JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. A10, PAGES 21,549-21,554, OCTOBER 1, 2001
New evidence for the brightness and ionization of blue starters
and blue jets
E.M. Wescott, • D.D. Sentman,
• H.C.Stenbaek-Nielsen,
1P.Huet,2M.J.Heavner,
3and
DR Moudry
•
Abstract.Bluejetsandbluestarters
arepartiallyionizedluminous
conesof primarilybluelight
thatpropagate
upward
outofthetopof thunderstorms
atspeeds
of order100kms . Bluejets
propagate
up-40 km,butbluestarters,
whichresemble
bluejets,terminate
abruptly
afteronlya
fewkilometers
of upwardtravel.Theories
ontheoriginof bluejetshaveproposed
thattheyare
dueto eitherpositive
ornegative
streamers
orrunaway
electrons.
Quantitative
analysis
of new
-I
multi-instrument
observations
of a bluestarterfrom an aircraftduringtheEnergeticsof Upper
Atmospheric
Excitation
byLightning,
1998(EXL98)campaign
of July1998,shows
thatthe
ionizationaccounts
for-3 % of theobservedintensity.Quantitativeanalysisof a remarkable
colorphotograph
of a bluejet takenfromR6union
IslandintheIndianOceanshows
thatthe
minimumopticalenergydeposition
was-0.5 MJ. Thesamephotograph
showsdetailsof
streamers never before seen.
1. Introduction
observations included a camera with a narrow band filtered at
Fifty-one bluejets and 30 blue starterswere first documented
and named from 1994 aircraft TV observations of a severe storm
near Texarkana,Arkansas[Wescott et al., 1995]. Blue jets
propagateupward out of the top of thunderstormsat a mean
ionizedN2+ (firstnegativegroup(1 NG)), 427.8-nm.The427.8
nm dataclearlydemonstrates
thatthe blue startersare partially
ionizedandby association
thatbluejets are alsoionizedbeams.
The knowledgeof the ionizationallowsa betterestimateof the
energyinvolvedandthe effectson the upperatmosphere
by these
verticalspeedof 112ñ 24 km s'z to a meanterminalaltitudeof phenomena.
37 ñ 5 km, basedon the triangulationof 34 examples[Wescottet
In this paper we presentthree new observationsof blue
al., 1998]. Blue starters,which resembleblue jets, terminate jets/starters.(1) A colorphotographobtainedin March 1997 by
abruptlyafteronly a few kilometersof upwardtravel,meanvalue one of the authors,P. Huet, from R6union Island. This beautiful
20.8 ñ 4.7 km [Wescottet al., 1996]. Priorto 1998,singleblue photograph,which shows the highest-resolutiondetails ever
jets hadbeenrecordedfrom aircraftoverKansas(July 4, 1994) captured,hasbeenquantitativelyanalyzedfor its brightness.(2)
and CentralAmerica (August29, 1995). What appearsto be a
Blue starter imagesobtainedduring the Energeticsof Upper
blue jet was capturedon film by an Australianphotographer, AtmosphericExcitation by Lightning, 1998 (EXL98) aircraft
Peter Jarver,and published[Lyons, 1997], but all the rest have
campaign
duringJuly1998andanalyzed
in detailto determine
beenseenon low light level TVs with only moderateresolution.
their ionizationproperties. (3) A blue jet, capturedover Iowa
Paskoet al. [1996] have proposeda theory which accountsfor also during the EXL98 aircraft campaign,exhibiting a timetheessential
featuresof bluejets andstarters
baseduponpositive varyingrate of ascent.
streamers,which seemsto agree with the fact that jets and
streamersseemto occurin regionsof thunderstorms
producing
negative cloud-to-ground(CG) flashes. The streamerswould 2. Analysisof Color Photograph
havea concentration
of positivecharges
at thefront.
A colorphotographof a bluejet Plate 1, was takenby one of
Sukhorukovet al. [1996] proposednegativestreamersas the
the authors, P. Huet, from St.-Denis, R6union Island, in the
causeof jets but requireda positiveCG (+CG) flash to initiate
IndianOcean(20ø 51.84' S latitude,55ø27.6' E longitude)while
the streamer.Roussel-Duprd
et al. [1998] andSymbalistyet al.
photographing
lightningfrom a largeeveningthunderstorm
over
[1998] havesuggested
upperatmospheric
discharges
dueto air
theoceanto thenorth.The photograph
is a 2-mintime exposure
breakdown
by runawayelectrons
asa possiblecauseof bluejets.
using400ISo Fujicolor
SG400film in a Zenithreflexcamera
SukhorukovandStubbe[1998] andRowland [1998] discussthe
with a 58-mm lens at,fi4 or 5.6. The date is uncertain,but the
theoriesof bluejet formation.
lightning photographywas definitely betweenMarch 3 and 7,
Wescottet al. [1998] havepreviouslypresentedevidencefrom
1997. Unfortunately,the weatherserviceon R6unionIslanddoes
colorTV imagesof bluejets andstartersthat suggested
thatthe
not keeprecordsof stormsoverthe ocean,so we havenot been
bluelightmusthaveanionized
N2+ component.
The 1998aircraft
Geophysical
Institute,
University
ofAlaska,
Fairbanks,
Alaska.
13Clermont
Ferrand,
France
LosAlamos
National
Laboratory,
LosAlamos,
NewMexico.
Copyright
2001bytheAmerican
Geophysical
Union.
able to obtain
further
information
about the thunderstorm
associated
with the event.Analysisof the starbackground
reveals
that if the datewasMarch 3, the time was 1745UT, by matching
the horizon with the rotation of the stars. The time would be 4
min later eachday after March 3, and the azimuthand elevation
would not change appreciably. Assuming March 3, we
determined that the azimuth was -13 ø east of north. The azimuth
Papernumber2000JA000429.
andelevationof the base,whereit appears
abovethe apparent
0148-0227/01/2000JA000429509.00
cloudtop, and the tip were determinedfrom the starfield. If we
21,549
21,550
WESCOTT ET AL.: BRIEF REPORT
assumereasonablevaluesfor the altitudeof baseandthe jet tip,
18, and 40 km, respectively[Wescottet al., 1998], the jet was
locatedat about19.6ø S latitudeand55.7ø E longitude,145 km
distant from the camera.
The image was digitized on a 24-bit color scannerand
separatedinto its constituentred, green,and blue components.
The bluejet is very evidentin all threecolor separations,
but red
and greenonly appearin the very brightestlower one third of the
jet. This is most likely due to saturationof the film and not to
real red or greenemissionsfrom the jet. In theory, there are red
I
:
/
----
I
0.8 J--
:
I
•.s^o,
"syn_lngl
.dat"
! ....
N22P
(X10Scale)
•,
0.6
0.4
I
I
o,,,,o,
II
/
JJ
I
'-j
II
,,
o
,
0
LU
I
425
:::3
I
428
•430
Wavelength (NM)
Figure 2. The band-passtransmissioncurvesfor the 427.8-nm
filter at normalincidence
andat 2.5ø off axis,andthe synthetic
spectra
of theN2+(1NG) ionized
emissions
andtheN2(2 PG)
neutralemissions
(plottedat a xl 0 scalefor comparison).
N2(firstpositive
group
(1 PG)emissions,
bhttheyarestrongly
quenched at these altitudes, and there are no likely green
emissions. We determinedthe brightnessof the blue jet in the
blue separationby comparisonwith the stars, which were not
saturated. Nineteen stars ranging from 3.89 to 7.4 magnitude
were color correctedand usedto provide a brightnesscalibration
for a single pixel, using the method of Johnson and Mitchell
[1975]. Of the blue jet pixels, 6.6 % in the lower one-third
portion of the jet were saturated,with a background-corrected
countof 190. Onepixelsubtends
3.38x 10-7srinthevicinityof
the jet. We calculate that a saturatedpixel has a minimum
brightnessof 6.76 MR, andnon saturatedpixelsare proportionate
to their backgroundcorrectedcountdividedby 190. A rayleighis
defined
as106photons
cm-2column
s-•,andcanbeconverted
to
opticalenergyin joules. We calculatedthe minimumvalue of the
optical energy in the blue portion of the film response by
summing all the background-corrected pixel values and
convertingto 0.5 MJ of optical energy.The saturatedpixels
accountfor 25% of the total opticalenergy. The true valueof the
total opticalenergyassociated
with thejet is higherbecauseof
18
i
0
I
I
I
2
'1
4
Soale km
Figure 1. Invertedblack and white image(5.9ø horizontalby
10.2ø vertical)of the bluecomponent
of the R•unionIslandblue
jet film photograph.Note the eightnarrowstreamers
branching
outfromthe mainjet andthe smallstreamercomingup abovethe
cloudtop to the right of the jet. The background,contrast,and
brightnesshave beenadjustedin the imageto showthesefaint
features.The streaksare startracksfrom the 2-min exposure.
the saturatedpixels and the band emissions(e.g., 391.4 nm)
belowthe film cutoffat 400 nm. •escott et al. [1996] estimated
thata singlebluestarterdepositedI MJ, basedon the energyof
an averagelightningflash.
Figure I showsthe blue image in reverseblack and white
which hasbeencontrastadjustedto showthe faint features.The
most interestingfeature consistsof eight faint filamentary
features (streamers)branchingoff the mainjet and one small
streamercoming up above the cloud top to the right. The
apparentwidthsof the small streamersrangefrom 50 to 100 m.
At the baseof the blue jet the apparentdiameteris •-400 m. The
diameterdoesnot vary appreciably
up to •-22-kmaltitude.At 30
km it broadens to •-2 km and is •-3 km at 35 km.
It is more
diffuse at the top but is estimatedto be •-4.5 km in diameterat
40-km altitude.Somestructurecan alsobe seenin the mainjet
abovethe saturated
portionabouthalf way up. Thetypicalhorn
shapeis evidentin the top one third, with a suggestionof limb
brighteningas if the conewerehollow.
WESCOTT
.....
....:
....
:-.
•.•:
ET AL.'
BRIEF
REPORT
21,551
.
Plate
1. Color
photograph
ofablue
jetover
alarge
thunderstorm
north
ofR6union
Island
intheIndian
Ocean
taken
by
oneoftheauthors,
P.Huet,inMarch
1997.Thefieldofviewis9.0øhorizontal
by12.4
øvertical
intheconstellation
Ursa
Major.
Thebase
oftheblue
jetwhere
itcan
beseen
above
thetopoftheclouds
is-18-km
altitude,
and
thetipisat-35
km.Thepi• streak
across
thebottom
isdue
toanaircraft
passing
during
the2-min
exposure.
21,552
I
45 ø
WESCOTT ET AL.' BRIEF REPORT
I
I
I
I
I•q•I
I
I
I
I
i
ß
i
i
_
WISCONSIN
Minneapolis/
St. Paul
/
-F
-
_
44.5 ø -
-t-
midway betweenDenver and Boulder, Colorado. It was
equipped
with six low lightlevelimagers:a colorTV system,a
wide-angle
unfiltered
low lightlevelTV (LLLTV), a narrowfield
of view LLLTV with a narrowband(1.25 nm) 427.8-nmfilter, a
near ultravioletcamera(NUV) filtered at 340 nm, and a near-
-
infrared (NIR) camera. All imagersviewed out left-side
windows,which were specificallyselectedto passUV or NIR
emissions.The 427.8-nmfilter was carefullychosento include
-
the N2+(1NG) bandat 427.8-nm,yet excludethe nearby
_
nonionizedN2(2 PG) neutralband.Figure2 showsthe filter
curves
andthesynthetic
spectra
of boththeionized
N2+(1NG)
ß
.
andtheN2 (2 PG) neutralemissions.
'
.
44 ø -
04:30:17
.
-'70 •
MINNESOTA
IOWA
ß 04:25:03
A/C location
km
-
50-
identifiedas a starter.See Figure 4. One more in cell B is
-
probably
a starter,
buttheremaining
five aremoreuncertain.
Figure4 (leftframe)shows
the3.1øvertical
by 3.5øhorizontal
field of view portionof the starterin theunfilteredvisiblelight
cameraframe,whichsaturated
theLLLTV. The saturation
of the
video signalcausesthe apparentsize of the starterto appear
_ largerthanit is in reality. The ionizedN2 lightat 427.8nm is
shownin Figure4 (rightframe)withthesamefieldof view.The
' blue starteris only0.8ø off axis,so onecanseefromthe filter
curvesin Figure2 thatthereis no contribution
fromtheN2 (2
-
43ø-94 ø i i i i-93.5
i øi i i i - 9I3 ø i i i i-92.5i øi i i I -92 ø
Figure 3. Map of National Lightning Detection Network
(NLDN) flashesin the time interval0425:03-0440:19UT, July
19, 1998, in the vicinity of blue starters.Negativeflashesare
filled circles,and positiveflashesare plus signs. Stormcell A
hadfive probablestarters,andcell B hadseven.Only onein cell
B is absolutelyidentifiedas a starter.SeeFigure4. One morein
cell B is probably a starter,but the remainingfive are more
uncertain.
3. MultispectralVideo Observationsof a Partially
Ionized
map,Figure3). StormcellA hadfiveprobable
starters,
andcell
B hadseven.Onlyoneat 0434:26.33UT in cell B is absolutely
'
-
43.5 ø-
Ontheeveningof July19, 1998,fifteenpossible
bluestarters
were recorded over two thunderstormcells in Minnesota (see
Blue Starter
The EXL98 spriteaircraftcampaigninvolveda Gulfstream2
PG) band.
This is conclusiveevidencethat the blue starterwas partially
ionized. No starswere detectedthrough the narrow 427.8-nm
filter in this frame, but at another time, (xVirgo (Spica), a
magnitude0.98, B 1 V spectraltype star, was visible throughthe
filter.
We
used this as a calibration
and calculated
that the
combination of the 427.8, 470.9, and 391.4-nm ionized bands in
the brightestpart of the starter would be 31 kR. The known
ratios of ionized N2 bands at 391.4- and 470.9-nm to 427.8-nm
emissionsare 3.28 and 0.2, respectively[Vallance-Jones,
1974].
These transitionsare all from the same level (v'=0) and are
independent
of the excitationmechanism.Althoughthe starterin
visiblewhite light,Figure3, wasvery saturated,
we were ableto
use the color TV image, which showedfive starsrangingin
magnitude
from2.23 to 4.24 with knownspectra,for calibration.
jet aircraftflying from the JeffersonCountyAirportlocated We calculatethat the maximum blue band starterbrightnesswas
Figure 4. The bluestartereventat 0434:26.33UT. The left frameshowsa 3.1øverticalby 3.5øhorizontalfield of view
portion of the unfilteredvisible light cameraframe, which is saturated.The ionizedN2(1 NG) emissionat 427.8 nm is
shownin the right frame with the samefield of view. This is clearproofthat the blue starterwas partiallyionized.The
unfilteredimageis very overexposed,
which makesit look much largerthan it is in reality. Quantitativeanalysisusing
starcalibrationsandthe colorTV frame showsthatthe ionizedN2 emissionsamountto 3% of the blue light.
WESCOTT ET AL.: BRIEF REPORT
21,553
40'
40'
Figure 5. A mapshowingtheNLDN cloud-to-ground
(CG) lightninglocationsfor a periodof 15 min preceding
the blue
jet, the aircraftposition,andthe probablelocationof thebluejet (a largedot), assumingit is centeredin the stormon the
azimuthdeterminedfrom the stars. The-CG lightningis shownas blackdots,and the + CG lightningis shown as
+pluses.
jets, but perhapsthe sort of blue jet that we recordedin 1998
from the aircraftfits the anecdotalobservations.This bluejet
In the color TV frame there is primarily blue cameraresponse was recordedat 0502:26 UT on July 22, 1998, asthe aircraftwas
(1 MR), presumably
fromtheneutralN2 (2 PG) band,butthereis flying parallel to a narrow (---50 km) frontal thunderstorm
also -70 kR or 7% of the blue total in both the red and green extendingover 500 km east-westacrossIowa. Figure5 showsa
channels. We have no explanationfor the green emissions,as map of the National Lightning Detection Network (NLDN)
there are no likely molecular sources,and the red N2 (1 PG) lightning,the aircraft position,and the probablelocationof the
bandsare stronglyquenchedat thesealtitudes. Startersare not blue jet, assumingit is centeredin the storm on the azimuth
upward tropospheric lightning, which would have atomic determinedfrom the stars. The locationin Iowa is not nearany
broadbandemissionsand thereforehave similar brightnessin all major populationcenter,and the recordsof the National Weather
three channels of the camera. Consistent with the observations of
Servicefor the two nearestcounties(Mahaskaand Keokuk)do
Wescottet al. [1996, 1998], no +CG or -CG flash can be not containany reportsof largehail suchashasbeenreportedfor
associated with this event although negative flashes were bluejets and startersby Wescottet al.,[1996, 1998]. The Storm
Data and UnusualWeather Phenomena[July1998;availableat
predominantin the regionof the starter.
http://www5.ncdc.noaa.gov/pubs/publications.htlnl]
reported
only urbanand small streamflooding,with somepropertyand
4. Blue Jet: "Rocket Lightning"
crop damagein centralIowa.
Mackensie and Toynbe [1886] and Corliss [1977, 1983],
Figure6 presentsa 350-mssequence
of unfilterednarrowfield
publishedanecdotal
reportsof whatwascalled"rocketlightning" TV images.The unusualfeatureof this bluejet is that at 0.200 s
propagatingupwardfrom the tops of thunderclouds.It is not the jet has separatedinto two parts. The jet startsat t=0 and
clear from thosereportsexactly what was seen,spritesor blue propagates
upward
at23kms-1,untilit brightens
considerably
at
1 MR. Thus the 31 kR of ionized light amountsto ---3%of the
blue total.
I
o
I
I
I
112o
I
!
260
i
300
Time (ms)
Figure6. A 350-mssequence
of unfiltered
narrowfieldTV images,
16.67msapart,of thebluejet eventat 0502:26UT
onJuly22, 1998. At t=0,thejet appears
at 15.5-kmaltitude
andpropagates
upwardat 23 km s-1,untilit brightens
considerably
att=0.133s. Thelowerportion
ofthejet disappears
completely
after-0.266s. Afterthattheupper
portion
ofthejetcontinues
upward
at91kms-1anddisappears
fromsight
after0.333satanaltitude
of-35 km.
21,554
WESCOTT ET AL.' BRIEF REPORT
t=0.133 s. Next the uppertip of the jet continuesupwardat 91
References
kms-• anddisappears
belowobserving
threshold
after0.300s at
Corrliss,W.R., Handbookof unusualnatural phenomena.The
sourcebook
Project,GlenArmMaryland,542pp.,1977.
upperportioncontinues
upward.The 91 km s'• motionhas Corrliss,W.R., Handbookof unusualnaturalphenomena,
Anchor
smearedout the apparentlengthof the upperportionduringthe
Books/Doubleday,
Garden
City,NewYork,423pp.,1983.
H.L. andR.I. Mitchell,Thirteen-color
photometry
of 1380
33-msintegrationof the TV field by 3 km, sothe actuallengthof Johnson,
brightstars,Rev.Mexicana
Astron.Astrofis..,
5, 299-324,1975.
the separatedionized streak is -3 km. This blue jet followed
an altitudeof-35 km. The lower part decaysaway while the
three -CG flashes and one +CG flash located-70
km to the west
Lyons,
W.A.,Sprites,
elvesandbluejets,Weatherwise,
50(4),19-23,
1997
and 4 s after a -CG flash at 14-km distance. This confirms the
Mackensie,
T., andH. Toyabee,
Meteorological
phenomena,
Nature,24,
previousconclusions
that bluejets are not associated
with any
particularCG flash.
Pasko,W.P.,
U.S.Inan,andT.F.Bell,Bluejetsproduced
by quasi-
5. Summary
The new observations
of bluejets andstartersreportedin this
paperconfirmsomepreviousconclusions
andrevealseveralnew
results:
245, 1886.
electrostatic
pre-discharge
thundercloud
fields,Geophys.
Res.Lett.,
23(3), 301-304, 1996.
Roussel-Dupre,
E., E. Symbalisty,
Y. Taranenko,
andV. Yukhimuk,
Simulations
of highaltitude
discharges
initiated
byrunaway
electrons,
J. Atmos.Sol. Terr. Phys.,60, 917-940, 1998.
ROwland,
H.L.,Theories
andsimulations
of elves,
sprites
andbluejets,,
J. Atmos.Sol. Terr.Phy2,.,60, 831-844,1998.
A.I.,E.VMishin,.,P. Stubbe,
andM.JRycroft,
Onbluejet
1. The total optical brightnessin the visible blue end of the Sukhorukov,
dynamics,Geophys,
Res.Lett.23, 1625-1628,1996.
spectrumof a large bluejet over the oceannorth of Reunion
Sukhorukov,
A.I.,andP.Stubbe,
Problems
ofbluejettheories,
0r.Atmos.
Island,recordedon a colorphotograph
by oneof theauthors,is at
Sol. Terr.Phys.,60, 725-732,1998.
least 6.76 MR.
Symbalisty,E.M., R.A. Roussel-Dupre,
V. Yukhymuk,and Y.
Taranenko,
Fulltimeevolution
of upperatmospheric
discharges
due
2. Jetsare due to streamers,
and from the arguments
of
to runaway
airbreakdown
(abstract),
. EosTrans.AGU,77(46),Fall
Wescottet al. [1998] they are probablypositive.The Reunion
Meet. Suppl.,F70, 1996.
photograph shows eight smaller streamers(50- to 100-m
Vallance-Jones,
A., Aurora,Geophys.
Astrophys.
Monogr.vol.9, p.129,
diameter)startingnearthebaseof themainjet.
D. Reidel,Norwell, Mass.,1974.
3. Blue starters,andby association
alsobluejets, emitonly a Wescott,E. M., D. D. Sentman,D. Osborne,D. Hampton,and M.
Heavner,Preliminary
resultsfromthe Sprites94
aircraftcampaign:
smallfractionof thetotallightfromtheionizedN2 (1 NG) band
Bluejets,Geophys.
Res.Lett.,22(10),1209-1212,1995.
at 427.8nm,provingthattheyareat leastpartiallyionized.From
Wescott,E.M., D.D. Sentman,D.L. Hampton,M.J. Heavner,D.L.
starcalibrations
a 1 MR starterrecordedin multiplewavelengths
Osborne,
andO.H. Vaughan,
Bluestarters:
Briefupwarddischarges
was found to have 3% ionization.
froman intense
Arkansas
thunderstorm,
Geophys.
Res.Lett.,23(6),
2153-2156, 1996.
4. In one remarkable blue jet event the streamer front
M.J. Heavner,D.L. Hampton,andO.H.
separated
fromthe sourceat thebaseandcontinued
upwardfor- Wescott,E. M., D.D. Sentman,
Vaughan
Jr.,BlueJets'Theirrelationship
to lightning
andverylarge
100 ms, reachingan altitudeof -3 5 km.
5. Bluejetsarenotassociated
withanyparticular
CG lightning
flash,but they do occurin regionsof thunderstorms
of mostly
-CG activity. They are not restrictedto areasof largehailfall
overland. Theyalsooccurovertropicaloceanthunderstorms.
hailfall,andphysicalmechanisms
for theirproduction,
0r.Atmos.Sol.
Terr. Phys.,60, 713-714, 1998.
M.J.Heavner,
LosAlamosNationalLaboratory,
LosAlamos,NM
Acknowledgments.
Thisworkwassponsored
byNASA grantNAS55125. We thanktheaircraftpilots,JeffTobolskiandBertZimmerly,and 87545.
P. Huet,13 RueVermenouze,
Apt 13, 63-000Clermond
Ferrand,
staffof Aero Air LLC for theirexcellentsupportof the EXL98 aircraft
flightsduringwhichthe blue startersandbluejet were imaged. D. France.
D.R. Moudry,D.D. Sentman,H.C. Stenbaek-Nielsen,
and E.M.
Osborne
of theU. of Alaskaprovidedinvaluable
planning,
construction,
Geophysical
Institute,
University
of Alaska,903 Koyukuk
installation,andoperationof the opticalinstrumentation
on the aircraft. Wescott,
We thankKenCummins
of GlobalAtmospherics
Inc.for permission
to
use the NLDN
Drive,Fairbanks,
AK99775-7320.
(gene.wescott•gi.alaska.
edu)
data. We thank Tom Hallinan for his discussions and
contribution
to thequantitative
analysis
of theR6union
bluejet.
Janet G. Luhmann thanks Nikolai G. Lehtinen and another refere,e for
theirassistance
in evaluating
thispaper.
(Received
November
16,2000:revised
January
16,2001'
accepted
February19,2001.)