Spring Discharge Research Articles

Hydrogeological assessment of a major spring discharging from a calcarenitic aquifer with implications on resilience to climate change

Groundwater is a vital source of freshwater, serving ecological, environmental, and societal needs. In regions with springs as a predominant source, such as the Northern Apennines (Italy), resilience of these springs to climate-induced recharge changes is crucial for water supply and ecosystem preservation. In this study, Nadìa Spring in the Northern Apennines is examined through an unprecedented array of multidisciplinary analyses to understand its resilience and unique characteristics. The Nadìa Spring's exceptional response, characterized by a sustained base flow even in the face of drought, is attributed to a combination of factors including a substantial groundwater reservoir, a complex network of faults/fractures, slope instabilities, and karst dissolution. The investigation reveals a dual porosity system in the aquifer, consisting of fast-flow conduits and a diffuse fracture network. While fast-flow conduits contribute to rapid responses during high-flow conditions, the diffuse system becomes predominant during low-flow periods. This dual porosity structure helps the spring maintain a consistent base flow in the face of climate-induced recharge fluctuations. The study shows that Nadìa Spring exhibits remarkable resilience to year-to-year variations in recharge, as evidenced by stable minimum discharge values. While the spring has undergone a decline in discharge over the past century due to long-term climate change, it is becoming more resilient over interdecadal timescales due to transition to a diffuse drainage system that mitigates the impact of reduced recharge. The availability of a century-long spring discharge monitoring was a crucial piece of information for understanding the spring's discharge response and drawing conclusions about its long-term resilience to recharge fluctuations. Continuing long-term monitoring and research in the future will be essential to validate and expand upon these findings in the context of changing climatic conditions. This research serves as a model for assessing strategic groundwater discharge points in geological settings similar to the Northern Apennines.

PENGARUH TUTUPAN LAHAN DAN CURAH HUJAN TERHADAP SIFAT FISIK TANAH SERTA DEBIT MATA AIR DI HUTAN CEMPAKA, PASURUAN, JAWA TIMUR

Climate change causes forest fires that impact land cover damage, which can cause damage to soil properties and disrupt hydrological characteristics such as loss of spring or discharge depletion. This study aimed to assess the influence of land cover and rainfall on soil physical properties and spring discharge in the Cempaka Forest area. The study was conducted on three land covers, i.e., shrubs, pine, and non-timber forest products (NTFPs) in the Cempaka Forest area, Curah Tangkil Block, Pasuruan Regency, East Java Province. The measured parameters were soil bulk density, permeability, porosity, infiltration, daily rainfall, and spring discharge. Data analysis used analysis of variance (ANOVA) test at a 5% level, correlation test, and regression. The results showed that differences in land cover had no significant effect on soil physical properties. The correlation and regression results between rainfall and spring discharge showed that rainfall influenced spring discharge by 4%. The correlation and regression results in the medium-heavy rainfall categories influenced spring discharge by 38%.

The Influence of Land Cover and Socio-Economic Aspects on Umbulan Spring Discharge in Pasuruan Regency

The Influence of Land Cover and Socio-Economic Aspects on Umbulan Spring Discharge in Pasuruan Regency

Лишайники восточного склона вулкана Берутарубе, о. Итуруп (Курильские острова, Дальний Восток России)

The paper presents an annotated list of lichens of the eastern slope of Berutarube volcano, one of the most inaccessible and poorly studied active volcanoes on Iturup Island. Samples of lichens were collected in August 2023 from eroded areas in the summit part of the volcano, as well as on the eastern slope along the Andrey Creek, in the upper reaches of which unnamed cold acidic springs discharge. In total, 53 species of lichens were identified from the sampling sites, of which 20 species were new to Iturup Island, 2 species were new to the Kuril Islands, and 1 species was new to the Sakhalin region. In the summit part of the volcano, lichens typical for areas of volcanic activity were identified: Cladonia crispata, Cladonia straminea, Huea confluens, Lecidea plana, Melanelia stygia, Pseudephebe pubescens, Rhizocarpon badioatrum, and Umbilicaria torrefacta. The influence of the solfataric gases of the volcano on the lichen biota of the Andrey Creek valley has not been revealed, which allows us to make the assumption that there is no direct influence of the products of the solfataric activity of Berutarube volcano on the natural environment of its eastern slopes.

Tracking flowpaths in a complex karst system through tracer test and hydrogeochemical monitoring: Implications for groundwater protection (Gran Sasso, Italy)

Groundwater in karst aquifers is frequently tapped for drinking purposes, due to frequent huge volumes of resources. Unfortunately, vulnerability of these aquifers can be high, due to possible fast transfer of recharge water on springs by the karst network. On Gran Sasso Mountain regional aquifer, several springs are subjected to drinking withdrawal and an updated evaluation of their potential is now a fundamental issue to be considered, facing climate change effects, which reflect on variation of discharge regimen and values. To distinguish between different contribution of spring recharge, a tracer test has been carried out on the Vitella d’Oro spring, fed both by the regional aquifer and by a local system exposed to karst features developed in the Rigopiano Conglomerates formation. Thanks to hydrogeological, hydrogeochemical and isotopic data, a conceptual model of spring recharge has been proposed and subsequently validated by the tracer test results. All information confirms the superimposition on the regional base flow, by a relevant contribution of the karst network, influencing the spring discharge in recharge periods. In detail, a fast flow component is responsible for discharge peaks and frequently of turbidity events, having a mean velocity ranging from 30 to 70 m/h in the aquifer. Besides of this fast flow, an additional aliquot of the recharge is due to the same local aquifer, but slower flow clearly identifiable by hydrochemistry and isotopic data. Thanks to these findings, a renewed management of the spring has been suggested, considering the different degrees of aquifer vulnerability (turbidity occurrence) directly related to the discharge regimen.

Hydrology Drives Crustal Deformation and Modulates Seismicity in the Matese Massif (Italy)

Abstract We analyze the interplay between hydrology, deformation, and seismicity in the Matese massif, located in the Italian Southern Apennines. We find that this area is characterized by the concurrent action of two hydrologically driven processes: the first is the deformation detected by the Global Navigation Satellite Systems (GNSS) data in the shallowest part (above the elevation of the major springs) of the Earth crust, in phase with the hydrological forcing; the second is the triggering of seismicity at depth with a delay suggesting a downward diffusive process. We study the first process by applying a principal component analysis to the GNSS displacements time series, aiming to identify a common signal describing the largest data variance. We find that the maximum horizontal displacements associated with the first principal component (PC1) are larger than 1 cm in two GNSS sites, and the PC1 temporal evolution is well correlated and in phase with the flow of the largest spring of the region, which we consider as proxy of the water content of the massif. This suggests that the main source of horizontal deformation is the water content fluctuations in the shallow portion of the Matese aquifer, in particular within fractures located in correspondence of the main mapped faults. The deformation rates caused by this process are one order of magnitude larger than the tectonic ones. Finally, we infer the second process by observing the correlation between the background seismicity and the spring discharge with a time lag of 121 days. In our interpretation, downward diffusive processes, driven by aquifer water content variations, propagate pore-pressure waves that affect the fault’s strength favoring the occurrence of microearthquakes. This is supported by the values of hydraulic diffusivity (1.5 m2/s) and rock permeability (3.2–3.8×10−13 m2), which are compatible with what is observed in karstified limestones.

НИЖНЕ-ЩАПИНСКИЕ ТЕРМАЛЬНЫЕ ИСТОЧНИКИ (КАМЧАТКА), КАК ПРИМЕР МАГНИЕВЫХ УГЛЕКИСЛЫХ ВОД

The results of hydrochemical studies of Nizhne-Shapinsky (Kipely) CO2-rich thermal (39 °C) springs in 2021 are presented. The springs discharge within the Shchapina graben near active Kizimen volcano. Results include macro- and trace-component composition, isotopic composition of spring water, some gas components and dissolved strontium. The results are discussed using previously published and reported data, taking into account the geological and structural setting of the area, water-rock interactions, and simple thermochemical calculations to explain the rare, but typical for the chemical composition of some CO2-rich waters, predominance of magnesium over calcium. It is shown that waters of Nizhne-Shchapinsky springs are formed as a result of mixing of two components: deeper and more heated sodium chloride water and more superficial, less heated water of Mg-Ca-HCO3- composition, formed due to interaction with carbonate-bearing rocks and CO2 of magmatic origin.

ENSO sentinels in the Americas' humid tropics: We need combined hydrometric and isotopic monitoring for improved El Niño and La Niña impact prediction

AbstractThis Scientific Briefing presents results from a nearly 10‐year hydrometric and isotope monitoring network across north‐central Costa Rica, a region known as a headwater‐dependent system. This monitoring system has recorded different El Niño and La Niña events and the direct/indirect effects of several hurricane and tropical storm passages. Our results show that El Niño‐Southern Oscillation (ENSO) exerts a significant but predictable impact on rainfall amount anomalies, groundwater level and spring discharge, as evidenced by second‐order water isotope parameters (e.g., line conditioned‐excess or line‐conditioned (LC)‐excess). Sea surface temperature anomaly (El Niño Region 3) is correlated with a reduction in mean annual and cold front rainfall across the headwaters of north‐central Costa Rica. During El Niño conditions, rainfall is substantially reduced (up to 69.2%) during the critical cold fronts period, limiting groundwater recharge and promoting an early onset of minimum baseflow conditions (up to 5 months). In contrast, La Niña is associated with increased rainfall and groundwater recharge (up to 94.7% during active cold front periods). During La Niña, the long‐term mean spring discharge (39 Ls−1) is exceeded 63–80% of the time, whereas, during El Niño, the exceedance time ranges between 26% and 44%. The regional hydroclimatic variability is also imprinted on the hydrogen and oxygen isotopic compositions of meteoric waters. Drier conditions favoured lower LC‐excess in rainfall (−17.3‰) and spring water (−6.5‰), whereas wetter conditions resulted in greater values (rainfall = +17.5‰; spring water = +10.7‰). The lower and higher LC‐excess values in rainfall corresponded to the very strong 2014–2016 El Niño and 2018 La Niña, respectively. During the recent triple‐dip 2021–23 La Niña, LC‐excess exhibited a significant and consistently increasing trend. These findings highlight the importance of combining hydrometric, synoptic and isotopic monitoring as ENSO sentinels to advance our current understanding of ENSO impacts on hydrological systems across the humid Tropics. Such information is critical to constraining the 21st century projections of future water stress across this fragile region.

Towards understanding formation of polytungstates in tungsten-bearing hot springs and its environmental implications

Polytungstates have toxicity, mobility, and bioavailability quite different from those of monomeric tungstates, the most common tungsten (W) species. To date, relevant studies of polytungstates have been conducted under laboratory conditions, and there is a lack of studies on such W species in natural waters. In the present study, polytungstates at environmentally relevant concentrations were detected in a number of investigated hot springs. All the hot springs with high polytungstates concentrations (> 10 μg/L) have vent temperatures above 70 °C and pH values above 8, and polytungstates are immeasurable in all the acidic hot springs. Further inspection showed that the detected polytungstates were not isopolytungstates but likely heteropolytungstates forming via incorporation of silicate or borate enriched in the hot springs into tungstates at elevated temperatures. These detectable polytungstates, however, might also be polymeric thiotungstates in view of their close correlation with aqueous sulfide and thiotungstates and their specific retention time during the chromatographic separation. More precise identification of the polytungstates depends on improvement of the used analytical technique to achieve a further chromatographic separation of different kinds of polytungstates. Upon the discharge of the hot springs, the aqueous W was transported to the hot spring sediments due to its adsorption onto the Fe-bearing minerals there. Although the pH values of the hot springs were the critical parameter for the W adsorption, the substantial formation of polytungstates in some neutral to alkaline hot springs could inhibit the adsorption to some degree, thereby contributing somewhat to the enrichment of W in such springs. This is the first study reporting polytungstates and their environmental behavior in natural hot springs. It is possible that polytungstates are widely present in high-temperature, neutral to alkaline hot springs hydrochemically similar to those involved in this study.

Magnesium and groundwater flow relationship in karst aquifers: a tool for exploitation management of springs

Karst aquifers are characterized by different types of groundwater flow, related to different types of permeability due to the simultaneous presence of matrix, fractures and conduits. The presence of a well-developed karst conduit system leads to a rapid circulation of groundwater within the aquifer and a pulse-type response of the spring flow to the rainfall inputs, with a potential fast transport of contaminants from the hydrogeological basin surface to the discharge zones. Supported by hydro chemical analyses of spring water samples and single discharge measurements, it was possible to develop specific mass balance models, correlating ion content to spring flowrates. Specifically, Mg2+ content revealed a reliable application for spring baseflow separation in karst settings. Once the local model has been set, its conservative behaviour, in mostly limestone-dominant aquifers, allows using Mg2+ as a natural tracer of groundwater flow, distinguishing conduit flow (overflow) and diffuse flow (baseflow) occurrence in the spring outlet, without additional discharge measurements. In karst settings, the difficulty in continuously monitoring the spring discharge values makes this application interesting for exploitation management. This study shows the results obtained for two springs located in Central Italy, confirming that monitoring groundwater quality in karst environments is often the key for successfully characterizing springs and assessing the total yield when direct measurements are not available.

Mediterranean springs: Keystone ecosystems and biodiversity refugia threatened by global change

AbstractMediterranean spring ecosystems are unique habitats at the interface between surface water and groundwater. These ecosystems support a remarkable array of biodiversity and provide important ecological functions and ecosystem services. Spring ecosystems are influenced by abiotic, biotic, and anthropogenic factors such as the lithology of their draining aquifers, their climate, and the land use of their recharge area, all of which affect the water chemistry of the aquifer and the spring discharges. One of the most relevant characteristics of spring ecosystems is the temporal stability of environmental conditions, including physicochemical features of the spring water, across seasons and years. This stability allows a wide range of species to benefit from these ecosystems (particularly during dry periods), fostering an unusually high number of endemic species. However, global change poses important threats to these freshwater ecosystems. Changes in temperature, evapotranspiration, and precipitation patterns can alter the water balance and chemistry of spring water. Eutrophication due to agricultural practices and emergent pollutants, such as pharmaceuticals, personal care products, and pesticides, is also a growing concern for the preservation of spring biodiversity. Here, we provide a synthesis of the main characteristics and functioning of Mediterranean spring ecosystems. We then describe their ecological value and biodiversity patterns and highlight the main risks these ecosystems face. Moreover, we identify existing knowledge gaps to guide future research in order to fully uncover the hidden biodiversity within these habitats and understand the main drivers that govern them. Finally, we provide a brief summary of recommended actions that should be taken to effectively manage and preserve Mediterranean spring ecosystems for future generations. Even though studies on Mediterranean spring ecosystems are still scarce, our review shows there are sufficient data to conclude that their future viability as functional ecosystems is under severe threat.

Linear and nonlinear ensemble deep learning models for karst spring discharge forecasting

Forecasting karst spring discharge is crucial for groundwater resource management in karst aquifers. These aquifers, with their inherent heterogeneity and complexity influenced by a myriad of environmental factors and hydrological processes, often show nonlinear and nonstationary behaviors. This study introduces novel multivariate multi-step ensemble models, both linear and nonlinear, to forecast karst spring discharge based on predictions from long short-term memory (LSTM), gated recurrent units (GRU) and one-dimensional convolutional neural network (1D-CNN). Linear and nonlinear ensemble learners including simple average method (SAM), linear regression (LR) and support vector regression (SVR) with Bayesian optimization (BO) are used to aggregate obtained results from base models and produce the final ensemble results. These proposed models, named as Ensem-SAM, Ensem-LR and Ensem-SVR-BO, are implemented for daily spring discharge forecasting across various lead times and time steps at Barton Springs, Texas, USA. Notably, these ensemble models outperform the individual base models in prediction accuracy and consistency. The results demonstrate that the ensemble framework can effectively leverage the strengths of diverse deep learning models and complement their limitations, thereby gaining strong generalization capabilities and robust performance, especially with diversiform and nonlinear data. The most marked disparities between individual base models and ensemble models arise with a short time step or a long lead time. Among all, Ensem-SVR-BO exhibits the best generalization capability and delivers accurate and robust prediction results even when one or all base models stuck in local optima.

Analysis of the Recharge Area of the Perrot Spring (Aosta Valley) Using a Hydrochemical and Isotopic Approach

The Perrot Spring (1300 m a.s.l.), located to the right of the Chalamy valley in the Monte Avic Natural Park (Valle d’Aosta, Italy), is an important source of drinking water for the municipality of Champdepraz. This spring is located on a large slope characterised by the presence of a Quaternary cover of various origins (glacial, glaciolacustrine, and landslide) above the bedrock (essentially serpentinite referred to the Zermatt–Saas Zone, Penninic Domain). Water emerges at the contact between the landslide bodies and impermeable or semi-permeable glaciolacustrine deposits. The aim of this study is to define the processes and recharge zones of this spring. The analysis of the data revealed the presence of two contributions to the Perrot Spring input: a spring thaw contribution defined by a small increase in flow and an autumn contribution from rainwater infiltration. The low average temperature and low variation of the annual temperature (4.8–6.5 °C) suggest a sufficiently deep flow circuit. Chemical analyses showed a groundwater chemistry consistent with the regional geology: the hydrochemical facies is calcium–magnesium bicarbonate and isotopic analyses (δ2H and δ18O) of rainfall and spring water suggested a recharge altitude of about 2100 m a.s.l. In conclusion, this study makes it possible to recognize the water inputs to the spring discharge and to delineate its recharge area, which can be proposed to implement strategies to protect the resource.

Effects of Irrigation Projects on the Classification of Yellow River Terrace Landslides and their Failure Modes: A Case Study of Heitai Terrace

The study of the classification and failure modes of Yellow River terrace landslides under the influence of irrigation projects is of key importance to alleviate the paradox between the rapid evolution of terrace landscapes caused by landslides and the survival of local residents. However, such studies remain controversial, despite it being widely recognized that a rise in groundwater level caused by irrigation is a key factor associated with landslide failure modes. In this paper, we take the Heitai terrace as a case study. Using aerial images and field investigations, we classify landslides in the Heitai loess layer into type A landslides (not related to groundwater) and type B1 and B2 landslides (related to groundwater). We analyze the failure modes and disaster-causing characteristics of each type of landslide, and our results indicate that the attenuation in soil strength is a key factor common to both type A and type B landslides, based on which type A landslides with small volume and short sliding distance are able to block the previous spring discharge, causing a rise in localized groundwater, which further contributes to type B landslides; the location of previous type B1 landslides with a large volume and long sliding distance and type A landslides may be more susceptible to type B2 landslides with a small volume and short sliding distance, where there are low confining pressures during the lower soil shear process. Therefore, we believe that the inevitable interaction effects between the failure modes of landslides during landslide evolution, which govern the geomorphological evolution of the Heitai terrace, are unavoidable. Combining these data with numerical analyses, we further demonstrate that a rise in groundwater level and discontinuous attenuation of soil strength caused by changes in soil properties during irrigation together control terrace landslides and their failure modes. From the results of interferometric synthetic aperture radar time-series monitoring of Yellow River terrace activity with and without irrigation projects, and electrical resistivity tomography groundwater detection, we conclude that in the future, Heitai terrace will continue to experience a high intensity of landslide activity, and conditions for the most catastrophic type of landslide (type B1) will remain, including the high localized groundwater caused by previous landslides, and the discontinuous attenuation of soil strength caused by the deterioration in soil properties. In this context, we believe that slope-cutting engineering will be one of the most economical means to achieve future landslide-type transformation on the Heitai terrace; this will mitigate the process of geomorphological evolution and improve the human living environment.

Quantifying the historic and future response of karst spring discharge to climate variability and change at a snow-influenced temperate catchment in central Europe

Climate change is projected to threaten groundwater resources in many regions, but projections are highly uncertain. Quantifying the historic impact potentially allows for understanding of hydrologic changes and increases confidence in the predictions. In this study, the responses of karst discharge to historic and future climatic changes are quantified at Blautopf Spring in southern Germany, which is one of the largest karst springs in central Europe and belongs to a regional aquifer system relevant to the freshwater supply of millions of people. Statistical approaches are first adopted to quantify the hydrodynamic characteristics of the karst system and to analyse the historic time series (1952–2021) of climate variables and discharge. A reservoir model is then calibrated and evaluated with the observed discharge and used to simulate changes with three future climate-change scenarios. Results show that changes in the annual mean and annual low discharge were not significant, but the annual peak discharge shifted to a low state (<13.6 m3 s−1) from 1988 onwards due to decreasing precipitation, increasing air temperature, and less intense peak snowmelt. The peak discharge may decrease by 50% in this century according to the projections of all climate-change scenarios. Despite there being no significant historic changes, the base flow is projected to decrease by 35–55% by 2100 due to increasing evapotranspiration. These findings show the prolonged impact of climate change and variability on the floods and droughts at the springs in central Europe, and may imply water scarcity risks at similar climatic and geologic settings worldwide.

Delay-aware karst spring discharge prediction

As an essential component of the hydrological cycle at both regional and global scales, karst aquifers store large amounts of water, serving as the main source of fresh water supply in many areas. Karst springs provide a natural drainage pathway for karst groundwater systems to participate in the hydrological cycle. Accurate prediction of karst spring discharge is of great value to the long-term planning and management of karst water resources. The mainstream methods proposed have achieved high prediction accuracy by introducing deep learning technology. However, these methods ignore the delay time of spring water discharge to water supply or drainage, so there is still room for improvement in prediction accuracy. To address this issue, this paper proposes a delay-aware spring discharge prediction (DSDP) framework that introduces the delay time to predict spring discharge. Specifically, the framework consists of three modules, including a time calibration module, a variation estimation module and a discharge prediction module. The time calibration module identifies the delay time of spring discharge to each element to calibrate the observations. And the variation estimation module utilizes a neural network to infer spring discharge variations. Finally, the discharge prediction module predicts spring discharge with derived discharge variations. This study validates the performance of the proposed method by using spring discharge data from Jinci Spring, a typical karst spring in China. The experimental results show that the proposed method outperforms baselines, with significant decreases in both MAE and RSM. The study will provide evidence for the precise management of Jinci Spring, and provide a methodological reference for the discharge prediction of other springs.

Prediction of karst spring discharge using LSTM with Bayesian optimisation hyperparameter tuning: a laboratory physical model approach

Prediction of karst spring discharge using LSTM with Bayesian optimisation hyperparameter tuning: a laboratory physical model approach

Impact of submarine karst sulfur springs on benthic foraminiferal assemblage in sediment of northern Adriatic Sea

PurposeThis work highlights the sedimentary characteristics and the role of submarine sulfur-rich karstic springs in the distribution of benthic foraminifera in the northern part of the Adriatic Sea (Bay of Koper). Little is known about how local conditions such as temperature and sulfur bursts may influence sediment properties, benthic habitat variability, and composition of foraminiferal assemblages. Here we compare the distribution of total and living benthic assemblages in surface sediment samples collected from a funnel-shaped depression created by submarine sulfur springs.Materials and methodsSampling was performed at water depths between 24.6 and 32.2 m in fine-grained sandy silt to silty sand (partially washed). Sedimentological, mineralogical, and geochemical analyses of the sediment were carried out and the distribution of benthic foraminifera living around the springs was studied.Results and discussionIn general, sediment characteristics (i.e., mineralogical, geochemical, and organic content) around the sulfur springs do not show prominent deviations from the marine surface sediment of the area; however, some differences exist among depressions of different depths. Deeper depressions in the lower parts probably extend to older continental sediments of Late Pleistocene age with alluvial features, while shallower depressions were formed entirely in Holocene marine sediments typical of a wider area. Only one of the five samples (M05) contained living foraminifera in sufficient abundance for biocenosis research. The benthic foraminiferal assemblages of moderate diversity are composed of opportunistic species. Elphidium translucens, Ammonia ex gr. tepida, Haynesina depressula, and Porosononion granosum dominate, while A. neobeccarii, Reussella spinulosa, and Textularia bocki are subordinate.ConclusionsThe distribution and diversity of foraminifera in the sediment near sulfur springs can be explained by several factors and their interactions. The intensity of the spring discharge affects the mixing/oxygenation of the sediment, the shape of spring depressions, and the granulometry of the coarser sediment around the springs. Sediment characteristics indicate different types of sediment origin. This is related to and can be explained by the depth of spring depressions.

Flow recession behavior of preferential subsurface flow patterns with minimum energy dissipation

Abstract. Understanding the properties of preferential flow patterns is a major challenge in subsurface hydrology. Most of the theoretical approaches in this field stem from research on karst aquifers, where two or three distinct flow components with different timescales are typically considered. This study is based on a different concept: a continuous spatial variation in transmissivity and storativity over several orders of magnitude is assumed. The distribution and spatial pattern of these properties are derived from the concept of minimum energy dissipation. While the numerical simulation of such systems is challenging, it is found that a restriction to a dendritic flow pattern, similar to rivers at the surface, works well. It is also shown that spectral theory is useful for investigating the fundamental properties of such aquifers. As a main result, the long-term recession of the spring draining the aquifer during periods of drought becomes slower for large catchments. However, the dependence of the respective recession coefficient on catchment size is much weaker than for homogeneous aquifers. Concerning the short-term behavior after an instantaneous recharge event, strong deviations from the exponential recession of a linear reservoir are observed. In particular, it takes a considerable time span until the spring discharge reaches its peak. The order of magnitude of this rise time is one-seventh of the characteristic time of the aquifer. Despite the strong deviations from the linear reservoir at short time spans, the exponential component typically contributes more than 80 % to the total discharge. This fraction is much higher than expected for karst aquifers and even exceeds the fraction predicted for homogeneous aquifers.

Simulation of spring discharge using graph neural networks at Niangziguan Springs, China

Over past decades, extensive groundwater development in karst aquifers has led to significant declines in groundwater levels and spring discharges. While the recent implementation of sustainable karst aquifer development policies appears to have lessened the declining trend of spring discharge, many aspects of karst aquifers remain to be explored. This paper applied two spectral-based graph neural networks (GNN) models of ChebNet and graph convolutional networks (GCN) for karst hydrological processes by considering the correlations among precipitation, spring discharge, and human impacts. This approach captures the spatial dependence of precipitation infiltration, groundwater propagation in heterogeneous karst aquifer and spring discharge from bared karst aquifer. This study proposed three graph structures (complete graph, information flow graph, and association graph of groundwater flow field) to depict relations between precipitation and spring discharge in the GNN models in Niangziguan Springs, China. The results show that the association graph of groundwater flow field is the optimal graph structure of GNN models. Based on the optimal structure, we investigated precipitation-driven-discharge models using ChebNet and GCN to predict the discharge. The results show that the high-order ChebNet is more adaptable to simulate the karst hydrological processes with nonlinear and nonstationary behaviors than that GCN. Moreover, the study confirms that the groundwater sustainable development policy has achieved a prominent contribution 2 m3/s in water conservancy in Niangziguan Springs catchment.