AI learning method to improve altimetry gridding
Image of the Month - November 2021
Current velocity over the Gulf Stream area. From top to bottom and left to right: model high-resolution output considered in the experiment as the "truth", Altimetry measurements (four-satellite configuration plus Swot), current output from optimal interpolation (Duacs processing), and using physics-informed AI scheme (credits IMT Atlantique, with support from Cnes, Idris and NVidia)
Newcomers to altimetry data are often surprised when they see along-track data maps. As all Earth Observation techniques, but with a much smaller "footprint" than most sensors, altimetry satellites can't see the whole Earth at once. The workaround has been to use several satellites' data, several days (including, if possible in the future as well as in the past of a given data file), and ocean physics to estimate what is missing by optimal interpolation. That is what is behind the "multi-mission gridded maps" that a number of altimetry users have been retrieving. However the approach has its drawbacks, including not being able to retrieve details below scales of around 100 km, even when the information were in the native along-track measurements.
Using artificial intelligence methods provides with a new approach to this problem. The deep learning scheme designed by OceaniX team at IMT Atlantique/Lab-STICC enables to better estimate the gradients between the measurements, using an a-priori knowledge of the involved physics.
Improving data processing to provide with the most accurate and most detailed datasets is among the studies ongoing. Swot data will help, but it won't provide data everywhere and anywhere, so such processings will continue to be in order.
See also:
- Applications / ocean
- Data processing: Duacs multimission processing
- Missions: Swot
Other website on this topic:
- OceaniX website
Reference:
- R. Fablet, M.M. Amar, Q. Febvre, M. Beauchamp, B. Chapron, 2021: End-to-end physics-informed representation learning for satellite ocean remote sensing data: applications to satellite altimetry and sea surface currents. ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2021, 295–302, 2021, https://doi.org/10.5194/isprs-annals-V-3-2021-295-2021
News
- Search.
- Front-page news.
Image of the month .- 2024 .
- 2023 .
- 2022 .
- 2021 .
- Dec. 2021: FSLEs help to forecast oil spill dispersion.
- Nov. 2021: AI learning method to improve altimetry gridding.
- Oct. 2021: Snow on sea ice.
- Sep. 2021: A buoy spiraling over Burdwood Bank.
- Aug. 2021: Understanding phytoplanktonic plumes around the Marquesas Islands.
- Jul. 2021: Surface Measurements for Oceanographic Satellites: the SUMOS in-situ and airborne campaign.
- June 2021: Swot and the lakes' winds.
- May 2021: Typhoon Surigae in the eyes of radars.
- Apr. 2021: Swot simulations of Sahelian ponds and lakes.
- Mar. 2021: Eddies are spreading salt in the ocean.
- Feb. 2021: Altimetry helps to monitor the Congo basin.
- Jan. 2021: A new "First light" in the Jasons' family.
- 2020 .
- 2019 .
- 2018 .
- 2017 .
- 2016 .
- 2015 .
- 2014 .
- 2013 .
- 2012 .
- 2011 .
- 2010 .
- 2009 .
- 2008 .
- 2007 .
- 2006 .
- 2005 .
- 2004 .
- 2003 .
- 2002 .
- 2001 .
- 2000 .
- 1999 .
- 1998 .
- Press review.
- Operational news and status .
- News from CTOH .
- Events calendar.
