Biologia, Bratislava, 61/Suppl. 19: S251—S254, 2006
Section Botany
DOI: 10.2478/s11756-006-0167-9
Simulated cadmium transport in macroporous soil
during heavy rainstorm using dual-permeability approach
Jaromír Dušek1, Tomáš Vogel1, Ľubomír Lichner2, Andrea Čipáková3 & Michal Dohnal1
Faculty of Civil Engineering CTU, Thákurova 7, CZ–16629 Prague, Czech Republic; email: [email protected]
Institute of Hydrology, Slovak Academy of Sciences, Račianska 75, SK–83102 Bratislava, Slovakia
3
Regional Public Health Authority, Ipeľská 1, SK–04220 Košice, Slovakia
1
2
Abstract: Numerical modelling is used to analyze the transport of cadmium in response to an extreme rainfall event. The
cadmium transport through the soil profile was simulated by the one-dimensional dual-permeability model, which assumes
the existence of two mutually communicating domains: the soil matrix domain and the preferential flow domain. The model
is based on Richards’ equation for water flow and advection-dispersion equation for solute transport. A modified batch
technique allowed us to consider domain specific sorption, i.e. each of the domains has its own distribution coefficient. The
dual-permeability model predicts that the cadmium can be transported substantially below the root zone after the storm.
On the other hand, classical single permeability approach predicted that almost all applied cadmium stays retained near
the soil surface.
Key words: cadmium transport, vadose zone, preferential flow, macropores, dual-permeability, sorption, particle-facilitated
transport
Introduction
The increased vulnerability of the fresh water resources,
with regard to a widespread use of agricultural fertilizers containing traces of cadmium (Lichner et al.,
2006), requires a careful assessment of the potential
paths of contamination from the soil surface to groundwater. Especially in lowland regions of Central and
Eastern Europe, such as the Danubian Lowland, the
use of such fertilizers has a long history (Lichner et
al., 2006) and thus poses significant threat for both terrestrial organisms and humans (Horn et al., 2006).
One of the processes that significantly affect the
transport of cadmium in soils is preferential flow (water and solutes travel at considerable high velocities
through preferential pathways concurrently bypassing
the porous matrix). This phenomenon has a direct influence on infiltration, drainage, and specifically on solute
transport. There are several types of preferential flow
which may be caused e.g. by textural interfaces, soil
water repellency, air entrapment, soil aggregation and
spatial heterogeneity (Walter et al., 2000; Ritsema
& Dekker, 2000; Wang et al., 1998; Lennartz et al.,
1999).
The aim of this study is to predict the cadmium
penetration in the soil of interest in response to an intense rainfall by means of dual-permeability modelling.
The dual-permeability model invokes local nonequilibrium in pressure head and solute concentration between
c
2006
Institute of Botany, Slovak Academy of Sciences
the two pore domains. This is achieved through dividing the liquid phase continuum into that of the preferential flow domain (further on abbreviated to PF-domain)
and the matrix flow domain (MF-domain). Moreover,
a distinct sorption coefficient for each of the two flow
domains of the dual-permeability system is used to estimate the effect of particle-facilitated transport of cadmium through macropores.
Material and methods
Experimental site
The soil at the Macov experimental station (Dunubian Lowland, Slovakia) has been classified as Calcari-Haplic Chernozem. The soil profile consists of five relatively homogeneous horizons. The field infiltration experiments accompanied by a dye tracer test have revealed significant preferential flow patterns.
The soil-water retention curves for the five soil horizons were measured by standard pressure plate apparatus
method on 100 cm3 undisturbed soil samples (with three
replicates for each layer), and consequently the hydraulic
parameters were obtained by fitting van Genuchten’s modified prediction model (VOGEL et al., 2001) to data points.
The measurements of the topsoil saturated hydraulic conductivity for the matrix and the preferential domain were
carried out by tension infiltrometer (LICHNER et al., 2006).
The saturated conductivities of remaining soil layers were
derived by a pedotransfer function model (SCHAAP et al.,
2001) based on textural classes. The volumetric portion of
Unauthenticated
Download Date | 6/18/17 10:11 PM
J. Dušek et al.
S252
Table 1. Measured soil hydraulic properties (volumetric portion of the fracture domain wf as well as the retention curve parameters
for the PF-domain were estimated).
Domain:
MF-domain
PF-domain
Depth (cm):
0–20
20–30
30–50
50–80
80–100
0–100
θs (–)
θr (–)
α (cm−1 )
n (–)
hs (cm)
Ks (cm d−1 )
wf
0.498
0.000
0.018
1.212
–2.48
23.2
–
0.486
0.000
0.042
1.176
–2.06
29.9
–
0.502
0.000
0.057
1.184
–0.80
18.9
–
0.452
0.000
0.026
1.215
–2.61
17.6
–
0.479
0.073
0.016
1.647
–2.88
31.7
–
0.600
0.050
0.145
2.680
0.00
201.3
0.1
the PF-domain was estimated as 10% of the bulk soil. Table 1 summarizes the soil hydraulic parameters for both PFand MF-domains.
Batch tests
Sorption distribution coefficient for the matrix domain,
Kdm , which describes the chemical partitioning between the
solid and the liquid phase, was experimentally determined
in the laboratory by standard batch test (SELIM et al., 1992;
ČIPÁKOVÁ & MITRO, 1997). In contrast, the sorption distribution coefficient for the PF-domain, Kdf , was determined
by modified batch technique (LICHNER & ČIPÁKOVÁ, 2002).
In case of the latter coefficient, the liquid phase contained
not only cadmium in aqueous phase, but also cadmium adsorbed on soil particles smaller than about 0.01 mm. Such
soil particles may be highly mobile in the macropore network (LAEGSMAND et al., 1999), thus may travel to considerable depths as suspension.
Numerical model S1D DUAL
The dual-permeability approach (GERKE & VAN GENUCHTEN, 1993; VOGEL et al., 2000) assumes that the porous
medium consists of two separate domains with specific hydraulic properties. One-dimensional variably-saturated water flow in the dual-permeability model is described by
a pair of Richards’ equations for the PF-domain and the
MF-domain pore systems. Similarly, a coupled pair of
convection-dispersion equations is solved to model solute
transport. The exchange of water and solute between the
matrix and the fracture domains is assumed to be proportional to the local pressure difference and the concentration
gradient between the two pore systems. An analogous approach to water flow modelling in macroporous soils was
also applied by KODEŠOVÁ et al. (2006) in this issue.
The dual sets of governing equations for flow of water and transport of cadmium are solved numerically by finite elements using the computer code S1D DUAL (VOGEL,
1999, Documentation of the S1D DUAL code – version 1.1,
CTU Prague, internal report), which is extended version of
the HYDRUS 5 code (VOGEL et al., 1996).
Domain specific sorption
When considering domain specific sorption, each pore domain (soil matrix domain and preferential flow domain) has
its own sorption properties. In our study, both pore domains
are assumed to contain only equilibrium sorption site characterized by the distribution coefficients Kdm (139 cm3 g−1 )
and Kdf (7 cm3 g−1 ), respectively (LICHNER & ČIPÁKOVÁ,
2002). The distribution coefficient for the PF-domain is approximately 20-times smaller than for the matrix due to the
fact that the soil particles smaller than 0.01 mm (loaded
with the adsorbed cadmium) are assumed to move with the
soil water, so that the coefficient Kdf is exclusively related
to the sorption on particles larger than 0.01 mm.
Initial and boundary conditions
At the beginning of the simulated period – 14 hours prior
to the rainstorm event of interest – the cadmium was applied in a definite pulse of water with the concentration load
2 µg/mL (i.e. 2000 µg/L). The pulse lasted 30 minutes and
comprised of 0.1 cm of water. The initial soil profile was
assumed to be free of cadmium. The bottom boundary condition was set to zero concentration gradient, in order to
allow the contaminant to pass freely the lower boundary at
the depth of 100 cm.
The soil surface was treated as the “atmospheric
boundary condition”. This type of condition allows for
switching between the Neumann and Dirichlet type conditions, i.e. when the top soil is not capable to transmit
water during heavy rain, the flux condition is changed to
pressure condition. In such a case, surplus water may either
generate surface runoff or stay retained at the soil surface.
The unit hydraulic gradient condition was used at the lower
boundary, allowing water to leave the soil profile at the rate
equal to unsaturated hydraulic conductivity. The initial condition for water flow was set close to residual water content
throughout the whole soil profile. No evaporation was taken
into account for the whole simulated period.
The rainfall event used in our analysis was recorded at
Pezinok-Myslenice weather station nearby the experimental
site at Macov. The recorded storm occurred in July, 1999
and total cumulative rate amounted to 15 cm rain during
17 hours.
Results and discussion
Figure 1 depicts the simulated infiltration rates during
the rainstorm. The simulation started about 14 hours
before the beginning of the major rainfall episode and
the infiltration capacity of the soil matrix became almost immediately exceeded, thus the surplus water was
diverted to the PF-domain. Since the influx to the
macropore domain was very high (Fig. 1) due to the
small proportion of the PF-domain compared to the
soil matrix domain, the preferential pathways became
Unauthenticated
Download Date | 6/18/17 10:11 PM
Simulated cadmium transport in macroporous soil
S253
Fig. 1. Simulated cumulative and infiltration rates during the natural rainfall event by dual-permeability model. The cumulative rates
are multiplied by the volumetric proportion of the respective flow domain.
concentration (µg/L)
1.2
0.9
PF-domain
0.6
0.3
MF-domain
0
0
3
6
9
12
15
time (hours)
18
21
24
Fig. 2. Simulated cadmium breakthrough curve in the matrix and
the PF-domain at the depth of 10 cm (liquid concentrations).
0
cadmium mass (log µg)
fully saturated as well, and hence were not able to transmit all rainwater into the soil profile. The fourth rain
pulse (17–18 hours) also saturated the matrix domain,
but at this time, the PF-domain transmitted all rainfall excess from the matrix. In total, the composite pore
system did not accept 5.8 cm of water during the whole
rainfall period. This amount of water was assumed to
be lost e.g. due to the surface runoff.
The breakthrough curves, simulated by the dualpermeability model, are shown in Fig. 2. Cadmium
residual concentration in the liquid phase from the PFdomain at the level of 10 cm beneath the soil surface
was much stronger compared to cadmium outflow from
the MF-domain, which showed only negligible increase
of concentration. Furthermore, this minor increase can
be attributed to the cadmium transfer from the preferential pathways to the matrix domain. In the model,
both the higher hydraulic conductivity and the lower
distribution coefficient of the PF-domain controlled the
deeper and more intense cadmium penetration in the
PF-domain leading to increased residual concentrations
down to the depth of about 40 cm below the soil surface. When classical single-permeability model was used
instead of the dual-permeability model, no increase of
cadmium concentration below the depth of 8 cm was
predicted (not shown in this paper).
The cadmium increase in soil sub-layers down to
40 cm beneath the soil surface is evident from Fig. 3.
The cadmium mass in the logarithmic scale was calculated by combining the contributions from the two flow
domains. The cadmium amount in the first sub-layer
(0–5 cm) mainly reflects the presence of cadmium retained in the soil matrix. The deeper cadmium contents
are associated with the PF-domain. The sudden mass
increase near the soil surface (horizon 0–5 cm) is related
to the cadmium application. Deeper cadmium penetration in macropores was caused by the significantly
smaller distribution coefficient Kdf for the PF-domain.
Note that the threshold of the numerical accuracy lies
in the order of magnitude 10−11 µg, hence the concentrations around this level are considered to be equal
0-5 cm
5-10 cm
10-15 cm
15-20 cm
20-25 cm
30-40 cm
-4
-8
-12
0
3
6
9
12
15
time (days)
18
21
24
Fig. 3. Cadmium increase in liquid and solid phase in the individual sub-layers as predicted by the dual-permeability approach.
to zero. Classical single-permeability approach would
predict that nearly all applied cadmium stays retained
near the soil surface in about 5 cm deep layer.
Conclusions
The effect of an extreme rainfall on cadmium transport
through macroporous soil was examined. It is demonstrated that cadmium, according to numerical model
Unauthenticated
Download Date | 6/18/17 10:11 PM
S254
based on dual-permeability approach, may percolate
below the root zone within a short period of time.
The simulation results create the most unfavourable situation: retention of cadmium in the root zone,
where it can be easily accessible by plants, and, at the
same time, quite deep cadmium penetration, which can
pose higher contamination risk for drinking water resources. Modelling with classical single permeability approach and without considering the particle-facilitated
transport, would lead to a significant underestimation
of cadmium penetration by disregarding the macroporous nature of soils.
Acknowledgements
The research has been supported by the European Union
FP6 Integrated Project AquaTerra (Project No. 505428)
under the thematic priority “sustainable development,
global change and ecosystems”, and by the research fund
of the Ministry of Education of the Czech Republic
(MSM 6840770002). Additional support was provided by
the Science and Technology Assistance Agency project
No. APVT-51-006502, and Slovak Scientific Grant Agency
project VEGA 2/6003/26.
References
ČIPÁKOVÁ, A. & MITRO, A. 1997. Influence of agrochemical characteristics on 85 Sr and 137 Cs sorption in soil samples from the
localities around nuclear power plants in Slovak Republic. J.
Radioecol. 5: 3–8.
GERKE, H.H. & van GENUCHTEN, M.Th. 1993. A dual-porosity
model for simulating the preferential movement of water and
solutes in structured porous media. Water Resour. Res. 29:
305–319.
HORN, A.L., REIHER, W., DÜRING, R.-A. & GÄTH, S. 2006. Efficiency of pedotransfer functions describing cadmium sorption
in soils. Water, Air, and Soil Pollution 170: 229–247.
KODEŠOVÁ, R., KODEŠ, V., ŽIGOVÁ, A. & ŠIMŮNEK, J. 2006. Impact of plant roots and soil organisms on soil micromorphology and hydraulic properties. Biologia, Bratislava 61(Suppl.
19): S339–S343.
J. Dušek et al.
LAEGSMAND, M., VILLHOLTH, K.G., ULLUM, M. & JENSEN,
K.H. 1999. Processes of colloid mobilization and transport
in macroporous soil monoliths. Geoderma 93: 33–59.
LENNARTZ, B., MICHAELSEN, J., WICHTMANN, W. & WIDMOSER, P. 1999. Time variance analysis of preferential solute
movement at a tile-drained field site. Soil Sci. Soc. Am. J. 63:
39–47.
LICHNER, Ľ. & ČIPÁKOVÁ, A. 2002. Cadmium distribution coefficients and Cd transport in structured soils. Plant, Soil
Environ. 48: 96–100.
LICHNER, Ľ., DLAPA, P, ŠÍR, M., ČIPÁKOVÁ, A., HOUŠKOVÁ, B.,
FAŠKO, P. & NAGY, V. 2006. The fate of cadmium in field
soils of the Danubian lowland. Soil Till. Res. 85: 154–165.
RITSEMA, C.J. & DEKKER, L.W. 2000. Preferential flow in water
repellent sandy soils: principles and modeling implications. J.
Hydrol. 231–232: 308–319.
SELIM, H.M., BUCHTER, B., HINZ, C. & MA, L. 1992. Modeling
the transport and retention of cadmium in soils: Multireaction and multicomponent approaches. Soil Sci. Soc. Am. J.
56: 1004–1015.
SCHAAP M.G., LEIJ, F.J. & VAN GENUCHTEN, M.Th. 2001.
ROSETTA: a computer program for estimating soil hydraulic
parameters with hierarchical pedotransfer functions. J. Hydrol. 251: 163–176.
VOGEL, T., HUANG, K., ZHANG, R. & VAN GENUCHTEN, M.Th.
1996. The HYDRUS code for simulating One-dimensional water flow, solute transport, and heat movement in variablysaturated media, Version 5.0. Research Report No. 140, U.S.
Salinity Lab., ARS, USDA, Riverside.
VOGEL, T., GERKE, H.H., ZHANG, R. & VAN GENUCHTEN,
M.Th. 2000. Modeling flow and transport in a two-dimensional dual-permeability system with spatially variable hydraulic properties. J. Hydrol. 238: 78–89.
VOGEL, T., VAN GENUCHTEN, M.Th. & CÍSLEROVÁ, M. 2001.
Effect of the shape of soil hydraulic functions near saturation
on variably-saturated flow predictions. Adv. in Water Resour.
24: 133–144.
WALTER, M.T., KIM, J.-S., STEENHUIS, T.S., PARLANGE, J.Y., HEILIG, A., BRADDOCK, R.D., SELKER, J.S. & BOLL,
J. 2000. Funneled flow mechanisms in a sloping layered soil:
Laboratory investigation. Water Resour. Res. 36: 841–849.
WANG, Z., FEYEN, J. & RITSEMA, C.J. 1998. Susceptibility and
predictability of conditions for preferential flow. Water Resour. Res. 34: 2169–2182.
Unauthenticated
Download Date | 6/18/17 10:11 PM