Total Terrestrial Water Storage Research Articles

GLOBGM v1.0: a parallel implementation of a 30 arcsec PCR-GLOBWB-MODFLOW global-scale groundwater model

Abstract. We discuss the various performance aspects of parallelizing our transient global-scale groundwater model at 30′′ resolution (30 arcsec; ∼ 1 km at the Equator) on large distributed memory parallel clusters. This model, referred to as GLOBGM, is the successor of our 5′ (5 arcmin; ∼ 10 km at the Equator) PCR-GLOBWB 2 (PCRaster Global Water Balance model) groundwater model, based on MODFLOW having two model layers. The current version of GLOBGM (v1.0) used in this study also has two model layers, is uncalibrated, and uses available 30′′ PCR-GLOBWB data. Increasing the model resolution from 5′ to 30′′ creates challenges, including increased runtime, memory usage, and data storage that exceed the capacity of a single computer. We show that our parallelization tackles these problems with relatively low parallel hardware requirements to meet the needs of users or modelers who do not have exclusive access to hundreds or thousands of nodes within a supercomputer. For our simulation, we use unstructured grids and a prototype version of MODFLOW 6 that we have parallelized using the message-passing interface. We construct independent unstructured grids with a total of 278 million active cells to cancel all redundant sea and land cells, while satisfying all necessary boundary conditions, and distribute them over three continental-scale groundwater models (168 million – Afro–Eurasia; 77 million – the Americas; 16 million – Australia) and one remaining model for the smaller islands (17 million). Each of the four groundwater models is partitioned into multiple non-overlapping submodels that are tightly coupled within the MODFLOW linear solver, where each submodel is uniquely assigned to one processor core, and associated submodel data are written in parallel during the pre-processing, using data tiles. For balancing the parallel workload in advance, we apply the widely used METIS graph partitioner in two ways: it is straightforwardly applied to all (lateral) model grid cells, and it is applied in an area-based manner to HydroBASINS catchments that are assigned to submodels for pre-sorting to a future coupling with surface water. We consider an experiment for simulating the years 1958–2015 with daily time steps and monthly input, including a 20-year spin-up, on the Dutch national supercomputer Snellius. Given that the serial simulation would require ∼ 4.5 months of runtime, we set a hypothetical target of a maximum of 16 h of simulation runtime. We show that 12 nodes (32 cores per node; 384 cores in total) are sufficient to achieve this target, resulting in a speedup of 138 for the largest Afro–Eurasia model when using 7 nodes (224 cores) in parallel. A limited evaluation of the model output using the United States Geological Survey (USGS) National Water Information System (NWIS) head observations for the contiguous United States was conducted. This showed that increasing the resolution from 5′ to 30′′ results in a significant improvement with GLOBGM for the steady-state simulation when compared to the 5′ PCR-GLOBWB groundwater model. However, results for the transient simulation are quite similar, and there is much room for improvement. Monthly and multi-year total terrestrial water storage anomalies derived from the GLOBGM and PCR-GLOBWB models, however, compared favorably with observations from the GRACE satellite. For the next versions of GLOBGM, further improvements require a more detailed (hydro)geological schematization and better information on the locations, depths, and pumping rates of abstraction wells.

Groundwater dominates terrestrial hydrological processes in the Amazon at the basin and subbasin scales

The hydrology of the Amazon River basin (ARB) has been extensively studied; however, critical scientific gaps remain regarding key processes that govern hydrologic dynamics and the resilience of the rainforest. This inhibits the understanding of hydrological considerations needed for sustainable forest management under climate change and growing human stressors. Here, using high resolution (∼2km), long-term simulations from a process-based hydrological model (LEAF-Hydro-Flood), we investigate the dominant hydrological processes across the ARB, their key roles in shaping basin functions, and the decadal evolution therein. Results indicate that shallow groundwater (<5m deep) strongly modulates the seasonality of the surface fluxes across the ARB and at least 34% of the Amazonian Forest is supported by groundwater during the dry season. A two-month lag between the seasonal peak of evapotranspiration (ET) and river discharge is a key mechanism that potentially prevents the rainforest from tipping into savanna. The dry season in the ARB is getting drier and the wet season is getting wetter, pointing to an accelerating hydrologic cycle. The ARB is dominantly energy limited, however, our results imply that in the absence of groundwater support, and with less than ∼125 mm/month of precipitation, the basin could have become water-limited over some regions. The long-term basin-averaged ET—dominated by transpiration—changed with a split pattern of ±9% in the past three decades. Similarly, water table depth (±19%) and runoff (±29%) changed with heterogenous patterns across the ARB. Further, the contribution of canopy interception loss and ground evaporation changed heterogeneously across the ARB in response to deforestation. River discharge did not change substantially due to the crucial buffering role of groundwater, but terrestrial water storage (TWS) decreased (increased) in the 2000s (2010s) compared to that in the 1990s. Although groundwater is the dominant contributor to total TWS, the dynamics of TWS over the major river channels are controlled by flood water, given relatively shallow groundwater. This study provides crucial insights on the dominant hydrological processes in the ARB to inform forest management practices.

Calibrating global hydrological models with GRACE TWS: does river storage matter?

Although river water storage contributes to Total Terrestrial Water Storage (TWS) variations obtained from GRACE satellite gravimetry, it is unclear if computationally expensive river routing schemes are required when GRACE data is used for calibration and validation in global hydrological modeling studies. Here, we investigate the role of river water storage on calibration and validation of a parsimonious global hydrological model. In a multi-criteria calibration approach, the model is constrained against either GRACE TWS or TWS from which river water storage is removed. While we find that removing river water storage changes the TWS constraint regionally and globally, there are no significant implications for model calibration and the resulting simulations. However, adding modeled river water storage a-posteriori to calibrated TWS simulations improves model validation against seasonal GRACE TWS variations globally and regionally, especially in tropics and Northern low- and wetlands. While our findings justify the exclusion of explicit river routing for global model calibration, we find that the inclusion of river water storage is relevant for model evaluation.

Contribution of vanishing mountain glaciers to global and regional terrestrial water storage changes

Mountain is called the water towers of the world. Due to high sensitivity and vulnerability to climate change, more attention is paid to the change of water resources in mountain regions, where provide water for environmental and human demands downstream. Mountains glacier, as one of the most important components of terrestrial water storage (TWS), effectively regulates and stabilizes surface water resources. TWS appears to be trending below previous ranges in recent years, and glacier retreating is becoming the primary cause of TWS depletion in mountain regions. However, how much mountain glacier changes contribute to TWS changes around mountain regions is unknown. Here, we combine multi-source datasets to quantify the contribution rate over high-elevation mountain regions. On a global scale, the glacial melting is equivalent to about 49% of the total TWS decline during 2006–2015. TWS and glacier show decreasing trends in most of mountain regions and watersheds of the Third Pole and surroundings (TPs), but the contribution ranges from −23% to 728% in mountain regions and −21% to 99% in watersheds of TPs. There are larger contributions in regions with more glaciers, while smaller in less glacier-covered regions. Glacier together with other storage components play diverse roles across different mountain regions and watersheds, but factors with great influence are glacier, groundwater, soil water, reservoirs and lakes.

Mass Variations in Terrestrial Water Storage over the Nile River Basin and Mega Aquifer System as Deduced from GRACE-FO Level-2 Products and Precipitation Patterns from GPCP Data

Changes in the terrestrial total water storage (TWS) have been estimated at both global and river basin scales from the Gravity Recovery and Climate Experiment (GRACE) mission and are still being detected from its GRACE Follow-On (GRACE-FO) mission. In this contribution, the sixth release of GRACE-FO (RL06) level-2 products applying DDK5 (decorrelation filter) were used to detect water mass variations for the Nile River Basin (NRB) in Africa and the Mega Aquifer System (MAS) in Asia. The following approach was implemented to detect the mass variation over the NRB and MAS: (1) TWS mass (June 2018–June 2021) was estimated by converting the spherical harmonic coefficients from the decorrelation filter DDK 5 of the GRACE-FO Level-2 RL06 products into equivalent water heights, where the TWS had been re-produced after removing the mean temporal signal (2) Precipitation data from Global Precipitation Climatology Project was used to investigate the pattern of change over the study area. Our findings include: (1) during the GRACE-FO period, the mass variations extracted from the RL06-DDK5 solutions from the three official centers—CSR, JPL, and GFZ—were found to be consistent with each other, (2) The NRB showed substantial temporal TWS variations, given a basin average of about 6 cm in 2019 and about 12 cm in 2020 between September and November and a lower basin average of about −9 cm in 2019 and −6 cm in 2020 in the wet seasons between March and May, while mass variations for the MAS had a relatively weaker temporal TWS magnitude, (3) the observed seasonal signal over the NRB was attributed to the high intensity of the precipitation events over the NRB (AAP: 1000–1800 mm yr−1), whereas the lack of the seasonal TWS signal over the MAS was due to the low intensity of the precipitation events over the MAS (AAP: 180–500 mm yr−1).

Multisource remote sensing data facilitate ecohydrological simulations without runoff calibration

AbstractObserved runoff is typically the only variable applied to calibrate ecohydrological models, which can simulate runoff well but cannot guarantee reliable reproduction of other hydrological components (e.g., evapotranspiration (ET) and total terrestrial water storage change (TWSC)). This study explores the potential of estimating hydrological components (runoff, ET, TWSC, groundwater, and root‐zone water storage.) using a conceptual model driven and calibrated by multisource remote sensing data only. The performances of the model were evaluated in 13 catchments with different geological, climatic and vegetation conditions. Results show that the model was encouraging in simulating monthly runoff (median KGE, 0.71; RMSE, 19.4 mm/mon), ET (median KGE, 0.86; RMSE, 9.0 mm/mon), and TWSC (median KGE, 0.58; RMSE, 26.6 mm/mon). The good positive correlation between root‐zone water storage and gross primary productivity also indicates the effectiveness of the model in simulating root‐zone water storage. This study offers new prospects for estimating key ecohydrological components precisely, especially in catchments that are difficult to monitor (e.g., karst regions) and catchments with limited observation data (e.g., ungauged regions).

When climate variability partly compensates for groundwater depletion: An analysis of the GRACE signal in Morocco

Study regionMorocco, Northwestern Africa. Study focusSince April 2002, the Gravity Recovery and Climate Experiment (GRACE) mission have opened new pathways for hydrologists to monitor the changes in terrestrial total water storage (TWS). Here, the Center for Space Research (CSR), Goddard Space Flight Center (GSFC), Jet Propulsion Laboratory (JPL), and the average (AVG) GRACE mascon solutions were used to examine the changes in TWS and groundwater storages (GWS) in Morocco, with an emphasis on natural replenishment events and their link to snow cover area (SCA) and rainfall variability. New hydrological insights for the regionThe results showed that GRACE TWS from AVG (TWSAVG) and GSFC (TWSGSFC) can fairly describe the temporal patterns of the groundwater level (GWL). Moreover, during 2002–2020, the TWS underwent a strong depletion relatively masked by natural recharge events. This was revealed as we identified two intermittent depletion episodes with statistically significant rates (−1.03 ± 0.11 to −0.31 ± 0.1 cm yr−1) higher than those obtained for the long-term trend lines (−0.28 ± 0.11 to −0.15 ± 0.07 cm yr−1). The TWS appeared to be strongly linked with the SCA metrics and rainfall indices with 1–3 months of lag. Our findings suggest that the rainfall distribution can be more insightful about changes in groundwater levels compared to the rainfall monthly totals.

Investigating different timescales of terrestrial water storage changes in the northeastern Tibetan Plateau

The northeastern Tibetan Plateau (NETP), bordering the endorheic lake basins and the Upper Yellow River region, has been disturbed by increasing human activities in recent years. The NETP water storage changes could be a combined effect of climate variability/change and human activities (e.g., reservoir operation). However, whether the human activities have evidently altered hydrological processes and become key drivers of total terrestrial water storage (TWS) changes in the NETP remains unclear. To explore the roles of human interventions in changing surface water storage (SWS) and thus influencing regional TWS changes in the NETP, in comparison with natural drivers, this study quantitatively disaggregated and compared the contributions of TWS changes from climate-dominated natural lakes and man-regulated reservoirs at different timescales. Time series of Gravity Recovery and Climate Experiment (GRACE) TWS anomalies (TWSA) exhibited an overall upward trend (0.78 ± 0.06 Gt/yr, p < 0.01) with evident periodic fluctuations from April 2002 to August 2020. Although the GRACE TWSA was more substantially influenced by changes in natural lake water storage (0.96 ± 0.02 Gt/yr) rather than reservoirs (0.54 ± 0.04 Gt/yr) in the long-term trend, the man-regulated reservoir water storage changes can significantly dominate the GRACE TWSA on interannual and intra-annual timescales, especially in the second sub-period (2013.01–2017.06; GRACE TWSA change rate: −1.82 ± 0.29 Gt/yr, p < 0.01, in comparison with the change rate of reservoir water storage of −1.28 ± 0.17 Gt/yr, and the natural lakes of 0.72 ± 0.07 Gt/yr). In some abnormal years, the reservoir storage changes were even close to the overall signal of region-wide GRACE TWSA. In addition, the increase in soil moisture storage (long-term linear trend: 0.65 ± 0.06 Gt/yr, p < 0.01) was also a key factor that cannot be neglected. Our results suggest that human activities are becoming one of the key factors influencing TWS changes in the NETP.

Insights into hydrological drought characteristics using GNSS-inferred large-scale terrestrial water storage deficits

The hydrological loading displacements measured by continuous Global Navigation Satellite System (GNSS) networks can provide critical constraints on total terrestrial water storage (TWS) anomalies. We invert sparsely distributed GNSS vertical positions for daily large-scale water heights based on Slepian basis functions and devise a novel GNSS-based drought severity index (GNSS-DSI) dataset for drought characterization in Brazil. The spatiotemporal patterns of GNSS-inferred water estimates agree with the TWS observations derived from Gravity Recovery and Climate Experiment (GRACE) spherical harmonic solutions. Both GNSS and GRACE capture notable annual water oscillations in the Amazon River Basin, with an annual amplitude close to 500 mm, larger than that of 100–200 mm in the other geographical divisions. The newly-developed monthly GNSS-DSI time series correlate well with the well-accepted GRACE-DSI dataset, and 95% of stations feature moderate-to-strong correlations (greater than 0.50) between these two DSI data sets. The new drought monitoring tool solely based on GNSS-inferred water storage deviations succeeds in identifying the well-documented historical droughts and provides a quantitative characterization of these drought extremes in the four large river basins in Brazil. Our results demonstrate that a sparsely instrumented continuous GNSS network could be taken as an independent tool to remotely monitor large-scale TWS variations and to quantitatively characterize regional-scale hydrological extremes.

Spatiotemporal Downscaling of GRACE Total Water Storage Using Land Surface Model Outputs

High spatiotemporal resolution of terrestrial total water storage plays a key role in assessing trends and availability of water resources. This study presents a two-step method for downscaling GRACE-derived total water storage anomaly (GRACE TWSA) from its original coarse spatiotemporal resolution (monthly, 3-degree spherical cap/~300 km) to a high resolution (daily, 5 km) through combining land surface model (LSM) simulated high spatiotemporal resolution terrestrial water storage anomaly (LSM TWSA). In the first step, an iterative adjustment method based on the self-calibration variance-component model (SCVCM) is used to spatially downscale the monthly GRACE TWSA to the high spatial resolution of the LSM TWSA. In the second step, the spatially downscaled monthly GRACE TWSA is further downscaled to the daily temporal resolution. By applying the method to downscale the coarse resolution GRACE TWSA from the Jet Propulsion Laboratory (JPL) mascon solution with the daily high spatial resolution (5 km) LSM TWSA from the Ecological Assimilation of Land and Climate Observations (EALCO) model, we evaluated the benefit and effectiveness of the proposed method. The results show that the proposed method is capable to downscale GRACE TWSA spatiotemporally with reduced uncertainty. The downscaled GRACE TWSA are also evaluated through in-situ groundwater monitoring well observations and the results show a certain level agreement between the estimated and observed trends.

Spatio-Temporal Evaluation of Water Storage Trends from Hydrological Models over Australia Using GRACE Mascon Solutions

The Gravity Recovery and Climate Experiment (GRACE) data have been extensively used to evaluate the total terrestrial water storage anomalies (TWSA) from hydrological models. However, which individual water storage components (i.e., soil moisture storage anomalies (SMSA) or groundwater water storage anomalies (GWSA)) cause the discrepancies in TWSA between GRACE and hydrological models have not been thoroughly investigated or quantified. In this study, we applied GRACE mass concentration block (mascon) solutions to evaluate the spatio-temporal TWSA trends (2003–2014) from seven prevailing hydrological models (i.e., Noah-3.6, Catchment Land Surface Model (CLSM-F2.5), Variable Infiltration Capacity macroscale model (VIC-4.1.2), Water—Global Assessment and Prognosis (WaterGAP-2.2d), PCRaster Global Water Balance (PCR-GLOBWB-2), Community Land Model (CLM-4.5), and Australian Water Resources Assessment Landscape model (AWRA-L v6)) in Australia and, more importantly, identified which individual water storage components lead to the differences in TWSA trends between GRACE and hydrological models. The results showed that all of the hydrological models employed in this study, except for CLM-4.5 model, underestimated the GRACE-derived TWSA trends. These underestimations can be divided into three categories: (1) ignoring GWSA, e.g., Noah-3.6 and VIC-4.1.2 models; (2) underrating both SMSA and GWSA, e.g., CLSM-F2.5, WaterGAP-2.2d, and PCR-GLOBWB-2 models; (3) deficiently modeling GWSA, e.g., AWRA-L v6 model. In comparison, CLM-4.5 model yielded the best agreement with GRACE but overstated the GRACE-derived TWSA trends due to the overestimation of GWSA. Our results underscore that GRACE mascon solutions can be used as a valuable and efficient validation dataset to evaluate the spatio-temporal performance of hydrological models. Confirming which individual water storage components result in the discrepancies in TWSA between GRACE and hydrological models can better assist in further hydrological model development.

Estimation of component contributions to total terrestrial water storage change in the Yangtze river basin

Terrestrial water storage (TWS) is a key variable in global and regional hydrological cycles. In this study, the TWS changes in the Yangtze River Basin (YRB) were derived using the Lagrange multiplier method (LMM) from Gravity Recovery and Climate Experiment (GRACE) data. To assess TWS changes from LMM, different GRACE solutions, different hydrological models, and in situ data were used for validation. Results show that TWS changes from LMM in YRB has the best performance with the correlation coefficients of 0.80 and root mean square error of 1.48 cm in comparison with in situ data. The trend of TWS changes over the YRB increased by 10.39 ± 1.27 Gt yr−1 during the 2003–2015 period. Moreover, TWS change is disintegrated into the individual contributions of hydrological components (i.e., glaciers, surface water, soil moisture, and groundwater) from satellite data, hydrologic models, and in situ data. The estimated changes in individual TWS components in the YRB show that (1) the contribution of glaciers, surface water, soil moisture, and groundwater to total TWS changes is 15%, 12%, 25% and 48%, respectively; (2) Geladandong glacier melt from CryoSat-2/ICESat data has a critical effect on TWS changes with a correlation coefficients of −0.51; (3) the Three Gorges Reservoir Impoundment has a minimal effect on surface water changes (mainly lake water storage), but it has a substantial effect on groundwater storage (GWS), (4) the Poyang and Doting Lake water storage changes are mainly caused by climate change, (5) soil moisture storage change is mainly influenced by surface water, (6) human-induced GWS changes accounted for approximately half of the total GWS. The results of this study can provide valuable information for decision-making in water resources management.

Ecological restoration impact on total terrestrial water storage

Large-scale ecological restoration (ER) has been successful in curbing land degradation and improving ecosystem services. Previous studies have shown that ER changes individual water flux or storage, but its net impact on total water resources remains unknown. Here we quantify ER impact on total terrestrial water storage (TWS) in the Mu Us Sandyland of northern China, a hotspot of ER practices. By integrating multiple satellite observations and government reports, we construct a TWS record that covers both the pre-ER (1982–1998) and the post-ER (2003–2016) periods. We observe a significant TWS depletion (P < 0.0001) after ER, a substantial deviation from the pre-ER condition. This contrasts with a TWS increase simulated by an ecosystem model that excludes human interventions, indicating that ER is the primary cause for the observed water depletion. We estimate that ER has consumed TWS at an average rate of 16.6 ± 5.0 mm yr−1 in the analysed domain, equivalent to a volume of 21 km3 freshwater loss during the post-ER period. This study provides a framework that directly informs the water cost of ER. Our findings show that ER can exert excessive pressure on regional water resources. Sustainable ER strategies require optimizing ecosystem water consumption to balance land restoration and water resource conservation. While ecological restoration has been promoted for curbing degradation and improving ecosystem health, the impacts on water flux and storage have been understudied. This article finds that large-scale restoration efforts can actually deplete terrestrial water storage, requiring more effort to balance ecosystem needs.

Evaluation of GRACE mascon solutions using in-situ geodetic data: The case of hydrologic-induced crust displacement in the Yangtze River Basin

Since the Gravity Recovery and Climate Experiment (GRACE) satellite mission was started in 2002, a variety of spatial products have been made available to further understanding of mass redistribution in the Earth system. Two such mascon (mass concentration) solutions were developed by the Center for Space Research (CSR-M) and the NASA Jet Propulsion Laboratory (JPL-M), which offers significantly improved spatial localization and more accurate amplitude measurements of changes in recovered terrestrial Total Water Storage (TWS). However, it is difficult to validate GRACE-derived TWS mascons due to the lack of independent measurements of water storage in various forms at larger scales.In this study, we present a simple framework to evaluate GRACE mascon products based on in-situ GPS measurements from the Yangtze River Basin (YRB) in China. We found that the mascons show a more pronounced spatial difference in TWS distribution and highlight more details as compared to smoother results from empirical post-processing filtering applied to spherical harmonics (SH) data. The prediction of vertical displacements from CSR-M and JLP-M is closer to GPS than that from SH. The residual analysis showed the reductions in WRMS (weighted root-mean-squares) from the GPS minus the CSR-M average were greater than those for JPL-M in 41 GPS stations, and the scaling factors from CLM4.0 used in JPL-M-sf had few improvements with respect to agreement with GPS measurements. Our findings indicated CSR-M solutions were more consistent with in-situ observations and more in line with actual surface mass transport in the YRB. These findings also suggested that when using GRACE mascons to detect local TWS changes or when combining GRACE-derived data with GPS-observed displacement to estimate crustal response to loadings, users should note the contributions from effects of load signal sources from atmospheric, non-tidal ocean, and difference sensitivity kernels on differences between TWS from satellite-based and in-situ observations.

GRACE and TRMM mission: The role of remote sensing techniques for monitoring spatio-temporal change in total water mass, Nile basin

Understanding the spatio-temporal changes of the terrestrial water mass behavior together with Enhanced models of natural water storage variation are imperative for better assessing, adapting better evaluable quantified hydrologic catchments.This study provides a reliable impact of the GRACE “gravity recovery and climate experiment” data to gain an independent water mass change on large regional scale. Using GRACE data sets, Terrestrial total Water Storage (TWS) are alternative index to investigate spatio-temporal water variation responses.This study focuses on the Nile basin and utilizes the integration of TWS extracted from GRACE data solutions and Tropical Rainfall Measuring Mission (TRMM) data sets to interpret the average annual/seasonal variation of Terrestrial Total Water Mass.Terrestrial total Water Storage (TWS) was obtained from GRACE data on monthly basis over the period of January 2003–Jul 2016.156 gravity field solutions (RL06 unconstrained solutions) were analyzed and processed as follows: (1) Removing of temporal mean, destriping, smoothing with 250 km radius, (2) Generation of Trend and Standard deviation (SD) images over the entire period, (3) All mass anomalies derived from GRACE data were explicated as an indicator for water mass changes, (4) Topographic, geological and hydrological data were compared to the preformed SDGRACE spatial distribution anomalies utilizing GIS workspace data pool to recognize areas addressing major and significant temporal variations in mass considering the affecting parameters that lead to such observed rise in those variations.Generally, both the analyzed, processed GRACE and TRMM data sets as well as the resulted TWS, SD and trend images displayed a statistically increasing trend in the seasonal water mass over the entire Nile basin as an average in the investigated period (2003–2016). The results show a significant increasing trend during 2013–2016-time segment especially over the northern region.

Combining Physically Based Modeling and Deep Learning for Fusing GRACE Satellite Data: Can We Learn From Mismatch?

AbstractGlobal hydrological and land surface models are increasingly used for tracking terrestrial total water storage (TWS) dynamics, but the utility of existing models is hampered by conceptual and/or data uncertainties related to various underrepresented and unrepresented processes, such as groundwater storage. The gravity recovery and climate experiment (GRACE) satellite mission provided a valuable independent data source for tracking TWS at regional and continental scales. Strong interests exist in fusing GRACE data into global hydrological models to improve their predictive performance. Here we develop and apply deep convolutional neural network (CNN) models to learn the spatiotemporal patterns of mismatch between TWS anomalies (TWSA) derived from GRACE and those simulated by NOAH, a widely used land surface model. Once trained, our CNN models can be used to correct the NOAH‐simulated TWSA without requiring GRACE data, potentially filling the data gap between GRACE and its follow‐on mission, GRACE‐FO. Our methodology is demonstrated over India, which has experienced significant groundwater depletion in recent decades that is nevertheless not being captured by the NOAH model. Results show that the CNN models significantly improve the match with GRACE TWSA, achieving a country‐average correlation coefficient of 0.94 and Nash‐Sutcliff efficient of 0.87, or 14% and 52% improvement, respectively, over the original NOAH TWSA. At the local scale, the learned mismatch pattern correlates well with the observed in situ groundwater storage anomaly data for most parts of India, suggesting that deep learning models effectively compensate for the missing groundwater component in NOAH for this study region.

Water Storage Trends in High Mountain Asia.

Changes in terrestrial water storage (TWS) in High Mountain Asia (HMA) could have major societal impacts, as the region’s large reservoirs of glaciers, snow, and groundwater provide a freshwater source to more than one billion people. We seek to quantify and close the budget of secular changes in TWS over the span of the GRACE satellite mission (2003–2016). To assess the TWS trend budget we consider a new high-resolution mass trend product determined directly from GRACE L1B data, glacier mass balance derived from Digital Elevation Models (DEMs), groundwater variability determined from confined and unconfined well observations, and terrestrial water budget estimates from a suite of land surface model simulations with the NASA Land Information System (LIS). This effort is successful at closing the aggregated TWS trend budget over the entire HMA region, the glaciated portion of HMA, and the Indus and Ganges basins, where the full-region trends are primarily due to the glacier mass balance and groundwater signals. Additionally, we investigate the closure of TWS trends at individual 1-arc-degree mascons (area ≈12,000 km2); a significant improvement in spatial resolution over previous analyses of GRACE-derived trends. This mascon-level analysis reveals locations where the TWS trends are well-explained by the independent datasets, as well as regions where they are not; identifying specific geographic areas where additional data and model improvements are needed. The accurate characterization of total TWS trends and its components presented here is critical to understanding the complex dynamics of the region, and is a necessary step toward projecting future water mass changes in HMA.

Detection of large-scale groundwater storage variability over the karstic regions in Southwest China

The 2003–2013 monthly groundwater storage (GWS) anomalies in the highly karstic region (HKR) and low karstic region (LKR) in the Southwest China are estimated from the Gravity Recovery and Climate Experiment (GRACE) terrestrial water storage (TWS) data by using the ancillary data of surface water storage (SWS) and soil moisture storage (SMS) from the WaterGAP model simulations. The leakage errors in the estimated GWS anomalies are corrected through using the iterative forward modelling approach. The estimated GWS anomalies compare well with in situ groundwater-level observations with the correlation coefficient r = 0.71 and the root-mean-square-error (RMSE) = 42 mm. For both HKR and LKR, ∼60% of temporal variability of TWS is contributed by GWS variability, while SMS contributes to 18% (HKR) and 28% (LKR), and SWS contributes to 22% (HKR) and 14% (LKR) of the TWS variability. Due to the higher permeability of the epi-karst zones and their better connection with the subsurface aquifers, GWS anomalies in HKR show larger correlations with SWS (r = 0.73, RMSE = 51 mm) and SMS (r = 0.68, RMSE = 47 mm) and a shorter lag to precipitation than that in LKR (SWS: r = 0.56, RMSE = 50 mm, and SMS: r = 0.48, RMSE = 49 mm). During the extreme drought in 2009, GWS loss in HKR (LKR) was 74.3 mm/yr (42.7 mm/yr), accounting for 66% (62%) of total TWS loss. The severe GWS loss was mainly due to larger discharges through the well-developed subsurface drainage system rather than human depletion, since groundwater resources are still under-exploited in Southwest China (∼4 km3/yr, 12% of the potentially exploitable amounts). A quicker recovery of GWS from 2009 drought can be observed in HKR than LKR due to the larger and earlier (approximately one month) precipitation and infiltration, and quicker response of groundwater to precipitation in HKR.

Water level changes of Hulun Lake in Inner Mongolia derived from Jason satellite data

Water levels in lakes can reflect changes in such bodies. Therefore, there is value in identifying the variations in water levels using observations from altimetry satellites and analyzing the possible causes. In this work, the water-level changes of Hulun Lake in Inner Mongolia during the period from 2002 to 2015 are monitored by the use of Jason satellite data, the results of which are compared with historical data. Landsat TM/ETM/OLI_TRIS remote sensing images are analyzed, and the surface area of the lake extracted from them and converted to the corresponding water level to verify the values obtained from the Jason observations. The results show a downward trend after 2000 (−0.98 mm/year) and a sharp increase after 2012 (3.07 mm/year). The root mean square error (RMSE) between the two methods was 0.2369 m, and the correlation coefficient was 0.986. By analyzing the various influencing factors, we draw the conclusion that the water level of Hulun Lake is affected by both natural factors (e.g., rainfall, runoff, evapotranspiration etc.) and anthropogenic influences (e.g., water consumption in coal mining, overgrazing, etc.). These are the main causes of the decrease in the area of Hulun Lake and other lakes in the Inner Mongolia Autonomous Region. By comparing the lake storage anomalies of Hulun Lake with the terrestrial Total Water Storage anomalies (TWSA) inverted from GRACE satellite data and the Surface Water Storage anomalies (SWSA) from WaterGAP Global Hydrology Model (WGHM) within the Hulun Basin, we find that not only do Hulun Lake and basin interact with each other, but also that Hulun Lake has an important function with regards to the changes within the basin as a whole. This work therefore provides a method for monitoring the dynamic changes of lake water levels, while analyzing the influencing factors based on multi-scale data. Such a method shows potential for being applied to efforts to ensure environmental protection.

The spatio-temporal variability of groundwater storage in the Amazon River Basin

In large Tropical River basins such as the Amazon, groundwater plays a major role in the water and ecological cycles with large influences on the rainforest ecosystems and climate variability. However, due to the lack of monitoring networks, Amazon groundwater storage and its variability remain poorly known. Here, we provide an unprecedented direct estimate of the spatio-temporal variations of the anomaly of groundwater storage over the period January 2003 -September 2010 in the Amazon Basin by decomposing the total terrestrial water storage measured by the Gravity Recovery and Climate Experiment (GRACE) mission into the individual contributions of other hydrological reservoirs, using multi-satellite data for the surface waters and floodplains and models outputs for the soil moisture. We show that the seasonal variations of groundwater storage represent between 20 and 35% of the terrestrial water storage seasonal volume variations of the Amazon. Larger seasonal amplitudes of groundwater storage (>450 mm) are found in the Alter do Chão and Iça aquifers in the central part of the Amazon Basin. Anomalies of groundwater storage exhibit a strong interannual variability (STD reaching 120 mm along the central corridor) during the study period in response to hydrologic variability and climatic events such as the extreme drought that occurred in 2005.