Recent numerical studies have shown that surface drag has a profound effect on the evolution of frontal structures within developing baroclinic waves: with surface drag included, warm frontogenesis is much reduced, while cold frontogenesis is not. In a previous study, Richard Rotunno, William Skamarock, and Christopher Snyder found that geostrophic warm frontogenesis is a more sensitive feature than is geostrophic cold frontogenesis, since the former depends on a rapid turning of the isobars in the vicinity of the surface cyclone center, while the latter depends on the diffluence of the isobars occurring in the zone between the surface cyclone and anticyclone. The inclusion of surface drag induces Ekman pumping, most strongly in the vicinity of the surface low; this pumping near the low vastly reduces the curvature of the isobars (since interior geostrophic vorticity is reduced), upon which the warm front depends for its existence. Furthermore, since the effects of Ekman pumping are most pronounced near the surface cyclone, the cold front, which forms in flow far from the cyclone center, is relatively unaffected.
Snyder and Daniel Keyser (State University of New York, Albany) have completed their study of how the planetary boundary layer modifies the structure and dynamics of surface fronts. Using a two-dimensional, non-hydrostatic, numerical model, they considered the "spin down" of a frontal zone, initially in thermal-wind balance, that is subjected to parameterized surface drag and eddy mixing. In order to simplify the problem, neither surface heat flux, nor large-scale forcing of frontogenesis, are allowed. For typical mesoscale fronts, the convergence of the frictionally driven flow overwhelms the frontolytic effects of mixing within the planetary boundary layer, resulting in robust frontogenesis. Thus, it seems unlikely that boundary-layer mixing can limit the scale of atmospheric fronts, except perhaps when the front reaches scales comparable to typical boundary-layer eddies, which, of course, are not explicitly represented in these calculations. Snyder and Keyser also found that, as the frontogenesis proceeds (and the model grid spacing is decreased below a few hundred meters in the horizontal), the leading edge of the surface front acquires characteristics of a density current, balancing between the cross-frontal acceleration and the pressure gradient, and a frontal head that breaks down into Kelvin-Helmholtz billows.
Christopher Davis [joint appointment with NCAR's Research Applications Program RAP] continued to study the effect of large-scale mountains on baroclinic waves. Additional nonlinear quasi-geostrophic (QG) simulations were performed which examined how the structure and growth rate of baroclinic waves were altered by the presence of a mountain. The main finding is that the mountain alters the mean distribution of surface potential temperature, and hence changes the propagation characteristics of the incident baroclinic wave. The mountain presence enhances the gradient of theta(*), the QG approximation to surface potential temperature, to the north of the peak and decreases it to the south. Thus, the component of the baroclinic wave that is identified with surface potential temperature perturbations propagates around the north side of the mountain and accelerates. This leads to a change in vertical structure of the incident wave which, for the wavelengths considered, systematically results in a smaller growth rate than one would expect without the mountain.
The structure of the baroclinic waves over and downstream of the mountain varies substantially depending on the location of the jet. For a jet to the north of the mountain, a strong, synoptic-scale "cold surge" develops in the lee, governed by QG dynamics. Upslope cooling reinforces horizontal temperature advection and an anticyclone intensifies, moving southward along the contours of theta(*). With the jet to the south, the anticyclone weakens and the cyclone in the lee dominates.
Xiaolei Zou, Ying-Hwa Kuo, and Simon Low-Nam performed a series of forecast experiments on the Experiment of Rapidly Intensifying Cyclone Over the Atlantic Intensive Observation Period 4 (ERICA IOP4) storm. For this they used a hemispheric version of MM5 (NCAR/Penn State University Mesoscale Model, version 5) at a horizontal resolution of 120 km. They showed that the accuracy of the model decreased progressively as the forecast duration was increased. A model forecast fracture occurred between days 4.5 and 5. The forecast that was initialized at 1200 UTC 31 December 1988 failed to predict the storm, while the experiment that was initialized 12 hours later captured the cyclogenesis. An interesting question is "What was missing in the model's initial conditions of 1200 UTC 31 December that prevented the development of the storm?" To answer this question, Zou and colleagues performed an adjoint sensitivity analysis based on the 12-hour model forecast error. The sensitivity fields suggest errors in the analysis of the initial conditions, an upper-level potential vorticity (PV) anomaly over the Gulf of Alaska and a cut-off low off the coast of California. After corrections, a perturbation field was derived, based on the minimization of 12-hour forecast errors, and added to the initial conditions of 1200 UTC 31 December. The perturbed run successfully reproduced the cyclogenesis over the Western Atlantic five days later. Analysis of the perturbed run shows that improvements in the analysis of these initial conditions are important in bringing the PV anomaly with sufficient intensity to the east coast of the U.S. to interact with the low-level baroclinic zone over the Western Atlantic, which is responsible for the development of the ERICA IOP4 storm. These results showed that the adjoint technique can be used to identify deficiencies in model initial conditions, and to improve the accuracy of medium range prediction of extratropical cyclogenesis.
Kuo, along with Shou-Jun Chen and Zu-Yu Tao (visitors from Peking University, China), and Bo Cui (visitor from the Chinese Meteorological Center) completed a modeling study of a heavy rainfall event that occurred in June 1991. A series of mesoscale convective systems developed successively and propagated eastward along the front, producing a rainfall maximum of 234 mm in 24 hours over the Jiang-Huai Basin. Using MM5 at a horizontal resolution of 54 km, Chen and colleagues simulated the sequential development of three rainstorms along the Mei-Yu front, two over the China mainland and one over the East China sea. The spacing, intensity, and size of the simulated rainstorms are very similar to the observed convective systems. The model also simulated the development of a meso-alpha-scale low pressure system associated with a rainstorm. As a rainstorm develops along the Mei-Yu front, vertical velocity increases by an order of magnitude (from 6 cm/s to 70 m/s). Relative vorticity and potential vorticity also increase rapidly. A mesoscale low-level jet forms to the south of the rainstorm in response to the strong latent heat release. The mesoscale low-level jet acts as a conveyer of warm, moist air and provides the additional dynamic lifting for the mesoscale convective system, establishing a positive feedback. These results suggest that mesoscale rainstorm systems on the Mei-Yu front can be viewed as a self-developing convective-baroclinic system. Latent heat release plays a crucial role in producing rapid spin-up of low-level cyclonic vorticity and pressure fall, acting on the low-level horizontal wind shear and broad baroclinic zone associated with the Mei-Yu front.
Kuo, Chen, An-Yu Wang (visitor from Zhong-Shan University, China), Alexis Lau [visitor from Hong Kong University of Science and Technology (HKUST)], and Jimy Dudhia performed a month-long simulation of the East Asian monsoon using the MM5 model for June 1994. The model domain extends from 55E to 175E, and from 20S to 60S with a horizontal resolution of 162 km. Because the model domain is fairly large, the model physical parameterization can exert a strong influence on the evolution of model climate, more so than the impact of lateral boundary conditions. They found that the MM5 model was able to reproduce the monthly mean features of the monsoon circulation and the rainfall of June 1994. Moreover, the changes of the large-scale circulation patterns and the evolution of the East Asian summer monsoon were successfully simulated, including: (1) the northward shift of the upper-level westerlies over the Tibetan Plateau and the development of the upper-level easterlies to the south of the Plateau, (2) the extension of the low-level southwesterly monsoon flow from the South China Sea to eastern China and the sequential jumps of the Mei-Yu front, and (3) the northward shift of the mid-tropospheric western Pacific subtropical high at 500 mb and the establishment of the Tibetan Plateau high at 200 mb. An additional experiment, with latent heating from cumulus parameterization excluded, showed that convective heating is crucial for the northward shift of the upper-level westerlies over the Tibetan Plateau, the establishment of the southwesterly monsoon flow, and the sequential jumps of the Mei-Yu front. These results indicate that MM5, even at a horizontal resolution of 162 km, is capable of regional climate simulation. This opens up future research opportunities for high-resolution regional climate simulation using the nested MM5 model at a considerably higher resolution and with explicit physical parameterization.
Kevin Knupp, Bart Geerts (University of Alabama, Huntsville) and John Tuttle (joint appointment with RAP) have completed a study investigating the evolution of mesoscale airflow and stratiform precipitation in a small, short-lived MCS observed in northern Alabama during the Cooperative Huntsville Meteorological Experiment (COHMEX). The COHMEX mesoscale observation network was situated such that it covered most of the MCS during its entire life cycle. The MCS developed from a cluster of convective cells in a barotropic environment characterized by moderate instability and weak shear. Even though the MCS was highly unsteady during most of its five-hour lifecycle, it exhibited many of the features commonly found in larger, long-lived systems. A descending rear inflow jet was observed before convective activity peaked, and remained strong even after the deep convection diminished. During the mature stage the MCS had a large trailing stratiform region which was separated from the leading squall line by a transition zone. The study demonstrated that even though the small, unsteady MCS developed in a weakly sheared environment that was insufficient to support a long-lived system, it assumed essentially the same precipitation and airflow characteristics as larger, long-lived MCSs.
Richard Carbone and Mitchell Moncrieff participated in the Maritime Continent Thunderstorm Experiment (MCTEX) which was conducted in November and December of 1995 on the Tiwi Islands north of the Australian Continent. Carbone's participation concerned the evolution and propagation of convection that is forced by islands with zonal orientation and small topography in an easterly jet regime. At least ten case studies of evolving convection were obtained over a six week period, including some that evolved from basic seabreeze forcing through cloud clusters and weak tropical squalls, to full fledged MCSs with strong meridional orientation and a dominant convective cell structure. Data have subsequently been calibrated and archived and a seasonal morphology is now being compiled. Early findings were presented in Reading, England by Thomas Keenan [Bureau of Meteorology Research Center (BMRC), Australia] and colleagues. Diagnostic studies of individual cases will be conducted with James Wilson (ATD) and Keenan in 1997.
Andrew Crook examined the development of convection over the Tiwi Islands (located just north of Darwin, Australia) using the Clark nonhydrostatic numerical model. Results are being compared with observations taken during MCTEX. An important feature of this study is that the simulations are being performed with sufficient resolution to capture some of the larger eddies in the convective boundary layer. Initial results from a case with large-scale westerly flow (along the major axis of the Tiwi Islands) indicate that storms initially develop along a low-level convergence line that extends the full length of the island complex. Evaporative cooling from these storms then forms a low-level cold pool which begins to propagate westward against the large-scale flow. Initial results also indicate the importance of ice processes in giving the explosive development, with updraft speeds approaching 50 m/s, that is often observed in storms over the islands.
Observations reveal that a variety of mesoscale convective vortices (MCVs) with horizontal scales ranging from tens to hundreds of kilometers, and time scales ranging from hours to days, are often generated within mesoscale convective systems. These MCVs play a variety of roles, from locally enhancing the strength of rear-inflow jets, which may contribute to the production of damaging surface winds, to initiating the development of long-lived, balanced, mid-tropospheric circulations, which can help trigger new convective episodes on subsequent days. Through the use of idealized non-hydrostatic numerical cloud model simulations, Morris Weisman and Christopher Davis have demonstrated that such vortices can originate via the tilting of system-generated horizontal vorticity at the ends of finite-length convective line segments. This process leads to the development of both cyclonic and anticyclonic line-end vortices to the left and right, respectively, of the mean low-level vertical wind shear vector. Over the longer-term, the convergence of Coriolis rotation at mid-levels may lead to the development of a preferred cyclonic vortex, which is frequently observed in asymmetric convective systems. The strength and scale of the resultant vortices was also found to depend on the strength of the ambient vertical wind shear, with the stronger, smaller vortices produced only in the more strongly sheared environments. These results emphasize that divergent, convective motion must be included in numerical or theoretical models that are designed to study or forecast the generation of such features.
Much confusion has arisen recently in both the forecast and research communities as to the role of vertical wind shear versus storm-relative environmental helicity in controlling supercell dynamics for straight versus curved hodograph environments. In an attempt to remedy this confusion, Weisman has completed a set of idealized numerical simulations of supercell storms that help to more clearly document the dynamical character of supercells in these various shear regimes. Analysis of these simulations clearly show that the dynamical forcing of the updrafts for both the straight and curved hodographs are fundamentally the same. This is in contrast to the helicity approach, which suggests that the curved-hodograph dynamics are significantly different than those with a straight hodograph, and that supercell potential is enhanced in curved-hodograph environments. The present results reinforce the universal character of the updraft-vertical wind shear interactions that lead to the production of long-lived, rotating convective storms, and further emphasize that supercell potential is related to the overall strength of the environmental vertical wind shear rather than just hodograph shape.
Eugene McCaul (NASA/Marshall Space Flight Center) and Weisman continued their studies into the dynamics of supercell storms by documenting the roles of the stratification of buoyancy and vertical wind shear in controlling supercell storm intensity. By varying the depth of the shear layer and height of maximum buoyancy over a range of values for given net magnitude of CAPE and shear, they have been able to demonstrate that the strongest storms are realized when the CAPE and shear are both concentrated at low levels in the sounding. These results highlight the importance of considering the vertical stratification of these forecast parameters when trying to anticipate convective storm intensity.
Weisman began collaborations with Nolan Atkins (Advanced Study Program) and David Dowell (University of Oklahoma) to simulate storms observed during the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX) 1994 and 1995 field seasons. Efforts to date have concentrated on the 16 May 1995 supercell (Atkins), which produced a tornado near Garden City, Kansas, and a series of supercells on 8 June 1995 (Dowell), which produced a series of strong tornadoes near Kellerville, Texas. The goal of these studies is to use the numerical simulations to clarify what aspects of the storm environments were critical for the production of the observed severe weather events. Preliminary results demonstrate the ability of the simulations to replicate some of the characteristics of the storms on these days, and also emphasize the strong sensitivities of these storm characteristics to the range of possible environments that the storms may have experienced.
Weisman began collaborations with Howard Bluestein (University of Oklahoma) on the study of supercell interactions within lines of convective storms. Preliminary simulations demonstrate that the ability to generate and sustain supercell storms within convective lines depends strongly on the orientation of the vertical wind shear profile relative to the line. For vertical wind shear parallel to a north-south oriented line, only the most southern or northern cells are able to develop supercell characteristics. However, for shear profiles oriented at significant angles to the line, a line of supercells may evolve. These simulations are being compared to observed supercell line cases to determine the applicability of these results to forecasting real supercell behavior in such situations.
The Tropical Ocean Global Atmosphere Coupled Ocean-Atmosphere Response Experiment (TOGA COARE) data analysis and modeling effort continues, with a focus on understanding how the environment determines structure and evolution, and in turn how structure influences the surface beneath and large-scale phenomena. The work explicitly on mesoscale convection appears here. The studies of impact that grew out of case studies discussed here appears in the section on the Role of Clouds in Climate.
Margaret LeMone (MMM) and Edward Zipser (Texas A & M University), aided by Michael Dey (former student assistant) and several of Zipser's A & M students, completed a study of the environment of different types of mesoscale convective systems in TOGA COARE, using primary aircraft data below 500 mb, and radiosonde data above. Of the 20 systems observed, nearly all occurring in environments with appreciable vertical shear below a low-level wind maximum were oriented near the normal to the shear beneath the wind maximum, consistent with what would be expected based on work by Rotunno, Joseph Klemp, and Weisman (RKW, 1988). When there was appreciable shear at middle levels but not at lower levels, the convective bands tended to align with the mid-level shear. Many of the slow-moving bands that were not oriented normally to the environmental low-level shear generated band-normal vertical shear near their leading edge, with the air at the lowest levels having the highest wind speeds toward the system.
A complementary modeling study, done by Skamarock and Stanley Trier, used the thermal stratification for 22 February 1993, (demonstrated to successfully produce convection), with two different wind hodographs, those for 22 February and 9 February. Using the same adaptive-grid model that was used for the 22 February simulations, they initialized both cases with a single cell to avoid predetermining line orientation. In both cases, the simulations replicated the observed line orientation and the mode of propagation (continuous for the 22nd, discrete for the 9th).
The environmental soundings for the COARE systems are both warmer and moister than those for the Global Atmospheric Research Programme Atlantic Tropical Experiment (GATE), the differences in surface conditions being consistent with the differences in sea surface temperatures (COARE 1-2 K warmer than GATE). However, a comparison of the COARE convection composite to the fast and slow line composites Barnes and Sieckman (1984) constructed for GATE shows the change in equivalent potential temperature between the surface and mid-levels, and the available buoyancy over the lowest 500 mb, are about the same for the two tropical experiments. Total convective available potential energy (CAPE) is higher for COARE mainly because of the larger depth of the convecting layer (level of free convection to equilibrium level).
Some differences are evident between COARE and GATE convection, and these are currently being explored by examining differences in the soundings. The soundings are available from the National Severe Storms Laboratory (NSSL) Mesoscale Research Division, Boulder, Colorado.
Rick Igau (Texas A & M University), Zipser, Dingying Wei (New Mexico State University), and LeMone have completed a statistical study of convective cores intercepted by NCAR's Electra airplane during TOGA COARE. They found tha t COARE updraft cores were similar to those documented in GATE and elsewhere (consistent with similar low-level available convective energies for the two experiments), but downdraft cores were slightly stronger in COARE. One interesting finding, being investigated further, is that the strongest downdraft cores were apparently positively buoyant.
20 February 1993 late-stage squall-line system. Sharon Lewis and David Jorgensen (both of NSSL), and LeMone evaluated the momentum transport associated with portions of both the leading squall line and one of the trailing convective bands. Particularly the latter matches the prediction of Moncrieff (1992) for a convective band propagating into the flow at all levels; surprisingly this band is propagating into the modified wake air behind the squall line. Attention shifted to an isolated cell which propagated into the wake region from the west, again in violation of conventional wisdom. Lewis tracked the cell back four hours to its inception 140-150 km behind the nascent squall line. Work will continue on a third convective line that formed to the rear of the squall line, and on the boundary-layer features associated with this complex system.
22 February 1993. LeMone, Jorgensen, and Trier completed the observational-modeling study of this squall-line system. The earlier papers emphasized comparison of the model with observations and sensitivity of the numerical model to inclusion of ice physics and surface fluxes and stress; current work involves the influence of this system on the atmospheric boundary layer and oceanic boundary layers and is discussed in the section on the Role of Clouds in Climate. The numerical results also will be used to look at the momentum transports by the whole system and various parts of the system, and to compare with Jorgensen's momentum fluxes for this day.