A Very Active Sprite-Producing Storm Observed Over Argentina

Eos, Vol. 88, No. 10, 6 March 2007

VOLUME 88 NUMBER 10 6 MARCH 2007

EOS, TRANSACTIONS, AMERICAN GEOPHYSICAL UNION PAGES 117–128

cal interest, there are numerous broader impli- A Very Active Sprite-Producing cations for studying sprites and other TLE events. In particular, they may be a significant driver or discharging mechanism for the global Storm Observed Over Argentina electrical circuit, the 200- kilovolt potential between the Earth and the ionosphere that is PAGES 117, 119 very low frequency (ELF/VLF) electric and thought to be maintained by thunderstorms magnetic field data recorded at the observa- and lightning. TLEs may also produce nitrous oxides (NO ), an important component gas for During the night of 22–23 February 2006, tory. We are currently analyzing these data x more than 400 middle- atmospheric optical further and plan to publish in-depth, ozone chemistry, in the middle atmosphere. discharges were observed above one large detailed studies of these TLEs and their par- Furthermore, these high- altitude discharges thunderstorm system over northeastern ent thunderstorms in the near future. This transfer energy from the troposphere to the Argentina. These transient luminous events region of South America is likely one of the middle and upper atmosphere, and might be (TLEs) were imaged during the Southern most active global locations for intense light- a significant driver of electron density Brazil Sprite Campaign, the first campaign to ning and TLE production, and these investi- changes in the lower ionosphere. Thus, focus on TLEs over southern Brazil, north- gations will be a major step toward a global TLEs may adversely affect radio communica- eastern Argentina, and Uruguay. All of the characterization of the electrical coupling tions systems, GPS, and aircraft. Finally, TLEs were imaged from the Brazilian South- between the troposphere and the middle terrestrial gamma-ray flashes (TGFs), of more ern Space Observatory (SSO) near Santa and upper atmosphere. than 1 mega-electronvolt in energy, have Maria, which is nearly in the center of the been detected by satellites orbiting Earth, such southernmost Brazilian state of Rio Grande as the Reuven Ramaty High Energy Solar do Sul. Although the fields of view of the The TLE Family Spectroscopic Imager (RHESSI) spacecraft imaging cameras were too narrow to view [Smith et al., 2005]. TGF production has been the entire storm, the more than 400 con- Sprites, halos, jets, and elves are members correlated with thunderstorms and might be firmed TLEs imaged indicate that this storm of a family of TLEs that illuminate the strato- related to TLEs. Determining the exact ranks as the third most active TLE producer sphere and mesosphere above thunder- source of these energetic photons may well ever reported. Hence, storms in this region storms, usually between 20 and 95 kilometers lead to some fundamental new understand- of South America might be some of the lead- in altitude. In addition to their intrinsic physi- ings of thunderstorms. ing TLE generators on Earth. The Southern Brazil Sprite Campaign is a joint project between the United States (Uni- versity of Washington, Utah State University, and Duke University) and Brazil (National Institute of Space Research; INPE). The pri- mary goal of the campaign is to measure the in situ electric field above sprite- parent lightning using sensors flown on zero- pressure stratospheric balloons. Unfortunately, for the two nights that TLEs were imaged during the campaign window of February–March 2006, the storms were well beyond the range of the predicted balloon flight path. Hence, no in situ measurements were attempted on these nights. In this article, we present a brief overview of the video data of TLEs occurring over Argentina and Brazil on 22–23 February, along with coincident lightning location data from the World Wide Lightning Location Net- work (WWLLN; see http://webflash.ess.washington.edu/) together with extremely low to

BY J. N. THOMAS, M. J. TAYLOR, D. PAUTET, M . B AILEY, N. N. SOLORZANO, R . H . H OLZWORTH, M . P. M CCAR- THY, M. KOKOROWSKI, F. S AO SABBAS, O. P INTO JR., S. Fig. 1. GOES-12 infrared image of a sprite-producing thunderstorm system to the west and south- A. CUMMER, N. J AUGEY, J. L I, AND N. J. SCHUCH west of the Brazilian Southern Space Observatory near Santa Maria (SM). Eos, Vol. 88, No. 10, 6 March 2007

Fig. 2. Halos and sprites over Argentina that occurred within about 0.60 seconds (from left to right) observed from the Brazilian Southern Space Observatory on 23 February 2006. The maximum altitude of these optical emissions is 80–90 kilometers.

Most of the TLEs imaged over Argentina meters to the west and southwest of the strokes. For example, the 194 coulomb kilo-

were sprites, which are optical discharges observatory near Santa Maria (indicated by meter iΔMq +CG lightning had a continuing that last up to tens of milliseconds that occur the letters SM). Note that the –40°C cloud current amplitude of about 20 kiloampere between 40 and 90 kilometers in altitude. shield had a maximum area of about 550,000 kilometers that lasted for about 200 millisec- These high-altitude discharges were first square kilometers, which is larger than most onds. Thus, the total charge moment change imaged by chance by the late Jack Winckler’s sprite- producing thunderstorms over the U.S. was several thousands of coulomb kilometers. group at the University of Minnesota [Franz High Plains that typically have a cloud shield Unfortunately, with a video rate of 30 frames et al., 1990]. Sprites have since been corre- of less than about 450,000 square kilometers per second, the timing uncertainty is too lated almost exclusively with positive cloud- [Lyons et al., 2006]. In fact, large thunder- great for the total charge moment change to- ground lightning (+CG) strokes with large storm systems are common in this region of associated with this long-delayed sprite to be charge moment changes (defined as the South America during the spring and sum- determined precisely. Nonetheless, it is likely total charge removed times the initial charge mer months, which further suggests that this that a large fraction of the charge transfer height). They usually occur during large is an active region of sprite production and a responsible for generating the carrot sprite thunderstorms, known as mesoscale convec- major driver for our experimental program. was provided by the long- continuing currents tive systems (MCSs), which have large cloud However, the details of how these South instead of just the return strokes. areas capable of maintaining thousands of American sprite- producing storms compare During this campaign, on 3–4 March 2006, coulombs of charge. and contrast with sprite- producing storms more than 100 TLEs from a second storm To date, the majority of sprites have been over the U.S. High Plains is not yet known were imaged. The storm was to the north- observed over MCSs that develop during the and is the topic of future analysis. west of Santa Maria, over parts of Brazil, spring and summer months over the High Figure 2 shows halos and sprites, imaged Argentina, and Paraguay, and had a maxi- Plains area of the U.S. Great Plains. However, from a single camera, that occurred within mum –40°C cloud shield of about 420,000 extremely low frequency (3 hertz to 3 kilo- about 0.60 seconds (from left to right) start- square kilometers. These data will be ana- hertz) magnetic field resonances of the ing at 0730:50.490 UT correlated with lyzed in future works. Earth- ionosphere waveguide, which are WWLLN- located lightning events. The left These observations are not the first time driven by large charge moment change light- frame of Figure 2 is a disc-like halo at a that sprites were imaged over this region of ning, suggest that this region of South Amer- range of about 585 kilometers with a diame- South America. The first imaging of sprites ica may be one of the most active areas for ter of about 66 kilometers. This halo is corre- from the space shuttle during the early 1990s sprite production globally [Sato and lated with +CG lightning with an impulse showed that they were occurring over Argen-

Fukunishi, 2003]. charge moment change (iΔMq) of 254 cou- tina [Boeck et al., 1995; Lyons and Williams, lomb kilometers, where iΔMq is the charge 1993], and Blanc et al. [2004] reported The February 2006 Thunderstorms moment change in the first 2 milliseconds of observing five sprites over this region from the stroke calculated from ground- based the International Space Station during Octo- During the late morning hours (local time) ELF/VLF measurements. ber 2002. of 22 February 2006, a large thunderstorm Recorded 66 milliseconds after the left system formed over the Pampas region of frame, the middle frame of Figure 2 shows a Questions to Be Addressed Argentina, east of Buenos Aires. This storm sprite accompanied by a halo at a range of system further developed and strengthened about 578 kilometers with a maximum width Our data analysis is still in progress, but throughout the day while moving northward. of about 82 kilometers correlated with a 368 these preliminary findings suggest that fur- By 0230 UT on 23 February (22 February, coulomb kilometer iΔMq +CG lightning. The ther observational campaigns in Argen- 11:30 P.M. local time), when the first sprites right frame, recorded 501 milliseconds later, tina, southern Brazil, Paraguay, and Uru- were observed from the Brazilian Southern shows dancing sprites along with a larger, guay are most warranted. Some specific Space Observatory, the storm encompassed carrot- shaped sprite at a range of about 611 questions that we hope to address are the an area of about 300,000 square kilometers kilometers correlated with 157 and 194 cou- following: (based on Geostationary Operational Envi- lomb kilometer iΔMq +CG lightning. 1. What were the meteorological condi- ronmental Satellite (GOES)- 12 infrared The width of this collection of luminous tions that gave rise to the two Argentine TLE- images of cloud top temperatures below – structures in the right frame is approximately producing thunderstorms, and how often do 40°C) and was producing significant 90 kilometers. Since the carrot sprite shown they occur? How do these conditions com- amounts of lightning as detected by the in the right frame occurred more than 50 mil- pare with TLE- producing storms observed WWLLN and the Brazilian Lightning Detec- liseconds after the two parent +CG lightning over the U.S. High Plains and elsewhere? tion network. Between 0230 and 0830 UT, strokes, the impulsive charge moment 2. How do the dynamical cloud top tem- 446 TLEs (mostly sprites and some halos or changes do not necessarily indicate the total peratures and cloud areas compare with the elves) were observed. charge moment changes. Indeed, very large TLE production rate and locations during the Figure 1 is a GOES 12 satellite infrared continuing currents, which are low- level cur- two Argentine storms? How do these parame- image of the sprite-producing thunderstorm rents that follow return strokes, were mea- ters compare with TLE-producing storms over at 0630 UT located about 500–800 kilo- sured by the ELF/VLF sensor following these the U.S. High Plains and elsewhere? Eos, Vol. 88, No. 10, 6 March 2007

3. What is the parent- lightning charge gulation of transient Schumann resonance events, moment change needed to drive sprites dur- thank students and staff at the Space Science Geophys. Res. Lett., 30(16), 1859, doi:10.1029/ ing these Argentine storms? How does this Laboratory of the Federal University of Santa 2003GL017291. Maria for their assistance throughout the Smith, D. M., L. I. Lopez, R. P. Lin, and C. P. Barrington- compare with the charge moment changes Leigh (2005), Terrestrial gamma-ray flashes measured over the U.S. High Plains and else- campaign. observed up to 20 MeV, Science, 307, 1085–1088. where? Are there some sprites that can be driven by relatively low charge moment References changes, which are too small to cause break- Author Information Blanc, E., T. Farges, R. Roche, D. Brebion, T. Hua, A. down using only the quasi-electrostatic field Labarthe, and V. Melnikov (2004), Nadir observa- Jeremy N. Thomas, Department of Earth and model? tions of sprites from the International Space Space Sciences, University of Washington, Seattle; Station, J. Geophys. Res., 109, A02306, doi:10.1029/ 4. Does the morphology of TLEs differ with E-mail: [email protected]; Michael J. Taylor, meteorological conditions and/or charge 2003JA009972. Boeck, W. L., O. H. Vaughan Jr., R. J. Blakeslee, B. Von- Dominique Pautet, and Matthew Bailey, Center moment changes? negut, M. Brook, and J. McKune (1995), Observa- for Atmospheric and Space Sciences and Physics Addressing these questions will help lead tions of lightning in the stratosphere, Department, Utah State University, Logan; Natalia N. to a new understanding of the lightning- and J. Geophys. Res., 100(D1), 1465–1475. Solorzano, Physics Department, Digipen Institute of TLE-driven coupling between the upper and Franz, R. C., R. J. Nemzek, and J. R. Winckler (1990), Technology, Redmond, Wash.; Robert H. Holzworth, Television image of a large electrical discharge Michael P. McCarthy, and Michael Kokorowski, lower atmosphere in all regions of the world, above a thunderstorm system, Science, 249, 48–51. Department of Earth and Space Sciences, Univer- not just the well-studied U.S. High Plains. Lyons, W. A., and E. R. Williams (1993), Preliminary investigations of the phenomenology of cloud-to- sity of Washington, Seattle; Fernanda Sao Sabbas stratosphere lightning discharges, paper presented and Osmar Pinto Jr., National Institute of Space Acknowledgments at the Conference on Atmospheric Electricity, Am. Research (INPE), Sao Jose dos Campos, Brazil; Meteorol. Soc., St. Louis, Mo., 4–8 Oct. Steven A. Cummer, Nicolas Jaugey, and Jingbo Li, Lyons, W. A., L. Andersen, T. E. Nelson, and This work was supported by the U.S. Department of Electrical and Computer Engineer- G. R. Huffines (2006), Characteristics of sprite ing, Duke University, Durham, N. C.; and Nelson National Science Foundation under grants producing electrical storms in the steps 2000 Jorge Schuch, Brazilian Southern Space Observa- domain, paper presented at the Second Confer- ATM- 0091825 and ATM- 0355190 to the Uni- tory, Southern Regional Space Research Center versity of Washington, and by the Brazilian ence on Meteorological Applications of Lightning Data, Am. Meteorol. Soc., Atlanta, Ga., 30 Jan. to (CRSPE/INPE), Santa Maria, Brazil. government FAPESP projects 02/01329-1 and 2 Feb. 04/12350-7 to the Brazilian National Institute Sato, M., and H. Fukunishi (2003), Global sprite for Space Research (INPE). The authors occurrence locations and rates derived from trian-

An 11.28- meter piston core (MD02-2579) Sea Level Rise in Tampa Bay recovered by the R/V Marion Dufresne in July 2002 from 9.14 meters of water in a depres- PAGES 117–118 conflicting patterns: those showing progres- sion in the central part of Tampa Bay (Figure sive submergence with a decelerating rate 1) contained three stratigraphic units charac- Understanding relative sea level (RSL) rise during the past 5000 years [Scholl et al., 1969] terized by distinct lithology and microfaunas. during periods of rapid climatic change is and those showing high sea level during the The lowermost unit 1 (11.2–7.2 meter core critical for evaluating modern sea level rise middle of the Holocene [Blum et al., 2001; depth) consists of shelly sands and yielded given the vulnerability of Antarctic ice Balsillie and Donoghue, 2004], where the radiocarbon ages from greater than 43,000 shelves to collapse [Hodgson et al., 2006], Holocene represents a geologic epoch that years ago. Amino acid racemization ages on the retreat of the world’s glaciers [Oer- extends from about 10,000 years ago to pres- mollusks, interglacial pollen assemblages, lemans, 2005], and mass balance trends of ent times. This discrepancy is emblematic of and marine ostracodes suggest it represents the Greenland ice sheet [Rignot and Kanaga- the uncertainty surrounding Holocene sea the last interglacial period, termed marine ratnam, 2006]. The first-order pattern of level and ice volume history in general. oxygen isotope stage 5. Unit 2 (7.2–2.9 global sea level rise following the Last Gla- Tampa Bay is a shallow (~4 meter depth) meters) consists of organic and calcareous cial Maximum (LGM, ~21,000 years ago) is estuary formed by dissolution of the Mio- muds containing nonmarine ostracodes that well established from coral [Fairbanks, cene Arcadia Formation [Hine et al., in were deposited in lakes, ponds, and wetlands 1989], continental shelf [Hanebuth et al., press] and deposition of Quaternary sedi- between the LGM and the end of the Younger 2000], and other records [Pirazzoli, 2000] ments in sinkholes and karst depressions Dryas cooling episode (~11,500 years ago) (D. and has been integrated into a global ICE-5G during glacio-eustatic sea level cycles. The Willard et al., Deglacial climate variability model of glacio-isostatic adjustment (GIA) Tampa Bay Study (http://gulfsci.usgs.gov/ from central Florida, USA, submitted to Pal- [Peltier, 2004]. However, uncertainty intro- tampabay/index.html), a collaborative proj- aeogeography, Palaeoclimatology, Palaeoecol- duced by paleo water depth of sea level indi- ect—between the U.S. Geological Survey ogy, 2006, hereinafter referred to as D. Willard cators, radiocarbon chronology (i.e., reser- (USGS), and the University of South Florida et al., submitted manuscript, 2006). Unit 3 voir corrections for marine shell dates), and Eckerd College, both in St. Petersburg— (2.9–0 meters) consists of marine sandy postglacial isostatic adjustment, and other is investigating the sea level, climatic, and muds, and represents late Holocene sedi- processes affecting vertical position of for- environmental history of Tampa Bay (Figure ments deposited during the past approxi- mer shorelines produces scatter in RSL 1). Here we report on the sedimentary mately 3500 years. curves, limiting our knowledge of sea level record of early Holocene sea level rise and Unit 2 pinches out laterally from the karst rise during periods of rapid glacial decay. its relationship to regional and global sea basin in central Tampa Bay, resulting in a One example of this limitation is the Gulf of level and polar ice volume. 7500-year hiatus between units 2 and 3 due Mexico/Florida region where, despite to erosion or nondeposition of lake sediment decades of study, RSL curves produce two prior to Holocene sea level rise. However, Quaternary Stratigraphy and Chronology cores from the nearby Hillsborough Bay region (VC-75, VC-77, VC-78) recovered a con- BY T. C RONIN, N. T. E DGAR, G. B ROOKS, D. H AST- Tampa Bay stratigraphy features alternating formable sequence of lacustrine and estua- INGS, R. LARSON, A . H INE, S. L OCKER, B. S UTHARD, B. episodes of estuarine- marine and nonmarine rine sediments continuously deposited in a FLOWER, D. H OLLANDER, J. WEHMILLER, sedimentation during periods of relatively separate karst- like basin, which allow us to D. WILLARD, AND S. SMITH high (interglacial) and low (glacial) sea level. pinpoint the age and elevation of early Holo-