1. Reconstruction of the Basin‐Wide Sea‐Level Variability in the North Sea Using Coastal Data and Generative Adversarial Networks
- Author
-
Emil V. Stanev, Zeguo Zhang, Sebastian Grayek, Zhang, Zeguo, 1 Institute of Coastal Research Helmholtz‐Zentrum Geesthacht Geesthacht Germany, and Grayek, Sebastian
- Subjects
551.46 ,numerical models ,business.industry ,Deep learning ,deep learning ,Numerical models ,Structural basin ,sea level ,Oceanography ,Adversarial system ,Geophysics ,Space and Planetary Science ,Geochemistry and Petrology ,tides ,Earth and Planetary Sciences (miscellaneous) ,Artificial intelligence ,business ,North sea ,tidal gauges ,Geology ,Generative grammar ,Sea level - Abstract
We present an application of generative adversarial networks (GANs) to reconstruct the sea level of the North Sea using a limited amount of data from tidal gauges (TGs). The application of this technique, which learns how to generate datasets with the same statistics as the training set, is explained in detail to ensure that interested scientists can implement it in similar or different oceanographic cases. Training is performed for all of 2016, and the model is validated on data from 3 months in 2017 and compared against reconstructions using the Kalman filter approach. Tests with datasets generated by an operational model (“true data”) demonstrated that using data from only 19 locations where TGs permanently operate is sufficient to generate an adequate reconstruction of the sea surface height (SSH) in the entire North Sea. The machine learning approach appeared successful when learning from different sources, which enabled us to feed the network with real observations from TGs and produce high‐quality reconstructions of the basin‐wide SSH. Individual reconstruction experiments using different combinations of training and target data during the training and validation process demonstrated similarities with data assimilation when errors in the data and model were not handled appropriately. The proposed method demonstrated good skill when analyzing both the full signal and the low‐frequency variability only. It was demonstrated that GANs are also skillful at learning and replicating processes with multiple time scales. The different skills in different areas of the North Sea are explained by the different signal‐to‐noise ratios associated with differences in regional dynamics., Plain Language Summary: The variability of sea level is one of the most important elements of the ocean dynamics. Basin‐wide observations are due to satellite altimeters, observations in coastal stations are provided by tidal gauges. The first are not very accurate in the coastal areas, the second do not provide basin‐wide coverage. The task in the present work is to use machine learning to reconstruct the sea‐level variability in the North Sea, which is an almost enclosed ocean region, using observations only. Using data from 19 coastal stations and data from numerical models as a representation of the true ocean (synthetic observations), we demonstrated that the generative adversarial networks reconstruct almost perfectly the sea level of the North Sea. The application of this technique, which learns how to generate datasets with the same statistics as the training set, is explained in detail to ensure that interested scientists can implement it in similar or different oceanographic cases., Key Points: Generative Adversarial Networks successfully reconstruct basin‐wide sea level in the North Sea using data from tidal gauges. Machine learning appeared successful when learning from different data sources. The proposed method is skillful at learning and replicating processes with multiple time scales.
- Published
- 2020