity. They concluded that the gravity wave triggered the precipitation band ahead of the wave trough.
In this paper we test the hypothesis, based on observations presented in Parts I and II, that evaporative processes acting on precipitation from a weak rainband above the frontal inversion led to gravity wave genesis. The role of the dry air mass in the development of the rainband and gravity wave is also explored. Given data limitations and analysis uncertainties at the time of wave genesis, this numerical study will complement the observational work presented in Parts I and II by revealing the early development of the dry air mass, rainband, and gravity wave. The focus of this paper is solely on wave genesis and evolution prior to the mature stage revealed in Parts I and II. Issues pertaining to wave maintenance, therefore, are beyond the scope of the present study.
We first describe our numerical approach, followed by the synoptic environment, and model representation of the wave. The structure of the surge of dry air and its role in rainband formation and wave genesis is then explored, and a conceptual model is presented. A comparison of the current study to related research is made, followed by a summary of our findings.
2. Methodology
Numerical simulations of the 14 February 1992 MGW were carried out with the nonhydrostatic fifth-generation Penn State–NCAR Mesoscale Model (MM5; Grell et al. 1995). The model initialization was obtained through analysis of upper-air and surface data from 0000 UTC 14 February 1992, 16 h prior to observed gravity wave genesis. This early time was chosen to allow sufficient duration for model spinup and to ensure that the MGW formation was not spurious or a consequence of initialization. National Centers for Environmental Prediction (NCEP) Global Data Assimilation System (GDAS) analyses were used as first-guess fields on which the STORM-FEST observations were analyzed. After the resulting analyses were interpolated to the model's vertical (sigma) coordinate, removal of integrated divergence was performed to reduce the initial condition noise.
Three nested grids were employed, with 54-, 18-, and 6-km horizontal grid spacing. The 6-km innermost grid spacing was considered adequate to resolve the gravity wave structure given the observed wavelength of 35–60 km from Doppler radar and Portable Automated Mesonet (PAM) observations presented in Part I. The outer two grids were initialized at 0000 UTC, while the innermost 6-km nest was initiated at 0600 UTC and moved eastward (at 1200 and 1800 UTC) to follow the lee cyclone and developing gravity wave. The three-grid configuration appears in Fig. 1. The lateral boundaries of the outer grid of the model were obtained by blending model-predicted values with those computed from a linear interpolation between objective analyses (Grell et al. 1995).
Since the emphasis in this paper is on gravity wave genesis, model simulations were carried out for 24 h, ending 0000 UTC 15 February. This included the period of observed wave genesis and of radar measurements of mature MGW structure. After 0000 UTC, the observed rainband and wave underwent significant morphological changes while moving from a near-neutral to an unstable environment after propagating out of the STORM-FEST dual-Doppler network. The evolution of the wave in this latter stage of development was poorly documented. We did not attempt simulation after 0000 UTC.
The MM5 configuration utilized in the control run of our study is summarized in Table 1. As in Powers (1997), explicit grid-scale microphysics followed the simple ice parameterization of Dudhia (1989). The Grell (1993) scheme, which includes the effect of moist downdrafts, was used for cumulus parameterization on all grids. The Blackadar (1979) parameterization was employed for the model's planetary boundary layer.
Traditional model fields were stored every 15 min during the simulation. In addition, we chose to save surface wind, pressure, temperature, moisture, low-level cloud, and rainwater mixing ratio, and four levels of vertical velocity at each 90-s model time step. These high temporal resolution surface fields were used to construct surface time series traces, which will be discussed later.
The four model simulations discussed in this paper are summarized in Table 2. The control experiment made use of full model physics, for the entire simulation. Experiment EXPLICIT omitted the Grell cumulus parameterization used in CONTROL on the innermost 6-km grid, allowing only explicit precipitation. In experiment NOEVAP, evaporative cooling was omitted after 1200 UTC. (Evaporation of cloud and rainwater continued, but no evaporative cooling took place. Only explicit microphysics were altered; the Grell scheme and its convective downdrafts were unaffected). Earlier simulations of other gravity wave events had shown significant alteration to cyclone intensity and structure when latent heating was omitted for the entire simulation (see Powers and Reed 1993, their Figs. 24b and 26a), thereby altering the wave genesis environment. We therefore chose to impulsively remove the impact of evaporation after 12 h, allowing enough time to establish most of the cyclone evolution deemed important in gravity wave development. Note that while evaporative cooling was omitted, latent heating was retained. In the NOEVAP experiment, evaporative cooling was omitted for the remainder of the simulation. Experiment RESTORE is similar to the NOEVAP experiment, except evaporative cooling was restored after 1630 UTC. Experiment RESTORE was used to test the hypothesis that the environment for gravity wave genesis was largely unaltered by the omission of evaporative cooling. If true, restoration of evaporative cooling (omitted in NOEVAP) would allow wave genesis to occur.
3. Synoptic environment and gravity wave genesis and evolution
The 14 February 1992 gravity wave event accompanied a lee cyclone, which moved from Colorado into Kansas during the day. At 1200 UTC (Fig. 2), the surface low pressure center was located in eastern Colorado in the simulation, slightly north of the observed position (Part I, Fig. 2). A warm front extended east from the cyclone center and a dryline was present to the south. A cool and moist air mass was present north of the warm front. Behind the dryline, strong westerly winds and dry air extended back to the Rocky Mountains of New Mexico and southern Colorado. In the model simulation and the observations, the cyclone tracked east-southeastward and then east along the Oklahoma–Kansas border during the day. During this time, the warm front lifted slowly northward, while the dryline swept eastward.
At 1200 UTC, a jet streak was present over Mexico and southwestern Texas. A surface cyclone developed under the left exit region of this jet streak (Fig. 2a). A well-developed comma-cloud structure was evident in the model relative humidity (RH) fields and infrared satellite imagery. A large-scale dry slot was located over Kansas, as indicated by the dry region surrounded by the area of 80% RH at 500 mb (Fig. 2b). A shallow moist air mass was located beneath this upper-level dry slot. In Fig. 2b, the hatched region indicates saturation at 900 mb, the top of the warm frontal inversion in central Kansas (Part I). The mesoscale gravity wave and associated rainband formed in both the model and observations within the larger dry slot. The close agreement between the observed and model-simulated positions of the surface lows and frontal boundaries and the overall flow pattern gives us confidence that we may use the MM5 fields to probe the role of evaporative processes in the genesis of the wave.
As the cyclone propagated eastward, the surface dryline swept east and northward. As in the observations, the dry air and strong winds behind the dryline, which originated as mountain downslope flow, ascended the warm front southeast of the surface cyclone center (Part I). Unlike the case reported by Karyampudi et al. (1995a,b), the dry air descending the Rocky Mountains was less dense than the cool, moist air mass it encountered north of the warm front. As a result, the dry air ascended the warm front.
The gravity wave first appeared in the MM5 surface pressure and wind fields near 1800 UTC in association with the development of a precipitation band aloft. Wave genesis took place north of the surface warm front under the left exit region of the jet, as in the UK87 climatology. The gravity wave structure on the innermost 6-km grid is documented in Fig. 3. At 2000 UTC, point E in Fig. 3a was along a wave of depression, which appeared as an inverted trough of lower surface pressure northeast of the lee cyclone center. The points C, D, E, and F were chosen near the wave axis at 1815, 1900, 2000, and 2100 UTC. MM5 surface data, at 90-s resolution, were used to construct time series profiles at these locations. The time series of wave-relative wind U′ and perturbation pressure P′ were detrended and plotted in Figs. 3c–f. While the simulated wave magnitude was smaller than observed (Part I, Fig. 4), the surface manifestation of the wave is clearly evident in the highly correlated wind and pressure perturbations found in the 6-km model data.
A northeast–southwest cross section through the wave appears in Fig. 3b. The bold dashed line in the figure identifies the pressure trough. The pressure perturbations associated with the wave are most prominent in the lowest 100 mb. The half-wavelength (λ/2 in the figure) is approximately 16 km, yielding a simulated wavelength of 33 km. This is in good agreement with the 35-km wavelength measured at Topeka, Kansas. In the observations, determined from radar data and individual pressure traces, the wavelength varied across the domain from 35 km at Topeka and 48 km at Hillsboro, Kansas, to 75 km at Columbia, Missouri, after 0000 UTC (Part I). Model-simulated wavelengths ranged from 36 to 55 km over Kansas. The simulation was halted at 0000 UTC 15 February, before the wave entered Missouri.
The pressure fall depicted in Fig. 3b was beneath a region of depressed isentropes, indicating warming below 800 mb. Elevated isentropes aloft identify cooling near 750 mb above the same region. This pattern of potential temperature perturbations and pressure falls had considerable temporal continuity (Fig. 4) as the wave propagated northeastward. The warm perturbation in Fig. 3b can be seen in Fig. 4 at 850 mb, tracking northeastward with the wave. After 2100 UTC, the isentrope depression diminished while surface pressures rose, signifying a transition to a wave of elevation, a transition that was also noted in the observational studies (e.g., Part I).
After 1800 UTC, the simulated gravity wave propagated through eastern Kansas. The wave's propagation is evident in the surface wind and pressure correlation data between 1800 and 2300 UTC. The area of correlation exceeding 0.9 for three overlapping 3-h time periods (Fig. 5a) progressed northeastward and maintained its coherence through 2300 UTC. When computed for a 3-h period centered on the detrended pressure minimum at each 6-km grid point, the result was the large area of high correlation seen in Fig. 5b.
The wave isochrones also appear in Fig. 5. Surface evidence of genesis first occurred near 1800 UTC in the simulation, as a wave of depression developed. The modeled wave propagated east-northeast at 12.5 m s−1 between 1900 and 2100 UTC, slower than the observed speed of nearly 20 m s−1. By 2100 UTC, precipitation began reaching the surface and pressure rises (bold lines in Fig. 5) became more prominent, consistent with the changes in the cross section in Fig. 4.
The genesis and evolution of the gravity wave in the model simulation was closely coupled with the formation and evolution of a precipitation band. At 2100 UTC (Fig. 6b), the isochrone marking the wave of depression (P−) was found immediately behind a region of precipitation, as observed in STORM-FEST (wave B- and rainband RB1 in KS99, their Fig. 6). While the modeled gravity wave trough was nearly 200 km long, the pressure falls and pressure gradient associated with the wave were strongest in the northern half of the trough axis. Surface pressure ridges were located within the model-predicted precipitation region (and in STORM-FEST observations by KS99), suggesting a hydrostatic response to evaporation of precipitation. The perturbation pressure field lagged the low-level vertical velocity field (shaded in Fig. 6b) by a quarter wavelength as expected (Eom 1975), with maximum descent located between the pressure maximum and minimum. This near-surface downdraft was coincident with the low-level rainwater peak in the surface time series (Fig. 6c). Weak low-level ascent occurred prior to the arrival of precipitation, and in the convergence region behind the trailing pressure minimum.
The vertical velocity field (Fig. 6a) was characterized by subsidence between 700 and 850 mb, to the rear of the precipitation region. This downdraft created the isentrope depression in Fig. 6a. The depressed isentropes are indicative of subsidence warming, consistent with the 850-mb warm anomaly in Fig. 6d. The 850-mb warming and isentrope depression are collocated with the 2100 UTC surface pressure minimum, and track with this pressure minimum over time (Fig. 4).
The cause of this subsidence was evaporation of precipitation. Near the trailing edge of the precipitation band, evaporation resulted in cooling aloft (e.g., the elevated 297.5-K isentrope in Fig. 3b), subsidence, and (provided the evaporation was total) adiabatic warming at lower levels. As the evaporative downdraft impinged on the warm frontal surface, the inversion depth was reduced in the model and hydrostatic surface pressure falls led to the development of a wave of depression. As the wave intensified, highly correlated surface wind and pressure perturbations formed (Fig. 5). Thus, the model fields suggest that evaporation aloft was responsible for wave genesis.
4. Dry airmass evolution and accompanying wave generation
The results of the previous section suggest a close relationship between wave genesis and evolution and evaporative processes. In this section, we address the origin of the precipitation band seen in Fig. 6 and the timing of the formation of the band and gravity wave within the context of the parent cyclone.
In both the observations and the model simulation, the rainband developed at the leading edge of a surge of dry air originating as downslope flow over New Mexico and southern Colorado. By 1400 UTC in the simulation, the surge of dry air had moved out over the Texas Panhandle (Fig. 7). This low- to midtropospheric dry air was distinct from and well behind the leading edge of the larger-scale dry slot seen in Figs. 2 and 7. The leading edge of the dry air mass moved through Oklahoma and southern Kansas, with the northern extent ascending the warm front east of the cyclone center. As the dry air mass ascended the warm front, the rainband formed in the simulation and moved into southwestern Kansas, remaining at the northern edge of the dry air mass. The rainband intensified, and precipitation to the rear of the band fell into drier air and completely evaporated, resulting in the formation of a downdraft. This downdraft led to both adiabatic warming and a depression of the height of the inversion. This resulted in surface pressure falls, which first became evident at 1800 UTC, the time of gravity wave genesis in the simulation. The rainband and gravity wave remained at the northern edge of the dry air mass through 2000 UTC, after which they became decoupled. This decoupling was also noted in the observational data in Part I, but not until after 0000 UTC.
After the dry air and precipitation had decoupled, precipitation intensified and no longer evaporated aloft but fell through to the cool air mass below the inversion. As a result, surface pressures rose as evaporative cooling (and possibly water loading) within the column became important, as discussed by Koch et al. (1988). This led to the weakening of the pressure trough while the pressure ridge ahead of the trough intensified, as in observations (Koch and O'Handley 1997), but at an earlier time. These pressure changes are evident in time series data after 2000 UTC (Figs. 3e,f) and in cross sections of perturbation pressure (Fig. 4). While the heavier precipitation reaching the surface (after 2030 UTC) was coincident with the transition to a wave of elevation, the decrease in pressure–wind correlation discussed by Koch et al. (1988) was only temporary in the simulation. High values of pressure–wind correlation were restored after 2200 UTC (Fig. 5).
A vertical cross section (Fig. 8) through the dry air mass, rainband, and developing gravity wave makes clear the order of evolution. The modeled rainband developed aloft between 1400 and 1500 UTC in the Oklahoma panhandle, with evaporative cooling evident shortly before 1600 UTC as it tracked into southwest Kansas. The band formed within a region of high relative humidity aloft and ahead of a sharp gradient in RH, marking the leading edge of the dry air just above 700 mb. By 1800 UTC, surface pressure falls appeared, consistent with the subsidence warming and isentrope depression noted in Fig. 4. The rainband and wave intensified by 2000 UTC. Note, however, that the dry air mass had begun to move east and out of the cross section by 2000 UTC.
Figures 7 and 8 suggest that lifting at the leading edge of the advancing dry air triggered the formation of the rainband. Once the precipitation band was established, the next significant feature noted was subsidence warming in the lower troposphere (not shown). Wave genesis occurred shortly thereafter.
Trajectories were utilized to determine the source of the dry air mass that triggered the rainband. Figure 9 shows the backward trajectory of an air parcel whose final position was located behind the modeled RH boundary associated with the leading edge of the dry air mass over southwest Kansas at 1930 UTC. The parcel passed south of the Oklahoma panhandle at 500 m above the surface at 1400 UTC, and over the Rocky Mountains of central New Mexico at 0700 UTC. Vertical motions along the trajectory are indicated. The parcel originated near mountaintop, descended sharply, turned northeast across the Texas and Oklahoma panhandles, and ascended the warm front. The dry air mass at the rear of the rainband enhanced the evaporation and associated subsidence above the warm front, thereby causing the inversion layer depression and hydrostatic pressure falls leading to wave formation.
STORM-FEST observations presented in Parts I and II are consistent with the relationship between the precipitation band and gravity wave derived from the simulations. A distinct cloud band was observed over southwest Kansas in the 1400 UTC infrared satellite imagery. The band was first evident on radar at 1600 UTC, followed shortly thereafter by indications of virga (Trexler and Koch 2000) near the time of the first observed surface pressure perturbation at Garden City in southwest Kansas (Part I, Fig. 3c). The developing wave of depression was found on the increasingly sharp trailing edge of precipitation (KS99), as has been noted in other cases (Bosart and Seimon 1988; Ramamurthy et al. 1993). Heavy precipitation did not reach the ground until after 1900 UTC (KS99). Although initial surface pressure falls occurred near the same time and location as radar indications of precipitation aloft, causality could not be unambiguously determined from the observations. However, a strong relationship between the precipitation band and gravity wave was clear from the observations. Our modeling results suggest that the cloud band developed into a weak rainband, and evaporative processes acting on precipitation aloft resulted in subsidence above the frontal inversion and gravity wave genesis.
Two additional experiments (summarized in Table 2) were carried out to further ascertain the role of evaporative processes in gravity wave genesis. These simulations were identical to the control run for the first 12 h (through 1200 UTC). Following this period, in experiment NOEVAP, evaporative cooling was omitted for the remainder of the simulation, while latent heating was retained. By 1800 UTC (Fig. 10), the rainband was present as in CONTROL but the subsidence warming at 850 mb was absent, as was any evidence of wave genesis in mean sea level (MSL) pressure fields at this or later times (not shown).
The second simulation (RESTORE) was similar to experiment NOEVAP except that evaporative cooling was permitted again after 1630 UTC. Isentrope depression again became evident as pressure falls developed at the surface. Experiment RESTORE confirmed that any changes in cyclone structure from altering the precipitation physics were minor and did not preclude wave genesis. Therefore, we conclude that evaporatively driven downdrafts from the precipitation band led to wave genesis in the model simulation.
The sensitivity of the modeled subsidence and wave genesis to the cumulus parameterization was tested with the EXPLICIT experiment. While CONTROL utilized the Grell parameterization on all three grids, no cumulus parameterization was used on the innermost 6-km grid in EXPLICIT. Therefore, all precipitation was produced by the (Dudhia) ice microphysics treatment in the model. The development of the rainband and gravity wave (Figs. 11a,b) were largely unchanged, indicating that most of the precipitation and evaporative processes were explicit (grid resolved) in the CONTROL case, rather than parameterized. Later in the simulations (Figs. 11c,d), the EXPLICIT run exhibited larger vertical velocities and greater amounts of precipitation reaching the ground, relative to CONTROL, and the surface wave pressure signature was stronger.
Finally, we compare the mature gravity wave and rainband structures in CONTROL to that observed and documented in Parts I and II as the wave crossed the STORM-FEST dual-Doppler domain. Figure 12 depicts the horizontal and vertical velocity fields in the vicinity of the rainband (bold contour). Note the cross section is slightly over 100 km wide, for comparison to profiler (Part I, Fig. 10) and radar (Part II, Figs. 9–12 and 15) observations. Higher momentum air is seen at and upstream of the rainband, as in the Doppler analysis, but at a higher level (closer to the Hillsboro profiler data in Part I), and speeds are reduced (consistent, perhaps, with the slower propagation speed of the rainband in the simulation). Below and behind the 700–800-mb downdraft, the surface pressure trough (denoted with an L) is collocated with the wind minimum, indicating the presence of correlated wind and pressure perturbations.
In both the model fields and radar analyses, higher-momentum air is present upstream of the rainband as well as immediately downwind of the band at upper levels, while lower-momentum air is evident in the rainband and upwind of the band at upper levels. The flow reaching the rainband decelerates and splits in the vertical, with ascent above the wind maximum and descent below it. The peak vertical velocities are somewhat under 1 m s−1 for the CONTROL case, and slightly over 1 m s−1 in the EXPLICIT case for the same cross section and time (not shown). These are weaker than the updraft maxima in Fig. 11 of Part II but not unreasonably so given the difference between the resolution of the radar analysis and the model simulation.
The convergence and perturbation pressure fields over the same domain (Fig. 13) also compare favorably to observations. The convergence axis extends from the surface up into the rainband. The maximum convergence is found to the rear, where higher momentum air reaches the rainband. The divergence magnitude and pattern are quite similar to those in the dual-Doppler analyses. The MM5 perturbation pressure field (Fig. 13) contains near-surface evidence of the gravity wave, including low pressure at the rear of the rainband. While this trailing pressure minimum would fall outside of the observed radar echo region, the observed surface pressure trace supports the model field (Part II, Fig. 5). Note that the retrieved P′ fields in Part II are perturbations from an undisturbed initial state (see, e.g., Hane et al. 1981), while the MM5 perturbation pressure represents deviations from an initial 3D structure (Dudhia et al. 2002). As a result, the MM5 P′ fields incorporate a substantial large-scale component that is absent from the Doppler-derived analyses in Part II. Features in the retrieved P′ field (Part II, Fig. 12) such as high pressure in the low-level downdraft region and at the top of the rainband are found in the model, but they are partly masked by the larger-scale features. The difference in horizontal resolution (6 km versus 300 m for radar analyses) likely contributes to the smoothness of the P′ field in the CONTROL experiment. We note also that the structures seen in Figs. 12 and 13 are similarly depicted in the EXPLICIT simulation (not shown). Slightly greater convergence and stronger vertical velocities characterized the case with no cumulus parameterization on the 6-km domain.
5. Conceptual model
Figure 14 illustrates a conceptual model for the genesis (left side) and mature structure (right side) of the STORM-FEST gravity wave. This model is consistent with the detailed observations presented in Parts I and II, analyses presented by Koch and O'Handley (1997) and KS99, and with the model simulations. The structure of the dry air mass, rainband, and gravity wave are shown in vertical cross section (Fig. 14, top panels) and plan view (Fig. 14, middle panels). Surface pressure traces are also shown. At the time of genesis (near 1800 UTC), a rainband had formed at the leading edge of the advancing dry air mass and begun precipitating into the dry air to its southwest. The dry air mass at that time had ascended the warm frontal inversion so that precipitation occurred above the warm frontal surface. Precipitation was evaporating entirely aloft, resulting in a downdraft aloft, subsidence warming, and depression of the inversion, creating a weak solitary wave of depression. Both the rainband and incipient gravity wave occurred well within the upper-level dry slot, whose leading edge had already swept through Kansas.
The mature gravity wave and rainband structures are illustrated in the right side of Fig. 14. By this time a well-defined pattern of correlated surface pressure and wind perturbations had become established. The surface pressure ridge developed in response to precipitation falling through the stable layer to the surface. Dry air from the southwest enhanced the evaporation, creating a downdraft and subsidence warming at the rear of the rainband. The downdraft also depressed the inversion, creating a surface pressure fall and a wave of depression. A transition to a wave of elevation was beginning, as evaporative cooling began to dominate adiabatic warming above it. With time, these processes led to increasing pressure rises and diminishing pressure falls after the time shown. Eventually, in both observations and in the simulation, the rainband decoupled entirely from the dry air and the pressure ridge dominated the surface pressure signature.
6. Discussion
The STORM-FEST event we studied corresponded to the UK87 synoptic environment and was a dynamically active event. Koch and O'Handley (1997) identified unbalanced flow in the left exit region of the jet streak on this day, and Jin (1997) suggested geostrophic adjustment may have occurred. Our modeling results, however, strongly indicate that the mechanism for wave genesis in this case was evaporatively driven subsidence acting on the inversion layer beneath the rainband. Latent heating was present in all experiments, as rainband formation took place and updrafts developed, However, the clear indication of the wave in the simulation with full physics (Figs. 3–5) and its absence when evaporative processes were omitted (Fig. 10) underscore the role of evaporative processes in wave formation.
The possible importance of this mechanism in other cases is of interest. Mesoscale gravity waves are often accompanied by convection. Given the mechanism discussed here, elevated convection is most relevant. Elevated convection has been found to occur most often east of the Rocky Mountains, northeast of surface cyclone centers, and north of surface warm fronts (Colman 1990a), a pattern consistent with the gravity wave climatology. Assuming the presence of jet maxima accompanying surface cyclogenesis, it seems likely that such environments may also be characterized by geostrophic adjustment and/or shearing instability. Key to ascertaining a possible role for the present mechanism in observational data would be evidence of weak (or newly formed) precipitation bands immediately in advance of the first surface evidence of wave genesis.
Several authors have discussed convection above low-level stable layers and the generation of surface pressure perturbations. Colman (1990b) examined surface pressure and precipitation data for elevated convection on 10 April 1979 and concluded, “it is clear that the ducted gravity waves … and the elevated thunderstorms are intimately related.” Bauck (1992) documented a case in which convective downdrafts above a stable marine layer triggered gravity waves, which propagated over 200 km from their source. Powers (1997) discussed wave generation in the context of elevated convection and the lifting and depression of the underlying stable layer. Stumpf et al. (1991) studied the wake low behind a mature mesoscale convective system (MCS) and concluded that downdrafts reduced the stable layer (convective cold pool) depth, lowering the surface pressure. Should such a downdraft descend through a stable layer to the ground, a heat burst can result (Johnson et al. 1989).
The surface perturbation pressure and wind fields accompanying the mature gravity wave in the model (Fig. 6b) strongly resemble the mesohigh and wake low (e.g., Johnson and Hamilton 1988, hereafter JH88; see their Fig. 25) accompanying an MCS. The wake low is known to result from subsidence warming accompanying descending rear inflow into the back of an MCS (JH88). In a two-dimensional modeling study, Gallus (1996) found that microphysical processes alone could result in the formation of a wake low when decreasing precipitation rates were specified over time. This mimicked a collapsing precipitation core or movement of a sharp reflectivity gradient, such that adiabatic (subsidence) warming dominated over latent cooling. In our modeling study, the gravity wave trough and mesolow appear to be one and the same, forming in response to evaporative processes.
Several authors have noted the relationship between gravity waves and wake lows. Bosart and Seimon (1988) examined a gravity wave at the trailing edge of a squall line and hypothesized that subsidence into the stable boundary layer could “organize and intensify” the gravity wave. Branick et al. (1988) discussed a long-lived wake trough and found evidence of a gravity wave in their surface time series data. KS99 noted the correlated pressure and wind fields in the JH88 conceptual model of the wake low and mesohigh, and found that positive pressure perturbations were located very near to rainbands in three cases from STORM-FEST.
This study indicates that evaporative processes associated with the rainband caused gravity wave formation on 14 February 1992. The environments in which gravity wave genesis and wake low formation occur are both characterized by subsidence. The wake low at the rear of a mature squall-line system typically forms as a result of descending rear inflow, in a region where subsidence warming dominates over evaporative cooling (JH88). The only stable layer present is usually that associated with the convective cold pool. Gravity wave genesis requires a stable layer, which usually takes the form of a warm frontal inversion, as in STORM-FEST. Rear inflow was not required, as unsaturated subsidence associated with evaporation of elevated precipitation resulted in downdrafts, which reduced the frontal inversion depth and resulted in gravity wave genesis.
7. Summary
We have successfully simulated the genesis of a gravity wave that occurred on 14 February 1992 during STORM-FEST. In both the simulations and the observations, a rainband formed at the leading edge of a surge of dry air that originated as downslope flow east of the Rocky Mountains. The dry air mass, rainband, and gravity wave remained tied together as they propagated over a warm frontal surface over northeast Kansas. Our results show that evaporative processes associated with the rainband resulted in subsidence warming and depression of the underlying warm-frontal inversion. The reduced inversion height produced surface pressure falls, which represented the surface manifestation of a developing gravity wave. Numerical experiments with and without evaporative processes isolated the dominant role of evaporatively driven downdrafts as key to wave genesis.
Several questions remain at the conclusion of this work. The role of this mechanism in wave maintenance, compared to ducting and wave–CISK, has not been assessed. The applicability of this mechanism to other cases in the central and eastern United States remains to be determined. Given the key role of evaporative processes in the STORM-FEST case, identifying and clarifying the relationship between developing, elevated convection and the earliest signs of wave genesis in other cases is key to determining the importance of this mechanism. Modeling work is underway to ascertain the possible role of this mechanism in other published cases of observed large-amplitude gravity wave events.
Acknowledgments
This work was supported by the National Science Foundation under Grants ATM-9708170 and ATM-0004274. Numerical simulations were carried out on the Cray J924 at the National Center for Atmospheric Research. Discussions with Dr. Steven Koch of NOAA/FSL are gratefully acknowledged. We thank Dr. Muqun Yang for his contributions to Parts I and II and discussions regarding this work.
REFERENCES
Balachandran, N. K., 1980: Gravity waves from thunderstorms. Mon. Wea. Rev., 108 , 804–816.
Bauck, B. H., 1992: An unusually strong gravity wave over western Washington. Wea. Forecasting, 7 , 389–397.
Blackadar, A. K., 1979: High resolution models of the planetary boundary layer. Advances in Environmental Science and Engineering, Vol. 1, No. 1. J. R. Pfafflin and E. N. Ziegler, Eds., Gordon and Briech, 50–85.
Bosart, L. F., and F. Sanders, 1986: Mesoscale structure in the megalopolitan snowstorm of 11–12 February 1983. Part III: A large-amplitude gravity wave. J. Atmos. Sci., 43 , 924–939.
Bosart, L. F., and A. Seimon, 1988: A case study of an unusually intense gravity wave. Mon. Wea. Rev., 116 , 1857–1886.
Bosart, L. F., W. E. Bracken, and A. Seimon, 1998: A study of cyclone mesoscale structure with emphasis on a large-amplitude inertia–gravity wave. Mon. Wea. Rev., 126 , 1497–1527.
Bradshaw, J. T., R. A. Murphy, and K. J. Pence, 1999: The Alabama gravity wave event of February 22, 1998: A classic mesoscale forecasting challenge. Preprints, Eighth Conf. on Mesoscale Processes, Boulder, CO, Amer. Meteor. Soc., 130–135.
Branick, M. L., F. Vitale, C-C. Lai, and L. F. Bosart, 1988: The synoptic and subsynoptic structure of a long-lived severe convective system. Mon. Wea. Rev., 116 , 1335–1370.
Clark, T. C., T. Hauf, and J. P. Kuettner, 1986: Convectively forced internal gravity waves: Results from two-dimensional numerical experiments. Quart. J. Roy. Meteor. Soc., 112 , 899–925.
Colman, B. R., 1990a: Thunderstorms above frontal surfaces in environments without positive CAPE. Part I: A climatology. Mon. Wea. Rev., 118 , 1103–1122.
Colman, B. R., 1990b: Thunderstorms above frontal surfaces in environments without positive CAPE. Part II: Organization and instability mechanisms. Mon. Wea. Rev., 118 , 1123–1144.
Cram, J. M., R. A. Pielke, and W. R. Cotton, 1992: Numerical simulation and analysis of a prefrontal squall line. Part II: Propagation of the squall line as an internal gravity wave. J. Atmos. Sci., 49 , 209–225.
Davies, H. C., 1979: Phase-lagged wave-CISK. Quart. J. Roy. Meteor. Soc., 105 , 325–353.
Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46 , 3077–3107.
Dudhia, J., D. Gill, K. Manning, A. Bourgeois, and W. Wang, cited 2002: PSU/NCAR mesoscale modeling system tutorial class notes and users' guide (MM5 modeling system version 3). NCAR. [Available online at www.mmm.ucar.edu/mm5/doc.html.].
Einaudi, F., W. L. Clark, D. Fua, J. L. Green, and T. E. VanZandt, 1987: Gravity waves and convection in Colorado during July 1983. J. Atmos. Sci., 44 , 1534–1553.
Eom, J. K., 1975: Analysis of the internal gravity wave occurrence of 19 April 1970 in the Midwest. Mon. Wea. Rev., 103 , 217–226.
Ferretti, R., F. Einaudi, and L. W. Uccellini, 1988: Wave disturbances associated with the Red River valley severe weather outbreak of 10–11 April 1979. Meteor. Atmos. Phys., 39 , 132–168.
Gallus Jr., W. A., 1996: The influence of microphysics in the formation of intense wake lows: A numerical modeling study. Mon. Wea. Rev., 124 , 2267–2281.
Grell, G. A., 1993: Prognostic evaluation of assumptions used by cumulus parameterizations. Mon. Wea. Rev., 121 , 764–787.
Grell, G. A., J. Dudhia, and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note TN-398+STR, 122 pp.
Hamilton, D. W., Y-L. Lin, R. P. Weglarz, and M. L. Kaplan, 1998: Jetlet formation from diabatic forcing with applications to the 1994 Palm Sunday tornado outbreak. Mon. Wea. Rev., 126 , 2061–2089.
Hane, C. E., R. B. Wilhelmson, and T. Gal-Chen, 1981: Retrieval of thermodynamic variables within deep convective clouds: Experiments in three dimensions. Mon. Wea. Rev., 109 , 564–576.
Jin, Y., 1997: A numerical model study of the role of mesoscale gravity waves in rainband dynamics in the central United States during STORM-FEST. Ph.D. dissertation, North Carolina State University, 318 pp. [Available from Dept. of Marine, Earth and Atmospheric Sciences, North Carolina State University, Campus Box 8208, Raleigh, NC 27695-8208.].
Johnson, R. H., and P. J. Hamilton, 1988: The relationship of surface pressure features to the precipitation and airflow structure of an intense midlatitude squall line. Mon. Wea. Rev., 116 , 1444–1473.
Johnson, R. H., S. Chen, and J. J. Toth, 1989: Circulations associated with a mature-to-decaying midlatitude mesoscale convective system. Part I: Surface features—Heat bursts and mesolow development. Mon. Wea. Rev., 117 , 942–959.
Kaplan, M. L., S. E. Koch, Y-L. Lin, R. P. Weglarz, and R. A. Rozumalski, 1997: Numerical simulations of a gravity wave event over CCOPE. Part I: The role of geostrophic adjustment in mesoscale jetlet formation. Mon. Wea. Rev., 125 , 1185–1211.
Kaplan, M. L., Y-L. Lin, D. W. Hamilton, and R. A. Rozumalski, 1998: The numerical simulation of an unbalanced jetlet and its role in the Palm Sunday 1994 tornado outbreak in Alabama and Georgia. Mon. Wea. Rev., 126 , 2133–2165.
Karyampudi, V. M., M. L. Kaplan, S. E. Koch, and R. J. Zamora, 1995a: The influence of the Rocky Mountains on the 13–14 April 1986 severe weather outbreak. Part I: Mesoscale lee cyclogenesis and its relationship to severe weather and dust storms. Mon. Wea. Rev., 123 , 1394–1422.
Karyampudi, V. M., S. E. Koch, C. Chen, J. W. Rottman, and M. L. Kaplan, 1995b: The influence of the Rocky Mountains on the 13–14 April 1986 severe weather outbreak. Part II: Evolution of a prefrontal bore and its role in triggering a squall line. Mon. Wea. Rev., 123 , 1423–1446.
Koch, S. E., 1979: Mesoscale gravity waves as a possible trigger of severe convection along a dryline. Ph.D. thesis, University of Oklahoma, Norman, OK, 195 pp.
Koch, S. E., and P. B. Dorian, 1988: A mesoscale gravity wave event observed during CCOPE. Part III: Wave environment and probable source mechanisms. Mon. Wea. Rev., 116 , 2570–2592.
Koch, S. E., and R. E. Golus, 1988: A mesoscale gravity wave event observed during CCOPE. Part I: Multiscale statistical analysis of wave characteristics. Mon. Wea. Rev., 116 , 2527–2544.
Koch, S. E., and C. O'Handley, 1997: Operational forecasting and detection of mesoscale gravity waves. Wea. Forecasting, 12 , 253–281.
Koch, S. E., and L. M. Siedlarz, 1999: Mesoscale gravity waves and their environment in the central United States during STORM-FEST. Mon. Wea. Rev., 127 , 2854–2879.
Koch, S. E., and S. Saleeby, 2001: An automated system for the analysis of gravity waves and other mesoscale phenomena. Wea. Forecasting, 16 , 661–679.
Koch, S. E., R. E. Golus, and P. B. Dorian, 1988: A mesoscale gravity wave event observed during CCOPE. Part II: Interactions between mesoscale convective systems and the antecedent waves. Mon. Wea. Rev., 116 , 2545–2569.
Koch, S. E., F. Einaudi, P. B. Dorian, S. Lang, and G. M. Heymsfield, 1993: A mesoscale gravity wave event observed during CCOPE. Part IV: Stability analysis and Doppler-derived wave vertical structure. Mon. Wea. Rev., 121 , 2483–2510.
Koch, S. E., D. Hamilton, D. Kramer, and A. Langmaid, 1998: Mesoscale dynamics in the Palm Sunday tornado outbreak. Mon. Wea. Rev., 126 , 2031–2060.
Koch, S. E., F. Zhang, M. L. Kaplan, Y-L. Lin, R. Weglarz, and C. M. Trexler, 2001: Numerical simulations of a gravity wave event over CCOPE. Part III: The role of a mountain–plains solenoid in the generation of the second wave episode. Mon. Wea. Rev., 129 , 909–933.
Koppel, L. L., L. F. Bosart, and D. Keyser, 2000: A 25-yr climatology of large-amplitude hourly surface pressure changes over the conterminous United States. Mon. Wea. Rev., 128 , 51–68.
Ley, B. E., and W. R. Peltier, 1978: Wave generation and frontal collapse. J. Atmos. Sci., 35 , 3–17.
Lin, Y-L., and R. C. Goff, 1988: A study of a mesoscale solitary wave originating near a region of deep convection. J. Atmos. Sci., 45 , 194–205.
Lindzen, R. S., 1974: Wave-CISK in the tropics. J. Atmos. Sci., 31 , 156–179.
Lindzen, R. S., and M. Fox-Rabinovitz, 1989: Consistent vertical and horizontal resolution. Mon. Wea. Rev., 117 , 2575–2583.
Miller, D. A., and F. Sanders, 1980: Mesoscale conditions for the severe convection of 3 April 1974 in the east-central United States. J. Atmos. Sci., 37 , 1041–1055.
Pecnick, M. J., and J. A. Young, 1984: Mechanics of a strong subsynoptic gravity wave deduced from satellite and surface observations. J. Atmos. Sci., 41 , 1850–1862.
Persson, P. O. G., and T. T. Warner, 1991: Model generation of spurious gravity waves due to inconsistency of the vertical and horizontal resolution. Mon. Wea. Rev., 119 , 917–935.
Pokrandt, P. J., G. J. Tripoli, and D. D. Houghton, 1996: Processes leading to the formation of mesoscale gravity waves in the Midwest cyclone of 15 December 1987. Mon. Wea. Rev., 124 , 2726–2752.
Powers, J. G., 1997: Numerical model simulations of a mesoscale gravity wave event: Sensitivity tests and spectral analyses. Mon. Wea. Rev., 125 , 1838–1869.
Powers, J. G., and R. J. Reed, 1993: Numerical simulation of the large-amplitude mesoscale gravity-wave event of 15 December 1987 in the central United States. Mon. Wea. Rev., 121 , 2285–2308.
Ramamurthy, M., R. M. Rauber, B. P. Collins, and N. K. Malhotra, 1993: A comparative study of large-amplitude gravity-wave events. Mon. Wea. Rev., 121 , 2951–2974.
Rauber, R. M., M. Yang, M. K. Ramamurthy, and B. F. Jewett, 2001: Origin, evolution, and fine scalestructure of the St. Valentine's Day gravity wave observed during STORM-FEST. Part I: Origin and maintenance. Mon. Wea. Rev., 129 , 198–217.
Raymond, D. J., 1975: A model for predicting the movement of continuously propagating convective storms. J. Atmos. Sci., 32 , 1308–1317.
Raymond, D. J., 1976: Wave-CISK and convective mesosystems. J. Atmos. Sci., 33 , 2392–2398.
Schmidt, J. M., and W. R. Cotton, 1990: Interactions between upper and lower tropospheric gravity waves on squall line structure and maintenance. J. Atmos. Sci., 47 , 1205–1222.
Snyder, C., W. C. Skamarock, and R. Rotunno, 1993: Frontal dynamics near and following frontal collapse. J. Atmos. Sci., 50 , 3194–3212.
Stobie, J. G., F. Einaudi, and L. W. Uccellini, 1983: A case study of gravity waves–convective storms interaction: 9 May 1979. J. Atmos. Sci., 40 , 2804–2830.
Stumpf, G. J., R. H. Johnson, and B. F. Smull, 1991: The wake low in a midlatitude mesoscale convective system having complex convective organization. Mon. Wea. Rev., 119 , 134–158.
Szoke, E. J., J. M. Brown, J. A. McGinley, and D. Rodgers, 1994: Forecasting for a large field program: STORM-FEST. Wea. Forecasting, 9 , 593–605.
Trexler, C. M., and S. E. Koch, 2000: The life cycle of a mesoscale gravity wave as observed by a network of Doppler wind profilers. Mon. Wea. Rev., 128 , 2423–2446.
Uccellini, L. W., and S. E. Koch, 1987: The synoptic setting and possible energy sources for mesoscale wave disturbances. Mon. Wea. Rev., 115 , 721–729.
Van Tuyl, A. H., and J. A. Young, 1982: Numerical simulation of nonlinear jet streak adjustment. Mon. Wea. Rev., 110 , 2038–2054.
Xu, Q., and J. H. Clark, 1984: Wave CISK and mesoscale convective systems. J. Atmos. Sci., 41 , 2089–2107.
Yang, M., R. M. Rauber, and M. K. Ramamurthy, 2001: Origin, evolution, and finescale structure of the St. Valentine's Day gravity wave observed during STORM-FEST. Part II: Finescale structure. Mon. Wea. Rev., 129 , 218–236.
Zhang, D-L., and J. M. Fritsch, 1988: Numerical simulation of the meso-β scale structure and evolution of the 1977 Johnstown flood. Part III: Internal gravity waves and the squall line. J. Atmos. Sci., 45 , 1252–1268.
Zhang, F., and S. E. Koch, 2000: Numerical simulations of a gravity wave event over CCOPE. Part II: Waves generated by an orographic density current. Mon. Wea. Rev., 128 , 2777–2796.
MM5 model grid configuration used in this study. The horizontal grid resolution for each domain appears in the lower-left corner. The initial location of the moving innermost 6-km nest is shown with a bold outline; the two later positions are hatched
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
(a)–(c) The 12-h MM5 forecasts and (d) infrared satellite image valid at 1200 UTC 14 Feb 1992. All model fields are from the outer (54 km) grid: (a) 250-mb height (thick contours, m) and isotachs (shaded, with 2.5 m s−1 contours dashed); (b) 500-mb RH (shaded, %) with the central plains region of 100% RH at 900 mb shown hatched, with a bold outline; and (c) sea level pressure (thick lines, P-1000 mb), surface temperature (thin contours, every 4°C), wind vectors (every second), and fronts
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
(a) The 2000 UTC MM5 surface perturbation pressure (every 0.25 mb, dashed negative) and wind vectors (every fifth grid point). Filled circles are locations of time series appearing in (c)–(f). (b) Cross section of potential temperature (dashed, every 2.5 K) and perturbation pressure (solid, every 0.2 mb) along axis in (a). Bold dashed line identifies pressure trough. (c)–(f) MM5 time series of detrended wave-relative surface wind (U′, dashed lines, m s−1) and perturbation pressure (P′, solid lines, mb) for 1500–0000 UTC. Locations of each trace appear in (a)
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
(top row) MM5 cross sections of potential temperature (light lines, every 2.5 K) and perturbation pressure (dark contours, every 0.2 mb) at 1815, 2000, and 2200 UTC. (bottom row) The 850-mb potential temperature (every 0.5 K) and wind vectors (every fifth grid point) in central and eastern KS for the same times. A contour identifying the warm perturbation is shaded in each figure. The cross-section axis is also outlined
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
(a) Isochrones, with regions of correlation exceeding 0.9 shaded. Isochrones are drawn with thin (thick) lines identifying the wave trough (ridge) at 1-h intervals, with hour shown as boxed text. Shaded correlation regions are for 3-h periods: 1800–2100, 1900–2200, and 2000–2300 UTC. Map at lower right shows analysis region. (b) Isochrones and region of correlation exceeding 0.9 (shaded) for the 3-h time window centered on the detrended surface pressure minimum at each MM5 6-km grid point
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
MM5 6-km fields illustrating the wave and precipitation. (a) Cross section [axis shown in (d)] of potential temperature (every 2.5 K, thin lines), rainwater (contoured and hatched), and vertical velocity (light shading, updraft; dark, downdraft). (b) Surface rainwater (contoured every 0.025 g kg−1) and 900-mb vertical velocity, (updraft, light shading; downdraft, dark) and 2100 UTC axes of the pressure trough (P−) and ridge (P+). (c) Time series [at circled locations in (b), (d)] of detrended pressure (P, mb) and wave-relative velocity (U, m s−1), σ = 0.9875 (approx 95 m AGL) vertical velocity (W, cm s−1) and low-level rainwater (Qr, g kg−1) from 1800 to 2300 UTC. (d) Surface fronts, 550-mb rainwater (light shading), surface rainwater exceeding 0.05 g kg−1 (dark shading) and wind vectors (every fifth grid point), and 850-mb potential temperature exceeding 293.5 K (hatched)
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
The 650-mb RH (every 15%, and shaded below 30% and 60%), winds (every fifth grid point, with maximum wind at lower left), and rainwater (contoured within dark shading every 0.05 g kg−1) on the 18-km grid. Isochrones of minimum pressure (from 6 km MSL pressure fields) are shown with light line segments at 1800, 2000, 2200, and 0000 UTC
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
Relative humidity (bold lines, every 15%, and shaded below 30% and 60%), rainwater (every 0.075 g kg−1 within dark shading and thick contour), and perturbation pressure (as defined in MM5; thin lines, every 0.2 mb) on the 6-km grid. Arrows in bottom two panels show the locations of the surface pressure troughs. The cross-section axis is shown in Fig. 7
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
MM5 trajectory path (thick line) ending at 1930 UTC 14 Feb. The 298-K RH (every 10%) is shown by thin lines, and 650-mb rainwater is contoured every 0.05 g kg−1 within dark shading. Along the path, time, elevation (MSL), and vertical velocity are listed. Inset: 3D view, looking west
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
Experiments showing the impact of evaporative cooling. Rainwater (every 0.05 g kg−1, solid) and potential temperature (dashed, every 2K) at (left) 1800 and (right) 1930 UTC are shown, with vertical velocity exceeding ±10 cm s−1 shaded. See discussion in text for details
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
Comparison of simulations CONTROL and EXPLICIT. (a) MM5 cross sections at 1815 UTC [axis in (b)] of potential temperature (light lines, every 2.5 K), rainwater outline (heavy black line is 0.05 g kg−1), and vertical velocity (shaded above/below 20 cm s−1, with descent dark, and ascent light) for CONTROL (left) and EXPLICIT (right). (b) Plan view of 650-mb (thin lines, every 0.075 g kg−1) and surface (shaded region exceeds 0.075 g kg−1) rainwater at 1815 UTC for CONTROL (left) and EXPLICIT (right). (c) As in (a), but at 2115 UTC. (d) As in (b) at 2115 UTC. Short line segment identifies the primary pressure minimum in (b) and (d)
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
The 6-km MM5 cross section of tangential wind (U, thin contours every 1 m s−1 and shaded over 10 m s−1), vertical velocity exceeding ±20 cm s−1 (W, hatched), and rainwater (outer 0.05 g kg−1 bold contour shown) valid at 2100 UTC. Arrows identify velocity maxima and minima, and H and L identify near-surface pressure extrema
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
The 6-km MM5 cross section of convergence (shaded and hatched exceeding ±0.2, 0.4, 0.6 × 10−3 s−1), perturbation pressure (PP, thin contours every 0.2 mb), vertical velocity exceeding ±20 cm s−1 (W, hatched), and rainwater (outer 0.05 g kg−1 bold contour shown) valid at 2100 UTC. Arrows show velocity maxima and minima from Fig. 12. Thin dashed line represents U = 10 m s−1, and the bold dashed line identifies the lower-tropospheric convergence axis. Letters C and D represent convergence and divergence maxima, and H and L identify near-surface pressure extrema
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
Conceptual model of (left) gravity wave genesis and (right) mature structure on 14 Feb 1992. (top to bottom) Cross section, surface pressure trace, plan view, and sequence of steps to wave genesis. Dark shading denotes the low-level stable layer and cloud and rainbands, and light shading the midtropospheric dry air mass and upper-tropospheric dry slot air. Airflow is indicated with dark arrows. The surface dryline and warm front are shown, with the bold dashed line denoting the leading edge of the dry air mass
Citation: Monthly Weather Review 131, 4; 10.1175/1520-0493(2003)131<0617:OEAFSO>2.0.CO;2
MM5 model parameters—CONTROL case
Summary of model simulations