1. U.A. Dyudina, A.P. Ingersoll, 150 -21 California Institute of Technology Pasadena, CA, 91125 Objectives We study lightning on Jupiter using spatially resolved (~25 km/pixel) Galileo images of the light scattered by the Jovian clouds (Little et al. 1999). We compare the lightning images with the images produced by our 3D Monte Carlo light scattering model. We vary cloud opacity, cloud geometry and the lightning location in the model, and fit the falloff of the brightness away from the brightness center in the images. From the fitting we learn about the energy, depth, shape, and vertical extent of the lightning and also about the size, shape, and radiative properties of the surrounding clouds. The data The lightning flashes were observed by the Galileo spacecraft on the night side of Jupiter in November 1997, Galileo orbit E11 (flashes 1 and 2) and in May 1999, orbit C20 (flashes 3 and 4). The flashes chosen for our analysis are the largest ones. Their size allows to resolve them spatially. The appearance of flashes 1 and 2 is discussed in Little et al. (1999) ( storms a and b in Fig.9). Flashes 3 and 4 are mentioned in Gierasch et al. (2000), Fig. 1 c) without discussing the lightning spot structure. The model The model calculates Monte Carlo scattering of photons in a 3D opacity distribution. During each scattering, light is partially absorbed. The new direction of the photon after scattering is chosen according to the Henyey-Greenstein phase function. Images from different directions are produced by accumulating photons emerging from the cloud in different solid angles. Lightning is modeled either as a point source or a vertical line source. Top: The image of the largest lightning spot observed by the Galileo spacecraft. The image is foreshortened because of the slant viewing geometry (emission angle is 54.5 degrees) The red arrow points toward the spacecraft. Bottom: The brightness falloff from the brightness center in the flash 1 image. The brightness is normalized by the brightest pixel in the image. Each colored circle corresponds to one pixel. Colors represent azimuthal angle when the image is projected on to the horizontal plane as shown in the color chart in the upper right corner. The chart is centered at the brightest pixel. Red arrow points toward the spacecraft. The green and red pixels (along the line of sight of the spacecraft) are brighter than the averaged black curve. The blue pixels (in perpendicular direction) are fainter. This color difference shows that the image is azimuthally asymmetric, namely elongated along the line of sight of the spacecraft. Top: Two examples of the opacity distribution in the modeled cloud. Model 1 shows 3-dimensional cloud: optically dense hemisphere (  = 10) within the plane-parallel haze (  = 1). The point-size lightning is located at the cloud base (yellow star). Model 2 shows plane-parallel cloud of optical depth  = 10. Middle: Modeled images corresponding to the models above. Bottom: Brightness falloffs for images from different models above them. The black lines are observational data - the averaged brightness falloff from the Flash 1. The scale in the model is chosen to match the half width at half maximum in the data brightness falloff. This scaling determines the depths of lightning below the cloud tops. The depth is 105  15 km for flash 1. The 3-D cloud model 1 shows a better fit than the plane-parallel cloud model 2 in both the shape of the curve and the azimuthal asymmetry i.e, green and red pixels along the line of sight are brighter than the blue pixels in perpendicular direction. New estimate of lightning depth. Lightning is found to be very deep - about 100 km below the cloud tops for the flash 1 and about 40 km for flashes 2, 3 and 4. Finding lightning as deep as 100 km is puzzling because it assumes that the discharge happens at the pressures of at least 8 bars. The water cloud which is considered to be likely candidate to produce lightning can exist at the pressures of about 5 bars, much higher than our results suggest. The lightning depth derived is proportional to the lightning spot size. As the largest flashes were studied, the derived depth estimates the upper limit for the lightning depth. Such large lightning spots were first time observed by Galileo (Little et al. 1999). Our results confirm the estimates of lightning depth given in Little et al. (1999). Cloud opacity. Some of the observed lightning flashes have smooth brightness distribution with no single pixel or line of pixels standing out (lightning flashes that occupy a single pixel cannot be distinguished from cosmic ray hits). Smooth distribution means that lightning is not seen directly. Instead, we see the light scattered by the clouds. From the model, only clouds with optical depth  >5 can block the direct light. Therefore, at least in some locations clouds above the lightning have  >5. New estimate of lightning energy. As the clouds are optically thick, most of the light is scattered back down instead of coming toward the spacecraft. According to our model, the fraction of light that can penetrate the clouds having  >5 is 10% or less. That factor of 10 increases the estimate for the optical energy of lightning at the cloud tops given in Little et al. (1999). Assuming the largest flashes have smooth brightness distribution, their optical energy would be of order 10 11 Joules. This is an order of magnitude larger than largest terrestrial lightning termed superbolts. Flash 1 Flash 1 Data Model 1 Model 2 Plane-parallel cloudHemisphere+haze layer Distance (km) 104 km 117 km data Flash 2 Flashes 2, 3 and 4: Three lightning images on the left are compared with the model 2 results on the right. Images and brightness falloff plots are constructed the same way as for the Flash 1. The emission angles are 69, 39.4 and 39.4 degrees for flashes 2, 3 and 4 correspondingly. The derived depth of lightning below the cloud tops is 45  10, 42  5 and 34  5 km correspondingly. References Gierasch.,P.J., et al. Observation of moist convection in Jupiter’s atmosphere 2000, Nature 403, 628-630 Little,B., et al. Galileo Images of Lightning on Jupiter 1999, Icarus 142,306-323 Brightness DataModel Flash 2 data Brightness Flash 4 DataModel Flash 4 data Brightness Flash 3 DataModel Brightness Flash 3 data Distance (km) Lightning scattered by 3D clouds. Monte Carlo simulation of Galileo images of lightning on Jupiter.