09:00 〜 09:15
[AGE29-01] Verify of Groundwater Storage Changes in the Zhuoshui River Alluvial Fan in Taiwan after Complementing the Satellite Gravity Measurement with Singularity Spectrum
キーワード：Singularity Spectrum Analysis, GRACE, GRACE-FO, Zhuoshui River Alluvial Fan, Groundwater Storage Variability
The Gravity Recovery and Climate Experiment (GRACE) approach allows for the estimation of groundwater storage in time and space, but 33 of the monthly gravity field solutions for GRACE and its follow-on mission (GRACE follow-up, GRACE-FO) have This inhibits the complete analysis and full utilization of GRACE gravity data. Therefore, the purpose of this study is to use 2,561 km2 of the Zhuoshui River Alluvial Fan in west-central Taiwan as the study site and to use Singular Spectrum Analysis (SSA) to compensate for missing data in GRACE and GRACE-FO gravity field solutions
In this study, we collected monthly data from 2002/04 to 2022/12 from NASA GRACE Center for Space Research at the University of Texas at Austin (CSR), and analyzed the principal component (PC) and empirical orthogonal function (EOF) of CSR data by SSA method, to create a new time series called Reconstructed Components RC for the GRACE and GRACE-FO gravity field solutions to compensate for missing data and at same time, we found the SSA iterations, the chosen window length, M, and the number of K have a direct impact on the resulting data, and M must be larger than the set number of K. In the tests conducted in this study, the highest R2 (0.891) was found for M = 20 and K = 20
In this study, we collected monthly data from 2002/04 to 2022/12 from NASA GRACE Center for Space Research at the University of Texas at Austin (CSR), and analyzed the principal component (PC) and empirical orthogonal function (EOF) of CSR data by SSA method, to create a new time series called Reconstructed Components RC for the GRACE and GRACE-FO gravity field solutions to compensate for missing data and at same time, we found the SSA iterations, the chosen window length, M, and the number of K have a direct impact on the resulting data, and M must be larger than the set number of K. In the tests conducted in this study, the highest R2 (0.891) was found for M = 20 and K = 20