Development of a Surface Reflectance Consistency Algorithm between KOMPSAT-3A and Sentinel-2A/Landsat Satellites

Surface reflectance consistency of multi satellite optical satellites is an essential factor for quantitative Earth observation and multi-satellite data integration. However, due to differences in spectral response functions (SRFs), band definitions, and preprocessing strategies, surface reflectance discrepancy between satellite sensors still exists.

In this study, the surface reflectance data of Korea Multi-Purpose Satellite (KOMPSAT)-3/3A were analyzed together with Sentinel-2 MSI, Landsat-8/9 OLI, MODIS, and New-space Earth Observation Satellite (NEONSAT) data. In order to minimize the influence of surface characteristics, radiometrically stable Pseudo-Invariant Calibration Sites (PICS), including desert areas and the North Pacific Ocean (NPO), were selected.

Before performing the surface reflectance consistency analysis, sensor-dependent reflectance differences were analyzed using spectral response functions (SRFs) based on the USGS Spectral Library. Spectral differences between sensors were evaluated by simulating band-equivalent reflectance through convolution of established hyperspectral surface reflectance spectra, including the USGS Spectral Library Version 7, with the spectral response function (SRF) of each sensor. Based on these simulation results, the Spectral Band Adjustment Factor (SBAF) was calculated by applying the median-based ratio method and the regression method constrained to pass through the origin. The calculated SBAF was evaluated using SRF-based simulated reflectance, and differences in reflectance between sensors before and after adjustment were quantitatively compared and analyzed using statistical indicators such as mean, standard deviation, and RMSE.

Surface reflectance differences showed sensor- and band-dependent patterns, with more evident deviations appearing in the near-infrared (NIR) region compared to other spectral bands. Based on the SRF-based SBAF evaluation, agreement among sensors generally increased, while the degree of improvement varied depending on the spectral band and adjustment strategy, resulting in residual discrepancies in some cases. Overall, these observations summarize the present characteristics of surface reflectance differences observed between KOMPSAT and other optical satellite sensors.

In future studies, the selected PICS will be used to apply radiative transfer model–based atmospheric correction using the 6S model, in order to further assess and improve surface reflectance consistency across multiple optical satellite sensors.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT)(RS-2025-00515357).