Life Cycle of Precipitating Weather Systems

Stanley Trier and Christopher Davis collaborated with investigators within NCAR/RAP (Cynthia Mueller, Daniel Megenhardt, and James Wilson) to examine the utility of mesoscale kinematic and thermodynamic information from RUC analyses and forecasts for short-range (0-3 h) forecasts of convective initiation and evolution. Testing on five widespread convective outbreaks, each of which caused significant disruption to United States air traffic operations, indicated that both absolute values and hourly trends of derived thermodynamic parameters, including convective available potential energy (CAPE) and convective inhibition (CIN) in multiple lower-tropospheric layers, were useful in determining when the local onset of deep convection would occur. They were also useful in identifying where, within preexisting areas of widespread convection, convection might subsequently weaken or decay. Future work is planned to examine these parameters, along with other thermodynamic and kinematic parameters (e.g., the vertical wind shear) as potential predictors for local initiation, areal growth, decay and movement of deep convection over a wider range of cases. Algorithms, based on the statistics from such studies, will be developed to aid in the automated short-range forecasting of these aforementioned aspects of deep convection.

a) Convective episodes

Studies on warm season precipitation "episodes" by Richard Carbone, John Tuttle, David Ahijevych, Christopher Davis, Stanley Trier and L. Jay Miller have progressed from their initial efforts to characterize the two-dimensional climatology. "Episodes" are defined as time/space clusters of heavy precipitation that often result from sequences of organized convection such as squall lines, mesoscale convective systems and mesoscale convective complexes. Understanding the dynamics of convective episodes is crucial for developing improved, non-local representations of convection in numerical models. The climatologies also provide the basis for statistical/dynamical prediction of warm season rainfall, perhaps leading to realistic probabilistic representations of precipitation in forecast models.

Three avenues of research and related service activities were undertaken during the past year:

John Tuttle and Richard Carbone are nearing completion of an investigation of a long-lived convective system that persisted for two days over the central U.S. on 14-15 July, 1998. The event featured an abrupt change in its orientation and propagation vector that occurred about midway through its life (see Fig 2). The convection initiated over the higher terrain of southwestern Montana (Figure 2-left image) and became loosely organized into a N-S line perpendicular to the W-E oriented, low level shear vector (Figure 2-right image). A cold pool formed and raced ahead of the storm resulting in a succession of discrete propagation events across Montana and North Dakota. Upon entering Minnesota the convection intensified and assumed an E-W orientation in response to the increased moisture, the strong southerly flow and the N-S oriented shear vector. There were no indications of any strong fronts that could have accounted for the abrupt changes or the longevity of the system. Following its reorientation, the line moved slowly southward initially, but in the highly unstable and favorably sheared environment, a strong rear-inflow jet developed and the system bowed southward into southern Minnesota and Iowa. The storm then decayed rapidly as it moved into dryer, more stable air due to subsidence aloft. It was concluded that favorable cold pool - low level wind shear interactions and changes in the lower tropospheric shear vector orientation can explain the life cycle transformation.

L. Jay Miller is studying a Mesoscale Convective System (MCS) that persisted for more than 2 days, 21-23 June 1998. A sequence of intensification, decay, and regeneration led to this relatively long-lived mesoscale convective event.

Trier and Davis completed an observational study of a serial mesoscale convective system (MCS) on 27-29 May 1998, that possessed a persistent mesoscale convectively generated vortex (MCV). Through novel trajectory diagnostics applied to Rapid Update Cycle (RUC) analysis output, they demonstrated that balanced lifting, resulting from the interaction of the MCV with the ambient vertical shear, contributed in large part to the thermodynamic destabilization that allowed the redevelopment of deep convection within the multi-day MCS/MCV event. The portion of the vortex located within the lower troposphere intensified during nocturnal episodes of organized MCS activity. This appeared to aid in the horizontal transport of conditionally unstable air toward the location deep convection. In these ways, the MCV was documented to be a crucial link between relatively quiescent periods characterized by balanced flow and intermittent periods of organized deep convection that produced flooding rains.

Using the MM5 model, Davis and Trier simulated the first full diurnal cycle of the May 1998 MCV/MCS. The simulation, initialized with a RUC analysis and nested to 1.5-km horizontal grid spacing over the area of convection, correctly reoriented convection from a north-south band to and east-west band overnight (see Figure 4) in response to northward transport of warm, conditionally unstable air within the nocturnal low-level jet. As in the RUC analyses, balanced vertical motion was found to contribute substantially to mesoscale lifting and thermodynamic destabilization, which localized the convection. Horizontal transport of moist, unstable air into the nocturnal convection was significantly modulated by the MCV. In contrast to other studies of MCVs, Davis and Trier found that the re-intensification of the MCV at night began in the lower troposphere with the formation of a line-end vortex on the northern end of the north-south oriented convective line. Intensification of the mid-tropospheric vortex followed in response to the development of a stratiform precipitation region (see Figure 5). Melting of hydrometeors appeared to contribute substantially to the development of the mid-level circulation.

David Ahijevych constructed Hovmoller diagrams for four warm seasons. He also computed radar-derived rainfall as a function of universal time and longitude for the 1997-2000 warm seasons and compared this to estimates obtained using different data sets.

Initial work on precipitation echo frequency was led by John Tuttle. Figure 6 is an animation of the July 1998 diurnal cycle of precipitation radar echo frequency. It reveals the monthly averaged genesis of convective systems over the western cordillera, propagation and regeneration of convection eastward and southward, and interaction with the Gulf of Mexico sea breeze initiated convection over the interior of the southeastern U.S.

(iii) Website Database and Data Access

Upon acquiring a new web server, Ahijevych developed the Episodes Project web page found at http://locust.mmm.ucar.edu/episodes. He also developed web pages which allow researchers to peruse weather images dating back to May 1998, http://locust.mmm.ucar.edu/case-selection.

In earlier work, Richard Rotunno, Morris Weisman, and Joseph Klemp formulated a theory suggesting that squall line structure, strength and longevity was most sensitive to the magnitude of the component of low-level (0-3 km AGL) vertical wind shear perpendicular to squall line orientation. An "optimal" state was proposed whereby the deepest leading edge lifting and most effective convective re-triggering occurred when these circulations were in near balance. This state was based on the relative strength of the circulation associated with the storm-generated cold pool and the circulation associated with the ambient shear. Following this work, many subsequent studies have brought into question the relevance of such an optimal state to observed squall lines. They note the existence of strong, long-lived systems in sub-optimal conditions and they raise the question of the potential role of deeper-layer shears in promoting system strength and longevity in such situations. In an attempt to clarify these issues, Weisman and Rotunno have completed and analyzed an extensive set of simulations. They used both a simplified two-dimensional stream-function model and a full two-dimensional and three-dimensional cloud model, and they have been able to re-confirm the primary role of the low-level shear in controlling squall line structure and strength. They further clarify that a wider range of environments other than strictly "optimal" support significant squall lines in the simulations. This is also evident from observations.

Collaborations have continued with Jeff Trapp (visitor, NSSL) and Nolan Atkins (Lyndon State College) on the observation and simulation of tornadic circulations within quasi-linear convective systems such as squall lines and bow echoes. A set of idealized simulations have been completed that reproduce many of the characteristics of such systems, including the tendency for surface mesocyclones to develop north of the apex of the bow for environments of moderate to strong low-level environmental vertical wind shear. Analyses show that the initial source of these low-level circulations is the downward tilting of the horizontal vorticity associated with the cold pool-updraft interface, forced by the downdrafts of strong, short-lived convective cells near the leading edge of the systems. The actions of Coriolis forcing then promotes the strengthening and upscale growth of the cyclonic member of the tilted vortex couplet, producing a significant mesocyclone at the surface that could support the development of a tornado. This process appears quite distinct from that associated with supercell tornadoes, whereby a deep, quasi-steady, dynamically forced rotating updraft usually precedes the development of the tornado. Attempts are also underway to simulate observed systems to compare with the idealized cases.

c) Field experiments

Davis continued to lead the coordination of the Bow Echo and MCV Experiment (BAMEX), now scheduled for May 20 - July 6, 2003. BAMEX is a collaboration among PIs at NCAR, NSSL, NWS and several universities (UCLA, Texas A&M, Penn State, CSU and U of Alabama). The goals of this experiment are: (1) to obtain kinematic and thermodynamic documentation of the development of system-scale circulation features behind the leading convective line in maturing and decaying MCSs; (2) to understand mechanisms of convective regeneration near MCVs and the dynamics of MCV intensification that appear critical for multi-day events; (3) to understand the cause of damaging surface winds in bow echoes; and (4) to assess the predictability of long-lived MCSs and their effects on weather. The planned observing facilities include two Doppler P-3s, a dropsonde aircraft and a movable ground-based observing system consisting of two Doppler radars, wind profiler, acoustic sounder radiometer, mesonet and soundings. Please refer to http://www.mmm.ucar.edu/bamex/science.html for more information.

a) Development of Hurricane Diana (1984)

Christopher Davis and Lance Bosart (University at Albany, SUNY) continued their study of the formation of Hurricane Diana (1984) by examining the behavior of numerous sensitivity simulations. Development was dependent on a pre-existing upper-tropospheric trough-ridge couplet that focussed vertical motion, grid-resolved condensational heating, and lower- tropospheric potential vorticity anomalies that merged to form a tropical storm. Simulations with cumulus schemes that allowed more grid-scale precipitation on the 9-km grid exhibit unrealistic grid-scale overturning and slower intensification, primarily due to production of cyclonic vorticity anomalies at large radii. Use of an innermost nest with 3-km grid spacing, without a cumulus scheme, generally improved the intensity prediction. Storm track depended primarily on synoptic-scale structure at upper levels. Cumulus schemes that allowed more grid-scale overturning enhanced the anticyclonic outflow aloft. The outflow deformed the tropopause, building an anticyclone poleward of the storm and facilitating cut-off low formation equatorward of the storm. Using PV attribution, it was shown that these upper-level changes were responsible for an enhanced easterly steering flow and more westward storm track.

Jordan Powers and Davis extended the analysis and simulation of Diana using regional, cloud-resolving simulations from the MM5 model. The simulation consisted of a single domain of 1.2 km grid spacing and dimensions of 1000 x 1060 x 37, run on 552 processors of SCD’s IBM SP. For comparison purposes, 3-km and 9-km grid simulations have been performed. Simulations using 1.2 km and 3 km grid spacing showed markedly similar overall storm evolution, whereas the 9-km grid produced a storm with an unrealistically extensive circulation and no tight inner core. The higher-resolution simulations show Diana to develop, consistent with observation, in three distinct phases: initial MCS activity, quiescence and persistence of the incipient vortex, and convective regeneration and tropical cyclone formation.

b) Simulation of Hurricane Danny (1997)

Ying-Hwa (Bill) Kuo, Wei Wang and Qinghong Zhang (ASP postdoctoral fellow) have performed a high-resolution numerical simulation of Hurricane Danny (1997) over a four-day period, from its genesis stage to its landfall. The simulation began at 0000 UTC 16 July 1997 when only a weak surface low was present over northern Gulf of Mexico. The PSU/NCAR MM5 model with triply-nested (81/27/9 km) grids was able to successfully simulate the development of a small tropical cyclone 72 h into the simulation, and its subsequent landfall over the Gulf coast. Subsequent numerical experiments at 3-km and 1-km grid resolution successfully captured interesting mesoscale structures of the storm, including the concentric eyewall, the eyewall replacement cycle and the trochoidal oscillations in storm track, as observed by the ground-based Doppler radars. Additional numerical experiments with 3-km and 1-km MM5 indicated that the simulation of the genesis of Danny was very sensitive to the choice of precipitation physics and planetary boundary parameterizations, and to the initial condition. The use of explicit cloud parameterization at cloud-resolving resolution (at least 3 km) is essential in simulating a realistic storm structure. Simulations with 9-km grid resolution using convective parameterization cannot properly reproduce the detailed storm structure as observed by the radars.

Richard Rotunno and R. Ferretti (University of L'Aquila, Italy) continued their analysis and simulation of Mesoscale Alpine Programme (MAP) cases. Although the large-scale flow was similar, important differences in mesoscale atmospheric structure made the difference between moderately intense rain in IOP2B, and relatively light rain in IOP8 of MAP. Rotunno and Ferretti have done a side-by-side analysis of these two cases with respect to precipitation, thermodynamic structure and wind. MM5 simulations of these cases agree well with the available data and, hence, provide a valuable interpretive tool. Analysis of the large-scale dynamics show that there was, in both cases, a moist tongue of southerly flow moving from west to east of the MAP area (northwestn Alps). The most important difference between the cases was the presence of a cold stable air mass in the Po Valley in IOP8, which persisted through the period in which the large-scale moist tongue was progressing eastward. The latter cold air mass prevented the most humid air from reaching the MAP area. Another important difference between the two cases occurred during the eastward passage of the cold front (the western boundary of the moist tongue) which is generally retarded at lower levels with respect to higher levels. In IOP2B this orographically induced differential advection of cold air produced strong conditional instability, and consequently, an additional episode of convective rain in the MAP area. In IOP8 the prefrontal air in the Po Valley was so cold that differential advection only reduced the already large static stability and, subsequently, there was no additional period of convective rain.

The Severe Thunderstorm Electrification and Precipitation Experiment (STEPS) was held in Eastern Colorado and Western Kansas in May through July 2000 with the goal of better explaining the relationship between kinematics, precipitation production and electrical characteristics of convective storms on the High Plains. Morris Weisman and L. Jay Miller focused on the analysis and simulation of supercell storms observed on 29 June and 5 July, which exhibited differing precipitation characteristics. Miller and Sarah Tessendorf (SOARS and CSU graduate student) have completed preliminary dual-Doppler analyses of the high-precipitation tornadic storm observed on 29 June. These analyses reveal many of the characteristic kinematic features associated with supercell storms, including a strong, quasi-steady rotating updraft and associated bounded weak echo regions (BWER) during the storm's mature phase. But they also reveal a more complicated multiple updraft configuration during other periods in the storms lifetime. Miller's preliminary dual-Doppler analyses of a more nearly low-precipitation supercell storm observed on 5 July reveal a more unicellular rotating updraft, reaching magnitudes of over 60 m/s. Weisman's initial simulations of the 5 July storm have been successful at replicating many of the observed storm characteristics, offering much hope that simulations in conjunction with the observations will provide useful insights into the precipitation mechanisms associated with the STEPS storms. Such analyses will then be used to improve the microphysical representations within cloud scale simulations of such events.

In collaboration with Andrzej Wierzbicki (University of South Alabama, Mobile) and Richard Laursen (Boston University), Charles Knight completed the development and application of a new etching technique that reveals the crystal surface orientations at which biological adsorb to ice, and whether or not the adsorbate is engulfed within the ice during growth. (These "antifreeze" molecules are peptides that adsorb to ice and prevent ice growth from supercooled water by a kinetic mechanism that is, as yet, poorly understood.) A major new finding of the first application of this technique has been that there is a lot of such adsorption to ice such that the adsorbate is not engulfed in the ice as it grows, but presumably is pushed ahead, along with the moving interface. Many synthesized peptides have been characterized in terms of their effectiveness as antifreezes, but the results have been ambiguous because there are several potential reasons for the variability. This new method can be used to remove some of that ambiguity.

Knight, in collaboration with K. Rider (Colorado School of Mines) completed a study of the crystallization of a clathrate hydrate from its pure melt. The interest in and motivation of the study has been to explain the high variability of crystal habit, which appears to be neither an impurity effect nor a result of crystal imperfections, though it is difficult to prove the latter with certainty. It has thus been a fundamental difficulty in crystal growth theory, because there are no other known reasons for such a variability. The hydrate is a cubic, completely faceted material, with only (111) faces, which, at small sizes and over a substantial range of supercooling, may grow either slowly as octahedra, or much faster as thin plates, or thin needles. The growth manifestations and the ways in which the different forms are initiated are complicated, but it is argued that the growth mechanism is probably dominated by layer nucleation at the face corners. When the faces are small, asymmetry of the corners may lead to very different growth rates on adjacent but crystallographically equivalent faces, causing the plate and needle habits. The early portion of this work was done in collaboration with a group researching clathrate hydrates (especially methane hydrate) for the petroleum industry at the Colorado School of Mines.