Filling the Gap between GRACE and GRACE-FO Data Using a Model Integrating Variational Mode Decomposition and Long Short-Term Memory: A Case Study of Northwest China
Environmental Earth Sciences(2021)
摘要
The Gravity Recovery and Climate Experiment (GRACE) mission operated
between March 2002 and October 2017, providing monthly global
terrestrial water storage anomalies (TWSA), which are essential in
investigating global climate change and the hydrological cycle. Its
continuation GRACE-Follow On (GRACE-FO) was launched in May 2018,
leading to an 11-months data gap between them. Therefore, it is of high
significance to fill the data gap. However, existing studies have not
yet achieved high prediction accuracy in filling the data gap over
Northwest China (NWC). Based on the CSR RL06 spherical harmonics (SH)
solution from April 2002 to June 2017, this study combined the
Variational Mode Decomposition (VMD) with the Long Short-Term Memory
(LSTM) model to fill the data gap over NWC. Two conventional LSTM models
were set as control groups: LSTM-1, for which GLDAS-based TWS,
temperature, and precipitation were used as input, and LSTM-2, for which
TWSA data of the previous 12 months were used as input. In general, the
two conventional LSTM models showed poor performance. In contrast, the
integrated VMD-LSTM model achieved significantly higher prediction
accuracy (NSE = 0.974, CC = 0.989, and RMSE = 0.808 cm) and exhibited
higher reliability compared to the single LSTM models. The proposed
method offers a means toward bridging the gap in GRACE-based TWSA data
over NWC. It also provides a reference for similar regions around the
world.
更多查看译文
关键词
Deep learning, GRACE, Long short-term memory, Northwest China, Terrestrial water storage, Variational mode decomposition
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要