0
Article Tier 2 Marine & Wildlife Sign in to save

A novel filtering method for geodetically determined ocean surface currents using deep learning

Environmental Data Science 2023 1 citation

AI summary Read the abstract

Researchers used deep learning to improve the accuracy of ocean current maps derived from satellite measurements of sea level and gravity. Better ocean current mapping helps scientists track where microplastics travel and accumulate in the ocean once they enter from rivers and coastlines.

Study Type Environmental

Abstract Determining an accurate picture of ocean currents is an important societal challenge for oceanographers, aiding our understanding of the vital role currents play in regulating Earth’s climate, and in the dispersal of marine species and pollutants, including microplastics. The geodetic approach, which combines satellite observations of sea level and Earth’s gravity, offers the only means to estimate the dominant geostrophic component of these currents globally. Unfortunately, however, geodetically-determined geostrophic currents suffer from high levels of contamination in the form of geodetic noise. Conventional approaches use isotropic spatial filters to improve the signal-to-noise ratio, though this results in high levels of attenuation. Hence, the use of deep learning to improve the geodetic determination of the ocean currents is investigated. Supervised machine learning typically requires clean targets from which to learn. However, such targets do not exist in this case. Therefore, a training dataset is generated by substituting clean targets with naturally smooth climate model data and generative machine learning networks are employed to replicate geodetic noise, providing noisy input and clean target pairs. Prior knowledge of the geodetic noise is exploited to develop a more realistic training dataset. A convolutional denoising autoencoder (CDAE) is then trained on these pairs. The trained CDAE model is then applied to unseen real geodetic ocean currents. It is demonstrated that our method outperforms conventional isotropic filtering in a case study of four key regions: the Gulf Stream, the Kuroshio Current, the Agulhas Current, and the Brazil-Malvinas Confluence Zone.

More Papers Like This

Article Tier 2

Drift Field Net: Learning Ocean Lagrangian advection fields from in-situ and satellite observations

AI summary Read the abstract

Scientists built an AI tool that predicts ocean currents more accurately than current forecasting systems, helping track where floating plastic debris drifts in the Pacific. Better predictions mean cleanup crews can target garbage patches more effectively, which matters because ocean microplastics eventually make their way into seafood and drinking water.

Article Tier 2

AI-driven prediction of marine microplastics from space: a case study of the Indian Ocean

AI summary Read the abstract

Scientists used satellite data and AI to accurately predict where microplastic pollution accumulates in the Indian Ocean, without needing to physically sample every location. This matters because microplastics can enter the food chain through seafood, and this cheaper monitoring method could help identify pollution hotspots faster, guiding cleanup and policy efforts to protect ocean health and, ultimately, ours.

Article Tier 2

Super-resolution GAN-enhanced detection of sub-pixel marine plastic debris: Leveraging Sentinel-2 time-series and physics-informed deep learning

AI summary Read the abstract

Scientists developed an AI tool that sharpens satellite images to spot floating plastic debris in coastal waters more accurately, even pieces too small to normally show up clearly from space. In tests along Indonesia's coast, the system correctly identified plastic patches over 96% of the time, which could help track ocean plastic pollution before it breaks down into microplastics that enter seafood and drinking water. Better monitoring means faster cleanup response and stronger data to guide policies that protect both ocean health and human health.

Article

Bridging gaps in ocean current modelling: a hybrid physics-driven deep learning approach

AI summary Read the abstract

PhyGOSCR-Net, a physics-driven deep learning model, reconstructed global ocean surface currents with 10–30 cm/s lower RMSE than existing models and successfully identified all five major oceanic microplastic accumulation zones including the Great Pacific Garbage Patch. Accurate ocean current reconstruction is fundamental to predicting microplastic transport, accumulation hotspots, and exposure risks to marine ecosystems and coastal communities.

Article Tier 2

GoogLeNet-Based Deep Learning Framework for Underwater Microplastic Classification in Marine Environments

AI summary Read the abstract

Researchers trained a GoogLeNet deep learning model on underwater images to classify microplastics into four categories, achieving strong classification performance for primary microplastics, secondary microplastics, non-microplastic debris, and marine biota in turbid coastal waters.

Research digests by email

When a large batch of papers lands in the Atlas, we read through it and send a short write-up of what stood out.

Email me about

Share this paper