Show Summary Details

Page of

Subscriber: null; date: 21 July 2018

# Formation and Development of Convective Storms

## Summary and Keywords

Cumulus clouds are pervasive on earth, and play important roles in the transfer of energy through the atmosphere. Under certain conditions, shallow, nonprecipitating cumuli may grow vertically to occupy a significant depth of the troposphere, and subsequently may evolve into convective storms.

The qualifier “convective” implies that the storms have vertical accelerations that are driven primarily, though not exclusively, by buoyancy over a deep layer. Such buoyancy in the atmosphere arises from local density variations relative to some base state density; the base state is typically idealized as a horizontal average over a large area, which is also considered the environment. Quantifications of atmospheric buoyancy are typically expressed in terms of temperature and humidity, and allow for an assessment of the likelihood that convective clouds will form or initiate. Convection initiation is intimately linked to existence of a mechanism by which air is vertically lifted to realize this buoyancy and thus accelerations. Weather fronts and orography are the canonical lifting mechanisms.

As modulated by an ambient or environmental distribution of temperature, humidity, and wind, weather fronts also facilitate the transition of convective clouds into storms with locally heavy rain, lightning, and other possible hazards. For example, in an environment characterized by winds that are weak and change little with distance above the ground, the storms tend to be short lived and benign. The structure of the vertical drafts and other internal storm processes under weak wind shear—i.e., a small change in the horizontal wind over some vertical distance—are distinct relative to those when the environmental wind shear is strong. In particular, strong wind shear in combination with large buoyancy favors the development of squall lines and supercells, both of which are highly coherent storm types. Besides having durations that may exceed a few hours, both of these storm types tend to be particularly hazardous: squall lines are most apt to generate swaths of damaging “straight-line” winds, and supercells spawn the most intense tornadoes and are responsible for the largest hail. Methods used to predict convective-storm hazards capitalize on this knowledge of storm formation and development.

# Introduction

Cumulus clouds are pervasive on earth, and play important roles in the energetics of earth’s atmosphere. Under certain conditions relatively shallow, nonprecipitating cumulus clouds may experience deep vertical growth; more restrictive conditions may allow such deep clouds to evolve into convective storms. Convective storms have many practical benefits, such as providing necessary rainfall for agriculture. But they also can be hazardous, as when they generate excessive precipitation leading to flooding, and cause hail, lightning, tornadoes, and nontornadic “straight-line” winds; in the tropics, collections of interacting storms can evolve into tropical cyclones.

A convective storm has vertical accelerations driven primarily, though not exclusively, by buoyancy over a deep layer. In the atmosphere, buoyancy arises from local density variations relative to some base-state density; the base state is typically idealized as a horizontal average over a large area, which is also considered the environment.1 The existence of vertical accelerations also means that the forces due to gravity and the vertical pressure gradient (which arises from the divergence of the stress tensor) are locally unbalanced. Convective clouds and storms effectively seek to restore this hydrostatic balance.

The qualifier “deep” implies that the cloud or storm occupies a significant fraction of the troposphere, which nominally has a depth of ~10 km; accordingly, the storm is essentially constrained by the tropopause. Internal storm motions—upward and downward currents or drafts—vertically transfer heat, moisture, and momentum over the storm depth. As just alluded to, this vertical mixing is a response to, and seeks to mitigate, hydrostatic imbalance. Particularly large imbalances and thus buoyancy yield particularly intense storms that can penetrate into the stratosphere. Cross-tropopause transport of water vapor and chemical species occurs within these “overshooting” storm tops.

The vertical and horizontal extent of convective storms are ultimately described in terms of cloud and precipitation particles, which allows for them to be visualized by weather satellite and radar. Cloud particles form on microscopic bits of dust or other cloud condensation nuclei when water vapor in the atmosphere is cooled to the point of condensation (or directly to the point of freezing); such cooling is facilitated by updrafts. In the lower levels of the atmosphere where the air is warmer than 0°C, the cloud particles are water droplets with radii of ~10 microns. At heights where the temperature is below 0°C, the cloud droplets begin to freeze, although a mixture of ice particles and liquid drops will still exist, especially when inside the rapidly rising air of strong updrafts. At heights where the temperature is below −40°C, the cloud particles are almost exclusively ice crystals with radii ≳ 100 microns. This is especially the case at storm top, within the characteristic anvil-shaped cirrus clouds. Because of the strong winds at this (tropopause) height in the atmosphere, anvil cirrus often extends far downstream of the actively growing and most intense parts of the storm.

Extensive areas of anvil cirrus reflect incoming solar radiation back into space and thereby prevent the shortwave radiative heating of the atmosphere and ground beneath the cirrus. Water phase changes within the cloud result in another form of diabatic heating and cooling: Consider that when cloud droplets form, the condensation of vapor into liquid releases a tremendous amount of heat that reinforces the buoyancy driven updrafts; in a similar vein, evaporation of liquid drops reinforces downdrafts. Thus the storm is influenced by but also influences the thermodynamic properties of its environment.

Such environmental modification by mesoscale systems of convective storms, which include tropical cyclones, can occur over length scales of ~100’s of kilometers and temporal scales of ~10’s of hours. Individual convective clouds and storms have length scales of ~1 to several 10’s of kilometers and time scales of hours, and the convective processes contributing to these clouds and storms are even smaller in scale. Owing to their relatively small scales, convective clouds and storms tend not to be explicitly represented in the current generation of global climate models and some weather prediction models. Consequently, their effects (e.g., mixing and heating) must be indirectly (and imperfectly) represented through “parameterizations”; an analogous statement can be made regarding the need to parameterize the formation and evolution of precipitation. One implication is that the parameterized amount and location of convective-storm-generated rainfall is a major source of error and uncertainty in weather and climate model predictions. This also causes errors in temperature (imagine how temperature near the ground would be different if beneath versus outside of a thunderstorm). Both errors depend on the storm intensity and areal coverage, which in turn depend on the storm formation and structure.

# Advancements Toward the Current State of the Science

The current understanding of the formation and structure of convective storms owes to a combination of theory, observations, and numerical modeling. While many of the gains in such understanding have been small and incremental, large gains have been facilitated by certain technological advancements, and certain nontechnological occurrences. For example, weather-related accidents associated with the emerging aviation industry during the 1930s and 1940s helped motivate efforts to deploy subsynoptic meteorological networks to observe thunderstorms, or at least the “noise” they induced in otherwise smooth fields of surface temperature, pressure, and winds (Fujita, 1986). The high-resolution data collected during the Thunderstorm Project in Florida in 1946, and then in Ohio in 1947, yielded the basis for the accepted conceptual model of (single-cell) thunderstorm evolution and structure, as well as the length, time, and velocity scales associated with thunderstorms.

In addition to instrumented aircraft, the Thunderstorm Project made use of meteorological stations at the ground, and balloon-borne radiosondes; a coarse network of these “surface” and “upper air” data were already available for operational weather forecasting, however. Indeed, efforts to predict the onset of thunderstorms exploited these data and existing theory (Corfidi, 1999), and led to the knowledge that certain environments (such as shown in Figure 1) and weather patterns were particularly favorable for intense convective storms and associated hazards like tornadoes (e.g., Fawbush & Miller, 1954).

The Thunderstorm Project also made early use of radar.2 At the time, this technology had been reserved mostly for military endeavors, but during the Thunderstorm Project, radar was used to identify the parts of storms that were thought too hazardous for in situ sampling by aircraft. This recognition of weather applications of radar-remote sensing eventually led to the first-generation radar network in the United States. A range of possible “precipitation formulations” (Fujita, 1986) was deduced from these conventional radars during this era (1950s–1960s), including what is now known as the “supercell” (Browning, 1964). Doppler weather radars developed in subsequent years provided further insight into internal thunderstorm motions, as did the organized storm-intercept teams who sought to confirm the radar-observed signatures associated with phenomena like tornadoes, damaging winds, and hail (Bluestein, 1999).

The birth of radar meteorology coincided roughly with the birth of numerical weather prediction (i.e., the use of digital computers to predict the future state of the atmosphere using mathematical equations). Further availability of computational resources, and further development of the mathematical equations and their discrete approximations, have allowed for simulations of the behaviors of individual convective clouds. The first of these numerical models represented convective clouds in a single dimension only; the current generation of cloud models are time dependent, fully three-dimensional, have become progressively more sophisticated, and tend to be freely available as open source codes (e.g., Bryan & Fritsch, 2002; Skamarock et al., 2008)

Numerical cloud models provide a virtual laboratory for hypothesis testing. In typical idealized model applications, the modeled atmosphere is assumed to have a vertical stratification in temperature, moisture, and winds, but otherwise is initially homogeneous over horizontal planes. Clouds are “initiated” in this initially balanced atmosphere by imposing, for example, a small perturbation in temperature. Experiments within this virtual laboratory seek to answer questions such as how the structure and intensity of the initiated cloud would change given some increase in the initial winds (e.g., Weisman & Klemp, 1982). Answers to this and related questions are revealed in the next section.

Highly controlled exploration of cause and effect is not well afforded in the real atmosphere, yet the basis for the explorations often stem from observed relationships, especially those derived from large datasets. The routine collection of Doppler radar data, reports of tornado and hail occurrence, and vertical profiles of temperature, moisture, and winds using radiosondes, have, for example, led to estimates of tornado frequency from supercells, and the average meteorological conditions that most likely support their formation (e.g., Smith et al., 2012; Thompson et al., 2012). Observational data collected during organized field campaigns, using mobile weather radars, mobile radiosonde systems, etc., have tended to enhance the resolution characteristics and thus detail in these datasets.

High-impact convective-storm events sampled using mobile or fixed platforms tend to undergo considerable analysis, motivated in part by the premise that conclusions drawn from the analyses can be generalized to other events exhibiting extreme behaviors. The validity of this premise is tested as new events are observed as well as simulated using nonidealized models like the Weather Research and Forecasting (WRF) model. Toward that end, process-based analyses from these real-data simulations are invaluable.

# Development and Structure of Convective Storms

As noted, convective clouds are driven largely by buoyancy, and quantifications of atmospheric buoyancy allow for an assessment of the likelihood that convective clouds will form. A vertical integration of buoyancy quantifies the potential energy available to a convective cloud or storm. Known as convective available potential energy (CAPE), this vertically integrated quantity also provides a thermodynamically based estimate of the maximum intensity of the updraft and thus storm.

Buoyancy and thus CAPE are evaluated using vertical profiles of temperature and humidity, which are key atmospheric variables that are measured twice daily (nominally) using balloon-borne radiosondes. An example of such a vertical profile (which, because of balloon drift during its ascent, is meant to be the temperature and humidity in a vertical surface or plane rather than along a vertical line) is shown in Figure 1.

Click to view larger

Figure 1. Atmospheric sounding in the environment of an intense convective storm. Blue line shows the temperature profile, and black line shows the moisture profile (via the dew point temperature). In this skew T-log p diagram, the ordinate is the logarithm of pressure (in hPa), and the abscissa is temperature (°C). A corresponding height (km) is included for reference.

In this atmospheric sounding, the temperature over some vertical depth is used to graphically assess the likelihood that massless air particles, or parcels, will be positively buoyant and thus accelerate vertically over this depth. The sounding diagram facilitates a comparison between the environmental temperature assumed to have been measured using a radiosonde and the temperature of an idealized, hypothetical air parcel: In essence, if the air parcel is warmer than its environment at some height, the air parcel will rise (accelerate upward); if cooler than its environment, the air parcel will sink (accelerate downward); and if the same temperature of its environment, the air parcel will be static (experience no vertical accelerations). The so-called parcel theory thus used to illustrate parcel tendencies has numerous shortcomings and limitations (e.g., Doswell & Markowski, 2004; Schultz et al., 2000), but it is sufficient for the purposes herein.

Especially within continental midlatitude regions, an air parcel may experience relatively warmer environmental air (and thus be negatively buoyant) at heights within the lowest ~1–3 km above the ground, but then be positively buoyant over the remainder of the atmosphere above this boundary layer. Often contributing to the negative buoyancy—and to the vertically integrated negative buoyancy quantified as convective inhibition (CIN)—are elevated layers of hot, dry air that are vertically well mixed. In the United States, elevated mixed layers originate over high-altitude desert plateaus (e.g., the Mexican Altiplano) and then are transported east and northward to overlie relatively cooler, moister air at lower altitudes. Elevated mixed layers and other layers of negative buoyancy over the lower troposphere create atmospheric “lids” or “caps” because they inhibit air from rising into the free atmosphere, in analogy to how a lid atop a pan of boiling water prevents the steam from escaping the pan.

Some mechanism by which parcels are lifted through the negatively buoyant layers is thus required for parcels to realize the positive buoyancy and contribute to deep convective cloud formation. One of the canonical lifting mechanisms is a topographic barrier: Consider how air flowing toward a two-dimensional mountain is forced to ascend upon encountering this mountain. The other is a weather front, which lifts parcels through a vertical circulation associated with the front’s horizontal temperature gradient . The dryline, a front-like boundary that is defined in terms of a horizontal gradient in moisture, has an analogous vertical circulation, and often aids in the initiation of deep convective clouds within the Great Plains region of the United States. Numerous other convection-initiating mechanisms have been identified, and are known to work in tandem (Weckwerth & Parsons, 2006).

Within the archetypal environment of an intense convective storm over the continental midlatitudes (Figure 1), the typically late-afternoon initiation can be fairly explosive (e.g., from shallow cumulus cloud to cumulonimbus on the order of an hour); within tropical oceanic regions, such a transition from shallow to deep convection tends to be more gradual. The subsequent evolution of the nascent convective storm depends on the details of its environment, which includes not only temperature and moisture but also the variation of the horizontal wind vector (speed and direction) with distance (or height) above the ground. This variation is termed vertical wind shear, which formally is the magnitude of the vertical derivative of the horizontal wind vector. Vertical wind shear over roughly the lowest half of the troposphere has been found to be particularly relevant, and is often computed as the magnitude of the vector difference between the wind at 6 km above ground level and the wind at ground level (i.e., $|V→6km−V→0km|=S06$.) This quantification of vertical wind shear is also known as “bulk shear” or “deep layer shear,” and is complementary to storm-relative helicity, which is an integrated quantity that explicitly accounts for the vertical changes in environmental wind speed and direction (e.g., see Rasmussen & Blanchard, 1998).

Consider the case in which the environmental wind is weak (e.g., ≲ 10 m s−1) and exhibits little change in speed and direction with height over the lowest half of the troposphere. Low S06 is common at low latitudes, but also exists at midlatitudes especially during summer. Low S06 and sufficiently high CAPE (e.g., ≳ 500–1000 J kg−1) support a convective storm that is most apt to be short-lived (duration of ~ 1 hr) and possess a single- or unicellular morphology. Borrowing from the biological sciences, this “cell” characterization stems from the storm’s appearance on a weather-radar display as a discrete entity with a single core of enhanced radar reflectivity factor (hereinafter, reflectivity). The reflectivity core represents a region of larger and/or more numerous hydrometeors such as raindrops, thus also representing a region of enhanced precipitation. It indirectly reveals information about the two convective components that help generate the precipitation, namely an updraft and downdraft.

Indeed, as depicted by Figure 2, the updraft is a localized current of rising air and a manifestation of the aforementioned positive buoyancy.

Click to view larger

Figure 2. Idealized evolution of a unicellular convective storm, based on a numerical model simulation. Vectors are airflow in a vertical cross section, and blue shading shows cloud outline: (a) early stage, (b) mature stage, and (c) decay stage. Thin dashed line indicates the freezing level. In (b), red shading shows the core of precipitation in downdraft. In (c), gray shading shows the cold pool.

A typical unicellular updraft speed is ~10 m s−1, within a several-kilometer wide core. Hydrometeor formation is initiated within an updraft as cloud droplets inwardly diffuse available water vapor, collide, and merge with other cloud droplets, and/or freeze while ascending to heights where the temperature is significantly below 0°C. Once the hydrometeors attain a size and thus mass that can no longer be suspended by the updraft, they begin to fall relative to the updraft, and, in this weakly sheared case, through it. The drag of the falling hydrometeors contributes to a downdraft, which is a localized current of sinking air (Figure 2); the speed and horizontal dimension of the downdraft tend to be comparable to those of the updraft. Further hydrometeor evolution, like the evaporation of falling raindrops and the melting of snow, graupel, and hail contributes to a negative buoyancy that reinforces the downdraft. As it must, the relatively cold downdraft terminates at the ground, forming a “pool” of air that spreads laterally (Figure 2). The leading edge of this horizontally expanding cold pool is often termed the gust front because, as with the relatively larger weather fronts, its passage ushers in cooler (and gustier) air. The cold-pool gust front also has a vertical circulation that serves as an air-parcel lifting mechanism: here, however, the lifting connects ambient parcels to an ongoing convective storm, thus fueling the storm with warm, buoyant air and promoting its sustenance. Accordingly, once—or if—a gust front has expanded horizontally too far from its updraft or downdraft source, the ambient or environmental air can no longer be connected to the ongoing storm, and the storm will dissipate.

Gust front expansion and loss of connection with updraft occurs relatively quickly when environmental wind shear is negligible: The cycle of unicellular storm growth through dissipation occurs over ~1 hr. Now imagine how this evolution is changed with the introduction of wind shear: Initially, the updraft is not vertically erect but rather is tilted in the direction of the wind shear vector. This is because the cloud droplets are differentially transported (advected) with height by the varying wind. But the additional ramification of the tilt is that the hydrometeors growing inside the tilted updraft will also be differentially transported and eventually will fall out of the updraft rather than through it. The resultant downdraft then feeds a cold pool that subsequently will lift parcels to replenish the existing updraft, or at least aid in the formation of a new updraft in proximity to the existing one.

A symbiotic relationship between updraft, downdraft, and cold pool in the presence of nonnegligible environmental wind shear has been used to explain (e.g., by Rotunno et al., 1988) the longevity of mesoscale convective systems (MCSs), which are storms comprised of multiple cells that are highly interactive. By virtue of the “mesoscale” qualifier, MCSs are by definition relatively large, with a major-axis (horizontal) dimension of at least ~100 km. Smaller versions of MCSs are simply known as multicell storms. In both versions, cell interactivity is manifest largely by the aggregation of the cold pools contributed by individual cells. This aggregate cold pool plays a dominant role in the storm’s dynamics and thus in its evolution and intensity.

The archetypal mature MCS resulting from the cold-pool dynamics is a quasi-two-dimensional structure that again exhibits a symbiotic relationship between the convective components, except now on a somewhat larger scale (Figure 3).

Click to view larger

Figure 3. Schematic of a squall-line type MCS, in (low-level) plain view and vertical cross-section (X-X’). Bold black line shows the outline of the squall line as viewed by radar. Grey-shaded region represents the updraft and associated reflectivity core. Blue-shaded region in vertical cross section shows the cold pool. Adapted from Smith et al. (2009).

In this MCS representation as a squall line, which has also become known as a quasi-linear convective system (QLCS), environmental air parcels lifted at the cold-pool gust front contribute to a vertically erect convective updraft that culminates in more gradual ascending air. This front-to-rear ascending air current deposits hydrometeors into a rear-to-front descending air current that feeds the cold pool and terminates at the gust front. The rear-to-front current itself is partly a reflection of the environmental flow but also a result of an acceleration by a horizontal gradient in pressure associated with internal storm processes, including the cold pool. Thus, this current simultaneously contributes to and is influenced by the cold pool. Because it also terminates at the gust front, the rear-to-front current enhances the parcel lifting at the gust front and thereby enhances the updraft. In turn, the enhanced updraft generates more/larger hydrometeors that further strengthen and cool the descent in the rear-to-front current. As long as this coupling is maintained, the MCS will persist. A MCS lifetime of several hours is typical, which is several hours longer than the lifetime of an individual cell in a weakly-sheared environment. A particularly long-lived and intense breed of an MCS is a derecho (Corfidi et al., 2016). The derecho event of May 8, 2009, for example, had a duration of more than fifteen hours and caused wind damage in five US states (Coniglio et al., 2011).

A simple environmental wind profile that supports existence of a long-lived, eastward-moving MCS is characterized by low-level easterly winds that become westerly with height. This unidirectional wind profile traces out a straight line when graphically presented using a hodograph (Figure 4a).3

Click to view larger

Figure 4. (a) Unidirectional and (b) curved wind profile, as represented in a hodograph plot. The bold black line is the hodograph curve, the light gray arrow is the surface wind vector, and the dark gray arrow is the wind vector at a height of 6 km.

With S06 ≳ 20 m s−1, both a straight-line hodograph as well as a curved hodograph (as shown in Figure 4b) can promote the occurrence of a supercell, which is a special or extraordinary type of a unicellular storm. As implied by the prefix, a supercell is larger, longer lived, and more intense than an ordinary unicell. The intensity and longevity in particular are intimately related to the existence of rotation about the vertical axis of the storm’s updraft. This rotation is the basis for the unique supercell dynamics, and accordingly is a salient characteristic of supercells.

The vertical rotation, which is a deep column of swirling, rising air (and thus is often visualized as a “barber’s pole”), originates from the horizontal rotation present in the environmental wind shear. Vorticity, which formally is the curl of the velocity vector, is used to quantify this horizontal and vertical rotation. Of specific interest here is the means by which horizontal components of vorticity are reoriented into the vertical. The process is known as tilting, and occurs within the updraft (and downdraft).4 In the case of the straight-line hodograph with an east-to-west environmental wind, the initial vorticity vector is horizontal and points toward the north (as can be verified using the right-hand-rule). An idealized circular updraft with a central core will tilt this horizontal vorticity such that vertical vorticity with a counterclockwise (clockwise) sense is generated in the southern (northern) half of the circular updraft (Figure 5).5

Click to view larger

Figure 5. Schematic of an early stage of supercell development. Gray ribbons show the presence of counterrotating vortices on the updraft flank, and yellow arrows indicate rotationally induced upward accelerations. The vortices are generated by tilting of environmental horizontal vorticity, as represented by thin black lines; the sense of the horizontal vorticity is indicated by the green ribbons. Storm splitting is initiated during this stage, as implied by the blue arrow; the red ribbon shows inflow and updraft prior to splitting. Adapted from Klemp (1987).

In meteorological vernacular (as applied in the Northern Hemisphere), cyclonic (anticyclonic) vertical vorticity now resides in the southern (northern) flank of the updraft; the cyclonic vorticity will ultimately become a mesocyclone, and the anticyclonic vorticity, a mesoanticyclone.

The schematic in Figure 5 represents an early stage of supercell development, as would occur shortly after a typically late-afternoon initiation of deep convection from an aforementioned lifting mechanism. The presence of counterrotating vortices on the updraft flank is most prominent at the middle levels of the storm, or ~3–7 km above the ground; the vortices (i.e., mesocyclone/mesoanticylone) have diameters of a few kilometers, and swirling windspeeds of ~10 m s−1. The pressure inside the vortex core is lower than outside, regardless of the sense of rotation. And because the vortices have intensities that increase with height to midlevels, the drop in core pressure also increases with height. This means that the force due to the vertical gradient in (vortex-reduced, nonhydrostatic) pressure is directed upward beneath the midlevel vortices, thereby accelerating air upward into the midlevel vortices (Figure 5). New updraft growth on the flanks of the existing updraft is dynamically promoted by the midlevel vortices, ultimately leading to a literal split of the early stage storm into two mirror-image storms.

When viewed by weather radar, the process just described is akin to biological cell division as viewed under a microscope. Subsequent storm-cell splits are possible but less common. Instead, the mirror image storms persist along opposing tracks, with the “southern” (“northern”) cell continuing to deviate southward (northward) from its initially easterly motion. The southern deviation—or rightward motion (relative to the initial motion)—is both forced by and reinforces the cyclonic updraft rotation; an analogous statement applies to the leftward motion of the anticyclonically rotating northern cell.

Such rotationally induced deviant motion is one supercell hallmark; another is the presence of updraft rotation itself. As seen in weather radar images, a “hook echo” is yet another salient characteristic, and owes to interactions between the precipitation and the wind field of the mesocyclone. Specifically, hydrometeors falling forward of the updraft, within the forward-flank downdraft (FFD), are transported by the mesocyclonic wind field to the rear of the storm to contribute to a rear-flank downdraft (RFD) (Figure 6).

Click to view larger

Figure 6. Plan view of idealized supercell. Red (blue) shading shows regions of low-level updrafts (downdrafts); FFD indicates core of forward-flank downdraft; and RFD indicates core of rear-flank downdraft. Streamlines are of (storm-relative) near-surface flow. Bold contour shows 40 dBZ reflectivity contour. Dashed bold red and green line is location of midlevel updraft and mesocyclone core. Adapted from Lemon and Doswell (1979).

The direct influence of environmental vertical wind shear in terms of differential advection of precipitation, as well as the indirect influence via the mesocyclone, are therefore realized in the development of supercell structure.

Supercells also have a symbiotic relationship between updraft, downdraft, and cold pool as facilitated by the environmental wind shear but, unlike MCSs, in a highly three-dimensional sense (e.g., Figure 6). Parcel lifting occurs along the rear-flank gust front and forward-flank gust front, helping to sustain the updraft. Both gust fronts represent zones between the warm air of the environment and the cool air generated by melting and evaporating precipitation. The forward-flank gust front has the additional distinction of being oriented parallel to the low-level environmental winds. Bearing in mind the simple buoyancy concept that “warm air rises, cold air sinks,” air parcels encountering this horizontal baroclinic zone acquire vector vorticity that is parallel both to the baroclinic zone and the low-level environmental winds (see Figure 6). This horizontal streamwise vorticity is tilted into the vertical as it encounters the updraft, therefore contributing to the updraft rotation. The vertical pressure gradient enabled by the rotation further accelerates air upward, enhancing hydrometeor formation and growth. The hydrometeors are differentially advected by the sheared environmental winds to fall forward of the updraft. As already noted, these hydrometeors and their associated evolutions (e.g., melting, evaporation) contribute to the forward-flank downdraft /gust front and, by virtue of the mesocyclone, the rear-flank downdraft /gust front. The statement made regarding MCS longevity can be repeated here nearly verbatim: as long as this coupling is maintained, the supercell will persist.

Supercells typically last for several hours, which again is much longer than an ordinary unicellular storm. A late-evening end to the lifecycle of the supercell morphology can be the beginning of the MCS morphology, especially in situations of more than one supercell: A natural progression during a diurnal cycle is the expansion of individual cold pools such that the cold pool of one supercell collides with the cold pool of another. The result is the formation of an MCS, as discussed previously, and a form of upscale convective growth. The development of a nocturnal low-level jet (e.g., see Shapiro et al., 2016) facilitates this upscale growth within the Great Plains region of the United States.

Neither supercells nor MCSs last indefinitely, however. This is partly due to the fact that their environments are not steady as tacitly assumed, nor are they geographically homogeneous. Indeed, the environmental buoyancy and vertical wind shear that strongly control the convective morphology and intensity are described by larger-scale three-dimensional temperature, moisture, and wind fields that evolve in time and space. The rotation of the earth about its axis, and the resultant diurnal cycle in solar heating, impose one temporal scale on the temperature and implicitly on the other variables; the seasonal cycle in the orbit of the earth about the sun imposes yet another. The characteristics of the earth-sun relationship in general account for Rossby waves as well as baroclinic and barotropic instabilities, which shape the weather patterns—and their associated three-dimensional distributions in temperature, moisture, and wind—that encircle and move around the globe. The basic consequence of the spatially and temporally varying weather patterns is that a convective storm can at one time and/or location thrive within a favorable environment, yet weaken at another time and/or location within an unfavorable environment.

It is in this light that the effects of the land surface should also be mentioned. Consider, for example, that a convectively generated cold pool will be heated from below upon moving over a hot, dry land surface. Because the cold pool behaves as a density current, and therefore has a motion that is proportional to its temperature deficit, this particular modification by vertical fluxes of heat (and moisture) reduces the deficit and slows the cold pool motion (e.g., Ross et al., 2004).6 In turn, a slowed and diluted cold pool weakens the coupling with the convective updraft and downdraft, and ultimately leads to the demise of the convective storm. Other land surface properties, such as the effective roughness, will also influence cold-pool motion and its degree of coupling with the other convective components.

A storm that has undergone a demise can still influence convective storm development during later periods, even on a subsequent day. This is particularly true for storms that persist well after sunset. Remnant cloudiness, such as from an “orphan anvil” (cirrus cloud that previously comprised the anvil-shaped thunderstorm top), can reduce solar insolation and its contribution to parcel buoyancy. A remnant gust front associated with a largely diluted cold pool can, on the other hand, aid in parcel lifting and otherwise help locally condition the atmosphere for deep convection. A remnant vortex—a mesoscale convective vortex (MCV)—can also serve to initiate a new MCS.

An MCV is a manifestation of an upscale convective feedback, and exemplifies a convection-storm proliferation of sorts. An MCV typically originates in an MCS with vertically rotating cores that form in much the same way as do mesocyclones. The cyclonically rotating cores are enhanced and aggregated in association with storm-scale convergence of the background cyclonic vertical vorticity due to the earth’s rotation. This process is alternatively explained in terms of generation of potential vorticity by diabatic heating—namely, the latent heat released when, for example, water vapor is condensed into liquid cloud drops within the convective updraft of the MCS (Raymond & Jiang, 1989). The vortex, which tends to reside within the middle levels of the troposphere, and has a diameter ~100 km, can persist into the next day and thus well after the demise of the MCS. The wind and thermal structure of the MCV provide a mechanism for parcel lifting. The now remnant MCV subsequently acts to initiate new deep convection during the next day and, if the environmental CAPE and wind shear still support it, an MCS may form out of this new convection. The diabatic heating within the convective updrafts reinvigorates the vortex, and the cycle continues, possibly for multiple days.

# Convective-Storm Hazards

Flooding is a particular hazard associated with MCVs and more generally with MCSs. This owes to the broad areas of convective rainfall generated by MCSs over a similar geographic region, and perhaps also over multiple days. Nontornadic “straight-line” wind damage over large areas is also most often caused by MCSs, especially by those with a “bow-echo” presentation on radar. In contrast, the most intense tornadoes are spawned by supercells, as is the largest hail (Smith et al., 2012).

The ability to attribute specific hazards to specific convective-storm morphologies has significant value in numerous research and forecasting applications, especially because of the aforementioned relationship between environmental conditions and morphology. Thus observational data on (and future predictions of) environmental temperature, moisture, and wind—and derived parameters like CAPE and S06—should equate to knowledge about morphology and thence to predominant hazard. In addition to its routine use in daily forecasts of hazardous weather, this approach has been used to construct future projections of convective-storm hazards under anthropogenic climate change (e.g., Trapp et al., 2007).

A troubling issue, however, is that none of these parameters have values that unambiguously relate to a specific morphology. This is due in part to the imperfect concept and thus evaluation of an environment; as already noted, this in turn is due to an atmosphere that is inherently nonsteady and heterogeneous. Moreover, given an ambiguous environment in the presence of certain atmospheric forcings, one morphology might be coerced into evolving into another. For example, within a marginal supercell environment, the “linear” forcing of a front is known to promote the evolution of a line of discrete supercells into a squall line (e.g., Dial et al., 2010).

According to the research methods described above, collection of meteorological data that are highly resolved in time and space will be helpful toward addressing these and related issues, as will numerical model simulations of convective storms in nonsteady, spatially heterogeneous settings.

## References

Bluestein, H. B. (1999). A history of severe-storm-intercept field programs. Weather Forecasting, 14(4), 558–577.Find this resource:

Browning, K. A. (1964). Airflow and precipitation trajectories within severe local storms which travel to the right of the winds. Journal of Atmospheric Science, 21(6), 634–639.Find this resource:

Bryan, G. H., & Fritsch, J. M. (2002). A benchmark simulation for moist nonhydrostatic numerical models. Monthly Weather Review, 130(12), 2917–2928.Find this resource:

Coniglio, M. C., Corfidi, S. F., & Kain, J. S. (2011). Environment and early evolution of the 8 may 2009 derecho-producing convective system. Monthly Weather Review, 139, 1083–1102.Find this resource:

Corfidi, S. F. (1999). The birth and early years of the storm prediction center. Weather Forecasting, 14(4), 507–525.Find this resource:

Corfidi, S. F., Coniglio, M. C., Cohen, A. E., & Mead, C. M. (2016). A proposed revision to the definition of “derecho.” Bulletin of the American Meteorological Society, 97, 935–949.Find this resource:

Dial, G. L., Racy, J. P., & Thompson, R. L. (2010). Short-term convective mode evolution along synoptic boundaries. Weather Forecasting, 25, 1430–1446.Find this resource:

Doswell, C. A., & Markowski, P. M. (2004). Is buoyancy a relative quantity? Monthly Weather Review, 132(4), 853–863.Find this resource:

Fawbush, E. J., & Miller, R. C. (1954). The types of airmasses in which North American tornadoes form. Bulletin of the American Meteorological Society, 35(4), 154–165.Find this resource:

Fujita, T. T. (1986). Mesoscale classifications: Their history and their application to forecasting. In P. Ray (Ed.), Mesoscale Meteorology and Forecasting (pp. 18–35). Boston: American Meteorological Society.Find this resource:

Klemp, J. B. (1987). Dynamics of tornadic thunderstorms. Annual Review of Fluid Mechanics, 19, 369–402.Find this resource:

Lemon, L. R., & Doswell, C. A. (1979). Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Monthly Weather Review, 107(9), 1184–1197.Find this resource:

Rasmussen, E. N., & Blanchard, D. O. (1998). A baseline climatology of sounding-derived supercell and tornado forecast parameters. Weather Forecasting, 13, 1148–1164.Find this resource:

Raymond, D. J., & Jiang, H. (1989). A theory for long-lived mesoscale convective systems. Journal of Atmospheric Science, 47(24), 3067–3077.Find this resource:

Ross, A. N., Tompkins, A. M., & Parker, D. J. (2004). Simple models of the role of surface fluxes in convective cold pool evolution. Journal of Atmospheric Science, 61(13), 1582–1595.Find this resource:

Rotunno, R., Klemp, J. B., & Weisman, M. L. (1988). A theory for strong, long-lived squall lines. Journal of Atmospheric Science, 45(3), 463–485.Find this resource:

Schultz, D. M., Schumacher, P. N., & Doswell, C. A. (2000). The intricacies of instabilities. Monthly Weather Review, 128(12), 4143–4148.Find this resource:

Shapiro, A., Fedorovich, E., & Rahimi, S. (2016). A unified theory for the Great Plains nocturnal low-level jet. Journal of Atmospheric Science, 73(8), 3037–3057.Find this resource:

Skamarock, W. C., Klemp, J. B., Dudhia, J., Gill, D. O., Barker, D. M., Wang, W., . . . Powers, J. G. (2008). A description of the Advanced Research WRF version 3. Boulder, CO: Mesoscale and Microscale Meteorology Division National Center for Atmospheric Research. [PDF Document]. Retrieved from http://www2.mmm.ucar.edu/wrf/users/docs/arw_v3.pdf.

Smith, A. M., McFarquhar, G. M., Rauber, R. M., Grim, J. A., Timlin, M. S., Jewett, B. F., & Jorgensen, D. P. (2009). Microphysical and thermodynamic structure and evolution of the trailing stratiform regions of mesoscale convective systems during BAMEX. Part I: Observations. Monthly Weather Review, 137, 1165–1185.Find this resource:

Smith, B. T., Thompson, R. L., Grams, J. S., Broyles, C., & Brooks, H. E. (2012). Convective modes for significant severe thunderstorms in the contiguous United States. Part I: Storm classification and climatology. Weather Forecasting, 27(5), 1114–1135.Find this resource:

Thompson, R. L., Smith, B. T., Grams, J. S., Dean, A. R., & Broyles, C. (2012). Convective modes for significant severe thunderstorms in the contiguous United States. Part II: Supercell and QLCS tornado environments. Weather Forecasting, 27(5), 1136–1154.Find this resource:

Trapp, R. J. (2014). On the significance of multiple consecutive days of tornado activity. Monthly Weather Review, 142, 1452–1459.Find this resource:

Trapp, R. J., Diffenbaugh, N. S., Brooks, H. E., Baldwin, M. E., Robinson, E. D., & Pal, J. S. (2007). Changes in severe thunderstorm environment frequency during the 21st century caused by anthropogenically enhanced global radiative forcing. Proceedings of the National Academy of Sciences, 104(50), 19719–19723.Find this resource:

Weckwerth, T. M., & Parsons, D. B. (2006). A review of convection initiation and motivation for IHOP_2002. Monthly Weather Review, 134(1), 5–22.Find this resource:

Weisman, M. L. (1992). The role of convectively generated rear-inflow jets in the evolution of long-lived mesoconvective systems. Journal of Atmospheric Science 49, 1826–1847.Find this resource:

Weisman, M. L., & Klemp, J. B. (1982). The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Monthly Weather Review 110, 504–520.Find this resource:

## Notes:

(1.) For a somewhat different perspective, see Doswell and Markowski (2004).

(6.) Strictly speaking, the density current speed is: $Vdc=kgdρ′/ρo$, where $k$ is an emperical constant, $d$ is the density current depth, $ρo$ is the assumed uniform density of the environment, and $ρ′$ is the difference between the respective densities of the current and environment.