soil inﬁltration characteristics and pore distribution

The Cryosphere 15 2133ndash2146 2021httpsdoiorg105194tc-15-2133-2021copy Author(s) 2021 This work is distributed underthe Creative Commons Attribution 40 License

Soil infiltration characteristics and pore distribution underfreezingndashthawing conditionsRuiqi Jiang123 Tianxiao Li123 Dong Liu123 Qiang Fu123 Renjie Hou123 Qinglin Li1 Song Cui123 andMo Li123

1School of Water Conservancy amp Civil Engineering Northeast Agricultural University Harbin 150030 China2Key Laboratory of Effective Utilization of Agricultural Water Resources of Ministry of Agriculture Northeast AgriculturalUniversity Harbin Heilongjiang 150030 China3Heilongjiang Provincial Key Laboratory of Water Resources and Water Conservancy Engineering in Cold Region NortheastAgricultural University Harbin Heilongjiang 150030 ChinaThese authors contributed equally to this work

Correspondence Dong Liu (liudong9599yeahnet) and Qiang Fu (fuqiang0629126com)

Received 24 September 2020 ndash Discussion started 27 October 2020Revised 8 March 2021 ndash Accepted 26 March 2021 ndash Published 2 May 2021

Abstract Frozen soil infiltration widely occurs in hydro-logical processes such as seasonal soil freezing and thaw-ing snowmelt infiltration and runoff Accurate measurementand simulation of parameters related to frozen soil infiltra-tion processes are highly important for agricultural watermanagement environmental issues and engineering prob-lems in cold regions Temperature changes cause soil poresize distribution variations and consequently dynamic infil-tration capacity changes during different freezendashthaw peri-ods To better understand these complex processes and toreveal the freezendashthaw action effects on soil pore distri-bution and infiltration capacity black soils meadow soilsand chernozem were selected as test subjects These soiltypes account for the largest arable land area in HeilongjiangProvince China Laboratory tests of soils at different tem-peratures were conducted using a tension infiltrometer andethylene glycol aqueous solution The stable infiltration rateand hydraulic conductivity were measured and the soil poredistribution was calculated The results indicated that for thedifferent soil types macropores which constituted approx-imately 01 to 02 of the soil volume under unfrozenconditions contributed approximately 50 of the saturatedflow and after soil freezing the soil macropore proportiondecreased to 005 to 01 while the saturated flow pro-portion decreased to approximately 30 Soil moisture frozeinto ice crystals inside relatively large pores resulting in nu-merous smaller-sized pores which reduced the number of

macropores but increased the number of smaller-sized meso-pores so that the frozen soil infiltration capacity was nolonger solely dependent on the macropores After the icecrystals had melted more pores were formed within the soilenhancing the soil permeability

1 Introduction

Over the last few decades the temperature changes causedby global warming have altered the freezing state of near-surface soils and in China changes in characteristic valuessuch as the extent of the mean annual area of the seasonal soilfreezendashthaw state and maximum freezing depth indicate thedegradation of frozen soil especially at high latitudes (Wanget al 2019 Peng et al 2016) Under the effect of temper-ature most frozen regions experience seasonal freezing andthawing of soil accompanied by coupled soil water and heatmovement as well as frost heave processes thus making thesoil structure and function more variable (Oztas and Fayetor-bay 2003 Fu et al 2019 Gao et al 2018) Parameters suchas the soil infiltration rate and hydraulic conductivity are keyfactors in the study of soil water movement groundwaterrecharge and solute and contaminant transport simulation(Angulo-Jaramillo et al 2000) In regard to unfrozen soilstemperature has been shown to change the soil structure andkinematic viscosity of soil water thereby affecting the unsat-

Published by Copernicus Publications on behalf of the European Geosciences Union

2134 R Jiang et al Soil infiltration characteristics and pore distribution under freezingndashthawing conditions

urated hydraulic conductivity of soils (Gao and Shao 2015)In terms of frozen soils the water infiltration characteristicsand pore size distribution are highly variable and difficult toobserve (Watanabe et al 2013) moreover the water move-ment in freezingndashthawing soils is complicated by the migra-tion of water and heat and the associated water phase change(Jarvis et al 2016 Hayashi 2013) The accurate measure-ment of water movement parameters and soil pore distribu-tion under freezendashthaw conditions is a necessary prerequi-site for the quantitative description of the water movementin frozen soil and the mechanism and degree of influence ofthe temperature on the infiltration rate hydraulic conductiv-ity porosity and other parameters in the different stages offreezendashthaw periods require further research

Currently quantitative studies of the infiltration processin freezingndashthawing soils can be mainly divided into experi-mental and model studies Field experiments have been per-formed less often because under natural conditions infiltra-tion water exhibits a preferential flow into the deep soil andthe alternating freezendashthaw effect forms ice crystals to blockthe flow path through large pores subsequently limiting wa-ter infiltration (Stadler et al 1997) while the melting effectof the infiltration water on ice makes it difficult to reacha steady infiltration state Therefore the current relevantachievements are mainly focused on the infiltration processof snowmelt water (Hayashi et al 2003) and the influence ofpreferential flow (Mohammed et al 2019) Controlled lab-oratory experiments provide new opportunities for the simu-lation of frozen soil infiltration processes and the measure-ment of infiltration parameters Williams and Burt (1974)conducted early direct measurements in the laboratory re-solved the water freezing problem by adding lactose and ap-plied dialysis membranes on both sides of soil columns andthey determined the water conductivity of saturated speci-mens in the horizontal direction (Burt and Williams 1976)Andersland et al (1996) measured the hydraulic conductiv-ity of frozen granular soils at different saturations using aconventional drop permeameter with decane as the permeantand concluded that the hydraulic conductivity was the sameas that of unfrozen soils with water as the infiltration solu-tion McCauley et al (2002) determined and compared thedifferences in hydraulic conductivity permeability and infil-tration rate between frozen and unfrozen soils using dieselmixtures as permeants and their results indicated that the icecontent determines whether soil is sufficiently impermeableZhao et al (2013) quantified the unsaturated hydraulic con-ductivity of frozen soil using antifreeze instead of water andadopted a multistage outflow method under controlled pres-sures in combination with the concept of the pore impedancecoefficient (Jame and Norum 1980) However most of thesestudies did not consider the differences in kinematic viscosityand surface tension between soil water and other solutionswhich often results in an overestimation of hydraulic con-ductivity and homemade devices in the laboratory are ofteninconvenient for generalization in the field Due to the dy-

namic changes in the temperature and moisture phase directmeasurement is difficult and hydraulic conductivity empiri-cal equations and models of frozen soil have been developedFirst the frozen soil hydraulic conductivity was simply con-sidered to follow a power exponential relationship with thetemperature (Nixon 1991 Smith 1985) while others con-sidered the hydraulic conductivity of frozen soil to be equalto that of unfrozen soil at the same water content and as-sumed that the hydraulic conductivity of frozen soil was afunction of the moisture content of unfrozen soil (Lundin1990 Flerchinger and Saxton 1989 Harlan 1973) On thebasis of Campbellrsquos model (Campbell 1985) Tarnawski andWagner (1996) proposed a frozen soil hydraulic conductivitymodel based on the soil particle size distribution and poros-ity Watanabe and Wake (2008) viewed soil pores as cylindri-cal capillaries and suggested that ice formation occurs at thecenter of these capillaries and established a model to describethe movement of thin film water and capillary water in frozensoil based on the theory of capillaries and surface absorption(Watanabe and Flury 2008) The similarity between freezingand soil moisture profiles has been demonstrated (Spaans andBaker 1996 Spaans 1994) and subsequently soil freezingcharacteristic curves have been applied to estimate the unsat-urated hydraulic conductivity of frozen soils (Azmatch et al2012) which has been combined with field tests and inver-sion models to achieve a high accuracy (Cheng et al 2019)

Understanding the distribution characteristics of the soilpore system is essential for the evaluation of the water andheat movement processes in soil Soil macroporosity hasbeen shown to impose a major impact on water cycle pro-cesses such as infiltration nutrient movement and surfacerunoff (Demand et al 2019 Jarvis 2007) Macroporosityis widespread in a variety of soils and produces preferen-tial flow in both frozen and unfrozen soils (Mohammed etal 2018 Beven and Germann 2013) and the pre-freezemoisture conditions affect the amount and state of ice in themacropores of frozen soils resulting in notable variability inthe infiltration capacity of thawed soils (Hayashi et al 2003Granger et al 1984) Field experiments on frozen soil havealso demonstrated that macropores accelerate the infiltrationrate (Staumlhli et al 2004 van der Kamp et al 2003) the num-ber and size of macropores affect the freezing and infiltrationcapacity of soil layers to different extents and low tempera-tures cause infiltration water to refreeze inside macropores(Watanabe and Kugisaki 2017 Stadler et al 2000) Re-search on frozen soil macroporosity has largely focused onthe qualitative analysis of its impact on the soil structure andinfiltration capacity With the development of experimen-tal techniques certain new methods and techniques such ascomputed tomography (CT) and X-ray scanning have beenapplied to measure the number and distribution of macrop-ores (Taina et al 2013 Bodhinayake et al 2004 Greverset al 1989) but the lack of sampling techniques targetingfrozen soil still restricts related research

The Cryosphere 15 2133ndash2146 2021 httpsdoiorg105194tc-15-2133-2021

R Jiang et al Soil infiltration characteristics and pore distribution under freezingndashthawing conditions 2135

Table 1 Basic physical and chemical properties of three kinds of soils

Bulk Organic Electrical Particle size SoilSoil types density content conductivity (sandndashsiltndashclay) texture

(g cmminus3) (g kgminus1) (s mminus1) ()

Black soil 131 2832 002 1264ndash7082ndash1654 silt loamMeadow soil 122 1651 001 952ndash7300ndash1748Chernozem 115 2652 001 3899ndash5030ndash1071

Table 2 Soil porosity and pre-freezing soil water content ofrepacked samples

Soil Pre-freeingSoil types porosity water content ()

() minus5 C minus10 C

Black soil 5276 3053 2927Meadow soil 5282 2983 3070Chernozem 5322 3007 3093

Many limitations and deficiencies remain in the directmeasurement of frozen soil infiltration characteristics andpore distribution and the relevant models also require a largeamount of measured data to meet the accuracy and applica-bility requirements In this paper the stable infiltration rateand hydraulic conductivity of three types of soils at differenttemperatures were measured by precise control of the soiland ambient temperatures and the macropore and mesoporesize distribution was calculated by using a tension infiltrom-eter and a glycol aqueous solution as the infiltration mediumThe conclusions provide a basis and reference for the numer-ical simulation of the coupled waterndashheat migration processof freezingndashthawing soil and related parameterization stud-ies

2 Materials and methods

21 Test plan

Referring to arable land area data of various regions of Hei-longjiang Province the three types of soils that dominate thecultivated land area in this province are black soils meadowsoils and chernozem (Land Administrative Bureau of Hei-longjiang Province 1992) Harbin Zhaoyuan and Zhaozhouwere selected as typical soil areas for sampling A 5 cmsurface layer of floating soil and leaves was removed andtopsoil samples were collected at depths ranging from 0 to20 cm After natural air drying and artificial crushing the soilwas sieved and particles larger than 2 mm in diameter wereremoved The remainder was used to prepare soil columnsThe basic physical and chemical parameters of the test soils

such as the bulk density organic content and mechanical pa-rameters are listed in Table 1

An artificial climate chamber was applied to control thetemperature of the soil column and infiltration solutionand four temperature treatments were established with threereplications for each treatment 15 C unfrozen soil repre-senting the soil before freezing which was recorded as 15 C(BF) minus5 C stable freezing minus10 C stable freezing andfreezing at minus10 C followed by thawing at 15 C represent-ing the soil after melting which was recorded as 15 C (AM)Each soil column was tested for only one treatment Thefreezing and thawing times were both 48 h When the soiltemperature was consistent with the set temperature in theclimate chamber the samples were considered to be com-pletely frozen and the effect of the number of freezing andthawing cycles was not considered in this test Accordingto the basic information of the original soil the volumetricmoisture content of the sieved soil was set to 30 plusmn 2 us-ing deionized water with a dry bulk density of 12 g cmminus3To reduce the inhomogeneous distribution of soil moisturea spray bottle was used to add water to the sieved and air-dried soil in small but repeated amounts while the change insoil moisture was detected using sensors so that the mois-ture content of the soil columns was controlled within theset range and the texture was uniform as well as making thepre-freeze moisture content relatively consistent between dif-ferent soil columns The soil porosity and pre-freezing soilwater content of the repacked samples are shown in Table 2

To ensure a homogeneous column the soil was loadedinto a polyvinyl chloride (PVC) cylinder at 5 cm depth in-tervals and petroleum jelly was applied to the sides to re-duce the sidewall flow (Lewis and Sjoumlstrom 2010) The PVCcylinder was 26 cm in diameter and 30 cm in height witha perforated plate at the bottom To prevent lateral seep-age the barrel occurred 5 cm above the soil surface and thethickness of the soil layer was 20 cm A HYDRA-PROBEII sensor (STEVENS Water Monitoring Systems Inc Port-land Oregon USA) was inserted in the middle of the bar-rel to observe the potential soil temperature and liquid wa-ter content change to determine whether ice melting oc-curred and the ice content of frozen soil was measured bythe drying method A 5 cm thick layer of sand and gravelwas placed below the soil column and a 5 cm thick layer ofblack polypropylene insulation cotton was wrapped around

httpsdoiorg105194tc-15-2133-2021 The Cryosphere 15 2133ndash2146 2021


the outer layer and the bottom of the soil column The sta-ble infiltration rate under tension levels of minus3 minus5 minus7 minus9minus11 and minus13 cm was measured with a tension infiltrome-ter and the infiltration time and cumulative infiltration wererecorded The detailed layout of the test apparatus is shownin Fig 1

The addition of a certain amount of lactose antifreeze orother substances to water greatly reduces the freezing pointof water (Zhao et al 2013 Williams and Burt 1974) so thatthe soil macropores are not quickly filled with ice with de-creasing temperature thereby maintaining better conditionsfor water flow To further verify the feasibility of the useof deionized water to prepare an aqueous solution of ethy-lene glycol at a mass concentration of 40 as the infiltrationmedium for the frozen soil measurements the surface ten-sion of the aqueous glycol solution at minus5 and minus10 C and itsrelationship with the temperature were measured with a con-tact angle measuring instrument (OCA20 DataPhysics In-struments Germany) and a surface tension measuring instru-ment (DCAT-21 DataPhysics Instruments Germany) re-spectively As an example the contact angle measurementprocess of the black soil at minus10 C with the aqueous ethy-lene glycol solution is shown in Fig 2 It is observed that thecontact angle decreases to 0 C within a few seconds afterthe liquid droplet is placed on the soil and the liquid dropletcompletely dissolves in the frozen soil which implies thatthe addition of glycol to water does not alter the wetting abil-ity of the soil particles (Lu and Likos 2004) For the un-frozen soils of the 15 C (BF) and 15 C (AM) treatmentsdeionized water was used as the infiltration solution For thefrozen soils of the minus5 and minus10 C treatments aqueous ethy-lene glycol solution was used as the infiltration solution Therelevant physicochemical properties of the aqueous ethyleneglycol solution and water are compared in Table 3

22 Measurement of the frozen soil hydraulicconductivity

Gardner (1958) proposed that the unsaturated hydraulic con-ductivity of soil varies with the matric potential as follows

K(h)=Ksatexp(αh) (1)

where Ksat is the saturated hydraulic conductivity cm hminus1and h is the matric potential or tension cm H2O Wood-ing (1968) considered that the steady-state unconfined infil-tration rate into soil from a circular water source of radius Rcan be calculated with the following equation

Q= πR2K

[1+

4πRα

] (2)

where Q is the amount of water entering the soil per unittime cm3 hminus1 K is the hydraulic conductivity cm hminus1 andα is a constant Ankeny et al (1991) proposed that imple-menting two successively applied pressure heads h1 and h2

Figure 1 Diagram of the test equipment

Figure 2 Process of the contact angle measurement between theaqueous ethylene glycol solution and black soil at minus10 C

could yield the unsaturated hydraulic conductivity and uponreplacingK in Eq (2) with Eq (1) the following is obtained

Q(h1)= πR2Ksat exp(αh1)

[1+

4πRα

] (3)


[1+

4πRα

] (4)

Dividing Eq (4) by Eq (3) and solving for α yields

α =ln

[Q(h2)Q(h1)

]h2minush1

(5)

where Q(h1) and Q(h2) can be measured h1 and h2 are thepreset tension values and α can be calculated with Eq (5)The result can be substituted into Eq (3) or (4) to calculateKsat When the number of tension levels is higher than 2parameter fitting methods can be applied to improve the ac-curacy of α and Ksat (Hussen and Warrick 1993)

The tension is controlled by the bubble collecting tube ofthe tension infiltrometer and different pressure heads h cor-respond to different pore sizes r By applying different pres-sure heads h to the soil surface water will overcome thesurface tension in the corresponding pores and will be dis-charged and the infiltration volume is recorded after reach-ing the stable infiltration state



Table 3 Comparison of the physicochemical properties of the 40 ethylene glycol aqueous solution and water

Dynamic SurfaceInfiltration Temperature Density viscosity tension Contactsolution (C) (g cmminus3) (mPa s) (mN mminus1) angle ()

Water 15 09991 114 7356 0

Ethylene glycol minus5 10683 718 4889 0aqueous solution minus10 10696 906 4910 0

Under the assumption that the frozen soil pore ice pres-sure is equal to the atmospheric pressure and that solutes arenegligible the ClausiusndashClapeyron equation can be adoptedto achieve the interconversion between the soil temperatureand suction (Konrad and Morgenstern 1980 Watanabe et al2013) which can be simplified as follows

ψ =minusLρwT

27315 (6)

where ψ is the soil suction kPa L is the latent heat of fu-sion of water 334times 105 J kgminus1 ρw is the density of water1 g cmminus3 and T is the subfreezing temperature C Afterthe unit conversion of the soil suction into h (cm H2O) theunsaturated hydraulic conductivity of frozen soil at differentnegative temperatures can be obtained via substitution intoEq (2)

23 Measurement of the pore size distribution in frozensoil

As a nonuniform medium soil consists of pores with variouspore sizes and the equation for the soil pore radius r can beobtained from the capillary model (Watson and Luxmoore1986)

r =minus2σ cosβρgh

(7)

where σ is the surface tension of the solution g sminus2 β is thecontact angle between the solution and pore wall ρ is thedensity of the solution g cmminus3 g is the acceleration of grav-ity m sminus2 and h is the corresponding tension of the tensioninfiltrometer cm H2O

The effective macroporosity θm can be calculated for vari-ous soil particle sizes based on the Poiseuille equation (Wil-son and Luxmoore 1988)

θm = 8microKmρgr2 (8)

where micro is the dynamic viscosity of the fluid g (cm s)minus1Km is the macropore hydraulic conductivity and is definedas the difference between K(h) at various tension gradientscm hminus1 and r is the corresponding equivalent pore size Theeffective porosity is equal to the number of pores per unitarea multiplied by the area of the corresponding pore size

For different pore sizes the maximum number of effectivemacropores per unit area N can be calculated with the fol-lowing equation

N = θmπr2 (9)

where N is the number of effective macropores per unit areaand Eq (7) calculates the minimum value of the pore radiuswhile the result obtained with Eq (9) is actually the maxi-mum number of effective macropores per unit area and themaximum porosity

Considering the differences in surface tension and den-sity between the aqueous ethylene glycol solution and wa-ter when calculating the frozen soil pore size distribution itis necessary to convert the tension into the equivalent poreradius according to Eq (7) which is classified and subdi-vided into large and medium pores according to the commonclassification method (Luxmoore 1981) the details of whichare listed in Table 4 while the corresponding tension valuesin Table 4 are substituted into the fitting curve equation tocalculate the corresponding stable infiltration rate q and un-saturated hydraulic conductivity K

3 Results

31 Infiltration characteristics of freezingndashthawing soils

Curves of the recorded cumulative infiltration and infiltra-tion rate were plotted over time as shown in Figs 3 and 4respectively The unfrozen water contents and ice contentsof the frozen samples are shown in Table 5 The constantα and saturated hydraulic conductivity Ksat were calculatedunder different tensions h and corresponding steady-state in-filtration rates q and the unsaturated hydraulic conductivityunder different tensions was calculated with Eq (1) The sta-ble infiltration rate and unsaturated hydraulic conductivity atdifferent temperatures are shown in Fig 5 and the details ofα and Ksat are listed in Table 6

As shown in Figs 4 and 5 under the different tensionconditions the infiltration capacity of the unfrozen soil isbasically consistent with the findings of field experimentsand is highly influenced by the tension value (Wang et al1998) Compared to the room-temperature soil the cumula-tive infiltration of frozen soil slowly increases and the infil-



Table 4 Tension and equivalent pore radius conversions

Tension conversion (cm)

Pore types Pore radius Water (15 C) Ethylene glycol Ethylene glycol(mm) aqueous solution aqueous solution

(minus5 C) (minus10 C)

Macroporous gt 05 0ndash3 0ndash186 0ndash187

Mesoporous 03-05 3ndash5 186ndash311 186ndash312015ndash03 5ndash10 311ndash622 312ndash62301ndash015 10ndash15 622ndash932 623ndash935005ndash01 15ndash30 932ndash1865 935ndash1870

Figure 3 Cumulative infiltration over time under the different treatments (a) Black soil (b) meadow soil (c) chernozem

tration rate always remains low while under the same nega-tive temperature treatment the influence of the tension valueis also greatly reduced When the temperature was reduced tominus10 C few major tension differences were observed exceptfor the maximum tension of minus3 cm Based on the change inthe slope of the two curves the time for the unfrozen soilto reach the stable infiltration rate usually ranges from 15ndash

20 min while the time for the frozen soil to reach the stableinfiltration rate is usually 10 min under higher tensions ofminus3 and minus5 cm and 5 min under lower tensions A compar-ison of the infiltration process before and after the freezingand thawing of the soil indicates that overall the cumulativeinfiltration and infiltration rate exhibited varying degrees ofincrease with increasing tension value and the increase am-



Figure 4 Infiltration rate over time under the different treatments (a) Black soil (b) meadow soil (c) chernozem

Table 5 Unfrozen water contents and ice contents of the frozensamples

Unfrozen water Ice contentsSoil types contents (cm3 cmminus3) (cm3 cmminus3)

minus5 C minus10 C minus5 C minus10 C

Black soil 0123 0101 0159 0179Meadow soil 0128 0109 0145 0167Chernozem 0119 0097 0151 0185

plitude expanded Moreover the difference in the cumulativeinfiltration and infiltration rate between the low tension lev-els ranging from minus9 to minus13 cm after soil thawing was largerthan that before soil freezing which also indirectly demon-strated that freezing and thawing could further stabilize thesoil pore distribution by affecting the homogeneity whichwill be detailed in subsequent sections

By combining Fig 5 and Table 6 we observe that the threetypes of soils exhibit a high infiltration capacity under normal

temperature conditions With increasing set tension valuethe suction force of the soil matrix gradually weakened theconstraint and maintenance capacity of the matric potentialwith respect to the soil water decreased the number of poresinvolved in the soil water infiltration process increased andthe unsaturated hydraulic conductivity and stable infiltrationrate of the three types of soils exhibited different degreesof increase When the temperature was lowered from 15 tominus5 C and remained constant for a certain time the soilreached the stable frozen state which usually indicated thatthere were no more drastic changes in temperature and mois-ture content the saturated water conductivity of the blacksoil meadow soil and chernozem soil decreased by 8842 8078 and 738 respectively When the soil temperaturewas decreased tominus10 C due to the presence of liquid waterin the pores the saturated water conductivity still exhibiteda certain decrease over the pre-freeze conditions and contin-ued to decrease by 143 1094 and 985 respectivelyAt negative temperatures the unsaturated hydraulic conduc-tivity decreased considerably and fluctuated within a small



Table 6 Infiltration parameters of the different temperature treatments of the three soil types

Soil types Temperature (C) α (cm hminus1) Ksat (cm hminus1) Permeability (m2)

Black soil 15 (BF) 02742 51480 166times10minus9

minus5 01993 05960 114times10minus9


15 (AM) 02629 74658 241times10minus9

Meadow soil 15 (BF) 03071 59232 192times10minus9



15 (AM) 02934 93757 303times10minus9

Chernozem 15 (BF) 02166 37185 120times10minus9



15 (AM) 02182 51283 166times10minus9

Figure 5 Variation curves of the unsaturated hydraulic conductivity and stable infiltration rate with the tension for the different treatmentsof the three soils (a) Black soil (b) meadow soil (c) chernozem The solid lines represent the stable infiltration rate and the dashed linesrepresent the unsaturated hydraulic conductivity

range mainly because the unfrozen water content and sat-urated hydraulic conductivity were low after soil freezingBased on a comparison of the minus5 and minus10 C treatmentsthe unsaturated hydraulic conductivity (ANOVA P = 072F = 014) and stable infiltration rate (ANOVA P = 071F = 015) of the black soil exhibited almost no significantchange indicating that most of its pores were filled with icecrystals at minus5 C and were no longer involved in water in-filtration The unsaturated water conductivity of the meadowand chernozem soils still exhibited a more significant reduc-tion when the freezing temperature was further reduced tominus10 C When the temperature was raised again to 15 Cand the soil was completely thawed the steady infiltrationrate and saturated hydraulic conductivity increased with in-creasing temperature and the values were higher than thoseof the soil at the same temperature before freezing The sat-urated hydraulic conductivity of the black soil meadow soiland chernozem increased by 4502 5863 and 3791 respectively relative to the 15 C (BF) treatment values

32 Pore distribution characteristics of thefreezingndashthawing soil

Considering the differences in the physical and chemicalproperties between the infiltration solutions infiltration pa-rameters such as the hydraulic conductivity and stable infil-tration rate alone do not fully reflect the infiltration charac-teristics and internal pore size of frozen soils According toEqs (7)ndash(9) the maximum number per unit areaN effectiveporosity θm and percentage of pore flow to saturated flowP corresponding to the different soil pore sizes of the threesoils under the different temperature treatments were calcu-lated as shown in Figs 6 and 7

Figure 6 shows that pores of different equivalent radiiwidely occur in all three soils and under all four temper-ature treatments the largest N value is that for the mediumpores with an equivalent radius of 005ndash01 mm andN grad-ually decreases with increasing equivalent radius size Un-der the two room-temperature treatments at 15 C (BF) and15 C (AM) the largest number of 005 to 01 mm mediumpores and the smallest number of gt 05 mm macropores dif-



Figure 6 Number of pores and effective porosity of the different equivalent pores (a) Black soil (b) meadow soil (c) chernozem

Figure 7 Percentage of the pore flow in the saturated flow for the different equivalent pore sizes (a) Black soil (b) meadow soil (c) cher-nozem

fered by 2 orders of magnitude The number of pores ofeach size exhibited different degrees of increase or decreasein the two treatments at minus5 and minus10 C where freezing oc-curred with the number of medium pores with an equivalentpore size of 005ndash01 mm significantly changing Increases ofmore than an order of magnitude were achieved in all threesoils while the macropores with an equivalent pore size ofgt 05 mm were generally reduced by an order of magnitudewith the difference in the number of pores of these two sizesreaching 4 orders of magnitude This indicates that freezingcaused by temperature change significantly alters the soil in-

ternal structure with ice crystals forming in the relativelylarge pores containing the internal soil moisture resulting ina large number of smaller pores Based on an assessment ofthe two treatments at minus5 and minus10 C separately when thetemperature was lowered from minus5 to minus10 C the numberof pores in each pore size interval of the meadow and cher-nozem soils exhibited a significant decrease while the blacksoil revealed a small increase which might be related to thehigh organic matter content of the black soil A comparisonof the two treatments at 15 C (BF) and 15 C (AM) indicatedthat the number of pores in all three soils increased to differ-



ent degrees after thawing and more pores were formed withthe melting of ice crystals after the freezendashthaw destructionof the soil particles which enhanced the soil water conduc-tivity

Comprehensive analysis of Figs 6 and 7 reveals that be-fore freezing the θm values of the various pore sizes of theblack soils meadow soils and chernozem with an equiva-lent radius of gt 05 mm were 011 013 and 007 respectively while the P value reached 5746 5565 and 4899 respectively with the values of the thawed soilbeing similar to these values This indicates that for all fivesoil pore sizes under unfrozen conditions although the num-ber of macropores with a pore size gt 05 mm was the small-est and the effective porosity was the lowest their contribu-tion to the saturated flow was usually more than half andthe macropores needed to represent only a small fraction ofthe pore volume to significantly contribute to the soil wa-ter flow For the frozen soil the P value of the gt 05 mmmacropores was significantly reduced and remained at ap-proximately 30 after the reduction while the P value ofthe smaller pore sizes such as 015ndash03 01ndash015 and 005ndash01 mm revealed different degrees of increase Moreover thesmaller the pore size the greater the P value increased andthe contribution of these pores eventually accounted for morethan 10 of the saturated flow The saturated flow becamemore evenly distributed across the pores of each size and thetotal proportion of medium pores exceeded that of the macro-pores This indicates that the freezing action caused obviouschanges to the soil structure pore size and quantity and al-though the macropores still played an important role the in-filtration capacity of the frozen soil no longer relied solelyon these macropores and the contribution of certain smaller-sized mesopores to the infiltration capacity of the frozen soilcould no longer be neglected Selecting black soil as an ex-ample the total effective porosity of the pores of each sizeunder the four treatments was 115 262 284 and180 and the P values were 9997 9756 9772 and 9996 respectively which implies that the soil waterinfiltrated almost entirely via the large and medium poresThe small micropores even in large numbers contributed lit-tle to the infiltration process

4 Discussion

41 Permeability and hydraulic conductivity of thefrozen soil

It is worth noting that the theory underpinning the tensioninfiltration analysis in this article is based on the assumptionthat larger pores only flow fully saturated which means noairndashwater interface inside the pore and excludes the forma-tion of an airndashwater interface with flowing water in largerpores (Perroux and White 1988) However recent work hasshown that this flow mode does indeed occur (Beven and

Germann 2013 Nimmo 2012 2010) In addition in un-frozen tension infiltrometer experiments the soil moisture isassumed to be that imposed by the applied tension whereasin the experiments with frozen samples the applied tensionis assumed to affect only the pores that are active during in-filtration

In the field environment although it is difficult to accu-rately measure the infiltration rate of frozen soils using tradi-tional instruments and methods such as single-loop infiltra-tors the obtained test results still demonstrate that the infil-tration capacity decreases by 1 or more orders of magnitudewhen the soil is frozen (Staumlhli et al 2004) Although thecumulative infiltration and infiltration rate of frozen soil arelow the presence of unfrozen water allows a certain amountof infiltration flow to be maintained in the soil When water isapplied as the infiltration solution the low temperature of thefrozen soil easily causes the infiltration water to freeze thusforming a thin layer of ice on the soil particle surface and de-laying the subsequent infiltration of water This phenomenonresults in a low infiltration rate after the freezing of soils witha high initial water content and a relatively high infiltrationrate after the freezing of dry soils (Watanabe et al 2013) be-cause the higher the ice content is the more latent heat needsto be overcome to melt any ice crystals resulting in a weak-ened propagation of the melting front thus limiting the in-filtration rate so that it is controlled by the downward move-ment of the melting leading edge of the ice crystals (Pittmanet al 2020) During the measurements using the tension in-filtrator in this study the sensor temperature always remainedconsistent with the soil temperature indicating that the useof an aqueous glycol solution could be a useful way to avoidthe problem of freezing of the infiltration solution In addi-tion the hydraulic conductivity of frozen soils with differentcapacities and at various water flow rates was demonstratednot to greatly differ (Watanabe and Osada 2017)

Whether water or other low-freezing point solutions areapplied as infiltration media the hydraulic conductivity offrozen soil significantly changes only within a limited tem-perature range aboveminus05 C depending on the unfrozen wa-ter and ice contents and at a soil temperature belowminus05 Cthe hydraulic conductivity usually decreases to less than10minus10 m sminus1 (Watanabe and Osada 2017 Williams and Burt1974) The unsaturated hydraulic conductivity in our experi-ments was measured at a set tension level and according toEq (6) the soil substrate potential increases by 125 m for ev-ery 1 C decrease in temperature (Williams and Smith 1989)while the frozen soil hydraulic conductivity calculated at minus5and minus10 C which corresponds to the actual matric poten-tial is much lower than 10minus10 m sminus1 close to the zero pointof the unsaturated hydraulic conductivity variation curves inFig 5 and can be ignored This suggests that even underideal conditions where no heat exchange occurs between theinfiltration solution and the soil and no freezing of the in-filtration water takes place to prevent subsequent infiltrationthe unsaturated hydraulic conductivity of frozen soil is so low



that the frozen soil at lower temperatures in its natural statecould be considered impermeable with respect to both waterand other solutions It should also be noted that we used dif-ferent liquids at different temperatures and that the differencein infiltration parameters such as Ksat with temperature is re-lated to the different viscosities of the liquid It is predictablethat if water is also used in frozen soils such differences willbe relatively insignificant as the ice melts

42 Effect of freezendashthaw cycles on soil poredistribution

In our study theN value after freezing for the different typesof soil was approximately 1000ndash2000 mminus2 based on the ten-sion infiltrator which agreed well with other studies such asconventional soil ring methods using X-ray and CT (Dinget al 2019 Holten et al 2018) and remained at the samemagnitude indicating that the method is generally reliableCompared with saturated frozen soils (Ding et al 2019) theeffect of freezendashthaw cycles on the pore distribution of un-saturated frozen soil is also significant Moreover it shouldbe noted that instruments such as CT are usually expensiveand difficult to carry tension infiltrators are affordable andwidely used and combined with temperature control equip-ment they will be helpful for field measurements of macro-porosity in frozen soils The freezendashthaw effect significantlyimproves the water conductivity of the different types of soilsbecause it increases the porosity decreases the soil compact-ness and dry weight and thus increases the soil water con-ductivity (Fouli et al 2013) On this basis we also foundthat the freezendashthaw process significantly altered the sizeand number of soil pores especially after freezing and thenumber of macropores decreased while the contribution ofmacropores to the saturated flow decreased The proportionof the saturated flow in the mesopores with a pore size oflt 03 mm approached or even exceeded the proportion in themacropores indicating that the soil water inside relativelylarge pores is more likely to freeze which in turn createsa large number of small pores whereas the water transferprocess in unfrozen soils primarily relies on the macroporeswith obvious differences (Wilson and Luxmoore 1988 Wat-son and Luxmoore 1986) The unsaturated water conductiv-ity of the frozen soils measured in this study was quite lowbut under human control (Watanabe and Kugisaki 2017) ornatural conditions in the field (Espeby 1990) water has beenshown to infiltrate frozen soils through macropores as longas the pore size is large enough Considering that the soil inthis experiment was disturbed soil that had been air dried andsieved although the macropores created by tillage practices(Lipiec et al 2006) and invertebrate activities (Lavelle et al2006) were excluded due to the inherent heterogeneity of thesoil particles macropores remain in the uniformly filled soilcolumn (Cortis and Berkowitz 2004 Oswald et al 1997)and these macropores still played a role in determining theinfiltration water flow In addition according to the equa-

tions of N and θm it can be found that the main source oftheir uncertainty is the value of the pore radius r and Eq (7)calculates the minimum value of pore radius while the resultobtained with Eq (9) is in fact the upper limit of effectivemacropores per unit area and effective porosity which mayalso lead to a certain overestimation Furthermore freezingof water may also have a potential effect on the microstruc-ture of soil pores (Wan and Yang 2020)

There are still only a few studies related to frozen soilmacropore flow and pore distribution consequently moredata should be acquired and more models should be devel-oped to better understand water movement in frozen soil re-gions The experiments in this paper are based on repackedsoils that were subjected to the first freezendashthaw cycle in thelaboratory and the conclusions may not be comprehensiveIn subsequent studies we will consider applying the methodsused in this paper to field experiments to examine the dynam-ics of the infiltration capacity and pore distribution in non-homogeneous soils during whole freezendashthaw periods underreal outdoor climatic conditions such as lower temperaturesand more severe freezendashthaw cycles but the infiltration so-lution must be carefully selected as ethylene glycol has lowtoxicity to prevent contamination of agricultural soils andcrops a certain concentration of lactose could be consid-ered (Burt and Williams 1976 Williams and Burt 1974)At room temperature an ethylene glycol aqueous solutionand water have similar densities and relatively similar vis-cosities We have compared these two infiltration solutions inunfrozen soil field experiments and the infiltration and poreconditions were basically similar so we still used an aque-ous glycol solution in the frozen soil laboratory experimentMeasurements should focus on frozen soil layers at differ-ent depths especially in the vicinity of freezing peaks andthe spatial variability in the distribution of frozen soil poresshould be investigated This work helps to improve the ac-curacy of simulations such as those of frozen soil water andheat movement or snowmelt water infiltration processes

5 Conclusions

In this paper the infiltration capacity of soil columns un-der four temperature treatments representing various freezendashthaw stages was measured and the distribution of the poresof various sizes within the soil was calculated based on themeasurements by applying an aqueous ethylene glycol so-lution with a tension infiltrator in the laboratory The re-sults revealed that for the three types of soils ie blacksoil meadow soil and chernozem the macropores whichaccounted for only approximately 01 to 02 of thesoil volume at room temperature contributed approximately50 to the saturated flow and after freezing the propor-tion of macropores decreased to 005 to 01 while theirshare of the saturated flow decreased to approximately 30 Coupled with the even smaller mesopores the large and



medium pores accounting for approximately 1 to 2 ofthe soil volume conducted almost all of the soil moisture un-der saturated conditions Freezing decreased the number ofmacropores and increased the number of smaller-sized meso-pores thereby significantly increasing their contribution tothe frozen soil infiltration capacity so that the latter was nolonger solely dependent on the macropores The infiltrationparameters and pore distribution of the black soil were theleast affected by the different negative freezing temperaturesunder the same moisture content and weight capacity con-ditions while those of the meadow soil were the most im-pacted

Data availability Data used in this study are available at Figshare(httpsdoiorg106084m9figshare12965123 Jiang 2020)

Author contributions RJ designed the research program TL andRJ built and deployed the soil column and instruments with assis-tance from QL and RH DL and QF provided funding for test equip-ment SC collected soil samples in the field RJ and TL analyzed thelaboratory data RJ prepared the manuscript with comments fromTL and DL

Competing interests The authors declare that they have no conflictof interest

Acknowledgements We are grateful for the guidance of the exper-iment scheme from Xianghao Wang We thank Hang Zhao for hisassistance in operating the apparatus We thank Xiaoying Lv for herhelp in drawing the diagram An earlier manuscript greatly benefit-ted from the helpful comments by the reviewer Aaron MohammedWe also appreciate the numerous insightful suggestions given byhandling editor Ylva Sjoumlberg

Financial support This research was supported by the NationalScience Fund for Distinguished Young Scholars (51825901) theJoint Fund of the National Natural Science Foundation of China(U20A20318) the Heilongjiang Provincial Science Fund for Dis-tinguished Young Scholars (YQ2020E002) the ldquoYoung TalentsrdquoProject of Northeast Agricultural University(18QC28) the ChinaPostdoctoral Science Foundation (2019M651247) and the Postdoc-toral Science Foundation of Heilongjiang Province(LBH-Z19003)

Review statement This paper was edited by Ylva Sjoumlberg and re-viewed by Aaron Mohammed and one anonymous referee

References

Andersland O B Wiggert D C and Davies S HHydraulic conductivity of frozen granular soils J Envi-

ron Eng 122 212ndash216 httpsdoiorg101061(ASCE)0733-9372(1996)1223(212) 1996

Angulo-Jaramillo R Vandervaere J-P Roulier S Thony J-L Gaudet J-P and Vauclin M Field measurement of soilsurface hydraulic properties by disc and ring infiltrometersA review and recent developments Soil Till Res 55 1ndash29httpsdoiorg101016S0167-1987(00)00098-2 2000

Ankeny M D Ahmed M Kaspar T C and HortonR Simple field method for determining unsaturated hy-draulic conductivity Soil Sci Soc Am J 55 467ndash470httpsdoiorg102136sssaj199103615995005500020028x1991

Azmatch T F Sego D C Arenson L U and BiggarK W Using soil freezing characteristic curve to es-timate the hydraulic conductivity function of partiallyfrozen soils Cold Reg Sci Technol 83 103ndash109httpsdoiorg101016jcoldregions201207002 2012

Beven K and Germann P Macropores and water flowin soils revisited Water Resour Res 49 3071ndash3092httpsdoiorg101002wrcr20156 2013

Bodhinayake W Si B C and Xiao C New method fordetermining water-conducting macro-and mesoporosity fromtension infiltrometer Soil Sci Soc Am J 68 760ndash769httpsdoiorg102136sssaj20040760 2004

Burt T and Williams P J Hydraulic conductiv-ity in frozen soils Earth Surf Proc 1 349ndash360httpsdoiorg101002esp3290010404 1976

Campbell G S Soil physics with BASIC transport models forsoil-plant systems Elsevier Amsterdam 1985

Cheng Q Xu Q Cheng X Yu S Wang Z Sun Y YanX and Jones S B In-situ estimation of unsaturated hydraulicconductivity in freezing soil using improved field data and in-verse numerical modeling Agr Forest Meteorol 279 107746httpsdoiorg101016jagrformet2019107746 2019

Cortis A and Berkowitz B Anomalous transport in ldquoclassicalrdquosoil and sand columns Soil Sci Soc Am J 68 1539ndash1548httpsdoiorg102136sssaj20041539 2004

Demand D Selker J S and Weiler M Influences of macroporeson infiltration into seasonally frozen soil Vadose Zone J 181ndash14 httpsdoiorg102136vzj2018080147 2019

Ding B Rezanezhad F Gharedaghloo B Cappellen P V andPasseport E Bioretention cells under cold climate conditionsEffects of freezing and thawing on water infiltration soil struc-ture and nutrient removal Sci Total Environ 649 749ndash7592019

Espeby B Tracing the origin of natural waters in a glacialtill slope during snowmelt J Hydrol 118 107ndash127httpsdoiorg1010160022-1694(90)90253-T 1990

Flerchinger G N and Saxton K E Simultaneous Heat and Wa-ter Model of a Freezing Snow-Residue-Soil System I The-ory and Development Am Soc Agr Eng 32 573ndash576httpsdoiorg1013031201331041 1989

Fouli Y Cade-Menun B J and Cutforth H W Freezendashthaw cycles and soil water content effects on infiltration rateof three Saskatchewan soils Can J Soil Sci 93 485ndash496httpsdoiorg104141CJSS2012-060 2013

Fu Q Zhao H Li T Hou R Liu D Ji Y Zhou Z andYang L Effects of biochar addition on soil hydraulic proper-



ties before and after freezing-thawing Catena 176 112ndash124httpsdoiorg101016jcatena201901008 2019

Gao B Yang D Qin Y Wang Y Li H Zhang Y and ZhangT Change in frozen soils and its effect on regional hydrol-ogy upper Heihe basin northeastern QinghaindashTibetan PlateauThe Cryosphere 12 657ndash673 httpsdoiorg105194tc-12-657-2018 2018

Gao H and Shao M Effects of temperature changes onsoil hydraulic properties Soil Till Res 153 145-154httpsdoiorg101016jstill201505003 2015

Gardner W Some steady-state solutions of the unsaturated mois-ture flow equation with application to evaporation from a watertable Soil Sci 85 228ndash232 httpsdoiorg10109700010694-195804000-00006 1958

Granger R J Gray D M and Dyck G E Snowmelt infil-tration to frozen Prairie soils Can J Earth Sci 21 669ndash677httpsdoiorg101139e84-073 1984

Grevers M Jong E D and St Arnaud R The characterizationof soil macroporosity with CT scanning Can J Soil Sci 69 629ndash637 httpsdoiorg104141cjss89-062 1989

Harlan R Analysis of coupled heat-fluid transport in par-tially frozen soil Water Resour Res 9 1314ndash1323httpsdoiorg101029WR009i005p01314 1973

Hayashi M The Cold Vadose Zone Hydrological and EcologicalSignificance of Frozen-Soil Processes Vadose Zone J 12 1ndash8httpsdoiorg102136vzj2013030064 2013

Hayashi M van der Kamp G and Schmidt R Fo-cused infiltration of snowmelt water in partially frozensoil under small depressions J Hydrol 270 214ndash229httpsdoiorg101016S0022-1694(02)00287-1 2003

Holten R Be F N Almvik M Katuwal S and Eklo O M Theeffect of freezing and thawing on water flow and MCPA leachingin partially frozen soil J Contam Hydrol 219 72ndash85 2018

Hussen A and Warrick A Alternative analyses of hydraulic datafrom disc tension infiltrometers Water Resour Res 29 4103ndash4108 httpsdoiorg10102993WR02404 1993

Jame Y W and Norum D I Heat and mass transfer in a freezingunsaturated porous medium Water Resour Res 16 811ndash8191980

Jarvis N A review of non-equilibrium water flow and solutetransport in soil macropores Principles controlling factors andconsequences for water quality Eur J Soil Sci 58 523ndash546httpsdoiorg101111j1365-2389200700915x 2007

Jarvis N Koestel J and Larsbo M UnderstandingPreferential Flow in the Vadose Zone Recent Ad-vances and Future Prospects Vadose Zone J 15 1ndash11httpsdoiorg102136vzj2016090075 2016

Jiang R Data for ldquoSoil infiltration characteristics and poredistribution under freezing-thawing conditionrdquo xlsx figshareDataset httpsdoiorg106084m9figshare12965123v4 2020

Konrad J-M and Morgenstern N R A mechanistic theory of icelens formation in fine-grained soils Can Geotech J 17 473ndash486 httpsdoiorg101139t80-056 1980

Land Administrative Bureau of Heilongjiang Province Hei-longjiang soil Agriculture Press Beijing 1992

Lavelle P Decaeumlns T Aubert M Barot S B Blouin M Bu-reau F Margerie P Mora P and Rossi J-P Soil inverte-brates and ecosystem services Eur J Soil Biol 42 S3ndashS15httpsdoiorg101016jejsobi200610002 2006

Lewis J and Sjoumlstrom J Optimizing the experimental design ofsoil columns in saturated and unsaturated transport experimentsJ Contam Hydrol 115 1ndash13 2010

Lipiec J Kus J Słowinska-Jurkiewicz A and Nos-alewicz A Soil porosity and water infiltration as influ-enced by tillage methods Soil Till Res 89 210ndash220httpsdoiorg101016jstill200507012 2006

Lu N and Likos W J Unsaturated soil mechanics Wiley Hobo-ken 2004

Lundin L-C Hydraulic properties in an operationalmodel of frozen soil J Hydrol 118 289ndash310httpsdoiorg1010160022-1694(90)90264-X 1990

Luxmoore R Micro- meso- and macroporos-ity of soil Soil Sci Soc Am J 45 671ndash672httpsdoiorg102136sssaj198103615995004500030051x1981

McCauley C A White D M Lilly M R and NymanD M A comparison of hydraulic conductivities permeabil-ities and infiltration rates in frozen and unfrozen soils ColdReg Sci Technol 34 117ndash125 httpsdoiorg101016S0165-232X(01)00064-7 2002

Mohammed A A Kurylyk B L Cey E E and HayashiM Snowmelt infiltration and macropore flow in frozen soilsOverview knowledge gaps and a conceptual framework Va-dose Zone J 17 1ndash15 httpsdoiorg102136vzj20180400842018

Mohammed A A Pavlovskii I Cey E E and Hayashi M Ef-fects of preferential flow on snowmelt partitioning and ground-water recharge in frozen soils Hydrol Earth Syst Sci 23 5017ndash5031 httpsdoiorg105194hess-23-5017-2019 2019

Nimmo J R Theory for Source-Responsive and Free-SurfaceFilm Modeling of Unsaturated Flow Vadose Zone J 9 295ndash306 httpsdoiorg102136vzj20090085 2010

Nimmo J R Preferential flow occurs in unsatu-rated conditions Hydrol Process 26 786ndash789httpsdoiorg101002hyp8380 2012

Nixon J Discrete ice lens theory for frost heave in soils CanGeotech J 28 843ndash859 httpsdoiorg101139t91-102 1991

Oswald S Kinzelbach W Greiner A and Brix G Observa-tion of flow and transport processes in artificial porous mediavia magnetic resonance imaging in three dimensions Geoderma80 417ndash429 httpsdoiorg101016S0016-7061(97)00064-51997

Oztas T and Fayetorbay F Effect of freezing and thaw-ing processes on soil aggregate stability Catena 52 1ndash8httpsdoiorg101016S0341-8162(02)00177-7 2003

Peng X Frauenfeld O W Cao B Wang K Wang H SuH Huang Z Yue D and Zhang T Response of changesin seasonal soil freezethaw state to climate change from 1950to 2010 across china J Geophys Res-Earth Surf 121 1984ndash2000 httpsdoiorg1010022016JF003876 2016

Perroux K M and White I Designs for Disc Per-meameters Soil Sci Soc Am J 52 1205ndash1215httpsdoiorg102136sssaj198803615995005200050001x1988

Pittman F Mohammed A and Cey E Effects of an-tecedent moisture and macroporosity on infiltration and wa-ter flow in frozen soil Hydrol Process 34 795ndash809httpsdoiorg101002hyp13629 2020



Smith M Observations of soil freezing and frost heave at InuvikNorthwest Territories Canada Can J Earth Sci 22 283ndash290httpsdoiorg1010160148-9062(85)90073-7 1985

Spaans E J The soil freezing characteristic Its measurement andsimilarity to the soil moisture characteristic PhD thesis Depart-ment of Soil Science University of Minnesota St Paul 126 pp1994

Spaans E J and Baker J M The soil freezing charac-teristic Its measurement and similarity to the soil mois-ture characteristic Soil Sci Soc Am J 60 13ndash19httpsdoiorg102136sssaj199603615995006000010005x1996

Stadler D Staumlhli M Aeby P and Fluumlhler H Dye trac-ing and image analysis for quantifying water infiltrationinto frozen soils Soil Sci Soc Am J 64 505ndash516httpsdoiorg102136sssaj2000642505x 2000

Stadler D Fluumlhler H and and Jansson P-E Modelling verticaland lateral water flow in frozen and sloped forest soil plots ColdReg Sci Technol 26 181ndash194 httpsdoiorg101016S0165-232X(97)00017-7 1997

Staumlhli M Bayard D Wydler H and Fluumlhler HSnowmelt Infiltration into Alpine Soils Visual-ized by Dye Tracer Technique Arct Antarct AlpRes 36 128ndash135 httpsdoiorg1016571523-0430(2004)036[0128SIIASV]20CO2 2004

Taina I A Heck R J Deen W and Ma E Y Quantifica-tion of freezendashthaw related structure in cultivated topsoils us-ing X-ray computer tomography Can J Soil Sci 93 533ndash553httpsdoiorg104141CJSS2012-044 2013

Tarnawski V R and Wagner B On the prediction of hydraulicconductivity of frozen soils Can Geotech J 33 176ndash180httpsdoiorg101139t96-033 1996

van der Kamp G Hayashi M and Galleacuten D Comparingthe hydrology of grassed and cultivated catchments in thesemi-arid Canadian prairies Hydrol Process 17 559ndash575httpsdoiorg101002hyp1157 2003

Wan X and Yang Z J Pore water freezing characteristic in salinesoils based on pore size distribution Cold Reg Sci Technol173 103030103031-103030103012 2020

Wang D Yates S and Ernst F Determining soil hydraulicproperties using tension infiltrometers time domain reflec-tometry and tensiometers Soil Sci Soc Am J 62 318ndash325httpsdoiorg102136sssaj199803615995006200020004x1998

Wang X Chen R Liu G Yang Y Song Y Liu J Liu ZHan C Liu X Guo S Wang L and Zheng Q Spatial dis-tributions and temporal variations of the near-surface soil freezestate across China under climate change Glob Planet Change172 150ndash158 httpsdoiorg101016jgloplacha2018090162019

Watanabe K and Flury M Capillary bundle model of hydraulicconductivity for frozen soil Water Resour Res 44 W12402httpsdoiorg1010292008WR007012 2008

Watanabe K and Kugisaki Y Effect of macropores on soil freez-ing and thawing with infiltration Hydrol Process 31 270ndash278httpsdoiorg101002hyp10939 2017

Watanabe K and Osada Y Simultaneous measurement of un-frozen water content and hydraulic conductivity of partiallyfrozen soil near 0 C Cold Reg Sci Technol 142 79ndash84httpsdoiorg101016jcoldregions201708002 2017

Watanabe K and Wake T Hydraulic conductivity in frozen un-saturated soil Proceedings of the 9th International Conferenceon Permafrost Fairbanks Alaska USA 29 Junendash3 July 20081927ndash1932 2008

Watanabe K Kito T Dun S Wu J Q Greer R C andFlury M Water infiltration into a frozen soil with simultaneousmelting of the frozen layer Vadose Zone J 12 vzj20110188httpsdoiorg102136vzj20110188 2013

Watson K and Luxmoore R Estimating macroporos-ity in a forest watershed by use of a tension in-filtrometer Soil Sci Soc Am J 50 578ndash582httpsdoiorg102136sssaj198603615995005000030007x1986

Williams P and Burt T Measurement of hydraulic con-ductivity of frozen soils Can Geotech J 11 647ndash650httpsdoiorg101139t74-066 1974

Williams P J and Smith M W The frozen earth fundamentals ofgeocryology Cambridge University Press 1989

Wilson G and Luxmoore R Infiltration macroporos-ity and mesoporosity distributions on two forestedwatersheds Soil Sci Soc Am J 52 329ndash335httpsdoiorg102136sssaj198803615995005200020005x1988

Wooding R Steady infiltration from a shallow cir-cular pond Water Resour Res 4 1259ndash1273httpsdoiorg101029WR004i006p01259 1968

Zhao Y Nishimura T Hill R and Miyazaki T Determininghydraulic conductivity for air-filled porosity in an unsaturatedfrozen soil by the multistep outflow method Vadose Zone J 121ndash10 httpsdoiorg102136vzj20120061 2013


Abstract

Introduction

Materials and methods

Test plan

Measurement of the frozen soil hydraulic conductivity

Measurement of the pore size distribution in frozen soil

Results

Infiltration characteristics of freezingndashthawing soils

Pore distribution characteristics of the freezingndashthawing soil

Discussion

Permeability and hydraulic conductivity of the frozen soil

Effect of freezendashthaw cycles on soil pore distribution

Conclusions

Data availability

Author contributions

Competing interests

Acknowledgements

Financial support

Review statement

References


urated hydraulic conductivity of soils (Gao and Shao 2015)In terms of frozen soils the water infiltration characteristicsand pore size distribution are highly variable and difficult toobserve (Watanabe et al 2013) moreover the water move-ment in freezingndashthawing soils is complicated by the migra-tion of water and heat and the associated water phase change(Jarvis et al 2016 Hayashi 2013) The accurate measure-ment of water movement parameters and soil pore distribu-tion under freezendashthaw conditions is a necessary prerequi-site for the quantitative description of the water movementin frozen soil and the mechanism and degree of influence ofthe temperature on the infiltration rate hydraulic conductiv-ity porosity and other parameters in the different stages offreezendashthaw periods require further research

Currently quantitative studies of the infiltration processin freezingndashthawing soils can be mainly divided into experi-mental and model studies Field experiments have been per-formed less often because under natural conditions infiltra-tion water exhibits a preferential flow into the deep soil andthe alternating freezendashthaw effect forms ice crystals to blockthe flow path through large pores subsequently limiting wa-ter infiltration (Stadler et al 1997) while the melting effectof the infiltration water on ice makes it difficult to reacha steady infiltration state Therefore the current relevantachievements are mainly focused on the infiltration processof snowmelt water (Hayashi et al 2003) and the influence ofpreferential flow (Mohammed et al 2019) Controlled lab-oratory experiments provide new opportunities for the simu-lation of frozen soil infiltration processes and the measure-ment of infiltration parameters Williams and Burt (1974)conducted early direct measurements in the laboratory re-solved the water freezing problem by adding lactose and ap-plied dialysis membranes on both sides of soil columns andthey determined the water conductivity of saturated speci-mens in the horizontal direction (Burt and Williams 1976)Andersland et al (1996) measured the hydraulic conductiv-ity of frozen granular soils at different saturations using aconventional drop permeameter with decane as the permeantand concluded that the hydraulic conductivity was the sameas that of unfrozen soils with water as the infiltration solu-tion McCauley et al (2002) determined and compared thedifferences in hydraulic conductivity permeability and infil-tration rate between frozen and unfrozen soils using dieselmixtures as permeants and their results indicated that the icecontent determines whether soil is sufficiently impermeableZhao et al (2013) quantified the unsaturated hydraulic con-ductivity of frozen soil using antifreeze instead of water andadopted a multistage outflow method under controlled pres-sures in combination with the concept of the pore impedancecoefficient (Jame and Norum 1980) However most of thesestudies did not consider the differences in kinematic viscosityand surface tension between soil water and other solutionswhich often results in an overestimation of hydraulic con-ductivity and homemade devices in the laboratory are ofteninconvenient for generalization in the field Due to the dy-

namic changes in the temperature and moisture phase directmeasurement is difficult and hydraulic conductivity empiri-cal equations and models of frozen soil have been developedFirst the frozen soil hydraulic conductivity was simply con-sidered to follow a power exponential relationship with thetemperature (Nixon 1991 Smith 1985) while others con-sidered the hydraulic conductivity of frozen soil to be equalto that of unfrozen soil at the same water content and as-sumed that the hydraulic conductivity of frozen soil was afunction of the moisture content of unfrozen soil (Lundin1990 Flerchinger and Saxton 1989 Harlan 1973) On thebasis of Campbellrsquos model (Campbell 1985) Tarnawski andWagner (1996) proposed a frozen soil hydraulic conductivitymodel based on the soil particle size distribution and poros-ity Watanabe and Wake (2008) viewed soil pores as cylindri-cal capillaries and suggested that ice formation occurs at thecenter of these capillaries and established a model to describethe movement of thin film water and capillary water in frozensoil based on the theory of capillaries and surface absorption(Watanabe and Flury 2008) The similarity between freezingand soil moisture profiles has been demonstrated (Spaans andBaker 1996 Spaans 1994) and subsequently soil freezingcharacteristic curves have been applied to estimate the unsat-urated hydraulic conductivity of frozen soils (Azmatch et al2012) which has been combined with field tests and inver-sion models to achieve a high accuracy (Cheng et al 2019)

Understanding the distribution characteristics of the soilpore system is essential for the evaluation of the water andheat movement processes in soil Soil macroporosity hasbeen shown to impose a major impact on water cycle pro-cesses such as infiltration nutrient movement and surfacerunoff (Demand et al 2019 Jarvis 2007) Macroporosityis widespread in a variety of soils and produces preferen-tial flow in both frozen and unfrozen soils (Mohammed etal 2018 Beven and Germann 2013) and the pre-freezemoisture conditions affect the amount and state of ice in themacropores of frozen soils resulting in notable variability inthe infiltration capacity of thawed soils (Hayashi et al 2003Granger et al 1984) Field experiments on frozen soil havealso demonstrated that macropores accelerate the infiltrationrate (Staumlhli et al 2004 van der Kamp et al 2003) the num-ber and size of macropores affect the freezing and infiltrationcapacity of soil layers to different extents and low tempera-tures cause infiltration water to refreeze inside macropores(Watanabe and Kugisaki 2017 Stadler et al 2000) Re-search on frozen soil macroporosity has largely focused onthe qualitative analysis of its impact on the soil structure andinfiltration capacity With the development of experimen-tal techniques certain new methods and techniques such ascomputed tomography (CT) and X-ray scanning have beenapplied to measure the number and distribution of macrop-ores (Taina et al 2013 Bodhinayake et al 2004 Greverset al 1989) but the lack of sampling techniques targetingfrozen soil still restricts related research













21 Test plan













Q= πR2K

[1+

4πRα

] (2)






[1+

4πRα

] (3)


[1+

4πRα

] (4)


α =ln

[Q(h2)Q(h1)

]h2minush1

(5)







Water 15 09991 114 7356 0



ψ =minusLρwT

27315 (6)





(7)






N = θmπr2 (9)



3 Results
































15 (AM) 02629 74658 241times10minus9




15 (AM) 02934 93757 303times10minus9




15 (AM) 02182 51283 166times10minus9
















4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References












21 Test plan













Q= πR2K

[1+

4πRα

] (2)






[1+

4πRα

] (3)


[1+

4πRα

] (4)


α =ln

[Q(h2)Q(h1)

]h2minush1

(5)







Water 15 09991 114 7356 0



ψ =minusLρwT

27315 (6)





(7)






N = θmπr2 (9)



3 Results
































15 (AM) 02629 74658 241times10minus9




15 (AM) 02934 93757 303times10minus9




15 (AM) 02182 51283 166times10minus9
















4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References








Q= πR2K

[1+

4πRα

] (2)






[1+

4πRα

] (3)


[1+

4πRα

] (4)


α =ln

[Q(h2)Q(h1)

]h2minush1

(5)







Water 15 09991 114 7356 0



ψ =minusLρwT

27315 (6)





(7)






N = θmπr2 (9)



3 Results
































15 (AM) 02629 74658 241times10minus9




15 (AM) 02934 93757 303times10minus9




15 (AM) 02182 51283 166times10minus9
















4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References




Water 15 09991 114 7356 0



ψ =minusLρwT

27315 (6)





(7)






N = θmπr2 (9)



3 Results
































15 (AM) 02629 74658 241times10minus9




15 (AM) 02934 93757 303times10minus9




15 (AM) 02182 51283 166times10minus9
















4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References




























15 (AM) 02629 74658 241times10minus9




15 (AM) 02934 93757 303times10minus9




15 (AM) 02182 51283 166times10minus9
















4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References










4 Discussion













5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References







5 Conclusions











References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References









References
















































































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References






























































Abstract

Introduction


Test plan



Results



Discussion



Conclusions

Data availability


Competing interests

Acknowledgements

Financial support

Review statement

References

soil inﬁltration characteristics and pore distribution

Documents