This repository contained an updated description of the data produced in the SuperIce project led by the NERSC and supported by ESA Φ-lab
More info on the project: (link to the website)[https://nansencenter.github.io/superice-nersc/]
Data are stored on the https://archive.sigma2.no/ archive.
The data is stored under the following paths:
model/
├── features/
└── predictions/
observations/
├── features/
└── predictions/
1. model/features
File format: YYYYMMDD.nc (YYYY: year, MM: month, DD: day in the month)
Time Period: From 2020 to 2023, only the freezing season (18 Oct. to 15 Apr.)
Contains pairs of low-resolution/high-resolution fields from the NeXtSIM simulation. The dataset generation is described in this document. The fields are listed here
| Dimensions | y = 1086 ; x = 1308 |
|---|
| Variable | Dimensions | Description |
|---|---|---|
time |
- | Time in days since 2022-01-12 00:00:00 |
longitude |
(y, x) | Longitude in degrees east |
latitude |
(y, x) | Latitude in degrees north |
sic |
(y, x) | Sea Ice Concentration |
sic_reprocessed |
(y, x) | Reprocessed Sea Ice Concentration |
sit |
(y, x) | Sea Ice Thickness |
sit_reprocessed |
(y, x) | Reprocessed Sea Ice Thickness |
divergence |
(y, x) | Divergence of sea ice motion |
divergence_reprocessed |
(y, x) | Reprocessed Divergence of sea ice motion |
shear |
(y, x) | Shear of sea ice motion |
shear_reprocessed |
(y, x) | Reprocessed Shear of sea ice motion |
2. model/predictions
File format: sit_hrai_YYYYMMDD.nc (YYYY: year, MM: month, DD: day in the month)
Time Period: From 2020 to 2023, only the freezing season (18 Oct. to 15 Apr.) matching dates in model/features
Contains ensemble (size n=30) of high-resolution fields generated by the diffusion model from features contained in model/features. The AI algorithm is described in this document.
The fields are listed here
| Dimensions | y = 1086 ; x = 1308 ; n = 30 |
|---|
| Variable | Dimensions | Description |
|---|---|---|
time |
- | Time in days since 2022-01-12 00:00:00 |
longitude |
(y, x) | Longitude in degrees east |
latitude |
(y, x) | Latitude in degrees north |
sit_ai |
(n, y, x) | Sea Ice Thickness generated by AI |
The members of the ensemble are spanning the possible high-resolution samples from a unique set of low-resolution features.
3. observations/features
File format (python numpy files): sic_sit_e1_e2_YYYYMMDD.npz (YYYY: year, MM: month, DD: day in the month)
Time Period: From 2020 to 2023, only the freezing season (18 Oct. to 15 Apr.)
Contains pairs of low-resolution fields derived from satellite observations. The longitude/latitude grid is the same as for the model files.
The dataset generation is described in this document.
The fields are listed here:
| Variable | Dimensions | Description |
|---|---|---|
sic |
(1086, 1308) | Sea Ice Concentration (OSI SAF) |
sit |
( 1086, 1308) | Sea Ice Thickness (CS2SMOS) |
divergence |
(1086, 1308) | Divergence of sea ice motion |
shear |
(1086, 1308) | Shear of sea ice motion (OSI SAF) |
2. observations/predictions
File format: sit_hrai_YYYYMMDD.nc (YYYY: year, MM: month, DD: day in the month)
Time Period: From 2020 to 2021, only the freezing season (18 Oct. to 15 Apr.).
Contains ensemble (size n=30) of high-resolution fields generated by the diffusion model from features contained in observations/features. The AI algorithm is described in this document.
The fields are listed here
| Dimensions | y = 1086 ; x = 1308 ; n = 30 |
|---|
| Variable | Dimensions | Description |
|---|---|---|
time |
- | Time in days since 2022-01-12 00:00:00 |
longitude |
(y, x) | Longitude in degrees east |
latitude |
(y, x) | Latitude in degrees north |
sit_ai |
(n, y, x) | Sea Ice Thickness generated by AI |
The members of the ensemble are spanning the possible high-resolution samples from a unique set of low-resolution observed features.