This research data is associated with the manuscript entitled “Long-term (1979-present) Total Water Storage Anomalies Over the Global Land Derived by Reconstructing GRACE data (https://doi.org/10.1029/2021GL093492)”. The study focused on the reconstruction of long-term GRACE-like gridded total water storage anomalies over the global land surface. Since 2002, the Gravity Recovery and Climate Experiment (GRACE) mission has monitored global total water storage anomalies (TWSA) with an unprecedented accuracy. Yet, many applications require a longer record, i.e. extending prior to the GRACE period. Besides, the Global Climate Observing System (GCOS) Steering Committee has made great efforts towards establishing TWSA as a new Essential Climate Variable (ECV). Here, we produced a new global (excluding Antarctica) total water storage anomaly data set by reconstructing the RL06 CSR mascons using precipitation, land temperature, sea surface temperature, soil moisture, evaporation, surface runoff, subsurface runoff, and several climate indices as inputs. The data set is provided with equivalent water height [unit: cm]. The grid resolution of this data set is 0.5° and the monthly time series covers the full period from July 1979 through June 2020. We compared our dataset to previously published products using the Satellite Laser Ranging (SLR) solution and the observed global mean sea level change as validations. The comparison suggests that we provided a more accurate dataset at the global scale than ever before. This dataset will contribute to the filling of the GRACE data gap and can contribute to the potential studies on testing climate model simulation, constructing the sea level budget, or understanding drought/flood events prior to the GRACE period.

GRID_CSR_GRACE_REC is derived based on the synthetic methodology framework developed by Li et al. (2020). The synthetic methodology framework contains seven data-driven methods, which could be classified into three groups of techniques – i.e, 1) two statistical decomposition techniques Principal Component Analysis (PCA) and Independent Component Analysis (ICA), 2) two time series decomposition techniques Least Square (LS) and Seasonal-Trend decomposition based on Loess (STL), and 3) three machine learning techniques Artificial Neural Network(ANN), AutoRegressive model with eXogenous variables (ARX), and Multiple Linear Regression (MLR).

One could read this global total water storage reconstruction data set by using any version of the software Matlab with a simple command 'load'.
====================
 Variable descriptions 
 
lat : Double type latitude.
long : Double type longitude.
time : Double type time.
str_month : String type month.
str_year : String type year.
grace_rec_full : Double type gridded global TWSA reconstructions.
grace_rec_detrended : Double type gridded global TWSA reconstructions, linear trends removed.
grace_rec_deseasonalised : Double type gridded global TWSA reconstructions, linear trend and seasonal signals removed.  

====================

When using the GRID_CSR_GRACE_REC data set, please cite

Li, F., Kusche, J., Rietbroek, R., Wang, Z., Forootan, E., Schulze, K., & Lück, C. (2020). Comparison of Data‐driven Techniques to Reconstruct (1992‐2002) and Predict (2017‐2018) GRACE‐like Gridded Total Water Storage Changes using Climate Inputs[J]. Water Resources Research, 56(5), e2019WR026551. https://doi.org/10.1029/2019wr026551.          
                                           

Li, F., Kusche, J., Chao, N., Wang, Z., & Löcher, A. (2021). Long-term (1979-present) total water storage anomalies over the global land derived by reconstructing GRACE data[J]. Geophysical Research Letters, 48, e2021GL093492. https://doi.org/10.1029/2021GL093492.