Spatiotemporal changes of terrestrial water storage and possible causes in the closed Qaidam Basin, China using GRACE and GRACE Follow-On data

JOURNAL OF HYDROLOGY

卷:598

Article

Wei, Linyong^1,2  Jiang, Shanhu^1,2  Ren, Liliang^1,2  Tan, Hongbing³  Ta, Wanquan⁴  Liu, Yi²  Yang, Xiaoli²  Zhang, Linqi²  Duan, Zheng⁵ 

[1] Hohai Univ, State Key Lab Hydrol Water Resources & Hydraul En, Nanjing 210098, Peoples R China

[2] Hohai Univ, Coll Hydrol & Water Resources, Nanjing 210098, Peoples R China

[3] Hohai Univ, Sch Earth Sci & Engn, Nanjing 210098, Peoples R China

[4] Chinese Acad Sci, Cold & Arid Reg Environm & Engn Res Inst, Lanzhou 730000, Peoples R China

[5] Lund Univ, Dept Phys Geog & Ecosyst Sci, S-22362 Lund, Sweden

关键词: GRACE;    Terrestrial water storage;    Long short-term memory;    Ensemble empirical mode decomposition;    Qaidam Basin;   

DOI  :  10.1016/j.jhydrol.2021.126274

来源: Elsevier

【 摘 要 】

Terrestrial water storage (TWS) is a crucial indicator of regional water balance and water resources changes. Due to limited hydrological observations, we combined the Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) products using the Long Short-Term Memory (LSTM) neural network to monitor the TWS changes from April 2002 to March 2020 over the closed Qaidam Basin in northwest China and examined the impacts of climate and meteorological changes on TWS variations. The results indicated that the LSTM model, driven by the cumulative precipitation, temperature, and Global Land Data Assimilation System datasets, was reliable for use in reconstruction of the GRACE products in the closed basin. The TWS variations featured seasonal variation characteristics and a significant upward trend at internal-annual scales, which were tested via linear statistics and a modified Mann-Kendall method. The increasing trend is likely to remain strongly sustainable in the near future with a Hurst index over 0.75 in most regions. Moreover, the TWS oscillation has a periodicity and nonlinearity increase trend of 0.43 mm/month as observed using ensemble empirical mode decomposition analysis, and the TWS components (including snow water equivalent, soil moisture, and groundwater) demonstrate discordant increasing trends in the basin. Under climate change conditions, tele-connection factors have strong impacts on TWS variability, particularly for the Pacific Decadal Oscillation index with a significant negative correlation by cross wavelet transform technology. Nonetheless, the increase in TWS is primarily influenced by precipitation increases and is more sensitive to the accumulated precipitation in this region. In this study, the GRACE products in combination with GRACE-FO data may help us to better understand the spatiotemporal characterization of TWS in Qaidam Basin, which will provide an important support for the water resource management and ecological environment protection in such data-scarce regions.

【 授权许可】

Free