This study developed a landslide susceptibility map (LSM) for the Angren area of Uzbekistan by comparing a frequency ratio (FR) modelwith a convolutional neural network optimized using the grey wolf optimizer (CNN-GWO). The study area features steep mountain slopes characterized by Neogene sedimentary lithology, overlain by Quaternary loess deposits, and is affected by intensive mining activities, making it highly susceptible to seasonal landslides. The susceptibility mapping was performed by integrating landslide inventorywith comprehensive topographic parameters,soil characteristics, land-use data, normalized difference vegetation index (NDVI), and geological information. The FR model assessed the statistical correlation between landslide occurrences and individual influencing factors, whereas the CNN-GWO model integrated allrasterlayersto identify complex spatial patternsin landslide susceptibility. The results show that the susceptibility maps generated by the FR and CNN-GWO models display comparable spatial patterns, with area under the curve values of approximately 78.11% and 80.11%, respectively. These findings provide a useful framework for spatial planning, infrastructure protection, and landslide risk reduction in the mountainous Angren area.

Unmanned aerial vehicles(UAVs) enable micro-scale thermal mapping in urban environments, yet thermal imagery records brightness temperature (BT) that is biased by land-surface emissivity (ε) and atmospheric effects. This study develops a single-channel (SC) land surface temperature (LST) retrieval framework that integrates UAV-based thermal and multispectral imagery (MicaSenseAltum-PT)with insitu observations. Pixel-wise ε was estimated using the empirical normalized difference vegetation index (NDVI)-ε relationship developed for the ASTER spectral response function (SRF), applied to NDVI derived from the Altum-PT RED (668 nm) and near-infrared (842 nm) bands. Background temperature (T_bkg), corresponding to the reflected apparent temperature, was measured using a crumpled aluminum foil temporarily placed within the imaging footprint, which served as a near-perfect reflector to capture ambient longwave radiation, while air temperature (T_a) from on-site loggers; atmospheric transmittance (τ) is fixed under clear-sky conditions. The workflow is demonstrated over the Namdong Industrial Complex (Incheon, Republic of Korea) at 11:00, 13:00, and 15:00 (UTC+9). We quantify the correction magnitude as ΔT=BT-LST within a LiDAR-derived walkable-surface mask and further stratify the results by land cover (Natural vs. Artificial pavements). ΔT is predominantly negative (median: -1.85, -0.95, -0.73 K), indicating systematic BT underestimation relative to LST. Artificial pavements exhibit larger negative ΔT and greater variability than natural cover, reflecting material-dependent emissivity and surface-condition heterogeneity. The magnitude of |ΔT| generally decreases from morning to afternoon. The results support the use of SC-retrieved LST rather than raw BT for threshold-based assessments(e.g., heat-risk mapping) in pedestrian contexts. Although validated on a single date and site without absolute blackbody references, the framework shows consistent behavior across times and land-cover types, highlighting its utility for urban micro-climate applications.

This study investigates the variation characteristics of the mass extinction efficiency (MEE) of particulate matter (PM; PM₁₀, PM_2.5-10, and PM_2.5) in the Busan port (Busan, South Korea) area using long-term Scanning Light Detection and Ranging (LiDAR) and in situ measurement data. The results show that the mean MEE of PM_2.5 (7.72 ± 3.86 m² g^-1) is significantly higher than that of PM₁₀ (5.19 ± 1.99 m² g^-1) and PM_2.5-10 (1.95 ± 1.26 m² g^-1), confirming the dominant optical role of fine particles. MEE exhibits a decreasing trend with increasing PM_2.5 mass concentration due to particle growth and aggregation under high-loading conditions, while it increases with a higher PM_2.5/PM₁₀ ratio, indicating the importance of particle composition. The combined analysis of MEE and relative humidity (RH) reveals a nonlinear interaction: MEE increases under moderate RH (60-70%) due to hygroscopic growth but decreases at high concentration or low RH (< 40%) as particle agglomeration dominates. These results demonstrate that MEE is controlled by the coupled effects of concentration, composition, and humidity rather than by a single variable. The findings provide a physical basis for developing region-specific correction factors for converting LiDAR extinction coefficients to PM mass concentrations in coastal and port environments.

Urbanization increases impervious surfaces and intensifies the surface urban heat island (SUHI). Satellite-based land surface temperature (LST) is widely used to assess SUHI, but limited high-resolution nighttime observations have led many studies to focus on daytime conditions. In this study, using 70-m ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) LST, we quantify seasonal (summer vs. winter) and diurnal (day vs. night) variability in surface urban heat island intensity (SUHII) in Suwon, South Korea, based on ten clear-sky acquisitions (2022-2025). A two-way repeated measures ANOVA confirms that SUHII is significantly stronger and more spatially heterogeneous in summer than in winter, and during the day than at night. To explain these time-dependent SUHII patterns in relation to urban form while minimizing scale-mismatch information loss, we aggregate 30-m local climate zone (LCZ) compositions within each 70-m SUHII pixel and estimate time-specific effects via compositional regression. Natural LCZs are generally associated with cooling, with tree-related classes showing cooling that varies with canopy density.Water(LCZ G) exhibits a pronounced day-night contrast, shifting from daytime cooling to nighttime warming. Built-up LCZs increase SUHII, with high-density development (LCZ 1-3) consistently producing warming across all times. Vertical structure effects are time-dependent, with stronger warming in low-rise areas during the day and in high-rise areas at night. This suggests that shading and solar exposure dominate daytime surface heating in low-rise buildings, whereas nocturnal warming in high-rise areas is more strongly shaped by radiative trapping and heat storage with in urban canyons. These results quantify time-dependent patterns in SUHII and LCZ-specific heat responses, supporting a basis for identifying season- and time-specific heat-vulnerable areas and tailoring mitigation by urban form.

Estimating photosynthetically active radiation (PAR) is crucial for quantifying ecosystem productivity through photosynthesis.In particular, providing continuous estimates of direct PAR (PAR_dif) and diffuse PAR (PAR_dif) is important for assessing ecosystem photosynthesis. The Breathing Earth System Simulator (BESS) has proven effective in estimating global PARdir and PAR_dif using an artificial neural network (ANN)that incorporates continuous satellite-based atmosphere products and elevation. However, BESS PAR does not account for topographic correction due to its relatively coarse resolution. The Perez-Driesse model is commonly used for topographic correction by partitioning incoming PAR into five components: PARdir, anisotropic PAR_dif, isotropic PAR_dif, horizontal brightness PAR_dif, and ground reflected PAR_dif. Among these, isotropic PAR_dif and ground reflected PAR_dif are strongly influenced by the surrounding visible sky and terrain fraction. However, the Perez-Driesse model oversimplifies the sky view factor(SVF) and ground-reflected PAR_dif calculations by considering only the slope at a given point. To address this issue, we leveraged horizon angles from a digital elevation model (DEM) by scanning azimuth directions from 0° to 360° at 3° intervals to more accurately estimate the SVF for isotropic PAR_dif, and we accounted for adjacent terrain slopes and distances for ground-reflected PAR_dif. We applied this improved SVF Perez-Driesse model to BESS PAR estimates, which combine 2 km-resolution atmospheric products from the GEO-KOMPSAT-2A satellite with the 30 m-resolution Copernicus DEM. Validation with LI-190R quantum sensor data from rooftop observations demonstrated significant improvements after applying topographic correction. For southward orientations, the coefficient of determination (R²) improved from 0.52 to 0.78,relative bias (rbias) from 28.17% to -0.78%, mean absolute error(MAE) from 429.3 to 299.4, and root mean squared error (RMSE) from 553.3 to 378.7. For westward orientations, R² increased from -0.36 to 0.55, rbias from -44.63% to 23.80%, MAE from 477.3 to 243.7, and RMSE from 566.3 to 340.0. Analysis of two study sites with different slopes on the Korean Peninsula in 2021 showed that seasonal variations in shadow ratios reduced PARdir by 40-80% under low solar positions. The rbias of PAR_dif was influenced by increased cloud cover and aerosols during summer. The effect of terrain aspect was pronounced in instantaneous topographic corrections, particularly due to differences between aspect and solar azimuth angle, whereas at the annual scale, north-south aspect differences were more evident. Our results showed that Perez-Driesse-based topographically corrected PAR can vary by up to 40% in rugged terrain, whereas SVF Perez-Driesse-based PAR varies by nearly 20%, leading the significant variation in photosynthesis estimation.

This study investigates spatiotemporal variations in column-average dry-air mole fractions of methane (XCH₄) over the Korean Peninsula from 2019 to 2024 using TROPOspheric Monitoring Instrument (TROPOMI)satellite observations. To assess data accuracy,satellite retrievals were compared with ground-based Total Carbon Column Observing Network (TCCON) measurements at Anmyeondo, Korea, and Saga, Japan. Despite limited data availability at Anmyeondo (N = 7) due to harsh coastal conditions, the validation showed a high correlation (R > 0.88) and consistent seasonal variations at both sites, confirming the accuracy of TROPOMI retrievals. The results revealed a distinct "west-high, east-low" spatial pattern of XCH₄ over the Korean Peninsula. This distribution reflects the concentration of major anthropogenic sources in western regions, including the densely populated Seoul Metropolitan Area, large-scale industrial complexes, and coal-fired power plants, showing strong qualitative and quantitative agreement with the Emissions Database for Global Atmospheric Research (EDGAR) 2025 emission inventory. In general, the mean XCH₄ concentration increased continuously across the Korean Peninsula, except for a temporary stagnation in 2022. Notably, a synchronized rapid surge was observed across all major regions in 2024, peaking at 1912.58 ppb. Regional analysis indicated that, while local industrial emissions (e.g., in Gyeongnam) drove regional heterogeneity until 2023, the sharp increase in 2024 was likely driven by strong external factors, such as the global rise in background methane levels or large-scale meteorological influences. These findings provide important insights into regional variations in methane levels and emissions across the Korean Peninsula. These results demonstrate that TROPOMI not only effectively captures anthropogenic emission patterns aligned with EDGAR but also identifies potential natural sources unique to the Korean Peninsula's environmental characteristics. By bridging the gap between existing emission inventories and real-world observations, our findings highlight the critical role of satellite-based monitoring in establishing comprehensive, data-driven strategies for regional methane management and achieving national climate goals.

Advances in Earth observation technologies have enabled mid-sized satellites to achieve high radiometric performance at relatively low cost, expanding their use to quantitative remote sensing applications. Compact Advanced Satellite 500-4 (CAS500-4), equipped with visible-to-near-infrared sensors including a red-edge band and designed to provide nearly 5-meter multispectral imagery every 3 days, requires a reliable radiometric monitoring framework. Cross-sensor comparison offers an efficient means of validating radiometric stability, particularly through reference sensors such as LANDSAT-8 Operational Land Imager(OLI),whose long-term calibration record ensures high radiometric consistency. This study evaluates whether forest-dominant regions on the Korean Peninsula, which represent the primary observation target of CAS500-4, can serve as suitable ground targets for cross-validation between LANDSAT-8 radiance and KOMPSAT-3 multispectral digital number (DN) values. Five KOMPSAT-3 scenes acquired near-nadir on 13 January 2025 were paired with the same-day Landsat-8 imagery. To mitigate geometric and spatial-resolution mismatches, zonal statistics and a pixel-based sliding-window sampling method (1,681 windows) were applied. Results showed that zonal-mean DN values from KOMPSAT-3 were highly stable across sliding windows, with standard deviations generally below 2 DN. Derived radiance conversion coefficients in the visible bands differed by less than 6% from previous KOMPSAT-3 radiometric studies, demonstrating the strong feasibility of the proposed method. Larger discrepancies in the near-infrared band were attributed to water-dominant landcover and nonlinear sensor response at low DN levels. Overall, forest regions exhibited stable reflectance characteristics suitable for radiometric comparison, suggesting that Korean forests provide a practical and effective foundation for CAS500-4 radiometric validation after launch.

Accurate rainfall information is fundamental for agricultural planning, climate risk management, and food security in rainfed systems, particularly in regions with sparse ground-based observations. Limited meteorological stations in southeastern Nigeria hamper access to reliable, spatially representative rainfall observations, necessitating the use of satellite and reanalysis precipitation products. However, the performance of these datasets varies across climatic and agro-ecological settings, requiring local-scale validation prior to application. In this study, we assess the accuracy and agricultural relevance of eleven widely used rainfall products over Southeastern Nigeria using long-term gauge observations from the Nigerian Meteorological Agency (NiMet) as reference data. We apply a comprehensive evaluation framework that combines statistical performance metrics, Taylor diagrams, analyses of seasonal and interannual rainfall variability, Standardized Precipitation Index (SPI)-based drought assessment, and trend detection using the Mann-Kendall test and Sen's slope estimator. Five satellite-based rainfall estimates-Global Precipitation Measurement Integrated Multi-satellite Retrievals for GPM [GPM-IMERG], Climate Hazards Group InfraRed Precipitation with Station Data [CHIRPS], Climate Prediction Centre Morphing Technique [CMORPH], Tropical Rainfall Measuring Mission Multi-satellite Precipitation Analysis [TRMM/TMPA], and Centre for Hydrometeorology and Remote Sensing product [CHRS]-four reanalysis products-ECMWF Reanalysis v5 [ERA5], ERA5-Land, Agrometeorological indicators from ERA5 [AgERA5], and Modern-Era Retrospective Analysis for Research and Applications v2 [MERRA-2]-and two gridded gauge datasets-Global Precipitation Climatology Centre [GPCC] and Climatic Research Unit Time Series [CRU]-were assessed over six meteorological stations in southeastern Nigeria from 2000 to 2025. The results reveal that ERA5-Land, CHIRPS, GPCC, ERA5, GPM-IMERG, and TRMM align closely with NiMet observations, effectively capturing rainfall intensity, variability, and seasonal patterns with strong correlations (r > 0.8) and minimal bias. These high-performing datasets also accurately represent drought and wet-year indicators, highlighting their suitability for agricultural risk assessment and climate-informed decision-making. Conversely, AgERA5 and MERRA-2 exhibit systematic biases, inflated variability, and weaker correlations, limiting their reliability for agricultural and climate applications in the region. Trend analyses reveal weak and largely insignificant long-term rainfall trends, suggesting that agricultural vulnerability in southeastern Nigeria is driven more by interannual variability and intra-seasonal extremes than by monotonic rainfall changes. Interestingly, this study provides a clear, evidence-based ranking of rainfall products relevant to Southeastern Nigeria and highlights the datasets best suited for climate-smart agriculture, drought monitoring, and hydroclimatic evaluation in data-scarce environments.

Mangroves are among the most effective vegetation types for carbon sequestration, storing carbon absorbed through photosynthesis in stems, leaves, roots, and substrates. Conventional methods for estimating Above Ground Carbon(AGC), such as destructive sampling, can damage mangrove ecosystems and are generally inefficient. Remote sensing offers a more efficient and less invasive alternative, enabling improved monitoring in terms of time, labor, and cost. This study maps and monitors mangrove AGC in the Perancak Estuary, Bali,Indonesia, using PlanetScope imagery. Field-based carbon measurements were conducted using the Diameter at Breast Height (DBH) method, which was converted to biomass using species-specific allometric equations and then to AGC using a conversion factor of 47% of total biomass. Three vegetation indices were evaluated: the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Red Edge Index (NDRE), and the Green Normalized Difference Vegetation Index (GNDVI). Model performance was assessed using the coefficient of determination (R²) and Root Mean Square Error (RMSE). NDRE yielded the best performance, with an R² of 0.341 and an RMSE of 0.039. The estimated AGC at the study site in 2025, derived from PlanetScope imagery using the NDRE index, ranges from 8.04 to 78.60 Mg C ha^-1. Multitemporal analysis shows that AGC remains relatively stable over time but increases significantly in 2025, indicating potential mangrove growth and enhanced carbon storage capacity.

In recent years, the urgency of flood management in West Java has increased due to the high frequency and intensity of flooding, especially in lowland areas connected to river basins, such as agricultural and urban areas. A comprehensive analysis of flood triggering factors is necessary to predict flood-prone areas more accurately. Thisstudy aimsto predict flood risk inWestJava by combining remote sensing methods, geospatial analysis, and machine learning approaches. A total of 114 historical flood locations that inundated several areas in West Java were identified. Flood inventory data was separated into 70%astraining data and 30%astesting data. Flood-triggering factors, including topography, hydrology, environment, and climatology, were collected and selected based on a multicollinearity assessment and a frequency ratio (FR) analysis. Machine learning models using a convolutional neural network (CNN) and long short-term memory (LSTM) algorithms were trained with flood locations as the dependent variable and triggering factors asthe independent variables. The resultsshowed that the cities of Bandung, Bekasi, Sukabumi, and Cianjur are at high to very high susceptibility levels. The model's prediction performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC) analysis, where CNN with an AUC value of 0.720 wassuperiorin conducting predictions compared to LSTM with an AUC value of 0.706. This research aligns with government policy and the Sustainable Development Goals and can be used as an effort to ensure public safety from flood disasters.

This study aims to establish a pre-launch validation framework for the Vegetation Index (VI) products of the Compact Advanced Satellite 500-4 (CAS500-4), scheduled for launch in 2026. To this end, Sentinel-2, internationally recognized for its high quality, was adopted as proxy satellite data, and a ground observation network covering South Korea's forest environments was established. All evaluation processes adhered to the Land Product Validation (LPV) protocol of the Committee on Earth Observation Satellites (CEOS) and the Quality Assurance Framework for Earth Observation (QA4EO), ensuring consistent objectivity and transparency in validation. Initial validation results using multispectral sensors from the established ground network (8 sites) showed high correlations (R² > 0.8) between Sentinel-2 Normalized Difference Vegetation Index (NDVI) and ground measurements in deciduous broadleaf and mixed forests, whereas notably lower correlations were observed in evergreen coniferous forests. These findings confirmed the operational readiness of the validation system. Furthermore, given that the current eight sites are insufficient to represent the geographical heterogeneity of forests nationwide, the necessity of expanding the network to an optimal 31 sites using geostatistical techniques was proposed. The standardized framework and the multi-platform (Tower- Unmanned Aerial Vehicle [UAV]-Satellite) triple-collocation strategy established in this study will serve as a key foundation for securing the reliability and maximizing the utility of future CAS500-4 data.