Water retention of repellent and subcritical repellent soils: New insights from model and experimental investigations

Journal of Hydrology 380 (2010) 104–111

Contents lists available at ScienceDirect

Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol

Water retention of repellent and subcritical repellent soils: New insights from model and experimental investigations H. Czachor a,*, S.H. Doerr b, L. Lichner c a

Institute of Agrophysics, PAS, ul. Doswiadczalna 4, 20-290 Lublin, Poland School of the Environment and Society, Swansea University, Singleton Park, Swansea SA2 8PP, UK c Institute of Hydrology, SAS, Racianska 75, 831-02 Bratislava 3, Slovakia b

a r t i c l e

i n f o

Article history: Received 7 March 2009 Received in revised form 18 October 2009 Accepted 26 October 2009

This manuscript was handled by P. Baveye, Editor-in-Chief, with the assistance of Renduo Zhang, Associate Editor Keywords: Soil Water retention Wetting angle Pore modeling Repellency

s u m m a r y Soil organic matter can modify the surface properties of the soil mineral phase by changing the surface tension of the mineral surfaces. This modifies the soil’s solid-water contact angle, which in turn would be expected to affect its water retention curve (SWRC). Here we model the impact of differences in the soil pore-water contact angle on capillarity in non-cylindrical pores by accounting for their complex pore geometry. Key outcomes from the model include that (i) available methods for measuring the Young’s wetting angle on soil samples are insufficient in representing the wetting angle in the soil pore space, (ii) the wetting branch of water retention curves is strongly affected by the soil pore-water contact angle, as manifest in the wetting behavior of water repellent soils, (iii) effects for the drying branch are minimal, indicating that both wettable and water repellent soils should behave similarly, and (vi) water retention is a feature not of only wettable soils, but also soils that are in a water repellent state. These results are tested experimentally by determining drying and wetting branches for (a) ‘model soil’ (quartz sands with four hydrophobization levels) and (b) five field soil samples with contrasting wettability, which were used with and without the removal of the soil organic matter. The experimental results support the theoretical predictions and indicate that small changes in wetting angle can cause switches between wettable and water repellent soil behavior. This may explain the common observation that relatively small changes in soil water content can cause substantial changes in soil wettability. Ó 2009 Elsevier B.V. All rights reserved.

Introduction The water retention curve is a fundamental soil characteristic, which determines important soil water storage and plant performance parameters including water content at field capacity, wilting point and plant available water content (i.e. water capacity) (Buckman and Brady, 1969; Givia et al., 2004). Soil water retention is a result of the solid phase wetting by the soil solution and is closely related to the capillary phenomena in soils. In the simplest approximation the soil pores are presented as a set of cylindrical micro tubes of different radii, known in the literature as the capillary bundle model (Kozeny, 1927; Kezdi, 1974; Mualem, 1976; Vervoort and Cattle, 2003). This approach is widely used in soil physics even though it is frequently acknowledged as ‘‘a vast oversimplification” (Letey et al., 2000). Given that the soil matrix is typically granular, it is indeed difficult to image a cylindrical pore structure to occur. Analysis of soil thin sections reveals that soil pores are not straight, but tortuous and not cylindrical, of variable cross section and interconnected (Imeson and Jongerius, 1974; * Corresponding author. Tel.: +48 81 7445061; fax: +48 81 7445067. E-mail address: [email protected] (H. Czachor). 0022-1694/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2009.10.027

VandenBygaart et al., 1999; Zhang et al., 2005; Vera et al., 2007; Adesodun et al., 2008). From a topological point of view, the soil pore network is typically very complex due to the heterogeneity arising from large differences in soil particle size and shape, and the disordered character of soil particle packing. In addition to pore geometry, the wettability (i.e. surface tension) of the soil solid phase is another fundamental feature influencing soil water retention. However, determining the soil solid phase surface tension of soils is problematic. The direct methods of wetting angle measurement developed for flat surfaces (e.g. sessile drop or Wilhelmy plate; Buczko et al., 2006; Bachmann et al., 2000) are problematic due to the lack of flat surfaces. Moreover many soils are wettable so that a sessile drop spreads and infiltrates shortly after its application. Indirect methods of wettability determination (e.g. capillary rise or repellency index) depend on the wetting front kinetics in a soil column and, using the Washburn theory (Washburn, 1921; Waniek et al., 2000), provide a value for an apparent wetting angle, which in turn is dependent on pore shape (Czachor, 2007). Other soil wettability indicators like the widely used water drop penetration time (WDPT) or the molarity of an ethanol droplet (MED) tests (Osborn et al., 1964; Letey, 1969) are indices of soil water repellency (SWR), which will only

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111

105

discriminate between soils exhibiting a certain degree of water repellency (de Jonge et al., 1999; Doerr et al., 2000, 2002). It is often stated that a water drop applied in these tests infiltrates if the wetting angle falls below 90° (Carillo et al., 1999). This would be true only in the case of cylindrical pores, but does not really apply to soils which have wavy pores with variable cross sections or soil surfaces with complex topography. For example, Shirtcliffe et al., 2006 have recently shown that infiltration into granular material such as soils may not occur for wetting angles P50°. Generally there is no fully accepted method of wetting angle measurement in granular porous materials such as soils. The aim of this paper is to (i) model the impact of non-cylindricity of the pores, and of the wetting angle on water retention in soils and (ii) test the model predictions experimentally using artificial ‘model soil’ material and arable soils of varying wettability. Capillarity in non-cylindrical pores As outlined above, the capillary bundle model (Kutilek and Nielsen, 1994) of soil pores is not able to explain some very fundamental properties of soils such as the hysteresis of the moisture retention curve. This phenomenon is the result of non-cylindricity of soil pores. As an approximation, the geometrical radius rg(x) of a pore can be described by the following equation

r g ðxÞ ¼ r1 þ r 2  sinðpx=hÞ

ð1Þ

where r1 is the mean geometrical radius of a wavy pore [L], r2 the amplitude of the pore geometrical radius [L] and 2h the length period [L]. In deriving Eq. (1), the soil pore space was considered as a set of interconnected necks and voids between particles of different size, with pore size being related to grain size distribution (GSD). In such a periodic pore, the geometrical radius changes from r1  r2 to r1 + r2 at a distance h (Czachor, 2006), with all the above parameters assumed to be related to bulk density and GSD of investigated material. Inside the pore, a liquid at a wetting angle h creates a meniscus described by a radius rm(x)

r m ðx; hÞ ¼ r g = cosðaðxÞ þ hÞ

ð2Þ

where a(x) is the wall slope angle in relation to the movement direction [°]. Supposing axis symmetry of the pore, one can derive a formula for the capillary pressure Pm(x), which is related to the meniscus radius via Laplace’s relation

Pm ðx; hÞ ¼ 2r cosðaðxÞ þ hÞ=r g ðxÞ

ð3Þ 1 2

where Pm (x) is the capillary pressure [M L T ] and r the liquid surface tension [M T2]. Substituting Eq. (2) into Eq. (3) and taking into account tga = drg/dx one gets

Pm ðx; hÞ ¼

2r cos h  pd2 sin h cosðpx=hÞ  r 1 ½1 þ d1 sinðpx=hÞsqrt½1 þ ðpd2 cosðpx=hÞÞ2 

ð4Þ

where d1 = r2/r1 and d2 = r2/h (Czachor, 2006). The capillary pressure Pm vs. meniscus position x for two such wavy horizontal pores are presented in Figs. 1 and 2 and concern an arbitrary chosen set of parameters d1, d2, h to show the impact of pore shape on capillary pressure. Calculations were made for the same set of wetting angles to show the difference between both pores. Fig. 2 corresponds to a more ‘‘wavy” pore, which means the maximum slope is greater compared to the pore of Fig. 1. This is schematically depicted in the inset graphs 1 and 2. The unit for the applied pressure (2r/r1) is a relative one, which is convenient for general considerations. It can be transformed into an absolute one where r and r1 values are known. The figures show that capil-

Fig. 1. Capillary pressure Pm vs. meniscus position x for a wavy capillary of d1 = d2 = 0.5 and the wetting angle values h = 0°, 17°, 32° and 90°. Inset graph 1: pore shape and a flat meniscus at the beginning of a pore for h = 32° corresponds to Pm = 0. Double arrow shows the hysteresis magnitude in the wavy capillary when h = 32°.

lary pressure inside a wavy pore depends on the position of the meniscus. Effects concern both the magnitude and the sign of the capillary pressure as well. From Fig. 1 it is evident that for wetting angles h between 0 and 32°, the capillary pressure Pm is always P0. The value h = 32° is thus the critical one (hcr) for such a pore and has the following physical meaning: if h > hcr then the liquid cannot spontaneously enter such a pore because at a certain meniscus position the capillary pressure Pm = 0 and the meniscus becomes flat (see the curve for h = 32° in Fig. 1). The critical wetting angle hcr is related to the pore shape parameter d2 by the following equation (Czachor, 2007) derived from Eq. (4), for the capillary pressure Pm = 0:

tgðhcr Þ ¼ 1=pd2

ð5Þ

The value of cos(p/hx) = 1 in Eq. (4) corresponds to the largest wall slope angle value a(x = 0) from Eq. (2). In the case of a wavier pore (i.e. a larger d2 value), as depicted in Fig. 2, the critical value of the wetting angle hcr = 17° < 32° and the capillary pressure is negative for the left and right part of the domain. This means that an external pressure is needed to push a liquid inwards until the meniscus reaches the position indicated in Fig. 2a by an arrow (x  0.3 h). Then the liquid will fill a pore until x  1.7 h even if there is no external pressure. The above considerations demonstrate that in the case of non-cylindrical pores, the critical wetting angle hcr can be much smaller than 90°. The model pores used in the above example have a hysteretic character, i.e. its drainage and wetting curves are not identical. This is a feature of soils and other porous media and clearly demonstrated in Figs. 1 and 2, where for a given wetting angle, there is

106

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111

domain depicted in Figs. 1 and 2. It is worth remembering here that in the case of cylindrical pores and a wetting angle h P 9°, the capillary pressure Pm < 0 independently of the meniscus position. Under such conditions a liquid would escape spontaneously from a cylindrical water repellent pore, whereas it would remain in a wavy pore because the maximum value of the capillary pressure > 0 even for a wetting angle h > 90°. This phenomenon is Pmax m exemplified in Figs. 1 and 2, where the curve corresponds to a wetting angle h = 90°. From the data in Fig. 3 it is evident that relatively occur over a small changes in the maximum capillary pressure Pmax m large wetting angle h range between 0° and 40–50°. The above effect becomes more pronounced when the capillary shape parameter d2 increases (see curves corresponding to d2 = 0.5 and d2 = 1 in Fig. 3). These relationships are of significance in the understanding of water repellent soil behavior. The importance of treating soil capillaries as non-cylindrical is therefore substantial. The values of wetting angles of 17° and 32° in the examples given in Figs. 1 and 2 are near-critical for wavy pores of d2 = 1 and 0.5, respectively. For the wavy pore, the ratio of the highest and the lowest pressure is about 50, while for a cylindrical pore it is 1. In contrast the impact of wetting angle on the ratio of the high  h¼32  1. est values is low Ph¼0 max =P max

Fig. 2. Capillary pressure Pm vs. meniscus position x for d2 = 2d1 = 1 and the wetting angle values h = 0°, 17°, 32° and 90°. Inset graph 2: pore shape and a flat meniscus position in a pore for h = 32° corresponds to Pm = 0. The double arrow shows the hysteresis magnitude in the wavy capillary when h = 32°.

a difference between the highest and the lowest value of the cap Pmin (see bold double arrow in the illary pressure DP ¼ P max m m upper right hand of Figs. 1 and 2). The lowest value determines the pressure at which the horizontal pore becomes wet, while the highest corresponds to its drying. The moisture retention curve of a soil (during the drying curve) can be seen as set of pressure– moisture content values, for which some pores become empty because their highest capillary pressures are smaller or equal than that of an adjacent one. The relationships in Figs. 1 and 2 can be applied as a useful tool towards a division of soils into wettable and water repellent, taking into account the lowest capillary pressure in a hypothetical wavy pore. This main point of the division concerns the minimum capillary pressure in a wavy pore, which is positive for wettable and sub-critically repellent (Goebel et al., 2008), and negative for water repellent soil. Moreover, it supports the existence of the category ‘sub-critically’ repellent. The term subcritical repellency has been introduced for soils that show reduced wettability, but not water repellency (Hallett et al., 2001). The latter is commonly defined as a drop of water not infiltrating within a few seconds (Letey et al., 2000). It implies that, for soils that are classed as wettable or non-repellent (i.e. WDPT < 5 s), their wettability may be far from a perfect. The data presented in Figs. 1 and 2 suggests a phenomenon that would seem to contradict conventional thought: that water retention is a feature of not only wettable, but also of soils that exhibit water repellency. This implies that removing water from a water repellent medium that may have been wetted under external pressure does not occur spontaneously, but requires external pressure as well. Drying of a water repellent medium will not occur without suction that equals the maximum capillary pressure in the wavy

It is widely accepted that certain types of soil organic matter modify the wetting properties of the solid soil phase (Doerr et al., 2000; DeBano, 2000). Soil water repellency, however, is also affected by soil moisture content and relative air humidity (Dekker et al., 2001; Doerr et al., 2002), which are related to the relationships described below. Soil wetting is affected by the minimum capillary pressure Pmin m , which is a strongly wetting angle-dependent value. This relationship is presented in Fig. 4 for the same capillaries used in Fig. 3. The imbibition (wetting) process can occur when the minimum is positive. The range of wetting angle capillary pressure P min m where this condition is fulfilled for a capillary characterized by d2 = 0.5 is relatively small (between 0° < h < 32°). However, the wetting angle value of a clean planar quartz surface is 21.4° (as described in more detail in Section ‘‘Methods for experimental verification”). This signifies that a wetting angle range fulfilling the imbibition condition is much smaller and equals 32–21.4 = 10.6° only. One can therefore speculate that relatively small changes in wetting angle can lead to substantial changes in soil wettability, encompassing also ‘switches’ between a wettable and water repellent status. This may explain the, in some cases substantial, changes in soil wettability that have been observed with relatively small changes in soil water content (de Jonge et al., 1999; Dekker

Fig. 3. Maximum capillary pressure P max vs. wetting angle h in two wavy capillaries m (d2 = 0.5 and d2 = 1) and in a cylindrical one (d2 = 0).

107

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111

Fig. 4. Minimum capillary pressure P max vs. wetting angle h in two wavy capillaries m (d2 = 0.5 and d2 = 1) and in a cylindrical one (d2 = 0).

et al., 2001). The above theoretical considerations have important repercussions for understanding of the SWRC, signifying that the wetting angle has little impact on the drying branch. It the calculations and their implications presented above, we have not considered the potential ‘decay’ of hydrophobicity (i.e. increase in surface tension) of individual soil particle surfaces during water contact. This is believed to occur due conformational changes of hydrophobic and hydrophilic parts of organic molecules present on soil particles during water contact (see review by Doerr et al. (2000)). Our calculations are therefore based on a simplified ‘permanent’ hydrophobicity of soil particle surfaces, however, its principles and implications also apply to more complex scenarios with changing particle surface hydrophobicity. Methods for experimental verification The theoretical considerations above were tested experimentally using materials of contrasting wettability including mixtures of hydrophobic and wettable ground glass particles used as ‘model soil’, and field soil samples. Preparation of ‘model soils’ with different water repellencies

subsample of this quartz sand was hydrophobized by means of wetting them to a near-saturated state with an octadecyloamine:propanol (1:500 v:v) solution followed by oven drying at 40–50 °C. This procedure was repeated three times resulting in highly water repellent grains. A water drop placed onto its surface was stable for several hours and eventually evaporated indicating, for the timescales involved here, a ‘permanent’ water repellent behavior of the individual grains, which the organic coating on this ‘model soil’ particles would retain during prolonged water contact. The water repellent and untreated sands were combined to get a set of ‘model soils’ containing 0%, 15%, 30%, and 45% of highly water repellent quartz grains. To obtain a comparable flat quartz surface, planar glass was cleaned using concentrated hydrochloric acid and methyl alcohol. An area of ca. 40 cm2 was treated with the same octadecyloamine:propanol solution to obtain a comparable hydrophobic agent load as applied to the quartz sand. The specific surface area of the sand (658 cm2 g1) was estimated from its grain size distribution assuming particle sphericity. More specifically the known grain size distribution (GSD) and density of this material allowed calculation of the number of each class of particles in a given sample mass. The specific surface of the material was then estimated based on the sum of the calculated particle surface areas from all classes. The applied octadecyloamine load thus equaled 0.010 g per m2 of quartz powder surface. Three 40-lL distilled water drops were placed on each of the surfaces of the clean and the hydrophobized quartz glass and imaged. Images were analyzed by using the image analysis software Aphelion 3.2 (Amerinex Applied Imaging, Inc., and ADCIS SA (France)) to determine the geometrical parameters of the water drops and respective wetting angles of both surfaces. The wetting angles of clean and hydrophobic quartz were 21.41° and 53.76°, respectively (average of three drops; Table 2). These values were used to represent approximations of actual wetting angles of the untreated and treated glass particle surfaces respectively (0% and 100% mixtures). The actual wetting angle for mixtures of 15%, 30% and 45% water repellent grains were calculated from the Cassie–Baxter equation (Cassie and Baxter, 1944)

cos h ¼

fi cos hi

ð6Þ

i¼1

where h is the contact angle of a liquid on the heterogeneous surface composed of fractional area fi, which fulfils the condition n X

Quartz glass material was chosen to produce ‘model soil’ as it is a very common soil mineral, and also available as a flat regular surface (allowing sessile drop measurements). Flat quartz glass (50  100  4 mm) and some kilograms of quartz clunker (small pieces of waste material) were obtained. The quartz clunker had been ground by means of an agate fine pulverizer and the particle fraction of 50–200 lm was separated for further experiments. A

n X

fi ¼ 1

ð7Þ

i¼1

where hi represents the Young wetting angle of ith fraction. In this case the number of fractions = 2. It is worth pointing out here that the wetting angle of the highly water repellent ‘model soil’ used here (fi = 45%) is 39° and thus much lower than 90°, which is the widely used demarcation between hydrophilic and hydrophobic

Table 1 Soil classification (ISSS, 1988), grain size distribution (dominant fraction in bold font), organic carbon content (C_org) and WDPT of the soils investigated following exposure to a relative humidity (RH) of 98.6%. Soil

% Fraction (mm)

WDPT (s)

1.0–0.1

0.1–0.05

0.05–0.02

0.02–0.005

Brown soil (Haplic Cambisol)

47

16

10

11

8

8

10.6

83

Podzol (Haplic Arenosol)

82

6

4

3

2

3

10.9

4100

Gray brown podzol (Haplic Luvisol)

19

11

38

14

7

11

9.8

<1

6

8

43

22

7

14

14.4

<1

13

14

7

11

18

37

12.3

<1

Chernozem (Haplic Phaeozem) Redzina (Haplic Leptosol)

0.005–0.002

C_org (g kg1) <0.002

108

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111

Table 2 Measured (bold font) and calculated (normal font) Young wetting angle h (with standard deviation; SD) and WDPT (mean of three measurements following exposure to a relative humidity of 98%) for ‘model soils’ (quartz powders of different water repellency). f1 (%) depicts the percentage of hydrophobized grains mixed with untreated grains. Fraction f1 (%)

0

15

30

45

100

Wetting angle h (°) SD (°) WDPT (s) SD (s)

21.41 0.82 0 0

28.40 – 5.3 0.47

34.00 – 136.7 17

38.92 – 263.3 54

53.76 2.21 >5000 N.a.

surfaces. Measured and calculated wetting angles for the hydrophobized model sand are shown in Table 2. Field soil samples Five arable soils, ranging from sandy to heavy clay soil, were sampled from the plough layer (0–30 cm) in the Lublin region in SE Poland (see Table 1 for details). Subsamples were heated in oven at 500 °C for 48 h to remove soil organic matter and any associated water repellency (King, 1981). Natural samples (prior to removal of organic matter) are identified by the suffix ‘OM+’, and following organic matter removal by ‘OM’. Six Kopecky cylinders (100 cm3 volume; 4.7 cm height; 5.2 cm ID) were prepared for every soil type. Determining sample hydraulic characteristics The persistence or stability of water repellency was measured in triplicate for all soil and ‘model soil’ samples using the water drop penetration time test (WDPT; Letey, 1969). Measurements were carried out following of exposure of samples to a relative humidity (RH) of 98.6%. Soil water repellency is known to be affected by changes in RH (Doerr et al., 2002), and this RH level was chosen as being most representative of conditions present in the soil matrix during wetting and drying phases. The Gray brown podzol, Chernozem and Redzina were fully wettable (WDPT < 1 s), whereas the Brown soil and the Podzol were water repellent (WDPTs 83 and 4100 s, respectively; Table 1). Following organic matter removal (OM) all soils were fully wettable (WDPT < 1 s; data not shown). WDPTs for the ‘model soil’ varied with hydrophobized fraction f1 in the mixture and was <1 s for pure quartz, 5.3 s for f1 = 15% (marginally water repellent), 136.7 s for f1 = 30%, 263.3 s for f1 = 45% and >5000 s for f1 = 100% (Table 2). Three replicate Kopecky cylinders were filled with an equal mass of each of the ‘model soil’ material. Samples of 0% and 15% repellent grains were saturated via capillary rise. Due to their high water repellency, the 30% and 45% samples had to be fully wetted via continuous 10–15 min mixing with water, after which they were placed into Kopecky cylinders. For each soil, three replicates for the natural soil state (OM+) and three for each of the organic matter free soils (OM) were prepared. The bulk density of the dry soils in the cylinders was identical for the OM+ and OM samples (except for the Podzolic soil; Table 3). All soil samples were inserted in water bath to a depth of 3 cm and wetted via capillary

Fig. 5. Drying curves for’ model soils’ (quartz sand; fraction 50–200 lm) with 0%, 15%, 30% or 45% of sample material consisting of particles treated with a octadecyloamine:propanol solution.

rise. After 24 h, it was assumed that the samples were sufficiently wetted to measure the drying branches of the SWRCs. This was done using commercial pressure plates (Soil Moisture Equipment Co., USA) for pF values 2, 2.5, 2.7, 3, and 3.7, and using our own system, which had been Taylor-made for low suctions based on the hanging column method for pF values 0, 1.0, and 1.5. In using the Soil Moisture Equipment system, the soil samples were placed inside a closed chamber using a porous plate of known capillary pressure. When sufficient pressure in the chamber is applied, water flows from the samples through the plate. At equilibrium state, the outflow becomes zero, allowing soil sample moisture to be determined. Our lower pF value system is composed of porous saturated plate connected by means of elastic tube with a water reservoir. Soil samples were placed on top of the porous plate. The vertical distance between the reservoir water level and the porous plate equals the applied water pressure. The equilibrium state and moisture content of measured soil samples were controlled via weight measurement. The wetting branch of SWRC was determined for the five soils (Table 1) and the ‘model soils’ with a 0% and 15% repellent grain fraction (Table 4). For ‘model soil’ material with higher repellent grain fractions, no capillary rise occurred. Measurements were performed using a simple device composed of several glass cylinders of 10 mm ID and lengths ranging from 10 to 100 mm. All segments were line-combined by means of adhesive tape and placed inside a 14 mm ID glass cylinder. The maximum length of the segmented columns was 500 mm. A column was then filled with ‘model’ or ‘real’ soil material. Micro-vibration was applied to the column wall to compact the material to a bulk density equaling that in the Kopecky cylinders. The column base was covered with a highly permeable textile and placed vertically in water. The column weight was recorded until it reached a stable value. When the equilibrium state was attained, the weight of cylindrical segments was mea-

Table 3 Bulk densities (BD in g cm3; average of three replications) and standard deviation (SD) of the five soils in their natural state (OM+) and after organic matter removal (OM). Soil

Redzina (Haplic Leptosol)

Gray brown podzolic (Haplic Luvisol)

Brown (Haplic Cambisol)

Chernozem (Haplic Phaeozem)

Podzolic (Haplic Arenosol)

OM+ SD OM SD

1.147 0.021 1.163 0.031

1.370 0.009 1.405 0.005

1.485 0.029 1.495 0.026

1.265 0.005 1.290 0.000

1.765 0.017 1.895 0.015

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111 Table 4 Bulk densities (BD in g cm3; average of three replications) and standard deviation (SD) of sand model material used for water retention measurements containing 0%, 15%, 30% or 45% hydrophobic grains. Hydrophobic grain content (%)

0

15

30

45

BD SD

1.280 0.020

1.277 0.021

1.353 0.034

1.357 0.025

sured. After oven drying at 50 °C for at least 16 h to equilibrium weight, the average moisture content was calculated for each segment. We did not use a higher drying temperature in order to avoid any heat-induced increases in sample water repellency (Doerr et al., 2005). The moisture content along the column at the equilibrium state was considered to represent the wetting branch of the water retention. Experimental results and relevance to theoretical considerations Drying branches of the SWRC for ‘model soils’ are given in Fig. 5 and the results for soils, exemplified by the Brown soil (Haplic Cambisol) and Redzina (Haplic Leptosol) are given in Fig. 6. Each curve represents an average value of three replications. The SWRCs of the three other soils, i.e. Podzol (Haplic Arenosol), Gray brown podzol ( Haplic Luvisol) and Chernozem (Haplic phaeozem), have similar features and are therefore not shown. Their pF curves in their natural soil state (OM+) and following organic matter removal (OM) are almost identical. The data in Fig. 5 indicates that the drying branch of water retention is almost independent of the wettability of the ‘model soil’ despite the fact that the 45% repellent grains were extremely water repellent with a WDPT > 1 h. (Note that the differences in saturated water contents of ‘model soils’ are related to different preparation procedures used for the lesser and more water repellent samples.) This observation is not consistent with the capillary bundle model, but supports the model of wavy capillaries introcan be duced earlier where the maximum capillary pressure Pmax m positive for a water repellent medium (wetting angles h > 90°). The impact of wetting angle h on the drying branch is thus relatively weak. This finding is consistent with experimental results obtained by Nieber et al. (2000) for silica sands with different water repellencies. From the theoretical considerations depicted in Fig. 3, it can be seen that for a wavy capillary of d2 = 1, the ratio max   P max m ðh ¼ 0 Þ=P m ðh ¼ 90 Þ ¼ 4=3, while in a cylindrical one it goes

Fig. 6. Drying curves for the Brown soil (Haplic Cambisol) (B_) and the Redzina (Haplic Leptosol) (R_) in their natural state (OM+) and without organic matter (OM).

109

to infinity. The estimated range of wetting angles for the ‘model soils’ (Table 2) is much smaller (from 21° to 39°). Thus, the ratio max   Pmax m ðh ¼ 21 Þ=P m ðh ¼ 39 Þ is slightly larger than 1, which explains the small differences between the curves seen in Fig. 5. This finding is relevant for the interpretation of the results concerning the soils. Another possible way to interpret the results obtained is also considered here. It is widely accepted that the persistence of water repellency, as measured by WDPT, is a function of moisture content. At a certain moisture content range (called a transition zone) an initially repellent soil becomes wettable (WDPT < 5 s; Dekker et al., 2001). This suggests that the wetting angle of a soil is also a function of soil moisture content with the wetting angle of a wet soil being lower than a static wetting angle measured at the flat dry surface. It should be approaching zero for a drop of water drop spreading on a soil surface that is already covered by a thin water layer. In such a case, a material that is repellent when dry, but would be wettable during drying as long as a water film is present, could be characterized by a wetting angle <39° (or even a zero). In other words, the curves presented in Fig. 5 can be interpreted to concern water repellent as well as wettable media. The above interpretation fits the observation that a receding wetting angle, which is relevant to the drying process, is always lower compared to the static one that was measured for the flat surface here. Both interpretations presented above are able to explain the similarity of the drying branches of repellent or sub-critically repellent porous media. However, in the case of the ‘model soils’ discussed above, it is likely that the repellent grain fraction contained therein would have retained their hydrophobic surface nature due to the ‘permanent’ water repellency invoked in the grain hydrophobization process. As regards wetting angles, when wetting occurs the angle involved is larger than a static one and is termed ‘advancing’ wetting angle (Bachmann et al., 2000). However, this wetting angle cannot be larger than the critical one introduced here. The calculations and data presented here suggest that wetting angles fulfilling imbibition conditions are likely to be larger than 10° when soils are not dry. According to the capillary bundle model, the absence of organic matter and hence the potentially greater surface tension of the mineral material, should change the soil wetting angle and the SWRC. However, the drying SWRCs of the ‘model soils’ and field soils investigated here were very similar for sample pairs with and without organic matter. It is feasible that for the soil samples, water repellency decayed sufficiently during the wetting process such that it would have little impact during drying. This is perhaps common in field soils and would, for example, help to explain the adsorption and desorption curves (and intermediate loops) described for an initially water repellent sandy soil by Ritsema et al. (1998). This, however, is unlikely to apply to the ‘model soils’ which exhibit a ‘permanent’ water repellency of the treated grain fraction. For these, the wetting curves are very different, which incidentally also applies to the Brown soil (Fig. 7). In fact the wetting curves in Fig. 7 of both respective pairs, the untreated (0%) ‘model soil’ and the Brown soil without organic matter (B_OM), and the water repellent (15%) ‘model soil’ and natural Brown soil (B_OM+), are remarkably similar, suggesting that the Brown soil has a very similar capillarity character to the artificially created ‘model soil’. It is worth underlining here that in contrast to the drying branch, the wetting branch is very sensitive to the wetting angle. Small changes in wetting angle can cause switches between wettable and water repellent soil behavior. This effect may be common and would help explaining how relatively small changes in soil water content can cause substantial changes in the WDPT of soils (Dekker et al., 2001; Doerr et al., 2002).

110

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111

Higher Education, and the EC Grant ‘Water Reuse’ (CS516731). This work does not necessarily reflect the European Commission’s views and in no way anticipates its future policy in this area.

References

Fig. 7. Wetting curve for ‘model soils’ (Q_, 50–200 lm fraction) with two hydrophobization levels (0% and 15% of sample material consisted of particles treated the octadecyloamine:propanol solution) and for the Brown soil (Haplic Cambisol) (B_) in its natural state (OM+) and with organic matter removed (OM).

Conclusions and implications Theoretical modeling, assuming a constant surface tension of soil particle surfaces, supported by experimental observations using materials with different water repellency levels (‘model soil’ soils containing grains that exhibit ‘permanent’ hydrophobicity and field soil samples), demonstrate that (i) water retention depends strongly on the solid-water wetting angle for the wetting curve, but has little impact on the drying curve and (ii) that water retention is a feature not of only wettable soils, but also soils that are in a water repellent state. This behavior cannot be explained on the basis of the commonly used, but vastly simplifying, capillary bundle model for soil pores. The new model presented here takes account of the wavy nature of soil pores and explains the experimental observations made on model and field soils. Although the model is based on a simplified (‘permanent’) hydrophobicity of soil particle surfaces, its principles and implications are relevant also to more complex scenarios in which particle surface hydrophobicity may change during wetting and drying phases. The currently used methods for measuring the Young’s wetting angle on soil samples are insufficient in representing the variable wetting angle in the soil pore space and hence its impact on the SWRC. The theoretical predictions and experimental results obtained here indicate that small changes in wetting angle can cause switches between wettable and water repellent soil behavior of the soil pore space. This may explain the common observation that relatively small changes in soil water content can cause substantial changes in soil wettability irrespective of changes in surface tension (i.e. hydrophobicity) of individual soil particle surfaces. The model derived here would also help to explain the substantial reductions in water repellency observed following soil compaction (Bryant et al., 2007), which results in changes in pore geometry and hence wetting angle. The above findings are not only relevant for soils, but also for predicting wettability and water retention of other, potentially water repellent, porous media such as reservoir rocks containing hydrocarbons, or the wide variety of man-made materials used in engineering, construction or medicine. Acknowledgements This investigation has been supported by the National Grant Frame No. 2 P06S 013 30 of the Polish Ministry of Science and

Adesodun, J.K., Davidson, D.A., Mbagwu, J.S.C., 2008. Soil faunal activity of an oilpolluted tropical alfisol amended with organic wastes as determined by micromorphological observations. Applied Soil Ecology 39, 46–57. Bachmann, J., Ellies, A., Hartge, K.H., 2000. Development and application of a new sessile drop contact angle method to assess soil water repellency. Journal of Hydrology, 66–75. Bryant, R., Doerr, S.H., Hunt, G., Conan, S., 2007. Effects of compaction on soil surface water repellency. Soil Use and Management 24, 228–234. Buckman, H.C., Brady, N.S., 1969. The Nature and Properties of Soils. The Macmillan Company, Collier-Macmillan Limited, London. Buczko, U., Bens, O., Hüttl, R.F., 2006. Water infiltration and hydrophobicity in forest soils of a pine–beech transformation chronosequence. Journal of Hydrology 331, 383–395. Carillo, M.L.K., Letey, J., Yates, S.R., 1999. Measurement of initial soil-water contact angle of water repellent soils. Soil Science Society American Journal 63, 433– 436. Cassie, A.B.D., Baxter, S., 1944. Wettability of porous surfaces. Transactions of the Faraday Society 40, 546–551. Czachor, H., 2006. Modelling the effect of pore structure and wetting angles on capillary rise in soils having different wettabilities. Journal of Hydrology 328, 604–613. Czachor, H., 2007. Applicability of the Washburn theory for determining the wetting angle of soils. Hydrological Processes 21, 2239–2247. de Jonge, L.W., Jacobsen, O.H., Moldrup, P., 1999. Soil water repellency: effects of water content, temperature, and particle size. Soil Science Society of America Journal 63, 437–442. DeBano, L.F., 2000. Water repellency in soils: a historical overview. Journal of Hydrology 231–232, 4–32. Dekker, L.W., Doerr, S.H., Oostindie, K., Ziogas, A.K., Ritsema, C.J., 2001. Water repellency and critical soil water content in a dune sand. Soil Science Society of America Journal 65, 1667–1674. Doerr, S.H., Shakesby, R.A., Walsh, R.P.D., 2000. Soil water repellency: its causes, characteristics and hydro-geomorphological significance. Earth-Science Reviews 51, 33–65. Doerr, S.H., Dekker, L.W., Ritsema, R.A., Shakesby, R.A., Bryant, R., 2002. Water repellency of soils: the influence of ambient relative humidity. Soil Science Society of America Journal 66, 401–405. Doerr, S.H., Douglas, P., Evans, R., Morley, C.P., Mullinger, N., Bryant, R., Shakesby, R.A., 2005. Effects of heating and post-heating equilibration times on soil water repellency. Australian Journal of Soil Research 43 (3), 261–267. Givia, J., Prasherb, S.O., Patel, R.M., 2004. Evaluation of pedotransfer functions in predicting the soil water contents at field capacity and wilting point. Agricultural Water Management 70, 83–96. Goebel, M.O., Bachmann, J., Woche, S.K., 2008. Modified technique to assess the wettability of soil aggregates: comparison with contact angles measured on crushed aggregates and bulk soil. European Journal of Soil Science 59, 1241– 1252. Hallett, P.D., Baumgartl, T., Young, I.M., 2001. Subcritical water repellency of aggregates from a range of soil management practices. Soil Science Society of America Journal 65, 184–190. Imeson, A.C., Jongerius, P.D., 1974. Landscape stability in the Luxembourg Ardennes as exemplified by hydrological and micropedological investigations of a catena in an experimental watershed. Catena 1, 273–295. ISSS (International Society of Soil Science), ISRIC (International Soil Reference and Information Centre), FAO (Food and Agriculture Organization of the United Nations), 1988. World Reference Base for Soil Resources. World Soil Resources Reports, 84. FAO, Rome. Kezdi, A., 1974. Handbook of Soil Mechanics. Soil Physics, vol. 1. Akademiai Kiado, Budapest. King, P.M., 1981. Comparison of methods for measuring severity of water repellency of sandy soil and assessment of some factors that affect its measurement. Australian Journal of Soil Science 19, 275–285. Kozeny, I., 1927. Über Kapillarleitfahigkeit des Wassers im Boden. vol. 136. Wiener Akademie der Wissenschaften, Wien. p. 303. Kutilek, M., Nielsen, D.R., 1994. Soil Hydrology. Catena, Cremlingen-Destedt. Letey, J., 1969. Measurement of contact angle, water drop penetration time, and critical surface tensions. In: Proc. Symp. Water Rep. Soils. University of California, Riverside, pp. 43–47. Letey, J., Carrillo, M.L.K., Pang, X.P., 2000. Approaches to characterize the degree of water repellency. Journal of Hydrology 231–232, 61–65. Mualem, Y., 1976. A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resources Research 12 (3), 513–521. Nieber, J.L., Bauters, T.W.J., Steenhuis, T.S., Parlange, J-Y., 2000. Numerical simulation of experimental gravity-driven unstable flow in water repellent sand. Journal of Hydrology 231–232, 295–307. Osborn, J.F., Pelishek, R.E., Krammes, J.S., Letey, J., 1964. Soil wettability as a factor in erodibility. Soil Science Society of America Proceedings 28, 294–295.

H. Czachor et al. / Journal of Hydrology 380 (2010) 104–111 Ritsema, C.J., Dekker, L.W., Nieber, J.L., Steenhuis, T.S., 1998. Modeling and field evidence of finger formation and finger recurrence in a water repellent sandy soil. Water Resources Research 34, 555–567. Shirtcliffe, N.J., McHale, G., Newton, M.I., Pyatt, F.B., Doerr, S.H., 2006. Critical conditions for the wetting of soils. Applied Physics Letters 89. Art. No. 094101. VandenBygaart, A.J., Protz, R., Tomlin, A.D., Miller, J.J., 1999. Tillage system effects on near-surface soil morphology observations from the landscape to microscale in silt loam soils of southwestern Ontario. Soil & Tillage Research 51, 139– 149. Vera, M., Sierra, M., Diez, M., Sierra, C., Martinez, A., Martinez, F.J., Aguilar, J., 2007. Deforestation and land use effects on micromorphological and fertility changes in acidic rainforest soils in Venezuelan Andes. Soil & Tillage Research 97, 184– 194.

111

Vervoort, R.W., Cattle, S.R., 2003. Linking hydraulic conductivity and tortuosity parameters to pore space geometry and pore-size distribution. Journal of Hydrology 272, 36–49. Waniek, E., Szatyłowicz, J., Brandyk, T., 2000. Determination of soil-water contact angle in peat-moorsh soils by capillary rise experiments. Suoseura (3), 149–154. Washburn, E.W., 1921. The dynamics of capillary flow. The Physical Review XXVII (3), 273–283. Zhang, X.X., Deeks, L.K., Bengough, A.G., Crawford, J.W., Young, I.M., 2005. Determination of soil hydraulic conductivity with the lattice Boltzmann method and soil thin-section technique. Journal of Hydrology 306, 59–70.