The 2020 eruption of the Nishinoshima volcano, Japan, emplaced new lava flows as well as numerous ash plumes from explosive eruptions. Several of these plumes were imaged by the Advanced Spaceborne Thermal Emission and reflection Radiometer (ASTER) as part of the Urgent Request Protocol (URP) during July 2020. Linear spectral deconvolution was used to model ASTER thermal infrared (TIR) data using the ASTER Volcanic Ash Library (AVAL), a spectral library containing volcanic ash compositional and particle size range end members. The airborne ash plume best matched the andesite library end member suites having a significant fine (<25 μm) fraction proximal to the vent, however particles >63 μm were also detected downwind, and distally around the cloud edges, assumed due to sampling of lower levels of the plume containing these larger particles even as the optical depth remains high. These results demonstrate that both composition and particle size variations within a plume can be estimated from the high spatial and spectral resolution ASTER TIR data using this approach, provided that representative spectral end members are present in AVAL and where the plume remains opaque to upwelling radiance.