hook echoes and rear-flank downdrafts: a...

852 VOLUME 130M O N T H L Y W E A T H E R R E V I E W

Hook Echoes and Rear-Flank Downdrafts: A Review

PAUL M. MARKOWSKI

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

(Manuscript received 19 March 2001, in final form 24 September 2001)

ABSTRACT

Nearly 50 years of observations of hook echoes and their associated rear-flank downdrafts (RFDs) are reviewed.Relevant theoretical and numerical simulation results also are discussed. For over 20 years, the hook echo andRFD have been hypothesized to be critical in the tornadogenesis process. Yet direct observations within hookechoes and RFDs have been relatively scarce. Furthermore, the role of the hook echo and RFD in tornadogenesisremains poorly understood. Despite many strong similarities between simulated and observed storms, somepossibly important observations within hook echoes and RFDs have not been reproduced in three-dimensionalnumerical models.

1. IntroductionPerhaps the best-recognized radar feature in a horizontal

depiction associated with supercells is the extension oflow-level echo on the right-rear flank of these storms,called the ‘‘hook echo.’’ Hook echoes are known to beassociated with a commonly observed region of subsidingair in supercells, called the ‘‘rear-flank downdraft.’’ Rear-flank downdrafts have been long surmised to be criticalin the genesis of significant tornadoes within supercellthunderstorms. In this paper, observations of hook echoesare reviewed first, beginning with the first documentationsof the 1950s and 1960s. Rear-flank downdrafts are re-viewed next, including discussion of pertinent Dopplerradar observations and theoretical and numerical simula-tion studies. Finally, relatively recent observations madeduring the Verification of the Origins of Rotation in Tor-nadoes Experiment (VORTEX; Rasmussen et al. 1994)and smaller subsequent field operations are reviewed. De-spite the well-known association among hook echoes, rear-flank downdrafts, and tornadogenesis, their dynamical re-lationship remains poorly understood. Our current con-fusion serves as a motivation for the collection of new, insitu observations having unprecedented spatial and tem-poral detail within hook echoes and rear-flank downdrafts.These data are the foundation of a companion paper (Mar-kowski et al. 2002).

2. Hook echoesa. Characteristics

The hook echo first was documented by Stout andHuff (1953; Fig. 1) in an Illinois tornado outbreak on

Corresponding author address: Dr. Paul Markowski, 503 WalkerBuilding, University Park, PA 16802.E-mail: [email protected]

9 April 1953, although van Tassell (1955) is given creditfor coining the term. The reflectivity appendage usuallyis oriented roughly perpendicular to storm motion. Hookechoes are typically downward extensions of the rearside of an elevated reflectivity region (Forbes 1981)called the echo overhang (Browning 1964; Marwitz1972a; Lemon 1982), with the region beneath the echooverhang termed a weak echo region (Chisholm 1973;Lemon 1977) or vault (Browning and Donaldson 1963;Browning 1964, 1965a). Browning and Donaldson(1963) and Browning (1965b) noted that the southernedge of the hook formed a wall of echo ‘‘which wasoften very sharp and sometimes rather upright.’’ Hookechoes are typically several kilometers in length andseveral hundred meters in width, at least as viewed byoperational radars (e.g., Garrett and Rockney 1962). Avariety of shapes that the hook echo may take werepresented by Fujita (1973; Fig. 2).

Fujita (1958a) documented hook echoes associatedwith other supercells on the same day Stout and Huffmade their observations. He inferred the concept ofthunderstorm rotation from viewing the evolution of thehook echoes, which he studied in unprecedented (andsince unparalleled) detail (Fig. 3). Brooks (1949) earlierhad referred to these circulations, having radii of ap-proximately 8–16 km, as tornado cyclones.1 Wind ve-locity data obtained following the installation of Dopp-ler radars in central Oklahoma in the late 1960s con-firmed an association between hook echoes and stronghorizontal shear zones associated with storm rotationand tornadoes (e.g., Donaldson 1970; Brown et al. 1973;

1 The tornado cyclone terminology now typically refers to distinctcirculations having a scale larger than a tornado but smaller than amesocyclone (e.g., Agee 1976).

APRIL 2002 853M A R K O W S K I

FIG. 1. Radar image from the first documentation of a hook echo.The hook echo was associated with a tornadic supercell near Cham-paign, IL, on 9 Apr 1953. [From Stout and Huff (1953).]

Lemon et al. 1975; Ray et al. 1975; Ray 1976; Brandes1977a; Burgess et al. 1977; Lemon 1977; Barnes1978a,b).

Garrett and Rockney (1962) were the first to relate acircular echo on the tip of a hook echo to the tornadoor tornado cyclone. They called this ball-shaped echoan ‘‘asc’’ (annular section of the cylinder of the vortex),but the authors did not offer an explanation for the exactcause of the appearance of the asc. Stout and Huff(1953) also observed a similar feature, but little wasdiscussed of it. Donaldson (1970) noted an echo holein the tornado he studied, and found that it was collo-cated with a tornado vortex. Forbes (1981) also ob-served similar reflectivity features during the tornadooutbreak of 3–4 April 1974, as did Fujita and Wakimoto(1982) in their study of the Grand Island, Nebraska,tornadoes of 3 June 1980.

Van Tassell’s (1955) images of a hook echo nearScottsbluff, Nebraska, on 27 June 1955 (it moved di-rectly over the radar) suggested the presence of a faintanticyclonic protrusion from the tip of the hook, ex-tending outward in a direction opposite that of the cy-clonic protrusion. An anticyclonic reflectivity flare alsohas been documented by Brandes (1981), Fujita (1981),and Fujita and Wakimoto (1982). Multiple Doppler ra-dar wind syntheses almost invariably have revealed aregion of anticyclonic vorticity on the opposite side ofthe hook echo as the more prominent (cyclonic) vorticityregion (Brandes 1977b, 1978, 1981, 1984a; Ray 1976;Ray et al. 1975, 1981; Heymsfield 1978; Klemp et al.1981; Fig. 4). It is perhaps surprising that the ubiquityof the vorticity couplet straddling the hook echo largelyhas been ignored, with the exception of Fujita and hiscollaborators. Fujita and Wakimoto (1982) documentedan anticyclonic tornado within the region of anticyclonicvertical vorticity (Fig. 5). The cyclonic member of thevorticity couplet also was associated with a tornado.

Fujita (1981) proclaimed ‘‘Mesoscale modelers shouldbe attracted by such a pair of cyclonic and anticyclonictorandoes which were evidenced in the Grand Islandstorm on 3 June 1980 and in the central Iowa storm on13 June 1977.’’ However, it was unclear why such at-tention should be given to the vortex couplet, and theorigin of the couplet was not well understood.

b. Formation

Fujita (1958a) originally attributed hook echo for-mation to the advection of precipitation from the rearof the main echo around the region of rotation associatedwith the tornado cyclone and updraft. Browning (1964,1965b) also documented hook echoes and attributedtheir evolution (Fig. 6) to essentially the same processdescribed by Fujita (1958a). Fujita (1965) later attri-buted hook echo formation to the Magnus force. Heexplained that this force pulled the spiraling updraft outof the main echo, resulting in the hook-shaped reflec-tivity appendage commonly observed on radar displays(Fig. 7).

Fulks (1962) hypothesized that hook echo formationwas due to a large convective tower extending into thelevels of strong vertical wind shear, which producedcyclonic and anticyclonic flows at opposite ends of thetower—the cyclonic flow to the southwest gave rise tohook echo development. No mention was made of thepossibility of an anticyclonic hook echo forming on thenorth side of the tower from the same mechanism.

Probably no one presented as many detailed Dopplerradar analyses of supercells as Brandes [1977a,b, 1978,1981, 1984a,b; Brandes et al. (1988)]. Brandes (1977a)looked at a nontornadic supercell on 6 June 1974. Hookecho formation was attributed to the ‘‘horizontal ac-celeration of (low-level) droplet-laden air’’ as the down-drafts intensified and the outflow interacted with theinward-spiraling updraft air. Apparently this hypothesiswas essentially that precipitation advection was respon-sible for hook echo formation, similar to the Fujita(1958a) and Browning (1964, 1965b) hypotheses. Athree-dimensional numerical simulation by Klemp et al.(1981) of the Del City, Oklahoma (20 May 1977) su-percell also suggested that the horizontal advection ofprecipitation was important for hook echo development.In some observations of hook echoes associated withtornadoes in nonsupercell storms, the hook echoes alsohave appeared to result largely from the horizontal ad-vection of hydrometeors (e.g., Carbone 1983; Robertsand Wilson 1995).

Other reliable radar observations have been made thatsuggest that hook echo formation, in at least some cases,appears to result from the descent of a rain curtain inthe rear-flank downdraft (e.g., Forbes 1981; E. N. Ras-mussen 2000, personal communication;2 L. Lemon

2 This was an oral presentation at the VORTEX Symposium inLong Beach, California.


TABLE 1. Summary of findings pertaining to hook echoes and RFDs.

Observation/Conclusion References

RFD originated at or above 7 km Nelson (1977), Lemon et al. (1978), Barnes (1978a), Lemon andDoswell (1979)

RFD originated below 7 km Klemp et al. (1981)Low uw at surface in RFD van Tassell (1955), Beebe (1959), Ward (1961), Browning and

Ludlam (1962), Browning and Donaldson (1963), Charba andSasaki (1971), Lemon (1976a), Nelson (1977), Brandes (1977a),Barnes (1978a,b), Klemp et al. (1981), Klemp and Rotunno(1983), Rotunno and Klemp (1985), Wicker and Wilhelmson(1995), Dowell and Bluestein (1997), Adlerman et al. (1999)

Warm air (but not necessarily high uw) at surface in RFD Tepper and Eggert (1956), Garrett and Rockney (1962), Williams(1963), Fujita et al. (1977), Brown and Knupp (1980), Bluestein(1983), Brandes (1984a), Johnson et al. (1987), Rasmussen andStraka (1996)

Hypothesized that the occlusion downdraft is driven by a down-ward-directed, nonhydrostatic pressure gradient arising from theintensification of low-level vertical vorticity

Klemp and Rotunno (1983), Brandes (1984a,b), Hane and Ray(1985), Rotunno (1986), Brandes et al. (1988), Trapp and Fied-ler (1995), Wicker and Wilhelmson (1995), Wakimoto et al.(1998), Adlerman et al. (1999), Wakimoto and Cai (2000)

Hypothesized the RFD is forced mainly thermodynamically fromaloft (which may result from stagnation)

Browning and Ludlam (1962), Browning and Donaldson (1963),Browning (1964), Nelson (1977), Barnes (1978a), Brandes(1981), Klemp et al. (1981)

Hypothesized the RFD is initiated by dynamic pressure excessaloft but maintained thermodynamically

Bonesteele and Lin (1978), Lemon and Doswell (1979)

Reflectivity gradients found on the upshear side of storms Nelson (1977), Bonesteele and Lin (1978), Barnes (1978a), Forbes(1981)

Tornadogenesis observed before hook formation Garrett and Rockney (1962), Sadowski (1969), Forbes (1975)Tornadogenesis observed at the time of overshooting top collapse Fujita (1973), Lemon and Burgess (1976), Burgess et al. (1977)Visual observations of clear slots accompanying tornadoes Beebe (1959), Garrett and Rockney (1962), Moller et al. (1974),

Peterson (1976), Stanford (1977), Burgess et al. (1977), Lemonand Doswell (1979), Marshall and Rasmussen (1982), Rasmus-sen et al. (1982), Jensen et al. (1983), Wakimoto and Liu (1998)

Vertical vorticity couplets associated with hook echoes Ray (1976), Ray et al. (1975, 1981), Brandes (1977b, 1978, 1981,1984a), Heymsfield (1978), Klemp et al. (1981), Fujita (1981),Fujita and Wakimoto (1982), Klemp and Rotunno (1983), Bran-des et al. (1988), Wicker and Wilhelmson (1995), Wurman et al.(1996), Straka et al. (1996), Bluestein et al. (1997), Dowell andBluestein (1997), Blanchard and Straka (1998), Gaddy and Blue-stein (1998), Wakimoto and Liu (1998), Wakimoto et al. (1998),Wakimoto and Cai (2000), Wurman and Gill (2000), Ziegler etal. (2001), Bluestein and Gaddy (2001)

Anticyclonic reflectivity flares on hook echoes van Tassell (1955), Brandes (1981), Fujita (1981), Fujita and Wak-imoto (1982), Wurman et al. (1996), Blanchard and Straka(1998), Wurman and Gill (2000), Ziegler et al. (2001)

Hook echo formation attributed to rotation (but not necessarily thesame mechanisms)

Fujita (1958a, 1965), Fulks (1962), Browning (1964, 1965b), Bran-des (1977a)

Hook echoes associated with strong horizontal shears or tornadoes(prior to 1980, in the interest of brevity)

Stout and Huff (1953), van Tassell (1955), Fujita (1958a, 1965),Garrett and Rockney (1962), Browning (1964, 1965b), Freund(1966), Sadowski (1958, 1969), Donaldson (1970), Forbes(1975), Ray et al. (1975), Lemon et al. (1975), Ray (1976),Brown et al. (1978), Lemon (1977), Burgess et al. (1977), Bran-des (1977a), Barnes (1978a,b)

Hook echoes associated with downdrafts (prior to 1980, in the in-terest of brevity)

Browning and Donaldson (1963), Haglund (1969), Fujita (1973,1975b, 1979), Lemon et al. (1975), Lemon (1977), Brandes(1977a), Burgess et al. (1977)

Hook echoes located in strong vertical velocity and temperaturegradients, somewhat behind the surface windshift associatedwith the RFD

Marwitz (1972a,b), Burgess et al. (1977), Lemon and Doswell(1979), Brandes (1981)

Air parcels that enter the tornado pass through the RFD Brandes (1978), Davies-Jones and Brooks (1993), Wicker and Wil-helmson (1995), Dowell and Bluestein (1997), Adlerman et al.(1999) [and implied by visual observations of Lemon and Do-swell (1979), Rasmussen et al. (1982), Jensen et al. (1983)]

Hypothesized that the RFD is important for tornadogenesis Ludlam (1963), Fujita (1975b), Burgess et al. (1977), Barnes(1978a), Lemon and Doswell (1979), Brandes (1981), Davies-Jones (1982a,b)


FIG. 2. Fujita introduced five variations on the shapes of hookechoes. [From Fujita (1973).]

FIG. 3. Development of the hook echo associated with the Cham-paign, IL, tornado from 1724–1738 CST 9 Apr 1953, as analyzed byFujita. [From Fujita (1958a).]

2000, personal communication). It may be worth theeffort in future field experiments to use temporally andspatially high-resolution mobile radar data to system-atically investigate the extent to which the hook echoforms from the above process, versus being due to astreamer of precipitation that is extruded from the mainecho core by horizontal advection. The relative contri-bution to hook echo formation from these two processesis not yet known. It is entirely possible that one mightdominate in one case, while the other dominates in adifferent case. Or perhaps the relative contribution fromthe two processes could be a function of time within asingle case.

For completeness, it is noted that at least one docu-mentation has been made of hook echoes not associatedwith updraft rotation. Houze et al. (1993) showed ex-amples of hook-shaped (in a cyclonic sense, with thehooks pointing toward the right with respect to stormmotion) reflectivity structures in left-moving severestorms in Switzerland. These features, termed falsehooks by the authors, apparently were associated withthe cyclonic downdraft regions on the right (southern)flanks of the anticyclonically rotating storms, in whichthe updrafts would have been on the left (northern)flanks (Klemp and Wilhelmson 1978a; Wilhelmson andKlemp 1978; Rotunno and Klemp 1982).

c. Tornado forecasting based on hook echo detection

The forecasting potential of hook echo detection be-gan to be explored in the mid-1960s. Sadowski (1958)documented a tornado that occurred after the hook echobecame visible (in fact, the hook echo was becomingless discernible on radar at the time of the reportedtornado formation). Sadowski might have been the firstto speculate that if hook echoes generally preceded tor-nadogenesis, then it might be possible to issue tornadowarnings in advance.

In Stout and Huff’s (1953) report, the evidence hadbeen inconclusive as to whether the hook echo precededtornadogenesis, or vice versa. In van Tassell’s (1955)summary, it was not mentioned whether the hook echo

developed before or after tornadogenesis. The tornadostudied by Garrett and Rockney (1962) apparentlyformed before the hook echo became prominent, unlessa narrow hook echo went undetected by the WeatherSurveillance Radar-3 (WSR-3) (48 beamwidth) prior totornadogenesis. The tornado dissipated when the hook‘‘closed off’’ or merged with the forward-flank echo.

Sadowski (1969) later documented a large amount ofsuccess using hook echoes to detect tornadoes withinthunderstorms. In a 1953–66 study, he computed anaverage time of 15 min between hook echo appearanceand tornadogenesis in a sample of 13 cases in whichhook echoes appeared before tornadoes were reported.Sadowski reported a false alarm rate of only 12%. Onthe other hand, Freund (1966) found that only 6 of 13tornadic storms near the National Severe Storms Lab-oratory in 1964 were associated with hook echoes, andGolden (1974) found that only 10% of waterspouts wereassociated with hook echoes.

The so-called Super Outbreak of tornadoes on 3–4April 1974 (Fujita 1975a,b) provided a large sample ofa variety of ‘‘distinctive echoes’’ that were studied byForbes (1975, 1981). Forbes (1975) found that 1) a ma-jority of hook echoes were associated with tornadoes,2) hook echoes often were associated with tornado fam-ilies, and 3) tornadoes associated with hook echoes tend-ed to be stronger than those from other echoes. Forbes(1975) also found that, on average, hook echoes ap-peared 25 min prior to tornadogenesis; however, muchvariance was present in his sample—10 of 27 (37%) ofthe hook echoes associated with the first tornado pro-duced by a supercell were detected after the reported3

tornado formation times. Forbes (1981) found a falsealarm rate of just 16% when using hook echoes to detecttornadoes. But because hook echoes were relatively rare

3 The accuracy of the reported tornado times may be questionablefor some of the tornadoes studied.


FIG. 4. Three-dimensional wind field relative to the Del City, OK, supercell 1845 CST 20 May 1977 at 400 m. Left panel shows horizontalwind vectors with radar reflectivity superposed. The right panel depicts vertical velocity (m s 21), with the 30-dBZ contour accentuated onboth panels. The tornado path is stippled. [From Brandes (1981).]

(as he defined them), a less restrictive shape (a ‘‘dis-tinctive echo’’, e.g., appendages, line-echo wave pat-terns, etc.) also was considered. Distinctive echoes wereassociated with a probability of detection of tornadoesof 65%. Forbes (1975, 1981) did raise concern aboutthe generality of his findings, since his statistics werebased on the events of a single day. Also, the statististicswere based on Weather Surveillance Radar-1957 (WSR-57) data, which may not have adequately resolved fi-nescale echo structures.

3. Rear-flank downdrafts

a. Association with hook echoes

Rear-flank downdrafts (RFDs) are regions of subsid-ing air that develop on the rear side of the main updraftof supercell storms, and these regions of descent havea well-established association with hook echoes. Thefirst documentation of an RFD, although not recognizedas such, probably was by van Tassell (1955). In thatcase study and in another by Beebe (1959) on the samestorm complex, three ‘‘reliable’’ reports of severe down-drafts on the south side of the Scottsbluff tornado (27June 1955) were made.

Browning and Ludlam (1962) and Browning andDonaldson (1963) also were among the first to mentionthe presence of a downdraft in the vicinity of the stron-gest low-level rotation, behind the main storm updraft.Browning and Donaldson (1963) noted that the hookecho itself may be associated with this downdraft region.Haglund (1969), Fujita (1973, 1979), Lemon et al.

(1975), Burgess et al. (1977), Brandes (1977a), Lemon(1977), and Forbes (1981) also documented an associ-ation between hook echoes and downdrafts. Accordingto Forbes (1981), ‘‘the hook represents a band of pre-cipitation accompanied by downdraft and outflow, sur-rounding a weak echo region (a region of inflow andupdraft).’’ Brandes (1977a) tentatively concluded thatthe hook echo reflected downdraft intensification. Hag-lund (1969) concluded that the hook echo slightly trailsthe surface wind shift associated with the outflow of theRFD, and that the hook echo is located near the bound-ary between updraft and downdraft. Surface analysesand aircraft penetrations have revealed that the hookecho is located in a region of large vertical velocity andtemperature gradients (Burgess et al. 1977; Marwitz1972a,b).

b. Visual characteristics

The number of visual and surface observations ofsupercells increased during the 1970s, largely becauseof organized storm intercept programs at the NationalSevere Storms Laboratory (Golden and Morgan 1972;Davies-Jones 1986; Bluestein and Golden 1993). Manyof these observations have advanced our understandingof the basic structures associated with tornadoes andtheir parent storms.

Golden and Purcell (1978) photogrammetricallydocumented subsiding air on the south side of theUnion City, Oklahoma, tornado (24 May 1973), ap-parently a visual manifestation of the RFD and also


FIG. 5. Radar image and analysis from Grand Island, NE, 0208 UTC 4 Jun 1980 showing a vortex couplet straddlinga hook echo, with tornadic circulations associated with both members of the vortex couplet. [From Fujita (1981) andFujita and Wakimoto (1982).]

evidence that the tornado occurred in a strong verticalvelocity gradient. Moreover, a clear slot was seen towrap itself at least two-thirds of the way around thetornado. Other observations of clear slots, which areprobably always visual manifestations of subsiding airin an RFD,4 have been presented by Beebe (1959; thiswas probably the first documentation), Moller et al.(1974), Peterson (1976), Stanford (1977), Burgess etal. (1977), Lemon and Doswell (1979), Marshall andRasmussen (1982), Rasmussen et al. (1982), and Jen-sen et al. (1983) (Fig. 8).

Burgess et al. (1977) found that the clear slot couldbe associated with a hook echo: ‘‘Perhaps large dropletsare present in the downdraft and are brought down fromthe echo overhang, even though the air contains onlyragged clouds or is visibly cloudless at low levels. Ifso, since radar reflectivity is more strongly dependenton the size rather than on the number of droplets, radarmay show substantial echo in the ‘clear’ slot.’’ Analysisof the 2 June 1995 Dimmitt, Texas, tornadic supercellalso indicated an association between the hook echo andclear slot, based on photogrammetrically determinedcloud positions (E. N. Rasmussen 2000, personal com-munication).5

c. Surface characteristics

Direct observations within RFDs have been scarce.Tepper and Eggert (1956) appear to have been the first

4 The absence of a clear slot does not necessarily indicate an ab-sence of subsiding air.

5 This was an oral presentation at the VORTEX Symposium inLong Beach, California.

to systematically analyze traces of thermodynamic datanear tornadoes, and consequently, within some RFDs.Data were obtained within 25 km of tornadoes in morethan 50 cases. Many of the thermograph traces measuredonly minor fluctuations during the passage of the tor-nadoes and associated RFDs, and other traces revealedcooling and moistening near the tornadoes. Only a fewobservations were available within 5 km of the torna-does, however.

Fujita (1958b) inferred the presence of a surface highpressure annulus encircling the Fargo, North Dakota,tornado cyclone (20 June 1957) from pressure traces inthe vicinity of the tornadoes (Fig. 9). Although Fujitaspeculated that the high pressure was associated with aring of subsiding air around the tornado, he was unableto verify this speculation. [Ward (1964, 1972) and Snowet al. (1980) found high pressure rings surrounding lab-oratory and numerically simulated vortices, but it is notclear whether these are the same phenomena inferredby Fujita, which appeared to be of a slightly largerscale.] Surface pressure excesses within RFDs of up toa few millibars also have been documented by subse-quent investigators (e.g., Charba and Sasaki 1971; Lem-on 1976a; Bluestein 1983), although no one else hasdocumented a high pressure ring as Fujita did.

The studies of the Scottsbluff tornado by van Tassell(1955) and Beebe (1959) contain some of the first de-scriptions, albeit qualitative, of surface temperaturewithin an RFD at close range from a tornado. At leasta couple of observers, located a few hundred meterssouth of the tornado, reported that the downdrafts felt‘‘cold.’’ Browning and Ludlam (1962) and Browningand Donaldson (1963) also reported cold temperature


FIG. 6. Evolution of the hook echo in an Oklahoma supercell on26 May 1963 studied by Browning. [From Browning (1965b).]

FIG. 8. Photograph of a typical clear slot associated with an RFD(2 Jun 1995 at Dimmitt, TX; photograph by the author).

FIG. 7. Fujita once hypothesized that the Magnus force led to theformation of hook echoes. [From Fujita (1965).]

and equivalent potential temperature (ue) measurements(with respect to the inflow) in the wakes of the Wok-ingham, England (9 July 1959) and Geary, Oklahoma(4 May 1961) supercells, presumably within the RFDs.Ward (1961) observed ‘‘cooler northwest winds a couplemiles southwest of the (Geary) tornado.’’

In a supercell (25 May 1974) investigated by Nelson(1977), the lowest wet-bulb potential temperature (uw)values observed at the surface were within the RFD,where they were ;6 K lower than the ambient uw values.Complete separation of the forward-flank downdraft(FFD) and RFD was evidenced by separate temperatureminima, as Lemon (1974) earlier had found in anothercase. Additional observations of relatively low-ue andlow-uw air at the surface within RFDs have been pre-sented by Brandes (1977a; supercell on 6 June 1974;.3 K uw decrease), Barnes (1978a,b; supercell on 29April 1970; 2–3 K uw decrease directly behind the RFD

gust front and $5 K farther behind the gust front), Lem-on (1976a; supercell on 25 June 1969; ;6 K uw de-crease), and Charba and Sasaki (1971; supercell on 3April 1964; ;7 K ue decrease). The relatively low ue

and uw values have been shown to be compatible withenvironmental air anywhere from 1–5 km above theground (e.g., Lemon 1976a; Charba and Sasaki 1971).

In contrast to the findings summarized above, Garrettand Rockney (1962) reported that a warm downdraftwas observed about 12–15 km south of a tornado nearTopeka, Kansas, on 19 May 1960. The observer de-scribed the air as ‘‘suddenly becoming noticeably hot,similar to a blast of heat from a stove.’’ Williams (1963)showed that RFD air can arrive at the surface warmerthan the surrounding air. He noted that when such anevent occurs, it may be south of the hook echo or wher-ever forced descent is less likely to encounter sufficientliquid water to maintain negative buoyancy.

Fujita et al. (1977) documented a warm RFD (onlytemperature data were available) near an F4 tornado inthe Chicago, Illinois, area on 13 June 1976. On the sameday, Brown and Knupp (1980) found nearly constant uw

3 km east of the Jordan, Iowa, F3 and F5 tornado pair,and those observations probably were in the RFD airmass, based on the pressure trace, which measured apressure excess of a few millibars.

In his summary of Totable Tornado Observatory(TOTO) observations, Bluestein (1983) documented arelatively warm RFD and pressure rise of .2 mb in anontornadic supercell on 17 May 1981. Bluestein alsopresented evidence of a 1.5-K temperature rise in anRFD approximately 1.3 km south of the Cordell,Oklahoma, tornado on 22 May 1981.6 Similar to whatFujita (1958b) first inferred, Bluestein also showed datathat suggested high pressure at least partly encircling atornado (his Fig. 7). In the violent Binger, Oklahoma,tornado on 22 May 1981, only small temperature fluc-

6 The reported pressure and temperature fluctuations were with re-spect to the storm inflow environment.


FIG. 9. Pressure field near the center of the Fargo tornado cyclone (20 Jun 1957); A and B represent barographstations. [Adapted from Fujita (1958b).]

tuations (,1 K) were observed as the tornado passedwithin a few hundred meters north of TOTO (Fig. 10).

Klemp et al. (1981) referred to ‘‘cold downdraft(RFD) outflow’’ in the Del City supercell, but no evi-dence was presented demonstrating that this air actuallywas cold—retrieved temperatures by Brandes (1984a)and observations by Johnson et al. (1987) suggested thatat least parts of the Del City storm’s rear-flank outflowwere warm, although uw values may not have been aslarge as in the inflow.

Although there have been surface observations ofwarm, high-ue air within RFDs, three-dimensional nu-merical simulations of supercells almost invariably haveproduced cold, low-ue RFDs at the surface (e.g., Klempet al. 1981; Rotunno and Klemp 1985; Fig. 11). Thepioneering numerical modeling studies of supercellsconducted in the 1970s (e.g., Schlesinger 1975; Klempand Wilhelmson 1978a,b; Wilhelmson and Klemp 1978)and the parameter space studies of the 1980s (e.g., Weis-man and Klemp 1982, 1984), with no known exceptions,used warm rain microphysics. The relatively simple pa-rameterization was not computationally demanding;thus, experiments requiring large numbers of simula-tions were feasible. However, the exclusion of ice mayhave promoted unrealistically excessive amounts of la-tent cooling near updrafts. When ice physics is included,

hydrometeors are distributed over a larger horizontalregion, and the intensity of outflow in close proximityto the updraft is reduced (Gilmore and Wicker 1998;Rasmussen and Straka 1998). In some recent, unpub-lished simulations, relatively warm downdrafts havebeen produced at the surface when ice physics and arelatively fine spatial resolution (,250 m in the hori-zontal directions) were used (M. Gilmore 2000, personalcommunication).

d. Characteristics above the surface

Johnson et al. (1987) presented observations of theRFD and FFD of the Del City storm collected by a 444-m tower as the storm passed overhead. The RFD wasassociated with ue values approximately 4 K lower thanthe ambient conditions; however, the temperature in-creased 1.5 K and the dewpoint temperature decreased2.5 K—if ue was nearly conserved, then air had subsidedfrom approximately 1 km (all heights are above groundlevel). Although the RFD was not sampled well by thetower, the data that were available suggested the pres-ence of a downward-directed perturbation pressure gra-dient within the lowest half kilometer.

Dowell and Bluestein (1997) documented the passageof another tornadic supercell over the same instrumented


FIG. 10. Pressure, wind speed, wind direction, and temperaturetraces from TOTO near the Binger, OK, tornado on 22 May 1981from approximately 1922–1931 CST. Only relative wind direction isknown; veering (backing) is indicated by a downward (upward)change. The tornado passed approximately 600 m north of TOTO.Note the small (less than ;1 K) temperature fluctuations recorded.It is inferred that the hook echo and RFD associated with this violenttornado were relatively warm. [From Bluestein (1983).]

tower on 17 May 1981. They found large ue and uy

(virtual potential temperature) deficits in the portion ofthe RFD sampled by the tower (.12 and .5 K, re-spectively), which was within a region of fairly high(.40 dBZ) radar reflectivity west of the circulation cen-ter. The deficits were prominent over the entire heightof the tower.

With the exception of the direct observations pre-sented by Johnson et al. (1987) and Dowell and Blue-stein (1997), thermodynamic quantities above theground within RFDs have been obtained only by indirectmeans. Brandes (1984a) and Hane and Ray (1985) wereamong the first to use the methods proposed by Gal-Chen (1978) and Hane et al. (1981) to retrieve ther-modynamic fields in supercells from three-dimensionalwind fields synthesized from multiple Doppler radars.Brandes (1984a) retrieved the pressure and buoyancyfields in the Del City and Harrah, Oklahoma (8 June1974) tornadic storms. At 3.3 km on the rear side ofthe updraft of the Del City storm, relatively cold tem-

peratures were retrieved within the RFD—radar reflec-tivity was a minimum here, possibly implying that evap-oration was occurring. Behind the rear-flank gust frontin the eastern mesocyclone quadrants,7 evidence wasfound of warm temperatures at low levels during thetornadic stage. The relatively warm conditions were at-tributed to subsidence within the RFD. In the Harrahstorm, Brandes (1984a) also retrieved negative buoy-ancy in the RFD aloft, but no mention was made of thelow-level buoyancy immediately behind the gust frontin the eastern quadrants of the mesocyclone.

Hane and Ray (1985) also completed a thermody-namic retrieval for the Del City storm. In the pretornadicstage, the pressure distribution included at each level ahigh–low couplet across the updraft with the maximumhorizontal pressure gradient generally oriented along theenvironmental shear vector at that altitude, in agreementwith linear theory predictions (Rotunno and Klemp1982). Although the orientation of the horizontal pres-sure gradient agreed relatively well with linear theory,its magnitude did not agree as well. The authors statedthat possibly the orientation and magnitude of the en-vironmental shear vector were not known exactly (also,calculations of the horizontal vertical velocity gradientmay have had significant errors). In the tornadic stage,the pressure field contained a pronounced minimum atlow levels coincident with the mesocyclone, probablydue to strong low-level vertical vorticity. Hane and Rayfound weak high perturbation pressure (;1 mb) in theRFD at low levels behind the gust front during the timeof the tornado. Vertical gradients of nonhydrostatic pres-sure perturbations may be relevant to the formation andevolution of the RFD, as will be discussed in the nextsubsection. The RFD contained significant negativebuoyancy at low levels in Hane and Ray’s analysis (tem-perature deficits as low as 24.5 K; Fig. 12).

The retrieval results of Brandes (1984a) and Haneand Ray (1985) in the Del City storm were in relativelyclose agreement with the direct measurements withinthe inflow and FFD regions reported by Johnson et al.(1987). But Johnson et al. cautioned that ‘‘noticeabledifferences in the RFD region suggested that there wasroom for improvement in the retrieval methods.’’ Bran-des and Hane and Ray had used considerable smoothingon the buoyancy field to eliminate noise; the details inthe retrieved low-level buoyancy fields may have beensuspect.

e. Origins and formation

In reviewing previous studies and hypotheses per-taining to RFD formation, it may be helpful to begin

7 The RFD had wrapped cyclonically around the updraft, so thatdowndraft air also was found in the eastern (forward) quadrants ofthe mesocyclone. This ‘‘occlusion process’’ will be discussed in sec-tion 4.


FIG. 11. Contours at z 5 250 m of (a) potential temperature fluctuation ( 5 303 K), in intervals of 1 K; and (b)uequivalent potential temperature fluctuation ( e 5 340 K), in intervals of 2 K, in the simulation conducted by Rotunnouand Klemp (1985). The updraft is represented by the shaded region. The thick circular line encloses the region wherethe cloud water is greater than 0.1 g kg21 (wall cloud). The thick dashed line encloses the region where the cloudwater is greater than 0.1 g kg21 at z 5 500 m. [Adapted from Rotunno and Klemp (1985).]

FIG. 12. Horizontal distribution of buoyancy and vector horizontalwind at 1 km at 1847 CST 20 May 1977 in the Del City, OK, storm.Buoyancy (potential temperature fluctuation) has been filtered to re-move noise, and is contoured at 1.58C intervals (negative valuesdashed). Updraft maxima are denoted by J and vorticity maximaare denoted by !. [From Hane and Ray (1985).]

with an inspection of the inviscid vertical momentumequation written as

dw ]w ]p5 1 v · =w 5 2c u 1 B, (1)pdt ]t ]z

where dw/dt is the vertical acceleration following a par-cel, ]w/]t is the local vertical acceleration, v 5 (u, y, w)is the three-dimensional velocity vector, 2v · =w is theadvection of vertical velocity, cp is the heat capacity,

is the mean potential temperature, p is the perturbationuExner function, and B is the buoyancy which can bewritten as

u9B 5 g 1 0.61q9 2 q 2 q , (2)y l i1 2u

where u9 and are potential temperature and waterq9yvapor mixing ratio fluctuations from the base state, re-spectively, and ql and qi are liquid water (includes cloudwater and rain water) and ice mixing ratios, respectively.

By taking the divergence of (1) and assuming that=p ; 2p (a reasonable assumption for ‘‘well-be-haved’’ fields away from the boundaries), it can beshown (Rotunno and Klemp 1982) that


2 2 2]u ]y ]w 1

2 2p } 1 1 1 [|D| 2 |h| ]1 2 1 2 1 2[ ]]x ]y ]z 2

]B1 , (3)

]z

where | D | and | h | are the magnitudes of the total de-formation and vorticity, respectively. The [(]u/]x)2 1 (]y/]y)2 1 (]w/]z)2] terms are referred to as the fluid exten-sion terms. If (3) is linearized about a base state con-taining vertical wind shear (primed velocity componentsrepresent fluctuations from the base state, which is givenby (z) 5 [ (z), (z), 0]), it may be rewritten asv u y

2 2 2]u9 ]y9 ]w9

p } 1 11 2 1 2 1 2[ ]]x ]y ]z

]y9 ]u9 ]w9 ]u9 ]w9 ]y91 2 1 11 2]x ]y ]x ]z ]y ]z

]v ]B1 2 · =w9 2 (4)

]z ]z

5 p 1 p 1 p (5)nl l b

5 p 1 p , (6)dn b

where pnl, pl, and pb are the contributions to p fromnonlinear, linear, and buoyancy effects, respectively,

2 2 2]u9 ]y9 ]w9

p } 1 1nl 1 2 1 2 1 2[ ]]x ]y ]z

]y9 ]u9 ]w9 ]u9 ]w9 ]y91 2 1 1 (7)1 2]x ]y ]x ]z ]y ]z

]vp } 2 · =w9 (8)l ]z

]Bp } 2 , (9)b ]z

and pdn collectively refers to the nonlinear and lineareffects as ‘‘dynamic’’ effects on p. Equation (7) indi-cates that nonlinear dynamic high (low) pressure per-turbations are associated with convergence and diver-gence and deformation (rotation). Equation (8) indicatesthat linear dynamic high (low) pressure perturbationsare located upshear (downshear) of an updraft. Equation(9) indicates that high (low) pressure perturbations dueto buoyancy are located above (below) the level of max-imum buoyancy.

Using (5) and (6), the vertical momentum equationcan be written as

]w ]p ]pnl l1 v · =w 5 2c u 2 c up p1 2]t ]z ]z

]pb1 2c u 1 B (10)p1 2]z

]p ]pdn b5 2c u 1 2c u 1 B , (11)p p1 2]z ]z

where 2cp ]pdn/]z sometimes is referred to as the dy-unamic forcing and (2cp ]pb/]z 1 B) sometimes is re-uferred to as the buoyancy forcing. If the vertical gra-dients of the fluid extension terms are neglected, alongwith the vertical gradients of the deformation and hor-izontal vorticity, then it can be shown that

2]p ]znl } 2 (12)]z ]z

]p ] ]vl } · =w (13)1 2]z ]z ]z2]p ] Bb } 2 , (14)

2]z ]z

where z is the vertical vorticity.From (11) it is evident that descent can arise owing

to negative buoyancy, which can be generated accordingto (2) by cold anomalies produced by evaporative cool-ing or hail melting, or by precipitation loading, and byvertical perturbation pressure gradients that can arisefrom, according to (12)–(14), vertical gradients of ver-tical vorticity, ‘‘stagnation’’ of environmental flow atan updraft,8 and pressure perturbations due to verticalbuoyancy variations (which are partially due to hydro-static effects), respectively.9 Research presented in thepast 40 years has found that all of the terms in (11) canbe significant.

Browning and Ludlam (1962) and Browning andDonaldson (1963) suggested that the RFDs in the Wok-ingham and Geary supercells might have been drivenby negative buoyancy (i.e., ‘‘thermodynamically’’forced) due to evaporation. Browning (1964) surmisedthat the rightward propagation of supercells increased

8 Note that the use of the term stagnation here does not imply thatsupercell updraft are solid obstacles, as as been suggested by severalinvestigators in the past (e.g., Newton and Newton 1959; Fujita 1965;Fujita and Grandoso 1968; Alberty 1969; Fankhauser 1971; Charbaand Sasaki 1971; Brown 1992). Theoretical studies have exposedserious weaknesses in the obstacle analogy (e.g., Rotunno 1981; Ro-tunno and Klemp 1982; Davies-Jones et al. 1994), and these studiesalso have shown that the pressure distribution around an updraft isnot what would be expected if the updraft was behaving as an ob-stacle, except at the storm top (Davies-Jones 1985). Some studieshave shown that updrafts occasionally can display behavior that ap-pears similar to how a solid obstacle might be expected to behave(e.g., Lemon 1976b; Klemp et al. 1981).

9 For additional discussion of downdraft forcings in terms of thevertical momentum equation, the reader is referred to the review byKnupp and Cotton (1985).


their midlevel storm-relative flow so as to increase evap-orative cooling, and ultimately aid in the genesis ofdowndrafts (both on the rear and forward storm flanks).These hypotheses were proposed at least partly becauseof findings by Browning and Ludlam (1962) and Brown-ing and Donaldson (1963) of low uw air in the wakesof the Wokingham and Geary storms, which apparentlyhad midlevel origins.

Brandes (1981) also concluded that RFDs are initiatedby the production of negative buoyancy aloft: ‘‘presum-ably the initiating downdraft (associated with the rear-flank gust front) is formed by precipitation falling fromthe sloping updraft . . . we suppose the intruding flowhas low uw, and when chilled by evaporation, becomesnegatively buoyant . . . because the entrained air pen-etrates well into the storm, evaporative cooling ratherthan perturbation pressure forces may initiate the down-draft.’’ Brandes (1984a) made a similar claim, based onretrieved buoyancy: ‘‘at 3.3 km, cool temperatures onthe southern fringe of the storm were suggestive of evap-orative cooling as environmental air mixed with stormair.’’

Klemp et al. (1981) attributed the RFD in the DelCity storm to water loading and evaporation based onprecipitation trajectories crudely approximated using es-timated terminal fall speeds. Moreover, midlevel flowapproaching the storm from the east flowed through theFFD—not through the RFD as Browning (1964) hadconceptualized. RFD air at the surface appeared to havecome from 1–2 km above the ground, directly behindthe gust front, based on trajectory analyses in their nu-merical simulation and observations of the storm. Airfrom higher levels reached the surface further behindthe storm.

Nelson (1977) found an erosion of the hydrometeorfield at and below 7 km, as well as a sharp reflectivitygradient on the west flank of an Oklahoma multicellstorm that evolved into a supercell on 25 May 1974—these radar observations were believed to have been amanifestation of RFD formation that apparently oc-curred at the start of the transition from multicell tosupercell. Nelson noted two mechanisms suggestive ofRFD formation—evaporative cooling and/or dynamicpressure perturbations (presumably he was referring tothose related to linear effects, i.e., stagnation). Nelsonbelieved that the evaporation-driven effect was morelikely because of the echo erosion aloft; he also citedstrong storm-relative winds (;16 m s21 in the 7–9 kmlayer) and a large dewpoint depression (;21 K) at thelevel of apparent RFD formation. Forbes (1981) foundsimilar radar signatures suggesting echo erosion andRFD formation during the 3–4 April 1974 tornado out-break.

Lemon et al. (1978) and Barnes (1978a) also con-cluded that the RFD forms at middle to upper levels.Lemon et al. based their findings on an analysis of theUnion City tornadic supercell. They analyzed a persis-tent difluent flow region in the 7–10 km layer northwest

(upshear) of the mesocyclone that they believed wasassociated with a downdraft. Barnes’ conclusion thatthe RFD formed between 6.0–7.5 km was based on hisstudy of tornadic storms in Oklahoma on 29–30 April1970. He surmised that the storm-relative midlevel flow(20–25 m s21) approaching the cyclonically rotatingupdraft was decelerated and deflected on the upwind(south) side while the relative upwind stagnation pointshifted to the left of the intercepting wind vector; thatis, toward the southwest flank. Here ‘‘stagnating’’ airexperienced the longest contact with the adjacent up-draft while mixing only slightly with it—both cloud andsmall precipitation drops chilled this air by evaporationand began its downward acceleration before saturationcould occur. Barnes added ‘‘We emphasize that the highhorizontal momentum and proximity to the updraftmake the RFD a potentially important interactant withthe gust front and updraft’s surface roots . . . We alsonote that the location and extent of such a downdraftprobably depends upon the ambient flow relative to thestorm, which very likely requires a specific verticalshear profile to place it on the rear flank of a stormwhere it attains an influential position.’’ Barnes inter-preted the large reflectivity gradient on the midlevelupwind (southwest) flank as indicating dry ambient airadjacent to a precipitation-laden updraft. Bonesteele andLin (1978) made a similar inference.

Lemon and Doswell (1979) developed a conceptualmodel of a supercell from an extensive compilation ofsurface, visual, and radar observations (Fig. 13). Thismodel included an FFD and RFD, a surface gust frontstructure resembling a midlatitude cyclone, a hook-shaped reflectivity region surrounding a cyclonically ro-tating updraft, and a tornado, if present, that residedwithin the vertical velocity gradient between the updraftand RFD. This model has undergone little modificationsince its presentation over 20 years ago. Based largelyon the work by Barnes (1978a,b), Lemon and Doswellinferred that the RFD typically originated between 7–10 km on the relative upwind side of the updraft [notethat they did not say upshear side; refer to (8); Rotunnoand Klemp (1982) showed that the linear forcing forpressure fluctuations depends on the vertical shear, andnumerical results confirmed this theoretical prediction,as did some later dual-Doppler radar findings (e.g., Haneand Ray 1985)]. The authors cited the observation ofan echo-free hole at 7.5 km, directly above a notchbehind the low-level hook echo—they believed this tobe the signature of the RFD. Lemon and Doswell pro-posed that storm-relative inflow impingement was theRFD source, because Darkow and McCann (1977)showed that the relative flow at these levels is muchstronger than the storm-relative flow minimum theyfound at 4 km, and because of the Barnes (1978a,b) andNelson (1977) observations. Lemon and Doswell alsohypothesized that the RFD initially is dynamicallyforced, and then enhanced and maintained by precipi-tation drag and evaporative cooling.


FIG. 13. Conceptual model of a tornadic supercell at the surfacebased on observations and radar studies. Thick line encompassesradar echo. The thunderstorm gust front structure and occluded wavealso are depicted using a solid line and frontal symbols. Surfacepositions of the updraft (UD) are finely stippled, forward-flank down-draft (FFD) and rear-flank downdraft (RFD) are coarsely stippled.Associated streamlines are also shown. Tornado location is shown byan encircled T. [From Lemon and Doswell (1979).]

Brandes (1984a) retrieved excess pressure aloft onthe rear of the Del City storm, which may have sug-gested that the RFD was partly forced by a downward-directed dynamic pressure gradient, as Lemon and Do-swell had proposed, but the vertical pressure gradientsin the stagnation region could not be examined due toa paucity of scatterers and corresponding lack of radarvelocity data. (A lack of data due to the pristine aircommon in RFD regions probably still is one of thebiggest obstacles in understanding the formation mech-anisms and the role of the RFD today.)

Klemp et al. (1981) simulated a supercell with a com-posite sounding derived from three ‘‘proximity’’ sound-ings on 20 May 1977, and compared the simulated stormcharacteristics to those observed in the Del City storm.‘‘Trajectory’’ analysis (these were not true trajectories,but rather streamlines—if the storm was assumed to bequasi-steady, then the streamlines would be similar totrajectories) in the simulated supercell showed obstacle-like flow at 7–10 km. Parcels at 7 km that impingedupon the upshear side of the updraft did not appear tosink [in contrast to observations made in different casesby Barnes (1978a), Nelson (1977), and Lemon and Do-swell (1979)], but those at 4 km did; that is, the RFDapparently was 4–7 km deep (parcels from 4–7 km didnot reach the surface, but negative vertical velocitiesextended to 4–7 km). Directly behind the gust front,RFD air appeared to come from 1–2 km aloft; farther

behind the gust front, air from higher levels reached thesurface.

For the sake of completeness, it might be worth men-tioning that Shapiro and Markowski (1999) recently in-vestigated the formation of downdrafts in simple two-layer vortices using an analytic model.10 The applica-bility of the idealized model to real atmospheric vor-tices, in which buoyancy, buoyancy gradients,precipitation, and asymmetries probably are important,is questionable. Their results demonstrated how the‘‘vortex valve’’ effect (Lemon et al. 1975; Davies-Jones1986) can transport vorticity from the top of a homo-geneous, axisymmetric, rotating fluid to low levels viaan annular downdraft and secondary circulation, whenthe top layer of fluid rotates with an angular velocitylarger than that of the bottom layer of fluid.

Prior to 1983, investigators sought forcing for theRFD from middle and upper levels, as has been re-viewed in this section. But downdraft forcing also canarise at low levels. This is the subject of the next section.

4. Occlusion downdrafts

a. Evolution as simulated in numerical models

Klemp and Rotunno (1983) investigated the transitionof a supercell into its tornadic phase through use of ahigh-resolution (250-m horizontal grid spacing) modelinitiated within the interior of the domain of the DelCity supercell simulation performed by Klemp et al.(1981). With the enhanced resolution, Klemp and Ro-tunno found that the low-level cyclonic vorticity in-creased dramatically and the gust front rapidly occludedas small-scale downdrafts developed in the vicinity ofthe low-level circulation center. They concluded that theintensification of the RFD during the occlusion processwas dynamically driven by the strong low-level circu-lation, that is, by way of a dynamic perturbation pressuregradient such as that given by (12). This was the firststudy to propose such a mechanism for downdraft gen-esis and intensification. Later, Brandes (1984a,b), Haneand Ray (1985), and Brandes et al. (1988; see section3) made the same conclusion based on Doppler radaranalyses of tornadic storms, as did Trapp and Fiedler(1995), Wicker and Wilhelmson (1995), and Adlermanet al. (1999) based on numerical simulations.

Klemp and Rotunno (1983) defined the RFD as thedowndraft ‘‘which supports the storm outflow behindthe convergence line on the right flank.’’ They statedthat since nontornadic storms often were observed topersist for long periods of time with a well-defined gustfront, these storm-scale downdrafts were not uniquelylinked to tornadogenesis within a storm. On the otherhand, noted Klemp and Rotunno, if a storm progressedinto a tornadic phase, the gust front became occluded

10 Idealized three-layer vortices also were investigated using a nu-merical model.


FIG. 14. Low-level flow field (z 5 1 km) in Klemp and Rotunno’s(1983) nested, 250-m horizontal resolution simulation of the Del City,OK, supercell (20 May 1977) at the time that an occlusion downdraftwas observed. The region in which the cloud water exceeds 0.4 g kg21

is shaded. The heavy black line encloses the region where the rainwater mixing ratio exceeds 0.5 g kg21. The circulation center isindicated with the black dot. Vertical velocity is contoured at 5 m s21

intervals. What has been referred to as the occlusion downdraft islocated within a broader region of negative vertical velocities, whichhas been referred to as the rear-flank downdraft. [Adapted from Klempand Rotunno (1983).]

and a strong downdraft formed directly behind the gustfront at low levels and also might divide the updraft atmidlevels. Klemp and Rotunno referred to this smaller-scale downdraft as the occlusion downdraft. The occlu-sion downdraft, which was associated with a local ver-tical velocity minimum, was situated within a broaderregion of negative vertical velocity, which was referredto as the RFD; that is, the downdraft region on the rearside of the storm, which partially wrapped around thecirculation center at the time when low-level verticalvorticity reached its peak, was a single, contiguous en-tity (Fig. 14). The occlusion downdraft also was locatedwithin a larger, contiguous downdraft region in the high-resolution simulations of Wicker and Wilhelmson(1995; their Figs. 5–9).

Klemp and Rotunno proposed that the occlusion pro-cess and its associated occlusion downdraft were dy-namically induced by the strong near-ground rotationthat evolved along the convergence line. The rotationinduced low pressure coincident with the center of cir-culation and dynamically forced air down from above—the downdraft formed first at low levels and then ex-tended upward as the flow adjusted to the nonhydrostaticvertical pressure gradient force. Rotunno (1986) alsohypothesized that the occlusion downdraft was initiatedby the explosive growth of vertical vorticity at low lev-els.

The finding of Klemp and Rotunno that the occlusiondowndraft is driven by low-level rotation sometimes hasbeen implied as being in conflict (e.g., Carbone 1983;Brandes 1984a; Klemp 1987) with their predecessors’early hypotheses that the RFD is driven from aloft ther-modynamically or dynamically and is responsible forincreasing low-level rotation (e.g., Fujita 1975b; Bur-gess et al. 1977; Barnes 1978a; Lemon and Doswell1979). However, it is the author’s opinion that the ap-parent conflict may be one of semantics. If the occlusiondowndraft and RFD are to be viewed as two distinctentities, as proposed by Klemp and Rotunno, then theformation mechanisms and roles of the occlusion down-draft and RFD should not be anticipated to be neces-sarily identical. Observations of RFD formation pre-ceding the increase of vorticity near the ground are plen-tiful (e.g., Barnes 1978a; Lemon and Doswell 1979;Brandes 1984a,b). Once a downdraft forms, the distri-bution of vortex lines invariably must be affected ac-cording to Helmholtz’s theorem, and feedbacks on thedowndraft by the new vorticity distribution would beprobable (e.g., when rotation near the ground becomessubstantial, a downward-directed vertical pressure gra-dient could become established, accelerating air towardthe ground). Early hypotheses that the RFD is respon-sible for bringing rotation to low levels never assertedthat once low-level rotation began to intensify that dy-namic effects could not feed back on the downdraft.Therefore, it is believed that the occlusion downdraftcan be viewed as a rapid, small-scale intensification ofthe RFD that occurs after the RFD transports largerangular momentum air toward the ground. Stated an-other way, the occlusion downdraft may be viewed asa by-product of the near-ground vorticity increase, andvorticity cannot become large next to the ground withoutthe RFD (Davies-Jones 1982a,b; Davies-Jones andBrooks 1993; Brooks et al. 1993, 1994; Wicker andWilhelmson 1995); that is, the RFD is ultimately nec-essary for occlusion downdraft formation. Perhaps notsurprisingly, observations have not been made in su-percell storms of an occlusion downdraft preceding oroccurring in the absence of an RFD, or even at a locationnot within an RFD.11

It might be tempting to argue that the RFD and oc-clusion downdraft should be considered separate down-

11 In some nonsupercell tornadoes in which preexisting verticalvorticity is present at the surface, an RFD is not needed to transportcirculation to low levels in order for tornadogenesis to occur. It mightbe possible for the low-level vorticity amplification associated withthis tornadogenesis process to dynamically induce a downdraft similarto that which Klemp and Rotunno (1983) called an occlusion down-draft in a supercell storm; thus, it might be possible for an occlusiondowndraft to occur in the absence of an RFD in such a case. Carbone(1983) compared a downdraft he observed in a nonsupercell tornadoevent to the occlusion downdraft defined by Klemp and Rotunno.However, the dominant forcing for the downdraft was uncertain; thus,it is not known whether the downdraft documented by Carbone (1983)should be regarded as an occlusion downdraft, at least as defined byKlemp and Rotunno.


drafts because the dominant forcings are different. It isthe author’s opinion that a contiguity criterion shouldbe considered when debating whether two phenomenahaving different forcings are regarded as ‘‘separate.’’Otherwise, the updraft of a supercell should be viewedas two separate updrafts, with one updraft being drivenby nonhydrostatic pressure gradient forces below thelevel of free convection, and another updraft being driv-en largely by buoyancy forces above the level of freeconvection. The evolution put forth in the paragraphabove is not at odds with early proposals that the RFDis initiated at middle to upper levels and is responsiblefor initiating rotation near the ground, nor is it in conflictwith contentions that strong subsidence develops nearthe tornado during or after its formation.

It is speculated that the clear slot may be a visualmanifestation of an intensifying RFD or occlusiondowndraft. Updrafts also have been shown to weakenduring the stage when low-level rotation rapidly in-creases, probably also due to the formation of a down-ward-directed dynamic pressure gradient induced by therotation (e.g., Brandes 1984a,b). Fujita (1973), Lemonand Burgess (1976), and Burgess et al. (1977) havedocumented the collapse of overshooting storm topsnear the time of tornadogenesis, which presumably is amanifestation of updraft weakening.

It also should be noted that although the occlusiondowndraft in the Klemp and Rotunno (1983) simulationwas found to be driven by low-level vertical vorticityamplification, the occlusion downdraft did not descendalong the axis of low-level rotation. An explanation wasnot offered. One might expect that the vertical pressuregradient associated with the vertical gradient of verticalvorticity would lead to a maximum acceleration alongthe rotation axis. Two reasons might account for theasymmetry: 1) the dynamic vertical perturbation pres-sure gradient associated with the vertical gradient ofvertical vorticity squared (]z2/]z) does not contribute tovertical velocity directly, but rather to vertical accel-erations—thus, dw/dt might be a minimum in the vor-ticity maximum center, but if this occurs within the up-draft (where w k 0), then a downdraft (w , 0) mayfirst appear on the periphery of the updraft, away fromthe center of rotation, where w is less positive; and 2)other terms in the vertical momentum equation, whencombined with the dynamic vertical perturbation pres-sure gradient force, may force the strongest downwardacceleration away from the axis of largest vertical vor-ticity—for example, the buoyancy forcing may favorascent in the updraft center, so that the net effect of thebuoyancy forcing and dynamic vertical perturbationpressure gradient may lead to the strongest downwardacceleration on the updraft periphery. Superposition ofthe fields of the vertical momentum equation forcingterms in Klemp and Rotunno’s (1983) simulation leadsto the strongest downward acceleration being to thesoutheast of the maximum low-level rotation (Fig. 15).Thus, it is entirely possible for an occlusion downdraft

to be ‘‘driven’’ by low-level rotation even if the occlu-sion downdraft is not collocated with the low-level ro-tation.12

b. Observations

Brandes (1978; the Harrah storm) appears to havemade observations prior to the Klemp and Rotunno(1983) simulation of a downdraft not becoming prom-inent until after low-level rotation became substantial.Brandes (1981) also stated, following his analysis of theDel City storm, that ‘‘the sudden appearance of strongrear downdrafts in storms persisting for hours may alsorelate to the intensity and distribution of updrafts andvorticity.’’ Brandes (1984a) attributed sudden occlusiondowndraft formation in the Del City and Harrah stormsto the vertical pressure gradient owing to the explosivegrowth of low-level vorticity as Klemp and Rotunno(1983) found. Furthermore, Brandes’s data also showedthat the occlusion downdraft did not descend along theaxis of the strong low-level vorticity. Brandes (1984b)claimed that the occlusion downdraft formed after theincipient tornado had been detected, and roughly co-incided with tornado formation. Hane and Ray (1985)also documented occlusion downdraft formation in theDel City storm.

Based on their analyses of the Lahoma and Orienta,Oklahoma, supercells (2 May 1979), Brandes et al.(1988) hypothesized that because RFDs possess weakpositive or negative helicity (because couplets of ver-tical vorticity straddle RFDs), the decline of storm cir-culation might be hastened by turbulent dissipationwhen the downdraft air eventually mixes into supercellupdrafts. As did Brandes (1984a,b) and Klemp and Ro-tunno (1983), Brandes et al. claimed that ‘‘the updraftminimum in the Lahoma storm and RFD in the Orientastorm apparently owed their existence to the build-upof low-level vorticity and related downward verticalpressure gradients.’’ Large downward pressure forcesexisted within the RFD and left-hand portions of thepersistent updraft region in the Orienta storm, and tothe rear of the persistent updraft in the Lahoma storm.Brandes et al. (1988) probably presented the most com-prehensive analyses, discussion, and insight into thepressure distribution in supercells to date.

5. Role of RFDs in tornadogenesis

a. Observations-based hypotheses

Ludlam (1963) was one of the first to write that down-drafts, especially those located on the rear flank of su-percells, actually may be important to tornadogenesis:‘‘It is tempting to look for the spin of the tornado in

12 Wakimoto and Cai (2000) proposed that it also may be possiblefor an occlusion downdraft to reach the surface away from the centerof strongest near-ground vorticity if a mesocyclone is vertically tilted.


FIG. 15. (a)–(c) Cross sections of the forcing terms in the vertical momentum equation at t 5 2 min andz 5 250 m in the nested, 250-m horizontal resolution simulation by Klemp and Rotunno (1983). Contoursare drawn at 1022 m s22 intervals: (a) dynamically induced pressure gradient, (b) advection terms, and (c)buoyancy forcing. The black dot indicates the location of the circulation center. [From Klemp and Rotunno(1983).] (d) A composite of the forcing terms in the vertical momentum equation. The hook echo and regionswhere w exceeds 0 and 3 m s21 are shaded according to the legend. The region where downward localvertical velocity accelerations (]w/]t) are largest also is shaded. This region is located where the superpositionof the buoyancy forcing, advection, and dynamic pressure forcing leads to the strongest downward localvertical velocity changes (]w/]t K 0). The regions where the buoyancy forcing and advection of verticalvelocity are positive, and where the dynamic vertical pressure gradient is most negative (directed downward),also are indicated in the legend. Note that the largest downward acceleration occurs southeast of the low-level circulation center; a downdraft driven by increasing low-level rotation need not be collocated with theaxis of strongest rotation. [Adapted from Klemp and Rotunno (1983).]

the vorticity present in the general air stream as shearand tilted appropriately in the vicinity of the interfacebetween the up- and down-motions.’’ Fujita (1975b)also proposed that the downdrafts associated with hookechoes may be fundamentally critical to tornado for-mation, in terms of his ‘‘recycling hypothesis’’: 1)downdraft air is recirculated into the (developing) tor-

nado, 2) this process results in an appreciable conver-gence on the back side of the (developing) tornado, and3) the downward transport of the angular momentumby precipitation and the recycling of air into the tornadowill create a tangential acceleration required for the in-tensification of the tornado. Research conducted withthe aid of coherent radars in the ensuing years led others


FIG. 16. Brandes’s (1978) conceptual model of low-level meso-cyclone characteristics during the tornadic phase included an oc-cluded mesocyclone with air parcels from the RFD feeding the tor-nado. [Adapted from Brandes (1978).]

(e.g., Burgess et al. 1977; Barnes 1978a; Lemon andDoswell 1979; Brandes 1981) to make the same generalspeculation. Burgess et al. (1977) believed that the RFD,hook echo, and tornadogenesis were intimately con-nected: ‘‘The formation and evolution of the RFD isjudged extremely important to tornado formation. . .thesevere tornado [the Oklahoma City tornado of 8 June1974] appears related to the increased vorticity sourceprovided by presumed downdraft intensification andgust front acceleration along the right flank.’’ Forbes(1981) also discovered signatures (e.g., a sharp reflec-tivity gradient along the upshear side of the updraft, andoccasionally a small echo mass several kilometers tothe right of the right-rear edge of the main echo) sug-gesting RFD formation (1–10 min) prior to tornado-genesis.

Lemon and Doswell (1979) noted that just beforetornadogenesis, the mesocyclone center shifted fromnear the updraft center to the zone of high vertical ve-locity gradient. The early mesocyclone apparently wasa rotating updraft, whereas the transformed mesocy-clone had a divided structure, with strong cyclonicallycurved updrafts to the east in the ‘‘warm inflow sector’’and strong cyclonically curved downdrafts to the westin the ‘‘cold outflow sector.’’ And while the tornado wasapparently found in a strong vertical velocity gradient,Lemon and Doswell noted that it probably was locatedon the updraft side of that gradient.

Lemon and Doswell explained the evolution of theRFD and tornadogenesis as follows: 1) air deceleratesat the upwind stagnation point, is forced downward, andmixes with air below, which then reaches the surfacethrough evaporative cooling and precipitation drag; 2)the initially rotating updraft is then transformed into anew mesocyclone with a divided structure, in which thecirculation center lies along the zone separating the RFDfrom the updraft (this process appears to result, in part,from tilting of horizontal vorticity); and 3) ‘‘descent ofthe mesocyclone circulation occurs simultaneously(within the limits of temporal resolution) with the de-scent of the RFD.’’

Observations of low-level vorticity couplets withinRFDs that seem to straddle the hook echoes (introducedin section 2a) may be indications that tilting of vorticityby the RFD is important in the formation of tornadoeswithin supercell storms, as hypothesized by Davies-Jones (1982a,b) and Davies-Jones and Brooks (1993).Furthermore, it may be worth noting that during thetornadogenesis phase in supercells, the parcels of airthat enter the tornado or incipient tornado regularlyseem to pass through the hook echo and RFD (Brandes1978; Klemp et al. 1981; Dowell and Bluestein 1997;J. Wurman et al. 2000, personal communication;13 Fig.16), which may have been the basis for Fujita’s (1975b)recycling hypothesis. Visual observations of mesocy-

13 This was a poster presented at the 20th Severe Local StormsConference in Orlando, Florida.

clones being nearly totally occluded by the RFD, asevidenced by observations of the clear slot during andjust prior to the tornadic stage, such as those made byLemon and Doswell (1979), Rasmussen et al. (1982),and Jensen et al. (1983), also may imply that the airentering the tornado comes from the RFD. The possibleimportance of the clear slot also did not escape theattention of Ludlam (1963): ‘‘. . . often the funnel isphotographed spectacularly against a segment of brightand practically cloud-free sky beyond the edge of thearch cloud.’’14 In addition to the observational evidence,simulations also have indicated that air parcel trajec-tories pass through the RFD en route to intensifyingnear-ground circulations (Davies-Jones and Brooks1993; Wicker and Wilhelmson 1995; Adlerman et al.1999).

Davies-Jones (1998) recently has questioned whetherthe hook echo is really a passive indicator of a tornado,since close-range airborne and mobile radar observa-tions during recent field experiments have revealed hookecho formation prior to tornadogenesis in every case.15

Davies-Jones hypothesized that the hook echo may ac-tually instigate tornadogenesis, either baroclinically, byway of buoyancy gradients within the hook echo (Da-vies-Jones 2000a), or barotropically, by redistributingangular momentum (Davies-Jones 2000b).

14 The ‘‘arch cloud’’ Ludlam refers to was that ‘‘on the forwardright flank of severe storms, outside of the precipitation region;’’ thatis, that which is on the leading edge of the trailing gust front.

15 Previous studies documenting cases in which tornadoes werereported prior to hook echo detection (e.g., Garrett and Rockney 1962;Forbes 1975) relied on temporally and spatially coarser radar datacompared to what was available in field experiments such as VOR-TEX, and reported tornado times may not always have been accurateto within a few minutes.


FIG. 17. Schematic diagram showing how cyclonic vorticity maybe generated from tilting of baroclinic horizontal vorticity in a down-draft. In the case of streamwise vorticity with flow to the right of thehorizontal buoyancy gradient and a southerly shear component, acombination of tilting and baroclinic generation causes the vorticityof parcels to change from anticyclonic (denoted by a) to cyclonic(denoted by c) while still descending. [Adapted from Davies-Jonesand Brooks (1993).]

b. Theoretical considerations

Most of the theoretical and numerical modeling stud-ies pertaining to supercell storms have investigated thedevelopment of midlevel and low-level rotation by wayof tilting of horizontal vorticity (either associated withthe large-scale mean vertical wind shear or generatedsolenoidally by a baroclinic zone) by an updraft (e.g.,Rotunno 1981; Rotunno and Klemp 1982, 1985; Lilly1982, 1986a,b; Davies-Jones 1984). However, Davies-Jones (1982a,b) noted that in order to obtain large ver-tical vorticity at the ground in an environment in whichvortex lines are initially quasi-horizontal, a downdraftwould be necessary. Tornadoes may arise in the absenceof a downdraft in environments containing preexistingvertical vorticity at the surface, such as in some casesof ‘‘nonsupercell tornadogenesis’’ (e.g., Wilson 1986;Wakimoto and Wilson 1989; Roberts and Wilson 1995;Lee and Wilhelmson 1997a,b, 2000).

Davies-Jones (1982a,b) concluded that in a shearedenvironment with negligible background vertical vor-ticity, an ‘‘in, up, and out’’ circulation driven by forcesprimarily aloft would fail to produce vertical vorticityclose to the ground [this conclusion depends on eddiesbeing too weak to transport vertical vorticity downwardagainst the flow; this was verified by Rotunno andKlemp (1985) and Walko (1993)]. If a Beltrami modelis crudely assumed to represent the flow in a supercell(Davies-Jones and Brooks 1993), then vortex lines arecoincident with streamlines and parcels flowing into theupdraft at very low levels do not have significant verticalvorticity until they have ascended a few kilometers. Oth-erwise, argued Davies-Jones, abrupt upward turning ofstreamlines, strong pressure gradients, and large verticalvelocities would be required next to the ground.

Davies-Jones (1982a,b) neglected baroclinic vorticityand suggested that the downdraft had the following rolesin near-ground mesocyclogenesis: 1) tilting of horizon-tal vorticity by a downdraft produces vertical vorticity,2) subsidence transports air containing vertical vorticitycloser to the surface, 3) this air flows out from thedowndraft and enters the updraft where it is stretchedvertically, and 4) convergence beneath the updraft isenhanced by the outflow. Davies-Jones also showed ki-nematically that the flow responsible for tilting and con-centrating vortex lines also tilts and packs isentropicsurfaces, thus explaining observations of strong entropygradients across mesocyclones near the ground.

c. Potentially relevant simulation results

Davies-Jones and Brooks (1993) showed that the ver-tical vorticity of air parcels descending in an RFD canbe reversed during descent, from anticyclonic initially,to less anticyclonic, then to cyclonic in the lowest 50–125 m of their descent. As air subsided in the downdraft,vortex lines turned downward due to the barotropic‘‘frozen fluid lines’’ effect (Helmholtz’s theorem), but

with less inclination than the trajectories because hor-izontal southward vorticity was being generated contin-uously by baroclinity within the hook echo. Because ofthe geometry, vortex lines crossed the streamlines fromlower to higher ones (with respect to the ground), andthe barotropic effect served to turn the vortex lines up-ward even during descent. The baroclinic effect actedto increase horizontal vorticity further but did not con-trol the sign of the vertical vorticity; thus, air with cy-clonic vertical vorticity appeared close to the ground.As this air passed from the downdraft into the updraft,its cyclonic spin was amplified substantially by verticalstretching (Fig. 17).

Brooks et al. (1993, 1994) found that the formationof persistent near-ground rotation was sensitive to thestrength of the storm-relative midlevel winds. Whenstorm-relative midlevel flow was weak, RFD outflowundercut the updrafts and associated mesocyclones.When storm-relative midlevel flow was too strong, thecold pool was not oriented suitably for vorticity gen-eration within the baroclinic zone immediately behindthe updraft, which was found to be needed for the de-velopment of near-ground rotation in their simulations.

Wicker and Wilhelmson (1995) used a two-way in-teractive grid to study tornadogenesis. During a 40-minperiod, two tornadoes grew and decayed within the me-socyclone. Wicker and Wilhelmson’s Fig. 9 depicted aspiraling, asymmetric RFD associated with tornadoge-nesis. Their figure also indicated anticyclonic verticalvorticity on the opposite side of the RFD as the cyclonicvertical vorticity. Furthermore, the RFD contained lowue values ( was as small as 215 K; was approxi-u9 u9e e

mately 25 to 28 K in the hook echo).Wicker and Wilhelmson found that parcels entered

the mesocyclone from the RFD (Fig. 18) and descended


FIG. 18. Vertical velocity and trajectories at 100 min (tornadolikevortex present in simulated storm at this time) for z 5 100 m in the120-m resolution simulation by Wicker and Wilhelmson (1995). Thetrajectories entering the vortex have come from the hook echo andRFD region. [Adapted from Wicker and Wilhelmson (1995).]

from ;500 m, as Davies-Jones and Brooks (1993) hadfound. Furthermore, these parcels initially containednegative vertical vorticity; however, vertical vorticityincreased to only weakly negative values, not to largepositive values as in the simulations of Davies-Jonesand Brooks. Trajectories into the tornado from the RFDrevealed that positive vertical vorticity was acquiredonly after parcels began to ascend, not while they werestill descending (Davies-Jones and Brooks 1993; Trappand Fiedler 1995; Adlerman et al. 1999). It is believedthat these differences should be regarded as minor, forthere is a time tendency for increasing positive verticalvorticity in the parcels passing through the downdraftin all of the studies.

Though numerical models have been instrumental inadvancing our understanding of supercell dynamics,models may have limited utility in exploring the ther-modynamic characteristics of RFDs, and the potentialsensitivity of low-level vorticity intensification to thesecharacteristics, owing to the unavoidable parameteri-zation of microphysical processes. For example, three-dimensional simulations have not been able to producethe warm and moist RFDs that often have been observednear some strong tornadoes—the RFDs of simulatedsupercells almost invariably have large potential tem-perature and equivalent potential temperature deficits,as discussed in section 3c. However, some idealizedsimulations have suggested that tornadogenesis may befavorable if downdrafts are not too cold. Eskridge andDas (1976) proposed that a warm, unsaturated down-draft could be important for tornadogenesis; however,

they did not specify whether the downdraft also couldbe cold, nor what the advantages of a warm downdraftover a cold downdraft were. Davies-Jones (2000b) re-cently has shown that an annular rain curtain can trans-port sufficient angular momentum from aloft to theground to result in tornadogenesis in an idealized axi-symmetric numerical model.16 No evaporation was per-mitted in the model; thus, no cold downdraft air waspresent. Furthermore, Leslie and Smith (1978) foundthat some vortices could not establish contact with theground when low-level stable air was present, even ifvery shallow. Remarkably, Ludlam (1963) many yearsearlier had argued that ‘‘at least a proportion of the airthat ascends in the tornado must be derived from thecold outflow; if this contains the potentially cold airfrom middle levels its ascent might be expected soonto impede if not destroy the tornado . . . it may be par-ticularly important for the intensification and persistenceof a tornado that some of the downdraft air may bederived from potentially warm air which enters the leftflank of the storm at low-levels.’’

Significant advances in our understanding of super-cells no doubt have been made by numerical models. Itis probable that some conclusions drawn from simula-tion results never could have been made from obser-vations or theory alone. However, the author shares theview expressed by Doswell (1985): ‘‘The RFD’s roleremains confusing with respect to tornadogenesis. Trulyconfirming evidence about the various aspects of nu-merical simulations awaits better observations, despitethe compelling similarities between simulations and realstorms.’’

6. Recent observations from VORTEX

New radar and surface observations having unprec-edented spatial and temporal resolution have been ac-quired by VORTEX and smaller, post-VORTEX fieldexperiments. The findings of these recent operations thatare pertinent to this review are highlighted below, withthe exception of analyses of new surface data withinhook echoes and RFDs, which will be presented byMarkowski et al. (2002).

Recent radar observations, which include data ob-tained from both airborne and ground-based mobileDoppler radars (Jorgensen et al. 1983; Ray et al. 1985;Bluestein and Unruh 1989; Bluestein et al. 1995; Daugh-erty et al. 1996; Wurman et al. 1997), have resolvedstructures within hook echoes that were barely resolv-able or unresolvable in the early radar studies of su-percell storms. For example, in high-resolution (,100m spatially) data of tornadoes presented by Wurman et

16 This simulation had some similarities with those conducted byDas (1983, unpublished manuscript), in which a precipitation-drivendowndraft was imposed upon a wind field in which angular momen-tum increased with height. The downdraft was found to be able totransport sufficient vorticity from aloft to result in tornadogenesis.


FIG. 19. Radar (left) reflectivity (dBZ ) and (right) velocity fields associated with a hook echo and tornado on 2 Jun 1995 near Dimmitt,TX. In the left image, the arrow labeled N points toward the north. In the right image, the arrow labeled R points toward the radar position.[From Wurman and Gill (2000).]

al. (1996), Wurman and Gill (2000), and Bluestein andPazmany (2000), the echo-free holes that were only mar-ginally resolved by Garrett and Rockney (1962) werewell resolved (Figs. 19 and 20). The images presentedby Bluestein and Pazmany (2000) even begin to mar-ginally resolve structures, possibly subvortices, withinthe echo-free hole itself. Moreover, the radar reflectivitydepictions ‘‘looked like a tropical cyclone, with con-centric inner bands and outer spiral bands’’ (Bluesteinand Pazmany 2000). Hook echoes as narrow as 100 mor less have been detected (Wurman et al. 1996; Blue-stein and Pazmany 2000), perhaps implying that pre-cipitation loading and evaporative cooling within thehook echoes of some storms may not be the most sig-nificant effects in driving the associated RFDs, at leastat low levels.

Just as the early multiple-Doppler radar analyses ofthe 1970s and 1980s revealed, radar observations ob-tained in the last decade also have detected vorticitycouplets straddling the hook echoes of tornadic storms(Rasmussen and Straka 1996; Wurman et al. 1996;Straka et al. 1996; Bluestein et al. 1997; Dowell andBluestein 1997; Dowell et al. 1997; Wakimoto and Liu1998; Wakimoto et al. 1998; Wurman and Gill 2000;Ziegler et al. 2001) and nontornadic storms (Gaddy andBluestein 1998; Blanchard and Straka 1998; Wakimotoand Cai 2000; Bluestein and Gaddy 2001). These vor-ticity doublets could be evidence that RFDs are involvedin a downward displacement of initially quasi-horizontalvortex lines, perhaps necessarily transporting rotation

to the surface during tornadogenesis, as many previousinvestigators (e.g., Ludlam, Fujita) have conjectured.

Using surface observations obtained from automo-bile-borne sensors (Straka et al. 1996), Rasmussen andStraka (1996) documented a relatively warm RFD southof the Dimmitt, Texas, tornado. In the same case, whichwas during VORTEX, the hook echo was collocatedwith the surface divergence maximum, implying an as-sociation between the hook echo and (at least) a low-level downdraft, as also had been suggested by numer-ous predecessors.

In two other VORTEX storms, Wakimoto et al. (1998)and Wakimoto and Cai (2000) concluded that the oc-clusion downdraft was driven largely by the reversal ofthe vertical gradient of dynamic pressure, owing to in-creasing vorticity at low levels. In the supercell docu-mented by Wakimoto et al. (1998; the 16 May 1995VORTEX storm), it was found that the precipitation-loading forcing of the occlusion downdraft was an orderof magnitude less than the forcing provided by the non-hydrostatic vertical pressure gradient. [In a differentcase, Carbone (1983) previously had suggested that pre-cipitation loading may contribute to occlusion down-draft genesis.] In the 12 May 1995 VORTEX stormstudied by Wakimoto and Cai (2000), a thermodynamicretrieval indicated that the occlusion downdraft was as-sociated with a warm core.

Perhaps the most remarkable observational findingduring the last 10 years is that the differences betweentornadic and nontornadic supercells may be subtle, if


FIG. 20. Radar reflectivity fields (dBZ) associated with a hook echo on 15 May 1999 (left) early in the life of a tornado, (center) duringthe mature stage, and (right) during the dissipation stage. [From Bluestein and Pazmany (2000).]

even distinguishable, even in dual-Doppler radar anal-yses of the wind fields just prior to tornadogenesis. Blan-chard and Straka (1998) documented a mobile radarsignature of a spiraling hook echo in a nontornadic su-percell having an appearance similar to those that havebeen associated with tornadic supercells (e.g., similarto the radar image in Fig. 20). Perhaps such near-groundcirculations are more common in nontornadic supercellsthan previously believed. Trapp (1999) and Wakimotoand Cai (2000) also documented circulations in non-tornadic supercells at levels close to the ground. Trappfound that the low-level mesocyclones associated withtornadogenesis had smaller core radii and were asso-ciated with more substantial vorticity stretching thanthose associated with tornadogenesis ‘‘failure.’’ Basedon a comparison of pseudo-dual-Doppler analyses, Wak-imoto and Cai concluded that the ‘‘only difference be-tween the Garden City storm and Hays storm (the 16May 1995 and 12 May 1995 VORTEX storms) was themore extensive precipitation echoes behind the rear-flank gust front for the Hays storm.’’

7. Concluding remarks

The association between hook echoes, RFDs, and tor-nadogenesis has been well documented for nearly 50

years; however, the precise dynamical relationship stillis not known today. The analysis of the three-dimen-sional wind structure of supercells afforded by Dopplerradar, along with speedy increases in the feasibility ofnumerical cloud modeling, led to relatively rapid gainsin knowledge of the recurrent storm structures and evo-lution associated with supercells. Within 30 years of thefirst radar image of a hook echo, we knew that the mostdamaging tornadoes were associated with supercells, weknew about the existence of the parent circulations oftornadoes (mesocyclones), we developed an understand-ing of the dynamics of midlevel storm rotation and stormpropagation, radar and visual features common to su-percells were well documented, and downdrafts wererecognized as being important in tornadogenesis. Yet inthe decades that followed the period of rapid advances,no breakthroughs emerged with respect to the role ofthe RFD in tornadogenesis. In fact, it is debatable wheth-er we can better anticipate tornadogenesis within su-percell storms today than we could 20 years ago.

If the hook echo and its associated RFD truly arecritical to tornadogenesis, as hypothesized for manyyears, then perhaps significant gains in understandingwill not be possible until more spatially and temporallydetailed observations of this region can be made, inaddition to numerical simulations with more realistic


representations of entrainment and microphysical pro-cesses. It is believed that some of the important out-standing questions include

R What are the dominant forcings for RFDs, as a func-tion of location within the RFD and stage in stormevolution?

R How do the dominant RFD forcings vary across thespectrum of supercell types (e.g., nontornadic vs tor-nadic, low-precipation vs heavy-precipitationstorms)?

R How do the thermodynamic and microphysical char-acteristics of hook echoes and RFDs vary across thesupercell spectrum, and why?

R How does the large-scale environment affect RFDcharacteristics?

R Is the tornadogenesis process sensitive to the ther-modynamic and microphysical properties of RFDs?

R What is the role of the RFD in tornadogenesis, anddoes the hook echo have an active role?

A number of direct observations have been reviewedherein; however, these observations have been relativelyscarce and often have been simply fortuitous. A newmobile surface observing system (Straka et al. 1996),introduced in 1994 for VORTEX, recently has collectedthe largest number of in situ measurements within su-percell storms to date. Analyses of these spatially andtemporally dense ‘‘mobile mesonet’’ observations with-in hook echoes and RFDs will be presented in a com-panion paper (Markowski et al. 2002), and these datamay begin to shed some light on at least a couple ofthe questions posed above.

Acknowledgments. I am indebted to Drs. Jerry Strakaand Erik Rasmussen for their encouragement and per-spectives during the course of this work. I also extendthanks to Dr. Chuck Doswell and two anonymous re-viewers, who noticeably improved the clarity and or-ganization of the paper. Finally, I also am grateful toDrs. Robert Davies-Jones, Fred Carr, and Brian Fiedler,who reviewed earlier versions of the presentation.

REFERENCES

Adlerman, E. J., K. K. Droegemeier, and R. P. Davies-Jones, 1999:A numerical simulation of cyclic mesocyclogenesis. J. Atmos.Sci., 56, 2045–2069.

Agee, E. M., 1976: Multiple vortex features in the tornado cycloneand the occurrence of tornado families. Mon. Wea. Rev., 104,552–563.

Alberty, R. L., 1969: A proposed mechanism for cumulonimbus per-sistence in the presence of strong vertical shear. Mon. Wea. Rev.,97, 590–596.

Barnes, S. L., 1978a: Oklahoma thunderstorms on 29–30 April 1970.Part I: Morphology of a tornadic storm. Mon. Wea. Rev., 106,673–684.

——, 1978b: Oklahoma thunderstorms on 29–30 April 1970. Part II:Radar-observed merger of twin hook echoes. Mon. Wea. Rev.,106, 685–696.

Beebe, R. G., 1959: Notes on the Scottsbluff, Nebraska tornado, 27June 1955. Bull. Amer. Meteor. Soc., 40, 109–116.

Blanchard, D. O., and J. M. Straka, 1998: Some possible mechanismsfor tornadogenesis failure in a supercell. Preprints, 19th Conf.on Severe Local Storms, Minneapolis, MN, Amer. Meteor. Soc.,116–119.

Bluestein, H. B., 1983: Surface meteorological observations in severethunderstorms. Part II: Field experiments with TOTO. J. ClimateAppl. Meteor., 22, 919–930.

——, and W. P. Unruh, 1989: Observations of the wind field in tor-nadoes, funnel clouds, and wall clouds with a portable Dopplerradar. Bull. Amer. Meteor. Soc., 70, 1514–1525.

——, and J. H. Golden, 1993: A review of tornado observations. TheTornado: Its Structure, Dynamics, Prediction, and Hazards,Geophys. Monogr., No. 79, Amer. Geophys. Union, 319–352.

——, and A. L. Pazmany, 2000: Observations of tornadoes and otherconvective phenomena with a mobile, 3-mm wavelength, Dopp-ler radar. Bull. Amer. Meteor. Soc., 81, 2939–2952.

——, and S. G. Gaddy, 2001: Airborne pseudo-dual-Doppler analysisof a rear-inflow jet and deep convergence zone within a supercell.Mon. Wea. Rev., 129, 2270–2289.

——, A. L. Pazmany, J. C. Galloway, and R. E. McIntosh, 1995:Studies of the substructure of severe convective storms using amobile 3-mm wavelength Doppler radar. Bull. Amer. Meteor.Soc., 76, 2155–2169.

——, S. G. Gaddy, D. C. Dowell, A. L. Pazmany, J. C. Galloway,R. E. McIntosh, and H. Stein, 1997: Doppler radar observationsof substorm-scale vortices in a supercell. Mon. Wea. Rev., 125,1046–1059.

Bonesteele, R. G., and Y. J. Lin, 1978: A study of updraft–downdraftinteraction based on perturbation pressure and single-Dopplerradar data. Mon. Wea. Rev., 106, 113–120.

Brandes, E. A., 1977a: Flow in a severe thunderstorm observed bydual-Doppler radar. Mon. Wea. Rev., 105, 113–120.

——, 1977b: Gust front evolution and tornado genesis as viewed byDoppler radar. J. Appl. Meteor., 16, 333–338.

——, 1978: Mesocyclone evolution and tornadogenesis: Some ob-servations. Mon. Wea. Rev., 106, 995–1011.

——, 1981: Finestructure of the Del City–Edmond tornadic meso-circulation. Mon. Wea. Rev., 109, 635–647.

——, 1984a: Relationships between radar-derived thermodynamicvariables and tornadogenesis. Mon. Wea. Rev., 112, 1033–1052.

——, 1984b: Vertical vorticity generation and mesocyclone suste-nance in tornadic thunderstorms: The observational evidence.Mon. Wea. Rev., 112, 2253–2269.

——, R. P. Davies-Jones, and B. C. Johnson, 1988: Streamwise vor-ticity effects on supercell morphology and persistence. J. Atmos.Sci., 45, 947–963.

Brooks, E. M., 1949: The tornado cyclone. Weatherwise, 2, 32–33.Brooks, H. E., C. A. Doswell III, and R. P. Davies-Jones, 1993:

Environmental helicity and the maintenance and evolution oflow-level mesocyclones. The Tornado: Its Structure, Dynamics,Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geo-phys. Union, 97–104.

——, C. A. Doswell, and R. B. Wilhelmson, 1994: The role of mid-tropospheric winds in the evolution and maintenance of low-level mesocyclones. Mon. Wea. Rev., 122, 126–136.

Brown, J. M., and K. R. Knupp, 1980: The Iowa cyclonic–anticy-clonic tornado pair and its parent thunderstorm. Mon. Wea. Rev.,108, 1626–1646.

Brown, R. A., 1992: Initiation and evolution of updraft rotation withinan incipient supercell thunderstorm. J. Atmos. Sci., 49, 1997–2014.

——, D. W. Burgess, and K. C. Crawford, 1973: Twin tornado cy-clones within a severe thunderstorm: Single-Doppler radar ob-servations. Weatherwise, 26, 63–71.

——, L. R. Lemon, and D. W. Burgess, 1978: Tornado detection withpulsed Doppler radar. Mon. Wea. Rev., 106, 29–38.

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


——, 1965a: Some inferences about the updraft within a severe localstorm. J. Atmos. Sci., 22, 669–678.

——, 1965b: The evolution of tornadic storms. J. Atmos. Sci., 22,664–668.

——, and F. H. Ludlam, 1962: Airflow in convective storms. Quart.J. Roy. Meteor. Soc., 88, 117–135.

——, and R. J. Donaldson, 1963: Airflow and structure of a tornadicstorm. J. Atmos. Sci., 20, 533–545.

Burgess, D. W., R. A. Brown, L. R. Lemon, and C. R. Safford, 1977:Evolution of a tornadic thunderstorm. Preprints, 10th Conf. onSevere Local Storms, Omaha, NE, Amer. Meteor. Soc., 84–89.

Carbone, R. E., 1983: A severe frontal rainband. Part II: Tornadoparent vortex circulation. Mon. Wea. Rev., 111, 2639–2654.

Charba, J., and Y. Sasaki, 1971: Structure and movement of the severethunderstorms of 3 April 1964 as revealed from radar and surfacemesonetwork data analysis. J. Meteor. Soc. Japan, 49, 191–214.

Chisholm, A. J., 1973: Alberta hailstorms. Part I: Radar case studiesand airflow models. Alberta Hailstorms, Meteor. Monogr., No.36, Amer. Meteor. Soc., 1–36.

Darkow, G. L., and D. W. McCann, 1977: Relative environmentalwinds for 121 tornado bearing storms. Preprints, 10th Conf. onSevere Local Storms, Omaha, NE, Amer. Meteor. Soc., 413–417.

Das, P., 1983: Vorticity concentration in the subcloud layers of arotating cloud. National Science Foundation Final Rep. ATM-8023825, 78 pp.

Daugherty, J. R., A. I. Watson, T. R. Shepherd, and C. L. Ziegler,1996: The evolution of tornadic supercells during VORTEX-95as observed by the NOAA P-3 lower-fuselage C-band radar.Preprints, 18th Conf. on Severe Local Storms, San Francisco,CA, Amer. Meteor. Soc., 128–132.

Davies-Jones, R. P., 1982a: A new look at the vorticity equation withapplication to tornadogenesis. Preprints, 12th Conf. on SevereLocal Storms, San Antonio, TX, Amer. Meteor. Soc., 249–252.

——, 1982b: Observational and theoretical aspects of tornadogenesis.Intense Atmospheric Vortices, L. Bengtsson and J. Lighthill, Eds.Springer-Verlag, 175–189.

——, 1984: Streamwise vorticity: The origin of updraft rotation insupercell storms. J. Atmos. Sci., 41, 2991–3006.

——, 1985: Dynamical interaction between an isolated convectivecell and a veering environmental wind. Preprints, 17th Conf. onSevere Local Storms, St. Louis, MO, Amer. Meteor. Soc., 216–219.

——, 1986: Tornado dynamics. Thunderstorm Morphology and Dy-namics. 2d ed., E. Kessler, Ed., University of Oklahoma Press,197–236.

——, 1998: Tornadoes and tornadic storms. Preprints, 19th Conf. onSevere Local Storms, Minneapolis, MN, Amer. Meteor. Soc.,185.

——, 2000a: A Lagrangian model for baroclinic genesis of mesoscalevortices. Part I: Theory. J. Atmos. Sci., 57, 715–736.

——, 2000b: Can the hook echo instigate tornadogenesis barotrop-ically? Preprints, 20th Conf. on Severe Local Storms, Orlando,FL, Amer. Meteor. Soc., 269–272.

——, and H. E. Brooks, 1993: Mesocyclogenesis from a theoreticalperspective. The Tornado: Its Structure, Dynamics, Prediction,and Hazards, Geophys. Monogr., No. 79, Amer. Geophys.Union, 105–114.

——, C. A. Doswell, and H. E. Brooks, 1994: Comments on ‘‘Ini-tiation and evolution of updraft rotation within an incipient su-percell thunderstorm.’’ J. Atmos. Sci., 51, 326–331.

Donaldson, R. J., 1970: Vortex signature recognition by a Dopplerradar. J. Appl. Meteor., 9, 661–670.

Doswell, C. A., 1985: The operational meteorology of convectiveweather. Vol. 2: Storm scale analysis. NOAA Tech. Memo. ERLESG-15, 240 pp.

Dowell, D. C., and H. B. Bluestein, 1997: The Arcadia, Oklahoma,storm of 17 May 1981: Analysis of a supercell during torna-dogenesis. Mon. Wea. Rev., 125, 2562–2582.

——, ——, and D. P. Jorgensen, 1997: Airborne Doppler radar anal-

ysis of supercells during COPS-91. Mon. Wea. Rev., 125, 365–383.

Eskridge, R. E., and P. Das, 1976: Effect of a precipitation-drivendowndraft on a rotating wind field: A possible trigger mechanismfor tornadoes? J. Atmos. Sci., 33, 70–84.

Fankhauser, J. C., 1971: Thunderstorm environment interactions de-termined from aircraft and radar observations. Mon. Wea. Rev.,99, 171–192.

Forbes, G. S., 1975: Relationship between tornadoes and hook echoeson April 3, 1974. Preprints, Ninth Conf. on Severe Local Storms,Norman, OK, Amer. Meteor. Soc., 280–285.

——, 1981: On the reliability of hook echoes as tornado indicators.Mon. Wea. Rev., 109, 1457–1466.

Freund, R. F., 1966: Radar echo signature of tornadoes. Preprints,12th Conf. on Radar Meteorology, Norman, OK, Amer. Meteor.Soc., 362–365.

Fujita, T. T., 1958a: Mesoanalysis of the Illinois tornadoes of 9 April1953. J. Meteor., 15, 288–296.

——, 1958b: Tornado cyclone: Bearing system of tornadoes. Proc.Seventh Conf. on Radar Meteorology, Miami Beach, FL, Amer.Meteor. Soc., K31–K38.

——, 1965: Formation and steering mechanisms of tornado cyclonesand associated hook echoes. Mon. Wea. Rev., 93, 67–78.

——, 1973: Proposed mechanism of tornado formation from rotatingthunderstorms. Preprints, Eighth Conf. on Severe Local Storms,Denver, CO, Amer. Meteor. Soc., 191–196.

——, 1975a: Color map of super outbreak tornadoes of 3–4 April1974. Weatherwise, 28, 55.

——, 1975b: New evidence from the April 3–4, 1974 tornadoes.Preprints, Ninth Conf. on Severe Local Storms, Norman, OK,Amer. Meteor. Soc., 248–255.

——, 1979: Objectives, operation, and results of Project NIMROD.Preprints, 11th Conf. on Severe Local Storms, Kansas City, MO,Amer. Meteor. Soc., 259–266.

——, 1981: Tornadoes and downbursts in the context of generalizedplanetary scales. J. Atmos. Sci., 38, 1511–1534.

——, and H. Grandoso, 1968: Split of a thunderstorm into anticy-clonic and cyclonic storms and their motion as determined bynumerical model experiments. J. Atmos. Sci., 25, 416–439.

——, and R. M. Wakimoto, 1982: Anticyclonic tornadoes in 1980and 1981. Preprints, 12th Conf. on Severe Local Storms, SanAntonio, TX, Amer. Meteor. Soc., 213–216.

——, M. R. Hjelmfelt, and S. A. Changnon, 1977: Mesoanalysis ofrecord Chicago rainstorm using radar, satellite, and raingaugedata. Preprints, 10th Conf. on Severe Local Storms, Omaha, NE,Amer. Meteor. Soc., 65–72.

Fulks, J. R., 1962: On the Mechanics of the Tornado. National SevereStorms Project Rep. No. 4, U. S. Weather Bureau.

Gaddy, S. G., and H. B. Bluestein, 1998: Airborne dual-Doppleranalysis of a supercell hailstorm. Preprints, 19th Conf. on SevereLocal Storms, Minneapolis, MN, Amer. Meteor. Soc., 60–63.

Gal-Chen, T., 1978: A method for the initialization of the anelasticequations: Implications for matching models with observations.Mon. Wea. Rev., 106, 587–606.

Garrett, R. A., and V. D. Rockney, 1962: Tornadoes in northeasternKansas, May 19, 1960. Mon. Wea. Rev., 90, 231–240.

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.

Golden, J. H., 1974: Scale interaction implications for the waterspoutlife cycle. Part II. J. Appl. Meteor., 13, 693–709.

——, and B. J. Morgan, 1972: The NSSL/Notre Dame tornado in-tercept program, spring 1972. Bull. Amer. Meteor. Soc., 53,1178–1179.

——, and D. Purcell, 1978: Airflow characteristics around the UnionCity tornado. Mon. Wea. Rev., 106, 22–28.

Haglund, G. T., 1969: A study of the severe local storm of 16 April1967. ESSA Tech. Memo. ERLTM-NSSL 44, 54 pp. [NTISPB188315.]

Hane, C. E., and P. S. Ray, 1985: Pressure and buoyancy fields derived


from Doppler radar data in a tornadic thunderstorm. J. Atmos.Sci., 42, 18–35.

——, R. B. Wilhelmson, and T. Gal-Chen, 1981: Retrieval of ther-modynamic variables within deep convective clouds: Experi-ments in three dimensions. Mon. Wea. Rev., 109, 564–576.

Heymsfield, G. M., 1978: Kinematic and dynamic aspects of theHarrah tornadic storm analyzed from dual-Doppler radar data.Mon. Wea. Rev., 106, 233–254.

Houze, R. A., W. Schmid, R. G. Fovell, and H.-H. Schiesser, 1993:Hailstorms in Switzerland: Left movers, right movers, and falsehooks. Mon. Wea. Rev., 121, 3345–3370.

Jensen, B., T. P. Marshall, M. A. Mabey, and E. N. Rasmussen, 1983:Storm scale structure of the Pampa storm. Preprints, 13th Conf.on Severe Local Storms, Tulsa, OK, Amer. Meteor. Soc., 85–88.

Johnson, K. W., P. S. Ray, B. C. Johnson, and R. P. Davies-Jones,1987: Observations related to the rotational dynamics of the 20May 1977 tornadic storms. Mon. Wea. Rev., 115, 2463–2478.

Jorgensen, D. P., P. H. Hildebrand, and C. L. Frush, 1983: Feasibilitytest of an airborne pulse-Doppler meteorological radar. J. Cli-mate Appl. Meteor., 22, 744–757.

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

——, and R. B. Wilhelmson, 1978a: Simulations of right- and left-moving storms produced through storm splitting. J. Atmos. Sci.,35, 1097–1110.

——, and ——, 1978b: The simulation of three-dimensional con-vective storm dynamics. J. Atmos. Sci., 35, 1070–1096.

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

——, ——, and P. S. Ray, 1981: Observed and numerically simulatedstructure of a mature supercell thunderstorm. J. Atmos. Sci., 38,1558–1580.

Knupp, K. R., and W. R. Cotton, 1985: Convective cloud downdraftstructure: An interpretive study. Rev. of Geophys., 23, 183–215.

Lee, B. D., and R. B. Wilhelmson, 1997a: The numerical simulationof non-supercell tornadogenesis: Part I: Initiation and evolutionof pretornadic misocyclone circulations along a dry outflowboundary. J. Atmos. Sci., 54, 32–60.

——, and ——, 1997b: The numerical simulation of non-supercelltornadogenesis: Part II: Tornado evolution along a weak outflowboundary. J. Atmos. Sci., 54, 2387–2415.

——, and ——, 2000: The numerical simulation of non-supercelltornadogenesis. Part III: Parameter tests investigating the roleof CAPE, vortex sheet strength and boundary layer vertical shear.J. Atmos. Sci., 57, 2246–2261.

Lemon, L. R., 1974: Interaction of two convective scales within asevere thunderstorm: A case study. NOAA Tech. Memo. ERL-NSSL, 71, 43 pp. [NTIS COM-74-11642/AS.]

——, 1976a: The flanking line, a severe thunderstorm intensificationsource. J. Atmos. Sci., 33, 686–694.

——, 1976b: Wake vortex structure and aerodynamic origin in severethunderstorms. J. Atmos. Sci., 33, 678–685.

——, 1977: Severe thunderstorm evolution: Its use in a new techniquefor radar warnings. Preprints, 10th Conf. on Severe Local Storms,Omaha, NE, Amer. Meteor. Soc., 77–83.

——, 1982: New severe thunderstorm radar identification techniquesand warning criteria: A preliminary report. NOAA Tech. Memo.NWS NSSFC-1, 60 pp. [NTIS PB-273049.]

——, and D. W. Burgess, 1976: Tornadic storm airflow and mor-phology derived from single Doppler radar measurements. TheUnion City, Oklahoma Tornado of 24 May 1973, R. A. Brown,Ed. NOAA Tech. Memo. ERL-80, National Severe Storms Lab-oratory, 85–106.

——, and C. A. Doswell, 1979: Severe thunderstorm evolution andmesocyclone structure as related to tornadogenesis. Mon. Wea.Rev., 107, 1184–1197.

——, D. W. Burgess, and R. A. Brown, 1975: Tornado productionand storm sustenance. Preprints, Ninth Conf. on Severe LocalStorms, Norman, OK, Amer. Meteor. Soc., 100–104.

——, ——, and ——, 1978: Tornadic thunderstorm airflow and mor-

phology derived from single Doppler radar measurements. Mon.Wea. Rev., 106, 48–61.

Leslie, L. M., and R. K. Smith, 1978: The effect of vertical stabilityon tornadogenesis. J. Atmos. Sci., 35, 1281–1288.

Lilly, D. K., 1982: The development and maintenance of rotation inconvective storms. Intense Atmospheric Vortices, L. Bengtssonand J. Lighthill, Eds. Springer-Verlag, 149–160.

——, 1986a: The structure, energetics, and propagation of rotatingconvective storms. Part I: Energy exchange with the mean flow.J. Atmos. Sci., 43, 113–125.

——, 1986b: The structure, energetics, and propagation of rotatingconvective storms. Part II: Helicity and storm stabilization. J.Atmos. Sci., 43, 126–140.

Ludlam, F. H., 1963: Severe Local Storms: A review. Severe LocalStorms, Meteor. Monogr., No. 27, 1–30.

Markowski, P. M., J. M. Straka, and E. N. Rasmussen, 2002: Directsurface thermodynamic observations within the rear-flank down-drafts of nontornadic and tornadic supercells. Mon. Wea. Rev.,in press.

Marshall, T. P., and E. N. Rasmussen, 1982: The mesocyclone evo-lution of the Warren, Oklahoma tornadoes. Preprints, 12th Conf.on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc.,375–378.

Marwitz, J. D., 1972a: The structure and motion of severe hailstorms.Part I: Supercell storms. J. Appl. Meteor., 11, 166–179.

——, 1972b: The structure and motion of severe hailstorms. Part III:Severely sheared storms. J. Appl. Meteor., 11, 189–201.

Moller, A., C. A. Doswell, J. McGinley, S. Tegtmeier, and R. Zipser,1974: Field observations of the Union City tornado in Oklahoma.Weatherwise, 27, 68–77.

Nelson, S. P., 1977: Rear flank downdraft: A hailstorm intensificationmechanism. Preprints, 10th Conf. on Severe Local Storms, Oma-ha, NE, Amer. Meteor. Soc., 521–525.

Newton, C. W., and H. R. Newton, 1959: Dynamical interactionsbetween large convective clouds and environmental verticalshear. J. Meteor., 16, 483–496.

Peterson, R. E., 1976: The Sunray tornado. Bull. Amer. Meteor. Soc.,57, 805–807.

Rasmussen, E. N., and J. M. Straka, 1996: Mobile mesonet obser-vations of tornadoes during VORTEX-95. Preprints, 18th Conf.on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc.,1–5.

——, and ——, 1998: Variations in supercell morphology. Part I:Observations of the role of upper-level storm-relative flow. Mon.Wea. Rev., 126, 2406–2421.

——, R. E. Peterson, J. E. Minor, and B. D. Campbell, 1982: Evo-lutionary characteristics and photogrammetric determination ofwindspeeds within the Tulia outbreak tornadoes 28 May 1980.Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX,Amer. Meteor. Soc., 301–304.

——, J. M. Straka, R. P. Davies-Jones, C. A. Doswell, F. H. Carr,M. D. Eilts, and D. R. MacGorman, 1994: Verification of theOrigins of Rotation in Tornadoes Experiment: VORTEX. Bull.Amer. Meteor. Soc., 75, 995–1006.

Ray, P. S., 1976: Vorticity and divergence fields within tornadicstorms from dual-Doppler observations. J. Appl. Meteor., 15,879–890.

——, R. J. Doviak, G. B. Walker, D. Sirmans, J. Carter, and B.Bumgarner, 1975: Dual-Doppler observation of a tornadic storm.J. Appl. Meteor., 14, 1521–1530.

——, B. C. Johnson, K. W. Johnson, J. S. Bradberry, J. J. Stephens,K. K. Wagner, R. B. Wilhelmson, and J. B. Klemp, 1981: Themorphology of several tornadic storms on 20 May 1977. J. At-mos. Sci., 38, 1643–1663.

——, D. P. Jorgensen, and S.-L. Wang, 1985: Airborne Doppler radarobservations of a convective storm. J. Climate Appl. Meteor.,24, 687–698.

Roberts, R. D., and J. W. Wilson, 1995: The genesis of three non-supercell tornadoes observed with dual-Doppler radar. Mon.Wea. Rev., 123, 3408–3436.


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

——, 1986: Tornadoes and tornadogenesis. Mesoscale Meteorologyand Forecasting, P. S. Ray, Ed., 414–436.

——, 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 simulatedsupercell thunderstorms. J. Atmos. Sci., 42, 271–292.

Sadowski, A., 1958: Radar observations of the El Dorado, Kansastornado, June 10, 1958. Mon. Wea. Rev., 86, 405–408.

——, 1969: Size of tornado warning area when issued on basis ofradar hook echo. ESSA Tech. Memo. WBTM Fcst. 10, 26 pp.[NTIS PB 184613.]

Schlesinger, R. E., 1975: A three-dimensional numerical model of anisolated deep convective cloud: Preliminary results. J. Atmos.Sci., 32, 934–964.

Shapiro, A., and P. M. Markowski, 1999: Dynamics of elevated vor-tices. J. Atmos. Sci., 56, 1101–1122.

Snow, J. T., C. R. Church, and B. J. Barnhart, 1980: An investigationof the surface pressure fields beneath simulated tornado cy-clones. J. Atmos. Sci., 37, 1013–1026.

Stanford, J. L., 1977: Tornado: Accounts of Tornadoes in Iowa. IowaState University Press, 120 pp.

Stout, G. E., and F. A. Huff, 1953: Radar records Illinois tornado-genesis. Bull. Amer. Meteor. Soc., 34, 281–284.

Straka, J. M., E. N. Rasmussen, and S. E. Fredrickson, 1996: A mobilemesonet for fine-scale meteorological observations. J. Atmos.Oceanic Technol., 13, 921–936.

Tepper, M., and W. E. Eggert, 1956: Tornado proximity traces. Bull.Amer. Meteor. Soc., 37, 152–159.

Trapp, R. J., 1999: Observations of nontornadic low-level mesocy-clones and attendant tornadogenesis failure during VORTEX.Mon. Wea. Rev., 127, 1693–1705.

——, and B. H. Fiedler, 1995: Tornado-like vortexgenesis in a sim-plified numerical model. J. Atmos. Sci., 52, 3757–3778.

van Tassel, E. L., 1955: The North Platte Valley tornado outbreak ofJune 27, 1955. Mon. Wea. Rev., 83, 255–264.

Wakimoto, R. M., and J. W. Wilson, 1989: Non-supercell tornadoes.Mon. Wea. Rev., 117, 1113–1140.

——, and C. Liu, 1998: The Garden City, Kansas, storm duringVORTEX 95. Part II: The wall cloud and tornado. Mon. Wea.Rev., 126, 393–408.

——, and H. Cai, 2000: Analysis of a nontornadic storm duringVORTEX 95. Mon. Wea. Rev., 128, 565–592.

——, C. Liu, and H. Cai, 1998: The Garden City, Kansas, stormduring VORTEX 95. Part I: Overview of the storm life cycleand mesocyclogenesis. Mon. Wea. Rev., 126, 372–392.

Walko, R. L., 1993: Tornado spin-up beneath a convective cell: Re-quired basic structure of the near-field boundary layer winds.The Tornado: Its Structure, Dynamics, Prediction, and Hazards,Geophys. Monogr., No. 79, Amer. Geophys. Union, 89–95.

Ward, N. B., 1961: Radar and surface observations of tornadoes of4 May 1961. Proc. Ninth Weather Radar Conf., Boston, MA,Amer. Meteor. Soc., 175–180.

——, 1964: The Newton, Kansas, tornado cyclone of May 24, 1962.Proc. 11th Weather Radar Conf., Boston, MA, Amer. Meteor.Soc., 410–415.

——, 1972: The exploration of certain features of tornado dynamicsusing a laboratory model. J. Atmos. Sci., 29, 1194–1204.

Weisman, M. L., and J. B. Klemp, 1982: The dependence of nu-merically simulated convective storms on vertical wind shearand buoyancy. Mon. Wea. Rev., 110, 504–520.

——, and J. B. Klemp, 1984: The structure and classification ofnumerically simulated convective storms in directionally varyingwind shears. Mon. Wea. Rev., 112, 2479–2498.

Wicker, L. J., and R. B. Wilhelmson, 1995: Simulation and analysisof tornado development and decay within a three-dimensionalsupercell thunderstorm. J. Atmos. Sci., 52, 2675–2703.

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.

Williams, D. T., 1963: The thunderstorm wake of May 4, 1961. Na-tional Severe Storms Project Rep. 18, 23 pp. [NTIS No. PB168223.]

Wilson, J. W., 1986: Tornadogenesis by nonprecipitation inducedwind shear lines. Mon. Wea. Rev., 114, 270–284.

Wurman, J., and S. Gill, 2000: Finescale radar observations of theDimmitt, Texas (2 June 1995), tornado. Mon. Wea. Rev., 128,2135–2164.

——, J. M. Straka, and E. N. Rasmussen, 1996: Fine-scale Dopplerradar observations of tornadic storms. Science, 272, 1774–1777.

——, ——, E. N. Rasmussen, M. Randall, and A. Zahrai, 1997:Design and deployment of a portable, pencil-beam, pulsed, 3-cm Doppler radar. J. Atmos. Oceanic Technol., 14, 1502–1512.

Ziegler, C. L., E. N. Rasmussen, T. R. Shepherd, A. I. Watson, andJ. M. Straka, 2001: The evolution of low-level rotation in the29 May 1994 Newcastle–Graham, Texas, storm complex duringVORTEX. Mon. Wea. Rev., 129, 1339–1368.

hook echoes and rear-flank downdrafts: a...

Documents