Solar energy, fundamental to Earth’s climate and life, plays a critical role in shaping the land surface energy budget, water cycle, and biogeochemical processes while providing direct feedback to the climate system (Zepp et al., 2007). Among the many components influencing solar energy dynamics, surface albedo stands out as a key element of Earth’s energy balance. Representing the fraction of solar radiation reflected by the surface, surface albedo directly influences the distribution of energy within the Earth system (Dirmeyer and Shukla, 1994). Variations in surface albedo create feedback loops within the climate system, exerting complex and long-term effects on the global environment (Lee et al., 2020). For instance, a reduction in vegetation cover can lead to an increase in surface albedo, resulting in amplified feedback through decreased soil moisture and nutrient availability, further exacerbating vegetation loss (Knorr and Schnitzler, 2006).

Surface albedo, which significantly impacts the climate system, is strongly modulated by interactions with aerosols and clouds. Aerosols and clouds influence surface albedo by altering the amount of solar radiation reaching the Earth’s surface (Jadhav et al., 2024). Aerosol Optical Depth (AOD), a widely used satellite-derived parameter, quantifies the extent to which aerosols in the atmosphere scatter and absorb solar radiation, serving as a critical indicator of air pollution levels (Alston et al., 2011). The distribution and trends of AOD and cloud cover have been extensively studied, with research by Benas et al. (2020) examining their variations in southern China from 2006 to 2015. Similarly, Cloud Optical Thickness (COT), a metric that characterizes the scattering and absorption properties of clouds, provides essential information about their radiative effects and physical structure (Aebi et al., 2020).

East Asia, a region particularly vulnerable to climate change and extreme weather events, has experienced significant atmospheric changes due to rapid industrialization and urbanization (Rahman et al., 2023). These changes have had profound effects on aerosols, clouds, and surface albedo, underscoring their critical roles in the region’s climate dynamics (You et al., 2022). However, existing studies have predominantly focused on localized regions, such as specific parts of India and China (Singh et al., 2023; Qin et al., 2018). This geographic bias limits a comprehensive understanding of the broader Northeast Asian climate and its underlying mechanisms.

To address these limitations, this study utilizes geostationary satellite data from the Communication, Ocean, and Meteorological Satellite (COMS) to analyze ground surface albedo, AOD, and COT—key indicators of solar radiation dynamics. By leveraging these datasets, this research aims to identify fluctuations in these parameters across Northeast Asia, offering insights into both seasonal variability and long-term trends. Through this approach, we seek to overcome the constraints of prior studies and provide a more holistic understanding of the region’s climate dynamics.

In this study, the research area encompasses Northeast Asia, ranging from 15°N to 50°N in latitude and 105°E to 150°E in longitude. The dataset utilized for this study was acquired from Communication, Ocean, and Meteorological Satellite (COMS), South Korea’s first geostationary multipurpose satellite which was launched in 2011. It is positioned at an altitude of 36,000km above the Earth’s equator. It serves multiple missions, including weather observation, ocean monitoring, and communication services. Among its payloads, the Meteorological Imager (MI) is dedicated to geostationary weather observation. The MI captures data across one visible light band with a spatial resolution of 1 km and four infrared bands with a spatial resolution of 4 km for each channel. The study period extends from January 2012 to December 2019.

Black-sky albedo refers to the surface albedo under conditions where all incoming sunlight is transmitted directly through the atmosphere without scattering. COMS surface albedo data are calculated globally on a daily basis with a spatial resolution of 1 km, using a single channel in the visible spectrum. The albedo calculation algorithm for COMS/MI involves three main steps: atmospheric correction, Bidirectional Reflectance Distribution Function (BRDF) modeling, and broadband conversion. In this study, Black-sky Albedo data, matched to the East Asian region, was used at a spatial resolution of 4 km.

AOD data derived from COMS/MI are based on observations from the broadband visible channel, which spans a wavelength range of 0.55–0.80 μm with a center wavelength of 0.67 μm, and are provided at a spatial resolution of 1 km. An essential step in the AOD retrieval process involves identifying clear-sky pixels that are free from cloud interference. For these clear-sky pixels, the observed values in the visible channel are converted into aerosol-induced reflectance values. AOD data are provided with a spatial resolution of 4 km and a temporal resolution of 1 hour over the Northern Hemisphere extended area (Lee, 2018). In this study, the AOD data were matched to the East Asian region with a spatial resolution of 4 km.

COT derived from COMS/MI is calculated using data from the visible channel (0.67 μm) and the shortwave infrared channel (3.75 μm). Due to the reliance on the visible channel, COT can only be retrieved during daytime. The calculation process is influenced by surface reflectance, making accurate surface reflectance data critical for determining both COT and the effective particle radius. During the daytime, the shortwave infrared channel includes contributions from Earth-emitted radiation, which are corrected using radiance from the 10.8 μm infrared channel. The radiances obtained from the visible channel (0.65 μm) and the shortwave infrared channel (3.75 μm), corrected for surface and atmospheric influences, are matched to a precomputed Look-Up Table (LUT) generated using the Radiative Transfer Model (RTM). This LUT enables the simultaneous determination of optimal COT and effective particle radius. The resulting COT data are provided at 1-hour intervals with a spatial resolution of 4 km across the Northern Hemisphere extended area. In this study, COT data were specifically matched to the East Asian region with a spatial resolution of 4 km.

In this study, surface albedo, AOD, and COT datasets utilized had differing spatial resolutions, necessitating a harmonization process to standardize the resolution across all datasets. For spatial resolution alignment, a preprocessing method based on the Nearest Neighbor approach, specifically the construction of a coreset KD-Tree (cKD-Tree), was employed to resample all data to a uniform resolution of 4 km.

In addition to spatial resolution harmonization, temporal resolution alignment was performed to facilitate simultaneous analysis of albedo data, provided as daily observations, with AOD and COT data, provided at hourly intervals. To enable comparison across datasets, monthly mean fields were generated for albedo, AOD, and COT. This was achieved by calculating the arithmetic mean of all daily and hourly data within each pixel corresponding to the same year and month for each dataset. Furthermore, to ensure comparability with albedo data, which is available only over land, a Land-Sea mask was applied to the monthly mean fields of AOD and COT to extract values exclusively over land areas.

In this study, the long-term trends of surface albedo, AOD, and COT—products derived from COMS/MI—were analyzed and compared through time series analysis and by examining the spatial distribution of monthly mean fields and trend distributions. Time series data often include seasonal variations, making it challenging to discern the underlying trends. To isolate the pure trends of surface albedo, AOD, and COT while excluding seasonal fluctuations, an additional Seasonal and Trend decomposition using LOESS (STL) analysis was performed. STL is a method for decomposing time series data into three components: trend, seasonality, and residual.

The trend represents the long-term pattern in the data, seasonality captures periodic variations occurring at regular intervals, and residual reflects random fluctuations that remain after removing the trend and seasonality. By separating these three components, STL facilitates a better understanding of the complex structure of time series data(Lee, 2023).

In this study, the monthly mean time series of surface albedo, AOD, and COT were analyzed. Fig. 1 presents the monthly mean time series of each variable during the study period, from January 2012 to December 2019. Over this period, surface albedo, AOD, and COT all showed a decreasing trend, with the rates of decline quantified as –0.0003, –0.0008, and –0.0328 per month, respectively.

Figure 1. Monthly time series of Surface albedo, aerosol optical depth (AOD), and cloud optical thickness (COT) over East Asia from 2012 to 2019.

The monthly mean time series revealed a repeating seasonal pattern combined with an overall downward trend. All variables maintained their annual minimum values, but their yearly peaks gradually declined. Surface albedo showed higher values during winter, influenced by snow cover, and lower values during summer, affected by vegetation growth. AOD increased from spring through autumn and decreased during winter, with the highest monthly mean AOD values consistently observed in October each year. COT exhibited higher values in winter and spring, decreasing rapidly through summer and autumn, with a clear downward trend over the study period.

Analysis of the spatial distribution of trends in the monthly mean fields revealed a significant decrease in albedo in regions north of 40°N, particularly in Inner Mongolia and Jilin Province of China. A decreasing trend in albedo was also observed in southern China (Fig. 2a). Similarly, AOD showed a general decreasing trend across most of the study area (Fig. 2b). However, in southern China and regions north of 40°N, some areas exhibited a slightly increasing trend. For COT, a consistent decrease was observed across all regions during the study period, with the strongest declining trends occurring in Southern China and areas north of 40°N.

Figure 2. Trend map of Surface albedo, aerosol optical depth (AOD), and cloud optical thickness (COT) over East Asia from 2012 to 2019, (a) Surface albedo, (b) AOD, and (c) COT. The red boxes in the figures show the increasing/decreasing trend in Southern China.

Fig. 3 presents the STL analysis results for the monthly mean time series of surface albedo, AOD, and COT during the study period. Panel (a) shows the STL result for surface albedo, panel (b) for AOD, and panel (c) for COT. From January 2012 to December 2019, all three variables–albedo, AOD, and COT–exhibited an overall decreasing trend in the East Asia region.

Figure 3. STL analysis result of time series over East Asia from 2012 to 2019: (a) surface albedo,(b) aerosol optical depth (AOD), and (c) cloud optical thickness (COT) with each row showing monthly mean, trend, seasonality and residual from the top to downwards, respectively.

AOD and COT displayed clear and continuous decreasing trends throughout the entire study period. In contrast, surface albedo showed fluctuations with alternating increases and decreases depending on the year. Notably, there was a sharp increase in albedo in 2012 and 2015, followed by a steep decline in 2013. From 2016 to 2018, albedo experienced a gradual decline, but this was followed by an increase in 2019. Despite these year-toyear variations, the results confirmed an overall decreasing trend in albedo over the entire study period.

This study analyzed the long-term trends of satellite-derived products in the East Asian region to enhance understanding of the local climate. Specifically, it examined the spatiotemporal trends of surface albedo, AOD, and COT over 8 years from 2012 to 2019.

During the study period, surface albedo, AOD, and COT exhibited decreasing trends of –0.0003, –0.0008, and –0.0328 per month, respectively. To isolate pure trends while excluding seasonal variations, an STL analysis was conducted, confirming the decreasing trends for all three variables.

The spatial analysis of increasing and decreasing trends revealed that surface albedo showed significant decreases in regions north of 40°N (Inner Mongolia and Jilin Province) and southern China. AOD also displayed a decreasing trend across most areas except for regions north of 40°N and southern China. COT demonstrated a consistent decreasing trend across the entire study area.

One limitation of this study lies in the lack of quantitative analysis to determine the causes of these observed trends. Nevertheless, this research is significant as it provides the first long-term trend analysis of solar radiation-related satellite products for the East Asian region, which has been underexplored in previous studies. Future research incorporating methods such as correlation analysis or Granger causality tests could further elucidate the causal relationships behind the observed decreases in albedo, AOD, and COT. Such efforts are expected to enhance the understanding of changes in and interactions among climate variables related to solar energy reflection in the East Asian region.

This research was supported by the “Technical Development for Utilizing Meteorological Satellite Data for Climate and Environment” (KMA2020-00123) of “Technical Development on Weather Forecast Support and Convergence Service using Meteorological Satellites” project funded by the National Meteoro-logical Satellite Center, Korea Meteorological Administration.

No potential conflict of interest relevant to this article was reported.