BCNN4GRACE

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

★ 21 Open GitHub ↗

A Bayesian Convolutional Neural Network for reconstructing GRACE TWSA signals

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

## [Hydrological droughts of 2017-2018 explained by the Bayesian reconstruction of GRACE(-FO) fields](https://doi.org/10.1029/2022WR031997)
[Shaoxing Mo](https://scholar.google.com/citations?user=b5m_q4sAAAAJ&hl=en&oi=ao), [Yulong Zhong](https://scholar.google.com/citations?user=A7D2Kv4AAAAJ&hl=en), [Ehsan Forootan](https://scholar.google.com/citations?user=Yaor7_UAAAAJ&hl=en), [Xiaoqing Shi](https://scholar.google.com/citations?user=MLKqgKoAAAAJ&hl=en&oi=sra), [Wei Feng](https://scholar.google.com/citations?user=B5oOckcAAAAJ&hl=en), Xin Yin, Jichun Wu

This is a PyTorch implementation of Bayesian Convolutional Neural Network (BCNN) for reconstructing the missing GRACE(-FO) TWSA fields of 2017-2018 from hydroclimatic predictors in an image-to-image (field-to-field) regression manner. It's revised after [the repo](https://github.com/cics-nd/cnn-surrogate) of [Dr. Yinhao Zhu](https://scholar.google.com/citations?user=89uRjBkAAAAJ&hl=en). BCNN can be felxiblely applied to other problems involving complex image-to-image mappings.

## Dependencies
* python 3
* PyTorch
* h5py
* matplotlib
* scipy

## BCNN network architecture
![](https://github.com/njujinchun/BCNN4GRACE/blob/main/imgs/BCNN_arch-1.png)

## Datasets
The BCNN reconstructed data are available at [https://zenodo.org/record/5336992](https://zenodo.org/record/5336992). We have also uploaded the datasets used for BCNN training to [Google Drive](https://drive.google.com/drive/folders/1TEEt8ssKEq9k5-VOe21DHBURKuFfhhqp?usp=sharing). One can download the datasets and train BCNN to reproduce our results.

## Network Training
```
python train_svgd.py
```
The parameter settings can be modified in args.py. Two network arechitectures, CBAM and RRDB, are provided. The latter is deeper and slower, but it can generally provide a slightly better performance. The CBAM and RRDB architectures are respectively employed in our [JoH paper](https://doi.org/10.1016/j.jhydrol.2021.127244) and [WRR paper](https://doi.org/10.1029/2022WR031997).

### Reference GRACE(-FO) TWSA field (a), BCNN's reconstruction (b) and predictive uncertainty (c)
![](https://github.com/njujinchun/BCNN4GRACE/blob/main/imgs/BCNN_preditions.gif)

### The BCNN-identified drought regions during the GRACE and GRACE-FO gap (July 2017-May 2018)
![](https://github.com/njujinchun/BCNN4GRACE/blob/main/imgs/BCNN_WSDI_201707_201805.png)

## Citation
See [Mo et al. (2022a)](https://doi.org/10.1029/2022WR031997) and [Mo et al. (2022b)](https://www.sciencedirect.com/science/article/pii/S0022169421012944) for more information. If you find this repo useful for your research, please consider to cite:

```
* Mo, S., Zhong, Y., Forootan, E., Shi, X., Feng, W., Yin, X., Wu, J. (2022a). Hydrological droughts of 
2017–2018 explained by the Bayesian reconstruction of GRACE(-FO) fields. Water Resources Research, 58, 
e2022WR031997. https://doi.org/10.1029/2022WR031997

* Mo, S., Zhong, Y., Forootan, E., Mehrnegar, N., Yin, X., Wu, J., Feng, W., Shi, X. (2022b). Bayesian 
convolutional neural networks for predicting the terrestrial water storage anomalies during GRACE and 
GRACE-FO gap. Journal of Hydrology, 604, 127244. https://doi.org/10.1016/j.jhydrol.2021.127244
```
Related article: [Zhu, Y., & Zabaras, N. (2018). Bayesian deep convolutional encoder–decoder networks for surrogate modeling and uncertainty quantification. J. Comput. Phys., 366, 415-447.](https://www.sciencedirect.com/science/article/pii/S0021999118302341)

## Questions
Contact Shaoxing Mo (smo@nju.edu.cn) with questions or comments.

BCNN4GRACE

Today's picks

Was this content helpful?

Recent Feedback

BCNN4GRACE

Today's picks

Was this content helpful?

Recent Feedback

Download README