Antecedent Groundwater Research Articles

Artificial intelligence models to evaluate the impact of climate change on groundwater resources

This study develops three different artificial intelligence (AI) models in order to investigate the effects of climate change on groundwater resources using historical records of precipitation, temperature and groundwater levels together with regional climate projections. In particular, the Non-linear Autoregressive Neural Network (NARX), the Long-Short Term Memory Neural Network (LSTM) and the Convolutional Neural Network (CNN) were compared. Considering an aquifer located in northern Italy as a case study, the neural networks were trained to replicate observed groundwater levels by taking as input precipitation and temperature records, and in the case of the NARX also antecedent groundwater levels, on a monthly scale. The trained networks were used to infer groundwater levels until the end of the century based on precipitation and temperature projections provided by an ensemble of 13 Regional Climate Models (RCMs) from the EURO-CORDEX initiative. Two emission pathways were considered: the RCP4.5 and RCP8.5. All the AI models show good performance metrics during the training phase, but NARXs perform poorly compared to the other models during validation and testing. For the future, the NARX and LSTM models predict a decline in groundwater levels, especially for the RCP8.5 scenario, while slight changes are expected using the CNN. As NARXs are not deep learning techniques and CNNs may not be able to extrapolate values outside the training range, LSTMs appear to be better suited for climate change impact evaluations.

Groundwater Level Trend Analysis and Prediction in the Upper Crocodile Sub-Basin, South Africa

Disasters related to climate change regarding water resources are on the rise in terms of scale and severity. Therefore, predicting groundwater levels (GWLs) is a crucial means to aid adaptive capacity towards disasters related to climate change in water resources. In this study, a Gradient Boosting (GB) regression modelling approach for GWL prediction as a function of rainfall and antecedent GWL is used. A correlation analysis carried out from 2011 to 2020 demonstrated that monthly GWLs can be predicted by antecedent GWLs and rainfall. The study also sought to understand the long-term effects of climate events on groundwater levels over the study area through a Mann–Kendall (MK) trend analysis. A total of 50% of the groundwater stations revealed declining trends, while 25% had no trends and the other 25% showed an increasing trend. Again, the correlation analysis results were used in justifying the trends. The GB predictive model performed satisfactorily for all groundwater stations, with the MSE values ranging from 0.03 to 0.304 and the MAE varying from 0.12 to 0.496 in the validation period. The R2 ranged from 0.795 to 0.902 for the overall period. Therefore, based on projected rainfall and antecedent groundwater levels, future GWLs can be predicted using the GB model derived in this study.

Groundwater‐Mediated Memory of Past Climate Controls Water Yield in Snowmelt‐Dominated Catchments

AbstractAccelerating warming, changes in the amount, timing, and form of precipitation, and rapidly growing populations highlight the need for improved predictions of snowmelt‐driven water supplies. Although decadal‐scale trends in reduced streamflow are common, minimal progress has been made in improving streamflow prediction on the annual time scales on which management decisions are made. Efficient allocation of dwindling supplies requires incorporating rapidly evolving knowledge of streamflow generation into parsimonious models capable of improving prediction on seasonal, annual, and multiyear time scales of water resource management. We address this need using long‐term streamflow and climate records in 12 catchments averaging 90 years of observations and totaling more than 1,080 site‐years of data. These catchments experience similar regional climate forcing each year, but are diverse enough to represent broad ranges in precipitation, temperature, vegetation, and geology characteristic of much of the western US. We find that January baseflow across all catchments exhibits a coherent, quasi‐decadal periodicity that presumably is indicative of groundwater response to decadal climate. Although the direct contribution of this discharge to streamflow is small, interannual variability in groundwater discharge is a consistently strong predictor of runoff efficiency suggesting that antecedent groundwater storage alters precipitation routing to streamflow. Incorporating antecedent groundwater storage with precipitation and melt dynamics in multiple linear regression models reduces uncertainty in annual runoff from approximately 40% to <5%. These simple models, using readily available data, provide immediately useful tools for water managers to anticipate and respond to streamflow variability on time scales of 1 to 10 years.

Characterising hillslope–stream connectivity with a joint event analysis of stream and groundwater levels

Abstract. Hillslope–stream connectivity controls runoff generation, during events and during baseflow conditions. However, assessing subsurface connectivity is a challenging task, as it occurs in the hidden subsurface domain where water flow can not be easily observed. We therefore investigated if the results of a joint analysis of rainfall event responses of near-stream groundwater levels and stream water levels could serve as a viable proxy for hillslope–stream connectivity. The analysis focuses on the extent of response, correlations, lag times and synchronicity. As a first step, a new data analysis scheme was developed, separating the aspects of (a) response timing and (b) extent of water level change. This provides new perspectives on the relationship between groundwater and stream responses. In a second step we investigated if this analysis can give an indication of hillslope–stream connectivity at the catchment scale. Stream water levels and groundwater levels were measured at five different hillslopes over 5 to 6 years. Using a new detection algorithm, we extracted 706 rainfall response events for subsequent analysis. Carrying out this analysis in two different geological regions (schist and marls) allowed us to test the usefulness of the proxy under different hydrological settings while also providing insight into the geologically driven differences in response behaviour. For rainfall events with low initial groundwater level, groundwater level responses often lag behind the stream with respect to the start of rise and the time of peak. This lag disappears at high antecedent groundwater levels. At low groundwater levels the relationship between groundwater and stream water level responses to rainfall are highly variable, while at high groundwater levels, above a certain threshold, this relationship tends to become more uniform. The same threshold was able to predict increased likelihood for high runoff coefficients, indicating a strong increase in connectivity once the groundwater level threshold was surpassed. The joint analysis of shallow near-stream groundwater and stream water levels provided information on the presence or absence and to a certain extent also on the degree of subsurface hillslope–stream connectivity. The underlying threshold processes were interpreted as transmissivity feedback in the marls and fill-and-spill in the schist. The value of these measurements is high; however, time series of several years and a large number of events are necessary to produce representative results. We also find that locally measured thresholds in groundwater levels can provide insight into the connectivity and event response of the corresponding headwater catchments. If the location of the well is chosen wisely, a single time series of shallow groundwater can indicate if the catchment is in a state of high or low connectivity.

Long-term rewetting of degraded peatlands restores hydrological buffer function

Precipitation is a key factor affecting shallow water table fluctuations. Although the literature on shallow aquifers is vast, groundwater response to precipitation in peatlands has received little attention so far. Characterizing groundwater response to precipitation events in differently managed peatlands can give insight into ecohydrological processes. In this study we determined the groundwater table response rate following precipitation events at a drained and a rewetted fen to characterize the effect of rewetting on hydrological buffer capacity. Multiple regression analysis revealed that the groundwater table at the rewetted fen has more than two times lower rate of response to precipitation events than that of the drained fen, even after adjusting for antecedent groundwater levels. Thus, the rewetted fen delivers a better hydrological buffer function against heavy precipitation events than the drained fen. We found that for the depths at which the groundwater interacts with incoming precipitation, the peat of the rewetted fen has a higher specific yield causing groundwater to rise slower compared to the response at the drained fen. A period of 20 years of rewetting was sufficient to form a new layer of organic material with a significant fraction of macropores providing storage capacity. Long-term rewetting has the potential to create favorable conditions for new peat accumulation, thereby altering water table response. Our study has implications for evaluating the success of restoration measures with respect to hydrological functions of percolation fens.

A simulation model for estimating root zone saturation indices of agricultural crops in a shallow aquifer and canal system

Shallow aquifers significantly impact crop growth as saturated soil conditions may occur. Canals are widely constructed in such areas to mitigate groundwater saturation or flooding. We applied a simulation model to estimate the occurrence of root zone saturation [root zone saturation index (RZSI)] for agricultural crops and to identify factors that influence root zone saturation in a shallow coastal aquifer and canal system. Results indicated that groundwater modeling combined with multiple linear regression can relate the influencing factors and root zone saturation durations in low lying farmland adjacent to canal systems. In our study, most areas had a low RZSI, but areas towards the northwest and southeast where the land surface elevation is generally low were predicted to have a greater RZSI. In general, positive correlations were found between the root zone saturation durations and rainfall amount, antecedent groundwater table elevation and average canal stages in areas where the higher RZSIs were predicted. Rainfall amount played a more important role than antecedent groundwater table elevation and canal stage in determining the root zone saturation during the wet season, while antecedent groundwater table elevation and canal stage played a more important role than rainfall amount during dry season. Correlations between the predicted root zone saturation duration and land surface elevation were negative and stronger during the wet season than the dry season, and the correlations were stronger in the deep (0–61 cm) root zone than with the shallow (0–18 cm) root zone. In area where the land surface elevation is relatively high, the root zone saturation duration was not influenced by rainfall amount, antecedent groundwater table elevation or canal stage, at least under the current management practices and climate conditions.

The role of antecedent groundwater heads in controlling transient aquifer storage and flood peak attenuation in karst watersheds

AbstractTransient storage of floodwaters in aquifers is known to attenuate peak flows in rivers and drive subsurface dissolution. Transient aquifer storage could be enhanced in watersheds overlying karst aquifers where caves facilitate surface and groundwater exchange. Few studies, however, have examined controls on, or magnitudes of, transient aquifer storage or flood peak attenuation in karstic watersheds. Here we evaluate flood peak attenuation with multiple linear regression analyses of 10 years of river and groundwater data from the Suwannee River, which flows over the karstic upper Floridan aquifer in north‐central Florida and experiences frequent flooding. Regressions show antecedent river stage exerts the dominant control on magnitudes of transient aquifer storage, with recharge and time to peak having secondary controls. Specifically, low antecedent stages result in larger magnitudes of transient aquifer storage and thus greater flood attenuation than conditions of elevated antecedent stage. These findings suggest subsurface weathering, including cave formation and enlargement, caused by transient aquifer storage could occur on a more frequent basis in aquifers where groundwater table elevation is lowered due to anthropogenic or climatic influences. Our work also shows that measures of groundwater table elevation prior to an event could be used to improve predictive flood models. © 2018 John Wiley & Sons, Ltd.

The impact of wildfire on baseflow recession rates in California

AbstractThe effect of wildfire on peak streamflow and annual water yield has been investigated empirically in numerous studies. The effect of wildfire on baseflow recession rates, in contrast, is not well documented. The objective of this paper was to quantify the effect of wildfire on baseflow recession rates in California for both individual watersheds and for all the study watersheds collectively. Two additional variables, antecedent groundwater storage and potential evapotranspiration, were also investigated for their effect on baseflow recession rates and postfire baseflow recession rate response. Differences between prefire and postfire baseflow recession rates were modeled statistically in 8 watersheds using a mixed statistical model that accounted for fixed and random effects. For the all‐watershed model, antecedent groundwater storage, potential evapotranspiration, and wildfire were each found to be significant controls on baseflow recession rates. Wildfire decreased baseflow recession rates 52.5% (37.6% to 66.0%), implying that postfire reductions in above‐ground vegetation (e.g., decreased interception, decreased evapotranspiration) were a stronger control on baseflow recession rate change than hydrophobicity. At an individual watershed scale, baseflow recession rate response to wildfire was found to be sensitive to intraannual differences in antecedent groundwater storage in 2 watersheds, with the effect of wildfire on baseflow recession rates being greater with lower levels of antecedent groundwater storage. Examination of burn severity for a subset of the study watersheds pointed to riparian zone burn severity as a potential primary control on postfire recession rate change. This study demonstrates that wildfire may have a substantial impact on fluxes to and from groundwater storages, altering the rate at which baseflow recedes.

Physical modelling of rainfall-induced flow failures in loose granular soils

The tragic consequences of the March 2014 Oso landslide in Washington, USA were particularly high due to the mobility of the landslide debris. Confusingly, a landslide occurred at that exact same location a number of years earlier, but simply slumped into the river at the toe of the slope. Why did these two events differ so drastically in their mobility? Considerable questions remain regarding the conditions required to generate flow failures in loose soils. Geotechnical centrifuge testing, in combination with high-speed cameras and advanced image analysis has now provided the landslides research community with a powerful new tool to experimentally investigate the complex mechanics leading to high mobility landslides. This paper highlights recent advances in our understanding of the process of static liquefaction in loose granular soil slopes achieved through observations of highly-instrumented physical models. In particular, the paper summarises experimental results aimed to identify the point of initiation of the chain-reaction required to trigger liquefaction flow failures, to assess the effect of slope inclination on the likelihood of a flowslide being triggered, and to quantify the effect of antecedent groundwater levels on the distal reach of landslide debris with the objective of beginning to explain why neighbouring slopes can exhibit such a wide variation in landslide travel distance upon rainfall-triggering.

Influence of slope inclination on the triggering and distal reach of hydraulically-induced flow slides

In the absence of water, a dry granular landslide mass released on an incline will experience greater acceleration than an otherwise identical mass released on a plane of lower inclination. However, when water is introduced to the system, the outcome is more complicated. While it requires less water pressure to trigger a landslide on a steep slope, the additional saturated soil in a gentle slope can make it more susceptible to flow liquefaction, which in turn can make it the slope of greater risk in terms of the landslide velocity and distal reach. Results from seven geotechnical centrifuge tests were evaluated to compare the behavior of flow slides triggered in a loose granular soil layer inclined at 20° and 30°. The results indicate that the influence of slope inclination on mobility is highly dependent on antecedent groundwater conditions. In low antecedent groundwater scenarios, the mobility of the failures triggered in the 20° and 30° were similar. However, in a high antecedent groundwater scenario, a flow slide triggered on the 20° slope was observed to travel twice the distance, and reaching nearly twice the maximum velocity than a flow slide triggered on the steeper 30° slope. These results provide experimental data documenting the complex interaction between slope inclination and antecedent groundwater conditions to control the triggering and distal reach of hydraulically-induced flow slides.

Effect of antecedent groundwater conditions on the triggering of static liquefaction landslides

Real-time early warning systems for shallow landslides are typically built upon real-time measurements and forecasts of rainfall and empirical correlations between past patterns of rainfall and landslide occurrence. Whereas these relationships describe whether certain combinations of rainfall and preexisting groundwater levels are of elevated risk of landslide triggering, not all combinations leading to landslide events necessarily have the same consequences in terms of landslide mobility (velocity and distal reach of the landslide). In this paper, the technique of geotechnical centrifuge modeling is used to quantitatively evaluate the hypothesis that the mobility of a landslide triggered under elevated antecedent groundwater conditions is higher than scenarios under drier antecedent conditions. Five identical slope models with a shallow depth to bedrock were subjected to different antecedent conditions ranging from zero groundwater flux to nearly saturated conditions prior to rainfall. The results from these scenarios show that higher antecedent groundwater conditions can result in landslides with velocities about three times higher and travel distances about eight times higher than low antecedent conditions due to static liquefaction of the soil at the base of the slope.

Changes in groundwater induced by water diversion in the Lower Tarim River, Xinjiang Uygur, NW China: Evidence from environmental isotopes and water chemistry

summary The Lower Tarim River in NW China is under severe ecosystem degradation due to stopped stream flow and diminished groundwater recharge. Since year 2000, eight water diversions from the upper stream and from the neighboring Kaidu–Kongque River have been implemented to alleviate the ecosystem disaster. In order to assess the effectiveness of the water diversion project and to identify proper tracers of groundwater dynamics, we sampled the riparian groundwater system in 2007 and 2008 along the 350 km-long river channel through the 40 monitoring wells situated along nine transects perpendicular to the river and three soil profiles. Measurements on the samples have included environmental isotopes ( 18 O, 2 H, 3 H) and water chemistry. The results show that remarkable changes have been induced by the water diversions. The observed response of riparian groundwater system includes general decrease in total dissolved solid (TDS) and rise of water table. Scope with greater than 1 m rise in water table is within �700 m from the riverbank in the upper segments and �300 m in the lower ones. Greater rise of water table occurs near the river bank. Tritium data show that the extent of modern recharge (since 1960s), including that from the diverted water, is limited to 600 m from the riverbank at the upper segments and 200 m at the lower ones. Stable isotopes show that groundwaters, regardless of modern or pre-modern, are enriched in heavy isotopes and are plotted in parallel to the meteoric water line in the d–d plot, attributed to evaporation during recharge. Groundwater is generally of Na–Mg–Cl–SO4 type and is formed by dissolution of minerals, such as halite, sulfate, and carbonates, based on component correlation matrices analysis. The salinity of groundwater is mainly affected by that of the diverted water and of the local antecedent groundwater, salts in the unsaturated zone, evapotranspiration during recharge. As the zone of smaller groundwater depth (less than 5 m) suitable for the most existing Populus euphratica and Tamarix ramosissima, the main species targeted by the rescue effort, restricts to 200 m from the riverbank, and narrows down towards downstream, long-term stability of the ecosystem cannot be achieved by the current water diversion scheme and regulating/saving water in source-streams and the Upper/Middle Tarim River is crucial for continuing water diversion.

Selection of an appropriately simple storm runoff model

Abstract. An appropriately simple event runoff model for catchment hydrological studies was derived. The model was selected from several variants as having the optimum balance between simplicity and the ability to explain daily observations of streamflow from 260 Australian catchments (23–1902 km2). Event rainfall and runoff were estimated from the observations through a combination of baseflow separation and storm flow recession analysis, producing a storm flow recession coefficient (kQF). Various model structures with up to six free parameters were investigated, covering most of the equations applied in existing lumped catchment models. The performance of alternative structures and free parameters were expressed in Aikake's Final Prediction Error Criterion (FPEC) and corresponding Nash-Sutcliffe model efficiencies (NSME) for event runoff totals. For each model variant, the number of free parameters was reduced in steps based on calculated parameter sensitivity. The resulting optimal model structure had two or three free parameters; the first describing the non-linear relationship between event rainfall and runoff (Smax), the second relating runoff to antecedent groundwater storage (CSg), and a third that described initial rainfall losses (Li), but which could be set at 8 mm without affecting model performance too much. The best three parameter model produced a median NSME of 0.64 and outperformed, for example, the Soil Conservation Service Curve Number technique (median NSME 0.30–0.41). Parameter estimation in ungauged catchments is likely to be challenging: 64% of the variance in kQF among stations could be explained by catchment climate indicators and spatial correlation, but corresponding numbers were a modest 45% for CSg, 21% for Smax and none for Li, respectively. In gauged catchments, better estimates of event rainfall depth and intensity are likely prerequisites to further improve model performance.

Statistical Analysis of Groundwater Table Depths in Upper Swarnamukhi River Basin

Development and effective utilization of groundwater resources is essential in semi-arid regions for activities such as water supply and irrigation. The present study aims to analyze statistically the groundwater data at the Chinnagottigallu, Yerravaripalem, Chandragiri, Sriramapuram, Tirupati, Renigunta, Karvetinagaram and Yerpedu piezometric stations of upper Swarnamukhi river basin in the drought prone Rayalaseema region of Andhra Pradesh, India and to develop models through multiple linear correlation and regression analysis. The monthly rainfall and groundwater data at the raingauge and piezometric stations of the basin for the period 2001-2006 were collected from the Groundwater and Irrigation Departments of the region. It is observed from the analysis that the groundwater table depth in any period is influenced by the rainfall in the period and, the rainfall and groundwater table depth in the previous period. The study also reveals that the effect of antecedent groundwater table depth is more pronounced than that of rainfall and antecedent rainfall. The models proposed may be adopted for the estimation of groundwater table depths to effectively plan and efficiently manage groundwater resources of the basin.

The Influence of Irrigation on Slope Movements, Ventura County, California, USA

Ancient landslides at two locations in Ventura County, California, USA, that have morphology consistent with thousands of years of stability were reactivated in 1988 and 1992–1994. The timing of reactivation was assisted by the burial or destruction of oil wells that penetrated through the landslides and had operated continuously since at least 1948. Slope movement at Grimes Canyon began during a period of lower-than-normal rainfall (compared to average precipitation since 1948), which suggests that irrigation triggered instability. At South Mountain, landsliding initiation overlapped a period of intense rainfall; however, the influence of irrigation, which may have contributed antecedent groundwater conditions conducive to slope failure, may not be ruled out. These case studies illustrate the complexity of identifying landslide triggering mechanisms, and they show that triggering of deep-seated landslides may be aided by climatic conditions or irrigation.

Anomalous behavior of soil mantle groundwater demonstrates the major effects of bedrock groundwater on surface hydrological processes

The formation of groundwater in the soil mantle has a dominant effect on rainwater discharge and shallow landslide occurrence in headwater catchments. Here, we report two completely different groundwater responses within a single well excavated into the soil mantle. One was an ephemeral‐type response that is well described by physical hydrology models based on a geographic information system (GIS). The other was a semi‐perennial‐type response, rarely reported in previous studies, which cannot be explained by the existing physical models. The semi‐perennial groundwater caused considerably high antecedent groundwater tables between storms, leading to an increased peak in the groundwater level during later heavy storm events and a likely increase in the risk of shallow landslides. Moreover, peaks in the semi‐perennial groundwater lagged considerably behind rainstorm events, which probably affected base flow discharge by forming a delayed peak. Geochemical and geothermal observations indicated that the source of the semi‐perennial groundwater was deep bedrock groundwater, demonstrating the considerable effects of bedrock groundwater on surface hydrological processes.

Hydrochemical behaviour of dissolved nitrogen and carbon in a headwater stream of the Canadian Shield: relevance of antecedent soil moisture conditions

AbstractThis paper examines the impact of contrasting antecedent soil moisture conditions on the hydrochemical response, here the changes in dissolved nitrogen (NO3−, NH4+ and dissolved organic nitrogen (DON)) and dissolved organic carbon (DOC) concentrations, of a first‐order stream during hydrological events. The study was performed in the Hermine, a 5 ha forested watershed of the Canadian Shield. It focused on a series of eight precipitation events (spring, summer and fall) sampled every 2 or 3 h and showing contrasted antecedent moisture conditions. The partition of the eight events between two groups (dry or wet) of antecedent moisture conditions was conducted using a principal component analysis (PCA). The partition was controlled (first axis explained 86% of the variability) by the antecedent streamflow, the streamflow to precipitation ratio Q/P and by the antecedent groundwater depth. The mean H+, NO3−, NH4+, total dissolved nitrogen and DOC concentrations and electrical conductivity values in the stream were significantly higher following dry antecedent conditions than after wetter conditions had prevailed in the Hermine, although the temporal variability was high (17 to 138%). At the event scale, a significantly higher proportion of the changes in DON, NO3−, and DOC concentrations in the stream was explained by temporal variations in discharge compared with the seasonal and annual scales. Two of the key hydrochemical features of the dry events were the synchronous changes in DOC and flow and the frequent negative relationships between discharge and NO3−. The DON concentrations were much less responsive than DOC to changes in discharge, whereas NH$_{4}^ +$ was not in phase with streamflow. During wet events, the synchronicity between streamflow and DON or NO3− was higher than during dry events and discharge and NO3− were generally positively linked. Based on these observations, the hydrological behaviour of the Hermine is conceptually compatible with a two‐component model of shallow (DON and DOC rich; variable NO3−) and deep (DON and DOC poor; variable NO3−) subsurface flow. The high NO3− and DOC levels measured at the early stages of dry events reflected the contribution from NO3−‐rich groundwaters. The contribution of rapid surface flow on water‐repellent soil materials located close to the stream channel is hypothesized to explain the DOC levels. An understanding of the complex interactions between antecedent soil moisture conditions, the presence of soil nutrients available for leaching and the dynamics of soil water flow paths during storms is essential to explain the fluxes of dissolved nitrogen and carbon in streams of forested watersheds. Copyright © 2007 John Wiley & Sons, Ltd.

Runoff Generating Processes in Boreal Forest Environments with Glacial Tills

This review describes the processes by which runoff is generated from rain or snowmelt in a boreal forest landscape with a shallow layer of glacial tills overlying a relatively impermeable bedrock. Glacial tills are the main surface deposits in the Nordic countries, and are also of frequent occurrence in other boreal forest environments in the Northern Hemisphere. Soil moisture and groundwater conditions control the partitioning of rain or snowmelt into evapotranspiration, runoff and temporary storage. Groundwater is viewed as the major determinant of runoff generation, both because it is the most important contributor to event flow, and because of its influence on saturation overland flow. A characteristic feature of till deposits is a saturated hydraulic conductivity which decreases rapidly with depth. As a consequence, topography has a significant impact on runoff by controlling movement and storage of water through convergence and divergence of flow. Modeling and field experience point to the idea of a continuum in both spatial and temporal occurrence of saturation overland flow and subsurface flow within individual catchments under different rainfall and snowmelt events and antecedent groundwater and soil moisture conditions. Results from studies using hydrological models with a structure which may be suitable for landscapes with glacial tills are presented.