Hasil untuk "Ocean engineering"

Jie Yu, Xin Chen, Yi Lin et al.

Spectral variability and nonlinear mixing interactions critically degrade spectral unmixing accuracy, especially in heterogeneous environments. To address these challenges, this study proposes a robust nonlinear spectral variability-aware unmixing model, AD-HKFCM, which integrates fuzzy clustering, kernel-driven nonlinear mapping, and intraclass/interclass affinity cohesion. The model introduces a hybrid kernel function combining polynomial and radial basis kernels to enhance linear separability in high-dimensional space. By replacing conventional fuzzy c-means prototypes with support vector data description-derived hypersphere centers, the model reduces dependency on pure pixels and adaptively suppresses outliers through adaptive penalty weight optimization. A physics-informed affinity distance metric is designed to explicitly quantify spectral variability by penalizing intraclass dispersion and amplifying inter-class separation, thereby enabling the precise inference of “virtual pure endmembers” from intimately mixed data. Experiments on simulated (including Orchard 2EM/3EM benchmarks and synthetic hyperspectral) and real satellite datasets demonstrate that AD-HKFCM achieves 5–26% lower abundance estimation errors compared to the best-performing comparative methods, particularly in densely mixed regions with seasonal vegetation variability. This work unifies spectral variability compensation and nonlinear unmixing into a cohesive architecture, offering a generalizable solution for robust unmixing in complex environments.

Jiamin Xu, Guangyi Kan, Juan Wang et al.

Abstract The encapsulation of curcumin in the emulsions has attracted much attention in functional food development. Herein, the fish oil-loaded silver carp scale gelatin-stabilized emulsions with vitamins were explored for the delivery of curcumin. The curcumin encapsulation had no obvious effect on the formation, storage stability, lipid oxidation, and in vitro droplet digestion behaviors of the emulsions. Both vitamin C (VC) and vitamin E (VE) additions had obvious effects on the peroxidation values of the emulsions: VC + VE < VC < VE < Control < fish oil. The accumulative free fatty acid release percentages were dependent on the vitamins: VC (89.6 ± 1.1%) > Control (86.1 ± 1.0%) > VC + VE (80.5 ± 0.8%) > VE (76.4 ± 1.2%). The emulsions’ curcumin retention at room temperature and in vitro digestion behaviors (transformation, bioaccessibility, and bioaccessibility index) depended on vitamin additions: VC + VE ≈ VC > VE ≈ Control. This work was beneficial for the development of fish gelatin-stabilized emulsions to deliver curcumin.

Junyu Yan, Kai Luo, Bo Wang et al.

The current study was conducted to investigate the influences of dietary compound acidifier (CA, Formic acid (45 %), malic acid (0.3 %), Fumaric acid (0.6 %) and Succinic acid (0.6 %), Biomin Feed Additives (Shanghai) Ltd) on the growth, anti-oxidation, immunity, and intestinal health of channel catfish (Ictalurus punctatus). Six experimental diets (about 36 % crude protein and 5 % crude lipid) were prepared with graded levels of dietary CA (0 (control), 0.05 %, 0.10 %, 0.20 %, 0.40 %, and 0.80 %). A total of 648 fish with similar initial body weight (143.77 ± 0.73 g) were randomly placed into 18 tanks, with 36 fish in each tank. Fish were hand-fed to apparent satiation thrice daily for 72 days (water temperature was 25–27℃). The results showed that the WGR (weight giant rate) and SGR (specific growth rate) of channel catfish in the 0.10 % group were significantly increased compared with the control group (P < 0.05). The broken line model revealed that 0.0863 % of dietary CA was recommended for optimal SGR. The SOD (superoxide dismutase) activity and MDA (malondialdehyde) content in the serum in the 0.10 % groups were remarkably increased and decreased compared with the control group, respectively (P < 0.05). The ACP (acid phosphatase), LZM (lysozyme) activities and ALB (albumin) contents in the serum were significantly increased after supplementation of 0.10–0.40 %, 0.10–0.20 % and 0.40–0.80 % of dietary CA, respectively (P < 0.05). The AKP (alkaline phosphatase) activity in the serum in the 0.05 % and 0.10 % groups was decreased compared with other groups (P < 0.05). The GOT (glutamic oxalacetic transaminase) and GPT (glutamic pyruvic transaminase) activities in the serum notably decreased by 0.10–0.20 % and 0.05 % of dietary CA, respectively (P < 0.05). The T-CHO (total cholesterol) and HDL-C (high density lipoprotein cholesterol) contents in the serum increased by 0.80 % and 0.05 %-0.80 % of dietary CA (P < 0.05). The TG (triglyceride) content in the serum was lowest in the 0.10 % group. The expression levels of gh (growth hormone), ghr (growth hormone receptor), igf-1 (insulin-like growth factors 1), igf-2, igfbp-1 (insulin-like growth factor-binding proteins-1), igfbp-2 and igfbp-3 in the liver of the 0.10 % group were significantly higher than the control and 0.08 % groups (P < 0.05). The level of tgf-β (transforming growth factor beta), il-10 (interleukin 10), caspase3, caspase7, caspase9, occludin, zo-1 (zonula occludens-1), zo-2, claudin-12, claudin-15a, claudin-15b in the intestine of the 0.10 % group significantly increased compared with the control group (P < 0.05). Supplementation of 0.10 % dietary CA could improve the microbial diversity in the intestine of channel catfish. In conclusion, adding 0.10 % of dietary CA to the diet could improve the growth performance, anti-oxidative capacity, immunity, and intestinal health of channel catfish. Based on the SGR, the recommended optimal level of CA for channel catfish was 0.0863 %. The mechanisms regulating the growth-promoting and health-improving effects of CA on channel catfish need to be further investigated.

Junqi Cui, Shuyi Zhou, Guangjun Xu et al.

The increasingly severe issue of marine debris presents a critical threat to the sustainable development of marine ecosystems. Real-time detection is essential for timely intervention and cleanup. Furthermore, the density of marine debris exhibits significant depth-dependent variation, resulting in degraded detection accuracy. Based on 9625 publicly available underwater images spanning various depths, this study proposes UTNet, a lightweight neural model, to improve the effectiveness of real-time intelligent identification of marine debris through multidimensional optimization. Compared to Faster R-CNN, SSD, and YOLOv5/v8/v11/v12, the UTNet model demonstrates enhanced performance in random image detection, achieving maximum improvements of 3.5% in mAP50 and 9.3% in mAP50-95, while maintaining reduced parameter count and low computational complexity. The UTNet model is further evaluated on underwater videos for real-time debris recognition at varying depths to validate its capability. Results show that the UTNet model exhibits a consistently increasing trend in confidence levels across different depths as detection distance decreases, with peak values of 0.901 at the surface and 0.764 at deep-sea levels. In contrast, the other six models display greater performance fluctuations and fail to maintain detection stability, particularly at intermediate and deep depths, with evident false positives and missed detections. In summary, the lightweight UTNet model developed in this study achieves high detection accuracy and computational efficiency, enabling real-time, high-precision detection of marine debris at varying depths and ultimately benefiting mitigation and cleanup efforts.

Mao-Hsu Yen, Yih-Hsia Lin, Tzu-Feng Lin et al.

Multithreading is widely used in microcontroller unit (MCU) chips. Multithreaded hardware is composed of multiple identical single threads and provides instructions to different threads. Using the concept of thread-level parallelism (TLP), pauses are compensated for during single-thread operation to increase the throughput at the same unit. The principle of pipelined management is to use instruction-level parallelism (ILP) to split the MCU into multiple stages. When an instruction is given in a certain stage, other instructions are provided to operate in other idle stages and improve their execution efficiency. Based on the four-thread and pipelined RISC-V MCU architecture, we analyzed the instruction types of three benchmarks, i.e., Coremark, SHA, and Dijkstra. A total of 94% of the instructions use the arithmetic logic unit (ALU). Based on the executable four-thread architecture, we developed two to four RISC-V architectures with different numbers of ALUs and a dispatch algorithm. This architecture allows for the simultaneous delivery of multiple instructions, enabling parallel processing of instructions and increasing efficiency. Compared to the traditional RISC-V architecture with only one ALU, the test results showed that the instructions per clock (IPCs) of RISC-V architectures with two, three, and four ALUs increased efficiency by 76, 128.9, and 154.3%, while the area increased by 12, 22.3, and 32.6% and the static power consumption increased by 5.1, 9.2, and 13.3%. The results showed a significant improvement in performance with only a slight increase in the area. Due to the limited area of chips, a two-thread microcontroller architecture was used for the IC design and tape-out. TSMC’s 180nm process with a chip area of 1190 × 1190 μm at 133 MHz was used in this study.

Chaoming Bao, Jingqi Zhong, Daiyu Zhang et al.

The evaluation of steady turning performance for dual-tail propulsion underwater gliders typically relies on high-fidelity unsteady CFD simulations, which remain computationally prohibitive for control optimization and multi-scenario analysis. To overcome this limitation, this paper proposes a rapid prediction framework integrating Kriging surrogate modeling with dynamic equilibrium constraints. The proposed method employs a physics-informed decoupling strategy that isolates the hydrodynamic behavior of the hull from the thrust generation of the propellers. Since these are governed by distinct physical mechanisms and operate at different spatial scales, the decoupling strategy enables efficient and targeted steady-state CFD analysis for each component subsystem. Latin Hypercube Sampling (LHS) is used to generate training data for highly accurate Kriging models, which are subsequently coupled with the glider’s balance equations to form a bidirectional solution system. The forward mode predicts turning performance from control inputs, whereas the inverse mode identifies propeller speeds required for desired trajectories. Validation via fully-coupled 6-DOF unsteady CFD simulations confirms that the framework achieves prediction errors below 10% for key turning parameters while improving computational efficiency by over an order of magnitude. The method provides an effective tool for rapid maneuverability evaluation, control system design, and real-time path planning in dual-tail propulsion underwater gliders.

Peiyan Guo, Yuxin Sang, Fengyi Li et al.

The influence of traditional Chinese ritual culture on courtyard spatial sequences is widely acknowledged. However, quantitative analytical methods, such as space syntax, have rarely been applied in studies of ritual–residential space relations. This study uses space syntax, specifically Visibility Graph Analysis (VGA) and axial maps, to conduct a quantitative study of the spatial relationship between ritual and residential areas in Prince Kung’s Mansion. The VGA results indicate a distinct gradient of visual integration, which decreases progressively from the outward-oriented ritual areas, such as the palace gate and halls, through the transitional domestic ritual areas to the inward-oriented residential areas, such as Xijin Zhai and Ledao Tang. This pattern demonstrates a positive correlation between spatial visibility and ritual hierarchy. The axial map results confirm that the central axis and core ritual spaces exhibit the highest spatial connectivity, reflecting their supreme ritual status. More importantly, spatial connectivity is intensified during ritual activities compared to in daily life, indicating that enhanced spatial connectivity is required during rituals. Ritual spaces are characterized by extroversion, high visibility, and connectivity, while residential spaces prioritize introversion and minimal exposure. The deliberately designed ritual–residential architectural spatial sequence of Prince Kung’s Mansion articulates Confucian ideological principles, such as centrality as orthodoxy, gender segregation, and hierarchy. This study visually and quantitatively illustrates the harmony between ritual and residential spaces in Prince Kung’s Mansion. It enhances our understanding of the mechanisms of expression of courtyard ritual cultural spaces, providing evidence-based guidance for functional adaptive transformations in heritage conservation practices. It also offers a fresh perspective on the analysis of courtyard ritual spaces.

Yebao Wang, Wenhao Liu, Chuntao Chen et al.

Typhoons pose a significant threat to China’s coastal regions, resulting in substantial economic losses and casualties. Understanding the vulnerability of these areas to typhoon stress is crucial for effective disaster management and risk mitigation. This study assesses the vulnerability of China’s coastal provinces to typhoon disasters by integrating three key factors: exposure, sensitivity, and adaptability. The primary methodologies employed are the Analytic Hierarchy Process (AHP) and Geographic Information System (GIS) techniques. A comprehensive risk assessment framework is developed using 17 indicators, with AHP applied for indicator weighting and GIS used for spatial analysis and visualization of vulnerability patterns. The findings indicate considerable spatial variation in vulnerability, with southern provinces such as Guangdong, Guangxi, and Hainan exhibiting high vulnerability due to frequent typhoons, dense populations, and lower adaptive capacity. Southeastern regions, like Fujian and Zhejiang, show moderate to high vulnerability, while northern provinces such as Jiangsu, Hebei, and parts of Shandong and Liaoning experience lower vulnerability, attributed to reduced exposure and stronger disaster preparedness systems. These results underscore the importance of targeted disaster management strategies tailored to the specific vulnerabilities of each region.

Christopher R. Glein, Ngoc Truong

Enceladus offers our best opportunity for exploring the chemistry of an ocean on another world. Here, we perform geochemical modeling to show how the distribution of phosphate species found in ice grains from Enceladus's plume provides a very straightforward constraint on the pH of the host solution. The ratio of HPO$_4$/PO$_4$ species serves as a pH indicator. We find evidence of moderately alkaline water (pH 10.1-11.6)--significantly more alkaline than current estimates (~8-9) of the pH of Enceladus's ocean. Nevertheless, the pH range from phosphates is consistent with the CO$_2$/H$_2$O ratio measured in the plume if CO$_2$ exsolves from ocean water according to its equilibrium solubility. A simple energy balance can be used to quantify volatile fractionation during gas transport inside Enceladus's tiger stripes; we deduce that ~83% of water vapor is removed as ice during transport between the liquid-vapor interface and where gases exit the subsurface. We also explore how CO$_2$ degassing may lead to an increase in the apparent pH of ocean water. We generate maps of allowed combinations of pH and dissolved inorganic carbon concentration of the source water for a wide range of scenarios. Our preferred interpretation, constrained by the observed heat flux, implies minimal CO$_2$ degassing from ocean water. Hence, the pH recorded by phosphates should closely approximate that of the ocean; our best estimate is pH ~10.6. Such a high pH seems to reflect a major role of silicates enriched in Na, Mg, or Fe(II) interacting extensively with ocean water. Silica nanoparticles would not form or would subsequently dissolve if the pH is too high (>10.5). The outgassing model presented here provides a new path to quantify the dissolved concentrations of volatile species.

Davide Maestrini, Giovanni Dematteis, Alvise Benetazzo et al.

Ocean wind waves are a fundamental manifestation of complex dynamics in geophysical fluid systems, characterized by a rich interplay between dispersion and nonlinearity. While linear wave theory provides a first-order description of wave motion, real-world oceanic environments are governed by nonlinear interactions that are responsible for a transfer of energy between waves of different lengths. The pioneering research by O. Phillips, K. Hasselmann, and V. Zakharov in the 1960s introduced the concept that four-wave resonant interactions serve as the primary mechanism for energy transfers among wave components in oceanic surface wave fields. Although the presence and efficiency of these resonant interactions have been demonstrated in controlled determinist wave tank experiments, their direct identification in the real ocean, where a large number of random waves interact, has remained elusive. Here, using a stereoscopic system that enables the measurement of surface elevation in both space and time, we provide experimental evidence of resonant interactions in ocean wind waves. Our data not only reproduce the well-known figure-eight pattern predicted by Phillips, but also reveal a continuum of different resonant configurations that closely match the theoretical predictions. These findings support the validity of third-generation ocean wave models, strengthening their ability to accurately capture wave dynamics in the ocean

Baoshan WANG, Qiang SHANG, Cheng MAN

Selective laser melting (SLM) is a powder-bed metal additive manufacturing technology that is extensively employed in the fields of marine engineering, biomedicine, and nuclear power due to its high processing precision and wide range of applicable materials. 316L stainless steel is one of the metal materials that have been researched earlier and has a more mature process in the field of SLM. Although SLM technology processing of 316L stainless steel parts (later referred to as SLM-316L stainless steel) has been conducted for industrial applications, it is rarely utilized for high temperatures, strong corrosion, complex loads, and other demanding conditions. Nonequilibrium solidification in the laser melt pool is an inherent mechanism of the SLM-316L stainless steel forming process, which contributes to the production of a nonuniform organizational structure and a high level of residual stress, which influence the reliability of SLM-316L stainless steel in long-term service. Heat treatment after preparation of SLM-316L stainless steel is the most effective approach to optimizing the organizational structure and reducing residual stress. SLM of 316L stainless steel is often employed for solid solution treatment to optimize the organization and reduce residual stresses to yield remarkable overall performance. The intergranular corrosion behavior of austenitic stainless steel highly depends on its organizational structure; thus, solid solution treatment is bound to enhance the intergranular corrosion performance of SLM-316L stainless steel. However, the law and mechanism of the effect of solid solution treatment on the intergranular corrosion behavior of SLM-316L stainless steel is still vague. Based on the mentioned above, in this work, solid solution treatment of SLM-316L stainless steel is conducted at 1150 ℃, its organizational and structural characteristics and morphology of nanooxidized particles are examined by Scanning electron microscope (SEM), Electron backscattered diffraction (EBSD), and Transmission electron microscope (TEM), and its intergranular corrosion behavior is investigated by double-loop electrochemical reactivation and ammonium persulfate electrolysis tests. The following conclusions can be drawn. (1) Recrystallization of SLM-316L stainless steel takes place after solid solution treatment, forming regularly shaped equiaxed grains and annealed twin crystals. (2) The nanooxidized particles are coarsened, and the maximum size at grain boundaries can attain the micrometer level. Meanwhile, the type of oxide particle also transforms from the rhodochrosite structure of MnSiO3 to the spinel structure of CrMn2O4. (3) Solid solution treatment results in a decrease in intergranular corrosion performance of SLM-316L stainless steel, together with a decrease in intergranular corrosion performance, which, in turn, is accompanied by sensitization time extension, and the type of intergranular corrosion changes from step-like to groove-like.

Yan Zhang, Rongbo Fan, PeiPei Duan et al.

Remote sensing image spatiotemporal fusion (STF) aims to generate composite images with high-temporal and spatial resolutions by combining remote sensing images captured at different times and with different spatial resolutions (DTDS). Among the existing fusion algorithms, deep learning-based fusion models have demonstrated outstanding performance. These models treat STF as an image super-resolution problem based on multiple reference images. However, compared to traditional image super-resolution tasks, remote sensing image STF involves merging a larger amount of multitemporal data with greater resolution difference. To enhance the robust matching performance of spatiotemporal transformations between multiple sets of remote sensing images captured at DTDS and to generate super-resolution composite images, we propose a feature fusion network called the multiscale deformable convolution distillation generative adversarial network (DCDGAN-STF). Specifically, to address the differences in multitemporal data, we introduce a pyramid cascading deformable encoder to identify disparities in multitemporal images. In addition, to address the differences in spatial resolution, we propose a teacher&#x2013;student correlation distillation method. This method uses the texture details&#x0027; disparities between high-resolution multitemporal images to guide the extraction of disparities in blurred low-resolution multitemporal images. We comprehensively compared the proposed DCDGAN-STF with some state-of-the-art algorithms on two landsat and moderate-resolution imaging spectroradiometer datasets. Ablation experiments were also conducted to test the effectiveness of different submodules within DCDGAN-STF. The experimental results and ablation analysis demonstrate that our algorithm achieves superior performance compared to other algorithms.

Yuting Zhu, Giuseppe Grieco, Jiarong Lin et al.

The accurate retrieval of sea-surface wind field data is crucial for weather forecasting and climate modeling. Despite this, the complexity of sea surface conditions poses significant challenges for satellite-based synthetic aperture radar (SAR) wind retrieval techniques. This study introduces a Bayesian inversion algorithm that incorporates azimuth cutoff wavelength information&#x2014;a parameter previously underutilized and highly sensitive to varying wind conditions. We aimed to enhance the accuracy of SAR-derived wind estimations to enable more reliable interpretations of marine atmospheric dynamics. The methodology probabilistically combines SAR data with ancillary meteorological information and optimizes the retrieval process through a cost function that leverages the sensitivity of the azimuth cutoff to changes in wind vector fields. The proposed method was comprehensively validated using Sentinel-1 and Gaofen-3 SAR datasets against buoy measurements and wind estimations from scatterometers. The results demonstrated that the proposed method significantly improved the accuracy of wind speed estimations, especially under low-wind conditions and different sea-state conditions, without substantially increasing the computational burden. Although the wind direction retrieval displayed limited enhancement, the improved accuracy in wind speed estimations provides considerable benefits for operational meteorological applications. These findings suggest that the integration of azimuth cutoff information could be a critical step toward obtaining more accurate and reliable wind field retrievals from SAR data, thereby advancing the field of remote sensing and oceanography.

Layrson de Jesus Menezes Gonçalves, Júlia Kaiser, Ronaldo Maia de Jesus Palmeira et al.

This study assesses the performance of the Weather Research and Forecasting (WRF) model using a high-resolution spatial grid (1 km) with various combinations of physical parameterization packages to simulate a severe event in August 2021 in the southeastern Brazilian coast. After determining the optimal set of physical parameterizations for representing wind patterns during this event, a year-long evaluation was conducted, covering forecast horizons of 24, 48, and 72 h. The simulation results were compared with observational wind data from four weather stations. The findings highlight variations in the efficacy of different physical parameterization sets, with certain sets encountering challenges in accurately depicting the peak of the severe event. The most favorable results were achieved using a combination of Tiedtke (cumulus), Thompson (microphysics), TKE (boundary layer), Monin-Obukhov (surface layer), Unified-NOAH (land surface), and RRTMG (shortwave and longwave radiation). Over the one-year forecasting period, the WRF model effectively represented the overall wind pattern, including forecasts up to three days in advance (72-h forecast horizon). Generally, the statistical metrics indicate robust model performance, even for the 72-h forecast horizon, with correlation coefficients consistently exceeding 0.60 at all analyzed points. While the model proficiently captured wind distribution, it tended to overestimate northeast wind speed and gust intensities. Notably, forecast accuracy decreased as stations approached the ocean, exemplified by the ATPM station.

Amita Biswal, Dah-Jing Jwo

One technique that is widely used in various fields, including nonlinear target tracking, is the extended Kalman filter (EKF). The well-known minimum mean square error (MMSE) criterion, which performs magnificently under the assumption of Gaussian noise, is the optimization criterion that is frequently employed in EKF. Further, if the noises are loud (or heavy-tailed), its performance can drastically suffer. To overcome the problem, this paper suggests a new technique for maximum correntropy EKF with nonlinear regression (MCCEKF-NR) by using the maximum correntropy criterion (MCC) instead of the MMSE criterion to calculate the effectiveness and vitality. The preliminary estimates of the state and covariance matrix in MCKF are provided via the state mean vector and covariance matrix propagation equations, just like in the conventional Kalman filter. In addition, a newly designed fixed-point technique is used to update the posterior estimates of each filter in a regression model. To show the practicality of the proposed strategy, we propose an effective implementation for positioning enhancement in GPS navigation and radar measurement systems.

Wonjun Jo, Insu Woo, Yoshiki Mikami et al.

This study examines the residual stress characteristics of spot welding in newly developed high-strength steel for automotive body construction through experimental and numerical methods. The effects of sheet thickness, nugget size, and the presence or absence of spacers on residual stress distribution and fracture stability were evaluated. Measurements using XRD and HDR revealed tensile residual stress below the yield strength at the nugget center. A numerical analysis system corroborated experimental findings, demonstrating that larger nugget sizes reduce tensile residual stress at the nugget center, enhancing fracture stability. However, for nugget sizes of 3<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msqrt><mi>t</mi></msqrt></mrow></semantics></math></inline-formula> (<i>t</i>: thickness), high tensile stress at the nugget edge compromised stability, while sizes of 3.5<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msqrt><mi>t</mi></msqrt></mrow></semantics></math></inline-formula> or larger improved fracture resistance. The study also found that thicker sheets increased fracture safety with larger nugget sizes, and the presence of spacers induced tensile stress through spring-back effects, which shifted to compressive stress as the nugget size increased. These results provide critical insights into optimizing welding parameters to improve the structural integrity of automotive components.

Bruno Buongiorno Nardelli, Daniele Iudicone

Revealing the ongoing changes in ocean dynamics and their impact on marine ecosystems requires the joint analysis of multiple variables. Yet, global observational records only cover a few decades, posing a challenge in the separation of climatic trends from internal dynamical modes. Here, we apply an empirical stochastic model to identify the emergent patterns of trends in six fundamental components of upper ocean physics. We analyze a data-driven reconstruction of the ocean state covering the 1993-2018 period. We found that including temporal derivatives into the state vector enhances the description of the ocean's dynamical system. Once Pacific oscillations are properly accounted for, averaged surface warming appears >60% faster, and a deep reshaping of the seascape is revealed. A clustering of the trend patterns identifies the main factors that drive observed trends in chlorophyll-a concentration. This data-driven approach opens new perspectives in empirical climate modelling.

Anass El Aouni, Quentin Gaudel, Charly Regnier et al.

Accurate ocean forecasting is crucial in different areas ranging from science to decision making. Recent advancements in data-driven models have shown significant promise, particularly in weather forecasting community, but yet no data-driven approaches have matched the accuracy and the scalability of traditional global ocean forecasting systems that rely on physics-driven numerical models and can be very computationally expensive, depending on their spatial resolution or complexity. Here, we introduce GLONET, a global ocean neural network-based forecasting system, developed by Mercator Ocean International. GLONET is trained on the global Mercator Ocean physical reanalysis GLORYS12 to integrate physics-based principles through neural operators and networks, which dynamically capture local-global interactions within a unified, scalable framework, ensuring high small-scale accuracy and efficient dynamics. GLONET's performance is assessed and benchmarked against two other forecasting systems: the global Mercator Ocean analysis and forecasting 1/12 high-resolution physical system GLO12 and a recent neural-based system also trained from GLORYS12. A series of comprehensive validation metrics is proposed, specifically tailored for neural network-based ocean forecasting systems, which extend beyond traditional point-wise error assessments that can introduce bias towards neural networks optimized primarily to minimize such metrics. The preliminary evaluation of GLONET shows promising results, for temperature, sea surface height, salinity and ocean currents. GLONET's experimental daily forecast are accessible through the European Digital Twin Ocean platform EDITO.

Daniel Holmberg, Emanuela Clementi, Teemu Roos

Accurate ocean forecasting systems are vital for understanding marine dynamics, which play a crucial role in environmental management and climate adaptation strategies. Traditional numerical solvers, while effective, are computationally expensive and time-consuming. Recent advancements in machine learning have revolutionized weather forecasting, offering fast and energy-efficient alternatives. Building on these advancements, we introduce SeaCast, a neural network designed for high-resolution, medium-range ocean forecasting. SeaCast employs a graph-based framework to effectively handle the complex geometry of ocean grids and integrates external forcing data tailored to the regional ocean context. Our approach is validated through experiments at a high spatial resolution using the operational numerical model of the Mediterranean Sea provided by the Copernicus Marine Service, along with both numerical and data-driven atmospheric forcings.

Harrison Nicholls, Tim Lichtenberg, Dan J. Bower et al.

Interactions between magma oceans and overlying atmospheres on young rocky planets leads to an evolving feedback of outgassing, greenhouse forcing, and mantle melt fraction. Previous studies have predominantly focused on the solidification of oxidized Earth-similar planets, but the diversity in mean density and irradiation observed in the low-mass exoplanet census motivate exploration of strongly varying geochemical scenarios. We aim to explore how variable redox properties alter the duration of magma ocean solidification, the equilibrium thermodynamic state, melt fraction of the mantle, and atmospheric composition. We develop a 1D coupled interior-atmosphere model that can simulate the time-evolution of lava planets. This is applied across a grid of fixed redox states, orbital separations, hydrogen endowments, and C/H ratios around a Sun-like star. The composition of these atmospheres is highly variable before and during solidification. The evolutionary path of an Earth-like planet at 1 AU ranges between permanent magma ocean states and solidification within 1 Myr. Recently solidified planets typically host H2O- or H2-dominated atmospheres in the absence of escape. Orbital separation is the primary factor determining magma ocean evolution, followed by the total hydrogen endowment, mantle oxygen fugacity, and finally the planet's C/H ratio. Collisional absorption by H2 induces a greenhouse effect which can prevent or stall magma ocean solidification. Through this effect, as well as the outgassing of other volatiles, geochemical properties exert significant control over the fate of magma oceans on rocky planets.