Reproduction of night sky brightness variations in urban areas of Japan caused by aerosols, artificial ground-based light, and surface reflectance

Light pollution has become an increasingly serious environmental problem, and quantitative assessment requires methods to estimate the spatial and temporal variability of night sky brightness. Previous global-scale models have provided valuable estimates of artificial night sky brightness; however, incorporating variations in atmospheric and surface environmental conditions remains challenging. In this study, we developed a regional-scale night sky brightness model that considers aerosols, ground-based artificial light, and land surface reflectance. The model includes radiative transfer calculations that account for multiple scattering in the atmosphere and multiple reflections between the atmosphere and the land surface. Three-dimensional light diffusion is represented using a point spread function derived from Monte Carlo photon transport simulations. The model reproduced the spatial distribution of hemispheric average radiance across Japan and showed improved agreement with ground-based observations compared with previous estimates. However, systematic biases remained, with overestimation under bright-sky conditions and underestimation under dark-sky conditions. Using the model, we analyzed long-term variations in night sky brightness in seven major Japanese cities from 2013 to 2023. Although regional night sky brightness generally increased with population size, differences among cities indicated the influence of urban structure. Night sky brightness decreased during the COVID-19 pandemic period and was associated with changes in both aerosols and ground-based artificial light emissions. This study provides a framework for quantitatively evaluating light pollution while accounting for atmospheric and surface environmental conditions and can be applied to regions with diverse environmental characteristics.