Received: 24 May 2024 Accepted: 9 September 2024

DOI: 10.1002/vzj2.20381

Vadose Zone Journal

ORIGINAL ARTICLE

Algebraic expressions for convective solute transport in a

homogeneous soil below a surface barrier

Zhuanfang Fred Zhang

INTERA Incorporated, Richland,

Washington, USA Abstract
Engineered surface barriers are expected to be used to cover sites containing buried
Correspondence
Zhuanfang Fred Zhang, INTERA, waste radionuclides after site closure. Understanding the transport of contaminants
Incorporated, Richland, WA 99354 USA. in the soil below a surface barrier is of importance in order to protect the underlying
Email: [email protected]
groundwater. The travel velocity of these contaminants is dependent on numerous
Assigned to Associate Editor Joaquin factors, such as the thickness, properties, the pre-barrier hydrologic condition of the
Jimenez-Martinez. vadose zone, and the initial depths and properties of the contaminants. Although
solute transport can be modeled with numerical modeling using site-specific data, by
nature a numerical method does not reveal the relationship between inputs and out-
puts. This work develops algebraic expressions of the velocities of convective solute
transport in the vadose zone below a surface barrier, time lag, solute safe depth, zero-
protection depth of a surface barrier, and solute travel time to the groundwater. This
development shows that solute convection in the soil with a surface barrier consists
of up to three stages: constant high velocity, decreasing velocity, and constant low
velocity. Solute transport based on the algebraic expressions was verified against
numerical simulations. The algebraic expressions may be used to assess the expected
impacts of a surface barrier to the convective transport of contaminants in the vadose
zone.

Plain Language Summary

Engineered surface barriers are like gigantic umbrellas above a waste site to intercept
precipitation so the rainfall will not enter the buried hazardous materials. Often the
movement of contaminants in the soil needs to be calculated using supercomputers.
This article develops some simple mathematical formulas for quick calculation. A
few formulas are established to determine how fast a contaminant moves, at what
depth a contaminant will be safe, at what depth a surface barrier has no impact to
contaminant movement, and how long it takes for a contaminant to move to the
groundwater. The formulas were verified against results from a supercomputer. These

Abbreviations: Bf, Backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sand; RMSE, root of mean squared error.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original
work is properly cited.
© 2024 The Author(s). Vadose Zone Journal published by Wiley Periodicals LLC on behalf of Soil Science Society of America.

Vadose Zone J. 2024;e20381. wileyonlinelibrary.com/journal/vzj2 1 of 15

https://doi.org/10.1002/vzj2.20381
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 of 15 Vadose Zone Journal ZHANG

formulas are useful to assess the impact of a surface barrier on the movement of the
buried hazardous materials.

1 INTRODUCTION may be effective in reducing the mobility of the underlying

contaminants.
Engineered surface barriers are expected to be used to cover The purpose of this article is to develop algebraic expres-
sites containing buried waste radionuclides after site clo- sions for estimating the convective transport in the vadose
sure, for example, as in the Hanford Site near Richland, WA. zone beneath a surface barrier based on the Brooks and Corey
Sophisticated surface barriers have been designed and demon- (1966) hydraulic retention model and the Burdine (1953) rela-
strated with an expected life of performance over 1000 years tive permeability model. These expressions are tested against
(Zhang et al., 2017). Regardless of how sophisticated a surface numerical simulations for four types of soils under two pre-
barrier is, it has its limitations. It usually takes a long time for barrier recharge rates. Both the terms solute and contaminant
the connate soil moisture to drain (Zhang et al., 2012). Zhang will be used depending on the perspective.
(2020) developed algebraic expressions to estimate drainage
velocities and barrier impact depths after the emplacement of
a surface barrier. Four impact depth terms are used to con- 2 MATHEMATICAL EXPRESSION
vey barrier impact: near-zero, 50%, average, and full-impact
depths. This section first summarizes the impact depths of a sur-
For a given waste site, the underground contaminants may face barrier on soil-moisture flow based on Zhang (2020).
reside at different depths. Hence, the slow drainage process Then, algebraic expressions are derived to quantify convec-
and large transition zone (from the near-zero to full-impact tive solute transport in the soil beneath a surface barrier. It is
depths as described in Zhang [2020]) have a few implications assumed that the soil is homogeneous and soil-moisture flow
on the transport of contaminants in the underlying vadose at barrier emplacement is at the steady-state condition cor-
zone: (1) The surface barrier will not deter the migration of responding to a constant pre-barrier recharge rate (qi ). This
contaminants residing in the vadose zone until after a time assumption generally can be met if a site is operated under
lag, (2) a contaminant resides above the safe depth is not the same conditions for a very long time. The surface barrier
expected to arrive at the groundwater before a target time, (3) has infinite size and a finite constant percolation rate (qf ).
the convection of contaminants residing below a certain depth
in the vadose zone will not be affected by the deployment of
a surface barrier before the contaminants enter the underly- 2.1 Water flow beneath a surface barrier
ing groundwater, and (4) the travel time to groundwater of the
contaminant residing above the zero-protection depth will be Before the emplacement of a surface barrier, the soil-water
extended by the use of a surface barrier. content (θ) and flux rate (q) are constant (Figure 1a). After
The time lag, solute safe depth, barrier protection depth, the emplacement of a surface barrier, the connate water in
and travel time to groundwater vary from site to site and can the soil starts draining. Because soil water resides in the
be estimated with numerical simulations. A disadvantage of pores of different size and moves at rates that are dras-
the numerical method is that the mathematical relationship tically different (e.g., by orders of magnitude), soil water
between inputs and outputs is usually given in complex dif- in larger pores drains much faster than that in smaller
ferential equations due to the nature of numerical simulation. pores. Zhang (2020) divided the soil profile below a sur-
Thus, based on the results of numerical simulations, it is usu- face barrier into three zones: the relatively “equilibrium
ally difficult to separate the characteristics that are common zone” at the top beneath the barrier, the relatively “unaf-
for all sites from those that are site specific. As a result, the fected zone” at the bottom, and a “transition zone” in the
findings from numerical simulations for one site may not be middle (Figure 1b). The lower end of the transition zone is
transferable to sites with different soil and climate conditions. referred to as the drainage front, and the upper end as the
If these variables are estimated with algebraic expressions, drainage tail of the transition zone. The drainage front is con-
the common characteristics can be easily revealed, as will sidered as the near-zero-impact depth of the surface barrier.
be shown in this investigation. The disadvantage of alge- As time (t) goes on, the depth of the drainage front becomes
braic expressions is their validity under specific conditions, larger.
for example, uniform soil and constant pre- and post-barrier Zhang (2020) defined the pore size-specific water veloc-
recharge rates. Despite these limitations, algebraic expres- ity (vw ) as the derivative of q versus θ under unit gradient
sions are useful for screening of sites at which a surface barrier condition, that is, 𝑣𝑤 = 𝑑𝑞∕𝑑𝜃. When the Brooks and Corey
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 3 of 15

(1966) soil-moisture retention and the Burdine (1953) relative

permeability models are used, vw can be calculated with Core Ideas
( )2+2/𝜆 ∙ Convective solute transport below a surface barrier
𝑞i (3 + 2∕𝜆) θ − θr
𝑣w = consists of three stages.
θi − θr θi − θr
∙ Algebraic expressions are developed to quan-
( )(2+2∕𝜆)∕(3+2∕𝜆)
𝑞 (3 + 2∕𝜆) 𝑞 tify the impacts of a surface barrier on solute
= i (1)
θi − θr 𝑞i convection.
∙ The impacts are reflected by the time lag, solute
where λ is the pore-size distribution parameter, θs and θr are safe depth, zero-protection depth, and travel veloc-
saturated and residual water content, respectively, qi is the pre- ity.
barrier recharge rate, and θi is the pre-barrier soil-moisture ∙ The algebraic expressions are verified against
content. numerical simulations.
Zhang (2020) characterized the impact of a surface bar-
rier with the velocities and depths of drainage front, 50% flux
reduction, average drainage, and drainage tail. The velocity of
𝑞f (3 + 2∕𝜆)
the drainage front (vw0 ) and drainage tail (vw1 ) are described 𝑣w1 = (3)
𝜃f − 𝜃r
by:
Zhang (2020; Equation 12) provided the expression that can
𝑞i (3 + 2∕𝜆)
𝑣w0 = (2) determine the water content distribution:
𝜃i − 𝜃r

F I G U R E 1 Schematics showing the drainage process beneath a surface barrier of infinite size. Definition of symbols GW: groundwater fringe;
z: depth; L: thickness of the vadose zone; Li : distance to groundwater fringe; t: time; q: water flux rate; θ: soil-water content. subscripts i: the initial
(pre-barrier) values (at t = 0); f: final (post-barrier) values; 1: at the end of stage 1, and 2: at the end of stage 2.
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 of 15 Vadose Zone Journal ZHANG

⎧ θf if 𝑧 ≤ 𝑣w1 𝑡 (equilibrium zone)

⎪ ( ) ( 𝑧 )λ∕(2+2λ)
θ (𝑧, 𝑡) = ⎨θr + θi − θr if 𝑣w1 𝑡 < 𝑧 < 𝑣w0 𝑡 (transition zone) (4)
𝑣w0 𝑡
⎪
⎩ θi if 𝑧 ≥ 𝑣w0 𝑡 (unaf fectived zone)

where z is the depth below a surface barrier and t is the time convection is also referred to as solute travel, as in solute
since the emplacement of a surface barrier. The three expres- travel velocity and solute travel time. Solute convection in
sions in (4) are for the equilibrium, transition, and unaffected the vadose zone below a surface barrier can be divided into
zones, respectively. The soil water flux at any time and depth three stages: constant high velocity, decreasing velocity, and
corresponding to the three zones is obtained by substituting v constant low velocity, within the corresponding unaffected,
= z/t into (1) and rearranging: transition, and equilibrium zones (Figure 1), respectively.

⎧ 𝑞f if 𝑧 ≤ 𝑣w1 𝑡 (equilibrium zone)

⎪ ( 𝑧 )(2+3𝜆)∕(2+2𝜆)
𝑞 (𝑧, 𝑡) = ⎨𝑞i if 𝑣w1 𝑡 < 𝑧 < 𝑣w0 𝑡 (transition zone) (5)
𝑣w0 𝑡
⎪
⎩ 𝑞i Otherwise 𝑧 ≥ 𝑣w0 𝑡 (unaf fectived zone)

2.2 Convective solute transport beneath a 2.2.1 Stage 1: Convective transport in the
surface barrier unaffected zone

The above analysis generally applies to semi-infinite soil. In During this stage, solute convection is unaffected by the sur-
the following analysis, it is assumed that the bottom of a face barrier as shown in Figure 1a because the drainage front
domain is the capillary fringe of the groundwater (to be con- has not reached the depth of the soil where the solute resides.
cise, it will be referred to as the groundwater). A question to be At this stage, solute travel velocity (i.e., pore-water velocity,
answered is whether a surface barrier will ever, and if so, when vsi ) is the same as that of the pre-barrier condition:
it will start to deter the migration of conservative or sportive
contaminants in the soil beneath the surface barrier because 𝑣(1)
si
= 𝑞i ∕𝑅𝜃i (6)
contaminants usually reside in some depth of the vadose zone
and, furthermore, they move with the draining water. Another where R = 1+ρb kd /θ is the retardation factor, with ρb being
question is how a surface barrier affects the travel velocity of dry bulk density [M L−3 ] and kd the distribution (or sorption)
contaminants. coefficient [L3 M−1 ]. When R = 1 the solute is conserva-
The movement of solute with water is referred to as convec- tive, and when R > 1 the solute is sportive. The subscript “s”
tion or advection. Because dissolved solutes move in a passive denotes for solute, and “i” denotes the initial condition. The
fashion, convective transport can be readily quantified when superscript “1” denotes stage 1.
the aqueous flux is known. In the following, the development Comparing Equation (6) with (2) gives the ratio (rvi ) of the
focuses on convective solute transport in an unconsolidated drainage front velocity to the solute convection velocity:
porous medium below a surface barrier. The surface barrier
also deters the penetration of plant roots through its bottom, 𝑣w0 𝑅𝜃i (3 + 2∕𝜆)
𝑟vi = = (7)
for example, by including an asphaltic concrete layer. In the 𝑣si 𝜃i − 𝜃r
following, algebraic expressions are derived for convective
solute transport in a homogeneous soil with and without a Equation (7) indicates that the drainage front velocity is at
surface barrier. In the derivation, solute dispersion is not con- least three times the solute travel velocity for the worst-case
sidered and the location of the solute is denoted by the center scenario (i.e., when θr = 0, λ → ∞, and R = 1). For typical
of solute mass. It is beyond the scope of this manuscript to soils at the Hanford Site, as will be shown later, vw0 is about
predict the spatial distribution of a contaminant plume. Solute 7 to 11 times of 𝑣si for a conservative solute. In other words,
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 5 of 15

the drainage front moves at a velocity 7 to 11 times the solute Because variables z and t have non-integer exponents and
travel velocity. The drainage water will catch up to the solute exist in both the numerator and denominator of the right-hand
plume as time goes on and then starts to deter the migration side of Equation (13), an explicit expression for z as a function
of the plume. This indicates that a surface barrier will, after a of t, or vice versa, cannot be obtained. In order to determine
time lag, start hindering solute convection in the vadose zone solute depth at any time, here an approximate expression is
beneath a surface barrier. According to Equation (6), the depth sought after assuming θr is zero. This assumption leads to an
of peak concentration without the impact of a surface barrier overestimation of 𝑣(2)
s . As a compensation, a correction factor
is b (<1) is introduced. The approximate solute travel velocity,
𝑣(2)
sa , is then obtained as follows:
s = 𝑧si + 𝑣si 𝑡 = 𝑧i + 𝑞i 𝑡∕𝑅𝜃i
𝑧(1) (8)
𝑏𝑣si 𝑧
𝑣(2)
sa (𝑧, 𝑡) = 𝑣w0 𝑡
if 𝑡1 ≤ 𝑡 < 𝑡2 , 𝑧1 ≤ 𝑧 < 𝑧2 (14)
The time needed for the center of contaminant mass to
arrive at the groundwater without the impact of a surface Parameter b is probably some kind of function of θr . It
barrier, 𝑡(1)
sgw , is appears that a constant b of 0.8 works reasonably well for the
soils used in this study. The accuracy of (14) will be evalu-
( )
𝑡(1)
sgw = 𝐿i ∕ 𝑣si = 𝐿 − 𝑧i 𝑅𝜃i ∕𝑞i (9) ated against numerical simulations of solute transport below
a surface under the same conditions. The solute travel velocity
where L is the total thickness of the vadose zone, Li is the dis- in the porous medium at any time and depth, by definition, is
tance of the center of solute to the groundwater at time zero dz/dt. Replacing the left-hand side of (14) by dz/dt and con-
(time of barrier emplacement), and subscript “gw” denotes ducting integration over z and t produces the expressions of
the groundwater. Equation (9) reveals that 𝑡(1)sgw is linearly pro- solute depth at time 𝑡 [𝑧(2)
s ] and solute travel time to depth z
(2)
portional to Li , meaning that if the solute resides farther away [𝑡𝑠 ]:
from the groundwater, it will take a proportionally longer time
( )𝑏𝑣si ∕𝑣w0
for the solute to reach there. The time that stage 1 ends (t1 ) and 𝑡
𝑧(2)
s (𝑡) = 𝑧1 if 𝑡1 < 𝑡 < 𝑡2 (15)
the corresponding solute depth (z1 ) can be calculated by: 𝑡1
𝑧i
𝑡1 = (10)
𝑣w0 − 𝑣si ( )𝑣w0 ∕𝑏𝑣si
𝑧
𝑡(2)
s (𝑧) = 𝑡1 if 𝑧1 < 𝑧 < 𝑧2 (16)
𝑧1
𝑣w0 𝑧i
𝑧1 = 𝑣w0 𝑡1 = (11) Equation (16) shows the following:
𝑣w0 − 𝑣si

1. Once the barrier starts to slow down solute convection

2.2.2 Stage 2: Convective transport in the after the time lag, the time needed for the solute to travel to
transition zone a certain depth is proportional to the depth ratio of z/z1 in
a power function (rather than a linear function as in stage
This stage starts after the time lag t1 (the solute is at z1 ), when 1).
the drainage front arrives at the center of solute mass, and ends 2. A smaller z1 (because of a smaller zi ) would significantly
at t2 (the corresponding solute depth is z2 ), when the drainage prolong solute travel time to the same depth.
tail arrives. Solute travel velocity decreases with time during 3. The ratio 𝑣w0 ∕𝑣si at the exponent of (16) is dependent on
this stage. Solute travel velocity (vs ) in this stage varies with the soil properties and pre-barrier recharge rate.
time:
Substituting zs in (15) for z in (14) yields the travel velocity
𝑣(2)
s (𝑧, 𝑡) = 𝑞 (𝑧, 𝑡) ∕𝑅𝜃 (𝑧, 𝑡) (12) of a given solute at time t during stage 2:

( ) 𝑏𝑣si
Substituting Equations (4) and (5) into (12) yields the solute 𝑣(2)
sa (𝑧, 𝑡) =
𝑏𝑣si 𝑧1 𝑡 𝑣w0
if 𝑡1 < 𝑡 < 𝑡2 , 𝑧1 < 𝑧 < 𝑧2
travel time for this stage: 𝑣w0 𝑡 𝑡1
(17)
( )(2+3𝜆)∕(2+2𝜆) In order to know the depth of the solute at the final equi-
𝑞i 𝑣 𝑧 𝑡
𝑣(2)
s (𝑧, 𝑡) = [ w0
( )𝜆∕(2+2𝜆) ] , librium stage, the time (t2 ) and depth (z2 ) at which stage 2
𝑅 θr +(θi −θr ) 𝑣 𝑧 𝑡 (13) ends need to be known. They are also the time and depth
w0
for 𝑡1 < 𝑡 < 𝑡2 , 𝑧1 < 𝑧 < 𝑧2 at which stage 3 starts. Once all the soil water in the pores
is larger than those corresponding to θf above the depth at
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 of 15 Vadose Zone Journal ZHANG

( )
which a solute resides drains out, solute convection is con- ⎧ 𝑧 − 𝑧i ∕𝑣si if 𝑧i ≤ 𝑧 ≤ 𝑧1 (Stage 1)
trolled by the percolation water from the surface barrier. After ⎪ ( )𝑣w0 ∕𝑏𝑣si
𝑡s (𝑧) = ⎨ 𝑡1 𝑧 if 𝑧1 < 𝑧 < 𝑧2 (Stage 2)
the drainage tail has reached depth z2 at time t2 , the solute will ⎪ (𝑧1 )
travel at a constant final velocity (vsf ). At t2 and z2 , according ⎩𝑡2 + 𝑧 − 𝑧2 ∕𝑣sf if 𝑧 ≥ 𝑧2 (Stage 3)
to Equation (14), the final convection velocity is GIVEN AS: (25)

𝑏𝑣si 𝑧2 ⎧ 𝑧i + 𝑣si 𝑡 if 𝑡 ≤ 𝑡1 (Stage 1)

𝑣sf = (18) ⎪ ( 𝑡 )𝑏𝑣si ∕𝑣w0
𝑣w0 𝑡2 𝑧s (𝑡) = ⎨ 𝑧1 if 𝑡1 < 𝑡 < 𝑡2 (Stage 2) (26)
𝑡1 ( )
⎪
Substituting Equations (15) to (18) for z2 produces the ⎩𝑧2 + 𝑣sf 𝑡 − 𝑡2 if 𝑡 ≥ 𝑡2 (Stage 3)
expression for t2 :
It is noted that Equations (25) and (26) do not apply any
( )𝑣w0 ∕(𝑣w0 −𝑏𝑣si ) more once the solute enters the groundwater. For an imperme-
𝑏𝑣si 𝑧1 ( )−𝑏𝑣si ∕(𝑣w0 −𝑏𝑣si ) able surface barrier (i.e., qf = 0), t2 and z2 approach infinity.
𝑡2 = 𝑡1 (19)
𝑣sf 𝑣w0 Hence, the solute convection will never enter stage 3, and
expressions of Equations (25) and (26) do not apply for an
Substituting Equations (16) to (18) for t2 produces the
impermeable surface barrier. The following steps may be used
expression for z2 :
to calculate solute travel time, depth, and/or velocity:
( )𝑏𝑣si ∕(𝑣w0 −𝑏𝑣si )
𝑏𝑣si ( )𝑣w0 ∕(𝑣w0 −𝑏𝑣si ) 1. Collect the hydraulic parameters, vadose zone thickness,
𝑧2 = 𝑧1 (20)
𝑣sf 𝑣w0 𝑡1 initial depth of solute, and solute retardation factor.
2. Calculate vw0 using (2), vsi with (6), and vsf with (21).
Both t2 and z2 become larger when vsf is smaller. When a
3. Calculate t1 with (10), z1 with (11), t2 with (19), and z2
drainage rate from the surface barrier approaches zero, vsf will
with (20).
approach zero, and hence both z2 and t2 approach infinity.
4. Calculate vs with (24), ts with (25), and zs with (26).

2.2.3 Stage 3: Convective transport in the 2.3 Other characteristics of solute

equilibrium zone transport below a surface barrier
After the drainage tail reaches the center of solute mass, the
A few other characteristics may also be used to understand
solute will travel at a final constant travel velocity:
the impact of a surface barrier on contaminant transport in
the vadose zone below the surface barrier.
𝑣sf = 𝑞f ∕𝑅𝜃f for 𝑡 ≥ 𝑡2 , 𝑧 ≥ 𝑧2 (21)

Solute travel time [𝑡(3) (3)

s ] and depth at stage 3 [𝑧s ] starts at 2.3.1 Time lag of a surface barrier
t2 and z2 , respectively, and can be calculated with:
( ) The time lag (tL ) of the surface barrier is defined as the dura-
𝑡(3)
s (𝑧) = 𝑡2 + 𝑧 − 𝑧2 𝑅∕𝑣sf , 𝑧 ≥ 𝑧2 (22)
tion from barrier emplacement to the time the surface barrier
starts to deter the migration of a contaminant residing in the
( ) vadose zone. At the time a surface barrier is deployed, the con-
𝑞f 𝑡 − 𝑡2 taminant in the underlying soil migrates at the initial velocity
𝑧(3)
s (𝑡) = 𝑧2 + , 𝑡 ≥ 𝑡2 (23)
𝑅𝜃f 𝑣si (Equation 6), while the drainage front moves at a velocity
of vw0 (Equation 2) that is at least three times larger than 𝑣si .
The expressions for solute travel velocity, time, and depth
It is the difference between the two velocities that determines
for all the three stages are summarized below with proper
the time lag of a surface barrier. The time lag of a surface
conditions included:
barrier needed for the drainage front to reach the center of
contaminant mass is the end of stage 1 (Equation 10), that
⎧ 𝑣si if 𝑡 ≤ 𝑡1 (Stage 1)
⎪( 𝑏𝑣 𝑧 ) ( ) 𝑣𝑏𝑣si is, tL = t1 . This means that the contaminant transport will be
𝑣s (𝑡) = ⎨ si 1 𝑡 w0
if 𝑡1 < 𝑡 < 𝑡2 (Stage 2) (24) unaffected by the surface barrier until after time tL of bar-
⎪ 𝑣w0 𝑡 𝑡1
rier emplacement. Equation (10) indicates that contaminant
⎩ 𝑣sf if 𝑡 ≥ 𝑡2 (Stage 3)
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 7 of 15

residing at a larger depth will have a larger time lag. For a tional to the difference between the drainage front velocity and
solute that initially resides at the ground surface, there is no solute travel velocity and the thickness of the vadose zone; it is
time lag and hence tL = 0. inversely proportional to the target time, meaning larger target
time will need the contaminant to reside in a smaller depth in
order that the contaminant will not arrive at the groundwater
2.3.2 Solute travel time to groundwater before the target time.
Equations (25) can also be used to estimate the travel time to
groundwater (tsgw ) when z is replaced by L: 2.3.4 Zero-protection depth of a surface
( ) barrier
⎧ 𝐿 − 𝑧𝑖 ∕𝑣si if 𝑧𝑖 ≤ 𝐿 ≤ 𝑧1 (Stage 1)
⎪ ( )𝑣w0 ∕𝑏𝑣si
𝑡sgw = ⎨ 𝑡1 𝐿 if 𝑧1 < 𝐿 < 𝑧2 (Stage 2) (27) Suppose a contaminant resides at a certain depth at the time
⎪ (𝑧1 ) of surface barrier deployment. The zero-protection depth (zp0 )
⎩𝑡2 + 𝐿 − 𝑧2 ∕𝑣sf if 𝐿 ≥ 𝑧2 (Stage 3)
is defined as the initial solute depth, below which the convec-
For a solute initially residing at the ground surface (zi = 0), tion of residing solute is not affected by the surface barrier
its convection will be in stage 3 all the time because t1 = t2 = 0. before the solute enters groundwater. The time needed for the
This solute will have the largest travel time to the groundwater drainage front to reach the groundwater is 𝑡wg = L/vw0 . The
(tsgw max ), which can be calculated with Equation (27): time needed for a contaminant to reach the groundwater with-
out the impact of a surface barrier is 𝑡(1)sg = 𝐿i ∕𝑣si . For the

𝑡max surface barrier to be effective, the drainage front must arrive at

sgw = 𝐿∕𝑣sf (28)
the groundwater no later than the contaminant, namely, 𝑡wg ≤
The effectiveness of a surface barrier can be evaluated 𝑡(1)
sg , and hence L/vw0 ≤ 𝐿i ∕𝑣si . The critical distance (𝐿c ) of
with the solute travel time-delay factor (ft ) defined as the contaminant to groundwater for the barrier to be effective is
solute travel time in the vadose zone with a surface barrier given as:
(Equation 27) to that without the impact of a surface barrier
Equation (9): 𝐿c = 𝐿𝑣si ∕𝑣w0 (31)

𝑓t = 𝑡sgw ∕𝑡(1)
sgw (29) Letting the zero-protection depth zp0 = L − Lc and substi-
tuting Equation (2) for vw0 and (6) for vsi in Equation (31)
The solute travel time-delay factor has the minimum value will produce the zero-protection depth of a surface barrier on
of one, which means no impact on solute convection by a sur- contaminant convection:
face barrier. Larger ft value means larger effect of a surface ( )
𝐿 𝜃
barrier on deterring the convective solute transport. 𝑧p0 = 𝐿 − 1− r (32)
𝑅 (3 + 2∕𝜆) 𝜃i

2.3.3 Solute safe depth This means that the convection of a contaminant resid-
ing below the zero-protection depth at the time of barrier
It is of interest to know at what depth a contaminant will not emplacement is not expected to be deterred by the surface bar-
pose a risk to the underlying groundwater during the life of a rier at all before the contaminant enters the groundwater. The
surface barrier. Solute safe depth [zsi (s) ] of a surface barrier is relative zero-protection depth (zpr0 ) is defined as the ratio of
defined as the initial solute depth, above which the center of zp0 to the thickness of the vadose zone, that is, zpr0 = zp0 /L.
the residing solute will not enter the groundwater before a tar- Hence,
get time (tg ), for example, the design life of a surface barrier. ( )
1 𝜃r
For such a contaminant not to enter the groundwater before the 𝑧pr0 =1− 1− (33)
target time, its transport is expected to be primarily in stage 𝑅 (3 + 2∕𝜆) 𝜃i
2. Substituting Equations (10) and (11) into (27), using tg for
tsgw and zsi (s) for zsi , and rearranging produce: 3 MATERIALS AND METHODS
( ) a
(𝑠) ( ) −1 𝐿 a−1 To verify the derived algebraic expressions, calculations were
𝑧si = 𝑣w0 − 𝑣si 𝑡g 𝑎−1 (30) conducted for flow and solute transport in typical soils at
𝑣w0
the Hanford Site beneath a surface barrier for 1000 years. It
where a = 𝑣w0 ∕𝑏𝑣si , and the superscript “s” denotes the safe was assumed that the vadose zone was homogeneous and the
depth. Equation (30) indicates that solute safe depth is propor- groundwater was 100 m below ground surface. The vadose
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 15 Vadose Zone Journal ZHANG

TA B L E 1 Hydraulic properties (after Table 4.5 of Last et al., 2006).

Soil class
Parametera Bf Hss Hfs Hcs
α (m−1 ) 1.9 0.8 2.7 6.1
n (−) 1.4 1.915 2.168 2.031
λ=n−1 0.4 0.915 1.168 1.031
θr 0.03 0.072 0.032 0.027
θs 0.262 0.445 0.379 0.349
Ks (m s−1 ) 5.98E-06 8.58E-07 3.74E-06 2.27E-05
a
α, inverse of the air-entry pressure; n and λ, pore-size distribution parameters; θr , residual water content; θs , saturated water content; Ks , saturated hydraulic conductivity.
Abbreviations: Bf, backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sand.

zone was at steady state under constant pre-barrier recharge capillary fringe. The initial conditions at the time of barrier
rates of 30 or 100 mm year−1 . The final drainage rate from emplacement and the final conditions are given in Table 2.
the surface barrier was assumed to be 0.5 mm year−1 , which
was a performance criterion for the demonstrative prototype
Hanford barrier (Zhang et al., 2017). Four types of sediments 4 RESULTS AND DISCUSSION
(Table 1), that is, Backfill, Hanford silt sand (Hss), Hanford
fine sand (Hfs), and Hanford coarse sand (Hcs), typical to the The algebraic expressions are evaluated by comparing the
Hanford Site were selected. center of solute mass determined by the algebraic expressions
Conservative solutes resided at the initial depths of 40, 50, against the results from numerical simulation for the scenar-
60, 70, and 80 m at the time of surface barrier emplacement ios without and with a surface barrier. The comparison was
and were named as “Szz,” where zz denotes the initial depth. conducted for five solutes at different depths of four types of
For example, S40 is the solute initially resided at the 40-m sediments under two pre-barrier hydraulic conditions. After
depth. The convection of the above five solutes was deter- that, the characteristics of convective solute transport below
mined using the algebraic expressions, while the transport, a surface barrier are presented based on analyses using the
including both convection and dispersion, of these solutes was developed algebraic expressions.
calculated with numerical simulation using the Brooks and
Corey (1966) soil-moisture retention and Burdine (1953) rela-
tive hydraulic conductivity models. For each of the scenarios, 4.1 Evaluation of the algebraic expressions
two cases were conducted, one without and one with a surface
barrier. In the numerical simulations, the domain was 100-m Figure 2 compares the depths of the center of solute mass
deep and discretized into 1000 cells with the vertical spac- determined with the algebraic expression (Equation 8) against
ing of 0.1 m and a cross-section of 1 × 1 m2 . Each of the those with numerical simulation for the five solutes in two
solutes uniformly filled a cell (at the concentration of 1 kg per selected types of sediments under two pre-barrier recharge
m3 of bulk soil) at the desired depth and has longitudinal dis- rates without a surface barrier. The predicted solute depths
persivity of 1.0 cm. Additional solutes at the initial depths of by the algebraic expressions are nearly the same as those by
0, 10, 20, 30, and 90 m were included only in the analytical numerical simulation for all the cases.
calculation. Numerical simulations were carried out using the Figure 3 compares the depth of the center of solute mass
subsurface transport over multiple phases simulator (White determined with the algebraic expression (Equation 26) and
et al., 2015). numerical simulation for two selected sediments with a sur-
The initial hydraulic condition was the steady-state condi- face barrier. The predicted depths are nearly the same with
tion corresponding to the pre-barrier recharge rate (i.e., 30 or the differences no >1.5% for all four types of sediments and
100 mm year−1 ) for the sediments. Five solutes with a unit two pre-barrier recharge rates. Relative to the results from the
concentration were placed in the one cell at the depths of 40, numerical method, the accuracy of solute depth calculated
50, 60, 70, and 80 m, respectively. The simulations started with the algebraic expressions is quantified with the root of
at the time of barrier emplacement. After barrier emplace- mean squared error (RMSE, Table 3). For the scenarios with
ment, the top boundary condition was set to the post-barrier a surface barrier, the average RMSE of solute depth over the
recharge rate of 0.5 mm year−1 and the bottom boundary con- four types of sediments was 0.54 m under the 30 m year−1 pre-
dition was unit gradient to mimic the top of the groundwater barrier recharge rate and was 1.18 m under the 100 m year−1
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 9 of 15

TA B L E 2 Initial conditions at the time of barrier emplacement and the final conditions.

Conditions Variables Bf Hss Hfs Hcs Average

Initial condition at θi (m3 m−3 ) 0.108 0.172 0.092 0.069 0.110
30 mm year−1 ψi (m) −8.10 −5.25 −1.67 −1.19 −3.03
Initial condition at θi (m3 m−3 ) 0.120 0.199 0.109 0.080 0.127
100 mm year−1 ψi (m) −5.56 −4.07 −1.34 −0.94 −2.31
Final condition at θf (m3 m−3 ) 0.077 0.118 0.057 0.045 0.074
0.5 mm year−1 ψf (m) −29.11 −12.43 −3.51 −2.65 −7.72

Abbreviations: Bf, backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sand.

F I G U R E 2 The depth of the center of solute mass in the soil without a surface barrier was determined with numerical simulation and the
algebraic expression (Equation 8) for two types of sediments. Each line is terminated at 100 years or when the solute enters the groundwater. Hcs,
Hanford coarse sand; Hss, Hanford silt sand.

recharge rate. The error became larger when the pre-barrier face barrier, such as the drainage front velocity, three stages
recharge rate was larger. The errors are considered small rel- of solute convection, time lag, solute safe depth, and zero-
ative to the large (i.e., 100 m) thickness of the vadose zone. protection depth of a surface barrier, and solute travel time
The relatively small error between the results indicates good to the underlying groundwater.
accuracy of the algebraic expression.

4.2.1 Drainage front velocity and solute travel

4.2 Characteristics of convective solute velocity
transport beneath a surface barrier
After the emplacement of a surface barrier, the soil beneath
The algebraic expressions can reveal the characteristics of the barrier starts to drain. For all sediments, the drainage
convective solute transport after the emplacement of a sur- front velocity is much larger than the pre-barrier solute travel
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 of 15 Vadose Zone Journal ZHANG

F I G U R E 3 The depth of the center of solute mass in the soil with a surface barrier determined with numerical simulation and the algebraic
expression (Equation 26) for two selected sediments. Hcs, Hanford coarse sand; Hss, Hanford silt sand.

TA B L E 3 Root of mean squared error of solute depth (m).

30 mm year−1 100 mm year−1

Soil No barrier With barrier No barrier With barrier
Bf 0.21 0.55 0.50 1.41
Hss 0.11 0.29 0.31 0.86
Hfs 0.26 0.70 0.60 1.32
Hcs 0.28 0.65 0.72 1.13
Average 0.21 0.54 0.53 1.18

Abbreviations: Bf, backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sand.

velocity (Table 4). For each of the sediment and recharge these two velocities is the reason a surface barrier can hinder
scenarios, each of the solutes initially traveled at a constant the migration of the underlying contaminants.
velocity as indicated by Equation (6), regardless of the ini-
tial depth of the solutes. The initial (i.e., pre-barrier) solute
travel velocities in the four types of materials varied between 4.2.2 Three stages of solute convection
0.17 and 0.44 m year−1 when the pre-barrier recharge rate was
30 mm year−1 and between 0.5 and 1.24 when the pre-barrier Figure 4 illustrates the three stages of the convection of two
recharge rate was 100 mm year−1 . solutes (i.e., S40 and S80) in two selected soils, as exam-
The drainage front velocity is roughly one order of mag- ples, after the emplacement of a surface barrier based on
nitude larger (by a factor of 7 to 11) than the solute travel Equation (24). Each line is terminated when the solute enters
velocity. This means that after the emplacement of a surface the groundwater. Each of the lines for S40 contains three
barrier, the drainage front will catch up to the solute and segments, meaning solute convection will experience three
then slow down solute convection thereafter. The difference of stages. However, each of the lines for S80 contains only two
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 11 of 15

TA B L E 4 Drainage front velocity (Equation 2), pre-barrier solute travel velocity (Equation 6), and final solute travel velocity (Equation 21).

Pre-barrier recharge
rate Variables Bf Hss Hfs Hcs
30 mm year−1 vw0 3.1 1.5 2.4 3.5
vsi 0.28 0.17 0.33 0.44
vw0 /vsi 11.1 8.8 7.3 8.0
vsf 6.5 × 10−3 4.3 × 10−3 8.8 × 10−3 1.1 × 10−2
vsi /vsf 43 41 37 39
100 mm year−1 vw0 8.9 4.1 6.1 9.3
vsi 0.83 0.50 0.91 1.24
vw0 /vsi 10.7 8.2 6.7 7.5
vsf 6.5 × 10−3 4.3 × 10−3 8.8 × 10−3 1.1 × 10−2
vsi /vsf 127 118 104 113

Abbreviations: Bf, backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sands.

F I G U R E 4 Solute travel velocity in the selected soils under a surface barrier. The initial depths of contaminants were 40 m for solute S40 and
80 m for S80. Each line is terminated at the time when the solute enters the groundwater. Time lag of a surface barrier. Hcs, Hanford coarse sand;
Hss, Hanford silt sand.

segments, suggesting solute S80 is expected to experience (Table 4) until the lagged time. After the lagged time, solute
only the first two stages before the solute enters groundwater. convection enters stage 2 (t1 < t < t2 ) and solute travel veloc-
Stage 1 starts from the emplacement of the surface to a ity decreases in a power function (Equation 24 and Figure 4).
lagged time. In this stage, solute travel is not affected by the For these cases, solute travel velocity in stage 2 decreases
surface barrier. Hence, a solute migrates at the initial velocity roughly by half when travel time is doubled. At stage 3, solute
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 of 15 Vadose Zone Journal ZHANG

F I G U R E 6 Solute safe depth in a 100-m thick vadose zone below

a surface barrier (Equation 30). Bf, backfill; Hcs, Hanford coarse sand;
Hfs, Hanford fine sand; Hss, Hanford silt sands.

F I G U R E 5 Time lag (year) before a surface barrier deters

contaminant migration (Equation 10). Bf, backfill; Hcs, Hanford coarse
sand; Hfs, Hanford fine sand; Hss, Hanford silt sands.

travels at the final constant velocity. The final solute travel

velocity is 7.65 × 10−3 m year−1 , on average, corresponding
to the percolation rate of 0.5 mm year−1 . The initial solute-
F I G U R E 7 The zero-protection depth (m) of a surface barrier
travel velocities are about 40 and 114 times the final velocities
(Equation 32) in a 100-m thick vadose zone. Bf, backfill; Hcs, Hanford
(Table 4), indicating the emplacement of a surface barrier coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt sands.
will considerably reduce the solute migration once the barrier
starts taking effect after stage 1.
The surface barrier will not hinder solute migration in groundwater of 500 years, which is the design life of a type
the vadose zone below the barrier until after a lagged time of surface barrier expected to be deployed at the Hanford
(Figure 5), at which the drainage front arrives at the depth Site. The safe depth ranged between 53.3 m and 66.6 m with
of the solute. For typical sediments at the Hanford site, the an average of 59.4 m when the pre-barrier recharge rate is
time lags for solutes residing at 90-m depth range are between 100 mm year−1 . This means that a contaminant initially resid-
29 and 65 years when the pre-barrier recharge rate is 30 mm ing above an average depth of 59.4 m is not expected to arrive
year−1 (Figure 5a) and between 11 and 25 years when the pre- at the groundwater during the life (e.g., 500 years) of the sur-
barrier recharge rate is 100 mm year−1 (Figure 5b). The time face barrier. The average safe depth increases to 68.4 m when
lag is linearly proportional to the solute depth at the time of the pre-barrier recharge rate is 30 mm year−1 . This means
surface barrier emplacement (Equation 10). For example, the that, if the pre-barrier recharge rate is smaller, the safe depth
time lag for a solute at the 80-m depth is double the value of becomes larger, and hence more contaminants can be safely
that at the 40-m depth. The time lag is inversely proportional protected from entering groundwater.
to the pre-barrier recharge rate. For a solute at a given depth,
the time lag under the pre-barrier recharge rate of 30 mm
year−1 (Figure 5a) is larger than that under the pre-barrier 4.2.4 Zero-protection depth of a surface
recharge rate of 100 mm year−1 (Figure 5b). barrier

If a contaminant resides at a relatively large depth and enters

4.2.3 Solute safe depth the groundwater before the time lag, the emplacement of the
surface barrier has no effect on the migration of this con-
Figure 6 shows the solute safe depth in a 100-m thick vadose taminant. Figure 7 illustrates the zero-protection depth for a
zone below a surface barrier with a target travel time to 100-m thick vadose zone. The zero-protection depth varies
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 13 of 15

F I G U R E 8 Solute travel time (year) to groundwater without a F I G U R E 9 Solute travel time (year) to groundwater with a
surface barrier (Equation 9) in a 100-m thick vadose zone. Bf, backfill; surface barrier (Equation 27) in a 100-m thick vadose zone. Bf,
Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt backfill; Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss,
sands. Hanford silt sands.

depending on the hydraulic properties of the sediment. The groundwater. Comparing with the corresponding values with-
values under the pre-barrier recharge rate of 30 mm/year are out a surface barrier (Figure 8), the travel time to groundwater
slightly larger than those under 100 mm/year. Overall, the with a surface barrier is extended by a few to thousands of
zero-protection depth was approximately between 85.0 and years for all the solutes except S90.
91.0 m with an average of 88.0 m. In other words, on average, Figure 10 depicts the travel-time delay factor for the solutes
when the solute depth is at or larger than approximately 88% residing at different depths in the vadose zone with a surface
of the thickness of the vadose zone at surface barrier emplace- barrier. A solute residing at a smaller depth (i.e., larger dis-
ment, the surface barrier will have no impact on the migration tance to the groundwater) has larger travel-time delay factor.
of these solutes. For example, when the pre-barrier recharge rate is 30 mm
year−1 , the travel time to groundwater would be extended by
a factor of 29 for S40 and only 1.6 for S80 in sediment Hcs
4.2.5 Solute travel time to groundwater (Figure 10a). For the solutes initially residing at the upper
half of the vadose zone, the travel-time delay factors are in
Figure 8 shows the solute travel time to groundwater with- the magnitude of 10s–100s. The travel time-delay factor is
out a surface barrier for solutes residing at different depths at or near unity for S90 in all sediments under the two pre-
in the four selected sediments under two pre-barrier recharge barrier recharge conditions, meaning surface barrier has no
rates. Depending on the initial residing depth and recharge or near-zero impact on the S90 convection before S90 enters
rate, solute travel time to groundwater varies from 8 years the groundwater because the zero-protection depth is about 88
(S90 in Hcs with 100 mm year−1 pre-barrier recharge) to 575 m on average.
years (S00 in Hss with 30 mm year−1 pre-barrier recharge). It
took more time for a solute to reach groundwater when either
the pre-barrier recharge rate is lower or the initial distance to 4.3 Implications and limitations
groundwater is larger, as indicated by Equation (9).
Figure 9 illustrates the solute travel time to groundwater 4.3.1 Implications
with a surface barrier for the solutes residing at different
depths in the vadose zone. The solute initially residing at the It is a challenge to contain nuclear waste buried in the vadose
ground surface has the longest travel time, while the solute zone to protect the underlying groundwater. For example,
residing near the water table has the shortest travel to the there are hundreds of waste sites at the U.S. Department of
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14 of 15 Vadose Zone Journal ZHANG

contaminants reside too close to the groundwater. Other reme-

diation technologies, for example, soil flushing followed by
pumping and treating of groundwater, will be needed for such
sites.

4.3.2 Limitations

There are some limitations of the developed algebraic expres-

sions. The expressions with the impact of a surface barrier
may be only applicable to a homogeneous vadose zone that
has been at steady state at the time a surface barrier is
emplaced. However, the expressions that are based on pre-
barrier flow conditions, for example, pre-barrier flow velocity,
time lag, and zero-protection depth, are applicable to each of
the layers in a layered soil that has been at steady state.
Approximation was used in the derivation of solute convec-
tion during stage 2 (Equation 14). This approximation leads to
a discontinuity of solute travel velocity at the transition time
from stage 1 to stage 2. Despite this, the error because of the
F I G U R E 1 0 Time-delay factor of solute travel to the
approximation appears to be small for the selected soils and
groundwater (Equation 29) in a 100-m thick vadose zone. Bf, backfill;
flow conditions. Their applicability to other soils and flow
Hcs, Hanford coarse sand; Hfs, Hanford fine sand; Hss, Hanford silt
conditions are to be tested.
sands.
Mechanical dispersion was not considered in the analysis.
Hence, the algebraic expressions are incapable of estimating
Energy’s Hanford Site. Surface barriers with a long design life solute concentrations. A path forward may be to estimate the
(e.g., 500 years or longer) are expected to be used over such distribution of solute concentration around the center of mass
waste sites. Usually, evaluation of the expected performance with an approximate analytical solution by considering only
of a surface barrier is carried out with complex numerical dispersion at the location of the center of mass using the soil-
simulations, which are often time-consuming and costly. A moisture content and flux rate averaged during the period of
major issue of the numerical approach is that the surface bar- solute transport. The approximate solution of concentration
rier must be designed before the performance assessment and distribution is expected to be accurate before the drainage
hence a poorly designed surface barrier must be modified. front arrives at the plume zone (i.e., for most of the time in
Another issue is how to know beforehand if a surface barrier stage 1). The approximate solution is subject to error for solute
is suitable to a site depending on the site-specific conditions transport during stages 2 and 3. However, this development
without doing complex numerical simulations. The analyti- and further testing are beyond the scope of this paper.
cal approach provides a simple alternative way to predict the
expected impact of a surface barrier on the transport of the
contaminants in the vadose zone. 5 SUMMARY
Based on the analytical analysis, the time lag of a sur-
face barrier predicts that a surface barrier will not take A set of algebraic expressions are developed to determine the
effect immediately after deployment but after a certain time. velocities of convective solute transport with and without a
For contaminants already in the vadose zone, it is better to surface barrier, time lag, solute safe depth, zero-protection
deploy a surface barrier as early as possible when the con- depth of a surface barrier, and solute travel time to groundwa-
taminants are still at relatively shallower depths. However, a ter. These expressions indicate that solute convection below a
dilemma is that the design, evaluation, and decision-making surface barrier may experience up to three stages. Solute travel
of a final permanent surface barrier take time before the velocity is at a constant high velocity during stage 1, decreases
barrier can be deployed. As a temporary solution, deploy- in a power function with time during stage 2, and is at a con-
ment of a short-term interim surface barrier (e.g., Zhang stant low velocity during stage 3. These expressions show that
et al., 2007, 2012) can fill the gap for better protection of the the drainage front moves at a velocity at least three times the
groundwater. solute travel velocity. For typical soils at the Hanford Site, the
The zero-protection depth informs that the surface barrier drainage front moves 7 to 11 times as fast as the travel of a con-
technology may not be suitable for some sites, in which the servative solute. The time lag linearly increases with the initial
 15391663, 0, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/vzj2.20381 by Test, Wiley Online Library on [18/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG Vadose Zone Journal 15 of 15

depth of solute. When a solute resides in a depth below 85%– Last, G. V., Gee, G. W., Freeman, E. J., Nichols, W. E., Cantrell,
91% of the vadose zone thickness, the surface barrier will K. J., Bjornstad, B. N., & Horton, D. G. (2006). Vadose zone
have near-zero effect on the convective transport of the solute. hydrogeology data package for Hanford assessments. Pacific North-
west National Laboratory. https://www.pnnl.gov/main/publications/
When a conservative contaminant resides above an average
external/technical_reports/PNNL-14702rev1.pdf
depth of 64 m in a 100-m thick vadose zone, the contaminant
White, M. D., Appriou, D., Bacon, D. H., Fang, Y., Freedman, V. L.,
is not expected to arrive at the groundwater in 500 years. For & Rockhold, M. L. (2015). STOMP/eSTOMP user guide (PNNL-
the solutes initially residing at the upper half of the vadose SA-108766). Pacific Northwest National Laboratory. https://stomp-
zone, the time-delay factors are in the magnitude of 10s to userguide.pnnl.gov/estomp_guide/eSTOMP_guide.stm
100s. The time lag, solute safe depth, zero-protection depth, Zhang, Z. F. (2020). Algebraic expressions for estimating the impact
and time-delay factor may be used to evaluate the expected depths of a surface barrier over a homogeneous soil. Vadose Zone
performance of a surface barrier on solute transport in the Journal, 19, e20003. https://doi.org/10.1002/vzj2.20003
Zhang, Z. F., Keller, J. M., & Strickland, C. E. (2007). T tank farm
vadose zone.
interim surface barrier demonstration–vadose zone monitoring plan
(PNNL-16538). Pacific Northwest National Laboratory. http://www.
AU T H O R C O N T R I B U T I O N S osti.gov/energycitations/servlets/purl/909251-sXhdht/
Zhuanfang Fred Zhang: Conceptualization; formal analysis; Zhang, Z. F., Strickland, C. E., Field, J. G., Parker, D. L., & Clayton, R. E.
investigation; methodology; resources; validation; visualiza- (2012). Evaluating the performance of a surface barrier for reducing
tion; writing—original draft; writing—review and editing. soil-water flow. Vadose Zone Journal, 11, 1–10. https://doi.org/10.
2136/vzj2011.0117
Zhang, Z. F., Strickland, C. E., & Link, S. O. (2017). Design and perfor-
C O N F L I C T O F I N T E R E ST STAT E M E N T
mance evaluation of a 1000-year evapotranspiration-capillary surface
The author declares no conflicts of interest.
barrier. Journal of Environmental Management, 187, 31–42.

ORCID
Zhuanfang Fred Zhang https://orcid.org/0000-0001-8676-
6426 How to cite this article: Zhang, Z. F. (2024).
Algebraic expressions for convective solute transport
REFERENCES in a homogeneous soil below a surface barrier. Vadose
Brooks, R. H., & Corey, A. T. (1966). Poperties of porous media affect- Zone Journal, e20381.
ing fluid flow. ASCE Journal of the Irrigation Drainage Division, 92, https://doi.org/10.1002/vzj2.20381
61–68.
Burdine, N. T. (1953). Relative permeability calculation from pore size
distribution data. Transactions of the American Institute of Mining and
Metallurgical Engineers, 198, 71–78.