In this paper, we analyze the predictability of the ocean currents using deep learning. More specifically, we apply the Long Short Term Memory (LSTM) deep learning network to a data set collected by the National Oceanic and Atmospheric Administration (NOAA) in Massachusetts Bay between November 2002-February 2003. We show that the current speed in two horizontal directions, namely u and v, can be predicted using the LSTM. We discuss the effect of training data set on the prediction error and on the spectral properties of predictions. Depending on the temporal or the spatial resolution of the data, the prediction times and distances can vary, and in some cases, they can be very beneficial for the prediction of the ocean current parameters. Our results can find many important applications including but are not limited to predicting the tidal energy variation, controlling the current induced vibrations of marine structures and estimation of the wave blocking point by the chaotic oceanic current and circulation.
Significant salinity anomalies have been observed in the Arctic Ocean surface layer during the last decade. Using gridded data of winter salinity in the upper 50 m layer of the Arctic Ocean for the period 1950-1993 and 2007-2012, we investigated the inter-annual variability of the salinity fields, attempted to identify patterns and anomalies, and developed a statistical model for the prediction of surface layer salinity. The statistical model is based on linear regression equations linking the principal components with environmental factors, such as atmospheric circulation, river runoff, ice p…
The Biogeochemical-Argo (BGC-Argo) program is building a network of globally distributed, sensor-equipped robotic profiling floats, improving our understanding of the climate system and how it is changing. These floats, however, are limited in the number of variables measured. In this study, we train neural networks to predict silicate and phosphate values in the Southern Ocean from temperature, pressure, salinity, oxygen, nitrate, and location and apply these models to earth system model (ESM) and BGC-Argo data to expand the utility of this ocean observation network. We trained our neural net…
We relate about the custom-made modification of a Benthos deep-ocean acoustic release into a deep caller, a hydrophone for calling ocean acoustic releases that cannot be reached from a standard deck-unit. The self-contained deep caller can be lowered down to 12 km on any non-conducting winch cable. It may prove useful to retrieve sub-surface instrumentation like a seafloor-lander hidden behind large rocks or in a narrow canyon, or moorings in very deep topographic depressions. We used it to retrieve a 7 km long mooring from 10,910 m depth in the Challenger Deep, Mariana Trench.
Munk (1966) showed that the deep (1000-3000 m) vertical temperature profile is consistent with a one-dimensional vertical advection-diffusion balance, with a constant upwelling and an interior diapycnal diffusivity of $\mathcal{O}(10^{-4})$ m$^{2}$ s$^{-1}$. However, typical observed diffusivities in the interior are $\mathcal{O}(10^{-5})$ m$^{2}$ s$^{-1}$. Recent work suggested that the deep stratification is set by Southern Ocean (SO) isopycnal slopes, fixed by SO eddies, that communicate the surface outcrop positions to the deep ocean. It is shown here, using an idealized ocean general circ…
In this paper we extend earlier results regarding the effects of the lower layer of the ocean (below the thermocline) on the baroclinic instability within the upper layer (above the thermocline). We confront quasigeostrophic baroclinic instability properties of a 2.5-layer model with those of a 3-layer model with a very thick deep layer, which has been shown to predict spectral instability for basic state parameters for which the 2.5-layer model predicts nonlinear stability. We compute and compare maximum normal-mode perturbation growth rates, as well as rigorous upper bounds on the nonlinear …
Naval Research Laboratory（海军研究实验室）；University of New Orleans（新奥尔良大学）
In this paper, we present a novel approach for the prediction of rogue waves in oceans using statistical machine learning methods. Since the ocean is composed of many wave systems, the change from a bimodal or multimodal directional distribution to unimodal one is taken as the warning criteria. Likewise, we explore various features that help in predicting rogue waves. The analysis of the results shows that the Spectral features are significant in predicting rogue waves. We find that nonlinear classifiers have better prediction accuracy than the linear ones. Finally, we propose a Random Forest …
The impact of large atmospheric disturbances on deep benthic communities is not well known quantitatively. Observations are scarce but may reveal specific processes leading to turbulent disturbances. Here, we present high-resolution deep-ocean observations to study potential turbulent mixing by large atmospheric disturbances. We deployed an array of 100-Hz sampling-rate geophysical broadband Ocean Bottom Seismometers (OBSs) on the seafloor. Within the footprint of this array we also deployed an oceanographic 1-Hz sampling-rate vertical temperature sensor string covering the water phase between…
The Atlantic Meridional Overturning Circulation (MOC) is a crucial part of the climate system because of its associated northward heat transport. The present-day MOC is sensitive to freshwater anomalies and may collapse to a state with a strongly reduced northward heat transport. A future collapse of the Atlantic MOC has been identified as one of the most dangerous tipping points in the climate system. It is therefore crucial to develop early warning indicators for such a potential collapse based on relatively short time series. So far, attempts to use indicators based on critical slowdown hav…
Data-driven Global Ocean Modeling for Seasonal to Decadal Prediction
School of Computer Science, Fudan University（复旦大学计算机科学学院）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；Potsdam Institute for Climate Impact Research（波茨坦气候影响研究所）；Japan Agency for Marine-Earth Science and Technology（日本海洋研究开发机构）；Institut de Recherche pour le Développement (IRD)（法国发展研究院）；French National Research Institute for Sustained Development (IRD)（法国可持续发展研究院）；Cooperative Institute for Research in Environmental Sciences（合作环境科学研究所）；NOAA Physical Sciences Laboratory（美国国家海洋和大气管理局物理科学实验室）；Nanjing University of Information Science and Technology（南京信息工程大学）；Technical University of Munich（慕尼黑工业大学）；University of Colorado Boulder（科罗拉多大学博尔德分校）
We present a comprehensive inter-comparison of linear regression (LR), stochastic, and deep-learning approaches for reduced-order statistical emulation of ocean circulation. The reference dataset is provided by an idealized, eddy-resolving, double-gyre ocean circulation model. Our goal is to conduct a systematic and comprehensive assessment and comparison of skill, cost, and complexity of statistical models from the three methodological classes. The model based on LR is considered as a baseline. Additionally, we investigate its additive white noise augmentation and a multi-level stochastic app…
The inertial subrange of turbulence in a density stratified environment is the transition from internal waves to isotropic turbulence, but it is unclear how to interpret its extension to anisotropic stratified turbulence. Knowledge about stratified turbulence is relevant for the dispersal of suspended matter in geophysical flows, such as in most of the ocean. For studying internal-wave-induced ocean-turbulence moored high-resolution temperature (T-)sensors are used. Spectra from observations on episodic quasi-convective internal wave breaking above a steep slope of large seamount Josephine in …
A general method for the derivation of asymptotic nonlinear shallow water and deep water models is presented. Starting from a general dimensionless version of the water-wave equations, we reduce the problem to a system of two equations on the surface elevation and the velocity potential at the free surface. These equations involve a Dirichlet-Neumann operator and we show that all the asymptotic models can be recovered by a simple asymptotic expansion of this operator, in function of the shallowness parameter (shallow water limit) or the steepness parameter (deep water limit). Based on this met…
The Pacific Ocean Neutrino Experiment (P-ONE) is a proposed cubic-kilometer scale neutrino telescope planned to be installed in the deep-sea of the north-east Pacific Ocean. In collaboration with the optical deep-sea data and communications network operated by Ocean Networks Canada, an international collaboration of researchers plans to install an array of kilometer-long mooring lines in a depth of around 2660 m to the relatively flat deep-sea region called Cascadia Basin, around 300 miles West of Vancouver Island. With the design and development ongoing, the P-ONE collaboration is interested …
Deep ocean acoustics, in the absence of shipping and wildlife, is driven by surface processes. Best understood is the signal generated by non-linear surface wave interactions, the Longuet-Higgins mechanism, which dominates from 0.1 to 10 Hz, and may be significant for another octave. For this source, the spectral matrix of pressure and vector velocity is derived for points near the bottom of a deep ocean resting on an elastic half-space. In the absence of a bottom, the ratios of matrix elements are universal constants. Bottom effects vitiate the usual "standing wave approximation," but a weake…
Evaluation of Deep Neural Operator Models toward Ocean Forecasting
Department of Mechanical Engineering, Center for Computational Science and Engineering（机械工程系，计算科学与工程中心）；Department of Electrical Engineering and Computer Science（电子工程与计算机科学系）；Massachusetts Institute of Technology, Cambridge, MA（麻省理工学院）
Simulating oil transport in the ocean can be done successfully provided that accurate ocean currents and surface winds are available -- this is often too big of a challenge. Deficient ocean currents can sometimes be remediated by parameterizing missing physics -- this is often not enough. In this chapter, we focus on some of the main problems oil-spill modelers face, which is determining accurate trajectories when the velocity may be missing important physics, or when the velocity has localized errors that result in large trajectory errors. A foundation of physical mechanisms driving motion in…
We use numerical climate simulations, paleoclimate data, and modern observations to study the effect of growing ice melt from Antarctica and Greenland. Meltwater tends to stabilize the ocean column, inducing amplifying feedbacks that increase subsurface ocean warming and ice shelf melting. Cold meltwater and induced dynamical effects cause ocean surface cooling in the Southern Ocean and North Atlantic, thus increasing Earth's energy imbalance and heat flux into most of the global ocean's surface. Southern Ocean surface cooling, while lower latitudes are warming, increases precipitation on the …
We present experimental and computational studies of the propagation of internal waves in a stratified fluid with an exponential density profile that models the deep ocean. The buoyancy frequency profile $N(z)$ (proportional to the square root of the density gradient) varies smoothly by more than an order of magnitude over the fluid depth, as is common in the deep ocean. The nonuniform stratification is characterized by a turning depth $z_c$, where $N(z_c)$ is equal to the wave frequency $ω$ and $N(z < z_c) < ω$. Internal waves reflect from the turning depth and become evanescent below the tur…
Deep Learning Weather Models for Subregional Ocean Forecasting: A Case Study on the Canary Current Upwelling System
Oceanografía Física y Geofísica Aplicada (OFYGA), ECOAQUA（应用物理海洋学与地球物理系（OFYGA）、ECOAQUA）；Centro de Tecnologías de la Imagen (CTIM), IUCES（图像技术中心（CTIM）、IUCES）；Universidad de Las Palmas de Gran Canaria（拉斯帕尔马斯大大学）
FuXi-Ocean: A Global Ocean Forecasting System with Sub-Daily Resolution
Artificial Intelligence Innovation and Incubation Institute, Fudan University（复旦大学人工智能创新与孵化院）；Shanghai Innovation Institute（上海创新研究院）；School of Marine Science and Technology,Tianjin University（天津大学海洋科学与技术学院）
Ocean dynamics constitute a source of incertitude in determining the ocean's role in complex climatic phenomena. Current observation systems have limitations in achieving sufficiently statistical precision for three-dimensional oceanic data. It is crucial knowledge to describe the behavior of internal ocean structures. We present the data-driven approaches which explore latent class regressions and deep regression neural networks in modeling ocean dynamics in the extensions of Gulf Stream and Kuroshio currents. The obtained results show a promising data-driven direction for understanding the o…
Simulation-informed deep learning for enhanced SWOT observations of fine-scale ocean dynamics
IMT Atlantique（大西洋高等电信学院）；CNRS（法国国家科学研究中心）；INRIA（法国国家信息与自动化研究所）；Ifremer（法国海洋开发研究院）；Jet Propulsion Laboratory（喷气推进实验室）；California Institute of Technology（加州理工学院）
IMT Atlantique（IMT大西洋学院）；UBL（布列塔尼卢瓦尔大学）；Lab-STICC（STICC实验室）；National Marine Data and Information Service（国家海洋信息中心）；Ocean University of China（中国海洋大学）；Mediterranean Institute for Advanced Studies (IMEDEA/CSIC-UIB)（地中海高等研究院（IMEDEA/CSIC-UIB））
This work presents EddyNet, a deep learning based architecture for automated eddy detection and classification from Sea Surface Height (SSH) maps provided by the Copernicus Marine and Environment Monitoring Service (CMEMS). EddyNet is a U-Net like network that consists of a convolutional encoder-decoder followed by a pixel-wise classification layer. The output is a map with the same size of the input where pixels have the following labels \{'0': Non eddy, '1': anticyclonic eddy, '2': cyclonic eddy\}. We investigate the use of SELU activation function instead of the classical ReLU+BN and we use…
We present a detailed analysis of the temperature statistics in an oceanographic observational dataset. The data are collected using a moored array of thermistors, 100 m tall and starting 5 m above the bottom, deployed during four months above the slopes of a Seamount in the north-eastern Atlantic Ocean. Turbulence at this location is strongly affected by the semidiurnal tidal wave. Mean stratification is stable in the entire dataset. We compute structure functions, of order up to 10, of the distributions of temperature increments. Strong intermittency is observed, in particular, during the do…
The Southern Ocean plays an integral role in the global climate system, exchanging heat, salt, and carbon throughout the major ocean basins via the deep, fast-flowing Antarctic Circumpolar Current. The Antarctic Circumpolar Current is bounded by spatially and temporally varying fronts that partition distinct water masses. Locating and quantifying cross-front transport is crucial for understanding global patterns of inter-basin exchange; however, this is challenging because fronts are typically defined by hydrographic properties, such as temperature or salinity, which are subject to external so…
Samudra 2: Scaling Ocean Emulators across Resolutions
Courant Institute School of Mathematics, Computing, and Data Science, New York University（Courant学院数学、计算与数据科学系，纽约大学）；Open Athena AI Foundation, Inc.（开放Athena人工智能基金会）；Program in Atmospheric and Oceanic Sciences, Princeton University（大气与海洋科学项目，普林斯顿大学）
During the last ice age several quasi-periodic abrupt warming events took place. Known as Dansgaard-Oeschger (DO) events their effects were felt globally, although the North Atlantic experienced the largest temperature anomalies. Paleoclimate data shows that the fluctuations often occurred right after massive glacial meltwater releases in the North Atlantic and in bursts of three or four with progressively decreasing strengths. In this study a simple dynamical model of an overturning circulation and sea ice is developed with the goal of understanding the fundamental mechanisms that could have …
Data-driven global ocean model resolving ocean-atmosphere coupling dynamics
Center for Climate and Carbon Cycle Research, Korea Institute of Science and Technology, Seoul, Republic of Korea（韩国科学技术院气候与碳循环研究中心，首尔，大韩民国）；Department of Environment and Energy, Jeonbuk National University, Jeonju, Republic of Korea（全南国立大学环境与能源系，全州，大韩民国）；School of Earth and Environmental Sciences, Seoul National University, Seoul, Republic of Korea（首尔国立大学地球与环境科学学院，首尔，大韩民国）；Department of Environmental Management, Seoul National University, Seoul, Republic of Korea（首尔国立大学环境管理系，首尔，大韩民国）
Generative Lagrangian data assimilation for ocean dynamics under extreme sparsity
Department of Applied Mathematics, University of California, Santa Cruz, CA 95064, USA（应用数学系，加州大学圣克鲁兹分校）；Department of Electrical and Computer Engineering, University of California, Santa Cruz, CA 95064, USA（电气与计算机工程系，加州大学圣克鲁兹分校）；Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695, USA（海洋、地球与大气科学系，北卡罗来纳州立大学）
Rainy years and dry years in SE Australia are known to be correlated with sea surface temperatures in the specific areas of the Indian Ocean. While over the past 100 years the correlation had been both positive and negative, it significantly increased in strength since the beginning of the 21st century. Over this period, Indian Ocean sea surface temperatures during the winter months, together with the El Niño variations, contain sufficient information to accurately hindcasts springtime rainfall in SE Australia. Using deep learning neural networks trained on the early 21st century data we perfo…
Ambient noise measured in the deep ocean is studied in the context of a search for signals from ultra-high energy cosmic ray neutrinos. The spectral shape of the noise at the relevant high frequencies is found to be very stable for an extensive data set collected over several months from 49 hydrophones mounted near the bottom of the ocean at ~1600 m depth. The slopes of the ambient noise spectra above 15 kHz are found to roll-off faster than the -6 dB/octave seen in Knudsen spectra. A model attributing the source to an uniform distribution of surface noise that includes frequency-dependent abs…
