Elsevier

Journal of Hydrology

Volume 599, August 2021, 126375
Journal of Hydrology

Research papers
Assessing the impact of tunnelling on karst groundwater balance by using lumped parameter models

https://doi.org/10.1016/j.jhydrol.2021.126375Get rights and content

Highlights

  • Tunnelling impact on karst aquifer discharge is assessed by lumped parameter models.

  • Dominant hydrogeological factors are identified by parameter sensitivity analysis.

  • Tunnel excavation is predicted to capture 19% of discharge from the aquifer system.

Abstract

Tunnelling activities may significantly alter the groundwater balance in a karst aquifer, but assessment of this effect remains challenging due to the complex flow geometries and strong hydrogeological heterogeneity. In this study, based on extensive site characterization of an independent hydrogeological unit (HU) in which a deep-buried tunnel is under construction, we present a modified lumped model to evaluate the impact of tunnel construction on the discharge from the aquifer system. The lumped model consists of four interacting flow compartments and the associated water balance equations. A series of simplified models is then developed, by merging or removing some of the flow components that represent different mechanisms and relationships between recharge and discharge. The dominant hydrological processes in the HU are identified by comparing the performance of the models of different structures and by screening analysis of parameter sensitivities. It is found that the tunnel excavation captured a mean portion of 19% discharge from the aquifer system into the tunnel, which is comparable to the result predicted by 3D numerical simulations. This study evidences that as a first and simple approximation, the lumped models provide a useful tool for characterizing the dominant factors that govern the groundwater response and for evaluating the groundwater budget changes induced by anthropogenic activities such as tunnelling and underground mining in karst regions.

Introduction

Karstic aquifers are characteristic of complex flow geometry and strong hydrogeological heterogeneity, caused mainly by double or triple porosity structures consisting of pores, fissures, fractures and conduits of various scales (White, 2002, Ford and Williams, 2007, Hartmann et al., 2014). Consequently, karst groundwater typically consists of base flow through fissure network and rapid flow through enlarged fractures and conduits (Jukić and Denić-Jukić, 2009, Hosseini et al., 2017). These characteristics give rise to significant challenges in groundwater flow modelling in karstic formations. Further complexity arises when anthropogenic activities, e.g., tunnel construction, are involved, where the groundwater flow paths and discharges may be changed drastically within a short period (Hao et al., 2016, Lv et al., 2020), even resulting in geological and environmental hazards such as mud and water inrush, spring depletion, and ground subsidence (Raposo et al., 2010, Pu et al., 2011, Mahler and Bourgeai, 2013, Grimmeisen et al., 2016, Han et al., 2017, Jiang et al., 2018, Liu et al., 2019, Chen et al., 2020, Zhou et al., 2021).

A variety of hydrological models (e.g., black-box, distributed, and lumped) have been proposed to confront the challenges and approximate the groundwater budget changes in karst formations (Beven, 1989, Fleury et al., 2007, Le Moine et al., 2008, Makropoulos et al., 2008, Jukić and Denić-Jukić, 2009, Hartmann et al., 2012). The black-box models treat an aquifer as a black-box system and focus on establishing an overall relationship between input (infiltration) and output (discharge), completely disregarding the structure and the physical mechanisms of the aquifer (Dreiss, 1983, Mangin, 1984). The distributed models solve flow equations with temporal and spatial discretization over the whole aquifer, hence requiring detailed information on the aquifer structure and hydrological properties for producing an acceptable simulation (Zhang et al., 2011, Zheng et al., 2020, Saller et al., 2013). The lumped models provide a compromised tool for karst groundwater simulation in which linear or nonlinear reservoirs/compartments are used to represent different physical elements of the aquifer structure and groundwater circulation (Fleury et al., 2007, Fleury et al., 2009, Jukić and Denić-Jukić, 2009). Each reservoir relies on a production function and a transfer function to simulate the process from input to output in each hydrogeological unit, and hence the lumped models are suitable for aquifer simulation with limited observation data and insufficient field characterization.

Spring discharge observations and trace element data are commonly used to calibrate the lumped models (Wagener et al., 2003, Hartmann et al., 2013, Ye et al., 2014, Mudarra et al., 2019). Particular concerns have been focused on the temporal variability in groundwater recharges and the effect of karst aquifers drainage induced by climate change or human activities (Jukić and Denić-Jukić, 2009, Chang et al., 2017, Hosseini et al., 2017). The construction of deep-buried tunnels in karst regions can profoundly change the water budget in the karst system. Hence there is a need to revisit the structure and parameters of the lumped model, for incorporating this strong impact of human activity.

This study is motivated by tunnel excavation-induced karst groundwater budget changes in the Jiayan Water Diversion Project (JWDP), located in Guizhou Province, Southwest China. In this work, a modified lumped model is presented to quantitatively estimate the impact of tunnel construction on the karst aquifers, by integrating a channel drainage compartment that represents the drainage from the karst aquifer to the tunnel during construction. The model structure is systematically investigated through comparison of simulations with varying levels of simplification by merging or removing some of the flow components. The dominant hydrogeological processes that control the aquifer discharge are identified by performance assessment of the models and sensitivity analysis of the input parameters. The impact of tunnel construction on the aquifer discharge is finally quantified with the calibrated models, which contributes to decision-making for appropriate engineering treatments during tunnel construction.

Section snippets

Study site and data

The study area is located in Guizhou Province, Southwest China (Fig. S1a in Supplementary material), where karst landscapes cover about 62% of the land area (Chen et al., 2020); it is bounded by the Baifu River in the east, the Liuchong River in the south, and a groundwater divide in the north. The terrains range from 1300 to 2300 m a.s.l., sloping from northwest to southeast. The study area lies in the subtropical monsoon climate zone with a marked seasonal pattern in the annual variations of

Model development

As depicted conceptually in Fig. 1, in the Shuchang HU, the rainfall infiltrates into the karst system by both diffuse flow through soil cover and quick flow through sinkholes and vertical shafts, where the recharge area often varies as the groundwater level fluctuates (White, 2002, Ford and Williams, 2007). The quick flow directly feeds into the underground river, while the diffuse flow recharges the epikarst zone characteristic of relatively lower degree of karstification at the site (

Model structures and performance

Both four- and three-compartment models (MLM1–8) are used to simulate the groundwater response in the Shuchang hydrogeological unit. The SCEM-UA method with the NSE criterion is used to calibrate the parameters of each model, after a warm-up period of one month to eliminate the influence of initial conditions on the result. As an example, Fig. 4 presents the box-whisker plot of the calibrated parameters for the model MLM1 having the most complex structure. Fig. 5, Fig. 6 plot the spring

Discussion

In Section 4, we have identified the dominant hydrogeological processes in the Shuchang HU, by comparing the performance of a series of models with different degree of simplifications and by screening analysis of parameter sensitivities. We find that the four-compartment models consistently perform better than the three-compartment counterparts. But as a result of the thin thickness of the soil cover, the calibrated soil storage capacity Vsmax is around 2 mm (MLM1), and is less sensitive to the

Conclusions

This study presents a series of modified lumped models with different structures or flow components to characterize the mechanisms and relationships between recharge and discharge in karst aquifers. A particular concern is placed upon the impact of tunneling activities on the aquifer discharge, by integrating a channel drainage compartment in the models. The lumped models are applied to an independent hydrogeological unit (HU) where a deep-buried tunnel is being constructed. By comparing the

CRediT authorship contribution statement

Bing-Qi Zhou: Investigation, Methodology, Formal analysis, Validation, Writing - original draft. Zhibing Yang: Investigation, Methodology, Formal analysis, Validation, Writing - review & editing. Ran Hu: Formal analysis, Validation. Xian-Jin Zhao: Investigation, Methodology. Yi-Feng Chen: Conceptualization, Supervision, Formal analysis, Resources, Funding acquisition, Writing - original draft, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Grant No. 51925906), the National Key R&D Program of China (Grant No. 2018YFC0407001), and the Guizhou Province Key Science and Technology Project (Grant No. [2017]3005-1).

References (49)

  • A. Hartmann et al.

    A multi-model approach for improved simulations of future water availability at a large Eastern Mediterranean karst spring

    J. Hydrol.

    (2012)
  • A. Hartmann et al.

    Progress in the hydrologic simulation of time variant recharge areas of karst systems – Exemplified at a karst spring in Southern Spain

    Adv. Water Resour.

    (2013)
  • D.M. Han et al.

    Alterations to groundwater recharge due to anthropogenic landscape change

    J. Hydrol.

    (2017)
  • Y. Hao et al.

    How does the anthropogenic activity affect the spring discharge?

    J. Hydrol.

    (2016)
  • S.M. Hosseini et al.

    Spring hydrograph simulation of karstic aquifers: impacts of variable recharge area, intermediate storage and memory effects

    J. Hydrol.

    (2017)
  • D. Jukić et al.

    Nonlinear kernel functions for karst aquifers

    J. Hydrol.

    (2006)
  • D. Jukić et al.

    Groundwater balance estimation in karst by using a conceptual rainfall–runoff model

    J. Hydrol.

    (2009)
  • J.C. Liu et al.

    Response of plants water uptake patterns to tunnels excavation based on stable isotopes in a karst trough valley

    J. Hydrol.

    (2019)
  • A. Mangin

    Pour une meilleure connaissance des systèmes hydrologiques à partir des analyses corrélatoire et spectrale

    J. Hydrol.

    (1984)
  • B.J. Mahler et al.

    Dissolved oxygen fluctuations in karst spring flow and implications for endemic species: Barton Springs, Edwards aquifer, Texas

    USA. J. Hydrol.

    (2013)
  • C. Makropoulos et al.

    A multi-model approach to the simulation of large scale karst flows

    J. Hydrol.

    (2008)
  • J.E. Nash et al.

    River flow forecasting through conceptual models. Part I – A discussion of principles

    J. Hydrol.

    (1970)
  • F. Pianosi et al.

    A simple and efficient method for global sensitivity analysis based on cumulative distribution functions

    Environ. Model. Softw.

    (2015)
  • J.R. Raposo et al.

    Quantitative evaluation of hydrogeological impact produced by tunnel construction using water balance models

    Eng. Geol.

    (2010)
  • Cited by (0)

    View full text