Meteorites Colliding Craters Inversion based on Kernel Gaussian Mixture Ridge Regression

Inverting kinetic energy impacts of high-speed meteoroids is an effective means to study the dynamics of celestial impact, thereby, evaluate the impact effects and hazards. This paper uses a Smoothed Particle Hydrodynamics method based Celestial Impact Dynamics (PySPH-CID) to simulate the ejecta and crater evolution process of meteorites colliding into Earth; and a Kernel Gaussian Mixture Ridge Regression (KGMRR) method is proposed to fit the one-to-many mapping between cause and effect of an impact with multiple sets of feasible solutions; a solution space is obtained using Gaussian Process (GP), for the formation conditions of craters after a small celestial body impacts. An inversion of colliding crater of Yilan in Heilongjiang Province, Northeast China, 46°23'03'' N and 129°18'40'' E, with 1.85km diameter, which located entirely on the cretaceous granite bedrock and preserved the remnants of two-thirds of the crater rim, was made. Assuming the impactor diameter is around 120m, preliminary experiment gives colliding cause vectors that impact velocities range between 17.8~15.0 km/s with 60.7°~66.6° obliquely incident, impactor density varying in scope of 2772~2815 kg/m3. This work provides a direct reference for effectively defending against the threat of foreign celestial bodies to the Earth, developing and transforming controlled impact technologies, as well as implementing precise and controllable impacts.