Preprints
https://doi.org/10.5194/egusphere-2024-455
https://doi.org/10.5194/egusphere-2024-455
20 Feb 2024
 | 20 Feb 2024
Status: this preprint is open for discussion.

Deep Learning for Super-Resolution of Mediterranean Sea Surface Temperature Fields

Claudia Fanelli, Daniele Ciani, Andrea Pisano, and Bruno Buongiorno Nardelli

Abstract. Sea surface temperature (SST) is one of the essential variables of the Earth climate system. Being at the interface with the atmosphere, SST modulates heat fluxes in and out of the ocean, provides insight on several upper/interior ocean dynamical processes, and it is a fundamental indicator of climate variability potentially impacting marine ecosystems’ health. Its accurate estimation and regular monitoring from space is therefore crucial. However, even if satellite infrared/microwave measurements provide a much better coverage than what achievable from in situ platforms, they cannot sense the sea surface under cloudy/rainy conditions. Large gaps are present even in merged multi-sensor satellite products and different statistical strategies have thus been proposed to obtain gap-free (L4) images, mostly based on the Optimal Interpolation algorithms. This kind of techniques, however, filter out the signals below the space-time decorrelation scales considered, significantly smoothing most of the small mesoscale and submesoscale features. Here, deep learning models, originally designed for single image Super Resolution (SR), are applied to enhance the effective resolution of SST products and the accuracy of SST gradients. SR schemes include a set of computer vision techniques leveraging Convolutional Neural Networks to retrieve high-resolution data from low-resolution images. A dilated convolutional multi-scale learning network, which includes an adaptive residual strategy and implements a channel attention mechanism, is used to reconstruct features in SST data at 1/100° spatial resolution starting from 1/16° data over the Mediterranean Sea. The application of this technique shows a remarkable improvement in the high resolution reconstruction, being able to capture small scale features and providing a root-mean-squared-difference improvement of 0.02 °C with respect to the L3 ground-truth data.

Claudia Fanelli, Daniele Ciani, Andrea Pisano, and Bruno Buongiorno Nardelli

Status: open (until 16 Apr 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2024-455', Anonymous Referee #1, 22 Feb 2024 reply
  • RC2: 'Comment on egusphere-2024-455', Anonymous Referee #2, 29 Mar 2024 reply
  • RC3: 'Comment on egusphere-2024-455', Anonymous Referee #3, 02 Apr 2024 reply
  • RC4: 'Comment on egusphere-2024-455', Peter Cornillon, 09 Apr 2024 reply
Claudia Fanelli, Daniele Ciani, Andrea Pisano, and Bruno Buongiorno Nardelli
Claudia Fanelli, Daniele Ciani, Andrea Pisano, and Bruno Buongiorno Nardelli

Viewed

Total article views: 331 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
229 86 16 331 5 5
  • HTML: 229
  • PDF: 86
  • XML: 16
  • Total: 331
  • BibTeX: 5
  • EndNote: 5
Views and downloads (calculated since 20 Feb 2024)
Cumulative views and downloads (calculated since 20 Feb 2024)

Viewed (geographical distribution)

Total article views: 339 (including HTML, PDF, and XML) Thereof 339 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 13 Apr 2024
Download
Short summary
Sea surface temperature (SST) is an essential variable in understanding Earth's climate system and its accurate monitoring from space is essential. Since satellite measurements are hindered by cloudy/rainy conditions, data gaps are present even in merged multi-sensor products. To address this, since Optimal Interpolation techniques tend to smooth out small-scale features, we developed a deep learning model to enhance the effective resolution of gap-free SST images over the Mediterranean Sea.