Abstract We present a global framework to infer sub-seasonal subsurface water storage dynamics from daily satellite surface soil moisture by optimizing the time constant of an exponential filter used for the depth extrapolation against GRACE and GRACE-FO terrestrial water storage (TWS) anomalies. The approach uses ESA CCI v9.1 surface and root-zone soil moisture products together with daily ITSG-Grace2018 gravity fields on a 1° global grid. For each grid cell, the time constant T of the filter is optimized to maximize the correlation between exponentially filtered soil moisture and GRACE-based TWS from which snow, surface-water, and seasonal components have been removed before. On average, the global area-weighted correlation increases from 0.19 for unfiltered surface soil moisture to 0.39 after optimization. The optimal T values decrease systematically with the depth of the soil moisture layer used as input and show physically consistent patterns related to climatic and hydrogeological controls such as aridity, soil characteristics, and depth to the groundwater table. In contrast to existing approaches that use in-situ soil moisture data to compute globally uniform T parameters, our approach allows to capture spatially varying infiltration dynamics into deeper soil layers. The resulting global T field thus provides an observation-driven proxy for subsurface storage dynamics at weekly-to-monthly time scales, offering a simple and transferable approach for linking satellite surface soil moisture to terrestrial water storage variations.

Changes in soil water storage can be studied on a global scale using a variety of satellite observations. With active or passive microwave remote sensing, we can study the upper few centimeters of the soil, while satellite gravimetry allows us to detect changes in the entire column of terrestrial water storage (TWS). The combination of both types of data can provide valuable insight into hydrological dynamics in different soil depths towards a better understanding of changes in subsurface water storage. We use daily Gravity Recovery and Climate Experiment (GRACE) data and satellite soil moisture data to identify extreme hydroclimatic events, focusing on prolonged droughts. To enhance our comprehension of the subsurface, we utilize not just surface soil moisture data but also integrate information on root zone soil moisture. Original level-3 surface soil moisture data sets of SMAP and SMOS are compared to post-processed level-4 data products (both surface and root zone soil moisture) and a multi-satellite product provided by the ESA CCI. We analyse the correspondence between high and low percentiles in TWS and soil moisture time series, which allows us to identify extreme events in different integration depths and storage compartments. Furthermore, we compute the rate of change of anomalies to assess how quickly the system accumulates storage deficits during drought conditions and recovers from them for different soil depths. Our investigation focuses on the temporal dynamics of near-surface soil moisture and TWS, highlighting the cascading effects that propagate from the surface into the subsurface. The results we obtained indicate characteristic patterns of the temporal dynamics of drought recovery in varying soil depths. Specifically, our analysis shows that surface soil moisture recovers faster than TWS, and that this recovery process slows down as soil integration depth increases.

The increasing frequency, intensity, and duration of extreme heat and drought events in a warming climate make it crucial to understand the relationship between surface and subsurface water storage dynamics during these events. Changes in water storage can be studied globally using satellite observations. Microwave remote sensing observes the upper few centimeters of the soil, while satellite gravimetry detects changes in the entire column of terrestrial water storage. We use daily data of the Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO), satellite-based surface soil moisture data and root zone products from Soil Moisture Ocean Salinity, Soil Moisture Active Passive, and European Space Agency Climate Change Initiative on a harmonized 1 degrees ${}<^>{\circ}$ global grid to study the evolution of water storage deficits across different soil layers. The joint analysis of the three types of data provides valuable insight into the hydrological dynamics in different soil depths and subsurface water storage compartments. To identify different dynamics, we compute the rate of change from de-seasonalized water storage anomaly time series to assess how quickly the system accumulates storage deficits during drought conditions and recovers from them for different integration depths in the subsurface. The results indicate characteristic patterns of the temporal dynamics of drought recovery with fast fluctuations and short recovery times for surface soil moisture, a prolonged behavior in the root-zone, and an even slower response in the entire water column. This highlights that the cascading propagation of drought dynamics from the surface to the subsurface can be quantified by remote sensing data with daily resolution at the global scale.

Water storage changes in the soil can be observed on a global scale with different types of satellite remote sensing. While active or passive microwave sensors are limited to the upper few centimeters of the soil, satellite gravimetry can detect changes in the terrestrial water storage (TWS) in an integrative way, but it cannot distinguish between storage variations in different compartments or soil depths. Jointly analyzing both data types promises novel insights into the dynamics of subsurface water storage and of related hydrological processes. In this study, we investigate the global relationship of (1) several satellite soil moisture products and (2) non-standard daily TWS data from the Gravity Recovery and Climate Experiment/Follow-On (GRACE/GRACE-FO) satellite gravimetry missions on different timescales. The six soil moisture products analyzed in this study differ in the post-processing and the considered soil depth. Level 3 surface soil moisture data sets of the Soil Moisture Active Passive (SMAP) and Soil Moisture and Ocean Salinity (SMOS) missions are compared to post-processed Level 4 data products (surface and root zone soil moisture) and the European Space Agency Climate Change Initiative (ESA CCI) multi-satellite product. On a common global 1∘ grid, we decompose all TWS and soil moisture data into seasonal to sub-monthly signal components and compare their spatial patterns and temporal variability. We find larger correlations between TWS and soil moisture for soil moisture products with deeper integration depths (root zone vs. surface layer) and for Level 4 data products. Even for high-pass filtered sub-monthly variations, significant correlations of up to 0.6 can be found in regions with a large, high-frequency storage variability. A time shift analysis of TWS versus soil moisture data reveals the differences in water storage dynamics with integration depth.

Information on water storage changes in the soil can be obtained on a global scale from different types of satellite observations. While active or passive microwave remote sensing is limited to investigating the upper few centimeters of the soil, satellite gravimetry can detect changes in the full column of terrestrial water storage (TWS), but cannot distinguish between storage variations occurring in different soil depths. Jointly analyzing both data types promises interesting insights into the underlying hydrological dynamics and may enable a better process understanding of water storage change in the subsurface.In this study, we investigate the global relationship of (1) several satellite soil moisture (SM) products and (2) non-standard daily TWS data from the GRACE and GRACE-FO satellite gravimetry missions on a sub-monthly time scale. The analysis of these GRACE data on a daily basis could be beneficial for identifying hydro-climatic extreme events such as heavy precipitation or flood events that occur on a sub-monthly basis.We sample all TWS and SM data sets to a common 1 degree spatial resolution and decompose each signal to sub-monthly frequencies by high-pass filtering. We find increasingly large correlations between the TWS and SM for deeper SM integration depths (root zone vs. surface layer). Even for high-pass-filtered (sub-monthly) variations, significant correlations of up to 0.6 can be found in regions with large high-frequency variability. Time spans with particularly large signal variability, that might hint at extreme events, are identified and compared in both in the TWS and the SM time series. Precipitation data were added to the analysis to provide further evidence for the causes/generation of SM and TWS variations.

Information on water storage changes in the soil can be obtained on a global scale from different types of satellite observations. While active or passive microwave remote sensing is limited to investigating the upper few centimeters of the soil, satellite gravimetry is sensitive to variations in the full column of terrestrial water storage (TWS) but cannot distinguish between storage variations occurring in different soil depths. Jointly analyzing both data types promises interesting insights into the underlying hydrological dynamics and may enable a better process understanding of water storage change in the subsurface. In this study, we aim at investigating the global relationship of (1) several satellite soil moisture (SM) products and (2) non-standard daily TWS data from the GRACE and GRACE-FO satellite gravimetry missions on different time scales. We decompose the data sets into different temporal frequencies from seasonal to sub-monthly signals and carry out the comparison with respect to spatial patterns and temporal variability. Level-3 (Surface SM up to 5 cm depth) and Level-4 (Root-Zone SM up to 1 m depth) data sets of the SMOS and SMAP missions as well as the ESA CCI data set are used in this investigation. Since a direct comparison of the absolute values is not possible due to the different integration depths of the two data sets (SM and TWS), we will analyze their relationship using Pearson’s pairwise correlation coefficient. Furthermore, a time-shift analysis is carried out by means of cross-correlation to identify time lags between SM and TWS data sets that indicate differences in the temporal dynamics of SM storage change in varying depth layers.