romeokienzler
commited on
Commit
•
ece812b
1
Parent(s):
a87dc64
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,21 @@
|
|
1 |
-
---
|
2 |
-
license: cdla-permissive-2.0
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cdla-permissive-2.0
|
3 |
+
---
|
4 |
+
# Model card for granite-geospatial-downscaling
|
5 |
+
|
6 |
+
`granite-geospatial-downscaling` is a fine-tuned foundation model for the downscaling of weather and climate data. It is based on the [Prithvi WxC foundation model](https://huggingface.co/Prithvi-WxC). `granite-geospatial-downscaling` has been used to downscale both MERRA-2 data as well as EURO-CORDEX climate simulations. The weights for the former are included here.
|
7 |
+
<img src="downscaling_T2M_coolwarm_animated.gif" alt="Downscaling of MERRA-2 T2M" width=462>
|
8 |
+
<p>
|
9 |
+
<b>6x downscaling of MERRA-2 2m temperature</b>
|
10 |
+
|
11 |
+
## Architecture
|
12 |
+
|
13 |
+
From an architecture point of view, we embed Prithvi WxC's transformer layers into a series of convolutional layers. That is, we typically increase resolution before and after the pre-trained transformer stages.
|
14 |
+
|
15 |
+
## Data - MERRA-2
|
16 |
+
|
17 |
+
As a reference and baseline how to use Prithvi WxC as well as the downscaling architecture, we have used `granite-geospatial-downscaling` for 6x downscaling of MERRA-2 2m temperature data. That is, we take MERRA-2 data of 0.5 x 0.625 degrees resolution, coarsen it by a factor of six along each axis and then apply an additional smoothing filter via a 3x3 convolution. Subsequently we fine-tune the above architecture to recover the high resolution data. The weights for this are included here.
|
18 |
+
|
19 |
+
## Further applications - EURO-CORDEX
|
20 |
+
|
21 |
+
In addition, we have used the same architecture with different hyperparameter choices for a 12x downscaling of a subset of EURO-CORDEX climate simulation.
|