During the aircraft flights of the Sprites94 Campaign over the central part of the U. S., two thunderstorm related phenomena were documented. In a companion paper, Sentman et al., [1994], discuss "Sprites" which are short duration red flashes in the D region, sometimes with bluish tendrils extending down to the apparent cloud tops. On flights 1 and 3 July, 1994 UT, we documented many examples of a second distinct phenomenon which is clearly what was reported by Hammerstrom [1993]. We have named them "Blue Jets" based upon their observed characteristics and color [Sentman and Wescott, 1994]. Observations were made from two aircraft instrumented with color and black and white TV cameras. See Sentman et al. [1994] for details of the Sprites94 campaign, the aircraft and optical equipment. The use of two aircraft allowed us to triangulate the positions and altitudes of the blue jets vs. time as they propagated upwards. The triangulations also allowed us to determine the velocities and the angular spreads.
We speculate on possible mechanisms, but none seem very satisfactory.
Upward lightning strokes similar to cloud to cloud lightning have been reported in the literature [Everett, 1903; Powell, 1968; Taylor, 1972; Wood, 1951]. During the first pass by the Arkansas storm we recorded nearly two dozen events which we interpret as upward lightning. They were brief and very bright, and extended upward from the cloud tops for much shorter distances than the blue jets. In some cases an upward stroke preceded blue jets by a few frames, and then the jet seems to have originated from the same location. There were some examples of two jets starting out close to each other separated by tens of ms in time. There were three examples of jets starting from the same apparent location and following the same track separated by 50 to 70 ms.
The typical blue jet is observed to appear from the apparent top of the anvil and propagate upwards in a narrow cone, and to flare out as it reaches maximum altitude such that it resembles a trumpet. We measured the cone angle on 18 examples and calculated a mean value of 14.7o ± 7.5o. The high value was 31.5o and the low value was 6.5o. Figure 3 shows a time sequence of stereo pair TV images 67 ms apart taken from the 90o field of view black and white cameras on the aircraft. This example has been triangulated and yields a terminal altitude of at least 44 km. In most cases the jet seems to fade away all along the cone simultaneously, about 200 ms after it begins. This can be seen clearly 233 ms after the first appearance in the last frames of Figure 3. The jet had a cone angle of 9.5o. All of the jets were quasi-vertical with differing angles. None were magnetic field aligned.
We have made some estimates of the brightness of blue jets using star calibrations to develop a brightness function vs. signal level in the pixel counts. Both the color TV and the B/W images were used. Assuming that the jet signal level falls at the peak blue TV response, then in early times the jet would be 6.7 kilo Rayleighs (kR). If the major emission is at 427.8 nm, the brightness would be 10 kR, from the color data. By use of the B/W camera data overall integrated response, we find that the estimated brightness is 16.6 kR. If the emission is at 427.8 nm, the brightness is 19.7 kR. It is more difficult to estimate the 391.4 nm contribution as the system response is falling much more steeply in that wavelength region in the color camera than in the B/W cameras.
The altitudes of the two aircraft are accurately known from GPS data recorded on the video frames. We digitize the TV frames using a True Vision Targa M8 board. The digitized frame is then corrected for lens distortion and for pincushion and off axis distortions introduced by the SIT camera. We then use a computer program called "Stereo" written by H.C. Stenbaek-Nielsen [personal communication] to fit the stars in the field of view to the positions in the Smithsonian Star Catalogue [1969]. By use of Stereo we can triangulate on any feature visible from both aircraft. The precision of the position determination depends upon several factors: 1) The number of stars in the field of view and star field fit; 2) the distance to the feature from both aircraft; and 3) The brightness and contrast of the feature with respect to the night sky brightness. The palpable front ends of the blue jets lose brightness and contrast with respect to the background as they propagate to altitudes above about 30 km, but we can reasonably triangulate the position of the front to about 40 km altitude.
Some information is lost in the digitization process. When the jets are observed in real time using the original video tapes the jets can be observed well above 40 km. On some we see what appears to be a hemispherical "shock wave" continuing at the original velocity to perhaps 50 km. We have triangulated the position vs. time of six separate blue jets. The altitude vs. time distribution of these jets are shown in Figure 4 along with the data points for one jet at 03:05:05 UT. The average vertical velocity, Vz is 98 ± 14 km/s, with a high value of 114.4 and a low value of 78.9 km/s.
The jets differ from sprites in color [Sentman et al. 1994]. We did not obtain a spectrum of a jet. But preliminary analysis of the color TV raw signal levels after subtracting background levels shows that ratio of blue to green in the brightest part of the jet was about 5:1 with no detectable red component. There will be further detailed calibrated analysis of color ratios.
Although the blue jets do not resemble stepped leaders, we note that the propagation velocity is close to that of type-a leaders, which have a uniform earthward speed of about 100 km/s [Uman, 1987]. Within the resolution of our video cameras there is little discernable spatial structure in the blue jets and no evidence of the sort of branching and forking observed in stepped leaders, but it may be possible that the jets are discharges following some sort of collimated quasi straight ion trails emanating from the clouds.
On some blue jets we can see what appears to be a faint hemispherical "shock front" ahead of the difuse leading termination with velocity about the same as the average velocity of the earlier front of the jet. If there is a sonic shock wave it would be at Mach 300, and might produce luminosity, but the excited molecules could not continue to emit for 200 ms. One would expect a travelling shock to produce a "roman candle" effect, and we wonder how a shock wave would be so well collimated.
If nature produces cones of runaway electrons in thunderstorms, then the blue jets might be a result of electron collisions with and ionization of the atmosphere. It is not clear however how the 100 km/s velocity would arise from runaway electrons. A 100 km/s electron has only 4 ´ 10-3 eV energy, so the electrons would have to have much higher energy and velocity to excite and ionize molecules and atoms. A 100 km/s proton would have about 7 eV energy, but with small mean free path in the atmosphere. Any mechanism producing the blue jets, including particle acceleration, would have to continue for at least 200 ms.
Based upon the preliminary analysis done so far, the following summarizes the salient characteristics of blue jets which a successful theory must explain:
Cole, Jr., R.K, R.D. Hill and E.T. Pierce, Ionized columns between thunderstorms and the ionosphere. Journal of Geophysical Research, 71, 959-964, 1966.
Corliss, W.R., Handbook of Unusual Natural Phenomena. The Sourcebook Project, Glen Arm, MD, 542 PP., 1977.
Corliss, W.R., Handbook of Unusual Natural Phenomena. Anchor Books/Doubleday, Garden City, New York, 423 PP., 1983.
Everett, W.H., Rocket lightning, Nature, 68, 599, 1903.
Fisher, J.R., Upward discharges above thunderstorms, Weather, 45, 451-452, 1990.
Hammerstrom, J.H., Mystery lightning, Aviation Week & Space Tech., 6, Aug. 30, 1993.
Malan, D., Sur les decharges orageuses dans la haute atmosphere, Academie des Sciences, November 3 Session, 1937. (original in French, translated by Liv Nordem Lyons).
Powell, G., Lightning, Marine Observer, 38, 173, 1968.
Sentman, D. and E.M. Wescott, Red Sprites & Blue Jets, Geophysical Institute Video Production, University of Alaska Fairbanks, 9 July, 1994.
Sentman, D., E.M. Wescott, D. Osborne, D. Hampton and M. Heavner, Preliminary results from the Sprites94 aircraft campaign: Red sprites, Geophys. Res. Lett., submitted Nov., 1994.
Smithsonian Astrophysical Observatory, Star Atlas of Reference Stars and Nonstellar Objects, MIT Press, Cambridge Massachusetts, 1969.
Taylor, T.J., Lightning, Marine Observer, 42, 12, 1972.
Uman, M., The Lightning Discharge, International Geophysics Series V. 39, Academic Press, Inc., Orlando, 1987.
Wilson, C.T.R., A theory of thundercloud electricity, Proc. Royal Meteor. Soc., London, 236, 32D-37D, 1956.
Wood, C.A., Unusual Lightning, Weather, 6, 64, 1951.
Wright, J.B., A thunderstorm in the tropics, Weather, 6, 230, 1951.
Figure 2. Rate of blue jet occurrence during 22 minute interval of observation during the first pass past the storm shown in Figure 1.
Figure 3. Time sequence of video frames, 67 ms apart, during a blue jet on July 4, 1994, from the Jet Commander, (left) and the Westwind 2, (right). These can be viewed in stereo. Note the brightness is starting to fade all along the jet at 03:06:12.21 in the last pair of frames.
Figure 4. Altitude vs. time plot for blue jet at 03:05:05 UT on July 1, 1994 with the least squares fit to the data, and the least squares lines for five other triangulated jets.