notes and correspondence a simulation of a supercell ...jyh10/pubs/marko-harring-anvilrad.pdf · b....

NOTES AND CORRESPONDENCE

A Simulation of a Supercell Thunderstorm with Emulated Radiative Cooling beneaththe Anvil

PAUL M. MARKOWSKI AND JERRY Y. HARRINGTON

Department of Meteorology, The Pennsylvania State University, University Park, Pennsylvania

(Manuscript received 10 September 2004, in final form 20 December 2004)

ABSTRACT

This note reports the preliminary results of an ongoing numerical study designed to investigate whateffects, if any, radiative transfer processes can have on the evolution of convective storms. A pair ofidealized three-dimensional simulations are conducted to demonstrate the potential dynamical importanceof shortwave radiation reductions within the large shadows cast by storms. One of the simulations (thecontrol) is run without surface physics and radiation. In the other simulation, radiative cooling due to cloudshading is emulated by prescribing a cooling rate to the skin temperature at any grid point at which cloudwater was present overhead. The imposed skin cooling rate is consistent with past observations. Low-levelair temperatures are coupled to the skin cooling in this second simulation by the inclusion of surface sensibleheat fluxes using simple bulk aerodynamic drag laws (latent and soil heat fluxes are not included). Signifi-cant differences are observed between the two simulated storms, particularly in the evolution of the verticalvorticity field and gust fronts. The storm simulated with emulated cloud shading develops substantiallyweaker low-level rotation than the storm in the control simulation.

1. Introduction

In this note, we present the preliminary results of anongoing investigation of the effects of radiative transferprocesses associated with long-lived convective stormson storm dynamics. This paper summarizes the out-come of a three-dimensional numerical simulation ofa supercell thunderstorm in which cloud shadinghas been crudely parameterized. Surface cooling dueto the cloud shading modifies the low-level tempera-ture field, and these modifications, in turn, alter thestorm kinematic fields in significant ways. A reviewof pertinent past research is presented in section 2.Section 3 describes the design of the numerical experi-ment and summarizes the results. Section 4 contains adiscussion of the results. Finally, some comments on thefuture work that is planned in this study appear in sec-tion 5.

2. Motivation

a. Long-lived convective storms and radiativetransfer processes

Long-lived convective storms tend to be among thestorms most likely to produce severe weather, such asflash floods, destructive straight-line winds, hail, andtornadoes. For this reason, long-lived storms, such assupercell thunderstorms and mesoscale convective sys-tems (MCSs), have been some of the most intenselystudied forms of atmospheric convection. Conse-quently, these types of storms are perhaps better un-derstood than less severe forms of convection. For ex-ample, considerable dynamical insight into supercellthunderstorms (e.g., the dynamics of storm propagationand the origins of midlevel storm rotation) was ac-quired once computing power became sufficient for thethree-dimensional numerical simulation of such stormsin roughly the mid-1970s (Schlesinger 1975; Klemp andWilhelmson 1978a,b; Wilhelmson and Klemp 1978; Ro-tunno and Klemp 1982, 1985; Klemp and Rotunno1983; Weisman and Klemp 1982, 1984). Although manyobservational studies have provided a foundation for

Corresponding author address: Dr. Paul Markowski, The Penn-sylvania State University, 503 Walker Building, University Park,PA 16802.E-mail: [email protected]

JULY 2005 N O T E S A N D C O R R E S P O N D E N C E 2607

© 2005 American Meteorological Society

JAS3497

our current understanding of MCSs (Newton andFankhauser 1964; Houze 1977; Zipser 1977; Maddox1980), numerical simulations also have been indispens-able in augmenting our comprehension of these larger-scale thunderstorm conglomerations (Hane 1973;Thorpe et al. 1982; Rotunno et al. 1988; Weisman et al.1988; Fovell and Ogura 1988).

Even though numerical simulations often have pro-duced convective storms that closely resemble thosethat have been observed, a potentially important forc-ing has frequently been ignored in these simulations—the influence of radiative effects on storm dynamics.Virtually all of today’s mesoscale weather predictionmodels include radiation parameterizations, in additionto representations of the other components of the sur-face energy budget, but the representation of the sur-face energy budget in numerical simulations of convec-tive storms done for research purposes has been farfrom commonplace, perhaps in the interest of limitingthe studies to the essential dynamics.

It is true that some two-dimensional numerical stud-ies of MCSs have previously included longwave radia-tive transfer processes. For example, it has been shownthat the formation of the well-documented transitionzone in MCSs (Smull and Houze 1985; Rutledge andHouze 1987; Biggerstaff and Houze 1993) can be sen-sitive to longwave radiation (Chen and Cotton 1988;Chin 1994; Braun et al. 1996). Furthermore, longwaveradiation effects are known to affect the circulationwithin the trailing stratiform regions of simulated MCSs(Chen and Cotton 1988; Dudhia 1989; Churchill andHouze 1991; Tao et al. 1991), ultimately enhancing pre-cipitation amounts within the stratiform regions (Tao etal. 1993), which typically account for a significant frac-tion (up to 50%) of the total precipitation produced byMCSs (Houze 1977; Zipser et al. 1981; Gamanche andHouze 1983; Rutledge and Houze 1987; Johnson andHamilton 1988). Although the studies cited above havemade important contributions toward our understand-ing of MCSs, the full range of possible dynamical effectsowing to radiative effects, especially those owing toshortwave radiative transfer processes, remains uncer-tain due to the model dimensionality of these previousstudies. In three-dimensional numerical simulations ofMCSs (e.g., Rotunno et al. 1988; Weisman et al. 1988;Dudhia and Moncrieff 1989; Weisman 1992, 1993; Ska-marock et al. 1994; Trier et al. 1997, 1998), radiativeeffects generally have not been considered, althoughTucker and Crook (1999) recently simulated an MCScase with and without solar radiation. They found thatthe inclusion of solar radiative effects reduced the in-tensity of the MCS, although the details of how solarradiative effects were included were not presented.

Radiative effects also have virtually always been ex-cluded from three-dimensional simulations of supercellstorms, even though computing capabilities have ad-vanced significantly since the seminal studies of the late1970s and early 1980s (e.g., Wilhelmson and Klemp1978; Rotunno and Klemp 1982; Klemp and Rotunno1983). Some recent three-dimensional case study simu-lations of supercellular convection have included radia-tion, but these parameterizations have been rathercrude. For example, often clouds were seen only asareas of very high water vapor content, with the radia-tive characteristics of condensed liquid and ice speciesnot taken into account (e.g., Finley et al. 2001), leadingto overestimates of solar fluxes reaching the surface incloudy regions and underestimates of longwave coolingat cloud tops. Furthermore, to our knowledge, in thetwo- and three-dimensional simulations (of both MCSsand supercells) that have included radiation, no sensi-tivity analyses were done to quantify the effects of ra-diation on storm dynamics. The exclusion of radiativeeffects in the majority of convective storm simulationstudies (particularly three-dimensional simulations) hasbeen justified with the assumption that radiative pro-cesses are unimportant on the time scales of the modelintegration (e.g., Trier et al. 1997, 1998) and that thestorms are “largely dynamically (not radiatively)driven” (Finley et al. 2001). Of course, all of the abovestatements are true in some sense, for the past modelingsuccesses likely would not have been achieved other-wise. However, what we argue here, and what we arepresently examining, is the possibility that convectivestorm dynamics may be modulated in certain, and per-haps significant, ways by radiative effects.

b. Observations of significant radiative effectsattributable to long-lived convective storms

It is well known that low-level baroclinity can havean important influence on convective storm dynamics.For example, low-level baroclinic zones (gust fronts)commonly are associated with the precipitation regionsof convective storms. Klemp and Rotunno (1983) andRotunno and Klemp (1985) showed that the forward-flank gust front of supercell thunderstorms (Lemon andDoswell 1979) may be an important source of horizon-tal vorticity for the updrafts of the storms. This hori-zontal vorticity is generated solenoidally along theboundary separating rain-cooled outflow from the rela-tively warm inflow. Large vertical velocity gradients as-sociated with the updraft may tilt this horizontal vor-ticity to give rise to significant low-level vertical vortic-ity in supercell storms. It also has been shown thatbaroclinic boundaries associated with the precipitation

2608 J O U R N A L O F T H E A T M O S P H E R I C S C I E N C E S VOLUME 62

regions of other storms (e.g., adjacent storms or evenstorms occurring earlier in the day) may have similarlyimportant dynamical consequences. For instance, whena storm crosses the outflow boundary left behind bysome other region of convection, the storm may ingestenhanced baroclinic horizontal vorticity residing alongthe outflow boundary, which may lead to rapid inten-sification of low-level rotation (and often tornadogen-esis), following tilting and stretching of the augmentedhorizontal vorticity (Purdom 1976; Maddox et al. 1980;Weaver and Nelson 1982; Markowski et al. 1998a; At-kins et al. 1999; Rasmussen et al. 2000). Even subtlelow-level static stability changes not associated withhorizontal temperature gradients are known to havesignificant impacts on storm behavior, and possiblyeven tornadogenesis (McCaul and Weisman 2001;Markowski et al. 2003).

Although the potential dynamical importance of low-level baroclinic zones associated with the precipitationregions of convection is well established, the dynamicalsignificance of low-level baroclinic zones arising fromradiative effects associated with convective storms havenot been previously examined. Yet the magnitude ofthe low-level temperature perturbations and baroclinitydue to anvil-generated radiative effects has been foundto be comparable to the magnitude of baroclinity asso-ciated with storm-scale precipitation regions in at leastsome cases (e.g., Markowski et al. 1998b). This may notbe entirely surprising. Even in stratiform precipitationevents (e.g., those associated with cold-air damming),radiative effects have been found, at least in a fewcases, to exert a greater influence on the low-level sta-bility and horizontal temperature gradients than latentcooling associated with the evaporation of precipitation(e.g., Fritsch et al. 1992).

Markowski et al. (1998b) showed that barocliniczones arising from surface cooling beneath opticallythick anvils may be capable of generating horizontalvorticity of the same order of magnitude as that whichis often generated along the more extensively studiedgust fronts, and also of the same magnitude as the hori-zontal vorticity associated with the mean vertical windshear of the ambient, large-scale environment [O(10�2)s�1]. Such horizontal vorticity can be converted to ver-tical vorticity within storm updrafts via tilting, then am-plified by stretching (Rotunno 1981; Davies-Jones 1984;Klemp 1987). Deep, persistent updraft rotation is thedefining characteristic of supercell storms; thus, meansof enhancing low-level horizontal vorticity may haveimportant implications for the dynamics of such storms,in addition to the dynamics of other types of convectivestorms.

In the cases documented by Markowski et al.

(1998b), temperature deficits as large as 5–6 K werefound to develop within the anvil shadows. The tem-perature differences were observed to occur over hori-zontal distances of �25 km. The rate at which horizon-tal vorticity is generated by such a baroclinic zone ex-ceeds 0.02 s�1 h�1. The temperature gradients andassociated vorticity generation rates owing to the anvilshadows may be smaller than temperature gradientsoften associated with the gust fronts of convectivestorms (�5 K km�1). However, parcel residence timeswithin anvil-generated baroclinic zones (Markowski etal. estimated the time scales to be �1 h) would gener-ally be larger than parcel residence times within thebaroclinic zones associated with storm-scale gust fronts(�5–15 min) due to the larger horizontal scale of theanvil shadow compared to the scale of the precipitationregion of the storm. Thus, the total horizontal vorticitygenerated baroclinically may be comparable to thatproduced by low-level outflow [O(10�2) s�1].

One question that Markowski et al. (1998b) wereunable to fully address was the depth over which thecooling and baroclinity developed. A few soundings in-dicated that the low-level baroclinity was likely severalhundred meters in depth. Although it is not known howdeep the baroclinity must extend for significant dy-namical effects to arise (this will be addressed by theresearch currently under way), there is growing evi-dence that modifications of just the lowest few hundredmeters of the environmental vertical wind profile mayhave profound effects on storm behavior (Wicker 1996;Markowski et al. 1998c). Low-level temperature modi-fications by anvils also affect the convective availablepotential energy (CAPE) and convective inhibition(CIN) of the storm environment. It is not known whateffects these alterations may have on convective stormbehavior when included in a numerical simulation. Fur-thermore, it is not known how the effects of baroclinicvorticity differ depending on whether the baroclinicvorticity has been generated within a region in whichequivalent potential temperature (�e) has been nearlyconserved (as might be expected to be the case alonggust fronts where cooling is largely due to evaporationof precipitation) or whether the baroclinic vorticity hasbeen generated within a region in which �e deficits alsohave been generated (as might be expected to be thecase where cooling owes to radiative effects). Even theforward speed of the convective storm might be ex-pected to affect the magnitude of the radiative effects,at least those related to surface cooling due to a reduc-tion of incident shortwave radiation. For example, arapidly moving storm with its attendant anvil will con-stantly be encountering warm ground that it will haveto cool in order to develop surface temperature


gradients. On the other hand, a stationary storm canshade the same ground for a longer period of time,thereby producing correspondingly larger surface tem-perature gradients. Conversely, the development ofsurface temperature gradients in storms due to low-level evaporative cooling is Galilean invariant since theprecipitation region (which produces the cooling)moves with the storm.

3. Numerical simulations with emulated radiativecooling

Using version 4.5.2 of the Advanced Regional Pre-diction System (ARPS; Xue et al. 2000, 2001), wepresent a pair of preliminary simulations in order toillustrate the potentially important effects of radiativetransfer processes on convective storm dynamics. Thesimulations are initialized with the composited sound-ing from the well-documented 20 May 1977 Del City,Oklahoma, storm (Ray et al. 1981; Johnson et al. 1987),which has been used in several other modeling studies(e.g., Klemp et al. 1981; Klemp and Rotunno 1983;Grasso and Cotton 1995; Adlerman et al. 1999; Adler-man and Droegemeier 2002). This sounding is charac-terized by a CAPE of approximately 2600 J kg�1 and a0–3 km storm-relative helicity of approximately 150m2 s�2 (Fig. 1). The hodograph is shifted by a mean

velocity so that the mature, cyclonically rotating stormis nearly stationary. One of the simulations (the con-trol) is run without surface physics and radiation. In theother simulation, radiative cooling due to anvil shadingis emulated by prescribing a cooling rate to the skintemperature of 5 K h�1 at any grid point at which cloudwater was present overhead. This skin cooling rate issimilar to that observed by Markowski et al. (1998b).Low-level air temperatures are coupled to the skincooling in this second simulation by the inclusion ofsurface sensible heat fluxes using simple bulk aerody-namic drag laws (latent and soil heat fluxes were notincluded). Though this emulation of radiative cooling isadmittedly simple, it should suffice to illustrate the po-tential effects that radiative cooling under the anvil mayhave on storm dynamics.

The domain is 200 km � 200 km � 20 km, the hori-zontal grid spacing is 1000 m, and the vertical grid spac-ing varies from 150 m in the boundary layer to 500 mnear the tropopause. Only warm-rain (Kessler) micro-physics is used for both of the idealized simulations.Though this microphysical simplification is not neces-sary for our demonstration, we choose to use warm rainonly for two reasons: computational expedience and,more importantly, neglecting ice processes leads to ananvil of smaller areal extent. Hence, if radiative coolingis important here, its importance will likely be ampli-

FIG. 1. (left) Skew T–logp diagram and (right) hodograph used to initialize the demonstration simu-lations. Units on the hodograph axes are m s�1, and open circles are placed along the hodograph at 1-kmintervals (every other one is labeled in km).


fied in the presence of a longer, thicker anvil. Neithersimulation includes the Coriolis force.

The evolution of both simulations in the first hour issimilar and will not be discussed in detail [see Klemp etal. (1981) for a more detailed discussion]. The initialupdraft splits into anticyclonically (left moving) and cy-clonically (right moving) rotating cells approximately20 min into the simulations. The cyclonically rotatingcell dominates throughout the rest of the simulations,becoming a mature supercell approximately after 40min have elapsed. By about 60 min, the supercell takeson a nearly steady-state character in both simulations,with a persistent midlevel (3–7 km) mesocyclone andupdrafts greater than 30 m s�1.

The purpose of the simulations is to show that, be-yond roughly the first 60 min, there are significant dif-ferences between the two supercells—differences thatcan only be due to effects associated with the emulatedanvil shading. Our goal is not to present detailed diag-nostics of the model output here, elucidating the pre-cise dynamics responsible for the simulation differ-ences, for this is one of the goals of our future research.Our goal here is only to show that radiative effectsassociated with cirrus anvils can have a significant im-pact on convective storm characteristics.

The differences between the two simulations do notbecome significant until after approximately an hourhas elapsed (Figs. 2, 3, and 4) by which time a well-developed anvil is present (Fig. 2). In the simulationhaving emulated cloud shading, near-ground air tem-perature deficits beneath the anvil outside of the for-ward-flank precipitation region of up to 3 K (4 K) de-velop by t � 2 h (t � 3 h). These deficits are similar inamplitude to those observed by Markowski et al.(1998b). Consistent with the prescribed skin coolingrate of 5 K h�1 beneath clouds, skin temperature defi-cits of approximately 8 K (13 K) develop by t � 2 h(t � 3 h; Fig. 5a). The minimum surface sensible heatflux beneath the anvil is approximately �450 W m�2

(Fig. 5b), which seems reasonable given that observednet radiation decreases of �500 W m�2 have been ob-served with the passage of optically thick anvils associ-ated with severe storms (Markowski et al. 1998b). Thedepth of the cooling (Fig. 6) is approximately 500 m,which also is similar to the observations documented byMarkowski et al. (1998b). The similarity of the horizon-tal air temperature fields, vertical temperature profiles,and sensible heat fluxes to those that have been ob-served beneath anvils provide some confidence that thesimple cloud shading parameterization has produced atleast a qualitatively realistic temperature field beneaththe anvil.

Although the temperature deficits may be considered

small and shallow compared to those frequently foundin the precipitation-induced convective outflows ofsimulated storms (often �8 K), the expansive region ofsurface cooling in the storm inflow has an effect onsubsequent storm morphology and evolution that maybe regarded as substantial. For example, the time seriesof maximum vertical velocity and low-level vertical vor-ticity (Fig. 4) have differences as large as 20% (100%)at some times. The rainwater fields also have obviousdifferences by t � 3 h (Figs. 3c and 3f).1

The supercell simulated without radiative effectsgenerally has much stronger low-level rotation and amore occluded gust-front structure compared to the su-percell simulated with emulated cloud shading (Figs. 3and 4). In fact, the differences in the near-surface vec-tor wind fields in the simulated storms are much largerthan those that have been observed between tornadicand nontornadic supercells (Blanchard and Straka1998; Trapp 1999; Wakimoto and Cai 2000; Markowskiet al. 2002). The low-level potential temperature fields,especially the orientation and magnitude of the gradi-ents in close proximity to the storms, have readily dis-cernible differences; thus, it is not surprising that largekinematic differences (e.g., the vertical vorticity differ-ences) arise, given the well-known relationship be-tween temperature gradients, horizontal vorticitymodulation, and vertical vorticity generation by tilting.

4. Discussion

We reiterate that a detailed investigation of the dy-namical causes for the differences between the twosimulations described in section 2 is beyond the scopeof this preliminary report, but that the results displayedin Figs. 3 and 4 quite clearly indicate that substantialmodifications of storm behavior can be brought aboutby radiative processes. In this pair of example simula-

1 An additional control simulation was conducted in which ra-diation was excluded but a surface sensible heat flux was included.Although the surface energy budget was somewhat unrealistic,the experiment was conducted in order to address concerns thatthe differences between the simulations with and without radia-tion and surface physics might be due largely to the presence of asurface sensible heat flux rather than radiative cooling in thecloud shadow. Although some differences were observed betweenthe two control simulations, the differences between each controlsimulation and the simulation that included radiation were con-siderably larger. Thus, the differences between the original con-trol simulation and the simulation that included radiation andsurface physics are much more strongly a consequence of theradiative cooling of the shaded ground (coupled to the air tem-perature by the sensible heat flux) than simply a consequence ofthe inclusion of the sensible heat flux.


tions, it appears as though low-level cooling beneaththe anvil has had a detrimental effect on the convectivestorm. In other cases, (e.g., cases in which the orienta-tion of the low-level inflow with respect to the low-levelbaroclinity is more favorable), it is quite possible thatanvils could have an enhancing effect on a convectivestorm. The research that is currently under way is de-

signed to determine precisely the situations in whichradiation has a significant effect on convective storms,as well as to determine whether the effects are positiveor negative and what is the dynamical nature of theeffects.

A few additional comments are in order concerningthe demonstration simulations. The long-lived supercell

FIG. 2. (a)–(c) Cloud isosurfaces and potential temperature perturbations at the lowest model level (75 m) at t � 1 h, t � 2 h, andt � 3 h, respectively, in the simulation in which no radiative parameterization was employed. View is from the south. Potentialtemperature perturbations are contoured from �1 to �8 K at 1-K intervals. (d)–(f) Same as in (a)–(c) but for the simulation in whichshortwave radiative cooling was emulated beneath cloud regions. The dashed rectangles in (a) and (d) enclose the regions depicted inFig. 3. The star in (e) indicates the location of the sounding shown in Fig. 6a.


FIG. 3. (a)–(c) Horizontal cross sections of the rainwater (3 km), potential temperature perturbation (75 m), and vertical vorticity (75m) fields at t � 1 h, t � 2 h, and t � 3 h, respectively, in the simulation in which no radiative parameterization was employed. The regionshown is indicated in Fig. 2a. Light (dark) gray shading denotes regions where rainwater exceeds 1 g kg�1 (3 g kg�1). The verticalvorticity field is analyzed at 0.004 s�1 intervals using bold contours (negative contours are dashed). Negative potential temperatureperturbations are analyzed at 1-K intervals using thin, dashed contours. Horizontal (ground relative) wind velocity vectors at 75 m havebeen drawn at every third grid point. (d)–(f) Same as in (a)–(c) but for the simulation in which shortwave radiative cooling was emulatedbeneath cloud regions. The region shown is indicated in Fig. 2d.


storms simulated in the pair of demonstration simula-tions are nearly stationary. Thus, it might seem that thesimulation with emulated radiative cooling perhaps ex-aggerates the effect of anvil shading due to shading ofthe same region for a long duration. However, the sur-face temperature deficit and associated horizontalbaroclinity beneath the anvil, even after 3 h (notshown), are not as large as has been observed. Further-

more, the hodograph structure (Fig. 1) is not ideal—notonly did low-level inflow parcels fail to spend muchtime in the anvil-generated baroclinic zone, but thebaroclinic zone itself was situated so that inflow enter-ing the updraft did not pass through the (southern)portion of the baroclinic zone that would have aug-mented the horizontal vorticity associated with thebase-state vertical wind shear. Instead, the horizontal

FIG. 5. (a) Horizontal cross sections of the rainwater (3 km) and skin temperature fields at t � 2 h in thesimulation in which shortwave radiative cooling was emulated beneath cloud regions. The region shown is indicatedin Fig. 2d. Light (dark) gray shading denotes regions where rainwater exceeds 1 g kg�1 (3 g kg�1). The skintemperature field is analyzed at 2-K intervals. Horizontal (ground relative) wind velocity vectors at 75 m have beendrawn at every third grid point. (b) Horizontal cross sections of the rainwater (3 km) and sensible heat flux fieldsat t � 2 h in the same simulation. The sensible heat flux field is analyzed at 100 W m�2 intervals (negative contoursare dashed).

FIG. 4. Time series of (left) maximum vertical velocity, wmax (m s�1), and (right) maximum vertical vorticity below 2 km,�max (�103 s�1), in the demonstration simulations without radiative effects (solid) and with emulated anvil radiative effects(dashed).


vorticity produced by the anvil-generated barocliniczone actually opposed the horizontal vorticity associ-ated with the base-state vertical wind shear. Perhapsthis is why the radiatively modified storm had weakernear-surface rotation, or perhaps there are other rea-sons (e.g., increased CIN in the inflow) that will bediscovered upon completing diagnostics of future ex-periments using a more complete radiation parameter-ization. It is also worth noting that the demonstrationsimulations do not include ice physics. The inclusion ofice physics leads to the production of more expansive

anvils (and therefore more expansive surface tempera-ture modifications) compared to the warm-rain micro-physical parameterization used for the demonstrationsimulations (Gilmore and Wicker 1998). Given theabove-noted departures from what might be consideredto be much more ideal conditions for magnifying theimportance of anvil radiative effects on convectivestorms, we believe that the demonstration simulationsquite conservatively indicate that radiative effects canhave important effects on convective storms.

It does not seem likely that radiative effects would beimportant in all convective storms. For example, short-lived storms probably would not be affected by radia-tively cooled shadow regions, owing to their relativelybrief duration and small anvils. On the other hand,long-lived storms might be more prone to radiative ef-fects on their dynamics, owing to their larger anvilcanopies and longer storm time scales. Furthermore,long-lived storms tend to occur in environments con-taining large vertical wind shear, which promotes stormorganization and longevity. Large vertical wind sheartypically is associated with strong storm-relative windsat anvil level, leading to the formation of long anvils inthe downstream direction, overlying the storm inflowregion where surface shading effects may have the larg-est impact. Also, the most dangerous convective stormsare often isolated (Browning 1964), and thus may havea well-defined anvil shadow edge and associated low-level baroclinity.

5. Future work

The preliminary work summarized herein representsthe first step of a larger effort to explore the effects ofradiative transfer processes on isolated convectivestorms. Our next step is to include more sophisticatedradiative and microphysical parameterizations in a se-ries of simulations designed to establish bounds on themagnitude of radiative effects on long-lived convectivestorms, as well as the nature of these effects and howthese effects depend on storm morphology and the am-bient environment. Since all radiation codes that areused in operational forecast and cloud models are onedimensional (1D), they only account for clouds that areoverhead. Consequently, true three-dimensional shad-owing effects are not possible. We plan to modify anexisting 1D radiation code so that 3D shadowing effectsare approximately accounted for by modifying the solardirect beam. Since the direct (as opposed to diffuse)radiative fluxes are the primary reason for the shad-owed region, it should be straightforward to modify theattenuation of the solar direct beam so that it accountsfor both solar zenith and azimuth (i.e., 3D) effects. The

FIG. 6. (a) Model sounding beneath the anvil at t � 2 h at thelocation indicated by the star in Fig. 2e. The base-state tempera-ture profile is indicated with the dashed line. (b) Observed sound-ing beneath the anvil of a thunderstorm complex on 8 Jun 1995.(c) Visible satellite image from 2315 UTC 8 Jun 1995. The starindicates the location of the sounding shown in (b).


results of these experiments will be reported in forth-coming articles.

Acknowledgments. We thank the anonymous review-ers and Drs. Jerry Straka and Erik Rasmussen for con-structive suggestions during the course of our work.This research was partially supported by the NationalScience Foundation (Grant ATM-0338661). J. Har-rington was supported by the Office of Science (BER),U.S. Dept. of Energy, Grant DE-F602-05ER64058.

REFERENCES

Adlerman, E. J., and K. K. Droegemeier, 2002: The sensitivity ofnumerically simulated cyclic mesocyclogenesis to variationsin model physical and computational parameters. Mon. Wea.Rev., 130, 2671–2691.

——, ——, and R. Davies-Jones, 1999: A numerical simulation ofcyclic mesocyclogenesis. J. Atmos. Sci., 56, 2045–2069.

Atkins, N. T., M. L. Weisman, and L. J. Wicker, 1999: The influ-ence of preexisting boundaries on supercell evolution. Mon.Wea. Rev., 127, 2910–2927.

Biggerstaff, M. I., and R. A. Houze, 1993: Kinematics and micro-physics of the transition zone of the 10–11 June 1985 squallline. J. Atmos. Sci., 50, 3091–3110.

Blanchard, D. O., and J. M. Straka, 1998: Some possible mecha-nisms for tornadogenesis failure in a supercell. Preprints, 19thConf. on Severe Local Storms, Minneapolis, MN, Amer. Me-teor. Soc., 116–119.

Braun, S. A., R. A. Houze, and M.-J. Yang, 1996: Comments on“The impact of the ice phase and radiation on a midlatitudesquall line system.” J. Atmos. Sci., 53, 1343–1351.

Browning, K. A., 1964: Airflow and precipitation trajectorieswithin severe local storms which travel to the right of thewinds. J. Atmos. Sci., 21, 634–639.

Chen, S., and W. R. Cotton, 1988: The sensitivity of a simulatedextratropical mesoscale convective system to longwave radia-tion and ice-phase microphysics. J. Atmos. Sci., 45, 3897–3910.

Chin, H.-N., 1994: The impact of the ice phase and radiation on amidlatitude squall line system. J. Atmos. Sci., 51, 3320–3343.

Churchill, D. D., and R. A. Houze, 1991: Effects of radiation andturbulence on the diabatic heating and water budget of thestratiform region of a tropical cloud cluster. J. Atmos. Sci., 48,903–922.

Davies-Jones, R. P., 1984: Streamwise vorticity: The origin of up-draft rotation in supercell storms. J. Atmos. Sci., 41, 2991–3006.

Dudhia, J., 1989: Numerical study of convection observed duringthe Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 3077–3107.

——, and M. W. Moncrieff, 1989: A three-dimensional numericalstudy of an Oklahoma squall line containing right-flank su-percells. J. Atmos. Sci., 46, 3363–3391.

Finley, C. A., W. R. Cotton, and R. A. Pielke, 2001: Numericalsimulation of tornadogenesis in a high-precipitation super-cell. Part I: Storm evolution and transition into a bow echo. J.Atmos. Sci., 58, 1597–1629.

Fovell, R. G., and Y. Ogura, 1988: Numerical simulation of a mid-latitude squall line in two dimensions. J. Atmos. Sci., 45,3846–3879.

Fritsch, J. M., J. Kapolka, and P. A. Hirschberg, 1992: The effectsof subcloud-layer diabatic processes on cold air damming. J.Atmos. Sci., 49, 49–70.

Gamanche, J. F., and R. A. Houze, 1983: Water budget of a me-soscale convective system in the tropics. J. Atmos. Sci., 40,1835–1850.

Gilmore, M. S., and L. J. Wicker, 1998: The influence of midtro-pospheric dryness on supercell morphology and evolution.Mon. Wea. Rev., 126, 943–958.

Grasso, L. D., and W. R. Cotton, 1995: Numerical simulation of atornado vortex. J. Atmos. Sci., 52, 1192–1203.

Hane, C. E., 1973: The squall line thunderstorm: Numerical ex-perimentation. J. Atmos. Sci., 30, 1672–1690.

Houze, R. A., 1977: Structure and dynamics of a tropical squallline system. Mon. Wea. Rev., 105, 1540–1567.

Johnson, K. W., P. S. Ray, B. C. Johnson, and R. P. Davies-Jones,1987: Observations related to the rotational dynamics of the20 May 1977 tornadic storms. Mon. Wea. Rev., 115, 2463–2478.

Johnson, R. H., and P. J. Hamilton, 1988: The relationship of sur-face pressure features to the precipitation and airflow struc-ture of an intense midlatitude squall line. Mon. Wea. Rev.,116, 1444–1471.

Klemp, J. B., 1987: Dynamics of tornadic thunderstorms. Annu.Rev. Fluid Mech., 19, 369–402.

——, and R. B. Wilhelmson, 1978a: The simulation of three-dimensional convective storm dynamics. J. Atmos. Sci., 35,1070–1096.

——, and ——, 1978b: Simulations of right- and left-movingstorms produced through storm splitting. J. Atmos. Sci., 35,1097–1110.

——, and R. Rotunno, 1983: A study of the tornadic region withina supercell thunderstorm. J. Atmos. Sci., 40, 359–377.

——, R. B. Wilhelmson, and P. S. Ray, 1981: Observed and nu-merically simulated structure of a mature supercell thunder-storm. J. Atmos. Sci., 38, 1558–1580.

Lemon, L. R., and C. A. Doswell, 1979: Severe thunderstorm evo-lution and mesocyclone structure as related to tornadogen-esis. Mon. Wea. Rev., 107, 1184–1197.

Maddox, R. A., 1980: Mesoscale convective complexes. Bull.Amer. Meteor. Soc., 61, 1374–1387.

——, L. R. Hoxit, and C. F. Chappell, 1980: A study of tornadicthunderstorm interactions with thermal boundaries. Mon.Wea. Rev., 108, 322–336.

Markowski, P. M., E. N. Rasmussen, and J. M. Straka, 1998a: Theoccurrence of tornadoes in supercells interacting with bound-aries during VORTEX-95. Wea. Forecasting, 13, 852–859.

——, ——, ——, and D. C. Dowell, 1998b: Observations of low-level baroclinity generated by anvil shadows. Mon. Wea.Rev., 126, 2942–2958.

——, J. M. Straka, and E. N. Rasmussen, 1998c: A preliminaryinvestigation of the importance of helicity “location” in thehodograph. Preprints, 19th Conf. on Severe Local Storms,Minneapolis, MN, Amer. Meteor. Soc., 230–233.

——, E. N. Rasmussen, and J. M. Straka, 2002: Direct surfacethermodynamic observations within the rear-flank down-drafts of nontornadic and tornadic supercells. Mon. Wea.Rev., 130, 1692–1721.

——, J. M. Straka, and E. N. Rasmussen, 2003: Tornadogenesisresulting from the transport of circulation by a downdraft:Idealized numerical simulations. J. Atmos. Sci., 60, 795–823.

McCaul, E. W., and M. L. Weisman, 2001: The sensitivity of simu-lated supercell structure and intensity to variations in the


shapes of environmental buoyancy and shear profiles. Mon.Wea. Rev., 129, 664–687.

Newton, C. W., and J. C. Fankhauser, 1964: On the movements ofconvective storms, with emphasis on size discrimination inrelation to water budget requirements. J. Appl. Meteor., 3,651–668.

Purdom, J. F. W., 1976: Some uses of high-resolution GOES im-agery in the mesoscale forecasting of convection and its be-havior. Mon. Wea. Rev., 104, 1474–1483.

Rasmussen, E. N., S. J. Richardson, J. M. Straka, P. M.Markowski, and D. O. Blanchard, 2000: The association ofsignificant tornadoes with a baroclinic boundary on 2 June1995. Mon. Wea. Rev., 128, 174–191.

Ray, P. S., B. Johnson, K. W. Johnson, J. J. Stephens, K. K. Wag-ner, R. B. Wilhelmson, and J. B. Klemp, 1981: The morphol-ogy of severe tornadic storms on 20 May 1977. J. Atmos. Sci.,38, 1643–1663.

Rotunno, R., 1981: On the evolution of thunderstorm rotation.Mon. Wea. Rev., 109, 577–586.

——, and J. B. Klemp, 1982: The influence of the shear-inducedpressure gradient on thunderstorm motion. Mon. Wea. Rev.,110, 136–151.

——, and ——, 1985: On the rotation and propagation of simu-lated supercell thunderstorms. J. Atmos. Sci., 42, 271–292.

——, ——, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45, 463–485.

Rutledge, S. A., and R. A. Houze, 1987: A diagnostic modelingstudy of the trailing stratiform region of a midlatitude squallline. J. Atmos. Sci., 44, 2640–2656.

Schlesinger, R. E., 1975: A three-dimensional numerical model ofan isolated deep convective cloud: Preliminary results. J. At-mos. Sci., 32, 934–957.

Skamarock, W. C., M. L. Weisman, and J. B. Klemp, 1994: Three-dimensional evolution of simulated long-lived squall lines. J.Atmos. Sci., 51, 2563–2583.

Smull, B. F., and R. A. Houze, 1985: A midlatitude squall linewith a trailing region of stratiform rain: Radar and satelliteobservations. Mon. Wea. Rev., 113, 117–133.

Tao, W.-K., J. Simpson, and S.-T. Soong, 1991: Numerical simu-lation of a subtropical squall line over Taiwan Strait. Mon.Wea. Rev., 119, 2699–2723.

——, ——, C. H. Sui, B. Ferrier, S. Lang, J. Scala, M. D. Chou,and K. Pickering, 1993: Heating, moisture, and water budgetsof tropical and midlatitude squall lines: Comparisons andsensitivity to longwave radiation. J. Atmos. Sci., 50, 673–690.

Thorpe, A. J., M. J. Miller, and M. W. Moncrieff, 1982: Two-dimensional convection in non-constant shear: A model ofmid-latitude squall lines. Quart. J. Roy. Meteor. Soc., 108,739–762.

Trapp, R. J., 1999: Observations of non-tornadic low-levelmesocyclones and attendant tornadogenesis failure duringVORTEX. Mon. Wea. Rev., 127, 1693–1705.

Trier, S. B., W. C. Skamarock, and M. A. LeMone, 1997: Struc-ture and evolution of the 22 February 1993 TOGA COAREsquall line: Organization mechanisms inferred from numeri-cal simulation. J. Atmos. Sci., 54, 386–407.

——, M. A. LeMone, and W. C. Skamarock, 1998: Effect of three-dimensional structure on the stormwide horizontal accelera-tions and momentum budget of a simulated squall line. Mon.Wea. Rev., 126, 2580–2598.

Tucker, D. F., and N. A. Crook, 1999: The generation of a meso-scale convective system from mountain convection. Mon.Wea. Rev., 127, 1259–1273.

Wakimoto, R. M., and H. Cai, 2000: Analysis of a nontornadicstorm during VORTEX-95. Mon. Wea. Rev., 128, 565–592.

Weaver, J. F., and S. P. Nelson, 1982: Multiscale aspects of thun-derstorm gust fronts and their effects on subsequent stormdevelopment. Mon. Wea. Rev., 110, 707–718.

Weisman, M. L., 1992: The role of convectively generated rear-inflow jets in the evolution of long-lived mesoconvective sys-tems. J. Atmos. Sci., 49, 1826–1847.

——, 1993: The genesis of severe, long-lived bow echoes. J. At-mos. Sci., 50, 645–670.

——, and J. B. Klemp, 1982: The dependence of numericallysimulated convective storms on vertical wind shear and buoy-ancy. Mon. Wea. Rev., 110, 504–520.

——, and ——, 1984: The structure and classification of numeri-cally simulated convective storms in directionally varyingwind shears. Mon. Wea. Rev., 112, 2479–2498.

——, ——, and R. Rotunno, 1988: The structure and evolution ofnumerically simulated squall lines. J. Atmos. Sci., 45, 1990–2013.

Wicker, L. J., 1996: The role of near surface wind shear on low-level mesocyclone generation and tornadoes. Preprints, 18thConf. on Severe Local Storms, San Francisco, CA, Amer.Meteor. Soc., 115–119.

Wilhelmson, R. B., and J. B. Klemp, 1978: A numerical study ofstorm splitting that leads to long-lived storms. J. Atmos. Sci.,35, 1974–1986.

Xue, M., K. K. Droegemeier, and V. Wong, 2000: The AdvancedRegional Prediction System (ARPS): A multiscale nonhydro-static atmospheric simulation and prediction tool. Part I:Model dynamics and verification. Meteor. Atmos. Phys., 75,161–193.

——, and Coauthors, 2001: The Advanced Regional PredictionSystem (ARPS): A multiscale nonhydrostatic atmosphericsimulation and prediction tool. Part II: Model physics andapplications. Meteor. Atmos. Phys., 76, 143–165.

Zipser, E. J., 1977: Mesoscale and convective-scale downdrafts asdistinct components of squall line circulation. Mon. Wea.Rev., 105, 1568–1589.

——, R. J. Meitin, and M. A. LeMone, 1981: Mesoscale motionfields associated with a slowly moving GATE convectiveband. J. Atmos. Sci., 38, 1725–1750.


notes and correspondence a simulation of a supercell ...jyh10/pubs/marko-harring-anvilrad.pdf · b....

Documents