Water freezing characteristics in granular soils: Insights from pore-scale simulations

https://doi.org/10.1016/j.advwatres.2020.103681Get rights and content

Highlights

  • A new model for pore-scale simulations of water freezing in pore space is proposed.

  • Soil freezing curves (SFCs) are characterized with the model using 3D pore images.

  • Simulated soil moisture curves (SMCs) agree with the literature data for sand.

  • Simulated SFCs curve and SMCs were successfully converted with equilibrium equation.

  • Pore scale hysteresis in water-ice contact angle can cause hysteresis in SFC.

Abstract

It has been observed in both laboratory and field studies that unfrozen soil moisture content is hysteretic between freezing and thawing conditions. While a few explanations have been provided for the causes of this hysteresis, there is no general agreement in the literature on underlying causal mechanisms. In this study, we use pore-scale simulations to investigate the role that contact angle hysteresis during freezing and thawing conditions may play in the observed differences in soil moisture content during freezing and thawing conditions. We also use these pore-scale simulations to better assess the similarities between freezing and water drainage processes in pore space as such realistic assessments are very limited in the literature. Specifically, we use 3D images of natural sand packs and synthetic silicates to generate pore-scale model domains and to simulate water drainage and water freezing in the soil pore space. The simulated water drainage curve and relative permeability-saturation relations for the sand packs agrees well with measured data available in the literature for sandy soils. The simulated soil freezing curves were successfully converted to simulated water drainage curve confirming conclusions of previous studies on coarse granular soils. Water freezing simulations using different contact angles for freezing (water receding) and thawing (water advancing) showed that the hysteresis in soil freezing curves could be driven by a hysteresis in the contact angle of water-ice interface at the pore wall (i.e. grain surface). In addition, simulations of water drainage in partially frozen soils revealed that soil freezing enhances liquid water retention in pore space due to reduction of mean active pore radius and increased frequency of pore corners which agrees with the observations of published studies in laboratory and field conditions.

Introduction

Shallow soil layer freezing and thawing play important roles in the annual hydrologic cycle of systems in cold regions because the degree to which soil freezes determines the infiltration rate of precipitated water or snowmelt, and thus affects the partitioning of water between surface runoff and infiltration (Hayashi, 2013). Understanding the fundamental processes governing the freeze-thaw dynamics in near surface soils and its impact on soils' hydraulic and transport properties are important for both hydrological and hydrogeological assessments including assessment of engineering designs, forecasting the migration of nutrients/contaminant leaching through frozen soil layers, and appropriately parameterizing soil freeze-thaw relationships for fully-integrated surface water and groundwater models such as HydroGeoSphere (Aquanty, 2019). Additionally, understanding the details of these processes at the pore-scale will facilitate the development of upscaling methodologies such that soil freezing processes can be effectively represented in local to large scale numerical models (e.g., Clausnitzer and Mirnyy 2016; Grenier et al. 2018; Berg and Sudicky, 2019; Langford et al., 2019). While several fundamental studies have been published on soil freezing processes and the hydraulic properties of frozen soils (e.g. Williams, 1964; Koopmans and Miller, 1966; Black and Tice, 1989; Spaans and Baker, 1996; Andersland et al., 1996; Stähli et al., 1999; Watanabe and Flury, 2008; Kurylyk and Watanabe, 2013), there still remain several fundamental questions that haven't been fully answered. For example, it not clearly known how partial freezing of soil could influence retention of unfrozen water; or, Is permeability saturation relation of unfrozen water in partially frozen soil different from that under partially drained condition? What role ice nucleation could play in soil freezing, and how important it could be in the natural shallow groundwater systems? How different is the transport of solutes (e.g. nutrient or dissolved contaminants) through partially frozen soil from their transport in partially saturated soil?

While analogies between the soil moisture curve (SMC; the relation between soil water content and water pressure in unsaturated conditions, sometimes also referred to as water retention curve) and the soil freezing curve (SFC; the relation between unfrozen water content and soil temperature in partially frozen soils) have been proposed (e.g. Williams, 1964; Koopmans and Miller, 1966; Black and Tice, 1989; Kurylyk and Watanabe,2013), the role of pore structure and connectivity in their similarity remains unclear. For a given soil type it is well known that shape of the SMC is determined by its pore-scale properties such as pore size distribution, pore connectivity, and aspect ratio (as the ratio of pore body size to that of connected pore throat); however, it is not clear if the SFC has the same sensitivity and functionality to these pore-scale properties. In addition, given that the distribution of water in the pore space determines the relative permeability-saturation (kr-S) relationship for the water phase, the similarities and differences of the kr-S relation of unfrozen water in partially frozen soils to that of water in partially saturated soils have remained uninvestigated.

The primary mechanisms controlling hysteresis observed in SFCs between freezing and thawing conditions are uncertain. Hysteresis in the SFC is observed as a difference in the unfrozen moisture content of the soil at a given temperature and depends on whether the soil is in freezing (decreasing temperature) or thawing (increasing temperature) conditions. This hysteretic behavior has been documented in several laboratory and field studies (e.g. Black and Tice, 1989; Spaans and Baker, 1996; Quinton and Blatzer, 2013), however, very few studies have attempted to explain the mechanisms for SFC hysteresis in granular soils where capillarity is dominant. In a study on gas hydrate growth in pore space, Anderson et al. (2009) explained the hysteresis in hydrate growth and desiccation curves with the ink-bottle effect. Tian et al. (2014) referred to the arguments of Anderson et al. (2009) and extended it to explain hysteresis in SFCs in soil. It should be noted that while there are similarities between gas hydrate growth and ice growth in pore space (i.e., both follow the Gibbs-Thomson equation for solid phase thermodynamic stability) there are fundamental differences between them. Gas hydrate is not exclusively formed out of water, and requires CH4 and/or CO2 for its growth. This means gas hydrate cannot start growing from the middle of a porous medium, and its growth must be piston-like, starting from the edges of the system and spreading into its center. In contrast, ice growth can occur anywhere in liquid water of the pore space (as ice can be formed out of water and without a need for a secondary molecule) during freezing. This means that thermodynamics determines the ice-water interface, and not the fluid transport (Wettlaufer and Worster, 2006). As a result, ice growth or melting in a given pore is not necessarily controlled by the connectivity of the pore to a continuous ice phase and thus is not subject to the effects of pore connectivity. Consequently, the ink-bottle effect on freezing and melting is negligible. In addition, Tien et al. (2014) argued that the freezing point depression during freezing is twice that in thawing (in identical unfrozen water contents) and used it as an evidence to support their explanations for SFC hysteresis; however, the results of Black and Tice (1989) and Spaans and Barker (1996) doesn't support this, which necessitates considering the possibility of other explanations. In another study, Spaans and Baker (1996) note that the hysteresis in SFC is similar to the hysteresis in SMC; however, comparisons of causal mechanisms in each case were not discussed.

Another potential mechanism that may play a role in SFC hysteresis is hysteresis of the water-ice contact angle. Hysteresis of the water contact angle for advancing and receding conditions has been vastly documented for multiphase systems (e.g., Morrow, 1975; Bikkina 2011; Andrew et al. 2014; Khishvand et al. 2016; Gharedaghloo and Price 2017), and it is a well-established point that a hysteresis in contact angle causes a hysteresis in the capillary pressure and soil moisture curves (e.g. Zhou, 2013; Ruspini et al. 2017; Sun et al., 2020). Such hysteresis and variability in the contact angle at the front of the solidifying phase (at its interface with the unsolidified liquid phase) has been observed and discussed before (Grosse et al. 1997; Wei et al. 2004). However, it has not been directly used to describe the hysteresis in soil freezing and thawing curves. It is worthwhile to assess whether the hysteresis of SFC can be explained by hysteresis in water-ice contact angles. This will provide insight for further investigations using high-resolution laboratory imaging techniques.

During soil freezing, when the frost front establishes and grows in the vadose zone, liquid water potential drops in the frozen zone causing a pressure gradient that drives liquid water to move from unfrozen depths to the freezing front (Hayashi, 2013); this is associated with an increase in total water content in the frozen portion of the vadose zone (Quinton and Blatzer, 2013), potentially providing water for additional freezing and leading to increased ice saturation. Ice saturation antecedent to snowmelt events significantly influences infiltration of meltwater (Iwata et al., 2010) and partitioning of meltwater between surface runoff and subsurface flow (Kane and Stein, 1983). An improved understanding of the upward migration of water during freezing could inform forecasts of pre-melt ice saturation, and support estimations of surface/subsurface water partitioning ratios.

Pore-scale numerical studies in porous media have been previously conducted to investigate other related phenomena, e.g. the relationship of pore size and pore connectivity to processes and properties including two-phase and three-phase flow (e.g. Ioannidis et al., 1993; Blunt, 2001; Valvatne and Blunt, 2004; Vogel et al., 2005; Culligan et al., 2006; Raeini et al., 2012; Blunt et al., 2013; Berg et al. 2016), solute transport (e.g. Bijeljic et al., 2004; 2013; Mehmani et al., 2012; Gharedaghloo et al., 2018), biofilm growth (von der Schulenburg et al., 2009; Peszynska et al., 2016), and phase change (Duval et al., 2004). Despite the advances in pore-scale simulation techniques, the micro-scale studies of soil freezing have been limited to the bundle of capillaries model (e.g. Watanabe and Flury, 2008; Lebeau and Konrad, 2012; Ming et al., 2020). Due to the complex nature of soil freezing/thawing, the simplifying assumptions used in the bundle of capillaries modelling approach are not necessarily valid for water freezing process in porous media. For example, these models lack micro-scale pore morphologies including pore bodies and pore throats, and connection between individual pores; capillary water and ice cannot be present simultaneously inside an individual pore. Additionally, the bundle of capillaries approach is over parameterized (e.g., tortuosity, hydraulic conductivity, and pore size distribution parameters) in comparison to realistic 3D pore space models which explicitly represent these features in their pore space representations and do not require such fitting parameters. Furthermore, bundle of tubes models cannot evaluate the possible causes of hysteresis in soil freezing.

Given the significant challenges associated with experimentally isolating the individual mechanisms that govern soil freezing (both in the field and the laboratory), we propose that pore-scale simulations of water freezing and water drainage with 3D pore space models helps addressing some of the abovementioned points. For example, with the 3D pore-scale simulations, the similarities between SFC and SMC, and between kr-S relations of soils undergoing freezing or drainage can be evaluated. Considering the uncertainty in the causes of SFC hysteresis and the difficulty in conducting pore-scale experiments, pore-scale numerical simulations may serve as an informative ‘laboratory’ for hypothesis testing and can be used to explore the effect of contact angle hysteresis on SFC hysteresis. Furthermore, it provides the ability to evaluate variations in water retention characteristics after a soil partially freezes, clarifying whether ice growth in pore space enhances or reduces the retention of liquid water.

With the use of pore-scale simulations the objectives of this study are: 1) to assess the analogies between SMC and SFC and kr-S relations of water in drainage and freezing conditions; 2) to study the influence of water-ice contact angle hysteresis on hysteresis of SFC; and, 3) to explore the influences of ice growth in the pore space on unfrozen water retention in the partially frozen soils. To our knowledge, this is the first study that proposes an algorithm and a model for simulating ice nucleation and ice growth in realistic 3D pore spaces. Here, for the first time, independent water drainage and water freezing simulation results are validated against an empirically validated phase equilibrium equation. The methods and workflow presented in this manuscript can also be extended to explore other phenomena including nutrients leaching, solute transport, and spilled non-aqueous phase liquids (NAPLs) percolation in frozen porous media.

Section snippets

Theory

The curvature at the interface of unfrozen water and frozen water (denoted as ice hereafter) causes a decrease in the water's freezing temperature. This is known as the Gibbs-Thomson effect (Wettlaufer and Worster, 2006) and in porous media it causes the water in micropores and pore corners (that necessitates a smaller water-ice interface curvature) to freeze in cooler temperatures in comparison to the water in the macropores. Eq. (1) describes this process (Liu et al., 2003) relating radius of

Water drainage

The modelling approach was verified by simulating water drainage for the 3 sand packs and comparing the simulated capillary pressure-saturation curves to previously published relations for similar sand packs in the literature. Fig. 1 illustrates the simulated (black and grey curves) capillary pressure-water saturation (Pc-S) curves (which is equivalent of SMC) and compares them to data measured for sand packs and sands available in literature (Brooks and Corey, 1964; Likos and Jaafer, 2013;

Conclusion

In this study we develop a pore-scale simulator capable of reproducing soil freeze/thaw and water drainage processes. This simulator was verified by simulating SFC and SMC for sand packs and synthetic silicate soil by illustrating that the adjusted phase equilibrium equation can calculate SMC out of SFC (and vice versa); this agrees with the results of previous studies. In the simulations, water drainage is controlled by the throat radii and requires connection of the air phase to one side of

CRediT authorship contribution statement

Behrad Gharedaghloo: Conceptualization, Methodology, Software, Validation, Visualization, Writing - original draft, Writing - review & editing, Resources. Steven J. Berg: Writing - review & editing, Resources, Supervision, Writing - review & editing, Resources. Edward A. Sudicky: Writing - review & editing, Resources, Supervision.

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.

Acknowledgements

We would like to thank all anonymous reviewers; their comments and suggestions significantly improved this work.

References (88)

  • S.C. Hardy et al.

    Surface tension and interface kinetics of ice crystals freezing and melting in sodium chloride solutions

    J. Cryst. Growth

    (1973)
  • W.B. Hillig

    Measurement of interfacial free energy for ice/water system

    J. Cryst. Growth

    (1998)
  • F. Kastanek

    Numerical simulation technique for vertical drainage from a soil column

    J. Hydrol.

    (1971)
  • M. Khishvand et al.

    Micro-scale experimental investigation of the effect of flow rate on trapping in sandstone and carbonate rock samples

    Adv. Water Res.

    (2016)
  • B.L. Kurylyk et al.

    The mathematical representation of freezing and thawing processes in variably-saturated, non-deformable soils

    Adv. Water Res.

    (2013)
  • F. Ming et al.

    Estimation of hydraulic conductivity of saturated frozen soil from the soil freezing characteristic curve

    Sci. Total Environ.

    (2020)
  • P. Mohammadmoradi et al.

    Petrophysical characterization of porous media starting from micro-tomographic images

    Adv. Water Res.

    (2016)
  • P. Mohammadmoradi et al.

    Toward direct pore-scale modeling of three-phase displacements

    Adv. Water Res.

    (2017)
  • A.W. Neumann et al.

    Thermodynamics of contact angles. I. Heterogeneous solid surfaces

    J. Colloid Interface Sci.

    (1972)
  • L.S. Penn et al.

    A study of the primary cause of contact angle hysteresis on some polymeric solids

    J. Colloid Interface Sci.

    (1980)
  • M. Peszynska et al.

    Biofilm growth in porous media: experiments, computational modeling at the porescale, and upscaling

    Adv. Water Res.

    (2016)
  • H.S. Rabbani et al.

    Effects of intermediate wettability on entry capillary pressure in angular pores

    J. Colloid Interface Sci.

    (2016)
  • A.Q. Raeini et al.

    Modelling two-phase flow in porous media at the pore scale using the volume-of-fluid method

    J. Comput. Phys.

    (2012)
  • A.Q. Raeini et al.

    Direct simulations of two-phase flow on micro-CT images of porous media and upscaling of pore-scale forces

    Adv. Water Res.

    (2014)
  • A.Q. Raeini et al.

    Modelling capillary trapping using finite-volume simulation of two-phase flow directly on micro-CT images

    Adv. Water Res.

    (2015)
  • L.C. Ruspini et al.

    Pore-scale modeling of capillary trapping in water-wet porous media: A new cooperative pore-body filling model

    Adv. Water Res.

    (2017)
  • D. Silin et al.

    Pore space morphology analysis using maximal inscribed spheres

    Physica A

    (2006)
  • C. Sun et al.

    Linking continuum-scale state of wetting to pore-scale contact angles in porous media

    J. Colloid Interface Sci.

    (2020)
  • O. Talabi et al.

    Pore-scale simulation of NMR response

    J. Petrol. Sci. Eng.

    (2009)
  • H. Tian et al.

    Freezing and thawing characteristics of frozen soils: Bound water content and hysteresis phenomenon

    Cold Reg. Sci. Technol.

    (2014)
  • K. Watanabe et al.

    Measurement of unfrozen water content and relative permittivity of frozen unsaturated soil using NMR and TDR

    Cold Reg. Sci. Technol.

    (2009)
  • P.S. Wei et al.

    Growths of bubble/pore sizes in solid during solidification—an in situ measurement and analysis

    J. Cryst. Growth

    (2004)
  • A.N. Zhou

    A contact angle-dependent hysteresis model for soil–water retention behaviour

    Comput. Geotech.

    (2013)
  • G. Zografi et al.

    Effects of surface roughness on advancing and receding contact angles

    Int. J. Pharm.

    (1984)
  • O.B. Andersland et al.

    Hydraulic conductivity of frozen granular soils

    J. Environ. Eng.

    (1996)
  • R. Anderson et al.

    Gas hydrate growth and dissociation in narrow pore networks: capillary inhibition and hysteresis phenomena

    Geol. Soc.

    (2009)
  • Aquanty. (2019). HydroGeoSphere: A three-dimensional numerical model describing fully-integrated subsurface and surface...
  • S. Berg et al.

    Real-time 3D imaging of Haines jumps in porous media flow

    Proc. Natl. Acad. Sci.

    (2013)
  • S.J. Berg et al.

    Toward large-scale integrated surface and subsurface modeling

    Groundwater

    (2019)
  • B. Bijeljic et al.

    Pore‐scale modeling of longitudinal dispersion

    Water Resour. Res.

    (2004)
  • B. Bijeljic et al.

    Predictions of non-Fickian solute transport in different classes of porous media using direct simulation on pore-scale images

    Phys. Rev. E

    (2013)
  • P.B. Black et al.

    Comparison of soil freezing curve and soil water curve data for Windsor sandy loam

    Water Resour. Res.

    (1989)
  • R. Brooks et al.
    (1964)
  • K.J. Buschow et al.

    Encyclopedia of materials

    Sci. Technol.

    (2001)
  • Cited by (0)

    View full text