Abstract:Ocean data timeseries are vital for a diverse range of stakeholders (ranging from government, to industry, to academia) to underpin research, support decision making, and identify environmental change. However, continuous monitoring and observation of ocean variables is difficult and expensive. Moreover, since oceans are vast, observations are typically sparse in spatial and temporal resolution. In addition, the hostile ocean environment creates challenges for collecting and maintaining data sets, such as instrument malfunctions and servicing, often resulting in temporal gaps of varying lengths. Neural networks (NN) have proven effective in many diverse big data applications, but few oceanographic applications have been tested using modern frameworks and architectures. Therefore, here we demonstrate a “proof of concept” neural network application using a popular “off-the-shelf” framework called “TensorFlow” to predict subsurface ocean variables including dissolved oxygen and nutrient (nitrate, phosphate, and silicate) concentrations, and temperature timeseries and show how these models can be used successfully for gap filling data products. We achieved a final prediction accuracy of over 96% for oxygen and temperature, and mean squared errors (MSE) of 2.63, 0.0099, and 0.78, for nitrates, phosphates, and silicates, respectively. The temperature gap-filling was done with an innovative contextual Long Short-Term Memory (LSTM) NN that uses data before and after the gap as separate feature variables. We also demonstrate the application of a novel dropout based approach to approximate the Bayesian uncertainty of these temperature predictions. This Bayesian uncertainty is represented in the form of 100 monte carlo dropout estimates of the two longest gaps in the temperature timeseries from a model with 25% dropout in the input and recurrent LSTM connections. Throughout the study, we present the NN training process including the tuning of the large number of NN hyperparameters which could pose as a barrier to uptake among researchers and other oceanographic data users. Our models can be scaled up and applied operationally to provide consistent, gap-free data to all data users, thus encouraging data uptake for data-based decision making.

Inferring ocean transport statistics with probabilistic neural networks

Spatiotemporal Weighted Neural Network Reveals Surface Seawater Pco2 Distributions and Underlying Environmental Mechanisms in the North Pacific Ocean

Parameterizing Vertical Mixing Coefficients in the Ocean Surface Boundary Layer using Neural Networks

Clustering of marine-debris- and \emph{Sargassum}-like drifters explained by inertial particle dynamics}

Probabilistic neural networks for fluid flow surrogate modeling and data recovery

ESTIMATION OF OCEANIC PARTICULATE ORGANIC CARBON WITH MACHINE LEARNING

Global Dynamic Analysis of the North Pacific Ocean by Data-Driven Generalized Cell Mapping Method

Learning functional priors and posteriors from data and physics

Spatio-temporal Local Interpolation of Global Ocean Heat Transport using Argo Floats: A Debiased Latent Gaussian Process Approach

Enhancing Global Maritime Traffic Network Forecasting with Gravity-Inspired Deep Learning Models

Lagrangian Drifter Path Identification and Prediction: SINDy vs Neural ODE

Launching Drifter Observations in the Presence of Uncertainty

Estimating Ocean Surface Currents With Machine Learning

Hierarchical probabilistic regression for AUV-based adaptive sampling of marine phenomena

Circumpolar Transport and Overturning Strength Inferred From Satellite Observables Using Deep Learning in an Eddying Southern Ocean Channel Model

Towards Optimally Weighted Physics-Informed Neural Networks in Ocean Modelling

Efficacy of Feedforward and LSTM Neural Networks at Predicting and Gap Filling Coastal Ocean Timeseries: Oxygen, Nutrients, and Temperature

Stochastic Reconstruction of Gappy Lagrangian Turbulent Signals by Conditional Diffusion Models

Bayesian neural networks for the probabilistic forecasting of wind direction and speed using ocean data

Predicting species distributions in the open ocean with convolutional neural networks.