Most satellite observations have gaps, for example, due to clouds. This paper presents a method to reconstruct missing data in satellite observation of the chlorophyll a concentration in the Black Sea. Rather than giving a single possible reconstructed field, the discussed method provides an ensemble of possible reconstructions using a generative neural network. The resulting ensemble is validated using techniques from numerical weather prediction and ocean modelling.

This paper presents a new satellite-derived gridded dataset, including 10 years of sea surface height and geostrophic velocity at monthly resolution, over the Arctic ice-covered and ice-free regions, up to 88° N. We assess the dataset by comparison to independent satellite and mooring data. Results correlate well with independent satellite data at monthly timescales, and the geostrophic velocity fields can resolve seasonal to interannual variability of boundary currents wider than about 50 km.

Earth-observing satellites provide routine measurement of several ocean parameters. However, these datasets have a significant amount of missing data due to the presence of clouds or other limitations of the employed sensors. This paper describes a method to infer the value of the missing satellite data based on a convolutional autoencoder (a specific type of neural network architecture). The technique also provides a reliable error estimate of the interpolated value.

The Mediterranean Sea exhibits an anti-estuarine circulation, responsible for its low productivity. Understanding this peculiar character is still a challenge since there is no exact quantification of nutrient sinks and sources. Because nutrient in situ observations are generally infrequent and scattered in space and time, climatological mapping is often applied to sparse data in order to understand the biogeochemical state of the ocean. The dataset presented here partly addresses these issues.

DINCAE is a method for reconstructing missing data in satellite datasets using a neural network. Satellite observations working in the optical and infrared bands are affected by clouds, which obscure part of the ocean underneath. In this paper, a neural network with the structure of a convolutional auto-encoder is developed to reconstruct the missing data based on the available cloud-free pixels in satellite images.