• Korean Journal of Remote Sensing
• /
• v.39 no.5_3
• /
• pp.933-948
• /
• 2023

Atmospheric aerosols not only have adverse effects on human health but also exert direct and indirect impacts on the climate system. Consequently, it is imperative to comprehend the characteristics and spatiotemporal distribution of aerosols. Numerous research endeavors have been undertaken to monitor aerosols, predominantly through the retrieval of aerosol optical depth (AOD) via satellite-based observations. Nonetheless, this approach primarily relies on a look-up table-based inversion algorithm, characterized by computationally intensive operations and associated uncertainties. In this study, a novel high-resolution AOD direct retrieval algorithm, leveraging machine learning, was developed using top-of-atmosphere reflectance data derived from the Geostationary Ocean Color Imager-II (GOCI-II), in conjunction with their differences from the past 30-day minimum reflectance, and meteorological variables from numerical models. The Light Gradient Boosting Machine (LGBM) technique was harnessed, and the resultant estimates underwent rigorous validation encompassing random, temporal, and spatial N-fold cross-validation (CV) using ground-based observation data from Aerosol Robotic Network (AERONET) AOD. The three CV results consistently demonstrated robust performance, yielding R²=0.70-0.80, RMSE=0.08-0.09, and within the expected error (EE) of 75.2-85.1%. The Shapley Additive exPlanations(SHAP) analysis confirmed the substantial influence of reflectance-related variables on AOD estimation. A comprehensive examination of the spatiotemporal distribution of AOD in Seoul and Ulsan revealed that the developed LGBM model yielded results that are in close concordance with AERONET AOD over time, thereby confirming its suitability for AOD retrieval at high spatiotemporal resolution (i.e., hourly, 250 m). Furthermore, upon comparing data coverage, it was ascertained that the LGBM model enhanced data retrieval frequency by approximately 8.8% in comparison to the GOCI-II L2 AOD products, ameliorating issues associated with excessive masking over very illuminated surfaces that are often encountered in physics-based AOD retrieval processes.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.29 no.1
• /
• pp.84-91
• /
• 1996

Heat flux of the East China Sea was estimated with the bulk method, the East China mount based on the marine meteorological data and cloud amount data observed by a satellite. Solar radiation is maximum in May and minimum in December. Its amount decreases gradually southward during the winter half year (from October to March), and increases northward during the summer half year (from April to September) due to the influence of Changma (Baiu) front. The spatial difference of long-wave radiation is relatively small, but its temporal difference is quite large, i.e., the value in February is about two times greater than that in July. The spatial patterns of sensible and latent heat fluxes reflect well the effect of current distribution in this region. The heat loss from the ocean surface is more than $830Wm^{-2}$ in winter, which is five times greater than the net radiation amount during the same period, The annual net heat flux is negative, which means heat loss from the sea surface, in the whole region over the East China Sea. The region with the largest loss of more than $400Wm^{-2}$ in January is observed over the southwestern Kyushu. The annual mean value of solar radiation, long-wave radiation, sensible and latent heat fluxes are estimated $187Wm^{-2},\;-52Wm^{-2},\;-30Wm^{-2}\;and\;-137Wm^{-2}$, respectively, consequently the East China Sea losses the energy of $32Wm^{-2}(2.48\times10^{13}W)$. Through the heat exchange between the air and the sea, the heat energy of $0.4\times10^{13}W$ is supplied from the air to the sea in A region (the Yellow Sea), $2.1\times10^{13}W$ in B region (the East China Sea) and $1.7\times10^{13}W$ in C region (the Kuroshio part), respectively.

• Korean Journal of Remote Sensing
• /
• v.30 no.1
• /
• pp.109-126
• /
• 2014

We improved the Land Surface Emissivity (LSE) data (Kongju National University LSE v.2: KNULSE_v2) over the Communication, Ocean and Meteorological Satellite (COMS) observation region using recent(2009-2012) Moderate Resolution Imaging Spectroradiometer (MODIS) data. The surface emissivity was derived using the Vegetation Cover Method (VCM) based on the assumption that the pixel is only composed of ground and vegetation. The main issues addressed in this study are as follows: 1) the impacts of snow cover are included using Normalized Difference Snow Index (NDSI) data, 2) the number of channels is extended from two (11, 12 ${\mu}m$) to four channels (3.7, 8.7, 11, 12 ${\mu}m$), 3) the land cover map data is also updated using the optimized remapping of the five state-of-the-art land cover maps, and 4) the latest look-up table for the emissivity of land surface according to the land cover is used. The updated emissivity data showed a strong seasonal variation with high and low values for the summer and winter, respectively. However, the surface emissivity over the desert or evergreen tree areas showed a relatively weak seasonal variation irrespective of the channels. The snow cover generally increases the emissivity of 3.7, 8.7, and 11 ${\mu}m$ but decreases that of 12 ${\mu}m$. As the results show, the pattern correlation between the updated emissivity data and the MODIS LSE data is clearly increased for the winter season, in particular, the 11 ${\mu}m$. However, the differences between the two emissivity data are slightly increased with a maximum increase in the 3.7 ${\mu}m$. The emissivity data updated in this study can be used for the improvement of accuracy of land surface temperature derived from the infrared channel data of COMS.

• Journal of the Korean earth science society
• /
• v.28 no.3
• /
• pp.298-310
• /
• 2007

Intercomparison between eight radiative transfer codes used for the studies of COMS (Communications, Ocean, and Meteorological Satellite) in Korea was performed under pure molecular, i.e., Rayleigh atmospheres in four shortwave fluxes: 1) direct solar irradiance at the surface, 2) diffuse irradiance at the surface, 3) diffuse upward flux at the surface, and 4) diffuse upward flux at the top of the atmosphere. The result (hereafter called the H15) from Halthore et al.'s study (2005) which intercompared and averaged 15 codes was used as a benchmark to examine the COMS models. Uncertainty of the seven COMS models except STREAMER was ${\pm}4%$ with respect to the H15, comparable with ${\pm}3%$ of Halthore et al.'s (2005). The uncertainty increased under a large $SZA=75^{\circ}$. The SBDART model generally agreed with the H15 better than the 6S model, but both models in the shortwave infrared region were equally good. The direct solar irradiance fluxes at the surface, computed by the SBDARTs of four different users, were different showing a relative error of 1.4% $(12.1Wm^{-2})$. This reason was partially due to differently installing the wavelength resolution in the flux integration. This study may be useful for selecting the optimum model in the shortwave region.