Preprints
https://doi.org/10.5194/essd-2022-236
https://doi.org/10.5194/essd-2022-236
 
19 Jul 2022
19 Jul 2022
Status: this preprint is currently under review for the journal ESSD.

Reconstructing ocean subsurface salinity at high resolution using a machine learning approach

Tian Tian1,2, Lijing Cheng2,3, Gongjie Wang4, John Abraham5, Shihe Ren6, Jiang Zhu2, Junqiang Song1, and Hongze Leng1 Tian Tian et al.
  • 1College of Meteorology and Oceanography, National University of Defense Technology, Changsha, 410073, China
  • 2Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing,100029, China
  • 3Center for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao, 266071, China
  • 4National Climate Center, Chinese Meteorological Administration, Beijing, 100081, China
  • 5School of Engineering, University of St. Thomas, Minneapolis, 55105, MN, USA
  • 6Key Laboratory of Research on Marine Hazards Forecasting, National Marine Environmental Forecasting Center, Ministry of Natural Resources, Beijing, 100081, China

Abstract. A gridded ocean subsurface salinity dataset with global coverage is useful for research on climate change and its variability. Here, we explore a machine learning approach to reconstruct a high-resolution (0.25° × 0.25°) ocean subsurface (0–2000 m) salinity dataset for the period 1993–2018 by merging in situ salinity profile observations with high-resolution (0.25° × 0.25°) satellite remote sensing altimetry absolute dynamic topography (ADT), sea surface temperature (SST), sea surface wind (SSW) field data, and a coarse resolution (1° × 1°) gridded salinity product. We show that the feed-forward neural network approach can effectively transfer small-scale spatial variations in ADT, SST and SSW fields into the 0.25° × 0.25° salinity field. The root-mean-square error (RMSE) can be reduced by ~11 % on a global-average basis compared with the 1° × 1° salinity gridded field. The reduction in RMSE is much larger in the upper ocean than the deep ocean, because of stronger mesoscale variations in the upper layers. Besides, the new 0.25° × 0.25° reconstruction shows more realistic spatial signals in the regions with strong mesoscale variations, e.g., the Gulf Stream, Kuroshio, and Antarctic Circumpolar Current regions, than the 1° × 1° resolution product, indicating the efficiency of the machine learning approach in bringing satellite observations together with in situ observations. The large-scale salinity patterns from 0.25° × 0.25° data are consistent with the 1° × 1°gridded salinity field, suggesting the persistence of the large-scale signals in the high-resolution reconstruction. The successful application of machine learning in this study provides an alternative approach for ocean and climate data reconstruction that can complement the existing data assimilation and objective analysis methods. The reconstructed IAP0.25° dataset is freely available at http://dx.doi.org/10.12157/IOCAS.20220711.001 (Tian et al., 2022).

Tian Tian et al.

Status: open (until 13 Sep 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on essd-2022-236', Anonymous Referee #1, 02 Aug 2022 reply

Tian Tian et al.

Data sets

IAP observational salinity gridded dataset at 0.25 resolution Lijing Cheng http://dx.doi.org/10.12157/IOCAS.20220711.001

Tian Tian et al.

Viewed

Total article views: 173 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
137 29 7 173 5 5
  • HTML: 137
  • PDF: 29
  • XML: 7
  • Total: 173
  • BibTeX: 5
  • EndNote: 5
Views and downloads (calculated since 19 Jul 2022)
Cumulative views and downloads (calculated since 19 Jul 2022)

Viewed (geographical distribution)

Total article views: 140 (including HTML, PDF, and XML) Thereof 140 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 08 Aug 2022
Download
Short summary
Based on in situ observations, satellite remote sensing data and coarse-resolution salinity grid data, a neural network (NN) was used to reconstruct a high-resolution ocean salinity dataset. New data is consistent with coarse resolution data in terms of large-scale variation, but shows more realistic signals in the regions with strong mesoscale variations. This shows that the NN can transform the small-scale signals of satellite remote sensing fields into a high-resolution salinity estimation.