Abstract
Lunar impact flashes are a rich source of data on highintensity, high-velocity impacts in the absence of significant atmosphere. Studies of these flashes have established that the source of much of the radiation emitted by the flash is through thermal emission from hot condensed debris produced by the impact. Here we present a quantitative forward flash model that uses high-fidelity simulations of the hypervelocity impact that are then handed off to a full radiative transfer code to calculate its radiative evolution and flash intensity. Our method uses machine learning methods to calculate the state of the debris cloud close to its radiating state. As a case study, a lunar flash observed during the January 2019 lunar eclipse is compared to a lunar flash generated using our method under the conditions estimated for that impact.
