Satellite-based modeling of radiation and energy fluxes over the Osaka Metropolitan Area and adjacent forests using Himawari-8 geostationary satellite data

Abstract

Radiation and surface energy exchanges are essential for understanding the urban heat island effect, with land surface temperature (LST) serving as a key indicator of surface thermal conditions. Geostationary satellites, such as Himawari-8, provide high-frequency observations that enable detailed analyses of diurnal thermal dynamics across urban and surrounding landscapes. This study developed a satellite-based framework that integrates radiation and surface energy-budget models to provide a physically constrained diagnosis of half-hourly and seasonal surface energy partitioning using Himawari-8-derived LST. Model performance was evaluated against eddy-covariance measurements at two urban and two forest sites in the Osaka Metropolitan Area and adjacent forests, Japan. Energy fluxes were calculated using bulk and single-source models. The single-source model generally improved turbulent-flux estimates, particularly at the forest sites, with root mean square errors of 36–69 W m^-2 for sensible heat flux and 27–48 W m^-2 for latent heat flux. Its improved performance was associated with reduced propagation of LST uncertainty. Regional simulations for the Osaka Metropolitan Area showed that urban areas had lower net radiation, higher ground heat flux, and lower aerodynamic conductance than forests, resulting in lower sensible heat flux. Among urban landscapes, low-rise buildings exhibited weaker turbulent coupling, whereas compact high-/mid-rise buildings partitioned more available energy into storage and consequently had smaller turbulent fluxes. These results demonstrate the utility of geostationary thermal remote sensing not only for monitoring urban energy fluxes at high temporal resolution but also for diagnosing the surface energy partitioning underlying observed thermal patterns.