Accumulations of spaceborne precipitation radar observations have enabled us to study characteristics of precipitation globally and statistically. Above all, structures of extreme precipitation are studied to renew our conventional assumption that intense precipitation comes from tall and intense convection. Further studies of wide-spread heavy rainfalls in very moist season in Japan clarified the contrasts between the severe rainfalls resulted from very unstable atmospheric condition and those from very wet atmosphere. The former is related to instability in terms of Convective Available Potential Energy (CAPE), while the latter is related to Moist Available Unstable Layer (MAUL) in very moist atmospheric conditions. Therefore, increase of moisture can increase the frequency of very heavy rainfalls without increasing the atmospheric instability in conventional sense. We introduce a series of studies that led us to restructure the knowledge of severe storms.

Granular matter is a ubiquitous component of various natural systems, ranging from planetary terrains to ecological environments. A defining feature of granular behavior is its inherent friction and yield (cohesive) strength, which lead to complex macroscopic manifestations. This article discusses the physical aspects of asteroid terrains, crab-burrow stability, and slow-earthquake statistics by focusing on these fundamental properties. Force-chain structures in granular beds are also explored to provide a microscopic foundation for these macroscopic landscapes.

Studies on formulation of correspondence between interacting fermions and non-interacting multi-component fermions is reviewed in examples of the Hubbard model, inspired by the successful descriptions of strongly correlated electron systems by fractionalized fermions. The formalism enables constructions of a neural network for a quantum many-body solver called Fermi machine represented by coupled noninteracting fermions with quantized hidden variables. Numerical algorithm of the Fermi machine for the Hubbard model is outlined with future outlook.