A decadal, hourly high-resolution satellite dataset of aerosol optical properties over East Asia

Abstract. Formerly known as one of the most polluted regions of the globe, East Asia underwent a dramatic improvement of air quality, especially for aerosols, starting in the 2010s. Numerous satellites have observed East Asia for a long time duration, but often with a low spatial or temporal resolution, limiting their ability to capture small-scale variabilities or provide continuous observations of long-range transport of aerosols. In this study, we provide an hourly aerosol optical property (AOP) dataset retrieved from the Korean Geostationary Ocean Color Imager (GOCI), with a high spatial resolution of 2 km at nadir, covering the entire operational period from March 2011–March 2021. The dataset is retrieved using the Yonsei Aerosol Retrieval Algorithm, providing aerosol optical depth (AOD) at 550 nm as the primary product, along with fine mode fraction, single scattering albedo, Ångström exponent, and aerosol type as ancillary products. Seasonal validation of AOD against the Aerosol Robotic Network (AERONET) showed that the fraction of data points within the expected error range of 0.05 + 15 % varied from 56.4 % in June-July-August to 64.5 % in September-October-December, with the mean bias generally within ±0.05. Compared to the operational version, the high-resolution product demonstrated improved retrieval capability in the presence of broken clouds, along complex coastlines, and in capturing AOD variability at the sub-district level. The decadal AOD exhibited a decreasing trend over four major cities within the observation domain. We expect this data to be widely used in climate modelling, reanalysis, atmospheric chemistry, marine optics, environmental health studies, variability and trend analysis, contributing to a more comprehensive understanding of the interactions between climate change, trace gases, human health, and AOPs. The dataset presented in this work is publicly available for download at https://doi.org/10.7910/DVN/WWLI4W (Lee et al., 2025).