Reactive Transport Model of Gypsum Karstification in Physically and Chemically Heterogeneous Fractured Media

Abstract

Gypsum dissolution leads to the development of karstic features within much shorter timescales than in other sedimentary rocks, potentially leading to rapid deterioration of groundwater quality and increasing the risk of catastrophes caused by subsidence. Here, we present a 2-D reactive transport model to evaluate gypsum karstification in physically and chemically heterogeneous systems. The model considers a low-permeability rock matrix composed mainly of gypsum and a discontinuity (fracture), which acts as a preferential water pathway. Several scenarios are analyzed and simulated to investigate the relevance for gypsum karstification of: (1) the dynamic update of flow and transport parameters due to porosity changes; (2) the spatial distribution of minerals in the rock matrix; (3) the time evolution of water inflows through the boundaries of the model; (4) the functions relating permeability, k, to porosity, ϕ. The average porosity of the matrix after 1000 years of simulation increases from 0.045 to 0.29 when flow, transport, and chemical parameters and the water inflows through the boundary are dynamically updated according to the porosity changes. On the contrary, the porosity of the matrix hardly changes when the porosity feedback effect is not considered, while its average increases to 0.13 if the water inflow occurs through the discontinuity. Moreover, the dissolution of small amounts of highly soluble sulfate minerals plays a major role in the development of additional fractures. The increase in hydraulic conductivity is largest for the power law with an exponent of n = 5, as well as the Kozeny-Carman and the modified Fair-atch k-ϕ relationships. The gypsum dissolution front propagates into the matrix faster when the power law with n = 2 and 3 and the Verma–Pruess k-ϕ relationships are used.

Comparison of saturated hydraulic conductivity estimated by empirical, hydraulic and numerical modeling methods at different scales in a coastal sand aquifer in Northern Ireland

“…However, some media approach this ideal state; this includes the coastal sand deposits at Magilligan. This aquifer was exhaustively characterized through geophysical, geotechnical and hydrogeological tests, confirming the great uniformity and high levels of homogeneity (Águila et al 2022b;McDonnell et al 2023). Consequently, the high homogeneity of this site makes it an excellent study area to investigate and increase the current state of knowledge on the scale effect of K when estimating this parameter using different measurement methods.…”

A physically based distributed karst hydrological model (QMG model-V1.0) for flood simulations

“…With the development of physical exploration technology and progress in mathematics, computing and other interdisciplinary disciplines, the level of modelling has gradually improved (Hartmann and Baker, 2017;Hartmann, 2018;Petrie et al, 2021), and distributed hydrological models have subsequently become widely applied to karst areas. The main difference between lumped and distributed hydrological models is that the latter divide the entire basin into many subbasins to calculate the runoff generation and confluence (Chang et al, 2021;Guila et al, 2022), thereby better describing the physical properties of the hydrological processes inside a karst water-bearing system (Jourde et al, 2007;Hartmann, 2018;Epting et al, 2018).…”

Comparison of saturated hydraulic conductivity estimated by empirical, hydraulic and numerical modelling methods at different scales in a coastal sand aquifer in Northern Ireland

“…Consequently, determining realistic values of K can prove essential in fields including civil engineering (Liu and Rowe, 2016;Tan et al, 2017;Chen et al, 2021), mining (McCoy et al, 2006;Meiers et al, 2011), chemical engineering (Francisca and Glatstein, 2010;Wu et al, 2020), agronomy (Peña-Haro et al, 2011;Picciafuoco et al, 2019) and environmental geology (Akcanca and Aytekin, 2014;Gao et al, 2018). Hydraulic conductivity typically varies within natural geological units (Mitchell and Soga, 2005), with this spatial heterogeneity giving rise to considerable scale-dependent uncertainties in determining an appropriate value (Mallants et al, 1997;Boschan and Noetinger, 2012;Piña et al, 2019;Águila et al, 2022a).…”