17:15 〜 19:15
[AHW24-P06] 70 years of GRACE-like Terrestrial Water Storage Reconstruction using a Statistical Method
キーワード：GRACE TWS, Reconstruction, Statistical Modeling, Climatic change, Krishna River Basin
Terrestrial water storage (TWS), comprising of water stored above and beneath the earth’s surface, plays a key role in various Earth system processes and meeting various natural and human-induced water demands. The lack of continuous and high-resolution records on various TWS components (mainly surface water, snow, soil moisture, and groundwater) has hindered its holistic quantification. Gravity Recovery and Climate Experiment (GRACE) and its successor, GRACE-Follow-On, have revolutionized TWS monitoring from space on required spatiotemporal scales with inherent limitations. However, GRACE data suffers from limited duration (starting April 2002), intermittent data gaps, and an 11-month gap between two missions from July 2017 to May 2018.
Here, we attempt to reconstruct monthly deseasonalized and detrended terrestrial water storage anomalies (Dst-TWSA) for the selected Krishna River Basin from 1951 to 2020 using a statistical modeling approach and forcing hydro-meteorological data (precipitation, temperature) from two different sources, i.e., IMD and ERA5. Two different GRACE Dst-TWSA data from CSR and JPL were also used to calibrate the reconstructed Dst-TWSA from 2003-2016. The reconstructed Dst-TWSA varies from ~-150 mm to ~200 mm similar to GRACE Dst-TWSA and possesses multiple increasing and decreasing trends in the reconstructed duration based on loess smoothing. The CSR and JPL-based reconstructed data show the highest decreasing trend of 6.45 km3 yr-1 and 9.03 km3 yr-1, respectively, from 1963 to 1972 and an increasing trend of 10.32 km3 yr-1 and 13.42 km3 yr-1, correspondingly, during 2016 to 2020 with IMD forcing data. While with the ERA5 forcing data, these showed a decreasing trend of 8.60 km3 yr-1 and 8.17 km3 yr-1, respectively, from 1967 to 1972 and an increasing trend of 8.17 km3 yr-1 and 13.42 km3 yr-1 from 2016 to 2020. Pearson correlation coefficient between GRACE TWS from multiple sources and different reconstructed data varies from 0.83 to 0.86. There are a few instances of overestimation and underestimation of the reconstructed TWS during extremely high and low precipitation. The reconstructed Dst-TWSA shows seasonal variation and short-term fluctuations pointing out the natural and anthropogenic impact on the study area. The results indicate a hydrological recovery in the basin, possibly because of climatic effects. The 70 years of reconstructed Dst-TWSA helps us better understand the long-term hydrological cycle and fill the various data gaps, and can subsequently be used for monitoring and forecasting the water extremes in the basin.
Here, we attempt to reconstruct monthly deseasonalized and detrended terrestrial water storage anomalies (Dst-TWSA) for the selected Krishna River Basin from 1951 to 2020 using a statistical modeling approach and forcing hydro-meteorological data (precipitation, temperature) from two different sources, i.e., IMD and ERA5. Two different GRACE Dst-TWSA data from CSR and JPL were also used to calibrate the reconstructed Dst-TWSA from 2003-2016. The reconstructed Dst-TWSA varies from ~-150 mm to ~200 mm similar to GRACE Dst-TWSA and possesses multiple increasing and decreasing trends in the reconstructed duration based on loess smoothing. The CSR and JPL-based reconstructed data show the highest decreasing trend of 6.45 km3 yr-1 and 9.03 km3 yr-1, respectively, from 1963 to 1972 and an increasing trend of 10.32 km3 yr-1 and 13.42 km3 yr-1, correspondingly, during 2016 to 2020 with IMD forcing data. While with the ERA5 forcing data, these showed a decreasing trend of 8.60 km3 yr-1 and 8.17 km3 yr-1, respectively, from 1967 to 1972 and an increasing trend of 8.17 km3 yr-1 and 13.42 km3 yr-1 from 2016 to 2020. Pearson correlation coefficient between GRACE TWS from multiple sources and different reconstructed data varies from 0.83 to 0.86. There are a few instances of overestimation and underestimation of the reconstructed TWS during extremely high and low precipitation. The reconstructed Dst-TWSA shows seasonal variation and short-term fluctuations pointing out the natural and anthropogenic impact on the study area. The results indicate a hydrological recovery in the basin, possibly because of climatic effects. The 70 years of reconstructed Dst-TWSA helps us better understand the long-term hydrological cycle and fill the various data gaps, and can subsequently be used for monitoring and forecasting the water extremes in the basin.