Published March 1, 2025
| Version v1
Dataset
Open
Reconstructing Historical Climate Fields With Deep Learning
- 1. UiT Norges arktiske universitet
-
2.
Potsdam Institute for Climate Impact Research
-
3.
UiT The Arctic University of Norway
- 4. Potsdam-Institut für Klimafolgenforschung eV
- 5. Technical University of Munich
- 6. University of Exeter
Description
Supplement to the article "Reconstructing Historical Climate Fields With Deep Learning".
Model checkpoints for LaMa, LaMa random, LaMa(years) on tas and LaMa trained on the sea ice concentration are provided.
Netcdf files for the sitewise and spatial RMSE of the different models on the test set are provided.
Additionally, we provide the reconstructed fields shown in the paper as well as an archived version of the original LaMa source code and our modified version.
For questions contact Nils Bochow (nils.bochow@uit.no)
Files
LaMa_checkpoints.zip
Files
(13.5 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:278e996d13a989046e741364f862cb60
|
3.3 GB | Preview Download |
|
md5:c550fdd6cda9f34c7a1fea2bdf7a0ebf
|
5.8 MB | Preview Download |
|
md5:14217174ef98658bb65b8e05b33efda5
|
3.9 GB | Preview Download |
|
md5:1ad64e14fc23c343b0760319499c894b
|
3.9 GB | Preview Download |
|
md5:6f235ded04f4df5c2d60583b30fbf0cc
|
3.9 MB | Preview Download |
|
md5:5977a9faee0fd35eccd1a9b6fae0f57e
|
74.6 MB | Preview Download |
|
md5:044d2c8f5fc9fbb615ea5ba0d33d6aee
|
157.7 MB | Preview Download |
|
md5:13852760fd137d6221ceaafca7df2e35
|
2.2 GB | Preview Download |
Additional details
Related works
- Is part of
- Software: https://github.com/NilsBochow/lama_reconstruction (URL)