Super-Resolving Ocean Dynamics from Space with Computer Vision Algorithms

Bruno Buongiorno Nardelli,Davide Cavaliere,Daniele Ciani,Elodie Charles

doi:10.3390/rs14051159

Abstract

Surface ocean dynamics play a key role in the Earth system, contributing to regulate its climate and affecting the marine ecosystem functioning. Dynamical processes occur and interact in the upper ocean at multiple scales, down to, or even less than, few kilometres. These scales are not adequately resolved by present observing systems, and, in the last decades, global monitoring of surface currents has been based on the application of geostrophic balance to absolute dynamic topography maps obtained through the statistical interpolation of along-track satellite altimeter data. Due to the cross-track distance and repetitiveness of satellite acquisitions, the effective resolution of interpolated data is limited to several tens of kilometres. At the kilometre scale, sea surface temperature pattern evolution is dominated by advection, providing indirect information on upper ocean currents. Computer vision techniques are perfect candidates to infer this dynamical information from the combination of altimeter data, surface temperature images and observing-system geometry. Here, we exploit one class of image processing techniques, super-resolution, to develop an original neural-network architecture specifically designed to improve absolute dynamic topography reconstruction. Our model is first trained on synthetic observations built from a numerical general-circulation model and then tested on real satellite products. Provided concurrent clear-sky thermal observations are available, it proves able to compensate for altimeter sampling/interpolation limitations by learning from primitive equation data. The algorithm can be adapted to learn directly from future surface topography, and eventual surface currents, high-resolution satellite observations.

Highlights

In the last decade, technological progress has opened new prospects for the application of deep-learning techniques in a wide range of fields
Due to the uncertainties in the initial conditions and parameterizations and the non-linearity of the dynamics, model predictions drift away from what is seen in the observations, unless observations are ingested within the simulation itself through data assimilation (DA), e.g., [4]
We aim to recover high-resolution sea surface dynamical features by combining low-resolution ocean absolute dynamic topography (ADT) fields based on satellite altimetry (resolving O(100 km) wavelengths) and high-resolution surface surface temperature temperature data data (SST) acquired by thermal imaging spaceborne radiometers (between O(1 km) and O(10 km) depending on the sensor)

Summary

Introduction

Technological progress has opened new prospects for the application of deep-learning techniques in a wide range of fields This revolutionary change originated from the concurrent increase of computational power at widely affordable costs and impressive growth of openly available data. To discover the laws governing Earth system processes and better predict their evolution over several spatial and temporal scales, a combination of precise observations and theoretical/numerical models is needed [2]. Even considering the significant increase in the number of acquisitions by remote sensing platforms and autonomous instruments, it will never be possible to describe and predict the state of the Earth system at all scales (or even of just one of its subsystems, such as the ocean) only through observed data. DA is mostly based on probabilistic approaches, and it is not rigorously tractable due to the huge number of variables and nonlinear processes involved, as well as the difficulty in simultaneously and properly characterizing model and observation errors

Objectives

Methods

Results

Conclusion