Daily GRACE satellite data evaluate short-term hydro-meteorological fluxes from global atmospheric reanalyses

Water cycle changes in reanalyses: a complementary framework

Climate reanalyses complement traditional surface-based measurements and offer unprecedented coverage over previously inaccessible or unmonitored regions. Even though these have improved the quantification of the global water cycle, their varying performances and uncertainties limit their applicability. Herein, we discuss how a framework encompassing precipitation, evaporation, their difference, and their sum could further constrain uncertainty by unveiling discrepancies otherwise overlooked. Ahead, we physically define precipitation plus evaporation to describe the global water cycle fluxes in four reanalysis data sets (20CR v3, ERA-20C, ERA5, and NCEP1). Among them, we observe four different responses to the temperature increase between 1950-2010, with ERA5 showing the best agreement with the water cycle acceleration hypothesis. Our results show that implementing the framework proposed can improve the evaluation of reanalyses' performance and enhance our understanding of the water cycle changes on a global scale.

Bridging the gap between GRACE and GRACE-FO missions with deep learning aided water storage simulations

The monthly high-resolution terrestrial water storage anomalies (TWSA) during the 11-months of gap between GRACE (Gravity Recovery And Climate Experiment) and its successor GRACE-FO (-Follow On) missions are missing. The continuity of the GRACE-like TWSA series with commensurate accuracy is of great importance for the improvement of hydrologic models both at global and regional scales. While previous efforts to bridge this gap, though without achieving GRACE-like spatial resolutions and/or accuracy have been performed, high-quality TWSA simulations at global scale are still lacking. Here, we use a suite of deep learning (DL) architectures, convolutional neural networks (CNN), deep convolutional autoencoders (DCAE), and Bayesian convolutional neural networks (BCNN), with training datasets including GRACE/-FO mascon and Swarm gravimetry, ECMWF Reanalysis-5 data, normalized time tag information to reconstruct global land TWSA maps, at a much higher resolution (100 km full wavelength) than that of GRACE/-FO, and effectively bridge the 11-month data gap globally. Contrary to previous studies, we applied no prior de-trending or de-seasoning to avoid biasing/aliasing the simulations induced by interannual or longer climate signals and extreme weather episodes. We show the contribution of Swarm and time inputs which significantly improved the TWSA simulations in particular for correct prediction of the trend component. Our results also show that external validation with independent data when filling large data gaps within spatio-temporal time series of geophysical signals is mandatory to maintain the robustness of the simulation results. The results and comparisons with previous studies and the adopted DL methods demonstrate the superior performance of DCAE. Validations of our DCAE-based TWSA simulations with independent datasets, including in situ groundwater level, Interferometric Synthetic Aperture Radar measured land subsidence rate (e.g. Central Valley), occurrence/timing of severe flash flood (e.g. South Asian Floods) and drought (e.g. Northern Great Plain, North America) events occurred within the gap, reveal excellent agreements.

Autoregressive Reconstruction of Total Water Storage within GRACE and GRACE Follow-On Gap Period

For 15 years, the Gravity Recovery and Climate Experiment (GRACE) mission have monitored total water storage (TWS) changes. The GRACE mission ended in October 2017, and 11 months later, the GRACE Follow-On (GRACE-FO) mission was launched in May 2018. Bridging the gap between both missions is essential to obtain continuous mass changes. To fill the gap, we propose a new approach based on a remove–restore technique combined with an autoregressive (AR) prediction. We first make use of the Global Land Data Assimilation System (GLDAS) hydrological model to remove climatology from GRACE/GRACE-FO data. Since the GLDAS mis-models real TWS changes for many regions around the world, we further use least-squares estimation (LSE) to remove remaining residual trends and annual and semi-annual oscillations. The missing 11 months of TWS values are then predicted forward and backward with an AR model. For the forward approach, we use the GRACE TWS values before the gap; for the backward approach, we use the GRACE-FO TWS values after the gap. The efficiency of forward–backward AR prediction is examined for the artificial gap of 11 months that we create in the GRACE TWS changes for the July 2008 to May 2009 period. We obtain average differences between predicted and observed GRACE values of at maximum 5 cm for 80% of areas, with the extreme values observed for the Amazon, Alaska, and South and Northern Asia. We demonstrate that forward–backward AR prediction is better than the standalone GLDAS hydrological model for more than 75% of continental areas. For the natural gap (July 2017–May 2018), the misclosures in backward–forward prediction estimated between forward- and backward-predicted values are equal to 10 cm. This represents an amount of 10–20% of the total TWS signal for 60% of areas. The regional analysis shows that the presented method is able to capture the occurrence of droughts or floods, but does not reflect their magnitudes. Results indicate that the presented remove–restore technique combined with AR prediction can be utilized to reliably predict TWS changes for regional analysis, but the removed climatology must be properly matched to the selected region.

Reconstruction of GRACE Mass Change Time Series Using a Bayesian Framework

Gravity Recovery and Climate Experiment and its Follow On (GRACE (-FO)) missions have resulted in a paradigm shift in understanding the temporal changes in the Earth's gravity field and its drivers. To provide continuous observations to the user community, missing monthly solutions within and between GRACE (-FO) missions (33 solutions) need to be imputed. Here, we modeled GRACE (-FO) data (196 solutions) between 04/2002-04/2021 to infer missing solutions and derive uncertainties in the existing and missing observations using Bayesian inference. First, we parametrized the GRACE (-FO) time series using an additive generative model comprising long-term variability (secular trend + interannual to decadal variations), annual, and semi-annual cycles. Informative priors for each component were used and Markov Chain Monte Carlo (MCMC) was applied to generate 2,000 samples for each component to quantify the posterior distributions. Second, we reconstructed the new data (229 solutions) by joining medians of posterior distributions of all components and adding back the residuals to secure the variability of the original data. Results show that the reconstructed solutions explain 99% of the variability of the original data at the basin scale and 78% at the one-degree grid scale. The results outperform other reconstructed data in terms of accuracy relative to land surface modeling. Our data-driven approach relies only on GRACE (-FO) observations and provides a total uncertainty over GRACE (-FO) data from the data-generation process perspective. Moreover, the predictive posterior distribution can be potentially used for "nowcasting" in GRACE (-FO) near-real-time applications (e.g., data assimilations), which minimize the current mission data latency (40-60 days).

Bridging the gap between GRACE and GRACE-FO using a hydrological model

Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO), two successive satellite-based missions starting in 2002, have provided an unprecedented way of measuring global terrestrial water storage anomalies (TWSA). However, a temporal gap exists between GRACE and GRACE-FO products from July 2017 to May 2018, which introduces bias and uncertainties in TWSA calculations and modeling. Previous studies have incorporated hydroclimatic factors as predictors for filling the gap, but most of them utilized artificial intelligence or pure statistical models that generally de-trended TWSA and had no physical foundation. Thus, a physically-based reconstruction is required for increasing robustness. In this study, we bridge the temporal gap by developing an empirical hydrological model. The "abcd" model, a T-based snow component, and linear correction are utilized to represent runoff generation, snow dynamics, and long-term trends. The testing results indicate that our hydrological model can successfully reconstruct TWSA in tropical, temperature, and continental climates, although further improvement is needed for arid climates. Our reconstruction for the gap achieves high accuracy and robustness as shown by the evaluations against sea-level budget and GLDAS-derived TWSA. Compared to previous studies using artificial intelligence or statistical techniques, our hydrological model performs similarly in the gap filling but does not involve de-trended or de-seasonalized transformations, which will facilitate the combination of GRACE and GRACE-FO products and improve the physical understanding of global TWSA.

An Innovative Slepian Approach to Invert GRACE KBRR for Localized Hydrological Information at the Sub-Basin Scale

GRACE spherical harmonics are well-adapted for representation of hydrological signals in river drainage basins of large size such as the Amazon or Mississippi basins. However, when one needs to study smaller drainage basins, one comes up against the low spatial resolution of the solutions in spherical harmonics. To overcome this limitation, we propose a new approach based on Slepian functions which can reduce the energy loss by integrating information in the spatial, spectral and time domains. Another advantage of these regionally-defined functions is the reduction of the problem dimensions compared to the spherical harmonic parameters. This also induces a drastic reduction of the computational time. These Slepian functions are used to invert the GRACE satellite data to restore the water mass fluxes of different hydro-climatologic environments in Africa. We apply them to two African drainage basins chosen for their size of medium scale and their geometric specificities: the Congo river basin with a quasi-isotropic shape and the Nile river basin with an anisotropic and more complex shape. Time series of Slepian coefficients have been estimated from real along-track GRACE geopotential differences for about ten years, and these coefficients are in agreement with both the spherical harmonic solutions provided by the official centers CSR, GFZ, JPL and the GLDAS model used for validation. The Slepian function analysis highlights the water mass variations at sub-basin scales in both basins.

Data Adaptive Analysis on Vertical Surface Deformation Derived from Daily ITSG-Grace2018 Model

With the widely used monthly gravity models, it is hard to determine the sub-monthly variations. Thanks to the high temporal resolution, a daily ITSG-Grace2018 gravity model is employed to derive the vertical deformation of the China region in 1.0° × 1.0° grids. The standard deviations of residuals between the daily and monthly averaged displacement range from 1.0 to 3.5 mm, reaching half of the median residuals, which indicates that a higher temporal resolution gravity model is quite necessary for the analysis of crustal displacement. For the signal analysis, traditional least square (LS) is limited in its analysis of signals with constant amplitude. However, geophysical signals in a geodetic time series usually fluctuate over long periods, and missing data happen. In this study, the data adaptive approach called enhanced harmonic analysis (EHA), which is based on an Independent Point (IP) scheme, is introduced to deal with these issues. To demonstrate the time-varying signals, the relative differences between EHA and LS are calculated. It illustrates that the median percentage of epochs at grids with a relative difference larger than 10% is 69.7% and the proportions for the ranges of 30%, 50%, and 70% are about 30.1%, 18.4%, and 13.0%, respectively. The obvious discrepancy suggests the advantage of EHA over LS in obtaining time-varying signals. Moreover, the spatial distribution of the discrepancy also demonstrates the regional characteristics, suggesting that the assumption of constant amplitude is not appropriate in specific regions. To further validate the effectiveness of EHA, the comprehensive analysis on the different noise types, number of IPs, missing data, and simultaneous signals are carried out. Specifically, EHA can deal with series containing white or color noise, although the stochastic model for the color noise should be modified. The signals are slightly different when selecting different numbers of IPs within a range, which could be accepted during analysis. Without interpolation, EHA performs well even with continuously missing data, which is regarded as its feature. Meanwhile, not only a single signal but also simultaneous signals can be effectively identified by EHA.