data preprocess

Author: JayKuo1

.gitignore

@@ -10,6 +10,8 @@ job
 
 tmp
 
+tmp*
+
 trainer
 
 output

README.md

@@ -1,4 +1,3 @@
-
 # Data-driven Global Ocean Modeling for Seasonal to Decadal Prediction
 
 <div align="center">
@@ -8,7 +7,7 @@
 </div>
 
 **This repository contains the official implementation of ORCA-DL**
----
+---------------------------------------------------------------
 
 <!-- ## 📌 Overview
 
@@ -22,9 +21,11 @@
 
 ### Installation
 
-```
+```bash
 git clone https://github.com/OpenEarthLab/ORCA-DL.git
 cd ORCA-DL
+conda create -n orca python=3.9.17
+conda activate orca
 pip install -r requirements.txt
 ```
 
@@ -33,14 +34,29 @@ pip install -r requirements.txt
 All the model weights and data can be found in  https://1drv.ms/f/c/49d761d10f0b201d/Emi9scIyaWBCrNTgRo6t12oBLnF2qGDRGj0M7-g0ekRM1A
 
 ### Quick Demo
+
 See [demo.ipynb](https://github.com/OpenEarthLab/ORCA-DL/blob/main/demo.ipynb)
 
-## 📋 Updates
+Please note that ORCA-DL initially uses [GODAS](https://psl.noaa.gov/data/gridded/data.godas.html) data as its starting point. Initialization with other data is also feasible, but it is necessary to ensure that the data is interpolated to the correct longitude and latitude range and resolution, as in [example_data](https://github.com/OpenEarthLab/ORCA-DL/blob/main/example_data). Here is an example of how to interpolate the original data using [CDO](https://code.mpimet.mpg.de/projects/cdo) (recommended):
 
-- Training data and code are coming soon.
+```bash
+wget https://downloads.psl.noaa.gov/Datasets/godas/sshg.1980.nc -O sshg-1980.nc   # download a 2D data from GODAS
+cdo -b f64 remapbil,grid sshg-1980.nc sshg-1980-processed.nc
 
-- **2025-03:** Model weights and demo code are released.
+wget https://downloads.psl.noaa.gov/Datasets/godas/salt.1980.nc -O salt-1980.nc   # download a 3D data from GODAS
+cdo -b f64 remapbil,grid salt-1980.nc tmp1.nc
+cdo intlevel,10,15,30,50,75,100,125,150,200,250,300,400,500,600,800,1000 tmp1.nc tmp2.nc
+cdo setzaxis,zaxis.txt tmp2.nc salt-1980-processed.nc
+rm tmp1.nc tmp2.nc
+```
+
+After the data interpolation is completed, you can refer to the [demo.ipynb](https://github.com/OpenEarthLab/ORCA-DL/blob/main/demo.ipynb) to run ORCA-DL.
 
+## 📋 Updates
+
+- Training data and code are coming soon.
+- **2025-03-21:** Data preprocessing processes are released.
+- **2025-03-04:** Model weights and demo code are released.
 
 ## 📄 Citation
 

grid

@@ -0,0 +1,13 @@
+gridtype  = lonlat
+xsize     = 360
+ysize     = 128
+xname     = lon
+xlongname = "longitude"
+xunits    = "degrees_north"
+yname     = lat
+ylongname = "latitude"
+yunits    = "degrees"
+xfirst    = 0.5
+xinc      = 1.0
+yfirst    = -63.5
+yinc      = 1.0

zaxis.txt

@@ -0,0 +1,5 @@
+zaxistype = height
+size = 16
+name = level
+units = m
+levels = 10,15,30,50,75,100,125,150,200,250,300,400,500,600,800,1000