Table of Contents | Director's Message | Executive Summary | MMM Achievements
Education and Outreach | Community Service | Awards | Publications | People | ASR 2004 Home


MMM Achievements

Lifecycles of Precipitating Weather Systems

Convection Initiation

The initiation of deep moist convection involves many time and space scales, including boundary-layer turbulence, interaction between the land surface and the atmosphere and moisture transport and lifting in the free troposphere. The studies herein are aimed at enhancing our knowledge of how convection begins and are closely linked with efforts in numerical prediction, statistical prediction (RAP), and data assimilation to better predict initiation in short-range (< 12 h) numerical forecasts. Observations from recent field campaigns such as the International H2O Project (IHOP) and the Bow-Echo and MCV Experiment (BAMEX) are keys to understanding convection initiation.

Dynamical Approaches

Stanley Trier completed a study with Fei Chen (RAP) and Kevin Manning that diagnosed simulations of convection initiation using a coupled land-surface mesoscale atmospheric model (MM5). This study isolated the importance of multiple scales of motion in the development of afternoon deep convection in the vicinity of a southern Great Plains dryline. Convection was initiated by fine-scale (1-10 km) circulations confined to ~100-km wide zone of enhanced early afternoon planetary boundary layer (PBL) depth, which itself resulted from horizontal differences on surface sensible heat flux and was reinforced by the ensuing mesoscale solenoidal circulation occurring over the sloped terrain with horizontal contrasts in soil moisture. This study illustrated that an improved forecast of convection initiation using a more detailed and accurate initialization of soil moisture resulted from subtle differences in lower-tropospheric thermodynamic structure even when the overall structure of the dryline moisture gradient remained relatively unaffected.

S. Trier and Christopher Davis have examined the kinematic and thermodynamic structure of mesoscale convective vortices (MCVs) generated by diabatic heating during the Bow-Echo and MCV Experiment (BAMEX) of 2003. MCVs have recently become of interest in convective research since model-produced convective forecasts have exhibited strong sensitivity to details of MCV location and structure. Composite analyses created using BAMEX field observations including dropsonde and mobile soundings combined with in-situ National Weather Service Profiler data constitute the most detailed look at mature MCV structure to date. A dipole pattern of lower tropospheric vertical motion with downshear (upshear) ascent (descent) was diagnosed in 3 of the 5 cases. The pattern of vertical motion was, however, more complicated for strong vortices occurring in weak vertical shear. Both the vertical motion pattern and vortex-induced horizontal advections resulted in large-differences in vertical shear and thermodynamic stability across the MCVs, which, in turn influenced the location of convective redevelopment. It appears that MCVs may act alone or with additional forcing mechanisms (e.g., frontal convergence) to focus convective redevelopment, with regeneration of deep convection requiring a precursor large-scale thermodynamically unstable environment.

Statistical Approaches

S. Trier and David Ahijevych continued collaboration with Cindy Mueller, Dan Megenhardt and Nancy Rehak (each of RAP) on using mesoscale information on thermodynamic stability and convective precipitation forecasts to influence 0-3 h nowcasts of deep convection. A combined interest field based on parameters derived from RUC output was implemented into the RAP auto-nowcaster algorithm to influence areal growth of deep convection. Results from a test period during the 2004 convective season are currently being analyzed. Examination of several convectively active periods from 2004 has established both the utility of 20-km RUC analyses in estimating the lower-tropospheric vertical shear and the strong relationship between the shear and convective behavior in subsequent 0-3 h periods.

Matt Parker (University of Nebraska, Lincoln) and Jason Knievel explored spatial patterns in thunderstorm frequency to probe the widespread belief that population centers (especially those with large populations of meteorologists) are weather “holes”, local minima in frequency of convection. Statistical analyses of data from the WSR-88D network demonstrated that many places thought to be weather holes generally are not. The analyses also revealed other basic characteristics about the shape, timing, and motion of groups of thunderstorms. M. Parker and J. Knievel recommend that their techniques and datasets be used to formulate probabilistic rainfall guidance for parts of the United States where radar coverage is good. A forecaster might then use such guidance to adjust deterministic predictions from numerical models.

Convection Evolution

Once initiated, convection tends to organize to larger scales and may become a mesoscale convective system (MCS). The study of MCS dynamics is a cornerstone of mesoscale meteorology. New areas of emphasis involve use of newly acquired datasets to validate numerical and theoretical models of MCSs, and to probe the complexity of organized convection structures and the affect of these structures on mesoscale and synoptic-scale motions.

Validation of RKW Theory

According to RKW theory (Rotunno et al., 1988) the evolution of a convective line is shaped by the relative values of horizontal vorticity contained in the environmental wind profile versus the cold pool. RKW theory was originally based on numerical simulations of idealized squall lines by the classic Klemp-Wilhelmson (K-W) Model. George Bryan (ASP), Matt Parker (University of Nebraska, Lincoln), and Jason Knievel tested how well simulations by other models bear out the theory. It turns out that RKW theory is not specific to the K-W Model. Indeed, the theory's primary points are consistent with simulations by all three models that were tested: the Advanced Regional Prediction System (ARPS), the Weather Research and Forecasting (WRF) Model, and the Bryan-Fritsch (BF).

G. Bryan, along with David Ahijevych, Christopher Davis, and Morris Weisman, sought observational confirmation of RKW theory by calculating the intensity of convectively generated cold pools within mesoscale convective systems (MCSs) using dropsonde data from the Bow Echo and MCV Experiment (BAMEX). The results support the notion that cold pool intensity plays an important role in MCS structure. In addition, the analyses from this study show that many idealized modeling studies have insufficiently deep and strong cold pools, owing to the neglect of ice processes. Because cold pool intensity is difficult to estimate in real-time, the calculations from the BAMEX dataset are being used to develop a method to determine cold pool intensity from only surface observations.

The North American Monsoon Experiment (NAME)

NAME was designed to study multiscale aspects of the North American monsoon. This international field project employed surface, airborne, and ship-based observations to document the continental-scale circulation that transports vast amounts of tropical moisture into western Mexico and the Southwest United States during the summer months. These instruments captured several phases of the monsoon, including its onset, peak, and termination. D. Ahijevych and James Done served as S-Pol radar scientists and implemented the most appropriate scanning strategy based on current scientific objectives and local conditions near Mazatlan, Mexico. This activity was also a valuable cross-divisional collaboration with ATD scientists, technicians and RAP software engineers.

Orographic Effects

The presence of orography induces numerous local enhancements and reductions of rainfall, and these effects are generally less predictable than thought, especially when the atmosphere is unstable to deep, moist convection. In addition orography induces numerous downstream effects on rainfall that must be quantified and properly represented in models.

Warm Season Orographic Influences

Click image for larger version.

Figure 25. 12 hour accumulated precipitation (contours) for northwest flow over the Rocky Mountains (grayscale). Contours are at 2, 4, 8, 16 and 32 mm.

Click image for larger version.

Figure 26. (top) Precipitation accumulation versus wind direction. (bottom) Wavelet correlations of terrain elevation versus wind direction.

Andrew Crook and Donna Tucker (University of Kansas) have completed a study of the flow over heated terrain. In Part I of this study, an analytical model was developed of the dry response to flow over heated topography. The individual responses to orographic and thermal forcing are first considered and then combined response is examined. It is shown that if the terrain is stretched preferentially in one direction, then the leeside lifting will be maximized if the large scale flow is along that direction. In Part II of this study, these results are applied to the complex topography of the Rocky Mountains. From Figures 25 and 26, it is evident that the precipitation accumulation is maximized for flow from the northwest and minimized for flow from the southwest. A wavelet analysis (Figure 25) applied to the terrain of the Rocky Mountains shows more topographic features that are stretched preferentially in the NW/SE direction than in the SW/NE direction. These results are important for understanding statistics of orographic rainfall and are potentially useful for validation of weather and climate prediction models.

Taking a larger-scale perspective on orographic, warm season rainfall, D. Ahijevych analyzed warm-season rainfall with spatally-averaged diurnal composites of precipitation frequency. The sharp coherence of propagating rainfall extending from the Central Massif of the Rockies through the Midwest was documented in Carbone et al. (2002). At first, the 2002 warm season coherence of rainfall looked very different from the previous six years, but after accounting for the east-west position of the Rocky Mountains, the characteristic timing and location of the diurnal precipitation peak reappeared. This illustrated the remote effect of orography on precipitation frequency hundreds of kilometers downstream. Plus, it highlighted two-dimensional aspects of the Rocky Mountain-Great Plains atmospheric circulation.

Cool Season Orographic Influences

Click image for larger version.

Figure 27.
Cloud water mixing ratio q_c where white indicates regions of q_c < 0.01 g/Kg, light grey, 0.01g/Kg <q_c<0.1g/Kg, medium gray, 0.1g/Kg<q_c< 0.5g/Kg, and dark gray, q_c>0.5g/Kg. Environmental flow is from left to right with initial q_c = 0.05 g/Kg. Panel a) shows that if the mountain height is small enough, a saturated flow can be maintained everywhere given sufficient initial cloud water. Panel b) shows that for tall mountains the atmosphere upwind of the mountain is maintained in a saturated state and transitions to an unsaturated downslope flow on the lee side (which has characteristics associated with downslope windstorms.) Panel c) shows that for mountains of intermediate height, the solutions have the unexpected feature of an upwind-propagating disturbance which has the effect of desaturating the atmosphere above the mountain..

Although a fairly common atmospheric condition in orographic-rain scenarios, there is relatively little known about moist neutral flows over a ridge from theory and modelling. To shed light on this important flow regime, M. Miglietta (Il Consiglio Nazionale delle Ricerche (CNR), Italy) and R. Rotunno have conducted numerical simulations of the orographic-flow modification occurring for a two-dimensional moist nearly neutral flow over a ridge. If an initially saturated moist neutral flow were to remain everywhere saturated as it flows over an obstacle (Figure 27a), then the expected solution would be the linear solution because then the condition for linearity (hill height less than the ambient wind velocity/static stability) is always met. However for tall mountains (Figures 27b and 27c), the numerical solutions indicate the development of areas of unsaturated air, with correspondingly larger values of local static stability. This internal switching from small to large values of static stability is an inherent nonlinearity which has far-reaching consequences for understanding the orographic-flow modification in this regime. M. Miglietta and R. Rotunno find that the numerical solutions fall into three basic categories as summarized in the figure.

M. Ralph (NOAA/ETL), P. Neiman (NOAA/ETL) and R. Rotunno used dropsonde observations to document the mean vertical profiles of kinematic and thermodynamic conditions in the pre-cold frontal low-level jet (LLJ) region of extratropical cyclones over the Eastern Pacific Ocean. This LLJ region is responsible not only for the majority of heavy rainfall induced by orography when such storms strike the coast (Figure 28a), but also for almost all meridional water vapor transport at midlatitudes. The data were collected from NOAA’s P-3 aircraft in ten storms during the CALJET experiment of 1998 (during El Niño) and in seven storms during the PACJET experiment of 2001 (during La Niña). Although the composite winds, temperatures and water vapor mixing ratios in 2001 differed markedly from 1998, Ralph, Neiman and Rotunno found that the moist static stability remained near zero from the surface up to 2.8-3.0 km altitude for both seasons (Figure 28b). Hence, they concluded that orographic precipitation enhancement is favored in this sector of the storm, regardless of the phase of ENSO. Although the total meridional water vapor transport in 1998 was about twice that in 2001, the vertical structure of the transport was nearly invariant.

Click image for larger version.

Figure 28. Conceptual representation focusing on conditions in the pre-cold-frontal low-level jet (LLJ) region of a land-falling extratropical cyclone over the northeastern Pacific Ocean. (a) Plan-view schematic showing the relative positions of a LLJ and trailing polar cold front. The average position of the 17 dropsondes used in this study is shown with a star (~500 km offshore of San Francisco), and the Cazadero microphysics site is marked with a bold white dot. The points A and A’ along the LLJ provide the approximate endpoints for the cross section in panel b. (b) Cross-section schematic along the pre-cold-frontal LLJ (i.e., along AA’ in panel a) highlighting the offshore vertical structure of wind speed, moist static stability, and along-river moisture flux at the location of the altitude scale. Schematic orographic clouds and precipitation are shown, with the spacing between the rain streaks proportional to rain intensity.

Basic Measures of Orographic Precipitation

Click image for a larger version.

Figure 29. Setup for idealized orographic precipitation studies.

Idar Barstad and Ronald Smith (both Yale University) and P. Smolarkiewicz quantified the atmospheric water cycle for idealized orographic flows to gain insight into orographic precipitation. They identified three measures of efficiency for characterizing orographic precipitating systems. The global measure, referred to as the drying ratio (DR), which determines the fraction of water vapor influx that turns into precipitation; while the interior measure, condensation ratio (CR) and the precipitation efficiency (PE), describe the conversion processes. In order to assess how these measures depend on external parameters (e.g., mountain width and surface temperature) two distinct mathematical models were used: i) a fully nonlinear numerical model that conserves water substance (EULAG), and ii) a linear orographic-precipitation model that has linearized micro-physics and airflow dynamics. A series of simulations of 2D flows past a Gaussian-shaped ridge, using a warm rain scheme in the nonlinear model, revealed that DR increases for decreasing temperatures and increasing mountain width. The idealized setup for the numerical experiments is shown in Figure 29. The re-evaporation efficiency of condensated water (PE) is strongest at low temperatures. Despite discrepancies between the two models, both show increased re-evaporation as the mountain width decreases.

Long-time-scale dynamics of mesoscale convective systems

Following the initiation of convection and its growth to a mesoscale convective system, longer-time-scale organization of convection is possible. This organization may take the form of continental-scale coherence of rainfall, involving multiple MCSs. Tropical cyclones are also an example of multi-day coherence of convection. In many problems, the key to understanding, and the crtical issue for representation in regional and large-scale models, is the upscale growth of vorticity and the emergence of balanced flow structures. Work described herein naturally links with global and regional climate modeling as well as short-range (1-2 day) precipitation prediction.

Continental Convection

Click image for larger version.

Figure 30. Time-latitude (Hovmoller) plots of radar derived rainfall rates averaged over a)110-105 W and b) 95-100 W longitude bands for July 20-31, 1998. The daily occurrence of convection over the higher terrain of the Rocky Mountains from Mexico to the Canadian border (Fig. 3a) is evident, but only a small fraction is long-lived and propagates into the central plains, arriving 8-10 hours later (Fig. 3b).

During the midsummer there often exists a well-defined corridor of precipitation episodes across the central U.S. Such corridors are, at best, weakly related to synoptic scale forcings. The corridor location typically persists 3-7 days with significant variability on the inter-seasonal timescale (Figure 30). A corridor will experience excessive cumulative rainfall while nearby regions may be well below normal. Understanding the nature and forcing mechanisms leading to the corridors has important implications for QPF and the
quantification of precipitation regimes.

Using the U.S. national composited radar data and RUC (Rapid Update Cycle) model analysis for the 1998-2002 warm seasons (July-August), John Tuttle examined the relationship of precipitation corridors to various forcing mechanisms over the central U.S. While the corridors show the expected correlation to CAPE and shear.

Click image for larger version.

Figure 31. Diurnally averaged a) rainfall b) meridional component of the 900 mb winds c) CAPE and low-level wind shear (between 600 and 900 mb) in time-latitude plots over the domain shown in Fig. 3b and d) rainfall in time-longitude format. During the averaging process the rain and RUC data were shifted in the north-south direction to 40 deg latitude using the centroid of the rain data as a reference. This allows for a presentation of the average RUC fields with respect to the convection.

A surprising find is the strong relatioship to the nocturnal low-level jet (Figure 31). Even in the five-year averaged data, the precipitation corridor is strongly associated with the exit region of the LLJ (the region of maximum moisture convergence). The precipitation maximum at 6:00-10:00 UTC is a combination of convective systems propagating into the area from the east slopes of the Rockies and convection forming locally in response to the enhanced convergence of the LLJ. On days when the LLJ is weak or non-existent, there are virtually no propagating convective systems. For days having a stronger LLJ, the total precipitation is stronger and a greater percentage of the precipitation develops locally.

A large-domain, convection-resolving simulations of one such corridor lasting 7 days over the Central U.S. was performed by S. Trier, C. Davis and Sherrie Frederick using the WRF model. The observed convection propagated over large longitudinal differences (~1000 km) but was confined to a relatively narrow latitudinal corridor, where deep tropospheric moisture and at least moderate vertical shear were collocated. High resolution simulations with explicit deep convection were able to accurately replicate characteristics of the observed coherent rainfall episodes including frequency, longevity, propagation speed and diurnal frequency of precipitation (Figure 32).

Click image for larger version.

Figure 32. Comparison of normalized diurnal frequency diagrams of hourly, latitudinally averaged (30-48 N) precipitation greater than 0.1 mm/h during 3-10 July 2003 derived from (a) stage IV rainfall observations, (b) the 4-km WRF simulation that used explicit deep convection, and (c) the 22-km WRF simulation that used a cumulus parameterization.

A broad spectrum of phase speeds were found with both the observed and simulated coherent rainfall episodes in the common large-scale environment suggesting that the propagation is likely to have an internally generated component. Simulations performed on a 22-km grid exhibited fewer rain streaks with a poor representation of the observed distribution of phase speeds. These latter results agree with those obtained by C. Liu, using the MM5 model.

Arlene Laing examined the propagation characteristics of African rainfall as part of a multi-year study of the lifecycles of precipitating systems globally. The study is aimed at understanding rainfall patterns in different regions of the world, and ultimately at validating and improving the prediction of precipitation in weather and climate models. Following Hovmoller techniques developed by John Tuttle for radar data, full resolution Meteosat infrared images were used to identify patterns of deep convection. These analyses showed that precipitating convection in Africa exhibit coherent patterns or episodes that are similar the continental United States (US) and East Asia (Figure 33). The mean zonal span and duration for May to August 1999 were 847km and 17.9h. Most episodes had phase speeds of 10 - 20 ms-1, generally near the maximum speed of the African Easterly Jet around 600 hPa. Convection showed orographic initiation and occasional marked coherence in diurnally averaged rainfall frequency (Figure 34). These early findings support the notion that that precipitation cycles may be predictable beyond one or two days.

Click image for larger version.

Figure 33. Hovmoller diagrams of cloud top brightness temperature <213K for the period of 1-15 July 1999.

Click image for larger version.

Figure 34. Mean diurnal cycle for (a) 1-15 May, (b) 16-30 June 1999. Times are UTC. The more rapid movement is the westward march of the solar heating maximum, while the streaks moving around 15 ms-1 are manifested by multiple convection systems.

Tropical Cyclones

C. Davis and Lance Bosart (University at Albany SUNY) continued their study of the tropical transition of Atlantic disturbances of extratropical origin. Tropical transition exhibits large interannual variability and tends to produce more storms in years when Cape Verde storms (formation from easterly waves) are less common. Tropical transition can occur close to land and make landfall before appropriate coastal preparations are made (e.g. Hurricane Alex of 2004). C. Davis and L. Bosart have examined the formation of Hurricane Humberto (2001), a storm forming in the subtropics in response to an upper-tropospheric cold low of extratropical origin. The Humberto simulations conducted with MM5 began before lower tropospheric cyclonic vorticity was evident, but produced a tropical storm in 96 h using a fully explicit treatment of convection (Figure 35). The development was episodic in the model, featuring numerous bursts of convection, initially triggered downshear from the upper low, and concomitant weakening of the vertical wind shear as relative vorticity increased in a stepwise fashion. A companion simulation, without the upper low, yielded a much broader region of cyclonic vorticity and no well-defined surface cyclone (Figure 36).

Click image to view the larger animation.

Figure 35. Animation of wind at 850 hPa and cloud-top temperature from MM5 simulation of the genesis of Humberto. Time increment is 0.5 h. Winds are plotted every 15 grid points

Click image for larger version.

Figure 36. 850 hPa wind and temperature at 66 h from (a) control simulation and (b) simulation with upper-tropospheric cold low removed.

Click image for larger version.

Figure. 37. Elfers Parkway at Elfers River, Pasco County, Florida, 9 September 2004. b) Storm total precipitation for 4-7 September 2004.

Heavy precipitation and associated flooding are among the deadly aspects of hurricanes at landfall. As part of a continuing study of the geographical influences of hurricane-related floods, A. Laing and Graham Tobin (University of South Florida) conducted a three-day field survey of flooding in the aftermath of Hurricane Frances (Figure 37). West-central Florida experienced widespread flash floods, urban floods, and several moderate to major river floods. Among the factors that influenced the greater flood impact were

  1. antecedent precipitation from two tropical cyclones and above normal precipitation during July and early August;
  2. Frances stalled along the east coast and moved very slowly across the peninsula;
  3. Frances was approximately twice the size of the previous system, thereby drawing moisture from both the Atlantic and the Gulf of Mexico; and (4) a second landfall along the Florida panhandle brought additional rain bands across the state.

The interaction of inertia-gravity waves and balanced flows

Improved prediction of precipitation, particularly in the short range and at the mesoscale, depends in part on predicting, and understanding, the occurrence of inertia-gravity waves in mesoscale flows. For example, the jet stream in the mid-latitudes is known from observations to be an important source of inertia-gravity waves (IGW), but the mechanisms responsible for such generation of IGW are poorly understood, let alone quantified.

In collaboration with David Muraki (Simon Fraser University, Burnaby, Canada) and C. Snyder, Riwal Plougonven (formerly ASP/MMM postdoc, now University of St. Andrews) is investigating how baroclinic instabilities can generate IGW. The idealized case of baroclinic instabilities on a constant vertical shear allows the analytical quantification of the amplitude of the gravity waves as a function of flow parameters, as shown in Plougonven et al. (2004). This fundamental study contributes to the understanding of the relation between balance and gravity waves in a shear flow.

D. Muraki and C. Snyder also developed exact solutions for a vortex dipole in the surface quasigeostrophic equations (Muraki and Snyder 2004). This dipole is a steadily propagating, balanced, coherent structure that decays rapidly with height. It represents an ideal starting point for examining the interaction of balanced flows with IGW, since any emitted waves will have clear signatures aloft owing to the decay of the dipole.

Cloud microphysics and precipitation

Microphysics studies, both laboratory and field studies, are critical for understanding processes leading to precipitation in clouds and ultimately the integrated effects of phase changes of water on the dynamics of precipitation systems.

Charles Knight completed a study first radar echoes using the STEPS data, emphasizing particularly new information that can be supplied by the differential reflectivity (ZDR) in the early stages of precipitation. It is not uncommon for there to be temporary, positive ZDR columns to 6 or 7 km MSL, indicating mm-sized water drops being elevated to these levels before freezing, just as the first precipitation echo is forming. The ZDR column is at the upshear side of the updraft, and generally separated from what has been conventionally called the first precipitation echo, which forms downshear of the updraft. The freezing of these drops may be important in the spread of the ice phase in these early cumulus clouds.

C. Knight started and completed a study of ice spikes. These form in the early stages of partly-confined freezing of liquid water, by the volume change squeezing water out the last opening in the surface. Sometimes this water does not spread out over the ice surface, but grows into a hollow, water-filled ice spike that grows only at its tip as the liquid water is extruded. These form on freezing drops in the atmosphere, and have been implicated in ideas of ice multiplication in clouds, and also form on ice growing in ice cube trays if pure water is used. The spikes are usually composed of either two or three crystals, and their crystal orientations are a critical factor in determining whether spikes form or not. This has been basic research, leading to a much better understanding of this particular phenomenon (which had not been studied previously in this way), but with no direct, immediate application.

Next Topic: Surface-Atmosphere Interactions


Table of Contents | Director's Message | Executive Summary | MMM Achievements
Education and Outreach | Community Service | Awards | Publications | People | ASR 2004 Home

National Center for Atmospheric Research University Corporation for Atmospheric Research National Science Foundation Annual Scientific Report - Home Atmospheric Chemistry Division Advanced Studies Program Atmospheric Chemistry Division Climate and Global Dynamics Division Environmental and Societal Impacts Group High Altitude Observatory Mesoscale & Microscale Meteorological Division Research Applications Program National Center for Atmospheric Research Scientific Computing Division