Characterisation and localisation of lightning and explosions by a flotilla of stratospheric balloons

On 3 August 2021, Sandia launched a flotilla of four Heliotrope solar hot air balloons (Bowman et al., 2020) from Belen regional airport in New Mexico (USA) to coincide with the launch of the Boeing Starliner rocket. Three balloons were equipped with two Gem microphones (Anderson et al., 2018) spaced vertically from 30 to 100 m apart, depending on the balloon. The fourth balloon had two infraBSU microbarometers mounted in opposite polarity in order to suppress local interference. These Heliotrope balloons allow level flights between 15 and 25 km altitude for several hours from sunrise to sunset. The rocket launch was cancelled after the balloons were launched, but eight chemical explosions of between 45 and 135 kg TNT equivalent and a thunderstorm took place near the balloons in the first few hours of cruise. After characterizing the measurements of explosions on balloons and two ground seismic stations, we evaluate the performance of a Bayesian method for locating explosions, taking into account local meteorology and whether or not seismic measurements were included. Using acoustic measurements under balloon conditions alone, we highlight the importance of network geometry and propagation between the explosion and the sensors in the localization error. We then characterize the lightning that occurred in a single thunderstorm cell located between 10 and 40 km from three of the four balloons. Several lightning flashes are clearly identified (using the method proposed by Farges and Blanc (2010) for ground thunder measurements and Lamb et al. (2018) for first stratospheric balloon measurements of lightning) and located in 3D with the method validated from explosion measurements. These 3D locations are compared with the 3D thunder model defined by Lacroix et al. (2010) and with the 3D distribution of thunder sound power described by Bestard et al. (2023). Finally, as the storm lasted approximately 45 minutes, continuous thunder emissions occurred. We calculated the cross-correlation time between measurements made under one of the balloons at 30 m apart during this period. We can see a change in the relative arrival delay as the balloon moves away from the storm cell, which does not move over this period. This suggests propagation through the AtmoSOFAR channel as defined by Albert et al (2023). SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.