THE SEVERE T HUN DERST ORM ELEC TRIFIC A TIO N A N D PRECIPIT A TIO N STUDY

BY TIM O T HY J . L A N G , L. J AY MILLER, M O RRIS W E I S M A N , STEVEN A . RUTLE D GE, LLYLE J . BARKER III, V . N . BRINGI, V . C H A N D R ASEKAR , A N D R E W D E TWILER , N O L A N D O E SKE N , J O H N HELS D O N , CHARLES K N IG H T , PAUL KREHBIEL, W A L T E R A . LY O NS, D O N M A C G O R M A N , ERIK RASMUSSEN, W I L L I A M RIS O N , W . D A VI D RUST , A N D R O N A L D J . T H O M A S

Data from a field project on the Colorado - Kansas border in summer 2000 is helping to improve our understanding of positive cloud-to-ground lightning and low- precipitation .

evere , because of their propensity as they exhibit not only a wide range of electrical ac- to injure, kill, and cause extensive property dam- tivity, but also diversity in precipitation type and Sage, are a primary concern to not only weather amount. One of the more intriguing severe storms forecasters but also the public. However, these storms types in this regard is the supercell remain a puzzling scientific and forecasting problem, (Browning 1964). In its most pristine state, a supercell is a unicellular thunderstorm comprised of a single, long-lived, rotating updraft, and it frequently pro- AFFILIATIONS: LANG, RUTLEDGE, BRINGI, AND CHANDRASEKAR— duces large , high winds, prolific lightning, and Colorado State University, Fort Collins, Colorado; MILLER, WEISMAN, occasionally tornadoes. While the basic dynamics of AND KNIGH T—National Center for Atmospheric Research, Boulder, supercells seem well understood (e.g., Klemp 1987), Colorado; BARKER—National Weather Service, Lincoln, Illinois; these storms exhibit a wide variety of precipitation DETWILER AND HELSD O N—South Dakota School of Mines and Technology, Rapid City, South Dakota; D OESKEN—Colorado Climate characteristics that are not understood. For instance, Center, Fort Collins, Colorado; KREHBIEL, RISON, AND TH OMAS— supercells have been classified as either low precipi- New Mexico Institute of Mining and Technology, Socorro, New tation (LP; Donaldson et al. 1965; Davies-Jones et al. Mexico; LY O NS—FMA Research, Inc., Fort Collins, Colorado; 1976; Burgess and Davies-Jones 1979; Bluestein and MACGORMAN, RASMUSSEN, AND RUST—National Severe Storms Parks 1983), classic or medium precipitation (MP), Laboratory, Norman, Oklahoma or heavy precipitation (HP; Doswell and Burgess CORRESPONDING AUTHOR: Timothy J. Lang, Department of 1993; Rasmussen and Straka 1998) based on visual Atmospheric Science, Colorado State University, Fort Collins, C O 80523 observations of the cloud and precipitation shafts. E-mail: [email protected] Perhaps the least-understood among these storms are D OI: 10.1175/BAMS-85-8-1 107 LP supercells, which characteristically produce some

In final form 26 March 2004 large hail but little . Potentially because of the dry ©2004 American Meteorological Society environment and lack of visible precipitation, the vis- ible cloud below the anvil is a skeleton compared with

A M ERI C A N M E T E O R O L O G I C A L S O C IE T Y A U G U S T 2004 BAfft I 1107

other supercell storms (Bluestein and Parks 1983; began with efforts to discriminate between hail and Bluestein and Woodall 1990). rain, but as these radars became more sophisticated, Another unusual aspect of some convective storms, the number of observed variables and thus the num- including supercells, is their tendency to produce ber of potential discriminants increased. Some algo- copious positive cloud-to-ground (+CG) lightning rithms distinguish between such diverse hydrometeor (e.g., Branick and Doswell 1992; Curran and Rust types as large and small hail, graupel, , and 1992; Carey and Rutledge 1998; Williams 2001), in mixed-phase precipitation. Hydrometeor identifica- contrast with most warm-season thunderstorms that tion can be useful in various applications to weather produce predominantly negative CG lightning (-90% forecasting and aviation weather warnings, as well as of all warm-season CGs are negative; Orville 1994; in fundamental studies of structure and evolu- Orville and Silver 1997; Orville and Huffines 2001). tion. However, like all remote sensing techniques, po- In fact, the percentage of CG lightning that is posi- larimetric hydrometeor classification needs in situ veri- tive in these storms [predominantly positive CG fication to establish and improve the scope of its validity. (PPCG) storms] can be far greater than 50%, some- During May-July 2000, the Severe Thunderstorm times approaching 100%. Many PPCG storms are Electrification and Precipitation Study (STEPS; severe, but not all severe storms are PPCG (Carey and Weisman and Miller 2000; http://box.mmm.ucar.edu/ Rutledge 2003; Carey et al. 2003), and we currently pdas/STEPS.html) took place near the Colorado- do not understand what exactly distinguishes PPCG Kansas border in order to achieve a better under- storms from other storms. standing of the interactions between kinematics, pre- Researchers typically observe two major positive cipitation production, and electrification in severe charge regions within the updrafts of ordinary thun- thunderstorms. Specific scientific objectives included derstorms, an upper one (above the -20°C isotherm 1) understanding the apparent major differences in altitude; the main reservoir for positive charge) and precipitation output from supercells that have led to a lower one (near 0°C altitude), with a major nega- their being classified as LP, MP, and HP; 2) under- tive charge region (thought to be the source region standing lightning occurrence and behavior in storms, for most -CGs) in between (i.e., between -10° and and how it differs among storm types, particularly to -20°C altitude; e.g., Krehbiel 1986; Koshak and better understand the mechanisms by which storms Krider 1989; Stolzenburg et al. 1998a,b). This typical produce predominantly +CG lightning; and 3) veri- structure is often thought of as a dipole (main upper fying and improving microphysical interpretations positive charge over midlevel negative) or tripole from polarimetric radar. Additionally, the emphasis (considering the commonly observed lower positive on +CG lightning enabled research into what is dif- charge; Williams 1989), although thunderstorm ferent about the small subset of +CGs, usually from charge structures can be more complex than a simple large thunderstorms known as mesoscale convective dipole or tripole, particularly outside the main updraft systems (MCSs), that trigger mesospheric transient (Stolzenburg 1998a,b). However, PPCG storms may luminous events (TLEs) such as sprites (Lyons et al. not have typical electrical structures. Indeed, out- 2000, 2003a,b; Williams 1998). This latter problem standing issues in +CG research include identifying has important implications for our understanding of the positive charge region that is the source for +CGs, how TLEs occur. and in addition, understanding if, how, and why the This paper is intended to be an overview of the charge structures in PPCG storms differ from ordi- STEPS field campaign and in addition provides a brief nary thunderstorms. Williams (2001) reviewed hy- examination of STEPS research in order to demon- potheses for positive CG storms, and testing these strate how the design and execution of the field cam- hypotheses requires more information on relation- paign helped address the project goals. Because of ships between precipitation formation, airflow kine- length constraints, this paper mainly, though not ex- matics, electrification, and lightning production in clusively, focuses on the positive CG issue (STEPS PPCG thunderstorms. goal no. 2) in discussing the STEPS project. The ability to understand these relationships, how- ever, requires sophisticated tools to observe and ana- S T E P S D E S I G N A N D E X E C U T I O N . T h e lyze thunderstorm characteristics. In particular, for STEPS project brought together a unique suite of precipitation, research with polarimetric radars has complementary observing platforms in eastern Colo- led to an emerging capability for identifying hydrom- rado and western Kansas. The basic geographical lay- eteor types remotely (Vivekandan et al. 1999; Liu and out of the project is shown in Fig. 1. This portion of Chandrasekar 2000; Straka et al. 2000). Such work the High Plains region of the United States has been

1 1 0 8 I BAF1 5 - AUGUST 2004

observed climatologically to favor supercell storms, ized, providing information on the size, shape, orien- particularly the LP variety (Bluestein and Parks 1983). tation, and thermodynamic phase of hydrometeors. This is primarily due to the warm-season presence in The network of three Doppler radars provided the this region of the dryline, the boundary between moist opportunity to examine the three-dimensional inter- air from the Gulf of Mexico and drier continental air, which has been strongly associated with the occur- rence of LP storms. This association exists possibly because LP storms form in environments that are drier and have less low-level shear than traditional supercell environments, characteristics that drylines could provide via proximity to dry air, as well as en- hanced mixing to reduce shear values (Bluestein and Parks 1983). However, little is understood about how exactly dryline proximity affects the kinematic and microphysical structure of LP storms. The STEPS region also is favorable for thunderstorms that pro- duce predominantly +CG lightning (Zajac and Rutledge 2001; Carey and Rutledge 2003; Carey et al. 2003) as well as severe hailstorms (Changnon 1977). Thus, the STEPS domain was ideal for studying the storms of interest. We specifically designed the field measurements and analysis for STEPS to explore the mechanisms of precipitation formation and lightning production in FIG. I. N o m i n a l a r e a s of c o v e r a g e (s h a d i n g ) by t h e supercell and other storms. The instrumentation, t r i p l e - D o p p l e r r a d a r n e t w o r k . O u t e r d u a l - D o p p l e r listed in Table 1, included two S-band (-10 cm wave- lo bes ( b e a m an gles g r e a t e r t h a n 3 0 ° ) a n d t h e i n n e r triple - D op pler trian gle a r e outlined in red. T h e second length) polarimetric radars, the Colorado State Uni- dual - Doppler lobe for t h e research radars C S U - C H I L L versity dual-polarization Doppler radar (CSU- ( C H I L ) an d S - Pol ( S P O L ) is o utlined in blue. T h e re - CHILL) near Burlington, Colorado, and the National gion wit h i n w h ic h vertical resolution is b et t er t h a n I k m Center for Atmospheric Research (NCAR) S-band f o r t h e L M A is o u t l i n e d in g r e e n ( ~ I 2 5 k m ra d i us). dual-polarization Doppler radar (S-Pol) near Idalia, T o p o g r a p h ic height co n t o u rs (black lines) a r e at 3, 4, Colorado, along with the S-band National Weather 5, a n d 6 kft. N W S ra d a rs a r e sh ow n for D e n v e r , C O Service (NWS) Weather Surveillance Radar-1988 ( K F T G ) , P u e b l o , C O ( K P U X ) , a n d G o o d l a n d , K S Doppler (WSR-88D) at Goodland, Kansas. Collec- ( K G L D ) , a l o n g w i t h t h e Y u c c a R i d g e F i e l d S t a t i o n ( Y R F S ) . La n d m a r k s a r e sh own at D e n v e r (d en), Colo - tively, these radars determined the internal airflow rad o Springs (csp), L i m o n (li m), an d A k r o n , C O (a k o), and precipitation structure of storms. The CSU- and at Mc C o o k , N E ( m c k ) . A l l distances a r e east - west CHILL and S-Pol radars are both dual-linearly polar- (x) an d north - south (y) f r o m t h e Go o d l a n d W S R - 8 8 D .

TABLE 1. List of STEPS i nstr u m e n tati o n an d t h eir p r i m a ry purposes d urin g t h e project.

P l a t f o r m P u r p o s e

Radar n e t w o r k C h aract eri z e precipitation struct ure and airflow in t h u n d erstor m s

L M A Map lightning discharges in t h r e e dimensions

N L D N Locate grou n d strik e locations and tim es of cloud - to - ground lightning

T - 28 ar m o r e d research aircraft In situ o bservations of t h u n d erst or m microphysics, electric fields, and win ds

Balloon - b orne EF M In situ o bservations of electric fields, win ds, and th erm o dyna mics

Mo bile m eso n et C h aract eri z e surface m esoscale e nviro n m e n t of t h u n d erst or m

M - G L A SS soundings C h aracteri z e th er m o dyn a mic struct ure and win d shear of e nviro n m e nt

YR FS Re c o r d o bservatio ns of T LEs o v e r t h u n d erstor m s

AMERIC A N METEOROLO GIC AL SO CIETY A UGUST 2004 B A I 1 5 * | 1 1 0 9

nal airflow of STEPS storms, via postproject synthe- NSSL has improved and used for 15 yr. An overview sis of the multiple-Doppler observations. of this instrument is provided by MacGorman and The deployable Lightning Mapping Array (LMA) Rust (1998, their section 6.2.3). The electric field data from the New Mexico Institute of Mining and Tech- yielded the full three-dimensional electric field vec- nology mapped the three-dimensional total lightning tor and can be used to infer layers of net negative or activity. The system located the sources of impulsive positive charge (e.g., Stolzenburg and Marshall 1994). very high frequency (VHF) radio signals from the The mobile laboratory also contained standard surface lightning by measuring the time that the signals ar- meteorological sensors and a surface electric field meter. rived at the 13 receiving stations deployed over a 60- NCAR mobile sounding systems [Mobile GPS/Lo- 80-km-diameter area in northwestern Kansas and ran Atmospheric Sounding System (M-GLASS)] and eastern Colorado. The LMA locates numerous VHF NOAA/NSSL mobile mesonet vehicles characterized sources from the lightning activity, which can be the storm environment. The mesonet vehicles, which grouped into individual flashes either manually using augment existing meteorological networks, consisted special analysis software or through an automated al- of meteorological instruments mounted on standard gorithm such as that of Thomas et al. (2003), by con- automobiles. They can provide accurate observations sidering the spatial and temporal proximity of VHF of pressure, temperature, and relative humidity, as sources to one another. The LMA is most sensitive to well as wind direction and speed, whether the vehicle VHF radiation from negative breakdown occurring is moving or stationary. The M-GLASS systems were within regions of net positive charge (Mazur et al. mounted on pickup trucks and made standard atmo- 1997; Rison et al. 1999). Thus, the LMA can be used spheric sounding and surface measurements. to identify major positive charge regions tapped by Finally, the Yucca Ridge Field Station (YRFS), lo- lightning based on analysis of VHF emission density. cated near Fort Collins, Colorado (275 km northwest The National Lightning Detection Network of the LMA centroid; Fig. 1), provided observations (NLDN) provided CG flash data. The NLDN consists of TLEs during STEPS. The YRFS, operated by FMA of a network of magnetic direction finder and time Research, made use of numerous on- and off-site in- of arrival sensors used to locate in space and time struments, including radio frequency (RF), telescopic, ground strike locations of CG lightning. Information and photometric sensors along with low-light imag- on CG polarity, peak current, and multiplicity also are ers. Data from up to six extremely low frequency available. The most recent NLDN upgrade, discussed (ELF) sensors on four different continents provided in Cummins et al. (1998), gives greater than 90% de- additional information on the parent lightning of TLE tection efficiency and 0.5-km location accuracy within phenomena in the STEPS domain. the STEPS domain. The combination of all of these observations pro- The South Dakota School of Mines and Technol- vided, along with an understanding of each storm's ogy (SDSMT) armored T-28 aircraft provided in situ mesoscale environment, a depiction of the coevolv- wind, microphysical, electric field, and particle charge ing kinematic, microphysical, and electrical structures data in the lower- to middle-altitude range within up- and lightning behavior of STEPS thunderstorms. drafts and hail shafts. The T-28 measured the com- Because of the detailed observing network, the STEPS plete spectrum of water and ice particles in clouds, data provide a unique opportunity to answer key ranging from cloud droplets a few micrometers in questions about precipitation formation and electri- diameter to about 5-cm-diameter hail. One of its three fication within severe storms. Additionally, the pres- precipitation particle imaging probes had the capa- ence of two polarimetric radars and in situ observa- bility to determine particle charge as the particle is tions provided an opportunity to evaluate and imaged (minimum sensitivity as low as 0.5 pC). In improve radar-based hydrometeor identification and addition, it carried a six-instrument electric field quantification algorithms. meter system that maps the total vector electric field The operations center for STEPS was at the CSU- inside and outside clouds. CHILL radar facility, temporarily relocated from its Mobile sounding systems from the National Oce- home base at Greeley, Colorado, to Burlington, Colo- anic and Atmospheric Administration/National Se- rado. Mobile facilities and STEPS personnel gener- vere Storms Laboratory (NOAA/NSSL) obtained bal- ally were based out of Burlington and Goodland, Kan- loon-borne measurements of electric fields inside sas. STEPS received significant support from the local storms [electric field meter (EFM) balloons]. The NWS forecast office in Goodland (see sidebar 1), and main instrument was the balloon-borne EFM [origi- daily forecast and observational platform status brief- nally developed by Winn and Byerley (1975)] that ings occurred each morning at this NWS facility.

1 1 1 0 I BAF1 5 - AUGUST 2004

Based on each briefing, we formulated operations tion, Table 2 lists the peak hourly percentage with plans for the afternoon and evening. The research positive polarity of all CGs that occurred within radars (CSU-CHILL and S-Pol) typically were run- 125 km of the LMA network centroid. This is used ning surveillance scans by noon. When convection as an indicator of positive CG production, as not all was likely, M-GLASS soundings were released at vari- storms have had their lightning data (both NLDN ous locations and vehicle platforms (mobile mesonet, and LMA) analyzed in detail yet. Only cases that had EFM balloons) were deployed in strategic locations 10 or more CGs per hour at peak have percentages where we expected activity. Once we identified con- computed. Note that since STEPS instrumentation vective targets, the vehicles and T-28 aircraft were focused on storms producing the most +CGs, this vectored to the storm via two-way radio communi- column in Table 2 tends to underestimate the maxi- cations with the operations center. In addition, the mum percentage of+CGs in the target storm because research radars would begin synchronized sector- of the potential inclusion of concurrent storms not based plan position indicator (PPI) and RHI scans of producing as many +CGs. Despite this limitation, the target storm. Table 2 demonstrates that most of the STEPS cases The main focus of observations was storms that oc- produced some sort of severe weather, and at least curred within or passed through the dual-Doppler half were associated with more than 50% -i-CGs at lobes formed by each radar pair within the STEPS peak (i.e., PPCG storms). network (see Fig. 1). Of these, the highest priority was Indeed, it was the ubiquity of PPCG storms in this supercell storms, especially those with LP character- area that was one of the distinctive results of STEPS, istics, as well as thunderstorms observed to be pro- in that the storms spanned vastly different organiza- ducing predominantly +CGs. tional categories, from small isolated convection to various types of large multicell storms, as well as E X A M P L E S O F S T E P S O B S E R V A T I O N S . supercells. Both storms with LP characteristics (de- Overview of STEPS cases. We were able to obtain data termined visually; total of three cases—31 May, on a number of different cases during STEPS. Table 2 3 June, and 5 July) and non-LP storms exhibited this lists the major STEPS operations days, along with a flash behavior. While most +CG storms were associ- short description of the most interesting storms of ated with severe weather, there was one PPCG case those days, and the most extreme severe storm re- (6 June) that had no severe weather reports. In addi- ports (if any) associated with those storms. In addi- tion, there were examples of low-CG storms (less than

[ T T ^ R I F T T I ATION 1 S J H I H A J *:E N RESEARCH AND FORECAST COMMUNI

T h e N W S office in Goodland, Kansas Web - based N C A R model output. Analysis and Nowcasting (SC A N)], and was in a unique position to provide the N W S short - term forecasters and severe weather reports to the O C STEPS ex periment with logistical project investigators collaborated on during field operations. assistance, forecast personnel, local forecast briefings. This allowed local T h e N W S benefited through ex pertise, and volunteer field -team ex pertise t o be integrated into the ex posure to unique datasets in near - participants. Indeed, demonstration of operational decision process. Local real time. Forecaster access to M- cooperation between the forecast and N O A A We a t h e r Radio stations GLASS soundings, timely reports fro m research com munities was a goal of the disseminated a daily briefing summary. the mobile mesonet, and Web - based ex periment. Preoperational phase Twenty - five volunteers fro m seven C SU - C HILL and N C A R S-Pol data all support included assistance in facility N W S offices participated in various contributed to an improved warning procure m ent, sensor placement, support positions. T h e roles of these program. Interaction with STEPS climatological research, lodging volunteers ranged fro m project researchers, including seminars assistance, and building local com munity nowcasters to field participants in the presented by project investigators, support for the project. In addition, mobile mesonet and EFM ballooning allowed N W S staff to increase their N W S personnel provided much of the operations. A two - way radio enhanced knowledge of convective processes and local media support during STEPS, com munications between the N W S severe convection forecasting. T h e including arranging of the STEPS media office and the STEPS O perations procedures and lessons learned during day. C e nter ( O C ). N W S relayed fixed STEPS provided a model for N W S During the operational phase, the mesonet data, output fro m N W S participation in the Bo w Echo and N W S office was the hub for planning analysis software [such as the Local Mesoscale C onvective V ort e x (M C V) and forecasting. Morning briefings Analysis and Prediction System (LAPS) Ex periment (B A M EX) field project, occurred at the office through the use and the System for C onvection which occurred in the summer of 2003. of both N W S com puter resources and

AMERICAN METEOROLOGICAL SO CIETY AUGUST 2004 BAI1 5* | 6 5

TABLE 2. O v e r v i e w of S T E P S cases. Seve r i ty based on re p o rts t o N W S . P e a k hourly positive C G p ercen ta g e based on all C G s occu rri n g d urin g o p eratio ns wit h i n 125 k m of t h e L M A n e tw o r k centroid . N / A: Less t h a n 10 C G s t o tal in any 1 h.

D a t e S t o r m s u m m a r y P e a k severity + C G peak hourly fractio n

25 May C o n v e ctiv e line N o re p orts 7%

26 May C o n v e ctiv e line 0.75 -in. hail 3 4%

31 May Isolated LP st or m l - in. hail 7 3% 70 - kt gust

3 Ju n Isolated LP st or m l - in. hail N / A

6 Ju n Isolated st or m N o re p orts 7 9%

9 Ju n C o n v e ctiv e line 60 - kt gust 3 7%

1 1 Ju n A sy m m e tric M C S 1.5-in. hail 6 1% 57 - kt gust

12 Ju n Isolated st or m 0.75 -in. hail N / A

19 Ju n Multicell st or m 65 - kt gust 10%

22 Ju n C o n v e ctiv e line F0 t orn a d o 7 4% 1.75-in. hail 60 - kt gust

23 Ju n Multicell st or m 0.75 -in. hail 8 0% 60 - kt gust

24 Ju n Classic su p ercell l - in. hail 9 6%

29 Ju n Classic su p ercell F l t orn a d o 6 8% 1.75-in. hail 61 - kt gust

1 Jul C o n v e ctiv e line N o re p orts 15%

5 Jul LP su percell Funnel clou d 9 3% 1.75-in. hail 65 - kt gust

10 Jul Multicell st or m 1.75-in. hail 7 5% 52 - kt gust

12 Jul Isolated st o r m N o re p orts 2 8%

17 Jul C o n v e ctiv e line N o re p orts 3 7%

18 Jul C o n v e ctiv e line l - in. hail 2 8% 60 - kt gust

19 Jul C o n v e ctiv e line F0 t orn a d o 8 8% 1.75-in. hail

20 Jul C o n v e ctiv e line Flash flo o d 8 8% 1.75-in. hail 63 - kt gust

10 total CGs per hour at maximum) that had inverted midlevels (i.e., near -20°C) and below the main nega- polarity electrical structures, in that the main positive tive charge region in contrast to what is most com- charge region appeared to lie within thunderstorm monly seen in other storms (e.g., Stolzenburg et al.

1 1 1 2 I BAF1 5 - A UGUST 2004

FIG. 2. (a) Photograph of 29 Jun classic supercell taken at 2323 U T C , a p p r o x i m ately 5 m i n before t o r n a d o touchdown. Photograph taken at S - Pol looking toward the E SE . Though taken too far away (60 - 70 k m) to resolve the precipitation shaft, the cloud does not have the structure of an L P supercell like 5 Jul [shown in (b)]. Photo by C. Knight, (b) Photograph of the mature phase (after 2330 U T C ) of the 5 Jul L P supercell. Notable visual characteristics are a striated, bell - shaped cloud that is often indicative of a rotating updraft, and very little precipitation to the north and east. Photo taken by M. W e i s m a n fro m the S S E direction, approximately 5 miles fro m the updraft base.

1998a,b). The most prominent example was 3 June, a Colorado, but a distinct dryline was not evident. A small isolated storm with LP characteristics that pro- short line of convective cells developed around duced no CGs of either polarity but had a possibly 2200 UTC in the northwest corner of Kansas. The inverted electrical structure based on inferences from convection subsequently moved southeastward, re- an EFM balloon sounding and LMA intracloud flash maining in a multicellular phase for nearly 1.5 h be- behavior (Rust and MacGorman 2002; Hamlin et al. fore making a 35° right turn as it became more 2003). Many of the PPCG storms also appeared to supercellular in character. The right turn is believed have inverted polarity structures based on EFM and to be the result of gust front convergence favoring LMA data (Rust and MacGorman 2002; Hamlin et al. growth on the storm's right flank. Around this time, 2003; Rust et al. 2003). Apart from the main cases re- storm size increased dramatically and a tornado first viewed here, there were smaller storms with low to- touched ground (2328 UTC). The tornado was on the tal flash rates (2-3 min - 1 ) that exhibited inverted elec- ground for about 16 min, and mobile mesonets trical behavior. One such storm, different from the tracked it throughout its lifetime. (A description and supercell on that day, occurred on 24 June. An inter- photogrammetric analysis of the tornado courtesy of esting contrast to these PPCG and potentially inverted E. Rasmussen is available online at www.nssl.no a a . storms is the severe multicell storm of 19 June, which gov /ssr /ind e x.ht m .) There also were multiple reports produced at most 10% +CGs. of large hail (up to 1.75 in. in diameter; Table 2), par- Based on Table 2, there are several case studies and ticularly after the midlife intensification and tornado comparisons that can be done to address the goals of touchdown. A photograph of the 29 June supercell the STEPS project. Many of these are either under- taken near the time of the tornado is shown in Fig. 2a. way or planned. A main focus of STEPS was The weather scenario on 5 July was consistent with supercells, and the project obtained data on three. To past conditions associated with LP supercell events, illustrate how the STEPS suite of observing platforms with a relatively quiescent synoptic environment and worked together to accomplish project goals, we have a strong dryline along the Kansas-Colorado border. selected two supercell cases for short overviews: a clas- CAPE and shear estimates east of the dryline sup- sic supercell that occurred on 29 June 2000, and an ported a forecast for supercellular storms. In addi- LP supercell that occurred on 5 July. tion, an outflow boundary from a convective system earlier that morning in eastern Nebraska had propa- Overv/ ews of the 29 June classic and 5 July LP supercells. gated westward into southwestern Nebraska. The sig- ME T E O R O L O GI C A L SCENARIOS. The afternoon of 29 June nificant storm on this day subsequently developed to 2000 saw an unstable air mass in western Kansas with the northeast of this dryline-outflow intersection, sufficient shear for supercell-type storms. Surface within the highly unstable air mass, and quickly de- dewpoints decreased toward the west into eastern veloped significant low and midlevel rotation and an

AMERICAN METEOROLOGICAL SO CIETY AUGUST 2004 BAI1 5* | 1 1 1 3

scribed in Thomas et al. (2003), except that only flashes consist- ing of five or more VHF sources were included in order to mini- mize contamination by noise. NLDN data provided CG flash rates. Both storms featured very high total flash rates (> 1 flash per second) and were dominated by +CG lightning for much of their lifetimes. The supercells often underwent pulsations in total flash rate, and sometimes also +CG flash rates (Fig. 3). Perhaps the most notable of these pulses is the intensification of 29 June between 2320 and 2340 UTC. This is also the period when the storm underwent its right turn and produced a tornado. Using polarimetric-based hy- drometeor identification with CSU-CHILL and S-Pol data, we can estimate the volume of radar echo occupied by hail as a func- tion of time and height. The hy- drometeor classification algo- rithm used is fuzzy-logic based, developed from Liu and Chandrasekar (2000), Straka et al. (2000), and Zrnic et al. (2001). In addition, by manually isolating FIG. 3. (a) T i m e series of t o tal an d cloud - to - ground lightning flash rates individual LMA flashes, we can for t h e 29 J u n 2000 classic su percell. ( b ) S a m e as (a), b ut for t h e 5 J ul determine the starting height of 2000 low - precipitation supercell. +CGs that occurred in the 29 June storm. The result of these efforts associated hook on the radar reflectivity field. A is shown in Fig. 4, which demonstrates that the in- photo taken during its mature phase (after 2330 crease in +CG production during 2320-2340 UTC UTC; Fig. 2b) shows many of the characteristics of was coincident with the presence of hail aloft. Addi- an LP supercell with a striated, bell-shaped c l o u d - tionally, the vast majority of positive CGs originated often indicative of a rotating updraft—and visually in the 0° to -20°C temperature range, which normally very little precipitation just to the north and east of is where the main negative charge region is found in the primary cloud. This contrasts with the more thunderstorms (Krehbiel 1986; Koshak and Krider amorphous cloud seen in the 29 June photo (Fig. 2a). 1989; Stolzenburg et al. 1998a,b). This storm produced a funnel cloud at 0000 UTC A more direct link between hail aloft and the oc- (Table 2), along with some large hail (up to 1.75 in. currence of +CG lightning is shown in Figs. 5 in diameter). (29 June) and 6 (5 July), which show horizontal and vertical cross sections of LMA VHF sources associ- LIG H T NIN G BEHAVIOR. We used the LMA and NLDN to ated with a single +CG flash from each storm, over- characterize the lightning behavior of observed storms. laid onto cross sections of radar reflectivity factor, Figure 3 shows total and CG lightning flash rate time multiple-Doppler-derived winds, and polarimetri- series for the (a) 29 June and (b) 5 July supercells. To- cally identified hail and graupel. These plots are typi- tal flash rates were determined by the methodology de- cal of many +CGs in these storms, which tended to

1 1 1 4 I BAF15 - A UGUST 2004

initiate in or near regions of hail and high-density by lightning discharges, or can confirm the lack of graupel aloft (initiation points are shown by white charge in parts of the storm untouched by lightning diamonds in Figs. 5 and 6), the latter being an inter- (such as the previously mentioned BWERs; e.g., mediate category between regular graupel and small Fig. 7). hail. We launched several EFM balloons into the These strong updrafts often coincided with 29 June and 5 July storms. To show how these sound- bounded weak echo regions (BWERs; Browning and ings are being used to improve our understanding of Donaldson 1963; Browning 1964, 1965) in the electrical structure in STEPS storms, we compare reflectivity field, as well as "holes" in VHF sources updraft soundings from these 2 days. Figure 8 shows detected by the LMA. Figure 7 shows an example of data from a 29 June updraft sounding, and Fig. 9 these phenomena from 29 June, with horizontal and shows a corresponding sounding for 5 July. Inside the vertical cross sections of LMA sources overlaid on updrafts of both storms, electric field magnitudes re- contours of reflectivity factor and multiple-Doppler- mained small and fairly constant (typically < 10 kV derived updraft speeds. The lack of VHF sources is m _ 1 in magnitude) until an altitude of ~8 km MSL, roughly collocated with the strong updraft in this confirming the inference of little charge in strong up- storm (note that we did not wind-advect VHF drafts based on LMA data (e.g., Fig. 7). In 29 June sources like the radar data during multiple-Doppler (Fig. 8), the electric field changed sign and indicated synthesis, so given the high translational speed of this a positive charge layer between 8 and 10 km MSL storm some mismatch is expected). The lack of VHF (near -20°C) as the balloon entered regions of heavier sources in strong updrafts has been observed before precipitation, including hail (reflectivity > 50 dBZ). by Ray et al. (1987) and Krehbiel et al. (2000) and sug- On 5 July (Fig. 9), however, the lowest charge region gests lack of a vigorous process of charge separation contained negative charge. But the next lowest charge in these regions. region on 5 July, encountered at a height of 8.5 km MSL, did have positive charge comparable to the low- ELE C T RI C FIELD OBSERVATIO NS. Information about the est positive charge on 29 June. electric fields in thunderstorms is an important The charge distribution on 5 July also was similar complement to LMA data, revealing information to that on 29 June in that both storms had more com- about the charge structure that would be otherwise plex charge structure during the descent in and near unavailable from lightning data alone (e.g., Coleman reflectivity cores than during the ascent through the et al. 2003). In particular, in situ measurements such strong updrafts, with charge extending to consider- as those from EFM balloons and T-28 storm penetra- ably lower altitudes in the reflectivity cores. Similar tions can be used to reveal more clearly regions of net differences in the complexity of charge structure be- negative charge, since as mentioned previously the tween strong updrafts and either weak updrafts or LMA is more sensitive to negative breakdown in re- other portions of storms have been reported previ- gions of net positive charge (Rison et al. 1999). ously by Marshall et al. (1995) and Stolzenburg et al. Electric field data also reveal charge layers not tapped (1998a,b; 2002).

FIG. 4. T i m e - h e i g h t cr oss sec - t i o n of 29 J u n 2000 hail e c h o v o l u m e ( n o t r a d a r r e f l ec t ivi ty factor) as d et er m i n e d f r o m both S - Po l a n d C H I L L p o l a r i m e t r i c data (r a d a r used at each t i m e is t h e o ne wi t h t h e best covera g e at t h a t t i m e ) via a h y d r o m e t e o r classification sc h e m e using fu z zy lo gic t e c h n i q u e s . A l s o s h o w n a r e t h e starting heights for indi - vidual + C G flashes p ro d uce d by t h e st o r m as d e t e r m i n e d by t h e L M A (black crosses), as well as notable t e m p e r a t u r e levels. T h e t o r n a d o occ u r r e d at 2328 U T C .

AMERIC A N METEOROLO GIC AL SO CIETY A UGUST 2004 BAI1 5* | 1 1 1 5

FIG. 5. (a) Hori z ontal cross section at 3 k m M S L of hyd r o m eteor identification ( H I D ) categories (shaded), radar reflectivity factor (line contours, starting at 10 d B Z with 15 d B Z increments), and L M A V H F sources (red dots, sources f r o m all altitudes projected onto t he hori z ontal plane) associated with a + CG flash (strike location denoted by a black cross) that occurred at 2338:18 U T C on 29 Ju n 2000. T h e location of the initial V H F source associated with the flash is indicated by t he large white diamond with the black border. T h e only H I D categories shaded are hail and high - density graupel related—lar ge hail mi x ed with rain (Lh + r; hail d ia meter > 2 c m), small hail mixed with rain (Sh + r; D < 2 c m), large hail ( L H ) , small hail ( S H ) , and high - density graupel ( H G ) . T h e polarimetric data are f r o m an S - Pol P P I vol u me starting at 2338 U T C , and the line indicates t he vertical cross section shown in (c). (b) S a m e as (a), but for 6 k m MSL . T h e T - 28 track for Fig. 10 is shown as t he thick blue curve, (c) Vertical cross section of H I D and radar reflectivity factor at 22 k m north of t h e K G L D radar, along wit h L M A sources and + C G strike location projected onto t he vertical plane. Legend is t he sa me as in (a) and (b).

The T-28 made several passes through the main draft core were free of hail (weak radar echo) and were updraft region of the 29 June storm at the 6 km MSL characterized by a weak negative vertical electric field

1

(-10°C) level during the transition period when the (magnitude < 10 kV nr ), suggesting relatively little storm produced a tornado and began producing fre- net charge in the updrafts. Unfortunately, the T-28 quent +CG lightning. The updraft and the electric was not available for the 5 July storm. field data from the last of these passes are shown in The T-28 provided the first in situ verification on Fig. 10. The corresponding track is superimposed on an LMA-mapped flash channel during an earlier the 6-km-horizontal cross section in Fig. 5 (aircraft 29 June penetration (not shown in Fig. 10). During a was moving northward). Although the multiple- pass across the downshear precipitation region, the Doppler wind syntheses (Fig. 7) suggest a single ex- T-28 encountered an intracloud lightning flash that tended updraft region around this time, the higher- was evident in both its electric field record and par- resolution aircraft data show multiple discrete tially imaged by the wing-mounted video camera. The updrafts at 6 km. Consistent with the balloon EFM LMA also detected this flash. Warner et al. (2003) observations (Fig. 8), all except the southernmost up- used the T-28 data to model the channel and estimate

1116 I BATLS* AUGUST 2004 FIG. 6. S a m e as Fig. 5, b ut for t h e 2332 U T C S - Pol r a d a r v o l u m e o n 5 J u l 2000. T h e LMA points a n d C G strike locatio n a r e for a + C G occ u rri n g at 2330:57 U T C . N o T - 28 t r a c k is sh own.

its location, charge density, and polarity. The analy- KINEMA TIC AND MICROPHYSICAL STRUCTURES. Figure 1 1 sis agreed with the LMA-inferred polarity and chan- shows the radar reflectivity factor and multiple- nel orientation. Doppler winds for the [(a),(b)] 29 June and Thus, the electric field data not only help to con- [(c),(d)] 5 July supercells during their mature phases firm inferences of the lack of charge in strong updrafts and demonstrates how similar the two supercells look in these storms, but also reveal complex charge struc- to a radar, despite obvious differences in their visible tures outside these main updrafts that are not evident cloud structures (Fig. 2). Both had strong updrafts in in the LMA data. In addition, the electric field data excess of 40 m s - 1 , midlevel rotation, low-level hook in concert with the LMA data show that both 29 June echoes, BWERs in the vicinity of the updrafts, and and 5 July had possibly inverted electrical structures, high peak reflectivity factors (> 50 dBZ), although with midlevel (near -20°C) positive charge in place overall, 5 July had lower peak reflectivity than 29 June. of midlevel negative charge. This inference is sup- One way to determine whether these gross radar- ported by preliminary modeling results for the based similarities mask important microphysical dif- 29 June storm by Kuhlman et al. (2003).These charge ferences is through the use of a numerical cloud structures appear to be related to positive CG produc- model. tion. We continue to use STEPS data to investigate Such models have been successful at reproducing these charge structures in more detail, as well as to the basic dynamical character of the observed con- investigate exactly how they arose and to understand vective storm spectrum (e.g., ordinary cells, their relationship to positive CG production. These multicells, supercells, squall lines, etc.; Weisman and efforts include not just the supercells of 29 June and Klemp 1986; Weisman et al. 1988), but have been far 5 July, but also other STEPS cases. less successful at reproducing the large variety of ob- served precipitation characteristics in any systematic

A MERIC A N METE OROLO GIC AL SO CIET Y A U G UST 2004 BAI1 5* | 1 1 1 7

FIG. 7. (a) Hori z ontal cross section of radar reflectivity factor (synthesi z ed f r o m both C H I L L and S - Pol; color contours) f r o m 2325 U T C on 29 J u n 2000. Also shown are LMA - d etecte d V H F source locations during 2325 - 2327 U T C and within 0.5 k m of the cross - sectional cut ( m a g en ta dots), as well as N L D N - detected + C G ground strike locations during the radar vo l u m e (black crosses). L M A data have not been advection - corrected. (b) S a m e as (a), except lacking lightning data and instead showing updraft speeds (black contours; every 10 ms" 1 starting at 10 m s 1 ) as estimated by multiple - Doppler synthesis, (c) Vertical cross section at sa m e t i m e showing radar reflectivity, LMA - d etecte d V H F source locations, and updraft speeds. Legen d is the sa m e as in (a) and (b).

or physically realistic manner (e.g., Weisman and Some initial idealized simulations have been com- Bluestein 1985). Additionally, numerical studies pleted for both the 29 June and 5 July supercell storms show great sensitivity in resultant convective struc- using the Weather Research and Forecast model ture, evolution, and precipitation output to relatively (WRF; http://wrf-model.org), with a 1-km (0.5 km) grid minor differences in microphysical schemes, casting spacing in the horizontal (vertical) direction over a much doubt on our current ability to forecast con- 120 km x 120 km x 22 km domain, and with the Lin vective precipitation in operational models (e.g., et al. (1983) microphysics parameterizations, which in- Gilmore et al. 2003). Numerical modeling of storms cludes six water species (water vapor, cloud water, cloud observed in STEPS is an important goal of the project ice, snow, rain, and graupel; Miller and Weisman in order to improve our understanding of the pre- 2002). Preliminary results indicate that the model is cipitation processes in supercells and other storms. able to replicate basic storm-scale properties, such as Observations from radar, the T-28, and soundings storm motion, orientation, and rotational character- can be used to "teach" the model to come as close as istics, but these same model results also highlight the possible (or as is practical) to the real storms. The difficulties in reproducing the microphysical charac- model results then can be used as the basis for a de- ter of the storms. For instance, while both storms ex- tailed analysis of precipitation formation. hibited low-level hook echoes and vaulted radar struc-

1 1 1 8 I BAF15 - AUGUST 2004

FIG. 8. Ve r t ic a l cross section of reflectivity at an a z i m u t h of 76° f r o m t h e C S U - C H I L L ra d a r at 0010 U T C on 30 J u n 2000, sh own w i t h t h e projectio n of electric field vect o rs in this plane for t h e balloon flight d urin g 0005- 0034 U T C . Electric field vectors, sh own in blue along t h e t r ac k (scale at t h e t o p), point f r o m t h e balloon t r ac k alo n g t h e d irecti o n a positive c h a r g e w o u l d m o ve ; t h e n u m b e r of vec t o rs has b ee n r e d uce d co nsi dera bly for clarity in t h e figure. R e d bars wi t h plus signs show t h e vertical e x te n t of positively charged regions inferred f r o m t h e electric field profile (including all available vect o rs) an d t h e lightning distribution, an d blue bars w i t h nega - tive signs sh ow t h e vertical e x t e n t of negatively char ged regions. T h e balloon location has b een c o r r ec t e d for st o r m m o t i o n t o d e t e r m i n e its path relative t o s t o r m str uct u re at t h e t i m e of t h e ra d ar scan. T h e vertical com - p o nent of electric field (£ z ), t e m p e r a t u r e ( T ), d ew p o i n t (T d ), ascent ra t e (Asc), an d relative hu midities ( R H an d

RH. c e) a r e sh own for t h e corresp o n d in g up an d d o w n soundings.

tures in the mid- to upper levels (Fig. 11), the simula- tions were not able to repro- duce the vaulted structures. As the model updrafts (> 50 m s - 1 ) were compa- rable to observations, the fault appears to lie with the bulk microphysical param- eterization used in the model, which requires that all like particles (e.g., grau- pel) fall with the same mean terminal velocity regardless of size. The simulations did

produce a much weaker FIG. 9. Ra d ar reflectivity, electric field, and inferred charge for t h e st o r m on 5 J ul low-level cold pool for 5 2000. (left) Ve r t ic a l cross section of reflectivity at an a z i m u t h of 4 5 ° f r o m t h e July than it did for 29 June, C S U - C H I L L radar at 0108 U T C on 6 Jul, shown with t h e projection of electric which may be consistent field vectors (electric field vect o r scale is t h e sa m e as in Fig. 8) in this plane for with the 5 July storm's hav- t h e balloon flight d urin g 0048 - 0127 U T C . T h e location of t h e balloon has been ing a more LP-type struc- correcte d for st o r m m otio n t o show t h e storm - relative track at t h e t i m e of t h e ture, but this result was sen- radar scan. Re d bars with plus signs show t h e vertical ex tent of positively charged regions in ferred f r o m t h e electric field profile an d t h e lightning distribution, sitive to changes in the an d blue bars w i t h negative signs show t h e vertical e x ten t of negatively charged microphysical parameters. regions, ( r i g h t) S t o r m - r e l a t i v e b allo o n t r a c k ( b l ac k line) s u p e r i m p o se d o n Future analyses will reflectivity at an elevatio n of 0.5° f r o m t h e N C A R S - Pol ra d ar at 01 19 U T C . consider observations from T h e origin in each panel is t h e location of t h e ra d ar t h a t acq uired t h e data.

AMERIC A N METEOROLO GIC AL SO CIETY A UGUST 2004 BAI1 5* | 1 1 1 9

STEPS documented more than 1200 TLEs, mostly sprites (Lyons et al. 2003a,b). Because STEPS featured a lightning mapper (the LMA) in close proximity to the Yucca Ridge Field Station, which recorded observations of TLEs, it pro- vided an opportunity to distin- guish sprite-parent +CGs (SP+CGs) from other light- ning flashes. During STEPS, a half-dozen remote locations coordinated measurements of ELF transient signatures, al- lowing for both global-scale geolocation (Price et al. 2002) and estimates of M of the FIG. 10. Vertical com ponent of the electric field and updraft are plotted vs SP+CGs (Hu et al. 2002). LMA t i m e f r o m the pass of the T - 28 through the core of the st o r m b etween mapping of SP+CGs provided 2339:00 and 2343:30 U T C . (b otto m) Four updraft cores are shaded in red. the height of the charge layers (top) T h e electric field magnitudes while the aircraft is in these cores are tapped by the CG strokes (Z g ; s h a d e d r e d w h e n p o si t ive a n d b l u e w h e n n e g a t i ve . In t h e f irst southeasternmost updraft there is hail and positive field, while in the re- Lyons et al. 2003b). Data from maining three cores, the last two of which are precipitation free, the field STEPS show that SP+CG tends to be negative. Field magnitudes are always less than 10 kV irr1 . A n flashes are associated with both abrupt field change due to nearby lightning is noted just before 2341:00 U T C . very high M q values (>1000 C " H a i l " indicates when hail was observed by the T - 28 microphysical sensors. km for 90% probability of sprites) and low-altitude Z^ the T-28 aircraft and inferences from the polarimet- values (~4 km AGL; Hu et al. 2002; Lyons et al. ric radar measurements to improve both the micro- 2003b). These results support the conceptual models physical parameterization schemes and (hopefully) of Williams (1998) and Huang et al. (1999), suggest- the simulated storm representations, especially cold- ing that the charge reservoir for SP+CGs would be pool production and distribution of precipitation rela- found within the lower portions of MCS stratiform tive to the updraft. In addition, we plan to examine regions and are consistent with past measurements of how the surrounding environment affects the kine- positive charge layer height in MCSs (e.g., Schuur et al. matic and microphysical structures of LP storms. 1991; Stolzenburg et al. 1994; Marshall et al. 2001). However, because the main differences between LP and other storms are likely to be microphysical, a key S T E P S O U T R E A C H A N D E D U C A T I O N . O u t remaining challenge in modeling LP storms is accu- reach to the general public was a key component of rate representation of the microphysics. STEPS. We scheduled a media day for the project that helped increase exposure to the general public. Sev- TLE observations. Within High Plains convection, eral reports on STEPS occurred in the national and sprites typically accompany only a small percentage international media. Locally, there were news broad- of +CG flashes, most often within the stratiform pre- casts on network-affiliate television stations in Colo- cipitation region of larger MCSs (Lyons 1996). Sprites rado and Kansas and stories in major regional news- appear to represent conventional dielectric break- papers. down in themesosphere (-70-75 km MSL) triggered Local community outreach efforts were organized by unusually large electric field transients from +CGs by the Colorado Climate Center at Colorado State below. Huang et al. (1999) noted that the key metrics University, in concert with an extension of the in sprite formation should be the magnitude of the CG Community Collaborative Rain and Hail Study lightning vertical charge moment (M ), the product (CoCoRaHS; www.cocorahs.org), which uses local of charge lowered to ground by a CG flash (C), and volunteer observers to report rain and hail measure- the height from which this occurs (Z ). ments. These efforts involved cooperation with local

1 1 2 0 I BAF15 - AUGUST 2004

FIG. II. (a) Hori z o ntal cross section at 3 k m M S L of ra dar reflectivity factor ( f r o m S - Pol; color shaded), mul - tiple - Doppler win d vectors, and in (b) and (d) updraft speeds (line contours; every 10 m s " 1 starting at 10 m s - 1) for the 29 J u n st o r m at 2325 U T C . (b) S a m e as (a), but for 8.5 k m M SL . (c) S a m e as (a), but for 2332 U T C on 5 Jul. (d) S a m e as (c) but for 8.5 k m M SL .

schools, whose students manufactured equipment for conducted. In addition, the Significant Opportunities deploying hail pads, which the volunteers used to in Atmospheric Research (SOARS; www.ucar.edu/ measure the number, size, shape, and density of hail- soars) program—cosponsored by the University Cor- stones. STEPS investigators also visited local schools poration for Atmospheric Research (UCAR), NSF, and gave presentations on the project to interested the U.S. Department of Energy (DOE), the National members of the community. Aeronautics and Space Administration (NASA), and STEPS provided research exposure to many un- NOAA—provided the opportunity for students to dergraduate and graduate students. Besides the par- work during the field campaign and do STEPS-related ticipation of students of STEPS investigators, a Na- research afterward. tional Science Foundation (NSF)-funded Research Finally, support from the NSF Informal Science Experience for Undergraduates (REU) program was Education program allowed production of a plan-

AMERIC A N METEOROLO GIC AL SO CIETY AUGUST 2004 BAI1 5* | 1 1 2 1

etarium program and home/classroom video, The A C K N O W L E D G M E N T S . The STEPS project was Hundred Year Hunt for the Red Sprite, featuring the funded by the National Science Foundation through the role of STEPS research in determining the atypical Physical , Aeronomy, and Lower Atmospheric nature of the sprite-parent lightning discharges (see Observing Facilities programs. In particular, STEPS would www.Sky-Fire.TV for details). not have occurred without the support and guidance pro- vided by Dr. Rod Rogers at NSF/ATM. NSF also funded C O N C L U D I N G R E M A R K S . The STEPS project the REU component of STEPS. NOAA and the National has provided the research community with observa- Center for Atmospheric Research provided significant sup- tions of the evolving kinematic, microphysical, and port for STEPS as well. In addition, the extensive collabo- electrical structures of a diverse array of thunder- ration with the National Weather Service, in particular the storms, including the primary targets of the experi- Goodland office, was a major key to the project's success. ment: supercells and predominantly positive CG The STEPS community appreciates the great cooperation (PPCG) storms. The project also provided both po- of the governments, schools, and general public of the cit- larimetric and in situ microphysical data to help im- ies of Burlington, Colorado, and Goodland, Kansas. Ken prove polarimetric radar-based hydrometeor classi- Cummins and Global Atmospherics, Inc., now part of fication and quantification schemes, as well as an Vaisala, generously provided the NLDN lightning data. opportunity to study the reasons transient luminous Data collection, as well as installation and teardown of events (TLEs) occur above thunderstorms. major facilities like the CSU-CHILL and S-Pol radars, was The cooperation between the NWS and atmo- primarily the result of the dedicated staffs of all of the in- spheric research communities, as well as outreach to strument platforms, as well as the tremendous number of the general public, were major goals of STEPS. These enthusiastic undergraduate and graduate students who two activities are increasingly identified as major fac- participated in the field campaign. People who made extraor- tors in a field project's overall success (e.g., Schultz dinary contributions to STEPS data collection and analysis et al. 2002), and the efforts to maximize outreach and include Eric Bruning, Larry Carey, Tim Hamlin, Jeremiah intercommunity cooperation during STEPS could Harlin, Sang-Hun Lim, Thomas E. Nelson, Walt Petersen, help provide a model for future field projects. Sarah Tessendorf, and Kyle Wiens. We thank Bruce Entwistle The combination of polarimetric and multiple- at the Goodland NWS office for providing additional severe Doppler radar observations, along with LMA-based storm reports not available in Storm Data. Steve Nesbitt pro- lightning mapping and in situ observations of elec- vided last-minute assistance with some figures in this paper. tric field structure, may provide new insights into the nature of predominantly +CG thunderstorms. For ex- ample, comparisons of the 29 June and 5 July REFERENCES supercells suggest important linkages between strong Bluestein, H. B., and C. R. Parks, 1983: A synoptic and updrafts, the development of large hail aloft, anomalous photographic climatology of low-precipitation severe charging in thunderstorm midlevels, and subsequent thunderstorms in the southern plains. Mon. Wea. production of positive CG flashes. These potential in- Rev., I l l , 2034-2046. terrelationships between thunderstorm kinematics, , and G. R. Woodall, 1990: Doppler-radar analysis microphysics, electrical structure, and lightning are of a low-precipitation severe storm. Mon. Wea. Rev., the subject of ongoing research, not only in the afore- 118, 1640-1664. mentioned supercells but also other STEPS storms. Branick, M. L., and C. A. Doswell III, 1992: An obser- Indeed, a key to unlocking the mystery of PPCG vation of the relationship between supercell structure thunderstorms, as well as other storms with inverted- and lightning ground strike polarity. Wea. Forecast- polarity charge structures, could be identifying and ing, 7, 143-149. understanding the c o m m o n kinematic, micro- Browning, K. A., 1964: Airflow and precipitation trajec- physical, and electrical features of all such thunder- tories within severe local storms which travel to the storms observed in STEPS. In addition, we need to right of the winds. /. Atmos. Sci., 21, 634-639. understand the key differences between these storms , 1965: Some inferences about the updraft within a and storms like that on 19 June 2000, which was se- severe local storm. /. Atmos. Sci., 22, 669-677. vere but produced mostly -CGs Because all of these , and R. J. Donaldson, 1963: Airflow and structure storms spanned different mesoscale environments, or- of a tornadic storm. /. Atmos. Sci., 20, 533-545. ganizational structures, and levels of severity, this is a Burgess, D. W., and R. P. Davies-Jones, 1979: Unusual major research task that is still ongoing. Look for more tornadic storms in eastern Oklahoma on 5 Decem- STEPS results in both current and future publications. ber 1975. Mon. Wea. Rev., 107, 451-457.

1 1 2 2 I BAF15 - AUGUST 2004

Carey, L. D., and S. A. Rutledge, 1998: Electrical and Electricity, Versailles, France, International Commis- multiparameter radar observations of a severe hail- sion on Atmospheric Electricity, 189-192. storm. /. Geophys. Res., 103, 13 979-14 000. Hu, W., S. Cummer, W. A. Lyons, and T. E. Nelson, , and , 2003: Characteristics of cloud-to-ground 2002: Lightning charge moment changes for the ini- lightning in severe and nonsevere storms over the tiation of sprites. Geophys. Res. Lett., 29, 1279, central United States from 1989-1998. /. Geophys. doi: 10.1029/2001GL014593. Res., 108, 4483, doi:10.1029/2002JD002951. Huang, E., E. R. Williams, R. Boldi, S. Heckman, W. A. , , and W. A. Petersen, 2003: The relationship Lyons, M. Taylor, T. E. Nelson, and C. Wong, 1999: between severe storm reports and cloud-to-ground Criteria for sprites and elves based on Schumann lightning polarity in the contiguous United States resonance observations. /. Geophys. Res., 104,16 943- from 1989 to 1998. Mon. Wea. Rev., 131, 1211- 16 964. 1228. Klemp, J. B., 1987: Dynamics of tornadic thunderstorms. Changnon, S. A., Jr., 1977: The climatology of hail in Ann. Rev. Fluid Mech., 19, 369-402. North America. Hail: A Review of Hail Science and Koshak, W. J., and E. P. Krider, 1989: Analysis of light- Hail Suppression, Meteor. Monogr. No. 38, Amer. ning field changes during active Florida thunder- Meteor. Soc., 107-128. storms. /. Geophys. Res., 94, 1165-1186. Coleman, L. M., T. C. Marshall, M. Stolzenburg, T. Krehbiel, P. R., 1986: The electrical structure of thun- Hamlin, P. Krehbiel, W. Rison, and R. J. Thomas, derstorms. The Earth's Electrical Environment, E. P. 2003: Effects of charge and electrostatic potential on Krider and R. G. Roble, Eds., National Academy lightning propagation. /. Geophys. Res., 108, 4298, Press, 90-113. doi: 10.1029/2002JD002718. , R. J. Thomas, W. Rison, T. Hamlin, J. Harlin, and Cummins, K. L., M. J. Murphy, E. A. Bardo, W. L. M. Davis, 2000: GPS-based mapping system reveals Hiscox, R. B. Pyle, and A. E. Pifer, 1998: A combined lightning inside storms. Eos, Trans. Amer. Geophys. TOA/MDF technology upgrade of the U.S. National Union, 81, 21-25. Lightning Detection Network. /. Geophys. Res., 103, Kuhlman, K. M., E. R. Mansell, C. L. Ziegler, D. R. 9035-9044. MacGorman, and J. M. Straka, 2003: Charging and Curran, E. B., and W. D. Rust, 1992: Positive ground lightning in simulations of the 29 June 2000 STEPS flashes produced by low-precipitation thunderstorms supercell. Proc. 12th Int. Conf on Atmospheric Elec- in Oklahoma on 26 April 1984. Mon. Wea. Rev., 120, tricity, Versailles, France, International Commission 544-553. on Atmospheric Electricity, 199-202. Davies-Jones, R. P., D. W. Burgess, and L. R. Lemon, Lin, Y.-L., R. D. Farley, and H. D. Orville, 1983: Bulk 1976: An atypical tornado producing cumulonimbus. parameterization of the snow field in a cloud model. Weather, 31, 336-347. /. Climate Appl. Meteor., 22, 1065-1092. Donaldson, R, A. Spatola, and K. Browning, 1965: Vi- Liu, H., and V. Chandrasekar, 2000: Classification of hy- sual observations of severe weather phenomena. A drometeors based on polarimetric radar measure- family outbreak of severe local storms—A ments: Development of fuzzy logic and neuro-fuzzy comprensive study of the storms in Oklahoma on systems, and in situ verification. /. Atmos. Oceanic 26 May 1963. Part I. Special Rep. 32, Air Force Cam- Technol., 17, 140-164. bridge Research Laboratory, 73-97. Lyons, W. A., 1996: Sprite observations above the U.S. Doswell, C. A., Ill, and D. W. Burgess, 1993: Tornadoes High Plains in relation to their parent thunderstorm and tornadic storms: A review of conceptual mod- systems. /. Geophys. Res., 101, 29 641-29 652. els. The Tornado: Its Structure, Dynamics, Prediction, , R. A. Armstrong, E. R. Williams, and E. A. Bering, and Hazards, Geophys. Monogr., No. 79, Amer. 2000: The hundred year hunt for the sprite. Eos, Geophys. Union, 161-172. Trans. Amer. Geophys. Union, 81, 373-377. Gilmore, M. S., J. M. Straka, and E. N. Rasmussen, 2003: , T. E. Nelson, R. A. Armstrong, V. P. Pasko, and Precipitation and evolution sensitivity in simulated M. A. Stanley, 2003a: Upward electrical discharges deep convective storms: Comparisons between liq- from thunderstorm tops. Bull. Amer. Meteor. Soc., 84, uid-only and simple ice and liquid phase microphys- 445-454. ics. Mon. Wea. Rev., in press. , , E. R. Williams, S. A. Cummer, and M. A. Hamlin, T., P. R. Krehbiel, R. J. Thomas, W. Rison, and Stanley, 2003b: Characteristics of sprite-producing J. Harlin, 2003: Electrical structures and storm sever- positive cloud-to-ground lightning during the 19 July ity inferred by 3-D lightning mapping observations 2000 STEPS mesoscale convective systems. Mon. during STEPS. Proc. 12th Int. Conf. on Atmospheric Wea. Rev., 131, 2417-2427.

AMERIC A N METEOROLO GIC AL SO CIETY AUGUST 2004 BAI1 5* | 1 1 2 3

MacGorman, D. R., and W. D. Rust, 1998: The Electri- storms. Proc. 12th Int. Conf. on Atmospheric Electric- cal Nature of Storms. Oxford University Press, ity, Versailles, France, International Commission on 422 pp. Atmospheric Electricity, 135-138. Marshall, T. C., W. D. Rust, and M. Stolzenburg, 1995: Schultz, D. M., and Coauthors, 2002: Understanding Utah Electrical structure and updraft speeds in thunder- winter storms: The Intermountain Precipitation Ex- storms over the southern Great Plains. /. Geophys. periment. Bull Amer. Meteor. Soc., 83, 189-210. Res., 100, 1001-1015. Schuur, T. J., B. F. Smull, W. D. Rust, and T. C. Marshall, , M. Stolzenburg, W. D. Rust, E. R. Williams, and 1991: Electrical and kinematic structure of the strati- R. Boldi, 2001: Positive charge in the stratiform cloud form precipitation region trailing an Oklahoma of a mesoscale convective system. /. Geophys. Res., squall line. /. Atmos. Sci., 48, 825-842. 106, 1157-1163. Stolzenburg, M., and T. C. Marshall, 1994: Testing mod- Mazur, V., E. Williams, R. Boldi, L. Maier, and D. E. els of thunderstorm charge distributions with Proctor, 1997: Initial comparison of lightning map- Coulomb's law. /. Geophys. Res., 99, 25 921-25 932. ping with operational time-of-arrival and interfero- , , W. D. Rust, and B. F. Smull, 1994: Horizontal metric systems. /. Geophys. Res., 102, 11 071-11 085. distribution of electrical and meteorological Miller, L. J., and M. Weisman, 2002: Comparison of conditions across the stratiform region of a mesos- radar-observed and WRF-modeled structures of two cale convective system. Mon. Wea. Rev., 122, 1777- STEPS storms. Proc., 21st Conf. on Severe Local 1797. Storms, San Antonio, Texas, Amer. Meteor. Soc., , W. D. Rust, and T. C. Marshall, 1998a: Electrical 299-302. structure in thunderstorm convective regions. Orville, R. E., 1994: Cloud-to-ground lightning flash Part II: Isolated storms. /. Geophys. Res., 103, characteristics in the contiguous United States: 1989- 14 079-14 096. 1991. /. Geophys. Res., 99, 10 833-10 841. , , and , 1998b: Electrical structure in thun- , and A. C. Silver, 1997: Lightning ground flash den- derstorm convective regions. Part III: Synthesis. /. sity in the contiguous United States: 1992-1995. Mon. Geophys. Res., 103, 14 097-14 108. Wea. Rev., 125, 631-638. , T. C. Marshall, W. D. Rust, and D. L. Bartels, 2002: , and G. R. Huffines, 2001: Cloud-to-ground light- Two simultaneous charge structures in thunder- ning in the United States: NLDN results in the first storm convection. /. Geophys. Res., 107, 4352, decade, 1989-1998. Mon. Wea. Rev., 129, 1179- doi: 10.1029/2001JD000904. 1193. Straka, J. M., D. S. Zrnic, and A. V. Ryzhkov, 2000: Bulk Price, C., M. Asfur, W. Lyons, and T. Nelson, 2002: An hydrometeor classification and quantification using Improved ELF/VLF method for globally geolocating polarimetric radar data: Synthesis of relations. J. sprite-producing lightning. Geophys. Res. Lett., 29, Appl. Meteor., 39, 1341-1372. 1031, doi: 10.1029/2001GL013519. Thomas, R., P. Krehbiel, W. Rison, J. Harlin, T. Hamlin, Rasmussen, E. N., and J. M. Straka, 1998: Variations in and N. Campbell, 2003: The LMA flash algorithm. supercell morphology. Part I: Observations of the role Proc. 12th Int. Conf. on Atmospheric Electricity, of upper-level storm-relative flow. Mon. Wea. Rev., Versailles, France, International Commission on 126, 2406-2421. Atmospheric Electricity, 655-656. Ray, P. S., D. R. MacGorman, W. D. Rust, W. L. Taylor, Vivekandan, J., S. M. Ellis, R. Oye, D. S. Zrnic, A. V. and L. W. Rasmussen, 1987: Lightning location rela- Ryzhkov, and J. Strata, 1999: Cloud microphysics tive to storm structure in a supercell storm and a mul- retrieval using S-band dual-polarization radar mea- ticell storm. /. Geophys. Res., 92, 5713-5724. surements. Bull. Amer. Meteor. Soc., 80, 381-388. Rison, W., R. J. Thomas, P. R. Krehbiel, T. Hamlin, and Warner, T. A., J. H. Helsdon Jr., and A. G. Detwiler, J. Harlin, 1999: A GPS-based three-dimensional 2003: Aircraft observations of a lightning channel in lightning mapping system: Initial observations in STEPS. Geophys. Res. Lett., 30, 1984, doi: 10.1029/ New Mexico. Geophys. Res. Lett., 26, 3573-3576. 2003GL017334. Rust, W. D., and D. R. MacGorman, 2002: Possibly in- Weisman, M. L., and H. B. Bluestein, 1985: Dynamics verted-polarity electrical structures in thunderstorms of numerically simulated LP storms. Preprints, 15th during STEPS. Geophys. Res. Lett., 29, 1571, Conf. on Severe Local Storms, Indianapolis, IN, Amer. doi: 10.1029/2001GL014303. Meteor. Soc., 167-171. , , P. R. Krehbiel, R. J. Thomas, E. C. Bruning, , and J. B. Klemp, 1986: Characteristics of isolated and S. A. Stroman, 2003: The status of our search for convective storms. Mesoscale Meteorology and Fore- inverted polarity electrical structures in thunder- casting, P. Ray, Ed., Amer. Meteor. Soc., 331-357.

1 1 2 4 I BAF15 - AUGUST 2004

, and L. J. Miller, 2000: An overview of the severe , 2001: The electrification of severe storms. Severe thunderstorm electrification and precipitation study Convective Storms, Meteor. Monogr. No. 49, Amer. (STEPS). Preprints, 20th Conf. on Severe Local Meteor. Soc., 527-561. Storms, Orlando, FL, Amer. Meteor. Soc., 273-276. Winn, W. P., and L. G. Byerley III, 1975: Electric field , , and R. Rotunno, 1988: The structure and evo- growth in thunderclouds. Quart. J. Roy. Meteor. Soc., lution of numerically simulated squall lines. /. Atmos. 101, 979-994. Sci., 45, 1990-2013. Zajac, B. A., and S. A. Rutledge, 2001: Cloud-to-ground Williams, E. R., 1989: The tripole structure of thunder- lightning activity in the contiguous United States storms. /. Geophys. Res., 94, 13 151-13 167. from 1995 to 1999. Mon. Wea. Rev., 129, 999-1019. , 1998: The positive charge reservoir for sprite-pro- Zrnic, D. S., A. Ryzhkov, J. Straka, Y. Liu, and J. ducing lightning. /. Atmos. Sol. Terr. Rhys., 60, 689- Vivekandan, 2001: Testing a procedure for automatic 692. classification of hydrometeor types. /. Atmos. Oceanic TechnoL, 18, 892-913.

Find out wkat is about to be THE published in A M S journals! Manuscripts that have been accepted for publication, received at AMS Headquarters, and that are currently in some phase of the production process are listed on the AMS Web PTA. At a minimum, for each AMS journal, the Web PTA provides the title of the paper, PTA the lead author, and the month the manuscript was received at AMS Headquarters. In addition, the corresponding author's e-mail address (when available) is provided, allowing those interested in a paper to contact the author directly. Abstracts and full text! All manuscripts accepted for the Journal of Climate "Letters" section are available in full text on the Web PTA, and many other authors are now providing the abstract and/or a link to the full text of their accepted manuscript on the Web PTA—allowing access to their research results months before formal publication in the journal. Come take a look to see kow A M S ' s Wet P T A can be or use to you! http:/ / www.anietsoc.org/journal__abstracts

AMERIC A N METEOROLO GIC AL SO CIETY AUGUST 2004 BAI1 5* | 1 1 2 5

WEATHERING THE STORNS SVERRE PETTERSSEN, THE D-D A Y FORECAST, AND THE RISE OF MODERN METEOROLOGY

Meteorology today is the beneficiary of Jfarftt: the fundamental work in weather analysis Jberw Pe / feritM) f/te Pvwtuf, and forecasting of Sverre Petterssen attrff/je jQ/je cfMto&n* Mefcow/cyy (1898-1974), a giant in the field and an Edited by James Rodger Fleming international leader in meteorology during its formative era. In this lively and insightful autobiographical memoir, written just before his death, Petterssen shares intimate memories from his childhood in Norway, his education and service with the famous Bergen school of meteorology, and his extensive experiences in polar forecasting and as head of the meteorology department at MIT. The crisis of World War II comes alive in his passionate recounting of how forecasts were made for bombing raids and special operations, including the contentious AMERIC A N METEOROLO GICAL SO CIETY forecasts for D-Day. Sverre Petterssen's complete autobiographical memoir, published here for the first time in English, offers a fascinating view of a man, an era, and a science. Anyone interested in weather, World War II, the history of science, or Norwegian history will enjoy this book.

Weathering the Storm: Sverre Petterssen, the D-Day Forecast and the Rise of Modem Meteorology, ISBN 1-878220-33-0, 326 pp., hardbound, $70 list/$50 member. To place an order, submit your prepaid orders to: Order Department, AMS, 45 Beacon Street, Boston, MA 02108-3693; call 617- 227-2425 to order by phone using Visa, Mastercard, or American Express; or send e-mail to [email protected]. Please make checks payable to the American Meteorological Society.

1 1 2 6 I BAF1 5 - AUGUST 2004