STEPS BAMS Article Working Draft

Last Updated 26 March 2004

***LATEST UPDATES***

Final Version - 26 Mar 04

• Manuscript

• Manuscript
• Original Figures

• The submitted article is now available as a Word document via this link. (2 MB download)
• The Adobe PDF version is here. (1 MB download)

Severe thunderstorms, due to their propensity to injure, kill, and cause extensive property damage, are a primary concern to not only weather forecasters but also the public. However, these storms remain a puzzling scientific and forecasting problem, as they exhibit a wide range of electrical activity as well as a diversity of precipitation type and amount. Indeed, the incomplete representation of precipitation in convective storms remains a significant impediment to improving the quantitative forecast of warm season precipitation nationwide (e.g., Fritsch et al. 1998, Droegemeier et al. 2000).

One of the more intriguing severe storm types in this regard is the supercell thunderstorm (Browning 1964). In its most pristine state, a supercell is composed of a single, long-lived, rotating updraft that frequently produces large hail, high winds, prolific lightning, and occasionally tornadoes. While the basic dynamics of supercells seem well understood (e.g., Klemp 1987), these storms exhibit a wide variety of precipitation characteristics that is not well understood. For instance, supercells have been classified as either low-precipitation (LP; Donaldson et al. 1965, Davies-Jones et al. 1976, Burgess and Davies-Jones 1979, Bluestein and Parks 1983), classic or medium-precipitation (MP), and heavy-precipitation (HP; Doswell and Burgess 1993, Rasmussen and Straka 1998) based on differences in overall precipitation characteristics. Perhaps the least understood among these storms are LP supercells, which characteristically produce a huge anvil, some large hail, but appear to produce little rain. The visible cloud is a skeleton compared with other supercell storms, and rarely has a visible rain shaft (Bluestein and Parks 1983, Bluestein and Woodall 1990).

Supercell updrafts generally are too strong to allow much precipitation growth in a single upward pass. Therefore, some form of recirculation of embryonic precipitation is required to produce larger-sized particles that fall out as raindrops, graupel, or hail (Browning 1977, Nelson 1983, Miller et al. 1988, 1990). One possible explanation as to why some supercells produce large hail with very little rain while others might produce large amounts of rain and hail of all sizes is that environmental shear (e.g., Marwitz 1972a,b) and storm-relative flow in the upper levels (Rasmussen and Straka 1998) modulate the recycling process. Consistent with the notion that airflow affects hail production, Nelson (1987) proposed that severe hailfall events are critically dependent on kinematic structure rather than microphysical factors. Therefore, better understanding of the workings of supercells should illuminate the mechanisms that influence storm precipitation efficiency in general, as well as the feedbacks between precipitation production and storm dynamics.

Another unusual aspect of severe convective storms, including supercells, is their tendency to produce large amounts of positive cloud-to-ground (+CG) lightning (e.g., Branick and Doswell 1992), in contrast with normal warm-season thunderstorms that transfer mostly negative charge to ground through their CG lightning (Orville 1994, Orville and Silver 1997). A major question is the location of the source charge regions for +CG flashes in these storms, and how those charge regions developed. Most normal convection is generally thought to have an approximately tripolar charge structure, with a small amount of lower positive charge below major mid-level negative (generally considered to be the origin location of most negative CG lightning) and upper-level positive charge layers (Williams 1989), although a more complex electrical structures may exist, particularly in thunderstorm downdrafts (Stolzenburg et al. 1998).

A detailed review of +CG hypotheses is provided in Williams (2001). Some researchers have posited an enhanced low-level positive charge layer as being responsible for most +CGs. A similar possibility is an inverted dipole, with mid-level positive charge underlying upper-level negative (MacGorman and Nielsen 1991, Williams et al. 1991). Other work points toward upper-level positive charge that is unshielded either due to falling precipitation (Carey and Rutledge 1998) or strong wind shear (tilted dipole; MacGorman and Nielsen 1991, Branick and Doswell 1992, Curran and Rust 1992). All of these hypotheses suggest interesting yet poorly understood relationships between precipitation formation, airflow dynamics, and lightning production in +CG thunderstorms.

The ability to understand these relationships, however, requires sophisticated tools to observe and analyze thunderstorm characteristics. In particular, for precipitation, research with polarimetric meteorological radars has led to an emerging capability for identifying hydrometeor types remotely (Vivekanandan et al. 1999, Liu and Chandrasekar 2000, Straka et al. 2000). Such work began with efforts to discriminate between hail and rain, but as these radars have become more sophisticated, the number of measurable variables and thus the number of potential discriminants has increased. Some algorithms distinguish between such diverse hydrometeor types as large and small hail, graupel, snow, and mixed-phase precipitation. Hydrometeor identification can be useful in various applications to weather forecasting, aviation weather warnings, as well as in fundamental studies of storm structure and evolution. However, like all remote sensing techniques, polarimetric hydrometeor classification needs in situ verification to establish and improve the scope of its validity.

During May-July 2000 a major field project was conducted in the High Plains, near the Colorado-Kansas border, in order to investigate all of the above issues. The broad goal of the Severe Thunderstorm Electrification and Precipitation Study (STEPS; Weisman and Miller 2000; http://box.mmm.ucar.edu/pdas/STEPS.html) was to achieve a better understanding of the interactions between kinematics, precipitation production, and electrification in severe thunderstorms. Specific scientific objectives included: 1) understanding the apparent major differences in precipitation output from supercells that have led to them being classified as LP, MP, and HP; 2) understanding lightning formation and behavior in storms, and how it differs among storm types, particularly to better understand the mechanisms by which storms produce predominantly +CG lightning; and 3) to verify and improve microphysical interpretations from polarimetric radar.

In addition to these major research objectives, STEPS provided an opportunity to examine some related issues. The emphasis on +CG lightning enabled research into what is different about the small subset of +CGs from certain storms which trigger mesospheric transient luminous events (TLEs) such as sprites (Lyons et al. 2000, 2003a,b; Williams 1998). In addition, the emphasis on polarimetric radar observations allowed research into how precipitation forms in growing cumulus clouds.

2. STEPS Design and Execution

The STEPS project centered on a unique suite of complementary observing platforms in eastern Colorado and western Kansas. This portion of the High Plains region of the United States has been observed to climatologically favor supercell storms, particularly of the LP variety (Bluestein and Parks 1983). This is primarily due to the warm-season presence in this region of the Dry Line, the boundary between moist air from the Gulf of Mexico and drier continental air, which has been strongly associated with the occurrence of LP storms (Bluestein and Parks 1983). This region also is favorable for thunderstorms that produce 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 ideally located for studying the storms of interest.

The field measurements and analysis for STEPS were specifically designed to explore the mechanisms of precipitation formation and lightning production in supercell storms. The instrumentation (Table 1) included two S-band polarimetric radars, the Colorado State University CSU-CHILL and the NCAR S-Pol - along with the NWS WSR-88D Doppler radar at Goodland, KS - for determining the internal airflow and precipitation structure of storms. The deployable Lightning Mapping Array (LMA) from New Mexico Institute of Mining and Technology was used to map the three-dimensional total lightning distribution, while the National Lightning Detection Network (NLDN) provided CG flash data. The South Dakota School of Mines and Technology (SDSMT) armored T-28 aircraft was used to penetrate storms and provide in situ microphysical, electric field, and particle charge data. Mobile sounding systems from NOAA/NSSL were used to obtain balloon-borne measurements of electric fields inside storms (EFM balloons). NCAR mobile sounding systems (M-GLASS) and NOAA/NSSL Mobile Mesonet vehicles were used to characterize the storm environment. Finally, the Yucca Ridge Field Station (YRFS), located a few hundred km NW of the STEPS domain, provided observations of TLEs during STEPS.

The basic geographical layout of the project is shown in Fig. 1. For more information on each of these observing platforms see Table 1. The combination of all of these observations provided a thorough depiction of the co-evolving kinematic, microphysical, and electrical structures of STEPS thunderstorms, along with an understanding of each storm's mesoscale environment. Due to the detailed observing network, STEPS provided the best opportunity to date to answer key questions about precipitation formation within and electrification of severe storms. Additionally, the presence of two polarimetric radars and in situ observations provided a unique opportunity to evaluate and improve radar-based hydrometeor identification and quantification algorithms.

The Operations Center for STEPS was the CSU-CHILL radar facility, which was temporarily re-located from its home base at Greeley, CO, to Burlington, CO, during the summer of 2000. Mobile facilities and STEPS personnel generally were based out of Burlington, CO, and Goodland, KS. STEPS received excellent support from the local NWS forecast office in Goodland, KS (see sidebar), and daily forecast and observational platform status briefings occurred each morning at this NWS facility.

Based on each briefing, operations plans were formulated for the afternoon and evening. The research radars (CSU-CHILL and S-Pol) typically were running surveillance scans by noon. When convection was forecast, M-GLASS soundings were released and vehicle platforms (Mobile Mesonet, EFM balloons) were deployed in strategic locations near where activity was expected. Once convective targets were identified the vehicles and T-28 aircraft were vectored to the storm via radio communications with the operations center. In addition, the research radars would begin synchronized sector-based PPI and RHI scans of the target storm.

The main focus of observations were storms that occurred within or passed through the dual-Doppler lobes formed by each radar pair within the STEPS network (see Fig. 1). Of these, the highest priority was given to supercell storms, especially those with LP characteristics, as well as thunderstorms observed to be producing predominantly +CGs. The two research radars, CSU-CHILL and S-Pol, provided both polarimetric and velocity information while the KGLD WSR-88D operational radar provided additional velocity measurements for wind syntheses, both dual- and triple-Doppler depending on storm location.

Despite a drought during much of the operations period, STEPS investigators were able to obtain unprecedented data on a wealth of diverse cases, including LP storms, supercells, and mesoscale convective systems (MCSs), among others (Table 2). Many of the storms produced predominantly positve CG lightning during all or a portion of their lifetimes, and also exhibited unusual electrical structures, such as a possibly inverted dipole.

3. Current STEPS Research

Although a variety of storms passed through the network, supercells were the main focus of the study. Therefore, we have selected two cases to represent the range of supercells observed: one a classic supercell that occurred on 29 June 2000, which was observed by every available platform, and the second an LP supercell that occurred a few days later on 5 July.

a. Overview of the 29 June Classic Supercell

The weather scenario for the afternoon of 29 June 2000 was characterized by an unstable airmass in western Kansas, with temperatures near 30 °C and dew points near 15 °C. Winds were 10-15 m s^-1 from the south at the surface, veering to 15-25 m s^-1 from the northwest aloft, producing sufficient shear for supercell-type storms. Surface dew points decreased toward the west into eastern Colorado, but a distinct dry line was not evident. A short line of convective cells developed around 2200 UTC in the northwest corner of Kansas as a weak upper-level disturbance moving southeastward out of Wyoming approached the more unstable airmass. The convection subsequently moved southeastward, remaining in a multicellular phase for nearly 1.5 hours before making a 35° right turn as it became more supercellular in character.

Around this time storm size and radar reflectivity increased dramatically and a tornado first touched ground (2328 UTC). The tornado was on the ground for about 16 min and was tracked by the Mobile Mesonets throughout its lifetime. (A description and photogrammetric analysis of the tornado courtesy of Erik Rasmussen is available at http://www.nssl.noaa.gov/ssr/index.htm.) The mid-life intensification of the storm radically altered its kinematics, microphysics, and lightning production, based on a detailed analysis of the radar and lightning data (Tessendorf et al. 2003, Wiens et al. 2003). Prior to the intensification, there was little CG lightning of either polarity and little radar evidence of hail aloft. After intensification and the right turn, large hail began to dominate radar returns aloft (as revealed by polarimetric data) and the storm simultaneously began producing large numbers of +CG flashes. There also was an increase in total lightning activity.

During its lifetime this storm underwent several convective surges, with updraft speeds peaking near 50 m s^-1 as estimated by multiple-Doppler synthesis (Tessendorf et al. 2003). The most important of these surges was the mid-life intensification mentioned earlier. Peaks in hail production aloft, largely around the altitude of -10 °C, were well correlated with the convective surges, as was +CG lightning production (Fig. 2; surges A and C in Fig. 3). The +CG discharges were usually initiated on the edge of hail regions and progressed into the hail regions and/or into the downshear part of the storm (e.g., Fig. 4). Trends in total flash rates for this storm closely followed trends in volumes of updraft, reflectivity, and hail with maximum flash rates near 300 min^-1.

Pulsations in updraft strength also closely matched observations of bounded weak echo regions (BWERs; Browning and Donaldson 1963, Browning 1964, 1965) in the reflectivity field, as well as "holes" in VHF sources detected by the LMA (Fig. 5). The ring-like structure of the lightning hole and the BWER indicates that charged precipitation particles were "wrapped around" the updraft by the storm rotation. The tornado occurred on the western side of the lightning hole and updraft region. Similar observations were obtained with the LMA from a storm producing an F0 tornado during the MEaPRS Project in Oklahoma in 1998 (Krehbiel et al. 2000).

The LMA observations indicate that, in its very initial stages, the storm contained a lower positive and mid-level negative charge; i.e., the lower half of a normal tripole structure (Fig. 6a; interval 'a' of Fig. 3). The inferred positive charge regions were well-correlated with the locations of significant graupel and hail concentrations, suggesting that the charging resulted from a charge reversal microphysics process (e.g., Takahashi 1978). As the storm evolved, the lightning flash rate quickly increased and indicated the rapid development of alternating positive and negative charge layers above the lower dipole (Fig. 6b; interval 'b' of Fig. 3). The storm then went on to develop a dominant, deep mid-level positive charge region, with negative charge above the positive (the inverse of a normal polarity dipolar structure) and produced numerous inverted polarity IC flashes between the two charge regions (e.g., Fig. 4). The inverted electrical structure developed less than an hour into the storm and persisted for the remainder of the storm's life.

Several EFM balloons were launched into the 29 June storm. The electric field data in Fig. 7 were obtained from the second balloon flight of the day. The launch was into the strong updraft of the storm's mesocyclone (sounding path shown in Fig. 7). Inside the updraft, electric field magnitudes remained small and fairly constant until an altitude of 8 km MSL. Both horizontal and vertical components of the electric field increased steadily from 8 km MSL to approximately 10 km MSL, with the change in electric field indicating the presence of positive charge in that region. This charge layer began as the balloon entered regions of heavier precipitation, including hail. At least three charge regions of alternating polarity were detected in the updraft below 12 km. The balloon subsequently descended into the downdraft portion of the storm, with the sounding showing a similar distribution of charge as in the upper part of the storm, but situated lower in altitude. Overall, these observations are consistent with the inferred charge regions based on the LMA data.

Data and conclusions from this flight are presented in detail in MacGorman and Rust (2003). As discussed in that study, one characteristic of all the mesocyclone soundings in significant updrafts is the almost complete absence of charge in the weak echo region of the storm, the absence extending much higher than observed in typical storm updrafts. This was consistent with lightning "hole" observed by the LMA (Fig. 5). MacGorman et al. (1989) hypothesized that this absence of charge at lower altitudes in the strong updraft of mesocyclones is caused by the relative lack of the precipitation, which appears to be the major carrier of charge at that altitude.

The T-28 made several passes through multiple flanking towers and the main updraft region of the 29 June storm at the 6 km (-10 °C) level during the 45-minute time interval when the storm produced a tornado and began to produce frequent +CG lightning. Consistent with the balloon EFM and LMA observations, the freshly developed updraft cores were basically free of ice and charge at this level, with negligible electric fields. A radar depiction of the storm during a pass through the main updraft region is shown in Fig. 8. The aircraft encountered four updraft maxima in this pass, separated by regions of reduced updraft and enhanced concentrations of precipitation-size particles (Fig. 9). The last two cores were precipitation-free at this level. Updrafts were estimated by the T-28 sensors to reach 35 m s^-1, in good agreement with a nearly concurrent multiple-Doppler synthesis and EFM sounding.

During a penetration between 2333 and 2338 UTC the T-28 encountered an intracloud lightning flash that was evident in both the electric field record and partially imaged by the wing-mounted video camera. This flash also was detected by the LMA. Warner et al. (2003) used the LMA data and video image, to model the channel as a line source oriented in front of and angled below the T-28 flight track. By adjusting the charge density and channel orientation combined with an assumed inverted tripole main charge distribution, they were able to match the observed electric field components, allowing them to estimate the channel's location, charge density, and polarity (negative). The analysis agrees with the LMA-inferred polarity and channel orientation and constitutes the first in situ verification of LMA observations.

b. Overview of the 5 July LP Supercell

The weather scenario on 5 July was consistent with past conditions associated with LP supercell events, with a relatively quiescent synoptic envrionment and a strong dry line becoming established along the Kansas-Colorado border by mid-afternoon. Surface temperatures peaked above 35 °C with dew points in the 0-5 °C range west of the dry line, with temperatures around 35 °C and dew points above 15 °C east of the dry line. More notably, an outflow boundary from a convective system earlier that morning in eastern Nebraska had propagated westward into southwestern Nebraska, with surface temperatures around 30 °C, and dew points above 20 °C. M-GLASS soundings taken south and west of this outflow boundary exhibited minimal CAPE and vertical wind shear. However, a sounding taken at 00 UTC within the modified airmass north of Goodland exhibited extreme instability, with CAPE values between 4000-5000 J kg^-1, despite a stable layer near the surface, and also exhibited significant vertical wind shear in the lowest several kms AGL, due in large part to the 10-15 m s^-1 ENE surface flow veering to modest SW flow aloft, allowing for the possibility of supercell-type activity.

The significant storm on this day subsequently developed to the northeast of this dry line-outflow intersection, within the highly unstable airmass, and quickly developed significant low- and mid-level rotation and an associated hook in the radar reflectivity field (Fig. 10). A photo taken during its mature phase (Fig. 11) shows many of the characteristics of a LP supercell with a striated, bell-shaped cloud - often indicative of a rotating updraft - and visually very little precipitation just to the north and east of the primary cloud. This storm did not produce a tornado (although tornado warnings were issued based on its radar structure), but did produce some large hail along with mostly +CG lightning. A smaller storm (hereafter referred to as the weaker storm) - further to the west and closer to the dry line - also exhibited LP characteristics, but dissipated as outflow from the primary storm further to the east apparently cut off its warm inflow.

The EFM ballooning crew launched three balloons into the 5 July storm. Figure 12 shows electric-field data from the second balloon flight of the day. This balloon was launched southeast of a large wall cloud, in outflow from the storm. Eventually it entered the inflow region and was caught in a strong updraft. At the location of the balloon, updraft speed increased steadily to a maximum of roughly 20 m s^-1 at an altitude of 8 km MSL. The balloon was damaged at an altitude of 13 km MSL and descended in and near a core of large reflectivity. Electric-field data continued to be collected down to at least 4 km MSL, and thermodynamic and tracking data were collected down to 1.5 km MSL.

The charge distribution inferred from the upward part of this sounding satisfied criteria suggested by Rust and MacGorman (2002) for classifying the storm's charge structure as inverted-polarity. The lowest charge encountered as the balloon rose in the updraft was similar to the lowest charge encountered by the updraft sounding for the 29 June storm in that neither was detected until the balloon reached an altitude of roughly 8 km MSL. On 5 July, however, the lowest charge region, instead of containing a large density of positive charge as on 29 June, contained a relatively small density of negative charge. The next lowest charge region on 5 July encountered at a height of 8.5 km MSL, did have a high density of positive charge, comparable to the density of the lowest positive charge on 29 June.

The charge distribution on 5 July was also similar to that on 29 June in that both storms had more complex charge structure during the descent in and near reflectivity cores than during the ascent through the weak-echo region of the updraft. As one might expect, charge extended to considerably lower altitudes in the reflectivity cores than in the updraft. Similar differences in the complexity of charge structure have been reported previously by Stolzenburg et al. (1998) and by Marshall et al. (1995), who compared soundings in strong updrafts with soundings in weak updrafts. However, these two previous studies did not compare soundings from the same storm. For more details and additional results from soundings of the 29 June and 5 July storms, see MacGorman and Rust (2003).

The 5 July case provided ample data with which to evaluate the three main applications of polarimetric radar: 1. retrieval of the parameters of the rain drop size distribution (DSD); 2. estimation of rain rate (R); and 3. classification of hydrometeor types (Bringi and Chandrasekar 2001; chapters 7 and 8). The radar measurements of reflectivity factor (Z_h), differential reflectivity (Z_dr), and specific differential propagation phase (K_dp) can be used to estimate the three parameters of a normalized gamma DSD form, namely the median volume diameter (D₀), shape (μ) and the generalized intercept (N_w). Radar retrieval of the gamma DSD parameters, are of great importance in characterizing different rain types (Bringi et al. 2003). Hydrometeor classification involves the use of a fuzzy logic methodology employing additional radar measurements such as linear depolarization ratio (LDR) and copolar correlation coefficient (ρ_co), as well as temperature, to infer the bulk hydrometeor type such as rain, graupel, hail, etc. (Bringi and Chandrasekar 2001; chapter 7).

The rain DSD during the mature phase of the weaker 5 July storm was analyzed using low-elevation-angle scans from SPOL. Figure 13 shows histograms of (a) D₀ and (b) log(N_w) within the main echo region of the storm using the methodology described by Bringi et al. (2002a). The standard deviation of D₀ (σ) is ~1 mm, significantly larger than values in more typical Colorado convective rain events (σ~0.5-0.6 mm). The mode of N_w in Fig. 13 is near 400 mm^-1 m^-3, which may be compared to 1700-2000 for continental convective rain (Bringi et al. 2003). Thus, the rainfall in this storm event of 5 July is characterized by an unusual DSD; that is, a very low concentration of larger-sized drops even during its mature phase - an interesting observation given the LP characteristics of this storm. The multiplicative coefficient a in the Z=aR^1.5 relation for this event would be in the range 600-800 as compared to the more usual range 150-300. Polarimetric radar can be used to continuously estimate the coefficient a and is a practical application of DSD retrieval relevant to the potential upgrade of the WSR-88D for dual-polarization (Bringi et al. 2002b).

Examples of hydrometeor classification using the CSU fuzzy logic scheme (Liu and Chandrasekar 2000) are presented in Fig. 14 for the two storms occurring on 5 July. Figure 14a is for the weaker storm (same as the DSD retrieval shown earlier) and Fig. 14b is for the strong LP supercell. The results for the weaker storm are typical for a thunderstorm; however, the difference between a normal thunderstorm and the case shown here is brought out in the DSD statistics shown earlier in Fig. 13. Thus, both classification and quantification are needed for characterizing storm types.

The classification results for the stronger LP supercell show some important features previously noted in severe hail-producing storms such as a large area of hail extending from surface to 10 km, significant graupel near the surface, a region of supercooled rain in the 3-7 km AGL altitude layer, and a region of hail situated above the supercooled rain. Within the main precipitation shaft, there are indications of heavy rain mixed with hail. These and other polarimetric radar features have been reported in greater detail by Conway and Zrnic (1993), Hubbert et al. (1998), Kennedy et al. (2001) and Loney et al. (2002), among others (see also chapter 7 of Bringi and Chandrasekar 2001 for a review).

The DSD statistics of rainfall as well as the hydrometeor type classification in storms using polarimetric radar data appear to be very useful for validation and possible improvement of the microphysical parameterization schemes used in idealized model simulations of such storms.

c. Additional STEPS Research

1) NUMERICAL MODELING

Numerical cloud models have been successful at reproducing the basic dynamical character of the observed convective storm spectrum (e.g., ordinary cells, multicells, supercells, squall lines, etc.; Weisman and Klemp 1986, Weisman et al. 1988), but have been far less successful at reproducing the large variety of observed precipitation characteristics in any systematic or physically realistic manner (e.g., Weisman and Bluestein 1985). Additionally, numerical studies show great sensitivity in resultant convective structure, evolution, and precipitation output to relatively minor differences in microphysical schemes, casting much doubt on our current ability to forecast convective precipitation in operational models (e.g., Gilmore et al. 2003). Thus, numerical modeling of STEPS storms is an important goal of the project. Observations from radar, the T-28, and soundings can be used to "teach" the model to come as close as possible (or practical) to the real storms. The model results then can be used as the basis for a detailed analysis of precipitation formation.

Some initial idealized simulations have been completed for both the 29 June and 5 July supercell storms using the Weather Research and Forecast model (WRF; http://wrf-model.org), with a 1-km (0.5-km) grid spacing in the horizontal (vertical) directions over a 120 km x 120 km x 22 km domain, and with the Lin et al. (1983) microphysics parameterization, which includes six water species (water vapor, cloud water, cloud ice, snow, rain, and graupel; Miller and Weisman 2002). Preliminary results indicate that the model is able to replicate basic storm-scale properties, such as storm motion, orientation, and rotational characteristics, but these same model results also highlight the difficulties in reproducing the microphysical character of the storms. For instance, while both storms exhibited low-level hook echoes and vaulted radar structures in the mid- to upper levels, the simulations were not able to reproduce the vaulted structures. The simulations did produce a much weaker low-level cold pool for 5 July than it did for 29 June, which may be consistent with the 5 July storm having a more LP-type structure, but this result was found to be sensitive to minor changes in the microphysical parameters. Future analyses will consider observations from the T-28 aircraft and inferences from the polarimetric radar measurements to improve both the microphysical parameterization schemes and (hopefully) the simulated storm representations, especially cold-pool production and distribution of precipitation relative to the updraft.

2) PRECIPITATION IN NASCENT CONVECTION

As mentioned earlier, investigation of precipitation development in growing cumulus clouds was an ancillary goal of STEPS. In a climate regime like that of the STEPS domain, the earliest-detectable precipitation echoes in the nascent stages of convective clouds have not been examined with an S-band polarimetric radar. An exploratory examination of the entire S-Pol data set is being undertaken to identify and study the first echo cases. The parameter of main interest besides Z_h is Z_dr, which should indicate whether the earliest precipitation echo comes from sizable liquid drops or not. During STEPS several of the early echoes do include a positive Z_dr column extending above the freezing level. This column is temporary, its upper portion disappearing after a few minutes. Presumably, it is composed of mm-sized liquid drops, the Z_dr signal disappearing when they freeze. In addition, during the course of this case survey, it was noted that on 23 June interesting observations were made of convective air motions being generated by descending anvil precipitation. This mechanism of generating instability by moistening air via falling precipitation could play a role in the maintenance of stratiform precipitation regions in certain MCSs.

3) TLE OBSERVATIONS

During STEPS more than 1200 transient luminous events (TLEs; mostly sprites) were documented (Lyons et al. 2000, 2003a). Within High Plains convection, sprites typically accompany only a small percentage of +CG flashes, most often within the stratiform precipitation region of larger MCSs (Lyons 1996). STEPS provided the ideal experiment to distinguish sprite parent +CGs (SP+CGs) from other lightning flashes. Sprites appear to represent conventional dielectric breakdown in the mesosphere (~70-75 km) triggered by unusually large electric field transients from +CGs below. Huang et al. (1999) noted the key metric in sprite formation should be the magnitude of the CG lightning charge moment (M_q), the product of charge lowered to ground by a CG flash (C) and the height from which this occurs (Z_q). During STEPS, multiple remote locations coordinated measurements of extremely low frequency (ELF) transient signatures providing estimates of M_q from SP+CGs. A new Israeli technique geolocated STEPS sprites within 200 km at a range of 11 Mm (Price et al. 2002). For a large sample of STEPS sprites, Hu et al. 2002 found that SP+CG M_q values reached 600 C km before there was a 10% probability of sprite occurrence. The probability approached 90% for M_q >1000 C km. These M_q values are far larger than those believed typical of most CG flashes. LMA mapping of the entire SP+CG discharge for the 19 July MCS revealed the charge layers tapped by the CG flashes (Z_q). Interestingly, sprites were not produced by +CGs until the upper-level (~10 km) maxima of LMA VHF emissions was replaced by sources closer to the melting layer. The average Z_q values for the 19 July sprite CGs (4.1 km AGL) and their M_q (800-1000 C km) imply mean charge transfers of ~200 C (Lyons et al. 2003b). This supports the conceptual models of Williams (1998) and Huang et al. (1999) suggesting the charge reservoir for SP+CGs would be found within the lower portions of the MCS stratiform region. Large M_q values appear to be necessary, though perhaps not sufficient, condition for sprite generation.

In contrast to MCSs, supercells rarely produce sprites, except during their dissipating stage, as stratiform debris develops. The 25 June supercell produced two sprites during its decaying phase when the storm's very last two +CG M_q values reached to 1800 C km (Fig. 15). As in the 19 July MCS, the sprites occurred as the altitude of the centroid of maximum VHF emissions was rapidly decreasing. During its supercell phase, the 29-30 June storm produced many high peak current +CG flashes, which while exhibiting large M_q values, did not exceed the 600 C km 10% sprite probability threshold. After sunset, the convection evolved upscale into an MCS in which large M_q +CGs produced 24 sprites above the stratiform region (0345-0546 UTC).

4. STEPS Outreach and Education

STEPS investigators were well aware of the importance of popularizing the project, and outreach to the general public was a key component of STEPS. A media day was scheduled for the project and yielded great exposure. Approximately 14 reports on STEPS occurred in the national and international media, including two spots each on NBC and ABC evening news programs as well as stories in USA Today and the New York Times. There also were STEPS reports on British and German television. Locally there were approximately 13 reports, including news broadcasts on network-affiliate television stations in Colorado and Kansas and stories in major regional newspapers such as the Rocky Mountain News.

Support from the NSF Informal Science Education program allowed production of a planetarium program, "The Hundred Year Hunt for the Red Sprite", featuring the role of STEPS research in determining the atypical nature of the sprite parent lightning discharges. The program will premiere at a major planetarium during spring 2003 (see http://www.Sky-Fire.TV for details).

Local community outreach efforts were organized by Nolan Doesken, research associate at the Colorado Climate Center, in concert with an extension of the Colorado Cooperative Rain and Hail Study (CoCoRaHS; http://ccc.atmos.colostate.edu/~hail/), which uses local volunteer observers to report rain and hail measurements. Approximately 120 volunteers in 3 eastern Colorado counties participated with rain and hail reports during STEPS. These efforts were organized through cooperation with local schools. In particular, Burlington High School students manufactured equipment for deploying hail pads, which were used by the volunteers to measure the number, size, shape, and density of hailstones. STEPS investigators also visited local schools and gave presentations on the project to interested members of the community.

STEPS provided research exposure to many undergraduate and graduate students. An NSF-funded Research Experience for Undergraduates (REU) program was conducted in conjunction with STEPS by Colorado State University professors Chandrasekar and Bringi. The program brought groups of undergraduate students to the STEPS field site. The students worked on radar observations as well as operated an instrumented chase vehicle to collect in situ data. The students also participated in the disassembly of the CSU-CHILL radar after the field experiment.

The Significant Opportunities in Atmospheric Research (SOARS; http://www.ucar.edu/soars) program co-sponsored by UCAR, NSF, DOE, NASA, and NOAA provided the opportunity for two students to experience weather nowcasting and EFM ballooning during the field campaign and to make short research presentations afterwards. Two SOARS proteges (including one not involved in the STEPS campaign) performed STEPS-related research during the 2001-2002 SOARS summer programs, and both are using this research in their university graduate programs.

5. Concluding Remarks

Overall, the STEPS project was a great success, providing the research community with comprehensive observations of the evolving kinematic, microphysical, and electrical structures of a diverse array of thunderstorms, including the primary targets of the experiment: LP supercells and +CG storms. The project also provided a wealth of both polarimetric and in situ microphysical data to develop and improve polarimetric radar-based hydrometeor classification and quantification schemes.

The combination of polarimetric and multiple-Doppler radar observations, along with LMA-based lightning mapping and in situ observations of electric field structure, is providing new insights into the nature of predominantly +CG thunderstorms. Results from 29 June 2000 suggest that development of a mid-level layer of positive charge, in the place of the usual negative charge layer (a so-called "inverted dipole"), plays a major role in production of +CGs as well as inverted polarity intracloud (IC) lightning flashes.

This conclusion is supported by EFM and LMA observations of several other STEPS storms (Table 2). Indeed, one of the more intriguing observations from STEPS was the large number of thunderstorms, even ostensibly weak ones, with apparently inverted-polarity electrical structures. Evidence of this phenomenon has been offered by Rust and MacGorman (2002). This unexpected result sheds new light on the climatological tendency for +CG thunderstorms in the High Plains (Zajac and Rutledge 2001, Carey and Rutledge 2003, Carey et al. 2003), and needs to be addressed by future +CG research.

A key remaining question is how such inverted electrical structures come to exist, in both weak and intense thunderstorms. For example, Wiens et al. (2003) show that hail production aloft and +CG flashes are well correlated for 29 June. However, the exact role hail may play in anomalous positive charging, if any, remains unclear and combined kinematic, microphysical, and electrical analysis of more STEPS cases is required. In addition, numerical modeling of key STEPS cases, using electrification schemes coupled to improved microphysical parameterizations, likely will be needed to resolve this issue.

The cooperation between the NWS and atmospheric research communities, as well as outreach to the general public, were major goals of STEPS. These two actvities are increasingly identified as major factors in a field project's overall success (e.g., Schultz et al. 2002), and the efforts to maximize outreach and inter-community cooperation during STEPS should help provide a model for future field projects.

The results of STEPS research are providing new insights into the physical relationships between thunderstorm kinematics, microphysics, and electrification. As the studies reviewed in this article are broadened to include more cases and more intercomparisons of different observing platforms, vastly improved understanding of these topics should occur. Because of STEPS, the current mysteries of supercell thunderstorms are slowly being unraveled.

Cooperation between Research and Forecast Communities

The Goodland, KS, National Weather Service (NWS) office was in a unique position to provide the STEPS experiment with logistical assistance, forecast personnel, local expertise, and volunteer field-team participants. Indeed, demonstration of cooperation between the forecast and research communities was a goal of the experiment. Pre-operational phase support included assistance in facility procurement, sensor placement, climatological research, lodging assistance, and building local community support for the project. In addition, much of the local media support during STEPS, including arranging of the STEPS media day, was provided by NWS personnel.

During the operational phase, the NWS office was the hub for planning and forecasting. Morning briefings were conducted at the office through the use of both NWS computer resources and web-based NCAR model output. Forecast briefings were conducted as a collaboration between NWS short-term forecasters and project investigators. This allowed local expertise to be integrated into the operational decision process. A briefing summary was disseminated daily over local NOAA Weather Radio stations.

Twenty-five volunteers from seven NWS offices participated in various support positions. The roles of these volunteers ranged from project nowcasters to field participants in the Mobile Mesonet and EFM ballooning operations. A two-way radio enhanced communications between the NWS office and the STEPS Operations Center (OC). Fixed mesonet data, output from NWS analysis software (such as LAPS and SCAN), and severe weather reports were relayed to the OC during field operations.

The NWS benefited through exposure to unique datasets in near real-time. Forecaster access to M-GLASS soundings, timely reports from the Mobile Mesonet, and web-based CSU-CHILL and NCAR SPOL data all contributed to an improved warning program. Interaction with STEPS researchers, including seminars presented by project investigators, allowed NWS staff to increase their knowledge of convective processes and severe convection forecasting. The procedures and lessons learned during STEPS will be used as a model for NWS participation in the Bow Echo and MCV Experiment (BAMEX) field project, scheduled for the summer of 2003.

The STEPS project was made possible through funding by the National Science Foundation. In addition, the extensive collaboration with the National Weather Service, in particular the Goodland branch, was a major key to the project's success. The STEPS community also appreciates the great cooperation of the governments and general public of the cities of Burlington, CO, and Goodland, KS. Data collection, as well as installation and teardown of major facilities like CSU-CHILL and S-Pol, was primarily the result of the dedicated staffs of all of the instrument platforms, as well as the tremendous number of enthusiastic undergraduate and graduate students who participated in the field campaign.

Bluestein, H. B., and C. R. Parks, 1983: A synoptic and photographic climatology of low-precipitation severe thunderstorms in the southern plains. Mon. Wea. Rev., 111, 2034-2046.

Bluestein, H. B., and G. R. Woodall, 1990: Doppler-radar analysis of a low-precipitation severe storm. Mon. Wea. Rev., 118, 1640-1664.

Branick, M. L., and C. A. Doswell, III, 1992: An observation of the relationship between supercell structure and lightning ground strike polarity. Wea. and Forecasting, 7, 143-149.

Bringi, V. N., and V. Chandrasekar, 2001: Polarimetric Doppler Weather Radar: Principles and Applications, Cambridge University Press, 636 pp.

Bringi, V. N., G.-J. Huang, V. Chandrasekar, and E. Gorgucci, 2002a: A methodology for estimating the parameters of a gamma raindrop size distribution model from polarimetric radar data: Application to a squall-line event from the TRMM/Brazil campaign. J. Atmos. Oceanic Technol., 19, 633-645.

Bringi, V. N., T. Tang, and V. Chandrasekar, 2002b: Evaluation of a new polarimetrically tuned Z-R relation. Proceedings, 2nd European Radar Conf., Delft, Netherlands.

Bringi, V. N., V. Chandrasekar, J. Hubbert, E. Gorgucci, W. L. Randeu, and M. Schoenhuber, 2003: Raindrop size distribution in different climatic regimes from disdrometer and dual-polarized radar analysis. J. Atmos. Sci., 60, 354-365.

Browning, K. A., and R. J. Donaldson, 1963: Airflow and structure of a tornadic storm. J. Atmos. Sci., 20, 533-545.

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

Browning, K. A., 1965. Some inferences about the updraft within a severe local storm. J. Atmos. Sci., 22, 669-677.

Browning, K. A., 1977: The structure and mechanism of hailstorms. Hail: A Review of Hail Science and Hail Suppression. Meteor. Monogr., 38, G. B. Foote and C. A. Knight, Eds., Amer. Meteor. Soc., Boston, Mass., 1-43.

Burgess, D. W., and R. P. Davies-Jones, 1979: Unusual tornadic storms in eastern Oklahoma on 5 December 1975. Mon. Wea. Rev., 107, 451-457.

Carey, L. D., and S. A. Rutledge, 1998: Electrical and multiparameter radar observations of a severe hailstorm. J. Geophys. Res., 103, 13979-14000.

Carey, L. D., and S. A. Rutledge, 2003: Characteristics of cloud-to-ground lightning in severe and non-severe storms over the central United States from 1989-1998. J. Geophys. Res., in press.

Carey, L. D., S. A. Rutledge, and W. A. Petersen, 2003: The relationship between severe storm reports and cloud-to-ground lightning polarity in the contiguous United States from 1989-1998. Mon. Wea. Rev., in press.

Changnon, S. A., Jr., 1977: The climatology of hail in North America. In Hail: A Review of Hail Science and Hail Suppression. Meteor. Monogr., 38, G.B. Foote and C.A. Knight, Eds., Amer. Meteor. Soc., Boston, Mass., 107-128.

Coleman, L. M., T. C. Marshall, M. Stolzenburg, T. Hamlin, P. Krehbiel, W. Rison, and R. J. Thomas, 2003. Effects of charge and electrostatic potential on lightning propagation. J. Geophysical Res., in press.

Conway, J. W., and D. S. Zrnic, 1993: A study of embryo production and hail growth using dual-Doppler and multiparameter radars. Mon. Wea. Rev., 121, 2511-2528.

Cummins, K. L., M. J. Murphy, E. A. Bardo, W. L. Hiscox, R. B. Pyle, and A. E.pifer, 1998: A combined TOA/MDF technology upgrade of the U. S. National Lightning Detection Network. J. Geophys. Res., 103, 9035-9044.

Curran, E. B., and W. D. Rust, 1992: Positive ground flashes produced by low-precipitation thunderstorms in Oklahoma on 26 April 1984. Mon. Wea. Rev., 120, 544-553.

Davies-Jones, R. P., D. W. Burgess, and L. R. Lemon, 1976: An atypical tornado producing cumulonimbus. Weather, 31, 336-347.

Donaldson, R., A. Spatola, and K. Browning, 1965: Visual observations of severe weather phenomena. A family outbreak of severe local storms - A comprehensive study of the storms in Oklahoma on 26 May 1963. Part I., K.A. Browning, Ed., Air Force Cambridge Research Lab., Spec. Rep. No. 32, 73-97. [AFGL, Hanscomb AFB, MA, 01731].

Doswell, C. A., III, and D. W. Burgess, 1993: Tornadoes and tornadic storms: A review of conceptual models. In The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophysical Monograph No. 79, American Geophysical Union, Washington, D.C., 161-172.

Droegemeier, K. K., J. D. Smith, S. Businger, C. Doswell, III, J. Doyle, C. Duffy, E. Foufoula-Georgiou, T. Graziano, L. D. James, V. Krajewski, M. LeMone, D. Lettenmaier, C. Mass, R. Pielke, Sr., P. Ray, S. Rutledge, J. Schaake, and E. Zipser, 2000: Hydrological aspects of weather prediction and flood warnings: Report of the Ninth Prospectus Development Team of the U.S. Weather Research Program. Bull. Amer. Meteor. Soc., 81, 2665-2680.

Fritsch, J. M., R. A. Houze, Jr., R. Adler, H. Bluestein, L. Bosart, J. Brown, F. Carr, C. Davis, R. H. Johnson, N. Junker, Y.-H. Kuo, S. Rutledge, J. Smith, Z. Toth, J. W. Wilson, E. Zipser, and D. Zrnic, 1998: Quantitative precipitation forecasting: Report of the Eighth Prospectus Development Team, U. S. Weather Research Program. Bull. Amer. Meteor. Soc., 79, 285-299.

Gilmore, M. S., J. M. Straka, and E. N. Rasmussen, 2003: Precipitation and evolution sensitivity in simulated deep convective storms: Comparisons between liquid-only and simple ice & liquid phase microphysics. Mon. Wea. Rev., In Press.

Hu, W., S. Cummer, W. A. Lyons, and T. E. Nelson, 2002: Lightning charge moment changes for the initiation of sprites. Geophys. Res. Lett., 29, DOI 10.1029/2001GL014593.

Huang, E., E. R. Williams, R. Boldi, S. Heckman, W. A. Lyons, M. Taylor, T. E. Nelson, and C. Wong, 1999: Criteria for sprites and elves based on Schumann resonance observations. J. Geophys. Res., 104, 16943-16964.

Hubbert, J., V. N. Bringi, L. D. Carey, and S. Bolen, 1998: CSU-CHILL polarimetric radar measurements from a severe hail storm in eastern Colorado. J. Appl. Meteor., 37, 749-775

Kennedy, P. C., S. A. Rutledge, W. A. Petersen, and V. N. Bringi, 2001: Polarimetric radar observations of hail formation. J. Appl. Meteor., 40, 1347-1366.

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

Krehbiel, P. R., R. J. Thomas, W. Rison, T. Hamlin, J. Harlin, and M. Davis, 2000. GPS-based mapping system reveals lightning inside storms, EOS, 81, 21-25.

Lin, Y.-L., R. D. Farley, and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 1065-1092.

Liu, H., and V. Chandrasekar, 2000: Classification of hydrometeors based on polarimetric radar measurements: Development of fuzzy logic and neuro-fuzzy systems, and in situ verification. J. Atmos. Oceanic Technol., 17, 140-164.

Loney, M. L., D. S. Zrnic, J. M. Straka, and A. V. Ryzhkov, 2002: Enhanced polarimetric radar signatures above the melting level in a supercell storm. J. Appl. Meteor., 41, 1179-1194.

Lyons, W. A., 1996: Sprite observations above the U.S. High Plains in relation to their parent thunderstorm systems. J. Geophys. Res., 101, 29,641-29,652.

Lyons, W. A., R. A. Armstrong, E. R. Williams and E. A. Bering, 2000: The hundred year hunt for the sprite. EOS, 81 (33), 373-377.

Lyons, W. A., T. E. Nelson, R. A. Armstrong, V. P. Pasko, and M. A. Stanley, 2003a: Upward electrical discharges from thunderstorm tops. Bull. Amer. Meteor. Soc., in press.

Lyons, W. A., T. E. Nelson, E. R. Williams, S. A. Cummer, and M. A. Stanley, 2003b: Characteristics of sprite-producing positive cloud-to-ground lightning during the 19 July 2000 STEPS mesoscale convective systems. Mon. Wea. Rev., in press.

MacGorman, D. R., D. W. Burgess, V. Mazur, W. D. Rust, W. L. Taylor, and B. C. Johnson, 1989: Lightning rates relative to tornadic storm evolution on 22 May 1981. J. Atmos. Sci., 46, 221-250.

MacGorman, D. R., and K. E. Nielsen, 1991:Cloud-to-Ground lightning in a tornadic storm on 8 May 1996. Mon. Wea. Rev., 119, 1557-1574.

MacGorman, D. R., and W. D. Rust, 1998: The Electrical Nature of Storms, Oxford University Press, New York.

MacGorman, D. R., and W. D. Rust, 2003: Electrical structure of the updraft and rainy downdraft of supercell storms. Geophys. Res. Lett., submitted.

Marshall, T.C., W.D. Rust, and M. Stolzenburg, 1995: Electrical structure and updraft sppeds in thunderstorms over the southern Great Plains. J. Geophys. Res., 100, 1001-1015.

Marwitz, J. D., 1972a: Precipitation efficiency of thunderstorms on the High Plains. J. Res. Atmos., 6, 367-370.

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

Miller, L. J., J. D. Tuttle, and C. K. Knight, 1988: Airflow and hail growth in a severe northern High Plains supercell. J. Atmos. Sci., 45, 736-762.

Miller, L. J., J. D. Tuttle, and G. B. Foote, 1990: Precipitation production in a large Montana hailstorm: Airflow and particle growth trajectories. J. Atmos. Sci., 47, 1619-1646.

Miller, L. J. and M. L. Weisman, 2002: Comparison of radar-observed and WRF-modeled structures of two STEPS storms. Preprints, 21st Conf. on Severe Local Storms, San Antonio, Texas, Amer. Meteor. Soc., Boston, 299-302.

Nelson, S. P., 1983: The influence of storm flow structure on hail growth. J. Atmos. Sci., 40, 1965-1983.

Nelson, S. P., 1987: The hybrid multicellular-supercellular storm - an efficient hail producer. Part II: General characteristics and implications for hail growth. J. Atmos. Sci., 44, 2060-2073.

Orville, R. E., 1994: Cloud-to-ground lightning flash characteristics in the contiguous United States: 1989-1991. J. Geophys. Res., 99, 10833-10841.

Orville, R. E., and A. C. Silver, 1997: Lightning ground flash density in the contiguous United States: 1992-1995. Mon. Wea. Rev., 125, 631-638.

Price, C., M. Asfur, W. Lyons, and T. Nelson, 2002: An Improved ELF/VLF method for globally geolocating sprite-producing lightning. Geophys. Res. Lett., 29, doi: 10.1029/2001GL013519.

Rasmussen, E. N., and J. M. Straka, 1998: Variations in supercell morphology. Part I: Observations of the role of upper-level storm-relative flow. Mon. Wea. Rev., 126, 2406-2421.

Rust, W. D., and D. R. MacGorman, 2002: Possibly inverted-polarity electrical structures in thunderstorms during STEPS. Geophys. Res. Lett., 29, (12), 1571, doi: 10.1029/2001GL014303.

Schultz, D. M., and co-authors, 2002: Understanding Utah winter storms: The Intermountain Precipitation Experiment. Bull. Amer. Meteor. Soc., 83, 189-210.

Stolzenburg. M., W. D. Rust, and T. C. Marshall, 1998: Electrical structure in thunderstorm convective regions. Part III: Synthesis. J. Geophys. Res., 103, 14097-14108.

Straka, J. M., D. S. Zrnic, and A. V. Ryzhkov, 2000: Bulk hydrometeor classification and quantification using polarimetric radar data: Synthesis of relations. J. Appl. Meteor., 39, 1341-1372.

Takahashi, T., 1978: Riming electrification as a charge generation mechanism in thunderstorms. J. Atmos. Sci., 35, 1536-1548.

Tessendorf, S. A., K. C. Wiens, S. A. Rutledge, and L. J. Miller, 2003: The 29 June 2000 supercell observed during STEPS. Part I: Kinematics and microphysics. Submitted to J. Atmos. Sci.

Vivekanandan, J., S. M. Ellis, R. Oye, D. S. Zrnic, A. V. Ryzhkov, and J. Straka, 1999: Cloud microphysics retrieval using S-band dual-polarization radar measurements. Bull. Amer. Meteor. Soc., 80, 381-388.

Warner, T. A., J. H. Helsdon, Jr., and A. G. Detwiler, 2003: Aircraft observations of a lightning channel in STEPS. Geophys. Res. Lett., in review.

Weisman, M. L., and H. B. Bluestein, 1985: Dynamics of numerically simulated LP storms. Preprints, 15th Conf. on Severe Local Storms, Indianapolis, Indiana, Amer. Meteor. Soc., pp. 167-171.

Weisman, M. L., and J. B. Klemp, 1986: Characteristics of Isolated Convective Storms. In Mesoscale Meterorology and Forecasting, Ray, P. (ed), Amer. Meteor. Soc., Boston.

Weisman, M. L, J. B. Klemp, and R. Rotunno, 1988: The structure and evolution of numerically simulated squall lines. J. Atmos. Sci., 45, 1990-2013.

Weisman, M. L. and L. J. Miller, 2000: An overview of the severe thunderstorm electrification and precipitation study (STEPS). Preprints, 20th Conf. on Severe Local Storms, Orlando, Fl., Amer. Meteor. Soc., Boston, 273-276.

Wiens, K. C., S. A. Tessendorf, and S. A. Rutledge, 2003: The 29 June 2000 supercell observed during STEPS. Part II: Lightning observations. Submitted to J. Atmos. Sci.

Williams, E. R., 1989: The tripole structure of thunderstorms. J. Geophys. Res., 94, 13151-13167.

Williams, E. R., R. Zhang, and J. Rydock, 1991: Mixed-phase microphysics and cloud electrification. J. Atmos. Sci., 48, 2195-2203.

Williams, E. R., 1998: The positive charge reservoir for sprite-producing lightning. J. Atmos. Sol. Terr. Phys., 60, 689-692.

Williams, E. R., 2001: The electrification of severe storms. In Severe Convective Storms, C. A. Doswell, III, Ed., Amer. Meteor. Soc., Boston, Meteorological Monograph Series, Vol. 27, No. 49, 570 pages.

Winn, W. P., and L. G. Byerley, III, 1975: Electric field growth in thunderclouds. Q. J. Roy. Meteor. Soc., 101, 979-994.

Zajac, B. A., and S. A. Rutledge, 2001: Cloud-to-ground lightning activity in the contiguous United States from 1995-1999. Mon. Wea. Rev., 129, 999-1019.

Table 1. List of STEPS instrumentation. Hit browser back button to return to text.

Platform	Description
Radar Network	The triple-Doppler network (Fig. 1) consisted of the CSU-CHILL polarimetric Doppler radar near Burlington, CO; the NCAR S-Pol polarimetric Doppler radar near Idalia, CO; and the NWS KGLD NEXRAD radar in Goodland, KS. All radars are S-Band (~10 cm wavelength). The CSU-CHILL and S-Pol radars are both dual-linearly polarized, providing information on the size, shape, orientation, and thermodynamic phase of hydrometeors. Post-project synthesis of the multiple-Doppler observations allowed reconstruction of the three-dimensional wind field within thunderstorms.
Lightning Mapping Array	The LMA was operated in the STEPS project area from mid-May to mid-August, 2000. Thirteen measurement stations were deployed throughout northwestern Kansas and eastern Colorado. The system locates the sources of impulsive VHF radio signals from the lightning by measuring the time that the signals arrive at the different receiving stations. The measurements enable accurate 3-dimensional pictures of the lightning inside storms. The LMA is good at identifying negative breakdown occuring within regions of net positive charge. Thus, the LMA can be used to identify major positive charge regions (and to a lesser extent, negative charge regions) tapped by lightning.
National Lightning Detection Network	The NLDN consists of a combined network of magnetic direction finder and IMPACT sensors used to locate in space and time ground strike locations from CG lightning. Information on CG polarity, peak current, and multiplicity also are available. The most recent NLDN upgrade, discussed in Cummins et al. (1998), gives greater than 80% detection efficiency and 0.5 km location accuracy within the STEPS domain.
T-28 Armored Research Aircraft	The SDSMT armored T-28 provided in situ observations of hydrometeors, winds, and electric fields in the lower to middle altitude range within updrafts and hail shafts. The T-28 is equipped to measure the complete spectrum of water and ice particles in clouds, ranging from cloud droplets a few micrometers in diameter to about 5-cm diameter hail. One of its three precipitation particle imaging probes (the HVPS) has the capability to determine particle charge as the particle is imaged. In addition, it is equipped with a 6-instrument electric field meter system that is used to map the total vector electric field inside and outside clouds.
Balloon-Borne Electric Field Meter	STEPS acquired usable electric field profiles from 23 balloon flights launched from a mobile facility. An overview of the balloon-borne electric field meter (originally developed by Winn and Byerly 1975) that NSSL has improved and used for 15 years is given by MacGorman and Rust (1998). The electric field data were processed to yield both the vertical and horizontal components of the vector. The mobile laboratory also was equipped with standard surface meteorological sensors and an electric field meter.
Mobile Mesonet	NSSL Mobile Mesonet vehicles were available for STEPS. These vehicles, which were designed to augment existing meteorological networks, consisted of meteorological instruments mounted on standard automobiles. They can provide accurate observations of pressure, temperature, relative humidity, as well as wind direction and speed whether the vehicle is moving or stationary.
M-GLASS Soundings	The NCAR Mobile GPS/Loran Atmospheric Sounding System facilities were contained within camper shells mounted on pickup trucks. The basic system consisted of hardware required to make soundings, including equipment to make supporting surface meteorological observations.
Yucca Ridge Field Station	The YRFS, located near Fort Collins, CO (275 km NW of the LMA centroid), and operated by FMA Research, made use of numerous on- and off-site instruments to study transient luminous events (TLEs) during STEPS. These included RF, telescopic, and photometric sensors along with low-light imagers. Data from ELF sensors scattered throughout the world provided additional information on TLE phenomena in the STEPS domain.

Table 2. Overview of STEPS cases. N, S, E, W, etc., refer to points of the compass. Date is the starting date; data collection sometimes continued past midnight. +CG: positive cloud-to-ground lightning. -CG: negative cloud-to-ground lightning. IC: intracloud lightning. LP: low precipitation. Hit browser back button to return to text.

Date	Storm Summary
25 May	Series of storms oriented SW-NE moved west to east across STEPS domain.
26 May	Scattered storms merged into a SW-NE-oriented squall line moving ESE with 0.5"-diameter hail.
31 May	Isolated storm approached from SW with 35 m s-1 surface wind. Then an LP storm to the north with a polarity switch from -CG to +CG was sampled.
3 June	Small LP storm provided good T-28 microphysics data. There were no CGs, but the electrical structure and IC lightning both were observed to be inverted-polarity.
6 June	Weak to moderate storms were observed in the western part of domain, while a later storm occurred north of the earlier convection and was nearly 100% +CG.
9 June	Small broken line of storms occured in the southeastern part of domain, with one of the cells being predominantly +CG.
11 June	Asymmetric mesoscale convective system (MCS) moved west to east across domain and featured significant +CG lightning.
12 June	A short-lived storm with possibly inverted charge structure occurred to the south.
19 June	Chaotic multicell storm moved west to east across domain with 30 m s-1 surface winds and mostly -CG lightning.
22 June	A severe line of storms featured a bow echo, an F0 tornado, 1" hail, 30 m s-1 winds, and mostly +CG lightning along with inverted-polarity IC flashes.
23 June	Disorganized convective cluster was observed to produce few, but mostly positive, CGs.
24 June	Observations were made on a small splitting cell, and then later a marginal supercell with 0.75" hail, +CG lightning, as well as inverted-polarity IC flashes and electrical structure. Last two +CGs of storm each produced sprites as storm dissipated around local midnight.
29 June	Supercell storm passed through dual-Doppler coverage and featured an F1 tornado, larger than golfball-size hail, predominantly +CG lightning, and inverted-polarity IC lightning. Mobile Mesonet and T-28 ended operations after this case.
1 July	Numerous narrow cells occurred along a SW-NE line and had very little lightning activity.
5 July	LP supercell storm was observed to the north, and featured golfball-size hail, mostly +CG lightning, and a possibly inverted-polarity electrical structure. EFM ballooning ended after this case.
8 July	Young cloud precipitation data were taken on fair-weather cumulus.
10 July	Several mostly short-lived severe storms occurred with golfball-size hail and predominantly +CG lightning.
12 July	Weak storm occurred in the first trip of radar, with a stronger +CG storm as a second trip that never moved within range.
17 July	Eastward moving linear MCS was observed SW of the CSU-CHILL radar
18 July	N-S-oriented linear MCS with frequent lightning, including +CGs, moved east across domain.
19 July	SW-NE-oriented convective line started in northern part of domain and moved ESE.
20 July	Scattered storms evolved into SW-NE-oriented squall line that moved SSE across domain and produced heavy rainfall.

Figure 1. Nominal areas of coverage (gray shading) by the triple-Doppler radar network. Outer dual-Doppler lobes (beam angles greater than 30°) and the inner triple-Doppler triangle are outlined in red. The second dual-Doppler lobe for the research radars (CHIL and SPOL) is outlined in blue. The region within which vertical resolution is better than 1 km for the LMA is outlined in green. Topograhic height contours (black lines) are at 3, 4, 5, and 6 kft. NWS radars are shown for Denver, CO (KFTG), Pueblo, CO (KPUX), and Goodland, KS (KGLD), along with the Yucca Ridge Field Station (YRFS). Landmarks are shown at Denver (den), Colorado Springs (csp), Limon (lim), and Akron CO (ako), and at McCook NE (mck). All distances are east-west (X) and north-south (Y) from the Goodland WSR-88D radar. Hit brower back button to return to text.

Figure 2. Time-height cross-section of 29 June 2000 hail echo volume as determined from both S-Pol and CHILL polarimetric data (radar used at each time is the one with the best coverage at that time) via a fuzzy hydrometeor classification scheme. Also shown are the mean starting heights for individual +CG flashes produced by the storm as determined by the LMA (black crosses), as well as notable temperature levels. Hit brower back button to return to text.

Figure 3. Height versus time of LMA-detected lightning sources during the first 3.5 hours of the June 29 storm, showing the occurrence of strong convective surges in the storm (examples of which are labeled A-D), the substantial increase in lightning activity associated with the third convective surge and the onset of the tornado, and the timing of the +CG lightning in the storm. Hit brower back button to return to text.

Figure 4. Examples of two inverted-polarity intracloud (IC) flashes (top) and the initial +CG discharge in the storm (bottom), at the time of the storm's initial convective surge (time A in Fig. 3). The IC discharges developed into inferred positive charge in the hail core of the storm, one from directly above the core and the other from below and off to the side of the core. The +CG was similar to the second IC but also went to ground, lowering positive charge to ground from within the hail core and from the downshear part of the storm. Hit brower back button to return to text.

Figure 5. (a) Horizontal cross-section of radar reflectivity factor (synthesized from both CHILL and S-Pol; color contours) from 2325 UTC on 29 June 2000. Also shown are LMA-detected VHF source locations during 2325-2327 UTC and within 0.5 km of the cross-sectional cut (magenta dots), as well as NLDN-detected +CG ground-strike locations during the radar volume (black crosses). (b) Same as (a) except lacking lightning data and instead showing updraft speeds (black contours; m s^-1) as estimated by multiple-Doppler synthesis. (c) Vertical cross-section at same time showing radar reflectvity, LMA-detected VHF source locations, and updraft speeds. Legend is the same as (a) and (b). Hit brower back button to return to text.

Figure 6. (a) Vertical cross-section of three IC flashes during the initial part of the storm (21:48:23-21:51:05 UTC), between inferred lower positive and mid-level negative charge in the storm. (b) Multilayered charge structure inferred from 35 seconds of lightning activity 11-13 minutes later, following the rapid onset of lightning into the upper part of the developing storm. Hit brower back button to return to text.

Figure 7. Vertical cross section of reflectivity at an azimuth of 76° from the CSU-CHILL radar at 0010 UTC on 30 June 2000, shown with the projection of electric field vectors in this plane for the balloon flight during 0005-0034 UTC. Electric field vectors, shown in blue along the track with a scale at the top, point from the balloon track along the direction a positive charge would move. Plus and minus symbols indicate the height at which positive and negative charge, respectively, were inferred from the electric field profile and the lightning distribution. The question mark indicates a more uncertain inference of charge. The depth of each inferred charge region is indicated by a vertical line beside the charge symbol. The balloon location has been corrected for storm motion to determine its path relative to storm structure at the time of the radar scan. The vertical component of electric field (E_z), temperature (T), dew point (Td), ascent rate (Asc), and relative humidities (RH and RH_ice) are shown for the corresponding up and down soundings. (Co-authors: Dotted lines will be removed in final figure.) Hit brower back button to return to text.

Figure 8. Horizontal cross-section of radar reflectivity factor (from both CHILL and S-Pol; line contours), along with horizontal winds (vectors) and updraft speed (color contours) estimated from multiple-Doppler synthesis by S. Tessendorf for the 2338 UTC radar volume. Also shown is the flight path for the T-28 aircraft from 2340 to 2343 UTC. Hit brower back button to return to text.

Figure 9. Vertical component of the electric field and updraft are plotted versus time from the pass of the T-28 through the core of the storm between 23:39:00 and 23:43:30. See Fig. 8 for a depiction of this path relative to storm structures at the aircraft altitude. Four updraft cores are shaded in red in the lower panel. In the upper panel the electric field magnitudes while the aircraft is in these cores are shaded red when positive and blue when negative. In the first southeasternmost updraft there is hail and positive field, while in the remaining 3 cores, the last 2 of which are precipitation-free, the field tends to be negative. Field magnitudes are always less than 10 kV m^-1. An abrupt field change due to nearby lightning is noted just before 23:41:00. Hit brower back button to return to text.

Figure 10. Horizontal sections of radar reflectivity factor (color scale on the right) in dBZ with overlaid updraft (dark contour lines) and ground-relative horizontal winds (scaled by the vector in the lower-right side of each panel) at (a) 2.5 km and (b) 9.0 km MSL for the 5 July LP supercell at 2332 UTC. Updraft contours are (a) 2, 5 m s^-1 and (b) 15, 25 m s^-1. Hit brower back button to return to text.

Figure 11. Photograph of the mature phase of the 5 July LP supercell. Notable visual characterisitics are: a striated, bell-shaped cloud that is often indicative of a rotating updraft, and very little precipitation to the north and east. Photo courtesy of Morris Weisman. Hit brower back button to return to text.

Figure 12. Radar reflectivity, electric field, and inferred charge for the storm on 5 July 2000. (Left) Vertical cross-section of reflectivity at an azimuth of 45° from the CSU-CHILL radar at 0108 UTC on 6 July, shown with the projection of electric field vectors in this plane for the balloon flight during 0048-0127 UTC. The location of the balloon has been corrected for storm motion to show the storm-relative track at the time of the radar scan. Red bars show the vertical extent of positively charged regions inferred from the electric field profile and the lightning distribution, and blue bars show the vertical extent of negatively charged regions. Other symbols are explained in the caption for Fig. 7. (Right) Storm-relative balloon track (lilac line) superimposed on reflectivity at an elevation of 0.5° from the NCAR-SPOL radar at 0119 UTC. The white dot indicates the horizontal point at which the vertical plane shown in the left panel intersected the corrected balloon track. The origin in each panel is the location of the radar that acquired the data. Hit brower back button to return to text.

Figure 13. Histograms of (a) the median volume diameter (D₀), and (b) log₁₀(N_w) for the weaker storm on 5 July. Note that N_w is in mm^-1 m^-3. Hit brower back button to return to text.

Figure 14. (a) Hydrometeor type classification results for the weaker 5 July storm based on one RHI sweep at 2322 UTC, azimuth angle=33.5°. (b) Hydrometeor type classification results for the 5 July LP supercell storm based on one RHI sweep at 2316 UTC, azimuth angle=57°. Hit brower back button to return to text.

Figure 15. Time line showing NLDN-detected CG flashes (by peak current and polarity) during the evolution of the 25 June supercellular storm. During its final 100 minutes no sprites were detected, except for the final two +CGs of the dissipating storm, both of which had very large charge moment changes. Hit brower back button to return to text.