Recently, water-related disasters such as floods and droughts are increasingly being impacted by climate change, characterized by growing uncertainty, interaction, and complexity. As a result, the scale of damage caused by these disasters continues to expand, with the impacts becoming more widespread and severe. Despite this, there is a lack of robust monitoring systems capable of responding to extensive disaster scenarios, making it difficult to swiftly and accurately assess disaster situations. Globally, the frequency and magnitude of floods and droughts are steadily rising, emphasizing the need for disaster management technologies that can ensure comprehensive observation, accurate forecasting, and prompt response capabilities.

Internationally, to address the challenges posed by climate change and water crises, continuous observation of water resources and water-related disasters is being conducted through the use of satellites, with countries developing and utilizing satellite infrastructure necessary to secure their water security. In South Korea, while satellites for oceanic, meteorological, and geographic observations have been developed and are in operation, these satellites were primarily developed for purposes such as meteorological observation, marine environment monitoring, communication, broadcasting services, and geographic information acquisition. Consequently, their application in water resource management faces practical challenges. Additionally, the demand for diverse and specialized applications in the water resource sector continues to grow, necessitating the continuous development of infrastructure and technology to efficiently manage water resources using satellites, monitor and respond to floods and droughts, and support the expansion of the water industry overseas.

This paper aims to analyze the utilization of remote sensing, particularly satellite technology, for water-related disaster management in South Korea and, based on this analysis, propose future development directions. Remote sensing has a crucial role in providing rapid and accurate data for the prevention and management of water-related disasters. This study focuses on how satellite remote sensing technology is being employed to address the increasingly frequent water disasters, particularly floods and droughts, driven by climate change and urbanization, and provides recommendations for effective disaster management and damage mitigation.

In developed countries, satellite-based information on hydrological phenomena is being acquired, and satellite imagery is directly integrated with modeling to be used for water resource management. Satellite technology enables the acquisition of global observation data, which is utilized to provide high-precision imagery for water resource environments, meteorological research, and disaster response. Furthermore, both geostationary and low-Earth orbit meteorological satellites are being operated complementarily to improve forecast accuracy and enhance climate change monitoring capabilities, with water-related satellites being developed and operated for these purposes.

In South Korea, the utilization of satellite-based information in the water-related field shows mixed results. The Korea Meteorological Administration (KMA) observes precipitation using the Communication, Ocean and Meteorological Satellite (COMS). However, due to the observational limitations of the infrared sensors onboard, the accuracy is relatively low, and the data is primarily used for hazardous weather monitoring and ultra-short-term forecasting. Automatic Weather Stations are used for accumulating rainfall data over South Korea and correcting deviations in radar rainfall estimates. The Korea Aerospace Research Institute takes satellite imagery of the Korean Peninsula and global regions using the Korea Multi-Purpose Satellite (KOMPSAT)-2, 3 satellites. The KOMPSAT-5, launched in August 2013, is equipped with X-band SAR imagery, enabling observation of water resources and flood areas without being affected by cloud cover. However, the lack of adequate data acquisition and system development limits its application in water-related agencies. The Geostationary Ocean Color Imager (GOCI) sensor, operated by the Marine Satellite Center, is capable of being used for drought monitoring through the calculation of vegetation indices in the Korean Peninsula, but calibration and correction technologies for this purpose are still under development.

In advanced countries with satellite capabilities, research is underway to apply data assimilation technologies using satellite data, enabling optimal integration of observation data and hydrological phenomenon simulations. The United States, Japan, and the Europe Union (EU) are at the forefront of hydrological research as understanding of meteorological and hydrological phenomena improves through the use of ground observation networks and remote sensing networks. National Aeronautics and Space Administration (NASA) launched the Soil Moisture Active-Passive (SMAP) satellite in 2015, which uses the Marshall Space Flight Center algorithm to acquire accurate information on surface water reflectance and flow models, using soil moisture absorption data to predict water availability. This satellite data is utilized for weather forecasting, flood and drought prediction, agricultural productivity enhancement, and climate change forecasting. For flood monitoring, efforts are being made to increase measurement accuracy through satellite technology, radar technology, satellite data processing, and algorithm development. Research is also being conducted to improve precipitation accuracy by utilizing multiple satellites including Global Precipitation Measurement (GPM), Defense Meteorological Satellite Program, TERRA, AQUA, and other geostationary satellites to calculate rainfall using the TRMM Multi-satellite Precipitation Analysis. Additionally, NASA, in collaboration with France, is utilizing the Surface Water & Ocean Topography satellite to monitor water levels in rivers, reservoirs, lakes, and wetlands.

In Japan, rainfall observation data obtained through satellites such as Tropical Rainfall Measuring Mission (TRMM) and GPM is integrated and analyzed with land observation data collected through Moderate Resolution Imaging Spectroradiometer (MODIS) and Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E), and this information is provided for public use. The Japan Aerospace Exploration Agency (JAXA) has a key role in the disaster management system utilizing space technology by supporting the establishment of Sentinel Asia, an international organization focused on disaster management in the Asia-Pacific region, where monitoring data is serviced via a dedicated website. Additionally, JAXA has been actively involved in the development and operation of the TRMM satellite in collaboration with NASA and has worked on the development of dual-frequency precipitation radar and joint algorithm research for the GPM satellite. To replace the decommissioned AQUA satellite, JAXA has launched the Global Change Observation Mission Water 1 satellite.

In Europe, the Joint Research Center has established a global flood monitoring system using AMSR-E satellite imagery, which provides real-time information on flood occurrences. Germany’s RapidEye satellite, specializing in agricultural applications, can capture images of the entire Korean Peninsula within two weeks and consists of five satellites, each with a 70 km swath width, providing satellite imagery suitable for large-scale water resource management. Additionally, Europe operates Sentinel-1, equipped with a SAR sensor, and Sentinel-2, equipped with a multispectral optical sensor, actively providing data for research related to floods and droughts.

Floods are among the most frequent and impactful natural disasters worldwide, causing significant damage to infrastructure, ecosystems, and human lives. In many countries, including South Korea, the intensity and unpredictability of floods are increasing due to climate change and urbanization, creating a demand for more advanced and efficient tools for flood monitoring, forecasting, and management. In this context, remote sensing has emerged as a crucial technology that enhances our understanding of flood events and improves response capabilities.

Remote sensing technology enables continuous and widespread observation of flood areas, providing data on key variables such as precipitation, soil moisture, water body detection, and land use changes. Through satellite imagery, radar, and other aerial platforms, remote sensing allows for large-scale monitoring of flood events, supporting the establishment of early warning systems and emergency response strategies. Additionally, remote sensing data, when integrated with hydrological and climate models, contributes to improving flood forecasting, ultimately helping to minimize the damage to communities and infrastructure caused by floods.

In South Korea, various remote sensing technologies are being employed for flood management, with SAR sensors, which are highly effective in detecting water bodies even under cloud cover, and optical sensors, which provide detailed images of flood-affected areas, being primarily used. The integration of these technologies into flood management systems enhances the ability to monitor and respond to flood risks, contributing to the reduction of damage and loss of life caused by disasters. This chapter will review the current applications of remote sensing in flood management in South Korea and analyze key case studies to address the remaining challenges in optimizing remote sensing tools for technological advancement and effective flood management. The two most widely used remote sensing technologies in flood management are Electro-Optical satellites (EO) and Synthetic Aperture Radar (SAR) satellites. These two technologies possess complementary characteristics, making them essential for effective flood monitoring and management.

Optical satellites provide high-resolution imagery that allows for visual observation of terrain changes, water distribution, and flood damage in affected areas. This capability is particularly useful for clearly identifying surface changes before and after a flood, making it valuable for damage assessment and recovery planning. However, optical satellites are sensitive to weather conditions, and they cannot provide clear images in cloudy or rainy situations. To compensate for this limitation, SAR satellites are used.

SAR satellites use radar waves to scan the Earth’s surface, offering the advantage of data collection regardless of weather conditions or time of day/night. SAR satellites can clearly distinguish flood-affected areas by utilizing the differences in the reflective properties of water and land surfaces, making it possible to observe even in cloudy conditions or at night. This capability is highly effective for urgent monitoring and response during flood events.

In the first section of this chapter, we analyze case studies of flood-related research based on optical satellites that have been actively conducted since the 2010s. Jung et al. (2013) simulated flood inundation areas using a traditional 1-dimensional hydraulic model and quantitatively presented the uncertainties of flood inundation simulations using optical satellite imagery. A flood inundation map visually represents areas that may be flooded during a flood event and serves as crucial data for pre-flood management and response. However, due to uncertainties in hydraulic model parameters and the lack of verification data, it is challenging to ensure the accuracy of these maps. In this study, flood inundation areas were constructed using the 1D hydraulic model Hydrologic Engineering Centers-River Analysis System (HEC-RAS), and water bodies were identified using Landsat 5 Thematic Mapper (TM) optical imagery with the Iterative Self-Organizing Data Analysis technique. This dataset was used for model verification and uncertainty estimation. The Generalized Likelihood Uncertainty Estimation method was employed to quantitatively assess the uncertainties in the hydraulic model’s roughness coefficients and flow rates, and floodplain areas were presented within a 5 to 95% confidence interval.

Hwang et al. (2016) developed an automatic system for estimating flood damage based on high-resolution satellite imagery, proposing a method to directly utilize satellite images for quickly and efficiently assessing damage during disasters. Traditional flood damage assessments relied heavily on

manpower and field surveys, which were time-consuming and labor-intensive. However, the study introduced a technique to efficiently estimate flood damage using high-resolution satellite imagery and Geographic Information System technology. In this research, 1-meter resolution KOMPSAT-2 satellite imagery was used to compare pre- and post-flood images, applying the Change Vector Analysis and Differential Normalized Difference Vegetation Index algorithms to estimate and validate the flood-affected area. The comparison of flood damage in Hongcheon, Yeoju, and Gyeonggi, which occurred in July 2013, showed an accuracy of 87% when compared to manual field surveys. The study suggested that utilizing satellite imagery for flood damage assessment can save time and manpower while providing highly reliable damage estimates, thereby improving the efficiency of manpower and resource allocation in actual recovery efforts.

Piao et al. (2018) utilized low-resolution optical satellites to detect large-scale flood areas. The MODIS satellite, mounted on the Terra/Aqua satellites, offers the advantage of capturing wide areas at high frequency despite its low resolution, making it particularly suitable for large-scale flood mapping. To detect inundated areas, the study employed several spectral indices, including Normalized Difference Water Index (NDWI), Enhanced Vegetation Index (EVI), and Land Surface Water Index (LSWI), which combine various bands. The detection results were validated using Landsat-5 TM optical imagery as a reference, and the study suggested that the combination of Shortwave Infrared and red bands provided the highest accuracy. Since the launch of European Space Agency’s Sentinel-2 optical satellite in 2015, numerous flood-related studies have been conducted. In particular, due to the unique situation of South Korea being the only divided country in the world, studies have been carried out on shared rivers with North Korea, focusing on unmeasured areas where access and acquisition of ground observation data are not collectible.

Kye et al. (2021) utilized Sentinel-2 optical imagery to detect water bodies using NDWI, with a focus on North Korea’s Hwanggang Dam. Although the Hwanggang Dam, located in North Korea, affects downstream areas in South Korea, access to data is restricted, making it impossible to obtain operational or water level information. This study proposed a method to minimize the impact of clouds, a limitation of optical satellites, in detecting water bodies in unmeasured reservoirs where access is restricted. By filtering out images with a high cloud cover ratio, the study overcame the issue of underestimating reservoir surface area. Out of 220 images taken between July 2018 and October 2021, 114 images were analyzed, and it was suggested that reliable water body area measurements could be obtained when the cloud cover ratio over the reservoir was less than 10%. The study presented a practical approach for managing and monitoring water resources in areas where ground observation is difficult, utilizing satellite data.

Kim et al. (2021a) used NDWI-based water body detection data to estimate water level changes in the unmeasured Hwanggang Dam reservoir in North Korea and subsequently estimated the inflow volume to the reservoir. To simulate reservoir water level and storage changes, a high-resolution Digital Elevation Model (DEM) was used to extract the reservoir’s stage-storage curve, and water level changes were estimated using water body data acquired from satellites. The study developed a model that combined a lumped hydrological model and a reservoir operation algorithm to calculate the inflow volume of the Hwanggang Dam and indirectly validated the satellite-derived water level change data from 2017 to 2020. The results indicated that satellite-observed data could serve as an effective tool for flood protection and water resource management modeling in unmeasured areas. Additionally, Kim et al. (2021b) extended the previous study by estimating the diverted water volume from the Hwanggang Dam to another basin through an analysis of the inflow to the dam and the water balance in the downstream area. The water balance analysis from January 2019 to September 2021 revealed that approximately 922 million tons of water per year were being diverted to the Yesong River basin, accounting for 45.5% of the dam’s annual average inflow. This significant volume indicates that a substantial amount of water was not flowing downstream but was diverted to other basins. The study also qualitatively analyzed the increase in discharge during the summer of 2020 and 2021 due to heavy rainfall, which raised the potential for flood damage in downstream areas. These findings confirm that satellite imagery can provide crucial baseline data for water resource management and flood response strategies in the border regions of the Korean Peninsula.

Table 1 . Summary of research on flood using electro-optical satellites.

Reference	Subject	Satellite	Methodology	Target area
Jung et al. (2013)	Flood map	Landsat 5 TM	Unsupervised classification	Mississippi
Hwang et al. (2016)	Flood damage	KOMPSAT-2	Change detection	Yeoju
Piao et al. (2018)	Flood map	MODIS	NIR classification	Morocco
Kye et al. (2021)	Waterbody	Sentinel-2	NDWI	Hwanggang Ddam
Kim et al. (2021b)	Water level	Sentinel-2	NDWI	Hwanggang Ddam
Kim et al. (2021a)	Water level	Sentinel-2	NDWI	Hwanggang Ddam
Kim et al. (2022b)	River network	Sentinel-2	NDWI	Han river
Lee et al. (2024)	Waterbody	CAS500-1	Geo-SAM	Daegu

NIR: Near Infrared..

Additionally, research has been conducted to create river networks in South Korea using satellites. Kim et al. (2022b) developed a technology to automatically extract river networks using Sentinel-2 optical imagery, focusing on the city of Seoul. River networks are essential geographical features for river management and flood disaster prevention, serving as parameters in hydrological models. Traditionally, river networks were constructed through field surveys or terrain measurements, but satellite technology now enables a more efficient approach. In this study, NDWI was used to differentiate water bodies from land, and morphological operations such as dilation and erosion were applied to automatically remove artificial structures like bridges, creating a continuous river boundary.

In March 2021, South Korea launched the Compact Advanced Satellite 500-1 (CAS500-1), a next-generation medium-sized satellite with a 0.5 m resolution optical payload, which is being used for mapping and land development in the country. Lee et al. (2024) conducted a study using the Segment Anything Model (SAM) technique to detect and extract water bodies from CAS500-1 satellite imagery. The accuracy of water body detection, when compared with ground data, achieved an evaluation metric score of 0.749, demonstrating the potential for automated water body detection across the Korean Peninsula through long-term observation. In the second section of this chapter, we analyze flood-related research case studies based on SAR imagery. Unlike optical imagery, SAR sensors have the ability to penetrate clouds due to the nature of microwaves, allowing for the observation of ground conditions regardless of weather conditions.

Seo et al. (2018) conducted a study to estimate the flow rate of small- and medium-sized rivers using Sentinel-1 SAR imagery. Due to the lack of field observation data for small- and medium-sized rivers, satellite-based flow estimation methods have emerged as a viable alternative. This study presented a basic method for estimating river flow using satellite imagery. By comparing the data from 14 rivers with actual observation data, the relationship between river surface area extracted from SAR imagery and flow rate was analyzed, and a power function-based flow estimation model was developed. Although the accuracy of flow estimation is influenced by spatial resolution and the geomorphological characteristics of the river, the study demonstrated the potential for estimating river flow over large areas.

Lee et al. (2019) proposed an efficient method for estimating the surface area of reservoirs using Sentinel-1 SAR imagery. The study utilized images captured between May 2015 and August 2019, applying the Radiometric Terrain Correction technique to correct image distortion and the Otsu thresholding method to classify water bodies. The relationship between the water bodies extracted from satellite images and field-measured reservoir volumes was analyzed for two large-scale and two small- to medium-scale reservoirs in South Korea. The results showed a strong correlation for large-scale reservoirs and a moderate correlation for small- to medium-scale reservoirs. The study suggested that overcoming the limitations of spatial resolution could improve accuracy even for small- to medium-scale reservoirs. Additionally, Jang et al. (2020) presented a similar method for estimating reservoir volumes using SAR satellite data for small reservoirs, collecting verification data through drone imaging for reservoirs without observational data. On average, the accuracy of reservoir surface area estimation was 75%, but factors like summer algal blooms reduced accuracy to as low as 60%. The study also found that for reservoirs smaller than 10,000 square meters, the resolution limitations of Sentinel-1 led to decreased accuracy. SAR imagery can improve accuracy through preprocessing and correction processes, and recently, the use of machine learning and artificial intelligence in water body classification has been increasing.

Table 2 . Summary of research on flood using SAR satellites.

Reference	Subject	Satellite	Methodology	Target area
Seo et al. (2018)	Discharge	Sentinel-1	Regression	Han River
Lee et al. (2019)	Waterbody	Sentinel-1	Thresholding	Reservoir
Kim et al. (2020)	Flood detection	Sentinel-1	U-net	East-South Asia
Jang et al. (2020)	Waterbody	Sentinel-1	Thresholding	Reservoir
Kim et al. (2022a)	Waterbody	Sentinel-1	U-net	River/Reservoir
Lee and Jung (2023)	Waterbody	Sentinel-1	U-net	River/Reservoir
Choi et al. (2023)	Waterbody	Sentinel-1/2	Thresholding	River/Reservoir

Kim et al. (2020) used Sentinel-1 imagery and deep learning techniques, SegNet and U-net, to detect flooded areas and compared the performance of the two models. The models were trained using manually classified flood data from major flood events in the Khorat basin in Thailand, the Mekong basin in Laos, and the Cagayan River basin in the Philippines. Both models are based on Convolutional Neural Networks (CNN), with SegNet offering relatively faster processing speed, while U-net provided higher classification accuracy.

Kim et al. (2022a), building on previous research that showed U-net-based water body classification achieving higher accuracy, proposed a method to improve water body detection accuracy by adding modules for Morphology operations and Edge-enhancement to the existing learning model. The Morphology module reduces noise and enhances shapes based on image brightness values, while the Edge-enhancement module helps detect water body boundaries more clearly. Based on the F1-score, the model showed a 9.81% performance improvement compared to the standalone U-net model, successfully detecting many areas that the original model had missed.

Lee and Jung (2023) developed a high-quality training dataset for inland water body detection in South Korea to support researchers in artificial intelligence and deep learning for water body detection. A total of 1,423 water body training datasets were created for the Han River and Nakdong River basins, using both VV and VH polarization images from Sentinel-1, along with Sentinel-2 optical images to ensure complementary data. The performance of U-net using this dataset achieved an F1-score of 0.987 and an Intersection over Union (IoU) of 0.955, demonstrating high accuracy. Even in validation areas not used for training or evaluation, the model showed a high F1-score of 0.941 and an IoU of 0.89, indicating excellent performance.

Drought is a gradually developing natural disaster that unfolds over an extended period, severely impacting agriculture, water resources, and ecosystems. In South Korea, prolonged droughts can lead to significant consequences, such as reduced crop yields, water supply shortages, and increased wildfire risk. Due to the gradual and often invisible nature of droughts, traditional monitoring methods may not be sufficient for rapid detection and response. To address these challenges, remote sensing has become a crucial tool in strengthening drought management strategies.

Remote sensing offers a comprehensive approach to monitoring the development of droughts over vast geographic areas, providing crucial data on variables such as soil moisture, vegetation health, and surface water availability. Through satellite imagery, remote sensing technology can capture subtle and long-term changes in surface conditions, which are key indicators of the onset of drought. When combined with meteorological data, this information can track the severity of droughts and contribute to early warning systems, helping to mitigate the impact on agricultural production and water resource management.

In South Korea, various remote sensing systems have been introduced to monitor the increasingly frequent and intense droughts caused by climate change. These systems are integrated into the national drought monitoring framework, helping to track water resource availability and support the development of reservoir management and agricultural planning. In this chapter, we will review current applications of remote sensing in drought management, focusing on key technologies, case studies, and ongoing efforts to improve drought monitoring strategies. Key remote sensing technologies in drought management include the use of MODIS sensors, satellite rainfall data, and soil moisture data. These three technologies have made significant contributions to the early detection and monitoring of droughts and are continually advancing to provide more accurate data.

The MODIS sensor, mounted on NASA’s Terra and Aqua satellites, is a powerful tool capable of observing wide areas of the Earth’s surface daily. MODIS can calculate several indices to monitor vegetation health, among which the NDVI and EVI play a critical role in detecting drought-affected areas. These indices measure photosynthetic activity, allowing for the assessment of vegetation health, and can quickly detect plant stress caused by water shortages. As a result, they provide early warnings before droughts spread, offering crucial information for agricultural and water resource management policies.

Satellite rainfall data is another valuable tool that provides essential information for drought management. It allows for the widespread monitoring of rainfall, offering reliable data even in areas where ground-based observation stations are scarce. This capability enables the tracking of changes in rainfall patterns before droughts occur, making it highly beneficial for long-term drought forecasting and management. Notably, NASA’s GPM mission has enhanced the accuracy of drought monitoring by providing high-resolution rainfall data on a global scale.

Soil moisture data is an essential element in drought monitoring. The moisture status of the soil provides early signals of drought, and monitoring it plays a crucial role in predicting the impact of drought on agriculture and ecosystems. Soil moisture monitoring is accomplished through technologies such as Passive Microwave Sensing, which detects the amount of moisture beneath the surface, allowing for an assessment of how vulnerable the soil is to drought. Soil moisture data is especially useful for predicting crop growth conditions and understanding plant stress due to water shortages. For example, understanding the soil moisture retention capacity during dry periods can help optimize water management and agricultural practices.

MODIS sensors, satellite rainfall data, and soil moisture data play complementary roles, enabling a comprehensive analysis of vegetation health, rainfall patterns, and soil moisture conditions. With recent advancements in technology, the resolution and data processing capabilities of optical sensors like MODIS have improved, while satellite rainfall data can now provide more precise and near real-time information. Additionally, soil moisture monitoring technology has become more accurate, allowing for better assessments of drought progression. These advancements are crucial for detecting droughts more quickly and establishing effective response strategies. In the future, MODIS sensors, satellite rainfall data, and soil moisture data will continue to be key tools in drought management, further strengthening drought prediction and response capabilities. In the first section of this chapter, we analyze case studies on drought-related research using multispectral imagery, which has been actively conducted since the late 2000s, primarily focused on the MODIS sensor.

Park et al. (2006) utilized MODIS imagery from NASA’s Terra satellite to analyze drought indicators from 2000 to 2005. In this study, well-known drought indicators such as NDVI, along with Land Surface Temperature (LST) and LSWI, were used to assess drought conditions. The key findings suggested that NDVI and LSWI were suitable indicators for evaluating spring droughts, whereas LST was found to be relatively less effective for drought detection.

Park and Kim (2009), building on previous research, evaluated the utility of NDVI by comparing it with ground-based data indicators such as the Palmer Drought Severity Index (PDSI) and the Standardized Precipitation Index (SPI). The study found that satellite-based NDVI showed a high correlation with the 6-month cumulative SPI, indicating that NDVI appropriately reflects vegetation conditions related to drought. In contrast, NDVI showed relatively lower correlation with PDSI, confirming that NDVI is a valid indicator for meteorological drought indices.

Kim and Park (2010) pointed out the limitations of traditional ground-based drought indices like SPI and PDSI in explaining the spatial distribution of droughts. They combined satellite-based NDVI and LST data and applied the Classification and Regression Trees (CART) method to calculate nine drought severity levels, which were then spatially distributed. This approach was proposed as a way to overcome the limitations of single drought indices and ground-based point data, providing a more detailed spatial analysis of droughts.

Rhee et al. (2014) utilized MODIS and TRMM satellite rainfall data to estimate evapotranspiration and precipitation and to evaluate rainfall conditions. Using LST data from the MODIS sensor, they estimated surface evapotranspiration through the Hargreaves method, while precipitation in the study area was estimated using satellite-based spatial rainfall data provided by TRMM. The severity of drought was assessed using the Precipitation-Potential Evapotranspiration (P-PET) index, which is based on the difference between precipitation and potential evapotranspiration. This study presented an effective method for hydrologically evaluating drought in unmeasured basins and areas with limited ground data, and it suggested that more precise analysis could be achieved with higher-resolution data.

Sur et al. (2014) analyzed the applicability of the Evaporative Stress Index (ESI), derived from MODIS satellite data, by comparing it with traditional drought indices such as PDSI and SPI. ESI is calculated as the ratio of Actual Evapotranspiration (AET) to PET, serving as an indicator that reflects moisture stress based on evaporation. When compared to SPI and PDSI during the 2013 drought in southern South Korea, ESI demonstrated superior drought detection capabilities and had the advantage of providing spatial distribution for drought analysis at both administrative and sub-watershed levels.

Nam et al. (2015) proposed the Vegetation Drought Response Index (VegDRI), which combines ground-based and satellite-based drought indices. VegDRI utilizes MODIS-based NDVI, along with ground-based data such as the Self-Calibrating (SC)-PDSI and SPI. Additionally, it incorporates DEM, land cover maps, and Antecedent Moisture Conditions (AMC) to enhance the accuracy of the index. VegDRI calculates drought indices for each grid cell using the CART algorithm, combining drought indices from various sources with regional conditions.

Park et al. (2015) evaluated the applicability of the MODIS satellite-based DSI by comparing it with SPI in domestic drought cases. DSI reflects vegetation conditions and moisture loss from the surface and is calculated by combining satellite-based NDVI and ET. This study assessed drought evaluation and prediction performance in the Dongducheon and Taebaek regions of South Korea, with a prediction accuracy of over 65%. The findings suggest that DSI can be used not only for current drought assessments but also for future drought predictions.

Baek et al. (2016) used MODIS satellite data to analyze and evaluate agricultural drought by utilizing not only NDVI but also EVI and Vegetation Stress Index Anomaly (VSIA). EVI is an index that reduces the soil effects and atmospheric influences that impact reflectance values compared to NDVI, while VSIA is a reconstructed index that removes the strong seasonal influences on vegetation indices by calculating EVI anomalies for sub-regions. When applied to the severe drought case in 2001, the results demonstrated that VSIA could show detailed regional vegetation stress compared to SPI, which is based solely on rainfall, highlighting its usefulness for agricultural drought monitoring.

Kim and Shim (2017) monitored drought conditions using various satellite observation data to conduct integrated drought monitoring across the Korean Peninsula, analyzing meteorological, hydrological, and ecological drought indices comprehensively. The study utilized the SMAP satellite to observe soil moisture, the Gravity Recovery and Climate Experiment (GRACE) satellite to monitor surface and groundwater storage changes, the MODIS satellite to calculate vegetation indices and evapotranspiration, and the TRMM and GPM satellites for satellite rainfall analysis. The study emphasized that satellite-based drought monitoring is a powerful tool for comprehensively assessing different types of droughts, while also stressing the need for data integration and advanced data analysis techniques.

Yoon et al. (2018) analyzed the applicability of the ESI for agricultural drought monitoring by comparing it with various other drought indices. The potential evapotranspiration used for ESI calculation was determined using the Food and Agriculture Organization (FAO) Penman-Monteith method, while actual evapotranspiration was estimated based on thermal infrared remote sensing. The study compared ESI with other drought indices such as NDVI, EVI, and Vegetation Health Index (VHI), focusing on the drought conditions of 2017. ESI was found to be the earliest indicator, detecting the onset of drought by mid-April, offering the advantage of early detection of agricultural drought. However, the study highlighted that the issue of spatial resolution remains a challenge that needs to be addressed.

Yoon et al. (2020) improved the spatial resolution of the ESI from 5 km to 500 m to overcome the resolution limitations identified in previous studies. By analyzing drought cases from 2001, 2009, 2014, and 2017, and comparing the low-resolution ESI with the SPI 6 index derived from ground observation data using Receiver Operating Characteristics (ROC) analysis, they confirmed that the 500 m resolution ESI was effective in detecting drought. This demonstrated that drought could be precisely detected even in small-scale agricultural areas in Korea. Furthermore, Lee et al. (2021) further validated the applicability of high-resolution ESI by scaling it down to individual rice paddies and comparing it with actual changes in water supply.

Kang et al. (2022) proposed a more refined drought index by spatially combining satellite-based and ground-based drought indices. In this study, they resampled MODIS satellite-based NDVI and ground-based SPI from rain gauge stations to the same resolution through spatial interpolation, then combined these indices to calculate the Scaled Drought Condition Index (SDCI). SDCI is derived by integrating the Normalized Precipitation Index (PCI), Temperature Condition Index (TCI), and Vegetation Condition Index (VCI). The study demonstrated that it is possible to calculate 1-, 3-, and 6-month cumulative SDCI indices using historical data, with the ability to detect droughts up to two months before they occur, thus enhancing the potential for drought prediction. In the second section of this chapter, we analyze drought-related research case studies that have been actively conducted since the late 2000s, focusing on satellite rainfall data from TRMM, GPM, and Climate Hazards Group InfraRed Precipitation with Stations (CHIRPS).

Table 3 . Summary of research on drought using multispectral satellites.

Reference	Satellite	Drought index
Park et al. (2006)	MODIS	NDVI / LSWI
Park and Kim (2009)	MODIS	NDVI
Kim and Park (2010)	MODIS	NDVI / LST
Rhee et al. (2014)	MODIS / TRMM	LST / P-PET
Sur et al. (2014)	MODIS	ESI
Nam et al. (2015)	MODIS	VegDRI
Park et al. (2015)	MODIS	DSI
Beak et al. (2016)	MODIS	EVI / VSIA
Kim and Shim (2017)	SMAP / MODIS / GRACE/ GPM	SPI / NDCI / EVI
Yoon et al. (2018)	MODIS	DSI / VHI / LAI / NDVI / EVI
Yoon et al. (2020)	MODIS	ESI
Lee et al. (2021)	MODIS	ESI
Kang et al. (2022)	MODIS	NDVI / SDCI

Jang et al. (2018) calculated the Effective Drought Index (EDI) using satellite-based rainfall data to apply it to unmeasured areas and evaluated its applicability by comparing it with ground observation data. The satellite rainfall data used in this study included Precipitation Estimation from Remotely Sensed Information-Climate Data Record (PERSIANN-CDR) using Artificial Neural Networks, TRMM, and Integrated Multi-satellite Retrievals for GPM (GPM IMERG) providing real-time rainfall data at 30-minute intervals since 2014 as the successor to the TRMM mission. EDI is an index that assesses drought severity by considering precipitation, runoff, and evaporation over time, similar to SPI, with the advantage of being simple to calculate using only rainfall data and the ability to assess daily drought severity. Using data from the Hoengseong Dam and Yongdam Dam basins in South Korea from 2001 to 2016, the study calculated the EDI index and compared it with ground rainfall data, showing correlation coefficients of 0.814 and 0.817, respectively. This highlighted the potential for calculating meteorological drought indices using satellite rainfall data and suggested that it could be applied to unmeasured areas lacking ground observation data.

Lee et al. (2018) used TRMM, GPM satellite rainfall data, and MODIS-based soil moisture to calculate an agricultural drought index. The rainfall data provided by TRMM/GPM, along with meteorological data from the KMA, were combined with the higher-resolution MODIS soil moisture data to estimate daily soil moisture. This information was then used as input for the Soil-Water-Atmosphere-Plant (SWAP) model to also estimate daily evapotranspiration. The calculated soil moisture and evapotranspiration were used to compute the Soil Moisture Percentile (SMP) and the Soil Moisture Deficit Index (SMDI). SMP evaluates relative soil moisture at a specific point in time, making it suitable for short-term drought assessment, while SMDI accumulates long-term soil moisture deficits, making it more appropriate for long-term drought evaluation and agricultural drought monitoring. This study indicated that satellite-based rainfall and soil moisture data could be effectively used for drought response and management in unmeasured areas.

Shin et al. (2019) proposed a decision-making model using a Bayesian network for meteorological drought forecasting based on satellite rainfall data. In addition to assessing the current drought status, it is also crucial to evaluate the onset, duration, and termination of droughts, and this study explored a probabilistic approach. Satellite rainfall data from PERSIANN-CDR, TRMM, and GPM were used, and a decision-making model based on a Bayesian network—a conditional probability model—was applied to forecast drought conditions. The drought prediction results were evaluated using ROC analysis, and the model showed higher predictive performance than the existing Multi-Model Ensemble (MME) model for 2- to 3-month drought forecasts. This study suggested that satellite rainfall and probabilistic drought forecasting using a Bayesian network can effectively predict meteorological droughts, aiding in the early detection of drought onset, duration, and alleviation.

Table 4 . Comparison of satellite precipitation data source.

Satellite data	Operating organization	Data available since	Spatial resolution	Temporal resolution
TRMM	NASA and JAXA	1997	0.25° (~25km)	3 hours
GPM	NASA and JAXA	2014	0.1° (~10km)	30 min to 1 hour
CHIRPS	CHG and UCSB	1981	0.05° (~5km)	Daily
PERSIANN-CDR	CHRS at UC Irvine	1983	0.25° (~25km)	Daily
GPCC	German Weather Service (DWD)	1891	1.0° (~100km)	Monthly

CHG: Climate Hazards Group, UCSB: University of California, Santa Barbara, CHRS: Center for Hydrometeorology and Remote Sensing..

Table 5 . Summary of research on drought using satellite rainfall.

Reference	Satellite	Drought index
Jang et al. (2018)	PERSIAAN-CDR / TRMM / GPM	EDI
Lee et al. (2018)	TRMM / GPM	SPI, SMP, SMDI
Shin et al. (2019)	PERSIAAN-CDR / TRMM / GPM	SPI
Nam et al. (2015)	CHIRPS	SPI
Mun et al. (2020)	CHIRPS / PERSIANN-CDR / GPCC	SPI

Nam et al. (2015) studied the applicability of CHIRPS rainfall data for evaluating meteorological drought indices on the Korean Peninsula. CHIRPS provides global high-resolution satellite rainfall data at approximately 5 km spatial resolution. The SPI index calculated using CHIRPS showed a correlation coefficient of over 0.7 when compared to SPI indices from ground-based observation stations in Korea, indicating sufficient applicability. Although CHIRPS has lower temporal resolution compared to TRMM and GPM, its higher spatial resolution makes it advantageous for long-term drought monitoring and suggests its potential for monitoring meteorological droughts across the entire Korean Peninsula, including North Korea.

Mun et al. (2020), building on previous research showing that global satellite-based rainfall data can be applied to meteorological drought assessment, calculated and evaluated the satellite rainfall-based SPI for major countries in East Asia (South Korea, China, Japan, Mongolia, etc.). The rainfall data used for comparison included CHIRPS, PERSIANN-CDR, and Global Precipitation Climatology Centre (GPCC), each with different spatial and temporal resolutions. The study found that SPI calculated using CHIRPS and GPCC had high correlation with ground observation data, whereas SPI based on PERSIANN-CDR showed somewhat lower accuracy. In particular, CHIRPS, with its higher resolution compared to other satellite rainfall data, was suggested to have effective applicability in the East Asia region.

In the field of flood management, we analyzed flood monitoring and damage assessment methods, particularly focusing on optical satellites and SAR satellites. This analysis reaffirms the crucial role that remote sensing holds in flood management.

First, optical satellites provide high-resolution image data, making them a highly useful tool for visually assessing the extent of damage and terrain changes in flood-affected areas. They are particularly effective in analyzing pre- and post-flood surface changes and visually assessing damage, which can be used to develop recovery plans. However, a drawback of optical satellites is their sensitivity to weather conditions. In situations with cloud cover or rainfall, capturing clear images becomes difficult, potentially limiting their ability to provide immediate response during flood events.

A tool that complements these limitations is the SAR satellite. SAR can collect data regardless of weather or time, making it ideal for rapid flood monitoring in emergencies. It can clearly distinguish flooded areas by utilizing the reflective properties of water and land, and one of its key advantages is its ability to operate even in cloudy conditions or at night. SAR satellites have been effectively used not only for large-scale flood damage monitoring but also for estimating water surface area and flow rates in small- to medium-sized rivers and reservoirs.

Additionally, research utilizing artificial intelligence and deep learning techniques has shown that flood damage detection based on SAR satellite images achieves very high accuracy. In particular, CNN-based deep learning models like SegNet and U-net have demonstrated the ability to detect flood-affected areas quickly and accurately, which will make significant contributions to future flood management.

In the field of drought management, we examined the impact of remote sensing tools such as MODIS, satellite rainfall data, and soil moisture data. The reviewed studies showed that drought monitoring using satellite data is highly effective in complementing traditional meteorological drought monitoring methods and demonstrated that combining various data sources allows for quicker detection of early drought signals.

Specifically, the MODIS sensor, utilizing NDVI and EVI, has proven to be an effective tool for monitoring vegetation health, showing a high correlation with meteorological drought indices (PDSI, SPI), confirming its validity as a drought detection indicator. Additionally, studies have combined MODIS sensor evapotranspiration data with land surface temperature to assess drought severity, further improving the accuracy of drought detection.

Satellite rainfall data serves as an effective tool for monitoring rainfall patterns, especially in areas with limited ground observation stations. Data from satellites like TRMM and GPM play an important role in global weather monitoring. This research demonstrated that using such satellite-based rainfall data to calculate drought indices like the EDI enables effective drought detection even in unmeasured areas.

Soil moisture data is an essential element in drought monitoring, as a lack of soil moisture serves as an early signal of drought and significantly impacts agricultural production and ecosystems. Soil moisture monitoring technology plays a crucial role in assessing surface and subsurface moisture conditions, allowing for the early detection of drought.

In conclusion, various remote sensing technologies have been confirmed to make significant contributions to the early detection, monitoring, and development of response strategies for drought. As the precision and resolution of remote sensing data improve, the analysis and prediction of drought impacts become more accurate, enhancing the ability to respond to increasingly frequent and severe droughts driven by climate change.

In South Korea, the Ministry of Environment and K-water are developing the CAS500-5 water resources satellite, as described in Fig. 1, to effectively monitor and respond to water-related disasters, including floods and droughts. The satellite is equipped with a C-band SAR, capable of observing the Korean Peninsula twice a day at a resolution of less than 10 meters, with a swath width of 120 km. This is expected to address the resolution limitations that have posed challenges in satellite utilization. Once operational, the satellite will provide high-resolution, high-frequency monitoring of the entire Korean Peninsula, and as illustrated in Fig. 2, related technologies are being developed to support its use in flood, drought, safety, and environmental applications.

Figure 1. Specifications of water resources satellite (CAS500-5).

Figure 2. Major applications (flood, drought, safety, environment) of water resources satellites.

Due to climate change, the frequency and intensity of water-related disasters such as floods and droughts are increasing in South Korea. Advanced and efficient management tools are essential for managing these disasters. Remote sensing technology plays a crucial role in improving the ability to monitor, predict, and mitigate the impacts of such disasters. In particular, optical and SAR satellites have made significant contributions to flood management, while multispectral and microwave sensors, along with satellite rainfall data, have been critical for drought management.

In flood management, integrating optical satellite imagery with SAR data enables more accurate and rapid detection of floods, even in challenging weather conditions. The combination of these technologies effectively supports early warning systems and emergency response. As technology advances, higher resolution and faster data processing speeds will further enhance South Korea’s flood management capabilities.

In drought management, optical and microwave sensors, satellite precipitation data, and soil moisture monitoring play crucial roles in tracking the progressive development of drought and its impacts on vegetation and water resources. Multispectral satellite-based vegetation indices (such as NDVI and EVI), satellite precipitation data, and soil moisture measurements have proven to be highly effective for providing early warnings, agricultural planning, and water resource management. With advancements in remote sensing technologies, the integration of real-time soil moisture data with precipitation and vegetation indices is expected to offer more comprehensive insights into the severity of droughts, enabling the formulation of more accurate and effective response strategies

In this paper, we offer several recommendations for managing water-related disasters using remote sensing.

Enhancing Data Integration: It is necessary to strengthen an integrated disaster management framework by combining various remote sensing data, including optical, radar, and soil moisture data. This approach will provide a more comprehensive view of flood and drought situations, enabling more informed decision-making and disaster response.

Investing in Technological Development: Continuous investment is required to improve the resolution, accuracy, and speed of remote sensing technologies. Enhancing satellite-based flood and drought monitoring capabilities will help South Korea further reduce the risks and impacts of these natural disasters.

Promoting Cross-Sector Collaboration: Strengthening cooperation among government agencies, research institutions, and the private sector is essential to effectively apply the latest remote sensing technologies to disaster management strategies. Expanding the use of remote sensing data beyond water-related disasters to sectors like agriculture, environment, and urban planning will also be a significant undertaking.

Strengthening Early Warning Systems: Especially in regions where floods and droughts are frequent, it is necessary to further develop and improve early warning systems. By integrating real-time remote sensing data to provide more accurate and timely warnings, better preparedness and response can be achieved.

By utilizing remote sensing technologies and continuously advancing their capabilities, South Korea can significantly enhance its response to water-related disasters. This will contribute not only to the safety and well-being of its citizens but also to reducing the economic and environmental damages caused by floods and droughts.

This research was supported by the Ministry of Environment, under the Development of Ground Operation System for Water Resources Satellite from K-water.

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