ScienceWatch@KIOST
Not a member yet
29648 research outputs found
Sort by
Summer Drying Trends in East Asia: Different Physical Mechanisms between Land and Ocean
No full textSummer precipitation-evaporation patterns across East Asia have shifted toward drier conditions over the past four decades, contrasting with the region's traditionally humid climate. We analyze precipitation- evaporation changes during April–September from 1980–2022 to understand regional mechanisms behind these drying trends. Results show significant drying over southeastern China, the Yellow Sea, and Korean Peninsula regions. Moisture budget decomposition reveals that the physical processes causing drying differ substantially between continental and oceanic areas. Over southeastern China, dynamic processes dominate the drying trend through enhanced moisture divergence. This results from strengthened subsidence linked to Indo-Pacific warm pool heating, driven by both anthropogenic warming and Pacific Decadal Oscillation patterns. In contrast, Yellow Sea drying occurs through thermodynamic processes involving moisture advection changes driven by global warming-induced humidity redistribution. Stronger humidity gradients between the Northwest Pacific and inland China enhance dry air transport from the continent to the Yellow Sea via prevailing southwesterly flows. These results show that East Asian summer drying involves spatially varying processes, improving our understanding of monsoon system changes under climate change.2
위성기반 서남해안 갯벌 지형 및 면적 장기 시계열 변화 양상 분석
No full text한국 서남해안 갯벌은 지속적인 개발, 침식, 매립, 해수면 상승 등 자연적 및 인위적 요인으로 인해 면적 변화가 뚜렷하게 나타나고 있다. 이러한 갯벌은 생태적, 경제적, 환경적으로 중요한 가치를 지니고 있어 갯벌 지형 및 면적 변화 모니터링을 통한 체계적인 관리가 필요하다. 최근 갯벌 조사는 주기적이면서 광역적으로 빠르게 관측이 가능한 위성영상을 활용한 방법이 많이 진행되고 있으며, 이러한 방법을 통해 전통적인 현장조사의 접근성 한계와 선박이나 비행기와 같은 플랫폼들의 높은 조사비용을 효율적으로 보완하였다. 그러나 갯벌지역은 조위 변동으로 인해 최저조위와 최고조위와 같은 특정 시점에서 위성영상을 확보하는 것은 어렵고, 또한 연안 지역에서 자주 발생하는 구름의 영향 등으로 갯벌 지형 및 면적 주제도를 제작하기 위한 입력 자료 확보에 있어 큰 제한점이 되고 있다. 본 연구는 갯벌 면적 주제도 제작 과정에서 구름과 조위 변동의 영향을 최소화하기 위해 기계 학습 기반 분류 기법을 적용하였다. 이를 통해 Normalized Difference Water Index (NDWI)와
Enhanced Vegetation Index (EVI) 등 시계열 영상에서 추출한 다양한 지수와 영상 밝기 정보를 합성하여 분석에 활용하였다. 1984년부터 2024년까지 40년간 Landsat 위성 시리즈 영상을 이용하여 5년 단위로 우리나라 서남해안 지역에 대한 갯벌 면적 주제도를 제작하였으며 기간별로 갯벌 면적 변화에 대한 사례 연구를 수행하였다. 또한 NDWI기반 시계열 갯벌 영상자료와 수륙경계선방법(Waterline method)을 이용해 5년 단위로 갯벌 지형도를 제작하고 갯벌 지형 변화에 대한
사례 연구도 같이 수행하였다. 제작된 갯벌 지형 및 면적 주제도의 정확도는 UAV-LiDAR 자료와 국립해양조사원(KHOA)에서 제공되는 갯벌 면적 자료를 통해 검증하였다. 사례 연구 결과, 위성기반 갯벌 지형도의 정확도는 RMSE 30 cm 이내로 제작이 가능함을 확인하였고, 갯벌 형태 특성에 따라 지형의 변화 양상이 다르게 나타났다. 기계 학습 기반의 방법은 구름과 조위 조건에도 불구하고 신뢰성 높은 갯벌 면적 주제도를 제작할 수 있었으며, 갯벌 면적의 장기적인 변화 양상을 효과적으로 모니터링할 수 있었다. 본 연구는 한국 갯벌의 체계적인 관리와 보존을 위한 중요한 과학적 근거를 제공하며, 향후 지속 가능한 갯벌 관리 정책 수립에 기여할 것으로 기대된다.2
Long-Term Monitoring of Tidal Flat Area Changes in the Korean West Coast Using Time Series Satellite Imagery
No full textThe tidal flats along the Korean west coast have experienced significant area changes due to both natural and anthropogenic factors, including ongoing development, erosion, land reclamation, and sea level rise. These tidal flats hold substantial ecological, economic, and environmental value, necessitating systematic management through monitoring area changes. Recently, tidal flat investigations have increasingly utilized satellite imagery, which allows for periodic and large-scale observations. This approach effectively addresses the accessibility limitations of traditional field surveys and the high costs associated with platforms like ships or aircraft. However, tidal flat regions pose challenges for satellite-based observations due to tidal variations, making it difficult to capture images at specific times such as lowest and highest tide. Furthermore, frequent cloud cover in coastal areas imposes significant constraints on acquiring input data for tidal flat mapping.
This study applied a machine learning based classification method to minimize the effects of cloud cover and tidal variations during the tidal flat mapping process. It utilized synthetic datasets derived from time-series imagery, incorporating indices such as the Normalized Difference Water Index (NDWI) and Enhanced Vegetation Index (EVI), along with brightness information from individual images. Using this approach, the study developed tidal flat area maps for the Korean west coast over a 40-year period from 1984 to 2024, using Landsat satellite series data in 5-year intervals. A case study was conducted to analyze area changes over time, and the accuracy of the generated tidal flat maps was validated against tidal flat area data provided by the Korea Hydrographic and Oceanographic Agency (KHOA).
The results demonstrated that the machine learning-based method produced reliable tidal flat maps, effectively mitigating the impacts of clouds and tidal conditions. Moreover, the approach successfully monitored long-term changes in tidal flat areas. This study provides essential scientific evidence for the systematic management and conservation of Korea tidal flats and is expected to contribute to the formulation of sustainable tidal flat management policies in the future.1
Due Regard Obligations in Areas beyond National Jurisdiction
No full textDue regard obligations require both States and non-State actors to reasonably consider the rights or interests of other States or non-State actors when exercising their own rights and performing their duties. This article examines how due regard obligations should be interpreted in areas beyond national jurisdiction (ABNJ) in light of the adoption of the Agreement under the United Nations Convention on the Law of the Sea on the Conservation and Sustainable Use of Marine Biological Diversity of Areas Beyond National Jurisdiction (BBNJ Agreement). As human activities in ABNJ increase, due regard obligations become crucial for balancing the competing rights and interests of States, non-State actors and relevant institutions, frameworks and bodies (IFBs). The literature and case law have mainly addressed conflicts between coastal and flag States concerning the application of due regard obligations within national jurisdictions. Different dynamics arise in ABNJ from potentially conflicting activities and disagreements between States with the same rights or interests, or those between States, non-State actors and IFBs. This article addresses this gap by analysing the new dynamics of due regard that are expected to arise concerning marine genetic resources and area-based management tools with the implementation of the BBNJ Agreement.11Nssciscopu
A New Species of Laophontella (Harpacticoida, Tetragonicipitidae) from Korea
No full textA new species of the genus Laophontella Thompson & Scott, 1903 was collected from an intertidal tide pool on Jeju Island, Korea. The new species is assigned to Laophontella based on diagnostic morphological features, including pointed posterolateral processes on the cephalothorax, a minute endopod of leg 4, and a foliaceous leg 5 in the female. The genus currently comprises three recognized species, one of which, Laophontella horrida (Por, 1964), includes three subspecies. Among these, L. horrida dentata Mielke, 1992 is most similar to the new species, sharing a backward projection on the first antennular segment and the armature pattern of legs 1–4. However, the new species differs by having seven (vs. eight) antennular segments, inner setae longer (vs. shorter) than apical setae on the female legs 2–4 exopods, a hoe-shaped (vs. typical) inner spine on the male leg 3 endopod, and the absence (vs. presence) of pointed processes on the outer margin of the female caudal rami. This study represents the first report of a new species of Laophontella from Korea, and provides a morphological key to species of Laophontella.2
Analysis of AI Resources Utilization in Maritime Environments
No full textWith the growing adoption of artificial intelligence (AI) technologies in various maritime platforms including autonomous surface vessels,
autonomous underwater vehicles (AUVs), and marine IoT buoys the importance of efficient hardware resource management is becoming
increasingly critical, especially in constrained resource environments. This study aims to analyze how the structural characteristics of AI
models affect the utilization of system resources in such constrained settings, particularly within maritime environments. We quantitatively
measured the utilization of CPU, GPU, memory, and storage for four AI models CNN, GPT, AlexNet, and Diffusion. Experimental results
reveal that the CNN model exhibited evenly distributed and relatively low resource usage, whereas the GPT model demonstrated
strong dependence on GPU computational performance. AlexNet exhibited simultaneous bottlenecks in both storage and GPU, while
the Diffusion model experienced complex resource bottlenecks across GPU, memory, and storage due to its repetitive computation
and intermediate data saving. These findings highlight the limitations of conventional model selection practices that focus primarily
on accuracy or efficiency, and emphasize the necessity of analyzing resource interactions and bottlenecks in advance for each model.
Ultimately, the study underscores the importance of selecting and deploying models suited to resource conditions model selection
and deployment strategies to ensure stable and reliable AI operations in environments with physical constraints, such as maritime
environments.1
Sea Level Reconstruction at Dokdo, East Sea (1993–2023): In-Situ, Altimetry, and Machine Learning Approaches
No full text2
Introducing the Korea–Latin America Cooperation Program for Climate Change Response
No full textThis side event aims to highlight Korea’s leadership in ocean climate monitoring and international cooperation to address climate change. Through case studies from the Northwest Pacific and the Indian Ocean, the event will showcase Korea’s contributions to global observation networks such as Argo, RAMA-K superstation, glider observations, and R/V Isabu campaigns. Peru’s national ocean observation efforts, led by IMARPE, will be introduced to highlight the role of ocean data in fisheries and climate resilience. In addition, the Korea–Perú Research Center for Ocean Science and Technology for Latin America (KOPELAR) will present its initiatives to strengthen regional cooperation in ocean science and capacity building across Latin America. The session will emphasize the strategic importance of ocean data for climate prediction, sustainable development, and international collaboration.1