Reconstructing high-resolution regional significant wave height (SWH) fields from sparse buoy observations is a critical challenge for ocean monitoring. We introduce AUWave, a hybrid deep learning framework that fuses a station-wise encoder with a multi-scale U-Net enhanced by self-attention to recover regional SWH fields. Trained and validated using NDBC buoy observations and ERA5 reanalysis over the Hawaii region, AUWave achieves high accuracy. It consistently outperforms a representative baseline, especially in configurations with more than a single buoy, demonstrating the benefit of its multi-scale architecture. Spatial error analysis shows performance is highest near observation sites, as expected. Further, buoy ablation studies identify critical anchor stations whose removal disproportionately degrades performance, offering actionable guidance for observational network design. AUWave provides a scalable pathway for gap-filling, creating high-resolution priors for data assimilation, and contingency reconstruction. Cross-basin evaluations in the Atlantic and Pacific confirm the model’s robustness and portability, highlighting its potential for operational use across diverse ocean regimes.

Tidal creeks are subject to the integrated effects of geomorphological and eco-hydrodynamical controls and there has been a lack of mechanism to reveal these effects on the formation and evolution of tidal creek systems. By taking the Yellow River Delta as a suitable example, this study presents a detailed analysis of the spatiotemporal adjustments in landforms of tidal creeks, tidal flats, vegetation cover, and the hydrodynamics. Temporally, the tidal flats adjacent to current estuary (Zone I) experienced significant seaward accumulation with a seaward migration rate of approximately 152 m/a in the coastlines, while that in the abandoned sub-delta (Zone II) experienced landward recession with an average rate of approximately 70 m/a. The density and bifurcation ratio of tidal creeks in Zone I and Zone II showed fluctuating upward trend, which increased rapidly during 2001 to 2016 and reached the maximum in 2016. Spatially, the density and bifurcation ratio of tidal creeks in the erosive Zone II is about 51% and 164% larger than the silted Zone I, indicating the denser and more complex tidal creek system developed in erosive tidal flats. Mechanistically, in Zone I, high sediment supply, weak hydrodynamics, and dense vegetation promoted tidal flat accretion and bank stability, favoring linear, low-bifurcation creeks. Vegetation reinforced banks and suppressed overland flow, inhibiting branching. In Zone II, limited sediment, stronger currents, and sparse vegetation led to erosion, enhancing lateral and headward erosion that formed dendritic, high-bifurcation networks. Hydrodynamic-sediment coupling controls branching, and ample sediment with weak flows suppresses bifurcation, whereas sediment deficit with strong flows promotes it. These insights clarify tidal creek formation and aid management of evolving deltaic systems.

In this study, the mechanisms governing sea ice accumulation and the dynamics of ice floes were investigated using a combination of experimental methods and a coupled CFD-DEM model. Based on experimental and numerical results, the kinematic characteristics of ice floes during the accumulation process were analyzed, and the equilibrium thickness of the accumulation profile at the ice-breaking cones and barrier net system was quantified. Additionally, the influences of incident flow velocity, water depth and weak wave disturbance on ice floe equilibrium accumulation, as well as the contact forces exerted by ice floes on the ice-breaking cones and barrier net system, were examined. The findings indicate that incident flow velocity dominates the equilibrium accumulation thickness, with incident flow velocity u < 0.4 m/s identified as a sufficiently safe velocity for cold-water intake operations. Shallow water depths lead to severe blockage and a rapid decline in the operational efficiency of the barrier net. Overall, the CFD-DEM simulation results show excellent agreement with the experimental data, and numerical simulations confirm that ice-breaking cones prevent direct ice floe impact on the barrier net, reducing the peak force acting on it by 16.5%. Weak wave disturbance has an insignificant influence on the contact forces exerted by sea ice floes on ice-breaking cones and the barrier net.

Cold waves are major source of winter disasters in China’s coastal regions, affecting local residents’ lives and economic development. We used the MIKE21 numerical model to simulate hydrodynamics in the Huanghe (Yellow) River Delta. By simulating the strong cold wave event in late November 2022, we analyzed the erosion and deposition changes in the delta. Results show that overall, the new Huanghe River estuary receives sedimentation, while the old estuary shows erosion. Affected by strong cold wave winds, the coastal water level in the delta rose significantly, with a change of about 0.50 m at the new estuary and around 0.57 m at the old estuary. Comparing residual currents with and without the cold wave, the main differences were observed near the new and old estuaries: 0.36 m/s to the southeast near the new estuary and 0.45 m/s to the southwest near the old one. Due to changes in flow velocity caused by the cold wave, erosion near the old estuary reached about 0.14 m, while the maximum sedimentation near the new estuary was around 0.25 m. As wind speed increased, the erosion area expanded in both estuaries; when wind speed decreased, flow slowed and erosion weakened, resulting in a smaller erosion area. This study provided a reference for responding to cold wave disasters in estuarine regions and enhanced understanding of the dynamic evolution of estuarine landforms.

Abstract Over the past four decades, the water and sediment fluxes of the Yellow River to the sea and the deltaic coastal morphology have changed markedly under the combined influence of human activities and natural processes. There is an urgent need to quantitatively investigate how shoreline adjustments driven by outlet migration and the embayment north of the river mouth affect the nearshore tidal‐current field. In this study, Landsat imagery and multi‐year bathymetric data were used to resolve shoreline changes, subaqueous topographic evolution, and outlet shifts of the Yellow River Delta in 1987, 2000, and 2023. A numerical hydrodynamic model was established to quantitatively evaluate the influence of different outlet and shoreline configurations on the nearshore tidal regime and to reveal the response mechanisms of the tidal‐current field to shoreline evolution, river‐mouth deflection and changes in runoff intensity. The results indicate the following: (1) Nearshore tidal currents are predominantly rectilinear, and an elliptical high‐velocity zone (>0.8 m/s) that migrates with the river mouth has persisted. Northward deflection of the river mouth has created an embayment between the Gudong seawall and the mouth, where current speed has decreased significantly, with a maximum reduction of 0.46 m/s, corresponding to a change rate of 78.23%. (2) Runoff intensity strongly modulates water levels and tidal currents near the river mouth. The runoff‐dominated area expands with increasing discharge, and under extreme runoff (4000 m 3 /s), the influence can extend up to approximately 10 km offshore beyond the mouth. (3) The distribution and strength of the tidal‐current field are significantly affected by the outlet course and shoreline configuration. Corresponding adjustments of the tidal‐current field are observed in regions of intense erosion and accretion, such as the Gudong nearshore area and the active and abandoned river mouths. These findings are of considerable engineering value for understanding delta morphodynamics and informing nearshore environmental management.

Tsunamis generated by submarine landslides pose significant threats to coastal populations and infrastructure. Their hazards are greatly related to wave propagation processes, such as wave refraction, reflection, and resonance, which can lead to unexpected wave elevations and catastrophic damage. Consequently, a comprehensive understanding of wave propagation and associated influencing factors is essential for enhancing tsunami prediction accuracy and mitigating hazards. In this study, based on the parameter constraints derived from multibeam bathymetric data, we modeled a realistic landslide scenario and associated tsunamis in the Xisha Islands of the South China Sea (SCS) by the coupled models of NHWAVE (Non-hydrostatic Wave Model) and FUNWAVE-TVD (Fully Nonlinear Wave model - Total Variation Diminishing). Through wavelet transform and model decomposition, we concentrated on investigating the energy variations and wave interactions during the wave propagation processes. Additionally, we analyzed the impact of island resonance on tsunami elevation. Our study highlights that landslide-generated tsunamis can excite wave resonance between islands, leading to sustained influence. By decomposing and analyzing the time series data, we identify two components of tsunamis: characteristic component 1 (CC1) and characteristic component 2 (CC2). CC1 is characterized by low-amplitude, high-frequency, short-period waves; conversely, CC2 consists of high-amplitude, low-frequency, long-period waves. CC1 represents the propagated tsunami waves that are directly generated by landslides; whereas CC2 encompasses those influenced by wave refraction and reflection during wave propagation. This study quantifies the contributions of refraction and reflection effects on tsunamis during wave propagation by using variance ratios and the Pearson correlation coefficients. The findings provide quantitative evidence elucidating how island morphologies influence the propagation of landslide-generated tsunamis. This study indicates that submarine landslide-generated tsunami represents a substantial geohazard for the islands, and will be also helpful for the tsunami hazard assessment in the regions where there are complex morphologies (e.g., islands and bays).

The research area is located in the western part of the Luntai uplift of the Tabei uplift in the Tarim Basin. It is in a favorable position for the southward migration of continental oil and gas in Kuqa. The area is rich in oil and gas resources. Previous exploration and development have revealed that there is a special type of buried hill reservoir in this area-carbonate-sandstone composite buried hill reservoir. At present, many reservoirs found in this area are gradually entering the development stage, so the demand for fine description of composite buried hill reservoirs is increasingly urgent. This study fully utilizes seismic data, implementing precise well-seismic fine calibration, horizon fine interpretation, fine identification of fracture, fine identification of buried hill strata, fine compilation of structural map—a series of five targeted technical measures. These efforts have achieved a refined description of the composite latent mountain oil reservoir in the study area. This fulfillment caters to the need for refined descriptions of oil reservoirs in older regions. It has also established a technical series for the seismic refinement description of composite latent mountain oil reservoirs in the Tarim Basin, and providing strong support for the fine evaluation and efficient development of the next phase of regional bands.

Seventeen polycyclic aromatic hydrocarbons (PAHs) and thirty-three n-alkanes were studied in 128 PM2.5 samples collected seasonally at urban and suburban sites of Wuhu, a rapidly developing city in the subtropical monsoon climate zone of China. The concentrations, spatiotemporal distributions, meteorological correlations, emission source apportionment, and spatial source regions were discussed. The average total PAHs concentrations were 9.72 ± 10.83ng/m3 and 9.90±10.29 ng/m3 at urban and suburban sites, respectively, while the corresponding annual average concentrations of n-alkanes were 36.59 ± 25.69 ng/m3 and 29.52 ± 28.00 ng/m3. No significant urban-suburban differences were observed for either TPAH or n-alkane concentrations. Such spatial homogeneity may be attributable to rapid urbanization-induced pollution homogenization or intensive regional pollutant transport, which offsets local urban-suburban concentration gradients. Both pollutants exhibited pronounced seasonal patterns consistent with regional climatic characteristics, with concentrations peaking in winter and declining to the lowest levels in summer. Meteorological correlation analysis further confirmed that temperature, atmospheric pressure, and wind speed were dominant factors modulating PAH and n-alkane variations. PMF-based source apportionment results revealed that the PAHs in Wuhu was predominantly derived from coal combustion, followed by vehicle exhaust and biomass burning, with petroleum volatilization contributing minimally. For n-alkanes, CPI and WNA analyses indicated that anthropogenic fossil fuel combustion constituted the dominant source across seasons. Spatial source analysis further demonstrated that regional pollutant accumulation was primarily attributed to emissions from surrounding urban agglomerations, with long-range transport from the northwest and northeast of Wuhu playing a critical role in pollutant loading. Collectively, these findings provide scientific support for seasonal and wind-direction-based early warning and targeted pollution control strategies for local atmospheric environmental management.

Delta morphological evolution, shaped by fluvial-sediment processes and human activities, is critical for estuarine-coastal ecological protection. While river discharge and sediment inflow are known to influence delta development, the century-scale effects of river morphological changes remain insufficiently understood. This study investigates the 350 years evolution of river deltas under runoff-tidal dynamics, using the Hydraulic Geometry Coefficient (HGC) to quantify river morphology and a Delft3D-based hydrodynamic-sediment-topography coupled model to assess HGC impacts. Validated against the Parana and Danube deltas, results confirm HGC as a pivotal control on delta stability and channel configuration: high HGC (wide-shallow channels) facilitates complex multi-distributary deltas, while low HGC (narrow-deep channels) leads to simpler, narrower forms. HGC's influence on riverbed fluctuations is time-dependent-weak initially, nonlinearly enhanced in the medium term, and weakened over the long term-with braided river complexity positively correlating with HGC. A critical morphological threshold (HGC = 5-10) and an "system memory effect" are identified: early morphological differences (e.g., channel estuary count) persist for centuries, with high HGC deltas maintaining complexity and low HGC deltas degenerating into single-channel systems. These findings highlight the significance of channel characteristics in delta formation, offering critical insights for estuary management and delta conservation.

A coupled CFD-DEM model was adopted to investigate the floating ice accumulation mechanism and its disturbance to the flow field in the pump house of coastal nuclear power plants in cold regions. Based on numerical simulations, the motion, accumulation, and flow interaction characteristics of floating ice under various release positions and heights were analyzed. The results indicate that the release height significantly governs the accumulation morphology and hydraulic response. The release height critically determines ice accumulation patterns and hydraulic responses. For inlet scenarios, lower heights induce a dense, wedge-shaped accumulation at the coarse trash rack, increasing thickness by 57.69% and shifting the accumulation 38.16% inlet-ward compared to higher releases. Conversely, higher releases enhance dispersion, expanding disturbances to the central pump house and intensifying flow heterogeneity. In bottom release cases, lower heights form wall-adhering accumulations, while higher releases cause ice to rise into mid-upper layers, thereby markedly intensifying local vortices (peak intensity 79.68, approximately 300% higher). Spatial release locations induce 2.7–4.8-fold variations in flow disturbance intensity across monitoring points. These findings clarify the combined impact of the release height and location on the ice accumulation and flow field dynamics, offering critical insights for the anti-ice design and flow safety assessment of pump houses.

Study region: The active Yellow River Delta (YRD) lobe. Study focus: Over recent decades, the morphology of the active delta lobe has changed frequently and unpredictably under a changing environment. To address this challenge, this study developed a lobe area-tidal level model leveraging satellite images combined with a machine learning (ML)-based approach to monitor the evolution of the lobe area and morphological changes since the implementation of the Water-Sediment Regulation Scheme. This framework enables consistent, large-scale extraction of lobes from long-term satellite images, resolving limitations of subjectivity and low efficiency in conventional methods. New hydrological insights for the region: The results from ML-based monitoring show that the overall area of the lobe has been expanding seaward at a rate of approximately 4.0 km2/yr, and its morphology has exhibited three stages: eastward development (2002-2009), northward development (2009-2017), and northward stabilization (2017-2022). The dynamic spatiotemporal patterns of the lobe reflect the complex interactions between channel dynamics, vegetation feedback, and sediment supply. The migration/bifurcation of the mouth channel have altered the redistribution of sediment, increasing the lobe land-building efficiency by 142-230 %. A critical sediment threshold ranging from 0.48 to 1.5 x 108 t is found to sustain the development of the lobe. This study clarifies the importance of multi-factorial interactions in the evolution of the lobe, and emphasizes the need for balanced intervention measures to maintain delta resilience.

Piles are common support elements for marine and coastal structures. The scour around pile foundations caused by currents is a major threat to the stability and safety of these structures. The empirical equations commonly used for estimating the equilibrium scour depth around pile groups are limited in their predicative capability, especially when the current approaches the pile group at an angle. This study applies a Multi-Layer Perceptron Backpropagation (MLP/BP) neural network to develop a general model for predicting the local maximum equilibrium scour depth around pile groups in steady currents. The input parameters for the model include all relevant non-dimensional hydrodynamic and structural variables taking full account of the effects of the pile group arrangement and its orientation relative to the approaching current. The model’s performance was evaluated by comparing its predictions against those generated by multiple other machine learning methods, as well as against results from widely used empirical formulas. A comprehensive sensitivity analysis is carried out to determine the importance ranking of the input parameters on model accuracy.

Laizhou Bay, a semi-enclosed bay, is prone to storm surges from cold waves due to its geographic and environmental characteristics. This study uses satellite data, in situ measurements, and the MIKE numerical model to analyze storm surges along Laizhou Bay’s coast under no-dike conditions. It examines the surges caused by cold waves with different intensities and directions. This study provides the storm surge disaster risk levels along Laizhou Bay’s coast. The results show that the maximum sustained wind speed during cold waves is distributed between the NW and NE. The NE wind direction causes the most severe storm surge along Laizhou Bay. Under NE-directed cold waves with level 12 wind, the maximum risk areas for Level III and IV are approximately 1341 km2 and 1294 km2, respectively. Dongying, Shouguang, and Hanting exhibit large Level I and II risk zones. The maximum seawater intrusion distance along the Kenli coast is about 41 km. The coastal segment from Kenli to Changyi is most severely affected by storm surges. It is recommended to effectively maintain and heighten seawalls along this segment to mitigate storm surge disasters caused by strong NE winds.

IntroductionOver the past three decades, approximately 16% of the world’s tidal flats have been lost. In the Yellow River Delta (YRD), the reduction in sediment supply due to decreased Yellow River discharge has raised concerns regarding the morphological stability of tidal flats.MethodsTo investigate the response of tidal flat development to reduced sediment input, a novel physical model experiment was conducted using natural silt from the YRD tidal flat. Offshore sediment concentration was decreased to simulate reduced sediment supply. An Argus system was deployed in the wave basin for the first time to capture the morphological changes during the experiment.ResultsThe results indicate that 94% of suspended sediment settles during its transport from offshore to the tidal flat. Suspended sediment concentration (SSC) in the subtidal zone increased during flood tides and decreased during ebb tides. With decreasing SSC, comb-shaped flow marks formed along the vertical shoreline in the intertidal zone, while the subtidal zone was dominated by the development of sand waves. For a given SSC, sand wave morphology and development patterns varied across different cross-shore profiles; conversely, for a given profile, different SSC levels led to distinct sand wave characteristics.DiscussionThis study demonstrates the significant influence of SSC on tidal flat morphology and sediment dynamics. The findings suggest that continued reductions in sediment supply could exacerbate erosion risks in the YRD, highlighting the need for sediment management strategies to preserve tidal flat stability.

Reconstructing high-resolution regional significant wave height fields from sparse and uneven buoy observations remains a core challenge for ocean monitoring and risk-aware operations. We introduce AUWave, a hybrid deep learning framework that fuses a station-wise sequence encoder (MLP) with a multi-scale U-Net enhanced by a bottleneck self-attention layer to recover 32×32 regional SWH fields. A systematic Bayesian hyperparameter search with Optuna identifies the learning rate as the dominant driver of generalization, followed by the scheduler decay and the latent dimension. Using NDBC buoy observations and ERA5 reanalysis over the Hawaii region, AUWave attains a minimum validation loss of 0.043285 and a slightly right-skewed RMSE distribution. Spatial errors are lowest near observation sites and increase with distance, reflecting identifiability limits under sparse sampling. Sensitivity experiments show that AUWave consistently outperforms a representative baseline in data-richer configurations, while the baseline is only marginally competitive in the most underdetermined single-buoy cases. The architecture's multi-scale and attention components translate into accuracy gains when minimal but non-trivial spatial anchoring is available. Error maps and buoy ablations reveal key anchor stations whose removal disproportionately degrades performance, offering actionable guidance for network design. AUWave provides a scalable pathway for gap filling, high-resolution priors for data assimilation, and contingency reconstruction.

Remote sensing has become an essential tool for monitoring the discharge range and thermal discharge temperature rise classification of nuclear power plants. The core of remote sensing-based thermal discharge monitoring is the accurate extraction of background water temperature. Existing methods for extracting background temperature mainly involve two approaches: one is based on expert prior knowledge to define a temperature range, with the average temperature within this range used as the background temperature. However, this method is somewhat arbitrary and heavily influenced by human judgment. The second approach is based on deep learning, which can accurately extract background temperature but requires a large amount of training data and needs to be retrained for different datasets. To address these issues, we propose a background temperature extraction method based on the temperature gradient algorithm. To validate the applicability and accuracy of this method, we utilized 991 scenes of Landsat data from January 1, 2023, to June 30, 2024, and performed validation across 65 nuclear power plants worldwide. The results show that, compared to methods such as the average temperature correction and adjacent-zone substitution method, the temperature gradient method can automatically and accurately extract background temperature and temperature rise areas. Moreover, this method demonstrates strong general applicability, making it suitable for both coastal and lakeside nuclear power plants.

This study investigates the applicability of Chronos, a tokenized and pretrained large-scale time series model, for forecasting Significant Wave Height across diverse buoy locations. We evaluate both the pre-trained version (ChronosZeroShot) and a domain-specific fine-tuned variant (ChronosFineTuned) over forecast horizons ranging from 1 to 120 h. Results demonstrate that ChronosFineTuned consistently outperforms a suite of statistical and deep learning baselines, achieving superior accuracy and generalization, particularly for mid-to-long-term forecasts. Interestingly, ChronosZeroShot exhibits competitive performance in short-term horizons (up to 24 h) despite lacking exposure to ocean wave data, highlighting the value of general-purpose pretraining. Performance variability across buoys is further analyzed, revealing that depth, latitude, and data volume influence predictive skill. Statistical significance testing confirms the robustness of ChronosFineTuned across most scenarios. The findings underscore the potential of attention-based, tokenized sequence models in geophysical forecasting and point to promising directions for extending LLM-based architectures in oceanographic applications.

Environmental magnetic analysis of profile NYQ-A from the second terrace of the Yellow River in Zoige Basin, Tibetan Plateau, provides insights into regional paleoenvironmental evolution. The results indicate that the magnetic minerals in these sediments are primarily composed of Pseudo-Single-Domain (PSD) magnetite and Single-Domain (SD) greigite. The paleo-deep lake deposits (A) represent sedimentation occurring before 51.82 f 2.34 ka, while paleo-deep lake deposits (B) formed between 39.18 f 2.03 and 36.77 f 1.66 ka. Higher values of chi, SIRM, chi ARM/chi, and chi ARM/SIRM indicate predominance of fine-grained magnetic minerals due to strong weathering in the watershed, suggesting a warm and humid climate with increased precipitation and meltwater. The paleo-shallow lake deposits (A) formed between 51.01 f 2.19 and 39.54 f 1.72 ka. Lower values of chi, chi ARM, SIRM, chi ARM/chi, and chi ARM/SIRM indicating dry and cold conditions that led to coarse-grained sediments with minimal weathering. The paleo-shallow lake deposits (B), formed between 36.73 f 2.04 ka and 31.44 f 1.97 ka, show increased hard magnetic minerals, reflecting enhanced Yellow River input of coarse sediments. Under the dual influence of a warm and humid climate and the neotectonics of the East Kunlun Fault, headward erosion of Yellow River has intensified, triggering a decline in paleo-lake water level. High magnetic minerals contents of two overbank flood deposits (OFD) suggest warm-climate weathering, while low chi ARM/chi and chi ARM/SIRM values indicate strong hydrodynamic conditions. The intensification of the East Asian Summer Monsoon and glacial melting initiated OFD1, which occurred between 30.43 f 3.76 ka and 27.47 f 3.57 ka, corresponding to MIS 3a period. OFD2 occurred between 15.30 f 1.04 ka and 12.93 f 1.63 ka, coinciding with the B & oslash;lling-Aller & oslash;d event. At these two periods, the water was mainly derived from the accelerated melting of mountain glaciers surrounding the basin and/or the large-scale precipitation, which led to the overbank floods. The research findings enhance the application of environmental magnetism in the study of river terrace sedimentary sequences.

Under the carbon neutrality framework, multiple coastal nuclear power plants in China have received construction approval. This development has drawn increased attention to the impact of thermal discharge on the marine environment. However, research on the diffusion effects caused by different thermal discharge configurations remains limited. This study focused on the Jinqimen Nuclear Power Plant. It employed the MIKE 3 (2014) three-dimensional numerical model, combined with field observations, to systematically investigate thermal plume dispersion. Specifically, it examined the effects of different jet angles at the discharge outlet (0°, 30°, 45°, 60°, 90°, and free diffusion conditions). The results indicate that the jet angle significantly influences the thermal rise envelope area and thermal stratification characteristics. Under free diffusion conditions (without jet velocity), the thermal rise area is the largest, with high-temperature zones concentrated near the surface. As the jet angle increases from 0° to 90°, the area of low-temperature rise gradually decreases, while the area of high-temperature rise expands. Among all tested configurations, the 30° jet angle exhibits the best overall performance. It demonstrates high thermal diffusion efficiency and strong heat dilution capacity. Moreover, it results in relatively smaller temperature rise areas at the surface, middle, and bottom layers. Additionally, tidal dynamics directly affect the thermal dispersion pattern. Smaller high-temperature rise areas are observed during peak flood and ebb tides. In contrast, heat accumulation is more likely to occur during slack tide periods. This study provides a scientific basis for optimizing the layout of nuclear power plant discharge outlets. It also serves as an important reference for mitigating thermal pollution and reducing ecological impacts of coastal nuclear power plants.