Cloud removal is often required to construct time-series sets of optical images for environmental monitoring. In regression-based cloud removal, the selection of an appropriate regression model and the impact analysis of the input images significantly affect the prediction performance. This study evaluates the potential of Gaussian process (GP) regression for cloud removal and also analyzes the effects of cloud-free optical images and spectral bands on prediction performance. Unlike other machine learning-based regression models, GP regression provides uncertainty information and automatically optimizes hyperparameters. An experiment using Sentinel-2 multi-spectral images was conducted for cloud removal in the two agricultural regions. The prediction performance of GP regression was compared with that of random forest (RF) regression. Various combinations of input images and multi-spectral bands were considered for quantitative evaluations. The experimental results showed that using multi-temporal images with multi-spectral bands as inputs achieved the best prediction accuracy. Highly correlated adjacent multi-spectral bands and temporally correlated multi-temporal images resulted in an improved prediction accuracy. The prediction performance of GP regression was significantly improved in predicting the near-infrared band compared to that of RF regression. Estimating the distribution function of input data in GP regression could reflect the variations in the considered spectral band with a broader range. In particular, GP regression was superior to RF regression for reproducing structural patterns at both sites in terms of structural similarity. In addition, uncertainty information provided by GP regression showed a reasonable similarity to prediction errors for some sub-areas, indicating that uncertainty estimates may be used to measure the prediction result quality. These findings suggest that GP regression could be beneficial for cloud removal and optical image reconstruction. In addition, the impact analysis results of the input images provide guidelines for selecting optimal images for regression-based cloud removal.

An attempt to derive the surface temperature from the Korea Multi-purpose Satellite (KOMPSAT)-3A mid-wave infrared (MWIR) data acquired over the southern California on Nov. 14, 2015 has been made using the MODerate resolution atmospheric TRANsmission (MODTRAN) radiative transfer model. Since after the successful launch on March 25, 2015, the KOMPSAT-3A spacecraft and its two payload instruments ― the high-resolution multispectral optical sensor and the scanner infrared imaging system (SIIS) ― continue to operate properly. SIIS uses the MWIR spectral band of 3.3-5.2 μm for data acquisition. As input data for the realistic simulation of the KOMPSAT-3A SIIS imaging conditions in the MODTRAN model, we used the National Centers for Environmental Prediction (NCEP) atmospheric profiles, the KOMPSAT-3A sensor response function, the solar and line-of-sight geometry, and the University of Wisconsin emissivity database. The land cover type of the study area includes water, sand, and agricultural (vegetated) land located in the southern California. Results of surface temperature showed the reasonable geographical pattern over water, sand, and agricultural land. It is however worthwhile to note that the surface temperature pattern does not resemble the top-of-atmosphere (TOA) radiance counterpart. This is because MWIR TOA radiances consist of both shortwave (0.2-5 μm) and longwave (5-50 μm) components and the surface temperature depends solely upon the surface emitted radiance of longwave components. We found in our case that the shortwave surface reflection primarily causes the difference of geographical pattern between surface temperature and TOA radiance. Validation of the surface temperature for this study is practically difficult to perform due to the lack of ground truth data. We therefore made simple comparisons with two datasets over Salton Sea: National Aeronautics and Space Administration (NASA) Jet Propulsion Laboratory (JPL) field data and Salton Sea data. The current estimate differs with these datasets by 2.2 K and 1.4 K, respectively, though it seems not possible to quantify factors causing such differences.

In this study a tendency of abnormal sea surface temperature (SST) occurrence in the seas around South Korea is analyzed from daily SST data from satellite and 14 buoys from August 2020 to July 2021. As thresholds 28℃ and 4℃ are used to determine marine heatwaves (MHWs) and abnormal low water temperature (ALWT), respectively, because those values are adopted by the National Institute of Fisheries Science for the breaking news of abnormal temperature. In order to calculate frequency of abnormal SST occurrence spatially by using satellite SST, research area was divided into six areas of coast and three open seas. ALWT dominantly appeared over a wide area (7,745 km2) in Gyeonggi Bay for total 94 days and it was also confirmed from buoy temperature showing an occurrence number of 47 days. MHWs tended to be high in frequency in the coastal areas of Chungcheongdo and Jeollabukdo and the south coastal areas while in case of buoy temperature Jupo was the place of high frequency (32 days). This difference was supposed to be due to the low accuracy of satellite SST at the coasts. MHWs are also dominant in offshore waters around Korean Peninsula. Although detecting abnormal SST by using satellite SST has advantage of understanding occurrence from a spatial point of view, we also need to perform detection using buoys to increase detection accuracy along the coast.

Mining activity causes environmental pollution and geological hazards such as ground subsidence or landslide of which continuous monitoring is necessary. In this study, the activity on the Fushun West Open- Pit Mine (FWOPM), one of the largest open-pit coal mines in Asia located in Fushun, Liaoning Province, China, was analyzed by using a time-series Sentinel-1 InSAR coherence dataset. By using the difference between the two Digital Elevation Models (DEM) of the area, it was possible to confirm that there was a stockpiling activity in the western area of the FWOPM while excavation activity in the eastern area. By using RGB composite images using the yearly-averaged InSAR coherence images, the activity of the mine was confirmed by period, which was confirmed by Google Earth optical images. As a result, it was possible to confirm three landslides and the related activities on the northwest slope and the dumping activity on the west slope of FWOPM.

The city of Khartoum, the capital of Sudan, was heavily damaged by the flood of the Nile in 2020. Classification using satellite images can define the damaged area and help emergency response. As Synthetic Aperture Radar (SAR) uses microwave that can penetrate cloud, it is suitable to use in the flood study. In this study, Random Forest classifier, one of the supervised classification algorithms, was applied to the flood event in Khartoum with various sizes of the training dataset and number of images using Sentinel-1 SAR. To create a training dataset, we used unsupervised classification and visual inspection. Firstly, Random Forest was performed by reducing the size of each class of the training dataset, but no notable difference was found. Next, we performed Random Forest with various number of images. Accuracy became better as the number of images increased, but converged to a maximum value when the dataset covers the duration from flood to the completion of drainage.

Coastal upwelling is a significantly imperative process for understanding the interactions between physical and ecological processes and has been investigated incessantly. In this study, we explored the upwelling index, specifically upwelling age (UA). UA enabled us to observe the initiating, sustaining, and decaying upwelling processes. Although the sensitivity of many other geophysical parameters to estimate UA has been investigated, the wind direction has not been evaluated. Thus, we assessed the appropriate wind direction for the UA and obtained efficient upwelling signals from the four coastal stations. Furthermore, we applied the UA and compared it with the satellite sea level anomaly, sea surface temperature, and chlorophyll-a changes to validate how UA depicts their spatial extents. Thus, UA can predict the timing of coastal upwelling events using predicted geophysical parameters.

The purpose of this study is to evaluate the classification performance and applicability when land cover datasets constructed for AI training are cross validation to other areas. For study areas, Gyeongsang-do and Jeolla-do in South Korea were selected as cross validation areas, and training datasets were obtained from AI-Hub. The obtained datasets were applied to the U-Net algorithm, a semantic segmentation algorithm, for each region, and the accuracy was evaluated by applying them to the same and other test areas. There was a difference of about 13-15% in overall classification accuracy between the same and other areas. For rice field, fields and buildings, higher accuracy was shown in the Jeolla-do test areas. For roads, higher accuracy was shown in the Gyeongsang-do test areas. In terms of the difference in accuracy by weight, the result of applying the weights of Gyeongsang-do showed high accuracy for forests, while that of applying the weights of Jeolla-do showed high accuracy for dry fields. The result of land cover classification, it was found that there is a difference in classification performance of existing datasets depending on area. When constructing land cover map for AI training, it is expected that higher quality datasets can be constructed by reflecting the characteristics of various areas. This study is highly scalable from two perspectives. First, it is to apply satellite images to AI study and to the field of land cover. Second, it is expanded based on satellite images and it is possible to use a large scale area and difficult to access.

DPR Korea has been creating cropland across the country due to its chronic food shortage. Cropland was about 17.4% at the end of the 1980s, but it increased steadily to 19.6% at the end of the 1990s, 24.8% at the end of the first decade of 2000s, and 25.4% at the end of the 2010s. On the other hand, the forest land declined from about 74.8% in the late 1980s to 69.5% in the late 2010s. Urbanization is also progressing, increasing from about 1.15% at the end of the 1980s to 1.68% at the end of the 2010s. Most of the deforestation that occurred in DPR Korea was caused by conversion to cropland. These characteristics of land cover changes in DPR Korea provide useful information and implications for international and inter-Korean cooperation for DPR Korea.

Fine particulate matter (PM2.5) has been the biggest environmental problem in Korea since the 2010s. The present study considers the value of urban forests and green infrastructure as an ecosystem service (ES) concept for PM2.5 reduction based on satellite and spatial data, with a focus on Seoul, Korea. A method for the spatial ES assessment that considers social demand variables such as population and land price is suggested. First, an ES assessment based on natural environment information confirms that, while the vitality of vegetation is relatively low, the ES is high in the city center and residential areas, where the concentration of PM2.5 is high. Then, the ES assessment considering social demand (i.e., the ESS) confirms the existence of higher PM2.5 values in residential areas with high population density, and in main downtown areas. This is because the ESS of urban green infrastructure is high in areas with high land prices, high population density, and above-average PM2.5 concentrations. Further, when a future green infrastructure improvement scenario that considers the urban forest management plan is applied, the area of very high ESS is increased by 74% when the vegetation greenness of the green infrastructure in the residential area is increased by only 20%. This result suggests that green infrastructure and urban forests in the residential area should be continuously expanded and managed in order to maximize the PM2.5 reduction ES.

Identifying ship types is an important process to prevent illegal activities on territorial waters and assess marine traffic of Vessel Traffic Services Officer (VTSO). However, the Terrestrial Automatic Identification System (T-AIS) collected at the ground station has over 50% of vessels that do not contain the ship type information. Therefore, this study proposes a method of identifying ship types through the Random Forest Classifier (RFC) from dynamic and static data of AIS and V-Pass for one year and the Ulsan waters. With the hypothesis that six features, the speed, course, length, breadth, time, and location, enable to estimate of the ship type, four classification models were generated depending on length or breadth information since 81.9% of ships fully contain the two information. The accuracy were average 96.4% and 77.4% in the presence and absence of size information. The result shows that the proposed method is adaptable to identifying ship types.