Real Ocean Research Articles

A Wave Prediction Framework Based on Machine Learning and the Third Generation Wave Model

Abstract To predict the evolution of wave spectrum in the real ocean, a machine learning framework is developed by training a long short-term memory (LSTM) neural network on a physics-based third-generation wave model (Simulating WAve Nearshore (SWAN)). Considering the realistic ocean waves are usually mixtures of windsea and swells, the wave spectrum is partitioned using a watershed algorithm, such that the windsea and swells are analyzed and predicted separately. Four parameters are selected to capture the wave spectrum of each system, including the significant wave height Hs, peaked wave period Tp, mean propagation direction θm, and directional spreading width σθ. The results demonstrate that the LSTM neural network can achieve accurate prediction of wave condition, the mean absolute error percentage (MAEPs) of 1-h prediction is less than 5.9%, 3.3%, 3.5%, and 3.3% for Hs, Tp, θm, and σθ, respectively, and accurate prediction of wave spectra is achieved. Even for the 10-h prediction, satisfactory results are obtained, e.g., the MAEP of Hs is less than 15.5%. The effects of output size (i.e., prediction duration), input data size (i.e., number of delays), as well as different combinations of input features on predictions of wave conditions are examined.

A comparative study of linear and step forcing temperature profiles in horizontal convection (a) (a)Contribution to the Focus Issue Turbulent Thermal Convection edited by Mahendra Verma and Jörg Schumacher.

We present a numerical study of horizontal convection using both step and linear forcing profiles as the imposed surface thermal driving, the former of which is commonly used in many studies but the latter is closer to the real ocean. Our investigation focuses on how forcing profiles can affect the theoretical and experimental results previously obtained based on the step forcing profile. We show that both types of forcing profiles exhibit similar behaviors in terms of global heat transfer (Nu), mean flow strength (Re), and boundary layer thickness. However, these two forcing profiles show significant differences for some other physical quantities, such as the generation rate of available potential energy G APE and global thermal dissipation rate, for which exact relationships with Nu can be derived for the step forcing profile but not for linear forcing. We show that a relationship between Nu and G APE can be developed for linear forcing by defining an effective forcing area, and such area is found to be approximately constant. We further derive an approximate relationship between Nu and the global thermal dissipation rate for linear forcing by assuming that the boundary heat flux is proportional to the local temperature deviation from the mean, which works best under the condition of moderate plume emissions. Differences are also found for the viscous dissipation rate; however, these differences diminish for sufficiently high Ra. Though step forcing is obviously less appropriate for modeling ocean circulations, we show that most differences arising from the forcing profile difference can be reconciled. Considering its convenience in experiment control and theoretical analysis, step forcing is a good configuration for most problems, but one should also be aware of the differences between step and other forcing types in quantities like the global thermal dissipation rate when extrapolating the results to more general situations.

Deep mixed ocean volume in the Labrador Sea in HighResMIP models

Simulations from seven global coupled climate models performed at high and standard resolution as part of the high resolution model intercomparison project (HighResMIP) are analyzed to study deep ocean mixing in the Labrador Sea and the impact of increased horizontal resolution. The representation of convection varies strongly among models. Compared to observations from ARGO-floats and the EN4 data set, most models substantially overestimate deep convection in the Labrador Sea. In four out of five models, all four using the NEMO-ocean model, increasing the ocean resolution from 1° to 1/4° leads to increased deep mixing in the Labrador Sea. Increasing the atmospheric resolution has a smaller effect than increasing the ocean resolution. Simulated convection in the Labrador Sea is mainly governed by the release of heat from the ocean to the atmosphere and by the vertical stratification of the water masses in the Labrador Sea in late autumn. Models with stronger sub-polar gyre circulation have generally higher surface salinity in the Labrador Sea and a deeper convection. While the high-resolution models show more realistic ocean stratification in the Labrador Sea than the standard resolution models, they generally overestimate the convection. The results indicate that the representation of sub-grid scale mixing processes might be imperfect in the models and contribute to the biases in deep convection. Since in more than half of the models, the Labrador Sea convection is important for the Atlantic Meridional Overturning Circulation (AMOC), this raises questions about the future behavior of the AMOC in the models.

The Role of the Gulf of Mexico Ocean Conditions in the Intensification of Hurricane Michael (2018)

AbstractHurricane Michael formed on October 7, 2018, in the Northwestern Caribbean Sea, and quickly traveled northward through the Gulf of Mexico, making landfall on the Florida panhandle as a devastating Category 5 hurricane only 3 days later. Before landfall, Michael underwent rapid intensification, despite unfavorable atmospheric conditions. Using observations, we characterized the key ocean features encountered by Michael along its track, which are known for favoring hurricane intensification: high sea surface temperatures, upper ocean heat content and low salinity barrier layer conditions. Ocean observations were consistent with suppressed hurricane‐induced upper ocean cooling, which could only be observed by underwater gliders, and showed that Hurricane Michael constantly experienced sea surface temperatures above 28°C. We carried out ocean Observing System Experiments, which demonstrate that the combined assimilation of in situ and satellite ocean observations into a numerical ocean model led to the most realistic representation of the ocean conditions. They also suggest that, when using the Cooper‐Haines (1996) method to assimilate altimetry observations, assimilating temperature observations is necessary to constrain the model upper ocean vertical structure. We also performed coupled hurricane‐ocean simulations to assess the impact of ocean initial conditions on forecasting Michael. These simulations demonstrate that the ocean conditions, in particular the high sea surface temperatures north of 24°N, played a crucial role in the intensification of Michael. Coupled simulations initialized with the most realistic ocean conditions, constrained by field and satellite observations, show a ∼56% error reduction in wind intensity prior to landfall compared to simulations initialized without data assimilation.

Are biodegradable plastics an environmental rip off?

While the use of biodegradable polymers is recognized as a global strategy to minimize plastic pollution, the technical standards (TS) used to attest their biodegradability may not be in compliance with most environmental parameters observed aquatic ecosystems. Indeed, through a careful assessment of the TS currently in use, this study evidenced that these guidelines cover only a fraction of the biogeochemical parameters seen in nature and largely disregard those that occur in the deep-sea. Thus, these TS may not be able to ensure the degradation of such polymers in natural environments, where microbial activity, pH, temperature, salinity, UV radiation and pressure are highly variable. This raises environmental concern, since relevant parcel of plastic ends up in the oceans reaching deep zones. Therefore, there is an urgent need to revise these TS, which must consider the actual fate of most plastic debris and include assessments under the challenging conditions found at these types of environment, alongside microplastic formation and ecotoxicology effects. Moreover, the next generation of biodegradability tests must be designed to enable a cost-effective implementation and incorporate accurate analytical techniques to assess polymer transformation. Furthermore, certification should provide information on time scale and degradation rates and, preferably, be globally harmonized.

Interannual to decadal variability of the Kuroshio extension: analyzing an ensemble of global hindcasts from a dynamical system viewpoint

The Kuroshio Extension (KE) is the inertial meandering jet formed by the convergence of the Kuroshio and Oyashio currents in the Northern Pacific. It is widely mentioned in the literature that the KE variability is bimodal on interannual to decadal time scale. The nature of this low frequency variability (LFV) is still under debate; intrinsic oceanic mechanisms are known to play a fundamental role in the phenomenon but there is also evidence from observations that the KE LFV is connected with changes in broader patterns associated with the Pacific Decadal Oscillation, which is in its turn generated by the dominant decadal mode of the sea level pressure variability in the North Pacific. We investigate the respective contributions of oceanic and atmospheric drivers of the KE variability by taking advantage of the OCCIPUT 1/4° global model dataset: it consists in an ensemble of 50 ocean–sea ice hindcasts performed over the period 1960–2015 (hereafter OCCITENS), and in a one-member 330-year climatological simulation (hereafter OCCICLIM). In this context, OCCITENS simulates both the intrinsic and forced variability, while OCCICLIM simulates the "pure" intrinsic variability. We explore several features of the KE, finding analogies between the OCCICLIM and OCCITENS datasets with autonomous and non-autonomous dynamical systems respectively. This approach aims to apply concepts from the dynamical systems theory on complex and realistic ocean simulations. In this framework, the results suggest that both oceanic and atmospheric drivers control the KE LFV, and that the effect of the low-frequency atmospheric forcing reduces the phase space region explored by the system through synchronization mechanisms. The system’s intrinsic variability can be paced, and therefore clustered over the system’s pullback attractor under the effect of the time dependent forcing.

The potential role of thermohaline–shear instability in turbulence production in the Bering Sea and the subarctic North Pacific

A recent linear stability analysis and a numerical simulation suggest that diffusive convection (DC), a regime of the double-diffusive convection, plays a potential role in onset of thermohaline–shear instability, implying that DC could contribute to turbulence production in the oceans. However, an existence of such a thermohaline–shear instability has not been examined in real oceans. We examine if this newly proposed instability mechanism exists in the subarctic North Pacific by analyzing our fine- and micro-scale turbulence measurement data. Vertical inversions were cautiously detected in seawater density profiles and used as a proxy for instability events for gradient Richardson number larger than the critical value of 1/4. We found that a portion of inversions were associated with active DC. Such DC-related inversions exhibited elevated levels of turbulent kinetic energy dissipation rate even for gradient Richardson number largely exceeding 1/4. Our estimate suggested that the thermohaline–shear instability contributes to roughly only 10% of the dissipation of turbulent kinetic energy in the diffusively convective layer in our observation site.

Using SAR imagery to survey internal solitary wave interactions: A case study off the Western Iberian shelf

Physical oceanography is increasingly relying on satellite remote sensing to survey the perpetually undersampled ocean, whereas the latest Synthetic Aperture Radars (SARs) are moving forward to provide a more continuous monitoring of the ocean. In this study we use a collection of SAR images to document the two-dimensional horizontal structure of Internal Solitary Waves (ISWs) propagating between two large submarine canyons off the Western Iberian Peninsula (between May and October 2018), which are observed to intersect approximately along the mid-shelf and originate a naturally-occurring interaction hotspot between different ISW packets. ISW interactions are well documented in theory and in laboratorial and numerical studies, but their observations in the real ocean are limited to airborne observations over the Strait of Georgia. The frequent SAR imagery of interacting ISWs in this region provides additional case studies to the literature, and we investigate if an energy proxy taken from their sea surface signatures can be used as an indicator for high-energy interaction events (e.g. when comparing with a non-interacting background). In particular, a quasi-synergetic event captured both in SAR and in a moored thermistor chain reveals that the often used weakly nonlinear theory for small-amplitude waves may underestimate the amplitudes measured in the waves’ interacting sections. ISWs provide the largest vertical displacements and velocities in the ocean. Understating how their vertical structure changes during wave-wave interactions may have important implications in the broader spectrum of ocean sciences, and SARs are shown in this study to be a first-approach tool to survey this frequent phenomenon in coastal regions.

Design and Implementation of Smart Ocean Visualization System Based on Extended Reality Technology

In the context of building a maritime power, building a smart ocean is one of the important means to promote ocean development. However, there is currently a lack of effective smart solutions for ocean development to integrate and manage ocean information. To solve the problem of insufficient development of smart ocean systems, a smart ocean visualization app based on extended reality technology has been developed, using Python crawler technology to collect ocean big data, and produce the system through Unity software. The app is built using C# programming, and the AR animation on the app is realized with the AR Foundation plug-in. Through terminals such as mobile phones or computers, it provides users with real-time ocean data query and expanded realistic ocean tourism services. The visualization function of the system realizes innovation in the way of querying marine data and makes up for the lack of development of smart marine apps.

Temporal unmixing-based cloud removal algorithm for optically complex water images

ABSTRACT Remote sensing has become an essential technique in water environment monitoring. However, the cloud contamination of remote sensing images breaks its spatial and temporal continuity in practical application. The current cloud removal algorithms always based on the hypothesis that the ground surface has no significant changes temporally, which is not proper for optically complex water body. Therefore, to overcome the weakness of current cloud removal algorithms on temporally fast-changing landscapes, a temporal unmixing-based cloud removal (TUBCR) algorithm is proposed to estimate the under-cloud information of optically complex water images. In this algorithm, the spatial information and temporal information were separated into temporal abundance images and temporal endmembers. By extracting the high temporal resolution endmembers, the fast-changing water information were estimated. The performance of the proposed algorithm was compared with those of two widely used cloud removal algorithms: the modified neighbourhood similar pixel interpolator (mNSPI) algorithm and the weighted linear regression (WLR) algorithm. The results of four simulated datasets (500 images in total) and two real Geostationary Ocean Colour Imager (GOCI) datasets (8 images in total) show that the proposed algorithm performed better in both quantitative indexes and visual effect compared with the mNSPI algorithm and the WLR algorithm. In satellite image time series, the proposed algorithm could accurately estimate under-cloud water information and correct the offsets of the statistical temporal trends that caused by cloud contamination. Hence, theoretically, the proposed TUBCR algorithm altered the similar pixel-based strategy of classic methods and got a satisfied performance in fast-changing water images. In applications, the TUBCR algorithm has great potential to improve the spatiotemporal coverage of optically complex water remote sensing.

Complexity of Mesoscale Eddy Diffusivity in the Ocean

AbstractStirring of water by mesoscale currents (“eddies”) leads to large‐scale transport of many important oceanic properties (“tracers”). These eddy‐induced transports can be related to the large‐scale tracer gradients, using the concept of turbulent diffusion. The concept is widely used to describe these transports in the real ocean and to represent them in climate models. This study focuses on the inherent complexity of the corresponding coefficient tensor (“K‐tensor”) and its components, defined here in all its spatio‐temporal complexity. Results demonstrate that this comprehensive K‐tensor is space‐, time‐, direction‐ and tracer‐dependent. Using numerical simulations with both idealized and comprehensive models of the Atlantic circulation, we show that these properties lead to upgradient eddy fluxes and the potential importance of all tensor components. The uncovered complexity of the eddy transports calls for reconsideration of how they are estimated in practice, included in the general circulation models and theoretically interpreted.

Seasonal variations in salinity of the North Pacific Intermediate Water and vertical mixing intensity over the Okinawa Trough

Seasonal variations in vertical mixing intensity in the intermediate layer of the northern Okinawa Trough was deduced using the seasonal variations in salinity in the North Pacific Intermediate Water (NPIW) using two distinct datasets: the World Ocean Atlas 2013 and historical conductivity, temperature, and depth data in the Okinawa Trough. Water mass analysis of the NPIW with a simple advection model showed that the vertical mixing intensity in the intermediate layer is enhanced in the winter at a rate of 20–50% of the annual mean amplitude value. To determine the driving force behind the seasonal variations in vertical mixing, we examined moored current-meter data in the northern Okinawa Trough during 2004–2011 and showed that the near-inertial internal wave energy in the intermediate layer was enhanced around January beneath the Kuroshio on the continental slope; this was due to strong southwestward flows associated with deep cyclonic eddies developing in the winter. We used density and velocity data from a realistic high-resolution ocean general circulation model simulation to hypothesize that the large-amplitude Kuroshio meander and associated deep cyclonic eddies in the winter generate near-inertial internal waves through centrifugal/inertial instability in the area of negative potential vorticity that appears not only beneath the Kuroshio near the western slope but also near the eastern slope of the trough. In addition, the symmetric instability along the offshore Kuroshio front and the wind energy input in the winter are suggested as possible sources of winter-enhanced near-inertial internal wave energy in the intermediate layer.

Analysis of ship maneuvering difficulties under severe weather based on onboard measurements and realistic simulation of ocean environment

The increasing demand for efficient, safe, and economic operation of ships has drawn attention to practical maneuvering behaviors for developing autonomous ships. Actual sailing conditions are reproduced to determine the relationship between environmental factors and ship steering records in rough seas. First, we generate realistic ocean environmental fields and analyze actual sea data. Then, we derive a modular maneuvering model reflecting environmental disturbances for further simulations. The correlation and multi regression analyses are performed based on measured data and environmental factors, which illustrate that the abnormal rudder angles are caused by reduced steering effectiveness. Finally, an attenuation function acting on the rudder normal forces is proposed to simulate this type of reduction. The time histories of maneuvering difficulties are selected as verification datasets. The maneuvers are simulated by adopting the attenuation function, and the simulation results show fair agreement with the measured data. The significant wave height, wind speed, mean wave period, current speed, wind apparent direction, and wave encounter angles are found to be the most statistically significant factors of rudder attenuation in the studied cases. The results and conclusions obtained from this study are of great significance for the further exploration of actual ship maneuvering behaviors in seas.

A Deep Learning Approach to Spatiotemporal Sea Surface Height Interpolation and Estimation of Deep Currents in Geostrophic Ocean Turbulence

AbstractSatellite altimeters provide global observations of sea surface height (SSH) and present a unique data set for advancing our theoretical understanding of upper‐ocean dynamics and monitoring its variability. Considering that mesoscale SSH patterns can evolve on timescales comparable to or shorter than satellite return periods, it is challenging to accurately reconstruct the continuous SSH evolution as currently available altimetry observations are still spatially and temporally sparse. Here we explore the possibility of SSH interpolation via Deep Learning by using synthetic observations from an idealized quasigeostrophic model of baroclinic ocean turbulence. We demonstrate that Convolutional Neural Networks with Residual Learning are superior in SSH reconstruction to linear and recently developed dynamical interpolation techniques. Also, the deep neural networks can provide a skillful state estimate of unobserved deep ocean currents at mesoscales. These conspicuous results suggest that SSH patterns of eddies might contain substantial information about the underlying deep ocean currents that are necessary for SSH prediction. Our training data are focused on highly idealized physics and diversification of processes needs to be considered to more accurately represent the real ocean. In addition, methodological improvements such as transfer learning and implementation of dynamically aware loss functions might be necessary to consider before its ultimate use with real satellite observations. Nonetheless, by providing a proof of concept based on synthetic data, our results point to deep learning as a viable alternative to existing interpolation and, more generally, state estimation methods for satellite observations of eddying currents.

Study on the Interaction of Nonlinear Water Waves considering Random Seas

AbstractThe nonlinear Schrödinger equation plays an important role in wave theory, nonlinear optics and Bose‐Einstein condensation. Depending on the background, different analytical solutions have been obtained. One of these solutions is the soliton solution. In the real ocean sea, interactions of different water waves can be observed at the surface. Therefore the question arises, how such nonlinear waves interact. Of particular interest is the interaction, also called collision, of solitons and solitary waves.Using a spectral scheme for the numerical computation of solutions of the nonlinear Schrödinger equation, the nonlinear wave interaction for the case of soliton collision is studied. Thereby, the influence of an initial random wave is studied, which is generated using a Pierson‐Moskowitz spectrum.

CFD Study on Penetration of Sonic Boom Applying a Wavy-Water Model

The penetration of sonic booms into the ocean and an estimation of the extent of their influence are important topics for supersonic and hypersonic aviation. Most prior studies were conducted assuming a flat-water surface. However, it is known that the effect of the surface waviness is not negligible in the case of real ocean environments. In the present study, the numerical method analyzes the two-dimensional flow field of sonic boom propagation across the air and water is proposed. When the flight Mach number of an aircraft is higher than 4.4, the underwater flow field becomes supersonic, and a sonic boom is generated there. The numerical results show that, in the underwater subsonic regime, the penetrating waves have the form of evanescent waves under either flat or wavy water surface, but the waviness-induced ripples with the smaller decay rate are responsible for deeper penetration. In contrast, in the supersonic underwater flow, the surface waviness hardly affects the penetration depth of underwater sonic booms, although it can slightly weaken the boom intensity. The effect of the amplitude of surface waviness on the underwater penetrating wave was also investigated.

High efficiency power extraction control method for direct‐drive linear wave energy converters based on forced resonance

It is commonly acknowledged that a direct-drive linear wave energy converter (DDLWEC) will achieve maximum power extraction in the resonance state, which can be realised through impedance matching under the assumption that the wave is regular. While the real ocean waves are hardly regular. Considering swell, a relatively regular type of real ocean wave, by approximating it with a sum of multiple frequency components, the DDLWEC can be forced resonance to the swell through velocity control, the optimal velocity is derived based on the dynamical model of the DDLWEC through frequency domain analysis. The stability of velocity control is proven and the float equilibrium position deviation due to the velocity control is analysed and countered through compensation control. In real-time control, the excitation force is approximated by the sum of four sinusoids, simulation and experiment results demonstrate that the proposed method is valid, stable, capable of achieving a high-power extraction efficiency, and superior to some existing methods.

The global biological microplastic particle sink

Every year, about four percent of the plastic waste generated worldwide ends up in the ocean. What happens to the plastic there is poorly understood, though a growing body of evidence suggests it is rapidly spreading throughout the global ocean. The mechanisms of this spread are straightforward for buoyant larger plastics that can be accurately modelled using Lagrangian particle models. But the fate of the smallest size fractions (the microplastics) are less straightforward, in part because they can aggregate in sinking marine snow and faecal pellets. This biologically-mediated pathway is suspected to be a primary surface microplastic removal mechanism, but exactly how it might work in the real ocean is unknown. We search the parameter space of a new microplastic model embedded in an earth system model to show that biological uptake can significantly shape global microplastic inventory and distributions and even account for the budgetary “missing” fraction of surface microplastic, despite being an inefficient removal mechanism. While a lack of observational data hampers our ability to choose a set of “best” model parameters, our effort represents a first tool for quantitatively assessing hypotheses for microplastic interaction with ocean biology at the global scale.

Autonomous boat driving system using sample‐efficient model predictive control‐based reinforcement learning approach

AbstractIn this article, we propose a novel reinforcement learning (RL) approach specialized for autonomous boats: sample‐efficient probabilistic model predictive control (SPMPC), to iteratively learn control policies of boats in real ocean environments without human prior knowledge. SPMPC addresses difficulties arising from large uncertainties in this challenging application and the need for rapid adaptation to dynamic environmental conditions, and the extremely high cost of exploring and sampling with a real vessel. SPMPC combines a Gaussian process model and model predictive control under a model‐based RL framework to iteratively model and quickly respond to uncertain ocean environments while maintaining sample efficiency. A SPMPC system is developed with features including quadrant‐based action search rule, bias compensation, and parallel computing that contribute to better control capabilities. It successfully learns to control a full‐sized single‐engine boat equipped with sensors measuring GPS position, speed, direction, and wind, in a real‐world position holding task without models from human demonstration.

Resolution dependency of sinking Lagrangian particles in ocean general circulation models.

Any type of non-buoyant material in the ocean is transported horizontally by currents during its sinking journey. This lateral transport can be far from negligible for small sinking velocities. To estimate its magnitude and direction, the material is often modelled as a set of Lagrangian particles advected by current velocities that are obtained from Ocean General Circulation Models (OGCMs). State-of-the-art OGCMs are strongly eddying, similar to the real ocean, providing results with a spatial resolution on the order of 10 km on a daily frequency. While the importance of eddies in OGCMs is well-appreciated in the physical oceanographic community, other marine research communities may not. Further, many long term climate modelling simulations (e.g. in paleoclimate) rely on lower spatial resolution models that do not capture mesoscale features. To demonstrate how much the absence of mesoscale features in low-resolution models influences the Lagrangian particle transport, we simulate the transport of sinking Lagrangian particles using low- and high-resolution global OGCMs, and assess the lateral transport differences resulting from the difference in spatial and temporal model resolution. We find major differences between the transport in the non-eddying OGCM and in the eddying OGCM. Addition of stochastic noise to the particle trajectories in the non-eddying OGCM parameterises the effect of eddies well in some cases (e.g. in the North Pacific gyre). The effect of a coarser temporal resolution (once every 5 days versus monthly) is smaller compared to a coarser spatial resolution (0.1° versus 1° horizontally). We recommend to use sinking Lagrangian particles, representing e.g. marine snow, microplankton or sinking plastic, only with velocity fields from eddying Eulerian OGCMs, requiring high-resolution models in e.g. paleoceanographic studies. To increase the accessibility of our particle trace simulations, we launch planktondrift.science.uu.nl, an online tool to reconstruct the surface origin of sedimentary particles in a specific location.