Soil micro-penetration resistance as an index of its infiltration processes during rainfall

Cho-Sheng Tng,Xue-Peng Gong,Zhengto Shen,Qing Cheng,Hilry Inyng,b,Cho Lv,Bin Shi

a School of Earth Sciences and Engineering,Nanjing University,163 Xianlin Avenue,Nanjing,210023,China

b Global Institute for Sustainable Development,Charlotte,NC,USA

Keywords:Micro-penetration test Soil wetting Rainfall infiltration Water content Hydro-mechanical behavior Microstructure

A B S T R A C T Rainfall infiltration is one of the most important driving factors of geological hazards,ecological environment problems,and engineering accidents.Understanding the principle of soil wetting during rainfall infiltration and its influence on soil mechanical properties is crucial for preventing geological hazards.In this study,micro-penetration tests coupled with moisture monitoring were performed to investigate the infiltration process during wetting through the measured change in mechanical characteristics.Results show that penetration resistance increases in the deep layer gradually.With increasing infiltration time,the wetting front keeps moving downward,and its range becomes wider.A slight increase of the penetration resistance in the shallow layer(d≤17.5 mm)is observed.However,the penetration resistance in the middle layer(22.5 mm≤d≤32.5 mm)decreases firstly before a slight increase.In the deep layer(d≥37.5 mm),the penetration resistance decreases continuously during infiltration.Based on the measured water content profile during infiltration,it is found that the evolution of soil mechanical characteristics is fully responsible by the infiltration-induced re-distribution of water content along depth.Generally,the penetration resistance decreases exponentially with increasing water content in the soil.When the water content is low,wetting can weaken soil strength significantly,whereas this effect diminishes when the moisture surpasses a certain threshold.The results highlight that the penetration curves and water content profile show close inter-dependency and consistency,which verifies the feasibility of using micro-penetration to investigate rainfall infiltration and wetting process in surface soil layer or laboratory small-scale soil samples.This method enables fast,versatile and high-resolution measurements of infiltration process and moisture distribution in soil.

1.Introduction

Water is an important component of natural soils.Changes in water content may significantly alter the hydro-mechanical behavior of a soil.These effects mainly include:(1)volumetric change such as swelling and shrinkage,and(2)variation of mechanical behavior such as strength and compressibility.These effects typically coupled to each other,giving rise to various ecological,environmental,geotechnical and geological problems(Tang et al.,2021).Generally,rainfall infiltration induced change in soil water content is responsible for most of the geo-environmental issues and geological hazards that are encountered in field projects.

Rainfall induced soil erosion is one of the common ecological environment problems in the earth(Cheng et al.,2021a;Liu et al.,2021;Shi et al.,2021).China is one of the countries with the most serious soil erosion problem,especially in Loess Plateau region(Wang et al.,2021).The serious soil erosion poses a threat to human life and significantly restricts the sustainable development of economy(Liu et al.,2008).Landslides are typical geological disasters caused by rainfall.On 25 October 2011,heavy rainfall in northwestern Italy caused thousands of shallow landslides,widespread erosive and depositional processes,several floods,13 casualties,the evacuation of about 1200 people,the interruption of both the A12 Highway and the Genova-La Spezia Railway,the closure of 43%of provincial roads,and the destruction of many bridges.On 16 and 17 June 2013,high-intensity rainfall(＞400 mm)in different parts of Garhwal Region,Uttarakhand,India,caused devastating flash floods and triggered widespread landslides.Heavy losses were recorded as infrastructure damage,agricultural losses,deaths and widespread destruction of natural resources(Mehta et al.,2016).On 7 August 2010,two debris flows induced by rainfall near the county of Zhouqu,Gansu,northwestern China,took the lives of about 1765 people living in the densely urbanized region(Tang et al.,2011).Many other rainfall induced landslide and debris flow events have been reported around the world each year(Wu et al.,2017).Increase in water content may decrease the shear strength of subgrade soil and its elastic modulus,resulting in earth structure damage including rutting,cracking,uneven settlement,and pitting(Ji et al.,2020).Slopes of foundation ditches are also affected by rainfall.Most of the foundation ditch deformation and collapse accidents recorded in practice are related to rainfall.Therefore,an in-depth understanding of the process of rainfall infiltration and soil wetting is of great significance for the understanding and mitigation of geological hazards and protection of geo-environment.

In the past few decades,a number of theoretical modeling studies on rainfall infiltration and soil wetting have been carried out(Cheng et al.,2021b).The process of rainfall infiltration refers to the migration and redistribution of rainwater in a soil profile driven by gravity and capillary force.The earliest model that described infiltration is the Green-Ampt model which is a transient“pistonflow’’model(Green and Ampt,1911).In the Green-Ampt approach,the soil moisture profile is assumed to be a step function with a water content discontinuity above which the soil is saturated(at residual air saturation),and below which the initial moisture content is maintained.However,assumptions in the Green-Ampt model are often invalid in practice and the suction at the front is difficult to accurately obtain.These constraints may limit the application of this model(Mein and Larson,1973).A probably more rigorous analytical model is the famous Richard’s equation(Richards,1931)which is a three-dimensional(3D)water flow partial differential equation that combines Darcy’s law with the law of conservation of mass.It is applicable to isotropic and uniformly unsaturated soil(Richards,1931).However,Richard’s equation is nonlinear and difficult to solve in either numerical or analytical methods.Taking the first two terms of its expansion as the analytical solution,Philip(1957)further modified the Richard’s model into thePhilip Equation,with a widely adopted analytical solution.With less restriction,it can be applied with the assumption of a vertically homogeneous soil,constant initial moisture content,and saturated soil surface with immediate ponding.

Although rainfall infiltration models play an important role in engineering practice,current research on infiltration model is more focused on the mechanism.The structures of most infiltration models are complex,the parameters are difficult to measure,and they are inconvenient for practical applications.Scholars have proposed a range of semi-theoretical,semi-empirical or purely empirical infiltration formulae(Horton,1941;Smith,1972).These models can only describe the variation of infiltration rate with time,and most of them pertain to single point infiltration,which is not adequate for analyzing complicated natural conditions.

In addition to model study,there are generally two experimental approaches to understand the infiltration process.One approach is to directly measure the water content profile during infiltration.In most cases,field engineers hope to monitor water content directly through the use of a certain apparatus,which operates with principles such as frequency domain reflectometry(FDR)and time domain reflectometry(TDR).Equipment such as tensiometer and psychrometers are also applied(Song et al.,2016;An et al.,2018).Due to the relatively large size of probes,the model tests are always designed at large-scale and are usually timeconsuming inevitably(Smits et al.,2011;An et al.,2017).In addition,TDR is highly susceptible to the salt content of soil(Wyseure et al.,1997),tensiometers cannot be applied in extreme drought conditions because the upper limit of its measurement range is relatively low(Toll et al.,2013),and these apparatuses are usually costly and difficult to install.More importantly,these methods are point tests such that the real spatial distribution of water content cannot be clearly reflected in surface soil layer(i.e.less than 10 cm depth),which is sensitive to climate change(Tang et al.,2016).Consequently,the hydro-mechanical behavior and especially the rainfall infiltration process of surface soil layer cannot be properly evaluated as it is exposed to changing climate.

Another approach is to investigate the infiltration process by examining the changes of soil mechanical behavior with increasing water content.For instance,Wei et al.(2018)performed a series of unconsolidated undrained triaxial compression tests using soil samples collected around the Three Gorges Reservoir Area in China.They found that for a given dry density,the cohesion increased first and then decreased as moisture content increased.This followed a quadratic curve with a turning point at a water content of about 11%.Pham et al.(2018)considered water content as one of the factors that can be addressed by machine learning methods.Kong et al.(2010)investigated the unsaturated expansion of soil and found that soil shear strength decreases with increasing water content,suggesting that suction reductions contribute to the decrease in shear strength.Guan et al.(2010)evaluated 13 predictive equations for unsaturated soil shear strength related to moisture content,using data from 11 published papers.The results show that the shear strength equations proposed by Vanapalli et al.(1996)and Garven and Vanapalli,(2006)agreed with more than half of the experimental results.Although the relationship between soil mechanical parameters and water content has been extensively investigated,the spatial and temporal dependence of mechanical parameters alone in a soil profile during infiltration still remains unclear due to the limitations of measurement techniques.As a result,the infiltration process inside a soil profile especially inside a small-scale soil sample cannot be accurately determined.

In order to improve the accuracy and precision of infiltration investigation for small-scale soil sample,a testing method for rainfall infiltration process based on a micro-penetrometer is proposed in this work,based on actual measurements of soil mechanical characteristics.It can directly reflect the dynamic process of rainfall infiltration and soil wetting,through rapid,simple and refined measurement of penetration resistance of a soil profile.The relationship between rainfall infiltration and changes in the measured soil water content can be established.This work aims to provide an alternative method for investigating soil wetting and rainfall infiltration under laboratory conditions or in field shallow soils.

2.Material and methods

2.1.Soil

The Xiashu clayey soil was collected from the Nanjing area,China.This soil widely distributes in the middle and lower reachesof the Yangtze River,and is an important foundation soil.The physical properties of the soil are shown in Table 1.The particle size distribution analysis shows that the soil is composed of sand(2%),silt(76%),and clay(22%).According to the Unified Soil Classification System(USCS)(ASTM D2487-11,2011),it is a low plasticity clay soil(CL).Xiashu clayey soil comprises quartz,feldspar and clay minerals.Illite is the main clay mineral followed by montmorillonite and kaolinite.

Table 1Basic physical properties of the Xiashu soil.

2.2.Apparatus

In geotechnical investigations,standard penetration test(SPT)has been widely adopted.Previous research has also proven the validity of using penetration resistance as an index of the mechanical properties of soils(Elbanna and Witney,1987;Low et al.,2010;Shin and Kim,2011).In this study,the micro-penetrometer,SMP-1,is developed and employed as a penetration test equipment during simulated rainfall infiltration process.The schematic diagram of SMP-1 is shown in Fig.1.It mainly consists of three parts:(1)load/displacement controlling and measuring system,(2)platen positioning system,and(3)data collecting system.

The load/displacement controlling and measuring system consists of a control box,a control panel,a displacement transducer and a load transducer.The control box is equipped with a motor that can apply constant displacement rate during the penetration test.The displacement rate can be adjusted through the control panel from 0 to 10 mm/min.The displacement transducer and load transducer are connected to the data collecting system:the former records the displacement of the platen,with a capacity of 50 mm to an accuracy of 0.01 mm;and the latter records the penetration resistance during penetrating,with a capacity of 100 N and an accuracy of 0.01 N.

The platen positioning system includes a platen and a position adjuster(in Fig.1).Soil specimen is settled on the platen and uplifted at a constant rate.The position adjuster can move the platen and sample,slowly and stably,by rotating the two sets of perpendicular nuts,up and down or left and right.By this means,it can be ensured that the probe penetrates at the specified position.

The displacement and resistance data obtained by transducers are then imported into the computer through a dedicated data logger,and the data can be analyzed using the customized software.

Fig.2 is a sketch of the mini-probe used in this investigation.The mini-probe is 60 mm long.The head of the mini-probe is conical.In order to reduce side friction,the diameter of the rod part is slightly smaller than that of the head part.The penetration resistance measured is taken to be equal to the end resistance of the probe(Gu et al.,2014;Wang et al.,2016a,b).

Fig.1.Schematic diagram of the micro-penetrometer(SMP-1).

Fig.2.Schematic diagram of the mini-probe.

2.3.Sample preparation

The collected Xiashu clayey soil was air-dried and crushed,and passed through a 2 mm sieve.A spray bottle was used to gently spray water on the surface of the soil with continuous stirring to ensure water uniformity.After reaching the designed initial water content of 15.8%(±0.2%),the soil was sealed in a plastic bag and placed in a moisture chamber for 24 h to ensure a homogeneous distribution of water.

The moist soil was then divided into two groups:one group was prepared for micro-penetration test under a simulated rainfall infiltration process(Group A),and the other group was prepared for profile sampling and water content test under the same simulated water infiltration process(Group B).The preparation procedure for the two groups was kept the same.The mold used in the test is shown in Fig.3.It has an inner diameter of 99.2 mm and a height of 105 mm.The soil was compacted into the container statically in 3 layers to a final height of 60 mm and a dry density of 1.6 g/cm3.There was also a circular porous positioning plate that matched the inner diameter of the mold.All the positioning holes were sealed with tape before testing to prevent evaporation.The corresponding positioning holes were opened upon testing,and the remaining holes were kept closed.The mold used in the test of group B has the same size with the mold in Group A.The same compaction levels were employed for the two groups:a final height of 60 mm and a dry density of 1.6 g/cm3.

2.4.Test procedures

The simulated rainfall infiltration tests for both Groups A and B were performed under the following procedures.

In Group A,a piece of full-sized and pre-saturated filter paper was placed on sample surface to ensure a homogenous wetting and aid the control of infiltration volume.The 20 mL of water was evenly sprinkled on the surface of the filter paper,and a clock was set to record the start ast0=0.The filter paper was taken out when all sprinkled water has infiltrated into the soil and the porous positioning plate was placed on the sample surface(Fig.3).It was observed that the sprinkled water was infiltrated in 1 min.The soil sample was then placed on the lifting platform of the SMP-1 and the penetration test started on a selected penetration position.The penetration rate was set at 10 mm/min and the real-time resistance and displacement data were recorded.The time when the first data were recorded was defined asti.When the penetration depth reached the predetermined 45 mm,the test was terminated.The position that is at least 10 times of the probe diameter away from the last penetration position was selected for the next penetration test.In this study,the sample was subjected to the infiltration for 2 h and the above penetration procedures were repeated.A total of nine penetrations were performed and started at timet1=4′47′′,t2=17′20′′,t3=29′28′′,t4=42′58′′,t5=56′10′′,t6=69′30′′,t7=82′50′′,t8=96′10′′andt9=117′23′′,respectively,as the penetration sequence illustrated in Fig.3.

Fig.3.Schematic diagram of the mold for sample preparation.

In Group B,these samples were used to determine the water content profile during the infiltration process.A total of 12 parallel samples were prepared.The preset infiltration times were 3 min,6 min,9 min,14 min,19 min,26 min,35 min,45 min,60 min,75 min,95 min and 120 min.The simulated rainfall infiltration process of Group B was strictly consistent with that of Group A.When the preset infiltration time was reached,the three-lobed mold was opened and the soils at the vertical profile were quickly sampled.The vertical soil profile was divided into 12 layers,i.e.soil samples were taken at every 5 mm depth.Then,these layer samples were oven-dried at 105°C for 24 h and the corresponding average water content at depths of 2.5 mm,7.5 mm,12.5 mm,17.5 mm,22.5 mm,27.5 mm,32.5 mm,37.5 mm,42.5 mm,47.5 mm,52.5 mm and 57.5 mm was measured respectively.

3.Results and analysis

3.1.Group A:the penetration test

The relationships between the penetration resistanceRpand depthd(i.e.the penetration curves)at different start times are shown in Fig.4.The pattern of each penetration curve is basically similar:as the depth increases,the penetration resistance increases monotonically and finally reaches stabilization.The following phenomenon occurs in most curves:after reaching a certain critical depth,penetration resistance increases dramatically.The transition zone where the penetration resistance increases dramatically is actually related to the wetting front,indicating that the strength of soil in this zone is weakened by water infiltration.Below this zone,the mechanical behavior of soil is not significantly influenced by infiltration and remains relatively stable penetration resistance.It was observed that the shorter the infiltration time,the more obvious the wetting front and the earlier the wetting front appears.For example,the wetting front of the first penetration test(t1=4′47′′)is located at the depths of 25-33 mm.As infiltration continues,the wetting front moves downward and the slope of the curve in this zone reduces gradually.

Fig.4.The penetration curves at different start times.

Fig.5 shows the variation of penetration resistance with infiltration time at different depths.With increasing in infiltration time,a slight increase of the penetration resistance in the shallow layer(i.e.d≤17.5 mm)is observed.However,the penetration resistance in the middle layer(i.e.22.5 mm≤d≤32.5 mm)decreases firstly before a slight increase.In the deep layer(i.e.d≥37.5 mm),the penetration resistance decreases continuously during infiltration.

3.2.Group B:the water content profile measurement

The water content profiles are shown in Fig.6.It can be seen from the figure that:(1)With the increase of depth,the water content decreases monotonously;(2)With the increase of infiltration time,the wetting front gradually moves downward,and the curve becomes smoother;and(3)The water content near the surface gradually decreases,while the water content in the deep gradually increases with increasing infiltration time.This suggests the transport of the sprinkled water from upper to deeper layer gradually.

These data points are plotted as a set of water content-time(ω-t)curves at different depths(Fig.7).The water content gradually decreases in the shallow layers(i.e.d≤17.5 mm),the water content increases first and then decreases in the middle layers(i.e.22.5 mm≤d≤32.5 mm),and the water content generally increases slightly throughout the test in the deep layers(i.e.d≥37.5 mm).

Combining the results of Groups A and B,it can be observed that the penetration test in the infiltration process can reflect the migration process of water in the soil profile.In general,the higher the water content,the smaller the penetration resistance.At the initial stage of infiltration,water mainly accumulates in the shallow layer and the penetration resistance in this area is low.The penetration resistance in the deep layer is relatively large as water has not entered into this area.As infiltration proceeds,the wetting front moves into the deeper layers.During this process,the surface water content decreases,and the deep water content increases.At the same time,the corresponding penetration resistance increases in the surface layer,and the penetration resistance in the deep layer significantly decreases with infiltration time.

Fig.5.The evolution of penetration resistance at different depths during infiltration process.

Fig.6.The water content profile at different infiltration times.

Fig.7.The evolution of water content at different depths during infiltration process.

3.3.Comparison of the results between groups A and B

As shown in Fig.8,the penetration resistanceRpand water content ω obtained from Groups A and B are plotted in a twodimensional(2D)space ofd-t.MATLAB’s cftool toolbox was used to perform polynomial fitting,and two binary functions,Rp(d,t)and ω(d,t),can be obtained:

Thus,Rp(d,t)and ω(d,t)have the same domain(d∈(0,60],t∈(0,120]).A series ofdvalues with tolerances of 0.3 in the interval[0,42)and a series oftvalues with tolerances of 1 in the interval[1,120]were selected to form a 2D array of 139×120.Then the above fitting functionsRp(d,t)and ω(d,t)were used to predictRpand ω on each point.AllRpand ω data can be drawn on a scatter plot,which is shown in Fig.9.These 16,680 data points were fitted in the form of linear,power,polynomial,exponential and logarithm functions.The exponential function best described the data and can be written as

Fig.8.3D diagram of penetration resistance function Rp(d,t)and water content function ω(d,t).

Fig.9.The relationship between soil penetration resistance Rp and water content ω.

According to the above fitting equation,the relationship between theRpand ω is therefore established.The real water content can be predicted using penetration resistance data,so as to indirectly study the pattern of water migration during infiltration.Fig.10 shows the prediction of water content cloud at each infiltration time based on the data of nine penetration tests(Fig.4).As shown in Fig.10,the water content of the soil gradually decreases in the shallow layer,increases first and then decreases slightly in the middle layer,and significantly increases in the deep layer,which is consistent with the measured moisture content(Figs.6 and 7).In addition,Huang et al.(2019)investigated the shear strength of the tested soil at different water contents but with the same dry density of 1.6 g/cm3.The changes of obtained cohesion and internal friction angle with water content are plotted in Fig.11,for comparing the penetration resistance.As expected,the shear strength parameters generally decrease with increasing water content.

Based on these results,it can be concluded that the soil penetration resistance exponentially decreases with the increase of water content.If the penetration curve is obtained,the spatial distribution and variation of the internal water content in soil during the infiltration process can be predicted using Eq.(3).However,due to the variability in soil type,sample preparation method,dry density,etc.,the coefficients in Eq.(3)may be different among different experiments.

4.Discussion

In previous studies,penetration resistance was typically used as a characterization parameter for the“structural strength”of soil(Low,1979;Gregory and O’Melia,1989;Mitchell and Soga,2005).In other words,the penetration resistance is the resistance that the probe is subjected to destruction of the soil structural unit.For compacted soil,the unit penetrated by the probe is mainly the aggregate,and the penetration resistance mainly reflects the cohesion within the aggregate or the connection strength among soil particles.In addition,the measured resistance is also derived from the resistance of the surrounding soil to the volumetric compression caused by the penetration of the probe,i.e.the compressive modulus.

According to the double porosity structure of a compacted soil(Sharma,1998;Romero et al.,1999),there are two types of pores in the soil:the larger pores are the inter-aggregate pores with diameters of several to tens of micrometers,and the smaller pores are the intra-aggregate pores with diameters of a few nanometers.At the initial stage of infiltration,the water content is low.Therefore,almost all water is present in the intra-aggregate pores(Lloret et al.,2003;Sivakumar et al.,2006).During the wetting process,the water preferentially enters into the intra-aggregate pores due to relatively high retention capacity.It is of note that when the intraaggregate pores are saturated,the redundant water can migrate into the inter-aggregate pores(Lloret et al.,2003;Sivakumar et al.,2006;Nowamooz and Masrouri,2010).

Soil wetting is not only a simple water migration process.It also affects the soil engineering properties:the increase in water content weakens the forces among soil particles.Generally,the cations on the surface of the clay sheets have a tendency to hydrate and adsorb water molecules during the wetting process.This increases the distance between clay sheets due to repulsion and weakens the van der Waals attractive force.This further decreases the cohesion between clay sheets and the aggregates will disperse and exfoliate into smaller soil particles.Therefore,the influences of these effects on the penetration tests are:(1)weakening of the aggregates after wetting such that the probe penetrates the soil easily;and(2)more susceptibility of the proximal aggregates of the soil to deformation and squeezing of the surrounding space when the probe penetrates the soil.

However,because of the limited number of cations and their low coordination number,the hydration of cations on the surface of clay sheets is limited.When the water content reaches this limit,the excess water does not separate the clay sheets from each other further or reduce the cohesion strength,but only migrates into the inter-aggregate pores.Therefore,the subsequent wetting may not weaken the soil strength further.

In summary,the wetting process can be divided into two stages.The first stage is the cation hydration stage when the water content is low,and the surface clay particles are not hydrated completely.The water molecules are mainly bound by the cations.The strength weakening effect on the soil is significant.The second stage is the inter-aggregate pore filling stage.After the cation hydration stage,the excess water only fills the inter-aggregate pores but has no effect on the cohesion within aggregates;therefore,the weakening effect on the soil strength is not significant.

In this study,the surface layer of each sample(e.g.at 2.5 mm)is saturated almost immediately after the initial addition of 5 mL of water.The ions between the clay sheets are sufficiently hydrated,and the inter-aggregate pores are also filled with water.Therefore,minimal penetration resistance was observed.During the infiltration process,the water in the inter-aggregate pores migrates first.As a result,during the 2-h test,the increase in the penetration resistance of the surface layer is very low(by about 0.8 N).In contrast,in the deep layer(e.g.37.5 mm),although water content just slightly increases(about 1%),the penetration resistance value is significantly reduced(by about 4 N).This is due to the migration of water from the upper layer.Water is first adsorbed by the cations on the surface of the clay sheet,resulting in an increase in the spacing of clay sheets and in the thickness of the electric double layer finally reducing the cohesive force and structural strength of aggregates.The penetration curve clearly confirms the occurrence of this process.

Although the exponential function is established to connect discrete points ofRpand ω(Eq.(3)),the distribution of data points is more similar to a piece-wise function(Fig.9).When ω＜16%,the slope ofRp(ω)is about-0.06 N.However,when ω＞16%,the slope is only about-0.01 N.Essentially,when the water content is lower than 16%,wetting has a significant weakening effect on the mechanical properties of the soil;and when the water content is higher than 16%(the degree of saturation is 73.6%),the weakening effect of the wetting is not significant.The critical water content may be related to the saturated water content of the intraaggregate pores in the soil sample as stated by Romero et al.(1999),who performed a series of experiments to study the fabric of Boom Clay.The mercury intrusion porosimetry(MIP)and inflow/outflow results indicated that the intra-aggregate water content is about 13%(the intra-aggregate pore is 54%of the total porosity)for the low-porosity fabric at a dry unit weight of 16.7 kN/m3which is close to the present tests.

Fig.10.Water content cloud chart at each infiltration time predicted from micro-penetration.

Fig.11.Changes of penetration resistance,cohesion and internal friction angle with water content.

5.Conclusions

Micro-penetration test was proposed to investigate the hydromechanical response of small-scale soil sample during simulated rainfall condition.The evolution of penetration resistance along depth was characterized by penetration curve at different rainfall infiltration times.The actual water content profile during infiltration process was measured and the relationship with penetration resistance was analyzed.The following conclusions can be drawn:

(1)During the short-term rainfall infiltration process,the soil water content at the top layer peaks at the initial stage.As the infiltration progress goes on,the water gradually migrates downward:the water content at the top layer decreases continuously;the water content in the middle layer increases first,and then decreases slowly;and the water content in the deep layer continues to increase slowly.The wetting front continues to gradually move downwards during the process.

(2)In the penetration test during wetting process,the penetration resistance is lower in the top layer,and increases in the deep layer gradually.The wetting front keeps moving downward,and its range also becomes wider with increasing infiltration time.

(3)For a given soil type,dry density and preparation method,the penetration resistanceRpis inversely proportional to the water content ω.When the water content is low,wetting has a significant weakening effect on the soil strength.However,when the water content exceeds a certain value,wetting process hardly weakens the soil.This can be explained by the two-stage wetting process:(i)at the cation hydration stage,water molecules are mainly adsorbed through the hydration of cations on the surface of a clay sheet,and the increase in water sharply reduces the aggregate structural strength;and(ii)at the inter-aggregate pore filling stage,the increase of water only fills the large pores among the aggregates and has little effect on the cohesive force or structural strength of the soil aggregates.It merely reduces the penetration resistance.

(4)The feasibility of using micro-penetration test to investigate soil infiltration by rainwater and wetting process was verified by comparative experiments and relevant mathematical analysis.In the process of simulated rainfall infiltration,the penetration curve and the water content profile have significant correlation,and the patterns reflected are basically consistent.In addition,due to the characteristics of rapidity,small disturbance and refinement of micro-penetration test,it has enormous advantages when used to investigate principles that govern soil wetting in the process of rainfall infiltration,especially in laboratory small-scale simulation tests or field surface soil tests.It can accurately measure the spatial and temporal distribution characteristics of the water field in shallow depth with compensation for the shortcomings of the traditional soil moisture monitoring methods.

Declaration of competing interest

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

Acknowledgments

This work was supported by the National Key Research and Development Program of China(Grant No.2020YFC1808101),National Natural Science Foundation of China(Grant No.41925012),Natural Science Foundation of Jiangsu Province(Grant No.BK20211087),and the Fundamental Research Funds for the Central Universities.