Ocean eddies play an important role in transferring and releasing thermal energy and momentum through horizontal and vertical advections. In order to understand their motions, the spatial boundaries are important to determine. Thus, this study validated the spatial boundaries of intrathermocline Ulleung Warm Eddy (UWE) as a case study. The surface spatial boundary was determined based on the Lagrangian Particle Tracking (LPT) method, which was evaluated based on subsurface thermal profile measurements collected by the National Institute of Fisheries Science (NIFS). We found that first, the surface spatial boundary was well determined by LPT when compared with the subsurface eddy shape determined by the isothermal 10°C. The mean and standard deviation of the differences between surface and subsurface-based UWE boundaries are 7.70?4.71 km. Second, when there are some non-uniform thermal structures in the upper layer, it is difficult to determine the spatial boundary based on LPT. Third, potential capabilities to validate UWE boundary based on Surface Water Ocean Topography (SWOT) were examined. Since the Sea Surface Height Anomaly (SSHA) Level 4 data and SWOT are not comparable yet due to being non-fully calibrated up to now, it is difficult to confirm the spatial boundary of UWE using SWOT.?Overall, this study can suggest how the altimetry to detect ocean eddies using LPT may need subsurface thermal structures to determine eddy shape.

There are approximately 17,080 agricultural reservoirs distributed across South Korea, and maintaining these reservoirs requires significant costs and time. Accordingly, time series tracking of water surface area changes using satellite imagery has been proposed as a more efficient and economical method for managing reservoirs. This study analyzes reservoir surface areas using Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 Multispectral Instrument (MSI) imagery and assesses the applicability of satellite data for reservoir maintenance. Sentinel-1 SAR analysis results show that except for Maengdong (0.58) and Gui (0.62), the Split and Merge technique demonstrated a correlation above 0.7. The Region Growing technique showed correlations above 0.7 for all reservoirs, while the threshold-based method maintained correlations above 0.7 for most reservoirs, except for Gui (0.68). Sentinel-2 MSI imagery was analyzed using Modified Normalized Difference Water Index (MNDWI), NDWI, and multi-band/multi-temporal NDWI approaches. MNDWI showed correlations above 0.7 only in select reservoirs, such as Maengdong, while NDWI demonstrated correlations above 0.7 for most reservoirs, except Gui (0.57) and Seongju (0.59). The multi-band/multi-temporal NDWI method exhibited correlations above 0.7 for all reservoirs except Seongju (0.52). This demonstrates the feasibility of monitoring reservoir surface area changes using satellite data and suggests its potential as a tool for supporting priority decisions in reservoir maintenance.

Change detection is essential for applications such as urban planning, environmental monitoring, and disaster response. Despite advancements in high-resolution satellite imagery, accurate change detection remains challenging due to increased landscape heterogeneity and variable atmospheric conditions. The Mamba model, an efficient state-space model-based architecture, has shown promise in capturing spatiotemporal relationships in high-resolution datasets, addressing the limitations of traditional methods that struggle with the diverse appearances of urban structures. This research investigates applying Mamba to multitemporal Korea Multi-Purpose Satellite (KOMPSAT) imagery, using both real and synthetic data from SyntheWorld, a dataset developed to simulate various change scenarios. This study introduces a synthetic data-augmented mamba-based change detection algorithm (SAMBA), designed to detect structural changes in urban environments using KOMPSAT-3A satellite imagery. The main objectives are to evaluate the Mamba binary change detection (MambaBCD) model’s ability to detect building changes in KOMPSAT-3A images and assess the impact of synthetic data augmentation on performance. Experimental results with MambaBCD-Small and MambaBCD-Tiny models indicate that synthetic data incorporation improves generalization in complex settings, achieving high performance across multiple data and model configurations. Notably, the MambaBCD-Tiny model, with or without synthetic augmentation, outperformed the larger-parameter MambaBCD-Small model, demonstrating enhanced sensitivity in detecting satellite image changes. Performance evaluation metrics yielded an overall accuracy of 99.73%, precision of 98.34%, recall of 96.54%, F1-score of 97.43%, intersection over union of 95.00%, and Kappa coefficient of 97.29%. These metrics were similarly used to test the SAMBA algorithm’s generalization on benchmark change detection datasets, showcasing its potential as a robust tool for highresolution image change detection.

The intensification of heatwaves and urban heat island effects due to recent climate change has raised concerns about public health and economic losses. This study quantitatively evaluated heat vulnerability using the Heat Vulnerability Index (HVI) for Busan and Daegu, two metropolitan cities in South Korea. ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) data were utilized to derive heat exposure variables that reflect the spatiotemporal variability of the thermal environment. These were combined with sensitivity and adaptive capacity variables, extracted through Principal Component Analysis (PCA) of socio-environmental indicator data, to calculate the HVI. The results identified Seomyeon Station in Busan and Dongseongro in Daegu as the most heat-vulnerable areas, driven by high heat exposure, a significant proportion of vulnerable populations, and low adaptive capacity. The derived HVI map demonstrated a strong correlation with heat-related illness statistics, validating its effectiveness in assessing heat vulnerability. This study provides a foundational dataset for establishing region-specific strategies to mitigate the impacts of heatwaves by identifying and characterizing highly vulnerable areas.

Crowdsourced drone LiDAR data, collected from diverse sensors and environments, often suffer from inconsistent data quality. Outliers in such datasets can distort the characteristics of terrain and structures, leading to inaccurate decision-making. This study aims to validate the quality of crowdsourced LiDAR data by developing a precise outlier detection method based on semantic segmentation. Using the open-source Semantic Terrain Points Labeling Synthetic 3D (STPLS3D) dataset, noise-augmented training data were generated through simulations. Subsequently, a Kernel Point Convolution (KPConv) model was trained. The trained model was applied to real-world crowdsourced LiDAR data and compared with existing methods, including Statistical Outlier Removal (SOR) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN). Experimental results demonstrated that the KPConv model outperformed SOR and DBSCAN in terms of accuracy and reliability, effectively identifying inherent noise in the original data. These findings highlight the utility of deep learning-based outlier detection methods for ensuring the quality of crowdsourced LiDAR data.

Monitoring crop growth changes and water content is crucial in the agricultural sector. This study utilized drones equipped with Short Wavelength Infrared (SWIR) sensors, sensitive to moisture changes, to observe soybeans’ growth and water content variations. We confirmed that as soybeans grow more vigorously, their water content increases and differences in irrigation levels lead to decreases in vegetation and moisture indices. This suggests that waterlogging slows down soybean growth and reduces water content, highlighting the importance of detailed monitoring of vegetation and moisture indices at different growth stages to enhance crop productivity and minimize damage from waterlogging. Such monitoring could also preemptively detect and prevent the adverse effects of moisture changes, such as droughts, on crop growth. By demonstrating the potential for early diagnosis of moisture stress using drone-based SWIR sensors, this research suggests improvements in the efficiency of large-scale crop management and increases in yield, contributing to agricultural production.

Current global climate change increases water-related disasters like floods, making the Korean Peninsula vulnerable to typhoons and rainfall damage. In North Korea, flood analysis relies on remote sensing due to poor infrastructure and observation systems. Synthetic aperture radar (SAR) imagery, unaffected by clouds, is effective for flood analysis. This study used Sentinel-1 SAR image differencing to assess 2023 flooding in North Korea. After differencing images from North and South Korea, random forest classification identified flooded areas. The accuracy of this method was confirmed by its kappa value of 0.9506, demonstrating its efficacy in the remote sensing of flood damage in areas with limited access, free from meteorological interference. The analysis results indicate substantial flood damage at the target sites compared with past cases. Specifically, Typhoon Khanun resulted in a flooded area of 7.8638 km², highlighting the susceptibility of this area to typhoon-related flooding. Additionally, the area flooded due to heavy rain exceeded 1 km², although it was less severe than the flood damage in South Korea during the same period, suggesting the need for further comparative research between North and South Korea. Future research could enhance the quality by combining the use of various satellite systems and more reliable classification maps.

This study was conducted to evaluate the potential of a multimodal model for land cover classification. The performance of the Clipseg multimodal model was compared with two unimodal models including Convolutional Neural Network (CNN)-based Unet and Transformer-based Segformer for land cover classification. Using orthophotos of two areas (Area1 and Area2) in Wonju City, Gangwon Province, classification was performed for seven land cover categories (Forest, Cropland, Grassland, Wetland, Settlement, Bare Land, and Forestry-managed Land). The results showed that the Clipseg model demonstrated the highest generalization performance in new environments, achieving the highest accuracy among the three models with an Overall Accuracy of 83.9% and Kappa of 0.72 in the test area (Area2). It performed particularly well in classifying Forest (F1-Score 94.7%), Cropland (78.0%), and Settlement (78.4%). While Unet and Segformer models showed high accuracy in the training area (Area1), they exhibited limitations in generalization ability with accuracy decreases of 29% and 20% respectively in the test area. The Clipseg model required the most parameters (approximately 150 million) and the longest training time (10 hours 48 minutes) but showed stable performance in new environments. In contrast, Segformer achieved considerable accuracy with the least parameters (about 16 million) and the shortest training time (3 hours 21 minutes), demonstrating its potential for use in resource-limited environments. This study shows that image-text-based multimodal models have a high potential for land cover classification. Their superior generalization ability in new environments suggests they can be effectively applied to land cover classification in various regions. Future research could further improve classification accuracy through model structure improvements, addressing data imbalances, and additional validation in diverse environments.

An industrial heat island (IHI) refers to the phenomenon in which industrial complexes exhibit higher temperatures than their surrounding areas. It is classified as a subtype of the urban heat island (UHI). Most IHI studies have focused on the local scale (0.1–10 km), limiting their integration with broader UHI research. Additionally, the factors that influence UHI within industrial complexes remain understudied. To address these gaps, this study proposes a geospatial framework to analyze IHI, strengthen its connection with UHI research, and identify key industrial factors. The primary dataset for this framework is land surface temperature (LST) obtained from Landsat-8 imagery. First, a hierarchical approach examines IHI at the mesoscale (10–100 km), using a chi-square test to determine its phenomenological independence at the city level. If IHI is present at an adequate scale, the LST profile method is applied to measure IHI extent and intensity at the local scale. Second, geographically weighted regression (GWR) quantifies the influence of industrial factors, including nitrogen dioxide (NO₂), sulfur dioxide (SO₂), digital elevation model (DEM), normalized difference built-up index (NDBI), soil-adjusted vegetation index (SAVI), automated water extraction index (AWEI), and workers. To validate the feasibility of this framework, it was applied to Incheon, South Korea, a city with diverse and aging industrial complexes. As a result, mesoscale analysis confirmed a significant association (p<0.05) between industrial complexes and UHI across all seasons. The local scale analysis indicated that IHI intensity was highest in summer but weakened in fall and winter, diverging from conventional UHI patterns. In addition, GWR results demonstrated varying impacts of industrial factors across complexes. The most influential variable was the industrial activity factor (SO₂ and NO₂). SO₂ exhibited a strong positive correlation with IHI, while NO₂ exhibited a negative correlation. For the industrial space factor, DEM indicated that lower elevations corresponded to higher IHI intensity. SAVI exhibited a moderate negative correlation, but its influence varied depending on vegetation type. However, NDBI and AWEI produced results that contradicted the trends identified in numerous UHI studies, likely due to coastal influences diminishing NDBI's effect and AWEI reflecting intensified IHI due to industrial wastewaterinduced water quality degradation. In the industrial workforce factor, workers also showed a strong positive correlation. A comparison of GWR and OLS models confirmed GWR's superior performance, with higher adjusted R² (0.9126 vs. 0.3677) and lower AICc values (111,146.4 vs. 171,605.4). In conclusion, this study establishes a scalable geospatial framework for IHI analysis, reinforcing its connection with UHI research. The findings underscore the need for tailored environmental measures addressing the unique characteristics of each industrial complex. Future studies should refine the framework by quantitatively connecting the results between the two scales and incorporating simulation models.

Recently, as the frequency of high-resolution satellite image collection and the amount of data have increased, the demand for image data with temporal and spatial accuracy has been rising in various fields. This image data serves an important role in various applications such as environmental monitoring, urban planning, and disaster management. However, geometric correction using ground control points for individual images is inefficient in terms of time and cost, and its practicality decreases, especially when processing large numbers of images. In this study, we propose a method to efficiently re-estimate the rational polynomial coefficients (RPCs) correction coefficients for a large number of uncorrected KOMPSAT- 3/3A images by setting a specific orthorectified CAS500-1 as the reference image and using bundle block adjustment. The proposed method applies the inverse geocoding technique to the set reference image to reproduce the image corresponding to the Level-2 Radiometric (L2R) and the RPCs information. Afterward, bundle block adjustment is performed with other Level-1 Radiometric (L1R) or L2R images to re-estimate the RPCs correction coefficients in bulk. This process improves the geometric correction accuracy for a large number of images while also saving time compared to the method of correcting each image independently. As a result of experiments using the proposed methodology, the initial relative error position error was reduced from 160 pixels to 1.3 pixels. This demonstrated significant improvement in accuracy and efficiency performance. Through the proposed method, high-accuracy single satellite images were used to perform precise corrections on multiple images. Moreover, the feasibility of bundle adjustment processing using various satellite image data was confirmed. As a result, it is expected that a large volume of satellite images processed quickly and accurately will be provided for various satellite image application fields.