Update README.md
Browse files
README.md
CHANGED
@@ -5,4 +5,24 @@ task_categories:
|
|
5 |
tags:
|
6 |
- climate
|
7 |
pretty_name: AgroFlux
|
8 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
tags:
|
6 |
- climate
|
7 |
pretty_name: AgroFlux
|
8 |
+
---
|
9 |
+
|
10 |
+
# AgroFlux: A Spatial-Temporal Benchmark for Carbon and Nitrogen Flux Prediction in Agricultural Ecosystems
|
11 |
+
|
12 |
+
This repository contains the official data of paper AgroFlux: A Spatial-Temporal Benchmark for Carbon and Nitrogen Flux Prediction in Agricultural Ecosystems.
|
13 |
+
Please refer to git repo for the usage of this data: https://github.com/qxc101/AgroFlux.git.
|
14 |
+
|
15 |
+
## Datasets
|
16 |
+
The benchmark works with several data:
|
17 |
+
|
18 |
+
- **Simulation datasets (T0)**:
|
19 |
+
- `mw`: Ecosys simulations
|
20 |
+
- `dc`: DayCent simulations
|
21 |
+
|
22 |
+
- **Observation datasets (T1)**:
|
23 |
+
- `n2o`: Nitrous oxide flux field measurements
|
24 |
+
- `co2`: Carbon dioxide flux field measurements (includes GPP)
|
25 |
+
|
26 |
+
Each Parquet file contains the data. The shape information and variable names are in metadata file. For the shape information, all data are in [year\*365, site\*N_exp_type, variables].
|
27 |
+
For the processing of this dataset, please refer to the file scale_data_t0.ipynb in the git repo.
|
28 |
+
The notebook file contains visualizations and value checks with actual feature names of all the data provided here.
|