Published December 12, 2022
| Version v2
Other
Open
4dvarnet models for the reconstruction of SSH fields from SST-SSH synergies
Description
Trained models and reconstruction results for multimodal 4dvarnets applied to the reconstruction of SSH fields from SST-SSH synergies
Associated preprint: https://arxiv.org/abs/2207.01372
For each model, this archive includes the selected checkpoint of the trained model (*.ckpt files) along with the reconstruction of the SSH fields for the tests dataset (*_res-testdata.nc files). Four models are available:
- 4dvarnet architecture with a multimodal state and a 20-dimensional non-linear multimodal observation term: modelSLA-4dvarnet-mmstate-sstobs-nolinear-tanh-bn_20.ckpt
- 4dvarnet architecture with a 3-dimensional linear multimodal observation term: modelSLA-4dvarnet-sstobs-linear-bn_03.ckpt
- 4dvarnet architecture with a 20-dimensional non-linear multimodal observation term: 4modelSLA-4dvarnet-sstobs-nolinear-tanh-bn_20.ckpt
- u-net architecture with a multi-modal SST-SSH input: modelSLA-unet-std-DirectInv-wSST.ckpt
Files
Files
(767.6 MB)
| Name | Size | |
|---|---|---|
|
md5:260ba0e1c24f282b96dbf3a5aa84bb6f
|
288.7 MB | Download |
|
md5:91102ee0c9586840dbf0d501eb6f4e55
|
20.8 MB | Download |
|
md5:06e7b7c8cf9bd96a92f044613e576ec0
|
71.1 MB | Download |
|
md5:62109895d2b53022da3fa65fbe140e1a
|
18.6 MB | Download |
|
md5:314979058aa2b7bbc158841e7e0e384b
|
19.2 MB | Download |
|
md5:075613d750266a4c3f47aece1c9b7124
|
288.7 MB | Download |
|
md5:dcb4dbd62c172842ee6c483cfc2654ab
|
28.0 MB | Download |
|
md5:5ed2bbe82abb1c1c78fe642fd0636b4a
|
32.7 MB | Download |