Preferential solute transport in soil: laboratory and field studies

Retrospective Theses and Dissertations 1989 Preferential solute transport in soil: laboratory and field studies Gerard J. Kluitenberg Iowa State Uni...
Author: Jonathan Pope
5 downloads 0 Views 2MB Size
Retrospective Theses and Dissertations

1989

Preferential solute transport in soil: laboratory and field studies Gerard J. Kluitenberg Iowa State University

Follow this and additional works at: http://lib.dr.iastate.edu/rtd Part of the Agricultural Science Commons, Agriculture Commons, and the Agronomy and Crop Sciences Commons Recommended Citation Kluitenberg, Gerard J., "Preferential solute transport in soil: laboratory and field studies " (1989). Retrospective Theses and Dissertations. Paper 9140.

This Dissertation is brought to you for free and open access by Digital Repository @ Iowa State University. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For more information, please contact [email protected].

INFORMATION TO USERS The most advanced technology has been used to photo­ graph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are re­ produced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University MIcrofiims international A Beli & Howeil information Company 300 Nortii Zeeb Road, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 9014915

Preferential solute transport in soil: Laboratory and field studies Kluitenberg, Gerard J., Ph.D. Iowa State University, 1989

UMI SOON.ZeebRd. Ann Alter, MI48106

Preferential solute transport in soil: Laboratory and field studies by Gerard J. Kluitenberg

A Dissertation Submitted to the Graduate Faculty in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Department: Agronomy Major: Soil Physics

Approved:

Signature was redacted for privacy.

In Charge of Major Work Signature was redacted for privacy.

or the Major Department Signature was redacted for privacy.

For^he Graduate College

Iowa State University Ames, Iowa 1989

ii

TABLE OF CONTENTS Page GENERAL INTRODUCTION Explanation of Dissertation Format

SECTION I. EFFECT OF MACROPORES ON THE SPATIAL VARIABILITY OF FIELD-MEASURED SOLUTE TRANSPORT PROPERTIES

1 4

5

ABSTRACT

6

INTRODUCTION

8

MATERIALS AND METHODS

12

Field Experiment

12

Transport Theory

17

Steady-State Analysis

19

Water Flux Analysis

22

RESULTS

23

Soil Moisture

23

Measured and Fitted Solute Concentration Profiles

27

Spatial Variability of Transport Parameters

30

Spatial Correlation Analysis

43

DISCUSSION

47

SUMMARY

55

REFERENCES

57

ill

Page SECTION II. EFFECT OF SOLUTE APPLICATION METHOD ON PREFERENTIAL TRANSPORT OF SOLUTES IN SOIL

61

ABSTRACT

62

INTRODUCTION

64

MATERIALS AND METHODS

66

RESULTS AND DISCUSSION

71

Characterization of the Soil Material

71

Pulse and Drip Application Experiments

79

Implications for Transport Experiments

83

Effect of Initial Soil Water Content

85

SUMMARY

89

REFERENCES

91

GENERAL SUMMARY

93

ADDITIONAL REFERENCES

96

ACKNOWLEDGMENTS

98

APPENDIX

99

\

1

GENERAL INTRODUCTION Contamination of groundwater by agricultural chemicals is a national problem. Nearly one half of the nation's drinking water supply comes from groundwater, and 95% of all rural households rely on groundwater for their drinking water (CAST, 1985). It is not surprising that the public is concerned about the quality of its water.

Locally,

in the state of Iowa, agricultural chemicals have been detected in many of the state's groundwater supplies. Hallberg (1986) indicates that many of the water supplies exhibit nitrate concentrations above the maximum contamination limit for public drinking water.

Kelley et al.

(1986) Indicate that trace amounts of a variety of pesticides have also been found in these waters. To combat this problem, the state of Iowa has recently enacted precedent-setting groundwater legislation. The discovery of agricultural chemicals in groundwater systems is surprising to many because the leaching potential of many of these chemicals was thought to be low.

Apparently, our current understanding

of the important physical processes is insufficient for sound predictive modeling of chemical movement. Recent improvements in our understanding of flow processes have come from realizing the Importance of non-uniform transport of water and solutes in soil, and from realizing the Importance of spatial variability of soil properties in transport phenomena. Thomas and Phillips (1979) and Seven and Germann (1982) have reviewed the results of a number of water and solute transport experiments that were conducted with soil material containing

2

accropores, or pathways of preferential flow. Their reviews Indicate that In many instances, water and solutes are able to move preferentially through cracks, structural planes, root channels, and channels formed by soil biota, bypassing a significant portion of the soil matrix. Both Thomas and Philips (1979) and Beven and Germann (1982) make reference to observations made in the late 1800s regarding the occurence of preferential flow, but the importance of preferential water and solute transport as a contribution to net transport has only recently been established, aided in part by concerns for the diminishing quality of our water resources. Non-uniform flow has not only been observed in the presence of macroporosity. Glass et al. (1989b) provide experimental evidence for the fingering of infiltrating water in homogeneous sands as the result of a textural discontinuity. In a companion paper. Glass et al. (1989a) review related experimental observations of fingering in homogeneous porous media. Krupp and Elrick (1969) as well as numerous others have also demonstrated that non-uniform flow can result from gradients in density and viscosity during the displacement of one miscible fluid with another. At present, we will limit our discussion to the preferential transport of water and solutes through soil macroporosity. Regardless of cause, preferential transport of water and solutes can result in spatial distributions of water and solutes that are not predicted by classical transport theories. Improvements in our ability to model the movement of chemicals in soil, and improvements in our ability to manage the application of agricultural chemicals to soil.

3

will come only after a more complete understanding of preferential flow processes is obtained. The second major improvement in our understanding of flow processes has come from realizing the importance of the spatial variability of soil properties in field-scale transport phenomena.

Early studies by

Biggar and Nielsen (1976), Bresler and Dagan (1979), Jury et al. (1982), and Amoozegar-Fard et al. (1982) demonstrated that a description of field-scale water and solute transport can not be achieved by using classical laboratory-scale transport theory with field-measured "average" transport coefficients. It is now generally recognized that the governing transport equations for water and solute transport require a description of how the pertinent soil transport and retention properties vary spatially over the area of interest. Jury et al. (1987) provide an excellent summary of work that has been conducted on this topic. It is clear from their summary, as well as from the work of others, that the spatial variability of soil properties within an area of interest controls to a large extent the shape of the area-averaged solute concentration profile or the shape of the area-averaged breakthrough curve observed at some depth below the surface.

Predictive

modeling of chemical movement at the field scale, and successful management of chemicals applied to agricultural fields requires that detailed investigations be made to characterize the spatial variability of soil transport and retention properties at different spatial scales. This study was focused on obtaining a more complete understanding of the process of preferential flow. Section I of the study describes

4

the results of a field study that provided a detailed characterization of chloride movement in a soil known to have a high potential for preferential flow. The spatial variations of chloride movement were characterized and contrasted with other published reports of spatial variability. Section II of the study describes a series of laboratory experiments that were conducted to examine the importance of the method of solute application on subsequent preferential flow of the solute. The experiments in Section II were conducted to answer specific questions raised in analyzing the results of the field experiments.

Explanation of Dissertation Format This study is presented in two sections.

Each section was prepared

as a complete article in a format acceptable for publication in a refereed scientific journal. The first section, "Effect of Macropores on the Spatial Variability of Field-Measured Solute Transport Properties", will be submitted for publication in the soil physics division of the Soil Science Society of America Journal or in Water Resources Research. The second section, "Effect of Solute Application Method on Preferential Transport of Solutes in Soil", has been accepted for publication in a special issue of Geoderma that will feature articles on preferential flow processes. Following these two sections is a General Summary of the results and conclusions of the two studies. Literature cited in this General Introduction and in the General Summary is listed under Additional References.

5

SECTION I.

EFFECT OF MACROPORES ON THE SPATIAL VARIABILTY OF FIELD-MEASURED SOLUTE TRANSPORT PROPERTIES

6

ABSTEIACT Numerous laboratory-scale experiments have been conducted to examine the importance of preferential water and solute movement in soils containing an abundance of preferential flow pathways. Experimental data are needed to ascertain the importance of preferential water and solute movement at larger scales. This paper presents the results of a solute transport experiment conducted within two 3-m by 3-m field sites. Objectives of the experiment were to examine the movement of surface-applied chloride in a soil that is known to contain many large and continuous macropores, and to obtain a high quality data set for use in field-scale modeling. Another objective of the field experiment was to characterize the observed spatial variability and spatial correlation structure of the measured solute transport properties. A chloride solution was sprinkled on the surface of field plots at two locations. The chloride pulse was followed by sprinkler irrigations of chloride-free water. Soil samples were collected in a square grid pattern so as to yield depth-concentrâtion profiles below each surface gridpoint. The convection-dispersion equation was fit to each concentration profile to yield estimates of D, the dispersion coefficient, and Vg, the average pore-water velocity.

The convection-

dispersion equation provided an adequate description of the individual concentration profiles, and yielded lognormal frequency distributions for both D and Vg.

The Matheron semivariogram estimator was used to

examine the spatial structure of the estimated transport parameters, and it revealed that both D and Vg displayed no spatial dependence for

7

separation distances of up to 3m. Mean values of v^ obtained for both sites were 0.8 times the pore-water velocity, v^, which was calculated by dividing the known flux at the surface by the average soil profile water content. This indicates that preferential movement of soil water through relatively large pore sequences partially bypassed chloride residing in smaller pores within the soil matrix. A ratio of v^/v^ less than one indicates a limitation in the ability of the convectiondispersion equation to describe chloride transport in this experiment.

8

INTRODUCTION Groundwater contamination resulting from the leaching of agricultural chemicals through soil is an increasingly important problem. Loss of fertilizers and pesticides to groundwater not only poses an environmental hazard but also represents an economic loss to farmers. The widespread detection of agricultural chemicals in groundwater supplies has challenged scientists because the leaching potential of many of these chemicals was thought to be low.

Chemical

transport modeling based on classical, laboratory-scale miscible displacement theory has not been able to account for the many observations of rapid and iar-reaching chemical movement. Research on solute movement in the last two decades has provided insight into the disparity between theory and observation. The pioneering work of Biggar and Nielsen (1976) showed that a description of field-scale solute movement must account for the spatial variability of soil physical properties. Studies by firesler and Dagan (1979) and Amoozegar-Fard et al. (1982) incorporated the spatial variability of soil physical properties and demonstrated that field-averaged solute profiles will differ in shape distinctively from the normal, or Gaussian-shaped solute profiles predicted by using classical miscible displacement theory. These workers calculated positively skewed concentration profiles with long tails, indicating the possibility of rapid and deep movement of solute at some field locations. In the work of Amoozegar-Fard et al. (1982), the skewed distributions were observed despite the fact that Gaussian-shaped solute profiles were assumed at

9

each individual spatial location. Initial efforts at characterizing the spatial variability of solute transport parameters focused on describing the probability density functions or the first two moments of the probability density functions of the transport parameters (Biggar and Nielsen, 1976; Van De Pol et al., 1977; Jury et al., 1982; Gish and Coffman, 1987). More recently, attention has focused on the fact that solute transport parameters may be spatially dependent, that is, parameters sampled or determined at neighboring spatial locations are correlated. Some characterizations of spatial dependence have been completed for soil hydraulic properties (Gajem et al., 1981; Sisson and Wierenga, 1981; Russo and Bresler, 1981), but similar direct characterizations of measured solute transport properties have not been reported. These characterizations are needed for several reasons. Descriptions of spatial correlation structure are needed in developing optimal sampling strategies, and form the basis for kriging (Matheron, 1963). Jury et al. (1987) also point out that disregard of the spatial structure of a soil property may lead to biased estimates of the moments of the probability derisity function of the property.

More important, however, is the realization that the spatial

correlation structure of a porous medium directly influences the spatial correlation structure of transport processes in that porous medium. This principle was demonstrated nicely in the work of Smith and Schwartz (1980) which involved modeling mass transport in an aquifer. They showed that changing the correlation lengths of the hydraulic conductivity field resulted in changes of correlation length of tracer

10

velocities. Research on solute transport in the last two decades has shown that rapid and far-reaching chemical movement can also be the result of preferential water and solute movement through macropores.

Reviews by

Thomas and Phillips (1979), Beven and Germann (1982), and White (1985) cite numerous examples in which water and solutes bypassed much of the soil matrix by moving preferentially through cracks, worm holes and root channels. In field studies, preferential flow has been detected by the use of tracers. The average pore-water velocity, Vg, can be determined by fitting an analytical solution of the convection-dispersion to a tracer breakthrough curve or concentration profile. This velocity can be compared with v^, the average pore-water velocity computed from a water balance with the assumption that water movement is uniform and that complete displacement occurs. Rice et al. (1986) and Bowman and Rice (1986) have reported results in which the ratio of Vg to v^ was significantly greater than 1.

Cassel (1971), on the other hand,

observed ratios that were less than 1 for the movement of chloride. Others have reported excellent agreement between Vg and v^ (Biggar and Nielsen, 1976; Gish and Coffman, 1987). The objectives of this field study were to examine the movement of chloride in a soil with an abundance of macropores or preferential flow pathways, and to obtain a high quality data set for the purpose of field-scale modeling. The field experiments were conducted on a soil type for which preferential flow has been well documented (Kanwar et al., 1985; McBride, 1985; Priebe and Blackmer, 1989; Kluitenberg and

11

Horton, 1990; Everts et al., 1989).

Another objective of this study was

to characterize the spatial variability of chloride movement.

12

MATERIALS AND METHODS

Field Experiment Two field sites (Fig. 1) were established in Story County, Iowa on the Clarion-Nicollet-Webster soil association. These soils have developed from loam-textured glacial till or till-derived sediments. One site (Upper Site) was located on the well drained Clarion soil series (fine-loamy, mixed, mesic Typic Hapludoll). The other site (Lower Site) was located on the somewhat poorly drained Nicollet soil series (fine-loamy, mixed, mesic Aquic Hapludoll).

The area at both

sites had been In continuous corn production with minimum tillage for the previous five years. Both areas received a light disking before establishing the sites, and the sites were then protected from subsequent wheel traffic. Both soils contained significant macroporosity. Kluitenberg and Horton (1990) have described some physical and morphological characteristics of the Nicollet soil. They counted worm and root channels in diameter classes of 0.5 - 2.0 mm and 2 - 5 nmi for 3 undisturbed soil cores (18 cm diameter, 254 cm^ cross sectional area). The average horizontal cross-section in the surface 30 cm contained approximately 8 channels from each size class. Dye experiments indicated that approximately half of the channels conducted dye solution during saturated flow. Vertical planar voids were also observed to conduct dye solution. After disking, corn residue and large clods were removed from the surface at both sites. Next, the soil surface was carefully smoothed.

Raised Border

Figure 1. Diagram of plot design used at both upper and lower experimental sites

14

leaving a thin layer of loose soil. Raised borders were installed to control surface water flow, and herbicide was applied to keep the sites free of vegetation. Twine was used to delineate 36 0.5-m by 0.5-m plots within each 3-m by 3-m site. •The sites were then left uncovered for a period of two months during which time approximately 6 cm of rainfall was recorded. Several days before beginning the leaching experiments, the upper and lower sites were flood irrigated with 100 mm of well water to establish an initially moist condition. After infiltration was complete, the sites were covered with polystyrene foam and polyethylene plastic to prevent evaporation and exclude rainfall. The sites remained covered throughout the experiment except to apply additional irrigation water and to collect samples. Experimentation was initiated with the application of a chloride solution. At both upper and lower sites, 10 mm of GaCl2 solution (C^ ~ 967 meq/L) was applied in 2.5 mm increments to each of the 36 plots, one at a time. To avoid disturbance of the soil surface, moveable bridges were used to access the experimental sites. apply the solution.

A sprinkler can was used to

A wind shield and back-and-forth motion was

employed to ensure uniform application of the solution. After the application of the CaCl2 solution, four uniform 50.8 mm irrigations of well water were applied to move the chloride pulse through the soil (Table 1). Chloride was present in the well water at only 0.02 meq/L. For each irrigation, water was added in 5 mm increments over the course of two days to minimize ponding. Shallow ponding did occur near the

15

Table 1. Schedule of Irrigation and soil sampling dates

Time

Flux-time®

Irrigation

t(d)

t*(d)

(mm)

(mc)

10.0 25.4 25.4

35.4 60.8

25.4 25.4

86.2 111.6

25.4 25.4

137.0 162.4

25.4 25.4

187.7 213.2

0 1 2 5 12 13 16 19 20 23 26 27 29

Cum. Irrig.

8.27

Sampling

X

15.18

X

22.09

X

29.00 ^Flux-time, t*, is computed from Eq. [9].

X

conclusion of both the chloride-free well water Irrigations and the chloride solution application. Soil samples were collected a few days after each irrigation (Table 1). Sixteen of the 36 plots were used for sampling (Fig. 1), and these were further subdivided into 4 subplots.

On each of the four sampling

dates, a set of 16 samples was collected from both upper and lower sites.

Samples corresponding to sampling dates I, II, III, and IV were

taken from subplots i, ii, iii, and iv, respectively.

A 102-nmi-diameter

auger was used to take samples at 0.05 m intervals to a depth of 0.25 m. Then, a 76-mm-diameter auger was used to obtain samples at 0.05 m intervals to 0.5 m and at 0.1 m intervals to a depth of 1.5 m. These sample sizes result in sample volumes of 405, 228, and 456 cm^ for the depth intervals 0 - 0.25, 0.25 - 0.50, and 0.50 - 1.5 m, respectively. Sample holes were back-filled with soil. Soil samples were collected in autoclavable bags so that the gravimetric water content of each whole sample could be determined. After determination of water content, the samples were ground and passed through a 2 mm sieve. Subsamples of 45 g were used to prepare 1:1 soilwater extracts, and the chloride concentration of each aqueous extract was determined by constant potential oculometry (Haake Buchler Digital Chloridometer). Soil water pressure potential was monitored at each site. Tensiometers were placed at 0.15 m intervals to a depth of 1.5 m in the border plots not used for soil sampling. In addition, soil bulk density measurements were taken at 10 locations within the border plots.

A

17

hydraulically driven tube was used to obtain continuous soil cores that were sectioned for determination of density.

Mean bulk density at each

site is shown in Fig. 2. Background concentrations of soil chloride were determined from soil samples that were collected at the time the experimental sites were established. Background concentrations were relatively uniform with depth, and an average concentration of

•> 1.3

meq/L was measured at both upper and lower sites.

Transport Theory An equation that describes longitudinal dispersion of a solute within soil at variable water content is d(dC)

[1]

at

where 0 is the volumetric water content (cm°/cm^), C is solute concentration (meq/L), x is depth below the soil surface (cm), t is time (d), D is the hydrodynamlc dispersion coefficient (cm^/h), and q is the water flux density (cm^/cm^d).

Equation [1] assumes that the displacing

and displaced miscible fluids are of the same density and viscosity, and that the solute does not Interact with the soil. If depth, both D and q become constants. ac at

a^c ax^

0

is constant with

Equation [1] can then be written ac ax

[2]

where V - q/6

is defined as the average pore-water velocity (cm/d).

[3]

18

BULK 1.0

1.2

DENSITY 1.4

1.G

(g/cm^)

I.B

2.0

0

25

E u

I hQ_ U a

50

75

1 00

125

150

Figure 2.

Mean soil bulk density at upper (solid line) and lower sites (dotted line). Each point is the mean of 10 observations

19

For a pulse application of solute to the soil surface, the Initial and boundary conditions that are needed to solve Eq. [2] are written as C(x,0) - C^;

X >0

ac vC(0,t) - D —(0,t)

[4]

r C*; 0 < t 3 t(

10;

ac (oo.t) - 0; ÔX

[5] t > t_ t> 0

[6]

The solution of Eq. [2] subject to Eqs. [4] - [6] is (Lindstrom et al., 1967) (CQ - Cj) A(x,t);

0 < t 3 t^

C(x

[7]

I

Ci + (Co - Cj) A(x,t) - Cq A(x,t-to); t > t^

where A(x,t) - — erfc

x-vt 2(Dt)

2 1 r VX V tl - 2 + "5 + 15-j

».

(^) fvxl

exp

-(x-vt) 4Dt x+vt 2(Dt)^.

Steady-State Analysis Strictly speaking, Eq. [2] describes solute dispersion in a steadystate water flow; however, Cassel et al. (1975) and Wierenga (1977) have presented experimental evidence from studies with large soil columns that suggest Eq. [2] can be used to approximate Eq. [1], that is, Eq. [2] can be used to approximate solute movement when water flow is not steady. In this approach, unsteady surface fluxes of water are replaced

20

with an equivalent steady flux, and vertical variations of soil water content are averaged. This steady-state approach has been used with some success to describe field-scale solute movement (Bowman and Rice, 1986; Gish and Coffman, 1987; Parker and van Genuchten, 1984). The steady-state approach is used in analyzing the results of the present study. Parker and van Genuchten (1984) suggest some helpful notation to assist with the spatial and temporal averaging that is necessary in carrying out the steady-state analysis.

Following their convention, a

time-averaged flux, q*k , can be defined as

rt,m q(s) ds

[8]

0

where q(t) represents the water flux density that varies with time from t-Otot-tjj|, a maximum time of interest. It is assumed that q(t) is constant with depth. A transformed time, or flux-time (Gish and Coffman, 1987) can be computed from * t-

JS q(s) ds ; q

[9]

In the field experiments, spatial variability of infiltration resulted in different water flux densities at different spatial locations; however, to use Eq. [9] for a specific spatial location, it is not necessary to know q(t) at that location. Consider an areallyaveraged water flux density, , and let

p

represent the fraction of

applied irrigation water that infiltrates at a specific spatial location. The flux density at that location can be represented by the

21

product f'. Upon substitution of p for q(t) in Eq. [9],

p

cancels since it appears in both the numerator and denominator, giving the result rt ds

N m

E)

E)

G) E) El El El El

— (vin"rinu)i^tD

tu en m en 01

CD CD • •

m

01 01

01 01 01

PERCENT UNDER

Figure 11. Normal probability plots for v* (a.) and In v* (b.) for 64 measurements obtained at the upper site

40

O (3 Q Q m TT in u)

CD m in

m m

00 m

• •

m mm m en m

PERCENT UNDER

J I + I + *+ +++44444# I I — o j u i — r\j • • •

m

(S —

J 1 I 1_ E El El El 13 E) E) ru m \r m lû m

_l

El in m m

1

1

L

mcninmcn m m en t.

[2]

3C (oo,t) - 0 dx the solution of [1] is (Lapidus and Amundson, 1952) ' Ci + (Co - C^) A t.

C(x,t) -

[3]

where X + vt

X - vt

A(x,t) - - erfc

+ — exp(vx/D) erfc L2(Dt)^J

L2(Dt)^J

Recommendation to apply the analytical solution for a semi-infinite system to effluent data collected from a finite soil column was given by van Genuchten and Parker (1984) after a detailed investigation of boundary conditions that can be used to solve Eq. [1]. Dimensionless time and concentration variables were used to express the results of the steady-fluid-flow miscible displacement experiments. The dimensionless time variable T (pore volumes) is defined by T — vt/L in which L is the length of the soil column. The dimensionless concentration C (relative concentration) is defined by C - (C - Ci)/(CQ - Ci).

Equation [3] was fit to the breakthrough curve

data by using a non-linear least squares inversion method (Parker and

69

van Genuchten, 1984). The dimensionless column Peclet number (P - vL/D) will also be used. The Peclet number provides a measure of the relative Importance of advectlve and dispersive processes. After the saturated, steady-fluid-flow displacement experiments were completed, the remaining chloride was removed from each column by extensive leaching with CaSO^ solution. Two transport experiments were then conducted with each of the three columns. In each experiment, the initial moisture condition was a gravity-drained profile.

Before the

start of a given experiment, a column was saturated with 0.01 N CaSO^ solution as described earlier and then allowed to drain for a 12-h period with the lower boundary open to the atmosphere.

The volume of

solution drained from each column was recorded and used to calculate the drained porosity (fraction of total porosity which drained) and the initial soil water content, 0^. In the first of the two experiments, the chloride solution was applied at the soil surface as a pulse (pulse application experiment); 200 ml of 0.05 N CaCl2 solution was ponded on the soil surface (equivalent to a depth of 0.79 cm) to initiate the experiment. Once this solution had infiltrated completely, 0.01 N CaSO^ solution was immediately ponded on the soil surface by using a constant head device. As in the steady-fluid-flow miscible displacement experiments, effluent exiting the column was collected in fractions for the determination of chloride concentration. At the completion of the first experiment, CaCl2 remaining in the columns was removed by leaching with CaSO^ solution, and the columns were again saturated with CaSO^ solution and drained. In the second of the two experiments, 40 ml of a

70

more concentrated solution (0.25 E CaCl2) was dripped onto the soil surface with a plpet (drip application experiment). The solution was dispensed uniformly over the soil surface during a period of 6-8 mln, and at no time did ponding occur.

Fifteen minutes after the initiation

of the solution application, CaSO^ solution was ponded on the soil surface precisely as it was in the first experiment. Again, effluent was collected in fractions, and chloride concentration was determined. During both the pulse and drip application experiments, when solution was ponded on the soil surface, the water content In each column changed rapidly from

the initial gravity-drained condition, to a nearly

saturated condition. Three pore volumes of a 0.025% methylene blue dye solution were passed through each soil column after the pulse and drip application experiments were concluded. The dye solution was ponded on the soil surface after saturating and draining each column by the same method as described previously.

Each column was drained and then sectioned

horizontally in 5-cm Increments to allow visual characterization of the stained and unstained soil voids. The oven-dry weight of the soil from each column was obtained for the determination of soil bulk density and water content. The total porosity of each column was estimated by using values of bulk density and particle density (mean particle density was measured as 2.59 g/cm^).

71

RESULTS AND DISCUSSION Characterization of the Soil Material The measured values of drained porosity (Table 1) indicate that the three columns contained different macropore volumes. Bullock and Thomasson (1979) and Germann and Seven (1981) have used similar measures of drained soil water to assign a volume fraction to the macroporosity of undisturbed soil, but such measures depend arbitrarily upon the water potential used for drainage. For this reason, the drained porosities listed in Table 1 are intended only to rank the macroporosity of the three columns. By applying capillary theory, drainage of soil water at a particular tension is often used as a measure of the volume fraction of pores that are greater than a certain diameter. It is important to remember that drainage of pore sequences also depends on pore continuity. A large pore cannot drain unless It is part of a pore sequence into which air has entered under the applied tension.

Thus, in

addition to macropore volume, the drained porosities reflect, to some degree, the continuity of macropores. The large Kgg^ values (Table 1) indicate the presence of wellconnected sequences of large pores. Together, the Kg^t values and drained porosities indicate that there are distinct differences in the amount and continuity of macroporosity among columns A, B, and C. A summary of the voids observed in the three columns is presented in Table 2. Stained and unstained channels in two diameter classes were

72

Table 1. Dimensions and physical properties of the three soil columns used. Mean particle density was determined to be 2.59 g/cm* Column

A

B

C

Kgat (cm/hr)

58

28

12

Diameter (cm)

18

18

18

Length (cm)

33

33

35

Bulk Density (g/cm®)

1.32

1.29

1.32

Total Porosity

0.49

0.50

0.49

gg (cm®/cm')

0.43

0.45

0.42

7.9

5.8

3.1

0.39

0.42

0.40

Drained Porosity (%)* 6^ (cm®/cm')

^Percentage of total porosity that drained under the influence of gravity.

73

Table 2.

Depth (cm)

Description of voids for columns A, B, and C. Total number of channels counted (stained and unstained) at each depth are shown in two size classes. Number of stained channels are Included in parentheses. Total number of stained planar voids observed at each depth are also shown. Values in parentheses are the respective lengths (cm) of the stained planar voids observed. Stained planar voids appear in cross section as continuous stained lengths. Unstained planar voids were not detectable Channels (.5-2mm) (2-5mm)

Stained Planar Voids

Column  5 10 15 20 25 30

17(10) 18(12) 8(7) 4(2) 3(2) 3(2)

7(3) 9(4) 4(2) 0(0) 3(2) 12(0)

0 2(3,5) 2(3,7) 1(13) 1(10) 1(6)

Column B 5 10 15 20 25 30

19(6) 5(4) 19(11) 6(1) 3(1) 2(1)

9(1) 3(0) 15(6) 4(0) 3(0) 14(0)

0 1(7) 0 0 0 0

Column C 5 10 15 20 25 30

16(9) 12(6) 5(2) 5(4) 3(3) 4(3)

11(2) 9(4) 2(2) 4(3) 13(3) 20(16)

2(2,2) 1(9) 3(8.4.4) 1(12) 3(5,2,2) 0

74

recorded for each cross section. In addition, planar voids exposed to the dye were recorded by measuring the stained length of each planar void appearing in the cross sections.

Unstained planar voids were not

detectable. There was an abundance of 0.5-mm to 5-mm-diameter channels and planar voids, and a large proportion of each effectively transmitted dye solution.

Many of the channels and planar voids were continuous

between the sections examined, and connections between channels and planar voids were observed as well. Column B contained fewer stained channels and planes than did columns A and C. In fact, dye was transmitted through the lower half of column B in one continuous 2-mmdlameter channel.

Additional small patches of dye were observed at the

5 cm depth in each column, but beyond the 5 cm depth, the presence of dye was limited to the voids shown in Table 2. Solute breakthrough experiments were also used to characterize the soil. Figure 1 shows the results of the miscible displacement experiments conducted using steady, saturated fluid flow.

All three

columns yielded highly asymmetric breakthrough curves with early first breakthrough and extensive tailing at relative concentrations approaching unity. Similar breakthrough curves have been obtained for a non-reactive solute in undisturbed soil columns by Anderson and Bouma (1977a), Kanchanasut et al. (1978), White et al. (1986), and Seyfried and Rao (1987). Equation [3] has been used extensively in the analysis of breakthrough curves. This equation was fit to the data of Fig. 1 to provide a characterization of the soil in a framework that has been

Figure 1. Chloride breakthrough curves for saturated, steady-fluid-flow miscible displacement experiments with Columns A, B, and C. Points indicate measured data and the solid lines represent a least-squares fit of Eq. [3] to the measured data

76

I .0 Iw

COLUMN A

0.8

OBSERVED a

F

u z o (_) Ul >

0.E 0.4

en

w CK

f

FITTED

f

0.2

1

0.0 . 0

1

1

1.0

1

0.8

0.4

en _) w cz

0.2

1

3.0

4.

COLUMN B

-

.

0/ y

U z o o w >

1

2.0

0 0

lu



OBSERVED a FITTED

0.5



V 1

0.0 .0

1

1

1.0

1

1

2.0

3.0

4.

1.0 0

lu

0

\

0 .B

1

u o

u w > cc LJ œ

0.B

COLUMN C



OBSERVED o

I

y

FITTED



0.4

0.2

0.0 0.0

,

1

1.0

,

1

,

1

2.0

PORE VOLUME,

3.0

T

,

4.0

77

commonly used to characterize the solute transport properties of soil materials.

Equation [3] was fit to each breakthrough curve by setting

X - L and by adjusting D to obtain a best fit. The average pore-water velocity, v, was taken as the volumetric solution flux (cm^/cm^h) divided by 0^.

The solid lines in Fig. 1 show the fitted concentration

values and the resulting parameter values are shown in Table 3. The large values of D and small values of F indicate large dispersive solute flux relative to convective solute flux. It is emphasized that Eq. [3] has been applied to the data of Fig. 1 only to provide a characterization of the transport properties of this soil material. It is doubtful that physical significance can be attached to the estimated parameters. Difficulties in assigning appropriate boundary conditions to columns of finite length have been pointed out by van Genuchten and Parker (1984). This problem is particularly acute for columns with small P. There were also experimental difficulties encountered in establishing the desired solute boundary conditions in the presence of extremely large soil water fluxes. More important perhaps is that the column lengths, L, used in this study were likely much shorter than the minimum mixing length, L^, or Lagranglan length scale for solute dispersion in this soil material. case of L
V

»

50

100

I

1

150 »

0

50

100

150

0.0 I—Î-

0.5

I.0

IlJUl

IIJU2 M O

I .5 150

•P-

0

50

100

0.0

0.5

IIJU3

1.0

I .5

-

IIJU4

150

CHLORIDE CONCENTRATION (meq/L) 150

IIJU5

150

IIJUB M O Ul

0 X H

50

^—'

100 —

150

150

1—

CL u a

0.5

1.0

I .5

: I IJU7

IIJUB

CHLORIDE CONCENTRATION (meq/L) 150

150

0.0

0.0

0.5

0.5

I.0 IIJU9

IIJU 0 o

o\

I .5 150

I

150 0.0

CL U

0.5

a

I .0

I IJU

£ IJUl2

I .5

CHLORIDE CONCENTRATION (meq/L) 0

0

50

100

150 0.0

5

50

100

0

IIJU13

5

IIJU14

5 0

50

100

150

0

50

100

5

0

5

150

5

0

0

0

IIJU15

I IJUIB

150

CHLORIDE CONCENTRATION (meq/L) 100

150

150

0.0

0.0

0.5

0.5

I.0

IllJUl

IIIJU2 M O 00

I .5 150

T

150

fQ_ u

Q

IIIJU3

IIIJU4 I .5

CHLORIDE CONCENTRATION (meq/L) 0

50

100

150

0

0

50

100

150

5

0

0

IIIJU5

5

IIIJUB

5 0

50

100

0

150

50

100

0.0

5

0

5

IIIJU7

0

IIIJUB

150

CHLORIDE CONCENTRATION (meq/L)

IIIJU9

IIIJU10 M O

IllJUlI

ÎIIJU12

CHLORIDE CONCENTRATION (meq/L) 0

150 0.0

50

,

,—

100

150

0.5

I .0

IIIJU13

. IIIJU14 1 .5 0

150

50

100

0.0

0.5

IIIJU15

1 . 0

-

IIIJUIS

1 .5

150

CHLORIDE CONCENTRATION (meq/L) 0.0

0

50

100

150

0

50

100

150

0.0

IVJUl

IVJU2 5

0

50

100

150

0

0.0

0.0

5

5

0

5

IVJU3

0

5

50

100

IVJU4

150

CHLORIDE CONCENTRATION (meq/L) 100

150

150

0.0

0.0

0.5

0.5

I.0

IVJU5

I .5

I .0 IVJUB I .5

W

150 X

0.0

H Q_ u a

0.5

IVJU7

I.0

IVJUB I .5

CHLORIDE CONCENTRATION 0

150 0.0

50 —

Cmeq/L) 100 1

150

0.5

1 .0

IVJU9

f' IVJU10

I .5 150

150 0.0

0.5

IVJUlI i .5

1 .0

1 .5

IVJUl2

CHLORIDE CONCENTRATION 0.0

0

50

100

1—

;

150

(meq / L ) 100

150

0.5

I .0 •

)' I V J U 1 3

IVJU14

I .5

Kn

0 % h-

0.0

50 100 150 1 —'—1—'

CL U

Q

0.5

I. 0

I .5

IVJU15

IVJUIB

CHLORIDE CONCENTRATION (meq/L) 0

50

100

0

150

50

100

0.0

5

5

0

0

IJLl

5

IJL2

5 0

50

100

150 0.0

0

I JL3

I JL4

150

CHLORIDE CONCENTRATION (meq/L) 0

50

100

150

0

0.0

0

0.5

5

0

50

0

IJL5

5

100

150

IJLB

5 50

100

150

0

50

100

0.0

UJ

5

a

IJL7

0

5

IJLB

150

CHLORIDE CONCENTRATION (meq/L) 0

50 1

0.0

0.5

I .8

100

150

0.5

!

1 .0

-

IJL9

1 .5

X

150 0.0

IJL10

I .5 150

00

0

0.0

0.0

0.5

0.5

50

100

>—• II —

h-

CL W

Q

1.0

1.5

IJLl I

I. 0

I .5

-

IJL12

150

CHLORIDE CONCENTRATION

( m e q / L )

150 0.0

0.5

0.5

t .0

I .0

IJL13 I .5

IJL14 \o

1 .5 0

ZC

150

0.0

50

0.0

-,—.• l»

100 '

150 '

150

0.0

h-

a_

u Q

0.5

1.0

0.5

-

IJL15 I .5

I JLIB

CHLORIDE CONCENTRATION (meq/L) 0

150

50

100

150

0.0

0.5

IIJLl

50

100

IIJL3

150

150

IIJL4

CHLORIDE CONCENTRATION (meq/L) 50

100

150

0

50

100

150

0.0

0.5

M N> M

150

150

Q_ w

a

IIJLB

CHLORIDE CONCENTRATION (meq/L) 150

150 0.0

0.5

\ .0

IIJL9

IIJL10 M

1 .5

HI

150

ro to

0

50

100

0.0

0.0

CL Ul

0.5

a

I.0

IIJL12

IIJLlI

I .5

150

CHLORIDE CONCENTRATION (meq/L) 0

50

100 —1

150

150 0.0

J

0.5

0.5

I.0

I.0

IIJL13

IIJL14

,

X 0.0 0_ U Q

0 «

50 1—

100 J

150 —

0.5

I .0

I .5

to

1 .5

1. 5

W

0

50

100

0.0

0.5

IIJL15

I.0

I .5

IIJLIB

150

CHLORIDE CONCENTRATION (meq/L) 150

150 0.0

0.5

I-0

IIIJLl

IIIJL2 N3 •P-

I .5 0

150

50

100

0.0

CL

U a

0.5

IIIJL3 1.5 ^

I .0

IIIJL4 I .5

150

CHLORIDE CONCENTRATION (meq/L) 150 0.0

0.5

IIIJL5

1.0

IIIJLB ro

I .5

Ln

0

50

100

0.0

CL u a

0.5

niJL?

t .0

I .5

IIIJLB

150

CHLORIDE CONCENTRATION (meq/L) 0.0

0

50

100

150

150 0.0

0.5

0.5

1.0

I.0

IIIJL3

I -5

IÏIJL10 M

to

1.5 ^ 0

50

100

150

0

0.0

0.0

0.5

0.5

50 '

100 —t

CL u a

'

I.0

I .5

IIIJLl1

I.0

IIIJL12 I .5

150

CHLORIDE CONCENTRATION (meq/L) 0 0.0 1''*'"*^

50

100

1

1

150

0.5

50

100

1.0

IIIJL13 1 .5

IIIJL14

1.5 ^ 0

50

100

0

150

0.0

0.0

0.5

0.5

NS -J

50

100

I—

CL U a

1. 0

IIIJL15 1.5

150

0.5

1.0

X

0 0.0

1.0 IIIJLIB I .5

150

CHLORIDE CONCENTRATION 150

0.0

(meq / L ) 100

150

0.5

1.0

IVJLl

IVJL2 H»

I .5

ro

00

JL \-

Q_ u Q

150

0.0

0.5

1.0

IVJL3 1 .5

IVJL4 I .5

CHLORIDE CONCENTRATION (meq/L) 150 0.0

0.5

IVJLS

IVJLS to

VO

150

IVJL7

IVJLS

CHLORIDE CONCENTRATION (meq/L) 150 0.0

0.5

1.0

IVJL9

IVJL 0

I .5

X 0.0

W

1.5 ^ 0

50 ,

100 —

1

o

150

150 0.0

CL U

a

0.5

0.5

f-

I .0

1.0

IVJLlI I .5

IVJL 2 1 .5

CHLORIDE CONCENTRATION (meq/L) 150

0.0

0.5

1.0

IVJL13

IVJLM

I .5 0

50

100

0

150

0.0

0.0

0.5

0.5

50

100

Q_

w a

1.0

1.0

IVJL15 I -5

I -5

-

IVJLIG

150

Suggest Documents