Search

Highlights • New methodology for addressing model error in Bayesian-MCMC stochastic inversions. • Construction of model-error basis using principal components analysis. • Projection of residual onto basis allows identification of the model-error component. • Identified model error is removed prior to computing the likelihood. • Application to synthetic vadose-zone infiltration problem. Abstract Proxy forward solvers are commonly used in Bayesian solutions to inverse problems in hydrology and geophysics in order to make sampling of the posterior distribution, for example using Markov-chain-Monte-Carlo (MCMC) methods, computationally tractable. However, use of these solvers introduces model error into the problem, which can lead to strongly biased and overconfident parameter estimates if left uncorrected. Focusing on the specific example of estimating unsaturated hydraulic parameters in a layered soil from time-lapse ground-penetrating radar data acquired during a synthetic infiltration experiment, we show how principal component analysis, conducted on a suite of stochastic model-error realizations, can for some problems be used to build a sparse orthogonal basis for the model error arising from known forward solver approximations and/or simplifications. Projection of the residual onto this basis during MCMC permits identification and removal of the model error before calculation of the likelihood. Our results indicate that, when combined with an informal likelihood metric based on the expected behaviour of the ℓ 2 -norm of the residual, this methodology can yield posterior parameter estimates exhibiting a marked reduction in bias and overconfidence when compared to those obtained with no model-error correction, at reasonable computational cost. [ABSTRACT FROM AUTHOR]

Highlights • Assessed impact of soil moisture on runoff in semi-arid area. • High antecedent moisture led to more runoff at plot and catchment scales. • Upscaling plot-scale parameters is difficult given high spatial variability. • Infiltration model performance depends on spatial scale. • Simplistic models may be preferable if data is limited. Abstract Antecedent soil moisture is an important factor in the generation of runoff, but guidance for modeling moisture conditions in semiarid catchments is limited and conflicting. In this study, the impact of antecedent moisture was assessed at the plot scale (2.8 m2) using a portable rainfall simulator, and at the catchment scale based on observed precipitation and discharge for a 2.8 km2 watershed in central New Mexico. Performance of three loss models commonly used for hydrologic analysis in the southwestern U.S. was tested at both scales. High initial moisture content led to substantially higher runoff ratios at both scales, confirming the importance of antecedent soil moisture for runoff predictions in semiarid drainages. Hydrologic parameters estimated based on plot experiments, however, were highly variable and cannot easily be upscaled to the catchment scale. Loss model performance was clearly scale dependent: more simplistic (one- and two parameter) loss methods out-performed a more complex (four parameter) model at the watershed scale, while the opposite held true for test plot simulations. At the catchment scale, all models performed poorly for small runoff events, but yielded acceptable results for storms causing large discharges if antecedent soil moisture was considered. Failing to account for antecedent moisture led to simulated runoff volume errors up to one order of magnitude. [ABSTRACT FROM AUTHOR]

Highlights • Temporal stability of soil water storage (SWS) in high-elevation forests are analyzed. • Temporal changes and spatial variation of SWS decreased with increasing soil depth. • One time-stable location (TSL) can be representative of the mean SWS. • These TSLs be with average LAI or relatively low canopy interception loss locations. • Temporal stability of SWS was mainly affected by soil and vegetation properties. Abstract Understanding soil water storage (SWS) dynamics in soil profiles is important for hydrological modeling and restoration of vegetation in semi-arid areas. Using the temporal stability method, we aimed to investigate the temporal stability of SWS at 0–40, 40–80, and 80–120 cm depth, and to identify representative sites for reliable estimates of the mean SWS in the permanent forest plot in the Qilian Mountains of China. Further, we wanted to identify correlations between temporal stability of SWS and soil, topography, and properties of the vegetation. Soil water content at soil depths 0–120 (for locations 1–40) and 0–70 cm (for locations 41–60) were measured using time-domain reflectometry (TDR) on 52 dates from 2016 to 2017. Results revealed that time-averaged mean SWS for the three layers differed significantly (P < 0.05), and the temporal changes and spatial variations of the mean SWS all decreased with increasing soil depth. Based on either the Spearman correlation coefficient or the standard deviation of relative difference (SDRD) index, the temporal stability of SWSs within the soil profiles under different soil layers were strong, and the number of time-stable locations increased with increasing soil depth, indicating that the SWS intended to be more temporally stable in deeper soil. One time-stable site can be representative of the mean SWS for multiple soil layers for the whole plot, and can accurately estimate the mean SWS for the three soil layers (R2 > 0.86, P < 0.001). Moreover, these time-stable locations should be those having average LAI, SBD or relatively low canopy interception loss compared to the corresponding field means. Soil bulk density, canopy interception rate, and aboveground biomass of mosses can significantly (P < 0.05) affect the stability of SWSs in this high-elevation forests. Such effects, however, differed among the different soil layers due to different conditions of soil characteristics, distribution of root biomass, and freeze–thawing processes. These results suggest that the influence of vegetation properties of the temporal stability of SWS should attract more attention and that both the relative importance of and interactions among different determining factors is helpful for better understanding the mechanistic determinants of SWS temporal stability in these areas. [ABSTRACT FROM AUTHOR]

Highlights • CRNPs were used in 3 different soil textures and compared with buried SWC sensors. • CRNPs estimate SWC in three soil textures with RMSE is 0.05, 0.10, 0.03 respectively. • Silt loam dam plot had the highest rates of change but held more water than others. Abstract The cosmic-ray neutron probe (CRNP) is a new tool for continuously monitoring mean soil–water content (SWC) on a hectometer scale, but the performance of CRNPs in highly heterogeneous distributions of SWC, vegetation and especially soil texture caused by the broken surface on the Loess Plateau in China remains uncertain. The study evaluated the application and accuracy of the method for three soil textures in a small watershed on the plateau to provide a reference and scientific basis for its application. Three CRNPs were installed in a loamy sand hills (LSH) plot, a silty loam dam (SLD) plot, and a sandy loam slope (SLS) plot, soil moisture sensors were buried at various depths in three plots. The footprints of the CRNPs for SWC observation had radii of 127–170, 126–174 and 148–174 m in SLD, SLS and LSH, respectively, and the mean measuring depth was about 20 cm for SLD and SLS and 40 cm for LSH based on general calculation equations. The accuracy of the CRNP method was high when validated using oven-dried method (RMSE = 0.068 cm3 cm−3). The mean SWCs estimated from the other types of sensors were consistent with the mean SWCs from the CRNP probes in the three plots. The coefficient of determination of the linear relationship was in the order: SLS > LSH > SLD. The coefficient of variations (CV) were higher in LSH than SLD and SLS, indicating that SWC varied most in LSH. During periods of rain, the variations were larger for SLD than SLS and LSH, with a CV of 32.2%. The CRNP for SLD had the highest rates of change during periods with and without rain for both increasing and decreasing SWC, but this plot held more water than LSH and SLS, consistent with the high clay and organic-matter contents in SLD. The rates of change were significantly correlated with the antecedent SWCs (P < 0.01), and the higher the antecedent SWC, the faster the rate of change for each plot. This study verified the application of the CRNP for monitoring field-average SWC of the broken surface on the Loess Plateau and represents a new approach for the study of the spatial distribution of SWC on the plateau. CRNP can measure SWC of the complex underlying surface accurately, so as to provide a more reliable assessment of the impacts of landscapes and vegetation changes, such as conversion of cropland to forest and grassland on the Loess Plateau on water resources. [ABSTRACT FROM AUTHOR]

Highlights • SWAT and DE were coupled to provide a dynamic calculation of WEC for NPS pollution. • Three different WECs were analyzed to reveal the impacts of upstream on downstream. • Seasons change had impacts on the WEC, and the smallest value occurred in wet season. • WEC of nutrients in organic species was more affected by seasons than that in inorganic. • Controlling NPS pollution and enhancing the WEC should be carried out simultaneously. Abstract The objective of this study was to advance a new method to dynamically analyze the water environmental capacity (WEC) of non-point source (NPS) pollution. A Soil and Water Assessment Tool (SWAT) model and differential evolution (DE) algorithm were coupled to automatically provide a calculation of the WEC in a daily step. The results showed that the largest ideal WECs occurred in the wet season with values up to 4500 t and 80 t for total nitrogen (TN) and total phosphorus (TP), respectively. The ideal WEC had a positive correlation with the rainfall intensity, while the opposite was true for actual WEC. The smallest value of actual WEC for TN was −150 t in June, and a negative value also occurred in June for TP. This is because NPS pollution discharge was more sensitive to rainfall and flow change than the ideal WEC. Compared to phosphorus, the WEC of nitrogen was more affected by the hydrological and NPS pollution discharge conditions. In addition, the WEC of nitrogen and phosphorus in organic species was more affected by seasonal change than that of nitrogen and phosphorus in inorganic species. Spatially, the channels in the middle and southeast of the study area had the smallest actual WEC and even negative values were observed. The results of remnant WEC showed that the smallest value occurred in July and June with values of −800 t and −140 t for TN and TP, respectively. Thus, significant attention should be paid to controlling NPS pollution during the wet season, and organic nitrogen and phosphorus should be set as the priority pollutants to control. Furthermore, controlling the NPS pollution discharge and enhancing the WEC should be carried out simultaneously. Compared to previously used methods, the calculation method employed in this study had the ability to simulate the transformation and interaction between nitrogen and phosphorus in organic and inorganic species, and better considered dynamic changes in the input conditions. [ABSTRACT FROM AUTHOR]

Highlights • GIS and factor analyses used to unravel factors causing As enrichment in the basin. • Sediment/soil As contents (0.3–13 mg/kg) showed no correlation with other elements. • As WHO limit exceeded in samples from wells (24.6%), springs (16%), and streams (10%). • As mobilization is governed by sulfide oxidation and aquifer salinization/oxygenation. • Both geogenic processes and human-induced perturbations affect groundwater As levels. Abstract The Uluova basin aquifer system (UBAS), comprised of fractured/karstic rock and multilayer basin-fill aquifers, provides water for both domestic and agricultural needs in the Elazığ province (Eastern Turkey). Although, the UBAS has been subject of large-scale hydrological modifications and controversial water management practices since late 1950s, thus far no studies have examined basinwide implications of such human perturbations in an integrated manner. In this study, GIS, geostatistics, R -mode factor analysis (R -mFA) and geochemical modeling techniques were employed to unravel the factors controlling the distribution and sources of arsenic (As) and selected potentially toxic elements (PTEs) in the stream sediment-soil and groundwater-surface water systems of the UBAS. As revealed by the results from R -mFA, three geogenic factors (F1: Clay minerals and Fe–Mn oxyhydroxides, F2: Weathering of parent materials, and F3: Sulfide oxidation in mineralized zones) govern the geochemical dynamics of the PTEs in the stream sediment/soil media. In stream sediment and soil samples, especially Ni, Cr, and Co presented significant enrichment relative to upper continental crust average composition, whereas As contents were relatively low, varying from 0.3 to 13 mg kg−1. Factors extracted from the combined water dataset (F1: Groundwater salinization and arsenic mobilization, F2: Clay minerals and Fe–Mn oxyhydroxides, F3: pH and redox conditions, and F4: Aquifer oxygenation and nitrate contamination) accounted for 72.59% of the total variance. Water-rock interaction (e.g. sulfide oxidation, carbonate dissolution, silicate hydrolysis, adsorption-desorption, ion exchange, and evaporite dissolution), dilution/mixing with fresh/saline water components, evapoconcentration, and human induced perturbations causing internal salinization and oxygenation of the UBAS were the key mechanisms controlling the chemistry of waters and mobilization of As. In the UBAS, majority of As-rich water samples are confined to central-northern half of the basin and typically display high levels of dissolved O 2 , inorganic oxyanions (HCO 3 −, SO 4 2− and of Si, B, Mo, Sb and V) and alkaline pH. Oxidation of sulfides (e.g. pyrite and arsenopyrite) found within the highly fractured Elazığ magmatics in the upland areas at north and subsequent competitive adsorption-desorption processes occurring under alkaline, oxidizing and high ionic strength aquifer conditions along the downgradient groundwater flow path play a pivotal role in the As-enrichment in the UBAS. As concentrations ranged between 0.02 and 367.2 μg L−1 in groundwater, 0.13–4842 μg L−1 in spring water, and 0.04–31.1 μg L−1 in the stream water samples, of which 20.83% exceeded the WHO provisional guideline value. In the water samples, As occurs mainly as As(V) species (HAsO 4 2− and H 2 AsO 4 −), indicated by the Eh-pH diagram and speciation calculations. The results of this study have shown that As enrichment in the UBAS can be attributed to both geogenic processes and anthropogenic activities that have modified the basin hydrology/hydrochemistry. [ABSTRACT FROM AUTHOR]

Highlights • Soil moisture and soil water tension showed seasonally different characteristics at a steep hillslope. • The distribution of soil moisture characteristics curve was different between spring and autumn seasons. • The soil moisture transfer process was seasonally evolved from spring to autumn. Abstract Understanding variations in soil moisture and soil water potential is important for describing hydrological processes at the hillslope scale. An intensive field campaign to measure soil moisture, soil water potential, and rainfall events was performed at many locations and depths to obtain seasonal and spatial distributions of hydrometric measurements along steep hillslopes. Soil moisture and soil water tension analysis showed seasonally distinctive features in statistics and probability distributions. The relationship between soil moisture and soil water tension was explored to characterize seasonal and spatial variations. The recession slopes of soil moisture at several given points of the relationship were also seasonally different. The Brooks and Corey model and the van Genuchten model were used to estimate parameters for the measured data to explain the spatial and seasonal variations. Soil water storage analysis not only indicated that lateral flow tends to be abundant at greater depths but also revealed a seasonal trend in development of subsurface flow in the downslope direction. The combined impact of topography and vegetation determined the seasonal differences in hydrometric responses along the hillslope locations. [ABSTRACT FROM AUTHOR]

Highlights • A coupled model was proposed for bed deformation and bank erosion in the JR. • The model covered morphodynamic, groundwater flow and bank stability modules. • Model performance was assessed by comparing calculations with measurements. • Seasonal changes in soil property and critical overhanging width were clarified. Abstract Recent significant channel evolution in the Jingjiang Reach has raised much attention, particularly the remarkable bank erosion. A coupled model for simulating bed deformation and bank erosion has been proposed in this study, which focuses on the erosion of the bank with a composite structure in the Lower Jingjiang Reach. In order to cover three contributing processes that may interact with each other, the proposed model integrates a one-dimensional morphodynamic module with a two-dimensional module of ground water flow and a bank erosion module for the cantilever failure of a composite riverbank. Model performance was evaluated through a detailed simulation of channel evolution along a 150.8-km subreach in the Jingjiang Reach over the 2005 hydrological year. Satisfying results were obtained from the simulation, showing relatively close agreement between the calculations and measurements in terms of hydrological data at the outlet section, bank erosion sites, longitudinal channel profile and typical cross-sectional profiles. In addition, investigations into temporal changes in bank soil properties and critical overhanging width at cantilever failure demonstrate that there was a seasonal variation in the volumetric water content of bank soil, which increased during the rising and flood periods and then decreased during the recession period, showing an impact on the occurrence timing of cantilever failures. The tensile strength and critical overhanging width had an inverse relationship with the water content, whereas the critical width sharply increased and then decreased during high flows affected by a rapid change in river stage. The temporal distribution of cantilever failure events indicates that cantilever failure primarily occurred in the flood and recession periods. The effects of bed roughness, water content variation and secondary flow on bank erosion were also discussed. [ABSTRACT FROM AUTHOR]

Highlights • Antecedent moisture conditions are shown to significantly influence flood volume. • Flood to antecedent precipitation elasticity is less than for event precipitation. • Flood to antecedent precipitation elasticity decreases as flood extremity increases. Abstract The magnitude of floods depends not only on the intensity and pattern of precipitation during the flood event (the 'flood-producing' precipitation), but also on the moisture stored in the catchment, which arises from antecedent hydrological processes over many preceding timescales. To characterise this effect, an empirical study is conducted on the influence of antecedent precipitation on significant flood events across multiple climate zones and catchment conditions, using 100 Australian catchments with hourly streamflow and precipitation. Antecedent conditions are shown to have a significant influence on flood volume, with three quarters of catchments having at least a 50% difference in flood volume depending on whether the catchment is wet or dry before the flood-producing precipitation event. The study considers the sensitivity of flow to antecedent precipitation by means of an 'elasticity' metric, which indicates the proportional change in flow for a change in antecedent precipitation or flood-producing precipitation. Flood-producing precipitation nevertheless remains the dominant flood driver across most catchments, with the elasticity of flow to antecedent precipitation typically being between 28% and 37% of the elasticity to flood-producing precipitation. Importantly, the elasticity of flow to antecedent precipitation relative to flood-producing precipitation decreases with increasing event magnitude, highlighting that conclusions of future change based on annual maximum streamflow may not be reflective of the processes that operate for more extreme floods that have the greatest impact on society. [ABSTRACT FROM AUTHOR]

• Groundwater mounds resulting from nearshore recharge is identified in a sandy hillslope. • The thickness of vadose zone determines fluxes across water table along the hillslope. • Nearshore recharge temporarily changes streamlines and increases tortuosity of flowpaths. • Flowpaths near the wetland are mainly controlled by regional GW discharge and strong ET. Groundwater mounds and depressions induced by rainfall and evapotranspiration near wetlands/lakes have been widely observed, but their effects on groundwater flow field are poorly known. Here, a sandy hillslope near a wetland in the Ordos Plateau, China, which has a semi-arid climate, is selected for investigation. After identifying temporary groundwater mounds/depressions based on three water level observation wells, a HYDRUS-2D model is setup to obtain the highly transient daily flow field. The variable thickness of unsaturated zone near the wetland, which controls soil water content and unsaturated hydraulic conductivity, is found to be responsible for groundwater mounds resulting from focused recharge and groundwater depressions resulting from nonuniform evapotranspiration. On nonrainy days, evapotranspiration in the lowland is supported by flowpaths from regional groundwater discharge, and in the slope is supported by a flow cell from the upland to the upper part of the slope, which causes water table recession in the upland. On rainy days, focused recharge in the lowland leads to two temporary local flow cells, which break the regional groundwater discharge flowpaths into two temporary flow cells. The two temporary flow cells toward the upland contribute to water table rises in the uplandMoreover, we find focused recharge in rainy days does not change the overall direction of flowpaths controlled by regional groundwater discharge and strong evapotranspiration, but only increase the tortuosity of flowpaths. By coupling daily atmospheric conditions and regional groundwater discharge, this study has implications for groundwater recharge estimation and future groundwater modelling in arid regions. [ABSTRACT FROM AUTHOR]

• ET and T from several models are evaluated by eddy covariance and sap flow data. • Comparison of remote sensing ET models with process based hydrological model. • Remote sensing models tend to underestimate ET in dry season but overestimate in the rainy season. Evapotranspiration (ET) and its components, transpiration (T) and soil evaporation (Es) are crucial to understand hydrological cycle and effective water loss in water resource management. Remote sensing models and process based hydrological models are two kinds of effective methods for evaluating ET and its partitions. Although the ability of remote sensing models to simulate ET has been widely verified, there is limited verification of ET components, especially T, which is the major component of ET. This research used ET obtained from eddy covariance system and T obtained from sap flow observation in an apricot orchard of north China to evaluate the accuracy of ET and its components from three remote sensing models (PT-JPL, PML, and SWH) and a process-based hydrological model (SHAW). The results showed that after calibrations, all the four models satisfactorily simulated total ET particularly at annual level. Among them, the PML model had the best fit efficiency, with the most satisfied R 2 and NSE values. The PT-JPL estimated average ET was the most accurate, with the slope nearest to 1.0. While for the simulation of T, SHAW model had a better performance owing to the fact that SHAW model simulate plant transpiration using changes in soil water content. The average T/ET of the models for the growing season was close to the observed value of 0.70, except that the T/ET of SWH is high due to the lack of Ei calculation. Furthermore, nearly all the three remote sensing models underestimated ET in the dry spring and overestimated ET in the rainy summer. In comparison with the process-based SHAW model, it is suggested that remote sensing ET relies more heavily on the seasonal change of meteorological factors such as precipitation but less capable in reflecting the gradually delaying process of soil moisture in the dry season. Therefore, it is indicated that for future improvement of remote sensing ET models, adjustment of soil water availability in a much longer term is suggested at least for semi-humid and semi-arid ecosystems. [ABSTRACT FROM AUTHOR]

• Snowmelt and soil moisture percolation recharge groundwater in freeze–thaw areas. • Spring snowmelt mainly affects groundwater in high altitude mountainous areas. • Seasonal frozen soil thawing impacts on groundwater in plain areas. • The snow and groundwater show a downward trend, while soil moisture increases. Due to the dense vegetation coverage in mountainous areas and difficulties in field monitoring in mid-high-latitude freezing-thawing zones, satellite and remote sensing technologies can be used to monitor long-term global hydrological changes. This study uses multiple satellite data of snow water equivalent (SWE), soil moisture (SM), and groundwater storage anomaly (GWSA) to analyse their spatial–temporal changes in the seasonal freezing-thawing Tumen River Basin. Water transformation during the thawing period is determined through groundwater level fluctuations and correlation analysis methods; thus, the recharge effects of snowmelt and seasonally frozen soil thawing on soil moisture and groundwater are identified. The snow begins to melt in February, and snow water equivalent decreases gradually from upstream to downstream after reaching the peak. Furthermore, seasonally frozen soil melts gradually from March. Soil moisture downstream is higher than that in the upper and middle reaches. Additionally, the GWSA increases gradually in the thawing period, which is high in mountainous areas and low in plains. The Spearman correlation coefficient among ΔSWE, ΔSM, and ΔGWSA is more significant than 0.5 and dependent on altitude. With snow and frozen soil melting, vertical infiltration and meltwater percolation increase soil moisture and GWSA. This phenomenon is predominant at low altitudes (<600 m) and pore phreatic groundwater in the middle and lower reaches. However, snowmelt is the dominant factor in groundwater recharge at high-altitude upper reaches. For seasonal freezing-thawing zones, a close water transformation was observed between snow melting, soil water thawing, and groundwater in spring. Thus, the environmental and ecological impact of spring snowmelt and floods in freezing-thawing zones can be monitored using multi-source data. [ABSTRACT FROM AUTHOR]

• A numerical model was proposed to study soil water movement in sprinkler irrigation. • Soil tank experiments were used to validate the accuracy of the simulation model. • Soil texture and initial soil water content influenced soil wetting patterns. • Soil moisture uniformity was higher than the irrigation uniformity above ground. Water movement and its distribution uniformity in the soil are essential for the design of sprinkler irrigation systems. They are both crucial factors influencing nutrient migration and the productivity of crops. In this study, a new numerical simulation method was proposed to investigate the soil water movement in sprinkler irrigation under nonuniform infiltration boundary conditions using COMSOL software. Besides, two soil tank experiments were conducted to test the reliability of the simulation model. Finally, the model was applied to evaluate the effects of sprinkler irrigation uniformity, soil texture, and initial soil water content on soil wetting patterns and soil water content uniformity. The results showed that the COMSOL-2D predictions of the vertical wetting fronts were in good agreement with the experimental data. The soil texture and initial soil water content influenced the soil wetting pattern and the risks of surface runoff and deep percolation in sprinkler irrigation systems. When the water application rate was 11.35 mm h−1 with an irrigation duration of 10 h, it was easy to cause surface runoff in silty clay loam while deep percolation in loamy sand. Additionally, when the initial soil water content is high, it should be better to avoid irrigation or cut down the irrigation duration to prevent percolation. The water within the soil was more uniformly distributed than that applied through a sprinkler irrigation system. The uniformity coefficient of soil water content distribution (CU s) increased from 87.75 % to 95.56 %, as the uniformity coefficient of sprinkler irrigation (CU) increased from 40.74 % to 82.41 %. Although a higher initial soil water content brought a higher soil CU s value, it also led to a higher risk of percolation. The experimental and simulation results indicated that the COMSOL-2D model could be used to accurately simulate soil water movement in sprinkler irrigation and determine the suitable operation mode of low-pressure sprinklers. [ABSTRACT FROM AUTHOR]

[Display omitted] • Piston flow dominates the water flux from thick vadose zone to loess aquifer. • Sulfate isotopes are useful to identify connectivity among different aquifers. • Unconfined and confined aquifers are not vertically hydraulically connected. • Multiple tracers can advance the conceptualization of water cycle processes. Low water flux and a thick vadose zone increase the likelihood that water bearing units are hydraulically disconnected in shallow groundwater systems. Particularly for the thick loess deposit of China, hydraulic connectivity among the vadose zone, unconfined and confined aquifers is still unclear since it is controversial whether the unconfined groundwater is recharged by precipitation or confined groundwater. Taking several loess tablelands as example, we collected precipitation, soil water, unconfined and confined groundwater to measure the contents of multiple tracers (e.g., hydrochemistry, δD, δ18O water , δ17O water , δ34S, δ18O sulfate , δ13C and 14C) to evaluate hydraulic connectivity among the vadose zone, unconfined and confined aquifers. Deep soil water can be recharged by precipitation with intensities ≥ 30–40 mm/day, and then vertically discharge to unconfined water primarily through matrix flow with an infiltration velocity of 0.15 m/year. Matrix flow dominates the water flux from the vadose zone to the unconfined aquifer with the rates of 85.4 ± 39.6 mm/year, reflecting the regional hydraulic equilibrium. The unconfined and confined aquifers exhibit negligible vertical hydraulic connectivity because of different recharge periods, flow patterns, water − rock interaction, water residence time and sulfate evolution behaviors. The identified connectivity supports local precipitation as the recharge source, which conflicts with the external source hypothesis. The utilization of multiple tracers in this study can complement each other to provide more comprehensive results, and the results greatly improve the conceptualization of vertical water exchange processes in the loess deposits. [ABSTRACT FROM AUTHOR]

• Intelligent adaptation parameter method improves the hydrological model performance. • LightGBM builds flood characteristics-parameter intelligent adaptation relationship. • The XAJ-IAP model can provide accurate and reliable flood simulation. • LightGBM analysis reveals importance of flood characteristics and parameters. Traditional hydrological modeling methods use a set of parameters to simulate flood processes with complex causes and variable intensity, which can easily lead to parameter instability. To address the problem of parameter instability, this study proposes an approach integrating the hydrological model with Intelligent Adaptation Parameters (IAP), whose intelligent adaptation relationship is established by the light gradient-boosting machine (LightGBM) based on individual calibration parameters by each flood event and flood characteristics including flood-caused rainstorm information and initial soil moisture. A widely used hydrological model, Xin 'anjiang (XAJ) model, is chosen to be integrated with IAP (XAJ-IAP) in this study, which has a relatively complex structure and a total of 15 model parameters. The obtained findings demonstrate that: (1) recalibrating the sensitive runoff concentration and separation parameters with a single flood leads to a notable enhancement in simulation accuracy, while simultaneously considering the model's physical significance; (2) the XAJ overestimates large floods and underestimates small floods. Compared with the XAJ, the XAJ-IAP has a better rain-flood response relationship and simulation accuracy for floods of different magnitudes, solving the problem of parameter instability that exists in XAJ; and (3) evaluated in terms of information gain, sensitive parameters contribute the most to the establishment of the intelligent adaptation relationship in the LightGBM compared to flood-caused rainstorm information and initial soil moisture, indicating that sensitive parameters are the most important input features of the LightGBM. It can be concluded that the intelligent adaptation system can not only solve the problem of parameter instability that exists when traditional hydrological models simulate complex and changeable floods, but also further reveal the relationship between the model and floods. [ABSTRACT FROM AUTHOR]

• In-situ soil moisture is used to characterize the spatiotemporal scales of precipitation, especially for those under 10 km. • A sliding window approach is adopted to capture the spatiotemporally varying rainfall events. • The Central TP is dominated by precipitation with relatively short durations (0–6 h) and small spatial extents (< 16 km). The spatiotemporal scale of precipitation is crucial for understanding the local hydrometeorological processes. It also plays a significant role in determining the layout of rain gauges. Convectional rainfalls are typical in the Tibetan Plateau (TP), but the spatiotemporal scales of precipitation events in the TP, particularly at scales less than 10 km, remain unclear, which is mainly due to limited in-situ precipitation observations and the relative coarse spatial resolution of satellite precipitation products. Building upon the strong response of soil moisture changes to precipitation, this study attempts to reveal the spatiotemporal scales of precipitation in the central TP during summertime (June–September) by investigating the inter-station correlation of accumulated positive soil moisture increments obtained from in-situ soil moisture measurements in a mesoscale region of the central TP. The soil moisture-implied spatiotemporal scales are further verified against precipitation from the Integrated Multi-satellitE Retrievals for GPM (IMERG) at spatial scales larger than 10 km. Results show that: 1) the variations of surface soil moisture increase and precipitation are fairly consistent, making positive increments in surface soil moisture an effective surrogate for reflecting the spatiotemporal scales of precipitation in central TP; 2) the spatial extents of precipitation as indicated by the precipitation-induced positive soil moisture increments are positively correlated with the temporal scales; 3) precipitation events with relatively short durations (0–6 h) and spatial extents of less than 16 km (and in particular, 0–3 h for scales less than 10 km) dominate the central TP. This study provides a novel approach for characterizing precipitation scales in regions with scarce in-situ precipitation observations, and is expected to inspire new way of assessing large scale soil moisture and precipitation products. [ABSTRACT FROM AUTHOR]

• An intercomparison of five RS SM products with in situ measurements was carried out. • Sixteen different European ecoregions were considered as test areas. • Product performances were evaluated in zones with diverse flora, fauna and climate. • SMAPL4 and ESA CCI provided the best performance as complementary products. • ASCAT products furnished the best results among the non-blended/modeled products. Several satellite microwave-based soil moisture (SM) products have become available in recent years, making possible to produce different datasets at continental and/or a global scale. Numerous variables, including geographic heterogeneity, weather and climate, and land cover, have an impact on these products' accuracy. In this study, we account for all these variables by using the ecoregions as the scale of analysis and conducting a thorough comparison of satellite products with ground-based SM measurements. The whole Europe has been selected as study area, considering it as the aggregation of several ecoregions, each of them classified as an homogeneous zones, in terms of climate, vegetation and potentially investigable soil cover. Selected SM satellite data includes: i) the National Aeronautics and Space Administration (NASA) SMAP L4 V5; ii) the European Space Agency (ESA) SMOS-IC V2.0; iii) the H115 and H116 SM products in time-series format generated by analyzing data collected by the Advanced Scatterometer (ASCAT) aboard MetOp satellites; iv) the CGLS SSM 1 km obtained by the radar onboard ESA's SENTINEL-1 platforms; and v) the European Space Agency's Climate Change Initiative for Soil Moisture (ESA CCI SM) "COMBINED" ESA CCI SM v06.1 product, created by merging satellite-based passive/active microwave measurements. Such a variety, in terms of technologies and main features, of publicly available online SM products (in their up-to-date version), allowed for a comprehensive intercomparison against in situ measurements of the International Soil Moisture Network (ISMN), that spreads across all of Europe within the above-mentioned terrestrial ecoregion, using four different metrics, i.e., r, bias, ubRMSE and r/r ANOM. Overall the intercomparison has underlined how SMAP L4 and ESA CCI show the best performance regardless of the considered metric. ASCAT has achieved best performances among non-modeled/blended products, while SMOS-IC has showed slightly better performance than the CGLS SSM 1 km. Using ecoregions allowed to further characterize the differences of each satellite products and identify the areas where all these products are more reliable (e.g., group (1) made by ecoregions as the Cantabrian Mixed forests or the Balkan Mixed forests; or group 2 made by ecoregions which belong to Mediterranean area) or less performing (e.g., the Central European mixed forests and the Pontic steppe). Hence, these classes have been further analyzed exploiting the ERA5-Land data, in terms of the SM, air temperature, precipitation, evaporation and Leaf area index. The research has provided helpful insights into how the performance of particular SM satellite products changes based on both the characteristics of the product under consideration and the area under investigation (i.e., one ecoregion compared to another). [ABSTRACT FROM AUTHOR]

• The two-reservoir model incorporated with SAS functions performs best in simulating streamflow δ18O. • The model reproduces temporal variations of δ18O in soil moisture and groundwater. • The SAS modeling results confirm occurrences of preferential flow in the Moso bamboo watershed. Understanding seasonal variations of the hydrological fluxes and preferential flow is crucial in exploring contaminant fate and transport processes in a watershed of interest. Here, performance of different StorAge Selection (SAS)-based rainfall-runoff models was compared for a Moso bamboo watershed under the East Asian monsoon climate. Most of these models are composed of a two-parallel-reservoir flow module and a companioned SAS module. The results show that the temporal variation in streamflow is well captured by a two-parallel-reservoir flow module. The tracer simulations from different models show that two-uniform-reservoir (SSU) model with no SAS function has the lowest NSE value (0.17), the model including an SAS in one of the two reservoir outflows (SSD(U) and SSD(L)) improves performance (NSE greater than 0.5). The performance in simulating streamflow δ18O is further improved when SAS functions are applied for outflows from both reservoirs (SSD (U, L), NSE = 0.76). Additional inclusion an SAS function for ET fluxes in the model (SSD (U, ET, L)) results in a minor improvement from SSD (U, L). The results of SSD (U, ET, L) have achieved the best transport performance (NSE = 0.77), with implication for preferential flow in the watershed. The preferential flow for deep drainage and lower reservoir outflow shows seasonal variation, with more young water in the East Asian summer monsoon period than that in the East Asian winter monsoon period. The upper reservoir storage and deep drainage rates are the main factors determining the temporal variation of preferential flow for deep drainage and lower reservoir outflow, respectively. [ABSTRACT FROM AUTHOR]

• Globally, RZSM showed a decreasing trend over 1981–2017. • RZSM over 50% of global land areas decreased. • Globally, decreased RZSM was mainly induced by the rising temperature. Root zone soil moisture (RZSM) has a direct impact on ecosystem function, vegetation growth and food security, and plays a vital role in global climate system, water and carbon cycles. However, large variations and uncertainties still exist in RZSM across the globe under the warming climate. In this study, we applied comparison map profile (CMP), Theil-Sen regression and partial correlation analysis to investigate the spatial and temporal changes of RZMS and its driving factors from 1981 to 2017 by using three soil moisture products–ERA5, GLDAS and MERRA-2. Results showed that RZSM derived from three products presented a similar spatial pattern that the highest RZSM values occurred in tropical forest and cold areas, followed by subtropical, while the relatively low RZSM values were observed in arid and semiarid regions. Globally, RZSM decreased in all of three datasets with a rate of −0.14 × 10-3 m3 m−3 yr−1 on average (p < 0.001), which was largely correlated with temperature anomalies. Spatially, the RZSM trends greatly varied, with 21–31% of global land areas experiencing a significant decreasing trend and 7–24% for an increasing trend, respectively, confirming their different sensitivities to climate change. Temperature-driven RZSM dominated 19–29% of global land areas and was primarily distributed in northern high-latitude areas. The areas dominated by evapotranspiration were mainly in arid and semiarid areas, accounting for 29–44% of global land areas. Precipitation dominates the remaining 36–45% of global land areas mainly in eastern America and Europe, suggesting variations in the dominance of environmental factors on the spatial patterns of RZSM trend. Our findings will deepen our understanding of the impacts of climate change on the long-term trend of global soil moisture, and will be greatly critical to global soil water resource protection and management under the warming climate. [ABSTRACT FROM AUTHOR]

• A new laboratory method for estimating soil water retention is introduced. • The method only requires soil water content-dependent SWIR reflectance data. • Validation for 21 vastly different soils indicates a good performance. The soil water retention curve (SWRC) is an essential soil property that relates soil water content and matric potential. It plays a crucial role in soil water dynamics and the understanding of various hydrological phenomena at the land surface, including infiltration, runoff, evaporation, and energy exchange processes. In recent years, proximal sensing methods have shown great potential for retrieving this challenging-to-measure property from spectral reflectance. However, a physically-based approach is still lacking as current methods rely on empirical data-driven algorithms. Here we propose a novel physics-based laboratory method that, for the first time, enables direct estimation of the entire SWRC from saturated to dry using soil water content/reflectance data pairs within the shortwave infrared domain. The main hypothesis underlying the new method is that soil optical properties not only vary with soil water content but also with the pore scale distribution of capillary and adsorbed soil water. For evaluation, retrieved soil water retention curves of 21 soils that vastly differ in physical and hydraulic properties were compared to direct measurements. The results suggest that the new method is a rapid and efficient alternative to established laboratory measurement methods. [ABSTRACT FROM AUTHOR]

• SWDI is good at detecting severe/extremely droughts in humid region. • Sand are more prone to agricultural drought than clay. • Impact of groundwater on agricultural drought differ with soil textures. • 95th percentile is the best estimator of field capacity for clay and loam. • Min(maxgs) is the best estimator of field capacity for sand, loamy sand. Soil moisture is an important indicator for monitoring agricultural drought, and is closely related to soil texture. However, the role that soil textures play in agricultural drought evolution in humid climates is not well understood. In the study, we systematically assessed the feasibility of the Soil Water Deficit Index (SWDI) for agricultural drought monitoring in humid regions, examined the impact of soil texture and groundwater depth on agricultural drought evolution, and estimated the field holding capacity for different soil textures in humid areas based on a large-scale regional soil moisture monitoring network. The results show that: (1) The SWDI has reasonably good performance in detecting severe and extreme agricultural drought, but it tends to overestimate the drought intensity during non-severe/extreme drought periods in humid regions, particularly for clay soils. (2) Sand soils are more prone to agricultural drought than clay soils, which are more resistant to severe and extreme agricultural drought with high air temperatures and low precipitation. (3) During drought aggravation, clay soil showed a delay of about 0–18 weeks from the lowest SWDI to the maximum groundwater depth, while the delay was only 0–5 weeks for non-clay soil textures. (4) The 95th percentile is proposed as the best estimator of field holding capacity for clay and loam, while Min(maxgs) is the best estimator for sand, loamy sand, and sandy loam. Overall, these findings provide insights into the role of soil texture and groundwater depth in agricultural drought evolution in humid regions, and offer practical guidance for field holding capacity estimation for different soil textures. [ABSTRACT FROM AUTHOR]

• OpenKarHydro with multilayer soil water interactions was proposed. • OpenKarHydro reveals how climate and land use affect water budget and cause dry soil. • OpenKarHydro performs well in simulating the soil water content and runoff dynamics. • OpenKarHydro is a useful tool for land use optimization and controlling dry soil. In water-limited areas, soil water content (SWC) is a critical factor for plant growth and ecosystem stability. To understand the long-term effects of three different land uses (Vigna radiata : VR, Stipa bungeana : SB, and Medicago sativa : MS) and two climate scenarios (representative concentration pathway 4.5 (RCP45) and nonclimate change (Non_CC)) on water budgets and the mechanisms of dried soil layer (DSL) formation, a new complete multilayered model, OpenKarHydro, was developed. The results showed that the model adequately captured the water budget for different land uses, with root mean square errors (RMSEs) of 0.058, 0.046, and 0.046 and coefficients of determination (R2) of 0.615, 0.605, and 0.582, respectively, for SWC dynamics. This model performance was also demonstrated by the verification of runoff. Sensitivity analysis indicated that the soil porosity (n) was the most sensitive parameter in all land uses. The DSL was primarily caused by higher evapotranspiration (VR: 83.16% vs. 82.48%, SB: 80.62% vs. 77.69%, and MS: 79.93% vs. 76.26%) and canopy interception (VR: 11.92% vs. 11.99%, SB: 17.56% vs. 23.15% and MS: 22.60% vs. 26.19%) in the Non_CC climate pattern than in the RCP45 climate pattern. Climate change has caused an increase in evapotranspiration and interception for all land uses, reducing bottom leakage. Rainfall is sufficient to meet the water demand in the VR plot regardless of climate change; however, the additional water consumption caused by climate warming cannot be offset by the increased rainfall for the SB and MS plots. Compared to Non_CC, RCP45 resulted in increases of 0%, 54.19%, and 57.62% in proportion with extreme drought and decreases of 4.70%, 38.74%, and 5.24% in proportion with no drought for the soil profiles of VR, SB, and MS, respectively. For MS, RCP45 (2026) caused the DSL to penetrate the soil profile (>400 cm) 4 years earlier than in the Non_CC (2030), which should take no more than 5 and 9 years, respectively. OpenKarHydro is a useful tool for optimal water resource management and ecological restoration in water-limited areas. [ABSTRACT FROM AUTHOR]

• Proactive drought responses are crucial in a drought vulnerable future. • Antecedent conditions need to be considered in context of anticipated conditions. • We combine rainfall, soil moisture and evapotranspiration data with forecasts for early warning. • Validation efforts found promising accuracy in the drought early warning maps. • Proof-of-concept drought early warning system can aid proactive drought management. Droughts are a cyclical feature of Australia's climate that have compounding impacts on agricultural productivity and the wellbeing of drought-affected communities. Understanding anticipated drought conditions in context of antecedent observations is critical to providing early warning of drought. In this study we paired probabilistic seasonal rainfall forecasts with precipitation, soil moisture and evapotranspiration data that were objectively combined using Principal Component Analysis (PCA). The final Drought Early Warning System (DEWS) maps overlay forecasting information with the multi-variate PCA-weighted maps at 1-, 3- and 6-month timescales over the common period of overlap between all datasets (1982–2018). In this study period, the 1982–1983 Ash Wednesday tinder drought and the 1997–2001 Millennium drought were investigated. We validated PCA-weighted maps with satellite vegetation data and found performance was strongest over the Murray Darling Basin region (R = 0.63, p = 0.009) and poorest over Central interior Australia (insignificant correlations). We also validated PCA-weighted maps using agricultural commodity data from ABARES. Significant negative correlations at 95 %, 99 % and 99.9 % confidence intervals were found between %-Area in drought category and crop cultivation area; export volume/value; crop yield; and rural debt. Our findings indicate that early warning of drought can be categorised by concern – wherein dry antecedent conditions and dry forecasted conditions are of the highest concern. The developed proof-of-concept DEWS contributes to the growing body of proactive drought management research. In a drought vulnerable future, operationalising and communicating drought early warnings will be critical to reducing the harmful impacts of drought on communities, economies, and environments. [ABSTRACT FROM AUTHOR]

• Combined information from geophysics and soil cores delineated soil stratigraphic heterogeneity. • Time-lapse electrical resistivity monitoring revealed preferential infiltration paths. • Hydrological states derived from resistivity monitoring improved a surface–subsurface flow model. • The hydrogeophysical framework captured subsurface flow in response to flooding. This study presents a framework to quantify changes in soil saturation in response to flooding caused by extreme hydrologic perturbation on coastal ecosystems at the interfaces and transition between terrestrial and aquatic systems. Subsurface heterogeneity limits the use of in situ measurements to quantify subsurface flow during flooding due to the spatial discontinuity in the measured data. While geophysical methods, including time-lapse electrical resistivity imaging (ERI), are increasingly used to monitor soil hydrological processes, their abilities to parameterize flow models have been underutilized. This study combines background ERI, ground penetrating radar (GPR), time-lapse ERI, soil characterization, and a numerical flow model developed using an Advanced Terrestrial Simulator (ATS) code to quantify the infiltration pathway and describe the hydrological dynamics during a simulated flooding experiment. We assessed the use of two conceptual models developed using [1] ERI and GPR data that described the stratigraphic distribution, and time-lapse ERI that mapped permeability contrast, and [2] information from a national soil database for capturing changes in saturation. Combining the ERI and GPR results with soil core data revealed the stratigraphic heterogeneity at the site with a silty clay layer from 1 to 2 m between an overlying loamy topsoil and an underlying saturated silty sand. This silty clay layer could restrict deep infiltration. The time-lapse ERI showed up to a 35% decrease in resistivity, which correlated with soil moisture data (R2 value > 0.53) and revealed preferential infiltration zones used to inform the flow model. Numerical simulation results from both the geophysics- and soil database-informed models quantified changes in soil saturation with calculated soil moistures that agreed with field data. The geophysics-informed model captured more of the system's variability, reflective of shallow subsurface heterogeneities. The framework presented will serve as a precursor for a robust ecohydrological model that can describe the impacts of extreme events induced by climate change on coastal ecosystems. [ABSTRACT FROM AUTHOR]

• Lognormal, unsaturated hydraulic conductivity model not requiring K s as input. • K s model derived from hydraulic parameters describing bimodal soil water retention function. • The novel hydraulic conductivity model does not require the use of integrals. • The model considers the clay-water interaction without using clay content data. • Lognormal distribution function parameters were constrained for physical consistency. We developed a novel, lognormal, pore-scale, unsaturated hydraulic conductivity model, K(ψ) M odel , which does not require saturated hydraulic conductivity, K s , as an input parameter. K(ψ) M odel is derived solely from hydraulic parameters describing a bimodal, lognormal, pore-scale, soil water retention curve θ(ψ). The K(ψ) M odel is based on the Hagen-Poiseuille equation, which represents the soil as a bundle of parallel, non-intersecting capillary tubes. To improve the modelling of fine-textured soils we introduced a novel model to consider the clay-water interaction. This model assumes that clay-water interaction occurs for soils having more than 30% of clay and an effective matrix porosity greater than 35%. Compared to previously developed models, the K(ψ) M odel does not require the use of integrals and can be computed from a spreadsheet and distinguishes between macropore (non-equilibrium) and matrix (equilibrium) flows. The K(ψ) M odel gives improved results when the hydraulic parameters are dynamically constrained and when θ(ψ) describes a bimodal, lognormal distribution. [ABSTRACT FROM AUTHOR]

[Display omitted] • From spatiotemporal dimension to analyze drought propagation of SPEI to SSMI on the TP. • There is a lag time of 2–3 months from SPEI to SSMI and increases from summer to winter. • Climate, vegetation and permafrost type have an impact on drought propagation. • In winter, drought duration decreases in 45% of the TP after drought propagation, and drought severity decreases in 52%, particularly in the Inner TP. • The direction of propagation from SPEI to SSMI is predominantly northwest. Meteorological drought signals trigger different types of droughts by propagating in the water and energy cycle processes. The understanding of the propagation process from meteorological to soil moisture drought is not clear on the Tibetan Plateau (TP). We used the standardized precipitation index (SPEI) and the standardized soil moisture index (SSMI) to represent meteorological and soil moisture drought. By using Mann-Kendall (MK) trend test, wavelet analysis, run theory, Moran's I and drought migration model, we evaluated the trend of SPEI and SSMI on the TP from 1980 to 2018, analyzed the response relationship, explored the temporal (lag time (LT), propagation rate, duration (DD) and severity (DS)) and spatial (spatial autocorrelation, trajectory, direction and distance) properties of drought propagation, and discussed the potential factors. The result indicated that SPEI intensified at winter and annual scales, SSMI mitigated at all time scales. Both of them showed a consistent dry trend in the southeastern TP. In the time dimension, SSMI usually lagged behind SPEI by about 2–3 months, and the LT increased from summer to winter. The LT tended to be shorter in summer in regions characterized by humid and semi-humid climates, meadows and shrublands, and seasonally frozen ground. Propagation rates were greater in the southeast than northwest. After propagation, DD and DS became weaker in winter (45 % and 52 % of regions decreased, respectively) and stronger in summer (80 % and 76 % of regions increased, respectively). In the spatial dimension, the LT of adjacent regions is closely related, dominated by High - High values and Low-Low values. SSMI had shorter migration trajectories and slower migration rates than SPEI. In addition, the direction of drought propagation was predominantly northwest. For the first time, this study has provided insights into the spatial and temporal propagation of meteorological drought to soil moisture drought on the TP and provides a theoretical basis for understanding drought propagation and predicting soil moisture drought. [ABSTRACT FROM AUTHOR]

[Display omitted] • A new Spatially Promoted SVM (SP-SVM) Model proposed to downscale GRACE from 0.5° to 0.25°. • The most influential fifteen variables, i.e., satellite- and gauged-based, are considered in GRACE downscaling. • The proposed SP-SVM model was compared with a commonly used statistical SVM. Satellite-based terrestrial water storage changes have been recorded using the Gravity Recovery and Climate Experiment (GRACE) satellite which causing it an important dataset in hydrology and other related fields. GRACE dataset is widely utilized in many studies, but its coarse spatial resolution is a limiting drawback. Machine-learning approaches (e.g., ANN and SVM) are commonly applied in spatially downscaling. However, their input formation, which is in vector form, is a limitation of considering neighbor relations between the gridded-based inputs, specifically in spatial downscaling. Thus, developing an appropriate, simple, fast, and novel model to spatially downscale GRACE resolution is initially necessary for its utilizations. In this study, a Spatially Promoted Support Vector Machine (SP-SVM) model is innovatively proposed for GRACE downscaling from 0.5° to 0.25°. This promotion is investigated utilizing the distances between the unknown target points (with 0.25°) and their surrounding GRACE-valued points (0.5°), called their Distance Effect Coefficient (DEC), as the SP-SVM model input. In addition, the efficiencies of different in-situ and satellite-based datasets (fifteen variables from May 2005 to August 2020) are evaluated as the inputs of the GRACE downscaling models. After finding the most influential datasets, showing the best correlation with the GRACE, their best combinations in GRACE downscaling are identified. Based on the results, the set of PERSIANN-CDR without delay, the in-situ evaporation with a 1-month delay, and the soil moisture in 0–10 cm depth with a 1-month delay show the best performance in GRACE downscaling. The results of GRACE downscaling by the SP-SVM approach are also compared with the ones based on a usual statistical SVM (S-SVM) model, consisting of an intermediate bias interpolation to improve the estimations through a bias correction step. The results show that the SP-SVM model outperforms the common statistical SVM-based. Thus, compared with the usual S-SVM approach, the proposed SP-SVM (linear) model could be used as a simpler and more accurate model for downscaling any variable in a hierarchical process. [ABSTRACT FROM AUTHOR]

• Remote sensing data useful for forecasting streamflow where in-situ data lacking. • Multiple linear regression produces accurate seasonal streamflow forecasts. • Incorporating soil moisture and groundwater data improves streamflow forecast accuracy. • Soil moisture, groundwater contributions to forecast accuracy vary across basins. Water managers need improved seasonal streamflow forecasts in drought-prone regions like the U.S. Great Plains, where the inter-annual variability in rainfall and streamflow is high and seasonal streamflow forecasting systems are not well developed. Recent research has demonstrated that in-situ soil moisture measurements can be used to produce accurate streamflow forecasts in the Great Plains where rainfall runoff is the predominant source of streamflow. However, an inadequate density of in-situ soil moisture monitoring locations makes producing such forecasts impossible in many locations. Therefore, our objective was to develop and quantify the accuracy of seasonal streamflow forecasts for rainfall-dominated watersheds using remotely sensed data. Precipitation and remotely sensed soil moisture and terrestrial water storage anomaly data were incorporated into a multiple linear regression (MLR) model to produce seasonal streamflow forecasts for five watersheds in the Great Plains. The best MLR models had Nash-Sutcliffe Efficiency (NSE) values ranging from 0.71 to 0.93, indicating good to very good model performance. In one watershed with nearby in-situ soil moisture data, the use of remotely sensed soil moisture data led to greater forecast accuracy than did the inclusion of the sparse in-situ data. Our results show that the use of remote sensing data in MLR models can provide accurate seasonal streamflow forecasts in rainfall-dominated regions which currently lack such information. [ABSTRACT FROM AUTHOR]

• Strategies of increasing sediment observations for SWAT calibration were evaluated. • Strategies integrating different portions of estimated sediment data were designed. • Model with different strategies performed differently for stations. • Representativeness of parameters for different strategies was analyzed. • Adopting portions of estimated data during typhoons can improve sediment simulation. Sediment yields in a watershed are usually affected by various natural disturbances and anthropogenic activities, and can be simulated by using hydrological models. However, due to the availability and type of data (i.e., continuous or discrete) of measured sediment data, model calibration and validation performances could be significantly affected and the calibrated parameters may not be representative for the area. In this study, the sediment rating curves (SRCs) were developed to increase measurements of sediments for model calibration, and five strategies of using measured and estimated sediment data on improving a hydrological model (Soil and Water Assessment Tool, SWAT) calibration and validation for multiple sediment stations were analyzed. The five strategies were: using measured sediment data (S1), using measured data and different portions of estimated sediment during typhoon events (S2 to S4), and using estimated sediment data during entire simulation period (S5). The results showed that although the S1 mostly performed better than other sediment strategies, S4 and S5 were also suitable for improving sediment simulation at the downstream stations (STN1, 2, and 3) and upstream station (STN4), respectively. Moreover, the impact of incorporating different portions of typhoon-induced sediment data (S2, S3, and S4) on the model performance showed differently at stations. In sum, the proposed analytical procedure is expected to be useful for calibrating sediment parameters of hydrological models with limited or highly unevenly distributed measured sediment data. [ABSTRACT FROM AUTHOR]

[Display omitted] • The top-down QPEs perform better over the humid and low-altitude regions. • IMERG-F (SM2RAIN-ASCAT) performs best (worst) in detecting drought occurrence. • The drought dynamic evolutions highly depend on the drought indices used. • All QPEs capture drought centroid trajectory better on smaller spatial scales. Satellite-based quantitative precipitation estimates (QPEs) have shown great potential in the precipitation-related extremes estimation. However, little is known about the capability of different QPEs to monitor the dynamic evolutions of drought on various spatio-temporal scales. Here, the drought utility of two top-down QPEs (i.e., Climate Hazards Group InfraRed Precipitation with Stations, CHIRPS; Integrated Multi-SatellitE Retrievals for Global precipitation measurement-Final run, IMERG-F) and one bottom-up QPE (i.e., Soil Moisture to Rain-Advanced SCATterometer, SM2RAIN-ASCAT), is evaluated comprehensively in mainland China, based on three commonly used drought indices (Standardized Precipitation Index, Standardized Precipitation Evapotranspiration Index, Composite index of meteorological drought). The results indicate that the top-down QPEs generally show better performance in detecting drought occurrence over the humid and low-altitude regions in eastern China than the arid and high-altitude regions in western China. IMERG-F performs best in detecting drought occurrence in the entire mainland China, followed by CHIRPS, and the least satisfactory by SM2RAIN-ASCAT. The trajectory of drought centroids and the temporal variability of drought characteristics are highly dependent on both the drought indices selected and QPEs used. IMERG-F and CHIRPS can better reproduce the trajectories of large-scale drought events (e.g., the locations of drought origin and termination) than SM2RAIN-ASCAT. The ability in tracking the drought centroid trajectory of all three QPEs is higher at the smaller spatial scales, also it improves from the semi-arid to humid regions. Our results highlight the superiority of the top-down QPEs in drought monitoring compared with the bottom-up QPE in mainland China, particularly over the humid regions. The application of SM2RAIN-ASCAT in drought monitoring to the large-scale arid areas should be cautious. [ABSTRACT FROM AUTHOR]

[Display omitted] • BW and GW are interlinked through dissimilar physical processes across ecoregions. • GW-BW relationship and their shift were projected under extreme warm and dry events. • In Mountains, future BW declines while GW maintains due to legacy moisture. • In natural lands, BW to GW shift is projected with a lag time and legacy moisture. • In crop lands, remarkable shortages in BW are projected with no shift to GW. As key drivers of terrestrial ecosystems and food production, green water (GW) and blue water (BW) are interlinked through natural and anthropogenic processes. Their relationship has reportedly been non-stationary, with an anticipated shift from BW to GW in the future. However, these potential shifts are poorly characterized, particularly in response to future extreme warm-dry events. To address this gap, we used a process-based agro-hydrological model to project GW-BW interlinkages and their potential shifts across three different ecohydrological regions – mountainous, crop, and natural lands – in the Nelson River Basin, the largest agricultural basin in Canada and a major breadbasket for the world. Our results indicate dissimilarity in the physical processes that link GW and BW across different ecohydrological regions, and therefore, disproportional projected changes in BW and GW. In mountainous and natural lands, with projected decreases in precipitation in extreme warm-dry years of up to 56% and 46% under SSP126 and SSP585 scenarios, BW decreases by 64% and 42% while GW decreases only by 37% and 29%. The decline in GW is expected to be less because of the availability of evaporative water supplied by legacy soil moisture and groundwater contributions from earlier seasons. This available water to supply evaporative demand represents a delayed overall shift of BW to GW. In crop lands, however, results show a significant decrease of precipitation (45%-55%), BW (up to 42%), and GW (up to 42%), with no remarkable shift of BW to GW in extreme warm-dry years in the future. The projected BW and GW declines in crop lands may impose severe water shortages on local and regional food production. Projected declines in BW from the mountains that serve as source water for downstream irrigation may exacerbate such possible shortages. [ABSTRACT FROM AUTHOR]

• Weathered bedrock thicknesses were thicker than soils in loess hilly regions. • Infiltration density was multiplier increased under thicker weathered bedrock. • Average evaporation intensity was inverse to weathered bedrock thickness. • Water distribution and storage were optimized by weathered bedrock. • Rock water storage was functionally weighty in water-limited ecosystems. Weathered bedrock ecohydrology is indispensable for terrestrial ecosystem stability; however, the influence on the hydrological processes is poorly understood in loess hilly-gully critical zone. Using simulation experiments and in-suit observations, this study evaluated the intervention of weathered bedrock particles on profile infiltration and evaporation, investigated rock core sample characteristics and detected the spatiotemporal water variation in hillslopes of woodland and shrubland. Results showed that adding weathered bedrock particles interfered with infiltration and evaporation processes. Infiltration simulation discovered that stable infiltration rates increased linearly with weathered bedrock thicknesses, which accelerated by 0.5–5.7 times more than bulk soil layers. Interestingly, due to water separation at soil-rock interface, there existed nonlinear initial and average infiltration intensity fluctuations. Evaporation simulation demonstrated that cumulative evaporation presented a linear decrease with weathered bedrock, while the initial daily evaporation intensity increased, reaching a maximum of 15.4 ± 2.2 mm d−1. Moreover, borehole investigations manifested that weathered bedrock layers presented a thickness of more than 5.0 m, and the weathering intensity gradually decreased with depth, which was related to terrains and locations in hillslopes. Furthermore, profile water movement monitoring between woodland and shrubland slopes demonstrated that weathered bedrock optimized profile hydrological allocation. The average rock moisture reached about twofold than soil moisture, with average soil and rock moisture of 0.07 ± 0.04 cm3 cm−3 and 0.15 ± 0.05 cm3 cm−3 in woodland, 0.08 ± 0.03 cm3 cm−3 and 0.15 ± 0.03 cm3 cm−3 in shrubland, respectively. Besides, the profile water storage changed dynamically in equilibrium and escalated linearly with weathering layer thickness. Fundamentally, a conceptual structure model covering soil and rock water balance, vegetation dynamics, and climate change was constructed and discussed throughout the Earth's critical zone, which was then applied in describing water circulation participating by rock moisture under relative water shortage and abundance scenarios in loess hilly-gully regions. This study would furnish insights for improving and deepening the water cycle in shallow soils, and provide new horizons for enriching water resources utilization under climate change. [ABSTRACT FROM AUTHOR]

• Divergent shift of normal alpine meadows altered the soil and meadow properties. • Ks of topsoil significantly reduced by normal meadow transition into severely degraded meadow. • Ks of sub topsoil significantly reduced by normal meadow transition into shrub meadow. • Divergent shift towards shrub and degraded meadows both decreased water retention and storage. Soil saturated hydraulic conductivity (Ks) is a critical ecohydrological parameter for assessing the capacity of soil layers in grassland ecosystems to respond to precipitation efficiency, recharge groundwater and supply freshwater to rivers, which is significantly impacted by shifts in vegetation composition. Nevertheless, few studies focused on the effects of the shifts in vegetation composition on Ks of alpine grassland ecosystems. This study examined the effects of normal alpine meadow transition into shrub and severely degraded meadow stages on Ks and further qualified the driving processes. The results showed that the divergent shift of normal alpine meadows altered the soil and meadow properties. Soil water retention and storage capacity of the topsoil was reduced due to the divergent shift of the normal meadow transition into shrub meadow or severely degraded meadow. Ks and field moisture capacity (FMC) reduced by 73.1% and 64.9% in the 0–10 cm soil layer during the transition from normal to degraded meadow, while their variation in the 0–10 cm soil layer were not significant but water consumption increased during the shift from normal to shrub meadow. At the state of normal alpine meadow shift toward severely degraded meadow, the dominant factors affecting Ks were meadow cover (MC), root mass density (RMD), clay content and total porosity, with total effect coefficients of 0.48, 0.42, 0.39 and 0.69, respectively. Meanwhile, when normal alpine meadow shift toward shrub meadow, dominant factors affecting Ks were MC , non-capillary porosity, clay and sand contents, with total effect coefficients of 0.47, −0.86, −0.54 and −0.50, respectively. This study helps understanding the impacts of the shifts in vegetation composition on the ecohydrological processes in alpine meadow ecosystems in response to climate change and overgrazing. [ABSTRACT FROM AUTHOR]

• The first study to comprehensively analyze and compare the long-term variations of global surface soil moisture among multiple datasets. • Pervasive surface soil moisture drying is found across the globe, especially in humid-arid transitional regions, in most datasets. • Warming plays a primary role in global surface soil moisture drying in most datasets, especially GLEAM. Accurate soil moisture datasets are essential to understand the impacts of climate change. However, few studies have evaluated the consistency and drivers of long-term trends in soil moisture among different dataset types (satellite, assimilation, reanalysis, and climate model) at the global scale. Here we analyze the spatio-temporal variations of global surface soil moisture and associated climate dynamics over 1980–2020 using multiple soil moisture datasets, i.e., multi-satellite assimilated remote sensing datasets (ESA CCI), simulated soil moisture based on LSMs (GLDAS, GLEAM, CMIP6), and reanalysis (ECMWF ERA5, MERRA2, CRA-Land). Most of these datasets indicate pervasive drying of global surface soil moisture over the last four decades. Prominent soil moisture drying is detected in North America, Europe, northeastern Asia, North Africa, and the Arabian Peninsula. The cross-correlations among the five synthetic soil moisture datasets are the highest between GLEAM and the reanalysis datasets. Using the Aridity Index (AI, the ratio between annual total precipitation and potential evapotranspiration), we find that soil moisture drying is the most intensive in the humid-arid transitional regions with AI ranging 0.8–1.2. Surface soil moisture drying is primarily driven by increases in temperature, followed by ENSO, as indicated by Maximum Covariance Analysis (MCA). However, the significance of the impact of ENSO on soil moisture variability is sensitive to the choice of soil moisture dataset used in the MCA. [ABSTRACT FROM AUTHOR]

• Model quantifying hydrological effects of moss and fallen leaves was presented. • Moss and pine needles were considered as mulch rather than part of vadose zone. • CoreKarHydro performed well for simulating the soil water content curves. • CoreKarHydro is useful to understand effects of moss on soil evaporation. Due to recent vegetation restoration projects in karst mountainous areas of China, the living moss and litter layer play important roles in controlling soil water dynamics in karst forests. However, the underlying mechanisms of the effects of moss and the litter layer on soil water dynamics are still controversial. To explore the underlying mechanisms, we employ the CoreKarHydro model, capturing soil water content dynamics driven by evapotranspiration and influenced by living moss and fallen leaves. This model integrates surface changes' impact on evapotranspiration, categorizing the domain into: no cover (S), moss (SM), pine needles (SP), and both (SMP). Model validation employs the MOCOM-UA method, using diverse metrics (RMSE , R2 , ACC , ED , KGE and SA) against lysimeter data with 4 moss biomass levels and 3 pine needle biomass levels. The results indicated: 1) The CoreKarHydro model exhibited reliable performance in simulating the soil water content curves, with an RMSE ranging from 1.49% to 4.21% and an R2 value ranging from 0.60 to 0.77. 2) Among the factors that influenced evaporation processes, soil hydraulic properties emerged as the most sensitive parameters, whereas the properties of the moss and pine needles played a relatively lesser role. 3) The impact of moss cover on evaporation processes varied depending on the soil texture and the stage of evaporation. Moss cover reduced evaporation loss in 8 out of 12 soil textures (loamy sand, loam, silt, silt loam, sandy clay loam, silty clay loam, sandy clay, and silty clay). On the other hand, moss cover increased evaporation loss in 4 out of 12 soil textures (sand, sandy loam, clay loam, and clay). This study contributes to the understanding of the mechanisms of coupled soil-moss(litter)-hydrological processes in karst areas. [ABSTRACT FROM AUTHOR]

• Established a precipitation manipulation experiment under a plantation (∼ 4000 m2). • Investigated DSLs response to precipitation gradients in the 21-m soil profiles. • Rainfall accession alleviated mean DSLFD (1.8 m compared with interception's 1 m). • Mean DSL-SWC responsed greater under the PAT (0.2%–0.9%) than under the PIT (0.1%–0.3%). • Increased rainfall in turn exacerbated DSLs during the transition period. Ongoing climate change alters the amount and frequency of precipitation, deteriorating the water status of soils and the ecosystem services that they provide. The decline of soil water formed the dried soil layers (DSLs), which is particularly relevant for dryland ecosystems like the forests planted on the Chinese Loess Plateau (CLP), because there is an increased water demand for vegetation transpiration. To assess the impacts of precipitation changes on DSLs, we conducted a large throughfall manipulation experiment comprising seven precipitation gradients, ranging from 30%, 50%, and 65% precipitation interception treatments (PIT) to the ambient condition and then to 30%, 50% and 65% precipitation accession treatments (PAT), under a Robinia pseudoacacia planting forest on the CLP. We measured changes in soil water content (SWC) and examined the development of DSLs for over a year. We found that: (1) mean SWC in the 0–2100 cm soil profile and in DSLs decreased by 0.44%–0.77% and 0.13%–0.36% under the PIT, respectively, while increasing by 0.13%–0.36% and 0.16%–0.92% under the PAT, respectively; (2) DSLs were observed c. 100 cm and c. 100–180 cm below the soil surface under the PIT and PAT, respectively; (3) DSLs varied seasonally, with the largest reduction occurring during the transition period between the dry and rainy seasons, while the reduction developed more rapidly under the PAT than the PIT; (4) the incremental rate of SWC in DSLs increased linearly (p < 0.05) with increasing precipitation, while the increase was faster when precipitation was above (PAT) than below the ambient conditions (PIT), indicating greater precipitation sensitivity of DSLs under enriched than limited water supply. These changes in DSLs provide experimental evidence for the effects of climate change on soil water, and may have strong effect on the sustainable development management of drylands. [ABSTRACT FROM AUTHOR]

Hillslopes are the fundamental building blocks of catchments, however, fully disentangling their hydrological response remains an outstanding challenge. In natural settings, hillslope hydrology is challenged by the interplay of several factors, such as, heterogeneous morphology, vegetation, and complex soils. Herein, we attempt to clarify the mechanisms that control the rainfall-runoff relationship by investigating the response of an artificial outdoor slope setup with uniform geometry, soil properties, topography, and in the absence of vegetation located in a Mediterranean climate. Based on the analysis of more than 100 storms across six consecutive hydrological years, our results demonstrate that either pre-event soil water content or rainfall conditions alone are not sufficient to explain the slope response. Seasonality played a major role through soil moisture conditions, and pre-event soil moisture, peak rainfall intensity and rainfall duration influenced the slope runoff response. Observations suggest that in fall and winter subsurface flow was the prevalent runoff mechanism, while infiltration excess overland flow mostly occurred in the dry season. Our findings show that continuous measurement of runoff could be beneficial to understand the peak rainfall-peak runoff relationship. • Rainfall-runoff studied in artificial slope through more than 100 storms. • Peak rainfall intensity, rain distribution and soil moisture influence runoff. • Occurrence of runoff before peak rainfall beneficial to predict peak runoff. [ABSTRACT FROM AUTHOR]

• MASCP considers the effect of ponding by introducing a developing saturated zone. • MASCP reveals scaling relations of infiltration for sloping and horizontal surfaces. • New numerical algorithm can model slope infiltration with non-uniform initial SWC. Slope rainfall infiltration is a critical component of the Earth's hydrologic processes. Although various infiltration models have been proposed to simulate rainfall infiltration, simple and accurate ones are still scarce. This work developed a non-integral infiltration model for sloping surfaces, called modified approximate solutions under constant pressure conditions (MASCP). New expressions were proposed to improve slope infiltration simulation by introducing a developing saturated zone in soil water content (SWC) profile to quantify the role of ponding water during infiltration. The MASCP was then extended to simulate the infiltration with uniform and non-uniform initial SWC distributions under complex rainfall patterns. The model performance was evaluated by comparing with Hydrus-1D and the sloping Green-Ampt model for four distinct textured soils. The results demonstrated that the ponding time and infiltration rate simulated by the MASCP agreed well with those by Hydrus-1D at all given conditions. Besides, the MASCP yielded higher simulation accuracy than the sloping Green-Ampt model, especially for coarse-textured soils, low rainfall intensity, and long rainfall hiatus. In conclusion, the MASCP is a simple and accurate computation tool for slope rainfall infiltration modeling. [ABSTRACT FROM AUTHOR]

[Display omitted] • An LSTM model outperforms a physics-based SWMM model in predicting recession rates. • A novel hydrological database provides continuous data for these models. • LSTM models can be powerful tools for GSI performance and maintenance needs forecasting. The expected performance of Green Stormwater Infrastructure (GSI) is typically quantified through numerical models based on hydrologic parameters and physics-based equations. With numerical models, the choice of a spatio-temporal discretization scheme for the computational domain is a strenuous task that requires extensive calibration and potentially lab-based parameters and experimentation. The performance of GSI has high temporal dynamics due to natural, anthropogenic, and climatic processes that are not well represented by the traditional physics-based hydrologic models, which are calibrated against only a few historical observations and have a user-defined and constrained set of computational outcomes. Deep learning-based predictive models, such as Long Short-Term Memory (LSTM) neural networks, offer an exciting opportunity to quantify GSI performance, accounting for its highly dynamic and constantly evolving nature by leveraging advancements in observational data. A LSTM regression can overcome some of the limitations associated with traditional hydrological models to aid the development of a fully data-informed GSI performance predictor. To demonstrate the LSTM and traditional model outcomes, both methods were applied to a rain garden in Villanova, PA, USA. Specifically, an LSTM model was used to predict the recession of ponded water depth in the rain garden using five years of observed data. A set of eight predictors (i.e., precipitation, air temperature, soil temperature, soil moisture content at a depth of 10 cm, 35 cm, 65 cm and 91 cm, water depth) and a target variable (i.e., recession rate) were considered for training/testing the LSTM model. A comparative study with the USEPA Storm Water Management Model (SWMM) was performed to observe the performance of a physics-based model and a LSTM model for continuous recession rate time series and specific storms. The LSTM model had a performance score, Root Mean Square Error (RMSE), of 0.081 for the continuous time series, outperforming the SWMM performance with a score of 2.173 when compared to observed data. In the case of storm-specific prediction, LSTM also outperformed SWMM simulation for four storms with lower RMSE values when compared to the observed data. The application of the LSTM model in predicting GSI performance is a crucial stride towards efficient real-time forecasting. [ABSTRACT FROM AUTHOR]

• Contrasting saturated hydraulic conductivity led to soil stratification at 30–50 cm. • Soil moisture was more correlated within layers above and below the interface. • Above the interface, preferential flow and soil moisture dynamics were more active. • Threshold was identified in runoff generation, controlled by soil moisture storage. Soil moisture is essential in runoff generation, chemical transport, vegetation growth, and geological hazards like landslides in mountainous regions. However, due to its difficulty of measurement and its significant spatial heterogeneity and temporal variability, the dynamics and influence of soil moisture in the rainfall-runoff process are still not fully understood. In this study, precipitation, runoff, and soil moisture at multiple depths were measured and analyzed in a small, post-seismic mountainous catchment. Our results suggest that the dynamics of soil moisture could be explained by the characteristics of soil. The drastic change and the anisotropy of saturated hydraulic connectivity in the upper soil of the loose deposits in our catchment would lead to soil stratification at depths between 30 cm and 50 cm. The temporal variability of soil moisture was more highly correlated within the soil layers above and beneath the interface. Above the interface, besides vertical infiltration, occurrences of preferential flow were also frequent, transient perched water table may emerge during the wet period, making it a more active runoff contributor than the soil layers below the interface. When the level of rainfall reached the threshold, the correlation between rainfall and runoff became linear. This threshold response was more significant at soil layers above the interface, especially at the site with large topographic gradient. High antecedent soil moisture could skew the partitioning of precipitation into discharge, leading to an earlier flow peak. Our findings provide guidance on observation spot selection and the soil properties needed for the interpretation of future measurements, and may help develop flood early warnings system in small ungauged mountainous catchments with high flood risk. [ABSTRACT FROM AUTHOR]

• Funneling ratio of live C. korshinskii branches was 1.8 times higher than dead ones. • Funneling ratio of dead S. psammophila branches was 1.6 times higher than live ones. • Rainfall intensity significantly affected stemflow intensity and time lags to rain. Shrub mortality due to self-thinning and climate change is a natural consequence of the long residence times of standing dead branches in unmanaged shrub plantations. Stemflow (SF) represents a key point source of input water and nutrients into the soil in shrub ecosystems. However, little information is available about how shrub mortality affects SF variables. In this study, we measured the SF volume (SFV), funneling ratio (FR), SF intensity, time lag to rain, and rainfall characteristics for live and standing dead branches of two major shrub species comprising Caragana korshinskii and Salix psammophila during the rainy season in 2022 on the Chinese Loess Plateau. The results demonstrated that the SF variables and temporal dynamic processes differed between standing live and dead branches of the two shrubs. For C. korshinskii , the mean SFV and FR were 1.9 and 1.8 times higher for live branches than dead branches, respectively, whereas for S. psammophila , the mean SFV and FR were 1.3 and 1.6 times higher for dead branches than live branches. These contradictory results can be explained by differences in the canopy structures of the species, such as the surface area that contributes to the generation of SF in the upper canopy, branch angles, and the degree of branch bark decomposition and decay. In addition, the SF variables were significantly temporally dependent on the rainfall characteristics. SFV, SFI, and SFD were positively and linearly correlated with the rainfall amount, intensity, and duration at the inter-event scale (P < 0.001). The influence of rainfall intensity on SF time lag to rain were not significant at the intra-event scale (P > 0.05). Considering the enrichment ratio due to canopy dry deposition and bark decomposition, the periodic flattening of dead S. psammophila branches to restore their growth is less significant than allowing them to die and renew naturally. These findings may serve as a reference for developing management practices for shrub species in areas with soil water and nutrient deficiencies. [ABSTRACT FROM AUTHOR]

• Rainfall amount, duration, and intensity significantly affect Ec from diurnal to seasonal timescales. • Rainfall timing has a considerable impact on the magnitude, timing, and intra-daily distribution of Ec. • The responses of Ec to rainfall vary among canopy phenology periods. • The driving mechanisms of biophysical variables to Ec are more complex in days without rainfall events. Rainfall patterns affect vegetation water use through canopy transpiration (Ec) and are becoming increasingly variable due to global warming. However, Ec and its controlling mechanisms in boreal forests have rarely been investigated under various rainfall patterns. In this study, we measured the sap flow using thermal dissipation method in a boreal larch forest (Larix gmelinii) and observed biophysical variables simultaneously (e.g., air temperature, solar radiation, relative humidity, vapor pressure deficit, wind speed, soil water content, rainfall, and leaf area index) over two consecutive growing seasons (May–September) in 2021 and 2022. The responses of Ec to rainfall patterns were evaluated among three rainfall categories (classified based on rainfall amount, duration, and intensity), five types of rainfall timing, and two phenology periods (i.e., leaf expansion and defoliation periods). The results showed that daytime Ec significantly decreased, while nighttime Ec was less variable with increasing rainfall amount, duration, and intensity. The responses of Ec to rainfall categories with different rainfall amount, duration, and intensity were inconsistent at the monthly scale. The magnitude and diurnal dynamics of Ec were greatly influenced by the rainfall timing, and these effects got more considerable with increasing rainfall amount, duration, and intensity. Surprisingly, Ec showed opposite responses to rainfall in leaf expansion and defoliation periods, suggesting an enhanced impact in the leaf expansion period while a diminished impact in the leaf defoliation period. The significantly affected biophysical factors of Ec varied dramatically among rainfall categories. The interactions between biophysical factors and Ec were more complex and diverse in days without rainfall events than days with rainfall days. These findings highlight the importance of rainfall patterns in affecting Ec, provide new insight into the interactions between rainfall patterns, Ec, and biophysical factors, and are important supports for improving evapotranspiration simulations in the context of climate change. [ABSTRACT FROM AUTHOR]

• RF-SM framework outperforms the four coarse-scale SM products. • RF-SM framework is a robust ML algorithm for SM retrieval and downscaling. • The spatial pattern of RF-SM results has high spatial heterogeneity at the field scale. As a key variable used to characterize the climate process between the land surface and atmosphere, the surface soil moisture (SM) plays an irreplaceable role in the fields of hydrology, meteorology and agriculture. However, the relatively coarse spatial resolution of SM products currently limit the application in water resource management at the field scale. In this study, we proposed a high-spatial-resolution SM retrieval framework based on the random forest algorithm (RF-SM) to integrate in situ SM dataset from in the International Soil Moisture Network (ISMN), Landsat 8 optical and thermal observations, soil properties from SoilGrids V2.0, meteorological variables from the fifth generation of the European ReAnalysis (ERA5) dataset and four coarse-scale SM products including the Soil Moisture Active/Passive (SMAP), the Soil Moisture and Ocean Salinity (SMOS), European Space Agency Climate Change Initiative (ESA CCI) and U.S. National Climate Assessment Land Data Assimilation System (NCA-LDAS). Compared to three other machine learning (ML) algorithms [extremely randomized trees (Extra-Trees), gradient boosting regression tree (GBRT) and extreme gradient boosting (XGBoost)], the random forest (RF) algorithm exhibited the best performance against a subset of 100 validation sites with a Kling–Gupta efficiency (KGE) of 0.69 and root-mean-square-error (RMSE) of 0.063 m3/m3. In terms of different land cover types and typical sites, RF-SM also showed a better accuracy than any of the individual SM product. Finally, the retrieval framework was applied to map the 30-m resolution SM spatial distributions in five substudy areas in the U.S. The results suggest that it is feasible to retrieve accurate SM at a 30-m spatial resolution from multiple satellite datasets based on the RF algorithm, which has important practical significance for agricultural drought monitoring at the field scale. [ABSTRACT FROM AUTHOR]

• Soil moisture, GWSC, and TWSC in the YRB increased during 2003 to 2019. • A new weighted GRACE drought standardisation index (WGDSI) was reconstructed. • The interannual variability was beneficial for detecting drought in the YRB. • The recovery times were 30 and 39 months in 2006 and 2011 droughts in the YRB. • Different signal spectrums components of TWSC are useful for drought monitoring. With the intensification of global climate change, droughts have frequently been occurring in the Yangtze River Basin (YRB), significantly affecting the production patterns, lives, and socioeconomic development of humans. Drought characteristics must be comprehensively identified and quantified to reduce the damage caused by droughts in the YRB. The spatiotemporal variability of precipitation, runoff, soil moisture, groundwater change (GWSC) and terrestrial water storage change (TWSC) in the YRB was comprehensively estimated using the Gravity Recovery and Climate Experiment (GRACE), hydrological and in situ observations by decomposing the variabilities into seasonal, sub-seasonal, trend, and interannual. The new weighted GRACE drought standardized index (WGDSI) and weighted groundwater standardized index (WGWSI) were reconstructed by weighing the seasonal, trend, interannual and sub-seasonal of TWSC and GWSC. They were compared with the standardized soil moisture index (SSI06), standardized precipitation index (SPI06), and standardized runoff index (SRI06). Additionally, the drought characteristics identified based on the observations of the water storage deficit, severity, peak, duration, and recovery time in the YRB were also quantified using the WGDSI/WGWSI over the YRB. The results indicated that soil moisture, TWSC, and GWSC in the YRB increased from 2003 to 2019, mainly based on seasonal and interannual signals. The correlation coefficients between the WGDSI and the SSI06, SPI06, and SRI06 at the monthly were 0.84, 0.56, and 0.75, respectively, representing increases of 9 %, 14 %, and 21 % compared to that with the unweighted GRACE drought standardized index, respectively. The interannual variability of the hydrologic variables was more consistent with the drought events in the YRB, which was beneficial for detecting droughts. Two severe serious droughts occurred in the YRB from 2003 to 2019. In 2006, a continuous seven-month-long drought occurred, with a peak at − 28.974 km3, a severity of − 174.767 km3∙month, an average drought recovery rate of 0.83 km3/month, and a recovery time of 30 months, while in 2011, a continuous five-month-long drought occurred, with a peak at − 18.384 km3, a severity of − 78.106 km3∙month, an average drought recovery rate of 0.40 km3/month, and a recovery time of 39 months. The results indicated that the WGDSI could effectively identify drought events and measure their severity accurately, offering a more holistic understanding of drought conditions compared to other available indices. The index proposed in this study can be applied to generate new datasets and methods for quantifying and detecting droughts on a global scale. [ABSTRACT FROM AUTHOR]

• Rainfall losses are a key input to event-based flood models widely used in practice. • Relationship between rainfall losses and antecedent soil moisture are quantified. • Climate change projections of soil moisture are used to project rainfall losses. • Both the magnitude and variance of future rainfall losses are projected to increase. • Climate change will affect rainfall losses and hence flood responses. Climate change is expected to impact the severity and frequency of floods, which has implications for flood risk management. Design floods and derived flood frequency curves obtained using event-based rainfall-runoff models are widely used in industry to assess flood risks for planning and design purposes. For these approaches it is necessary to have (a) rainfall inputs, and (b) rainfall losses specified, the latter representing the amount of rainfall that is either intercepted, stored on the surface, or infiltrated into the soil and does not contribute to the flood hydrograph. There is extensive research on changes in flooding under climate change that focus on projections of rainfall. However, there is little research into projections of rainfall losses under climate change, despite the knowledge that their changes will modulate the flood response. Here, we present one of the first studies seeking to quantify how rainfall losses, as represented by estimates of initial and continuing losses used in event-based models, are projected to change under climate change. We identify dependencies between rainfall losses and antecedent soil moisture in around half (over 200) of the largely unregulated catchments (i.e. watersheds) in Australia analysed in this study and use these relationships to project rainfall losses under climate change. Near universal increases in both the mean and variance of both initial losses and continuing losses are projected in these catchments, suggesting that increased rainfall losses could offset the impact of increased rainfalls for frequently occurring floods and result in an increased variance in flood responses. [ABSTRACT FROM AUTHOR]

• A Soil-Water Conservation parameterization scheme was developed based on the GBEHM. • It revealed the spatiotemporal hydrological changes in the YRB during 1982–2019. • Impacts of the climate change and SWC measure on the runoff change were quantified. Intensive Soil and Water Conservation (SWC) has been conducted in the Yellow River basin (YRB) since the 1950s, which has greatly altered the eco-hydrological processes and particularly led to the runoff reduction. However, previous hydrological models performed in the YRB rarely simulated the SWC explicitly to quantify its spatiotemporal impacts. This study developed a SWC parameterization scheme based on the distributed physically-based model named Geomorphology-Based Ecohydrological Model (GBEHM). The hillslope SWC (HSWC) was parameterized as an additional surface storage capacity and simulated together with hillslope hydrological processes. The check dams along the river networks (i.e., river-networks SWC, RSWC) were parameterized as reservoirs and simulated together with the flow routing. The enhanced model (GBEHM-SWC) was first calibrated and validated using natural streamflow and remote sensing-based evapotranspiration. Then it was applied to analyze the hydrological changes from 1982 to 2019 in the YRB. Our simulation showed that, compared with the first period (1982–2000), the climatic factors (especially an increase of annual precipitation by 27.3 mm) would cause an increase of annual natural runoff by 2.9 mm (2.17 billion m3) in the second period (2001–2019). But in fact, the observed annual natural runoff decreased by 4.3 mm in the second period, which was caused by the reduction effects of the HSWC and RSWC by 3.8 and 3.7 mm (2.84 and 2.74 billion m3), respectively. Regarding the spatial distribution, the simulated natural runoff increased above Lanzhou and decreased on the Loess Plateau. The SWC impacts showed high spatial heterogeneity over the YRB, which offered useful information to evaluate the SWC measures. [ABSTRACT FROM AUTHOR]

• A surrogate model for distributed hydrological models is developed using deep networks. • The surrogate model can handle high-dimensional input–output maps of VIC. • The surrogate model can capture the dynamic relationship between the time-varying input and output. • The surrogate model outperforms LSTM in approximating spatial patterns of VIC output. Surrogate modelling methods have been used to estimate spatially distributed parameters of land surface hydrological models due to their computational efficiency. However, traditional surrogate techniques have trouble in modelling high-dimensional input–output maps. In addition, their applications to a distributed model over large spatial and temporal domains are very unlikely to produce consistent spatial pattern and dynamic relationship of model input and output, leading to unbalanced model performance across different spatiotemporal domains. To address these issues, we employ a deep autoregressive neural network to construct an accurate and reliable surrogate system of distributed dynamic model outputs. The network uses a deep convolutional encoder-decoder architecture to take full advantage of the deep convolutional layers in high-dimensional image processing and spatial feature extraction. Moreover, the autoregressive strategy makes the network good at handling dynamic relationship between the time-varying model input and output. We apply this deep learning (DL) network to building a surrogate model of the Variable Infiltration Capacity (VIC) model simulated soil moisture over the Huaihe river basin of China. The results show that the adopted deep autoregressive neural network can provide an accurate surrogate model of the high-dimensional dynamic relationship between 15 input fields and 1 output field, each covering 408 grids, over multiple years. This surrogate model significantly outperforms the popular long short-term memory (LSTM) model, achieving an average mean squared error (MSE) of 0.26 and an average R 2 of 0.76 using the test datasets. The surrogate's MSE and R 2 performance represents an average improvement of 43% and 33%, respectively, compared to the LSTM model. In addition, it shows better spatial performance than the LSTM model, demonstrating its superior performance in approximating spatial patterns of VIC simulated soil moisture. The accuracy of this surrogate method and its ability to handle high-dimensional data is promising for advancing parameter uncertainty quantification of distributed land surface hydrological models, which usually have high dimensionality due to the large number of variables involved, including distributed parameters, input forcing variables, and output variables. [ABSTRACT FROM AUTHOR]

• The variations of unfrozen water in seven soils are studied from −80 to 0 °C. • The water phase transition process is divided into three phase transition zones. • A prediction model is proposed by introducing residual unfrozen water. • The model has good applicability to different soils in a wide temperature range. The unfrozen water content in frozen soil is the critical factor for liquid water migration, frost heaving and thawing settlement during freeze–thaw process, and its content largely depends on soil texture and temperature. Herein, the unfrozen water content in different soils was measured by low-field nuclear magnetic resonance (NMR) at extremely low temperature ranging from −80 to 0 °C. The effect of soil properties on the variation of pore water in soil during freezing is analysed by combining T 2 relaxation curve and freezing characteristic curve. The results indicate that soil properties affect the phase transition process between liquid water and solid ice in frozen soil, and the pore size distribution plays a leading role in capillary water freezing in the severe phase transformation zone. When unfrozen water develops from the transitional phase transformation zone to the frozen stability zone, the freezing of adsorbed water in soil is affected by clay content. In addition, by analogy with the soil freezing characteristic curve, a prediction model of unfrozen water content of soil is established in accordance with the phase transition process of liquid water, and the accuracy of the model is further confirmed by comparing the test results. Sensitivity analysis of the proposed model parameters revealed that the parameters are closely related to the pore size distribution and the clay content. The proposed model expands the response range between the variation of unfrozen water and temperature by experimental means, and realizes the prediction of unfrozen water content in different types of soil at extremely low temperature. This study would provide significant reference for exploring the heat-moisture migration and mechanical strength in frozen soil. [ABSTRACT FROM AUTHOR]

• The P-EMA model parameters are linked to soil water retention curves. • Parameter m of the van Genuchten model is calculated from the critical water content of the P-EMA model. • Parameter α of the van Genuchten model is estimated from soil properties and P-EMA parameters. • Three approaches to estimate SWRCs from λ(θ) measurements are evaluated. • The λ(θ) approaches provided reliable estimates of SWRCs for six independent soils. A soil water retention curve (SWRC) describes the relationship between soil water content (θ) and suction (h , also the absolute value of pressure head). Earlier work indicated that correlations existed between the percolation-based effective medium approximation (P-EMA) thermal conductivity (λ) model parameters and soil hydraulic properties. In this study, the critical water content (θ c) of the P-EMA model was related to the pore size distribution parameter (m) of the van Genuchten model, water content at the inflection point of a SWRC (θ i) and hydraulic continuity water content (θ hc). And a pedo-transfer function was established to estimate the van Genuchten model parameter α from soil properties and P-EMA parameters. Based on these relationships, three approaches were developed to estimate the van Genuchten models parameters from λ(θ) measurements, porosity, sand and clay contents. The three approaches were then validated on six independent soils, and results showed that all of the approaches estimated θ well at selected h values, with the average root mean square errors from 0.025 to 0.029 cm3 cm−3, the average mean relative absolute errors ranging from 0.111 to 0.157, and the average Akaike Information Criterion from −18.3 to −16.2. Two new approaches outperformed the original Fu et al approach but with fewer input parameters (no need for organic carbon content), thus also facilitating their broader application. [ABSTRACT FROM AUTHOR]

Highlights • Soil moisture sensing networks are typically deployed under grassland vegetation. • Conditions under grassland can differ greatly from those under adjacent cropland. • A neural network translated grassland soil moisture into cropland soil moisture. • This efficient new approach does not rely on temporal stability or complex models. • The approach revealed the dichotomous soil moisture pattern within a SMAP grid cell. Abstract Large-scale in situ soil moisture monitoring networks are becoming increasingly valuable research tools, but existing networks feature almost exclusive deployment of stations in grassland vegetation. These grassland soil moisture observations are unlikely to adequately represent the real soil moisture patterns in landscapes with intermixed land cover types. Here we demonstrate the severity of the problem for one particular landscape and introduce a flexible new method for solving the problem. The specific objectives of this study were (i) to compare root-zone soil moisture dynamics of two dominant vegetation types across Oklahoma, grassland (observed) and winter wheat cropland (simulated); (ii) to relate the soil moisture dynamics of grassland and cropland vegetation using an artificial neural network (ANN) as an observation operator; and (iii) to use the resulting ANN to estimate the soil moisture spatial patterns for a landscape of intermixed grassland and wheat cropland. Root-zone soil moisture was represented by plant available water (PAW) in the top 0.8 m of the soil profile. PAW under grassland was calculated from 18 years of soil moisture observations at 83 stations of the Oklahoma Mesonet, whereas PAW under winter wheat was simulated for the same 83 locations using a calibrated and validated soil water balance model. Then, we trained an ANN to reproduce the simulated PAW under winter wheat using only six inputs: day of the year, latitude and longitude, measured PAW under grassland, and percent sand and clay. The resulting ANN was used, along with grassland soil moisture observations, to estimate the detailed soil moisture pattern for a 9 × 9 km2 Soil Moisture Active Passive (SMAP) grid cell. The seasonal dynamics of root-zone PAW for grassland and winter wheat were strongly asynchronous, so grassland soil moisture observations rarely reflect cropland soil moisture conditions in the region. The simple ANN approach facilitated efficient and accurate prediction of the simulated PAW under winter wheat, RMSD = 21 mm and normalized RMSD = 0.17, using observed PAW under grassland as an input. This new method for estimating soil moisture under adjacent, contrasting land covers at a relatively low computational cost could be employed for any region and land cover pairing with training data available, and it may significantly enhance the applications of existing large-scale soil moisture monitoring networks. [ABSTRACT FROM AUTHOR]