A wake is traditionally defined as the whole region of nonzero vorticity on the downstream side of a body in an otherwise uniform stream. In a stratified fluid, the motions and density surfaces downstream of an obstacle become primarily horizontal; the vertical component of the vorticity associated with the horizontal motion, coexisting with the stable vertical density stratification, implies that there is potential vorticity (PV) in the wake.

Recent work by Richard Rotunno, Vanda Grubišic (visitor, Yale University), and Piotr Smolarkiewicz demonstrated that dissipation aloft, associated with a breaking mountain wave over an isolated peak, produces a dipole in PV downstream; the dipolar vertical vorticity of the wake is associated with the PV dipole. Left unanswered by this argument is the question, "Where does the vorticity come from?" To answer this question, they analyzed a weakly nonlinear model for PV production and wake formation in the case of a small-amplitude mountain, and carried out numerical simulations pertaining to the strongly nonlinear large-amplitude case. The simple model indicates that even with dissipation in the system, the vertical vorticity of the wake arises through the tilting of baroclinically generated horizontal vorticity by the dissipating mountain wave. Their analysis shows that there need not be any direct effect of friction in the vorticity equation to produce a wake; dissipation enters indirectly through its effect on the tilting term. Analysis of numerical simulations of the large-amplitude case shows that the conclusions from the weakly nonlinear model continue to hold in the strongly nonlinear regime.

Joseph Prusa (Iowa State University), Smolarkiewicz and Rolando Garcia (ACD) continued their collaborative study of gravity wave activity in the upper mesosphere and lower thermosphere. New results from full three-dimensional simulations continue to provide strong support for the saturation hypothesis of wavebreaking. Wave activity analyses show initial propagation and late time development that is completely consistent with earlier two-dimensional results. In particular, once vigorous wavebreaking has been established, wave activity shows an exponential drop-off with scale height equal to the density scale height of the atmosphere. This is predicted by wave saturation arguments. Analysis of the three-dimensional simulations also revealed that equipartition is still roughly satisfied even in the wavebreaking region. As should be expected, in the linear wavefields outside of the breaking region, wave energy density is evenly divided (precisely) between kinetic and potential types. In the wavebreaking region, the kinetic energy is consistently greater than the potential energy by about 40 percent, with a typical variation of 20 percent. Thus either field can be used to roughly estimate the total energy density to within about 20 percent. An approximate equipartition holds.

Richard Carbone, John Tuttle (joint appointment with RAP), Grubišic, William Cooper, and Wen-Chau Lee (ATD) completed a study of Hawaiian rainfall forcing that permits a definitive evaluation of breeze and blocking hypotheses advanced in the literature. These hypotheses were viewed in a competitive light since the Clark-Hall Model experiments by Smolarkiewicz, Roy Rasmussen, and Terry Clark in the 1980s. The current study depicts Hovmoller-type rainfall distributions by time and distance from the windward shore for 19 days during HaRP and stratifies these by island Froude number (Fr). The results are strongly confirmatory of both blocking influences and diurnally induced thermal influences on the production of precipitation. When integrated over the entire windward island domain (up to 70 km offshore), Fr is shown to explain a very large fraction of daily rainfall variance as predicted by the Clark-Hall Model experiments. On the other hand, diurnal forcing is shown to strongly influence rainfall location and proportion for a given phase of the diurnal cycle. An unexpected result of the study was the strong inference that oceanic rains, beyond the influence of the island, occur almost exclusively at night and thus further enhance nocturnal rainfall maxima over the island.

Grubišic, Smolarkiewicz, and Carbone continued the modeling study of Hawaiian circulations with the goal of quantifying sensible and latent thermal contributions to rainband forcing. Numerical predictions, by the Eulerian/semi-Lagrangian anelastic model, of a flow reversal initiation and the separation-line position over a one-and-a-half diurnal cycle showed a good agreement with observations, particularly when thermal forcing is purely radiative. The effect of evaporative cooling due to precipitation from orographic clouds remains to be addressed with higher resolution simulations.

Jian-Jian Wang (University of Illinois), Robert Rauber (University of Illinois), and Carbone conducted case studies of offshore rainbands for five cases observed in HaRP. The emphasis of these studies was the convective scale and mesoscale dynamical structure, principally deduced from dual-Doppler and other radar diagnostic techniques. Initiation and amplification of rainbands was observed to occur regularly along four distinct types of convergence lines. The strongest convection occurs along the diurnally-forced flow separation line that partitions island-induced westerly flow from the tradewinds. Moderate amplification of oceanic rainfall occurs 20-30 km upstream of the flow separation line along sharply deformative discontinuities in the upstream easterlies. Episodic strong convection was forced offshore by a specific terrain feature under E, ESE flow associated with Cape Kumakai. Finally, banded convection appears to be triggered by nearby mature convection, presumably by a cold pool process.

Using data from the HaRP, Carbone, in collaboration with Tuttle, Grubišic, Lee (ATD), and Cooper, completed a study of the morphology of Hawaiian rainfall under tradewind conditions. The principal objectives of the study were to determine the dependence of rainfall amount and distribution on the environment and the relative roles of dynamical flow blocking and diurnally induced effects (land-seabreeze forcing) in rainfall production.

Results showed a strong diurnal cycle of rainfall in which most rainband amplifications occurred over a mesoscale convergence line that separates the easterly tradewinds from an island induced westerly flow during the night hours. The westerly flow is driven by cooperation between the thermal (radiative and evaporative cooling from orographic rainfall) and dynamical forcings of the island. The flow separation line (FSL) was typically located 10-20 km offshore during the predawn hours. Total rainfall was found to be strongly correlated with the island Froude (Fr) number U/Nh, where U is the tradewind speed, N the Brunt-Vaisala frequency and h the characteristic height of the island. Average total domain rainfall increased by a factor of 2.4 when Fr increased from 0.23 to 0.35 and, in the coastal areas, rainfall increased by a factor of 3.5. For most of the diurnal cycle, the position of the FSL was relatively insensitive to Fr. The major exception to this was in the hours around sunrise when the FSL and the associated rain were predominantly offshore under ordinary Fr conditions. When Fr was elevated, the FSL was nearly 10 km closer to shore and a much larger fraction of rainfall was deposited on the windward coastal areas. These features can be seen in figures showing in Hovmoller format the radar-derived rainfall as functions of time and distance offshore for ordinary and elevated Fr number days.

A secondary upstream divergence zone was discovered that often maintained a position about 30 km upstream from the FSL. This may be a consequence of blocking dynamics and resulted in a doubling of the oceanic rainfall between 20 and 45 km offshore. Beyond 50 km, offshore background oceanic rainfall was observed with no island-induced effects apparent. The various rainfall regimes are easily seen in a figure showing the integrated rainfall amounts as functions of distance offshore and Fr. These findings, while specific to Hawaii, are applicable to other regions of the tropics where there are mountainous islands and the Fr number concepts are applicable to orographic precipitation with small conditional instability.

Clark, Janice Coen, Hsiao-Ming Hsu (RAP), Teddie Keller (RAP) and William Hall concluded work on the Tropical Storm Russ case study. This is believed to be the first numerical simulation of terrain-induced mechanical turbulence. The separation bubble was captured and successfully analyzed using a new analysis technique, the Bernoulli energy form. This type of work differs distinctly from channel flow in that it is more in the vein of "terrain-effects" rather than surface roughness, which characterizes channel flow. ( See also under Three-Dimensional Stereo Visualizations).

With the availability of the new 24-processor Gigaword J9, Ouray, in SCD, Clark, Hall and Robert Kerr continued their study of the 9 December 1992 incident over Evergreen where a DC-8 lost an engine in clear air turbulence. The simulation used five domains, the innermost using 240x240x74 mesh points with 200-meter horizontal resolution. See High Performance Computing and Communications for more details on how this was implemented. Earlier lower resolution simulations showed both the onset and demise of gravity wave breaking as well as the first signs of turbulence in the wave-breaking region. The higher resolution simulations showed more of the origin of turbulence, including small-scale vortex streaks at 10 km MSL coincident with the position of the DC-8 incident. The most intense vortex structures were aligned with the mean flow and were in clusters with a crosswind spacing of about 14 km. In cooperation with NOAA/ETL, Clark et al. are comparing simulation data directly with LIDAR observations of the day that the incident occurred. LIDAR data have confirmed the existence of these horizontal vortex streaks, and a 14-km cloud spacing appears in satellite data. The origin of these streaks is being investigated. (See also under Three-dimensional Stereo Visualizations).

Associated with this program, MMM hosted the annual meeting of the United Airlines Safety Panel in August 1997 to demonstrate NCAR capabilities in modeling and visualizing situations in mountain windstorms that are hazardous to aviation. Future interactions with United and Boeing are being planned in collaboration with Larry Radke (ATD).

Roelof Bruintjes, Clark, and Hall wrote a paper summarizing the 1995 Arizona Program. Two main objectives of this program were to improve the understanding of the processes that determine the spatial and temporal distribution of precipitation and to access the potential for artificially enhancing winter precipitation in a mountainous region of central Arizona. Of special interest to the program was the interaction of topographically-induced gravity waves with the ambient upslope flow where, it has been suggested, these waves may serve to augment the upslope-forced precipitation that falls onto the Mogollon Rim.

Janice Coen and Bruintjes conducted modeling studies with the Clark-Hall Model to provide guidance in the design and execution of a precipitation enhancement program initiated by the Korea Water Resources Corporation. The modeling work emphasized improving the understanding of the factors determining the spatial and temporal distribution of precipitation at the surface; specifically, the Andong Reservoir, in east central South Korea. On 8 June 1989, a case study of heavy rain (over 7 cm was measured at some stations) was performed.

Here, Clark-Hall Model simulations were initialized and the boundary conditions updated using a 24-hour forecast of the MM5 numerical model to represent the background environment and changing large-scale conditions as a cyclone over the Yellow Sea moved eastward over the mountainous Korean peninsula. With the Clark-Hall Model, three nested domains focussed from the large-scale to cloud-scale resolution, centered over Andong Reservoir and its source region. The simulations captured the broad synoptic-scale uplift ahead of the low-pressure system, a broad band of clouds comprised of layer clouds at the leading edge, and a line of deep, strong convective elements at the rear. The study examined the vastly different mechanisms of precipitation formation in the stratiform region and convective bands, where the stratiform region was characterized by a typical condensation-coalescence process below the melting level, supplemented by the melting of snow and ice crystals as they fell through the melting level. The convective elements were characterized by an efficient combination of a condensation-coalescence process combined with an intense ice process, which likely led to the high precipitation rates that were observed at some stations. Modeled and rain gauge-measured precipitation rates, accumulations, and timing compared extremely well, suggesting that in addition to capturing the large-scale forces shaping precipitation, the model represented the smaller-scale mountain and valley flows that concentrated precipitation in certain areas, making the model a useful tool for the future for understanding the causes of enhanced precipitation in locations with complicated terrain.

Three-dimensional stereo animated visualizations are now an integral part of the analysis of MMM terrain-following simulations. Three flows, described elsewhere in the current report (Tropical Storm Russ, forest fires, and mountain windstorms), were visualized by Don Middleton (SCD) using Vis5D. Vis5D can simultaneously represent isosurfaces, volume renderings, and two-dimensional slices of several fields. Middleton made extensive modifications so that the state-of-the-art graphics workstations in SCD can handle the large flow fields generated by Clark, Hall, and Kerr. They also used the package on their local graphics workstations for individual frames and for helping select the best sequences, fields, and perspectives for the full three-dimensional animations. Important new physical insights were gained from these new graphics tools. For the case of Tropical Storm Russ, the role of the topography in generating vorticity and turbulence was demonstrated. For mountain windstorms, the 14-km instability became apparent and the new thin, vortex tubes were seen.

A Vis5D stereo animation of the simulated case of dynamic fingering described in a paper by Clark et al. showed how the fire line bowed into a conical shape (the "convective fingering" phenomenon observed by firefighters) and developed a pair of counter-rotating vertical vortices at the fire line front that touched down within the fire line and broke it up, accompanied by fingers of hot gas shooting out from the fire line front ("dynamic fingering"), providing a revealing view into fire line dynamics. (See the visualizations done by Don Middleton on the SCD Visualization Gallery web page, including A Numerical Model of a Forest Fire)

Kevin Manning and Christopher Davis (joint appointment with RAP) examined the numerical simulation of coastal fog using the PSU-NCAR Mesoscale Model (MM5). The efforts this year have been largely directed toward testing a PBL scheme, developed at the Pennsylvania State University by George Gayno and Nelson Seaman, within MM5. The Gayno-Seaman PBL scheme was adapted to interact more closely with the radiation parameterization in MM5. Idealized tests were performed with a 2-dimensional version of MM5, and a single case was simulated with the full 3-dimensional MM5. Preliminary results suggest some ability of MM5 to simulate the spatial distribution of coastal fog, although the Gayno-Seaman scheme shows a tendency to predict somewhat less cloud water at the surface than the standard PBL scheme.

Davis and Simon Low-Nam, in collaboration with Clifford Mass (University of Washington) and Ola Persson (NOAA/ETL) continued to examine the dynamics of observed southerly surges trapped against the topography along the West Coast of the U.S. The simulations previously reported on were improved (from the standpoint of agreement with available observations). In the case of 10-11 June 1994, southerlies propagate northward a few hundred kilometers. The southerlies systematically occur along the east side of an extensive, north-south elongated anomaly of potential vorticity that appears to be produced by dissipation in the model. The ability of the surge to propagate northward depends upon the preconditioning of the larger-scale environment by warm advection and adiabatic descent in easterlies along the coast, which pushes the coastal baroclinic zone and associated northerly jet offshore, leaving little adverse pressure gradient along the coast.

Simulations of the Catalina Eddy event of June 1988 suggest a similar dynamical mechanism. A strong, positive PV anomaly is shed by the coastal topography and rolls up into a cyclonic eddy downstream from the mountains. A notable diurnal signal appears in the generation of anomalous PV with a maximum during late evening, suggesting that the occurrence of these coastal phenomena may favor certain times of day (i.e., nighttime or early morning).

Scott Braun (ASP), Rotunno, and Joseph Klemp have examined the role of surface friction in modifying cold fronts as they make landfall in regions of steep coastal orography by means of idealized simulations. The simulations show that surface friction primarily modifies the flow over the orography by increasing the upstream flow deceleration, and reducing the magnitude of the barrier jet. Hence, with surface friction, upstream frontal-motion retardation is strong and upstream frontogenesis is enhanced. They also found that there is an enhancement of the frontal updraft at the coast which results from the combined effects of the orography and the surface roughness change, however, large parcel displacements are due primarily to the orographic forcing. The simulations indicate that along-coast winds in the coastal zone during frontal passage are approximately determined by a superposition of the southerly barrier jet and the frontal jets.

Under funding provided by the ONR, Rotunno, Klemp, and William Skamarock examined the dynamics of spring and summer season coastally-trapped disturbances (CTDs) observed along the U.S. west coast. Simulations using a two-dimensional shallow-water equation model and three-dimensional primitive equation model show that an imposed offshore flow (observed in the climatological synoptic-scale CTD evolution) advects the marine layer offshore, weakens the prevailing northerly flow in the marine layer and lowers the pressure at the coast. The nearly balanced marine-layer flow around this low pressure raises the marine layer to the south of the low as the flow encounters the coastal mountains.

This region of elevated marine layer begins progressing northward as a Kelvin wave and later steepens into a bore or gravity current, these features being identified as the CTD. Many observed features associated with CTDs are captured in the idealized numerical simulations, including the formation of a mesoscale pressure trough offshore and deep southerlies in the CTD at the coast. They also found that including stability above the inversion, while allowing upward energy radiation from a CTD, does not lead to wave decay; steady solutions are produced for wave amplitudes sufficient to produce significant nonlinear steepening, although the propagating disturbances are less likely to form bores (discontinuities). The upper-level stability also supports a topographically trapped Rossby wave that can act to propagate signals northward and can affect the amplitude and structure of the CTD itself (see figure).

NCAR | UCAR | NSF | NCAR FY96 ASR | NCAR FY97 ASR