README.md 2.96 KB
Newer Older
tpys's avatar
fuxi  
tpys committed
1
2
## FuXi

tpys's avatar
tpys committed
3

tpys's avatar
fuxi  
tpys committed
4
5
6
7
8
9
10
11
This is the official repository for the FuXi paper.

[FuXi: A cascade machine learning forecasting system for 15-day global weather forecast
](https://arxiv.org/abs/2306.12873)

by Lei Chen, Xiaohui Zhong, Feng Zhang, Yuan Cheng, Yinghui Xu, Yuan Qi, Hao Li


tpys's avatar
tpys committed
12

tpys's avatar
fuxi  
tpys committed
13
14
15
16
17
18
19
20
21
## Installation

The Google Drive folder contains the FuXi model, code, and sample input data, all of which are essential resources for this study. Access to these resources is restricted by a password-protected link. To obtain the password, users must complete the provided [Google Form](https://docs.google.com/forms/d/e/1FAIpQLSfjwZLf6PmxRvRhIPMQ1WRLJ98iLxOq_0dXb87N8CFNPyYAGg/viewform?usp=sharing). For inquiries regarding the password, please contact Professor Li Hao at the email address: lihao_lh@fudan.edu.cn.

The downloaded files shall be organized as the following hierarchy:

```plain
├── root
│   ├── data
tpys's avatar
tpys committed
22
│   │    ├── 20210101
23
│   │         ├── input.nc
tpys's avatar
tpys committed
24
25
26
27
28
29
│   │         ├── output
│   │               ├── 006.nc
│   │               ├── 012.nc
│   │               ├── ...
│   │               ├── 360.nc
│   ├── model
30
31
32
33
34
35
│   |    ├── short
│   |    ├── short.onnx
│   |    ├── medium
│   |    ├── medium.onnx
│   |    ├── long
│   |    ├── long.onnx
tpys's avatar
tpys committed
36
|   |   
tpys's avatar
tpys committed
37
38
39
40
41
│   ├── fuxi.py
│   ├── util.py
│   ├── make_era5_input.py
│   ├── make_hres_input.py
│   ├── make_gfs_input.py
tpys's avatar
fuxi  
tpys committed
42
43
44

```

45
1. Install xarray 
tpys's avatar
fuxi  
tpys committed
46

47
48
```bash
conda install -c conda-forge xarray dask netCDF4 bottleneck
tpys's avatar
fuxi  
tpys committed
49
```
tpys's avatar
tpys committed
50

tpys's avatar
tpys committed
51
2. Install onnxruntime
tpys's avatar
tpys committed
52

53
```bash
tpys's avatar
fuxi  
tpys committed
54
55
56
57
58
pip install -r requirement.txt
```

## Demo

59
```bash 
tpys's avatar
tpys committed
60
python fuxi.py --model model_dir --input input_file --num_steps 20 20 20
tpys's avatar
tpys committed
61
62
```

tpys's avatar
fuxi  
tpys committed
63

tpys's avatar
tpys committed
64
## Data preparation 
tpys's avatar
fuxi  
tpys committed
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85

The `input.nc` file contains preprocessed data from the origin ERA5 files. The file has a shape of (2, 70, 721, 1440), where the first dimension represents two time steps. The second dimension represents all variable and level combinations, named in the following exact order:

```plain
'Z50', 'Z100', 'Z150', 'Z200', 'Z250', 'Z300', 'Z400', 'Z500', 'Z600', 'Z700', 'Z850', 'Z925', 'Z1000', 
'T50', 'T100', 'T150', 'T200', 'T250', 'T300', 'T400', 'T500', 'T600', 'T700', 'T850', 'T925', 'T1000', 
'U50', 'U100', 'U150', 'U200', 'U250', 'U300', 'U400', 'U500', 'U600', 'U700', 'U850', 'U925', 'U1000', 
'V50', 'V100', 'V150', 'V200', 'V250', 'V300', 'V400', 'V500', 'V600', 'V700', 'V850', 'V925', 'V1000', 
'R50', 'R100', 'R150', 'R200', 'R250', 'R300', 'R400', 'R500', 'R600', 'R700', 'R850', 'R925', 'R1000', 
'T2M', 'U10', 'V10', 'MSL', 'TP'
```

The last five variables ('T2M', 'U10', 'V10', 'MSL', 'TP') are surface variables, while the remaining variables represent atmosphere variables with numbers representing pressure levels.


**_NOTE:_**

- The variable 'Z' represents geopotential and not geopotential height.
- The variable 'TP' represents total precipitation accumulated over a period of 6 hours.