Phytoremediation: modeling plant uptake and contaminant transport in the soil–plant–atmosphere continuum

Phytoremediation: modeling plant uptake and contaminant transport in the soil–plant–atmosphere continuum

Journal of Hydrology 266 (2002) 66–82 www.elsevier.com/locate/jhydrol Phytoremediation: modeling plant uptake and contaminant transport in the soil– ...

535KB Sizes 6 Downloads 40 Views

Journal of Hydrology 266 (2002) 66–82 www.elsevier.com/locate/jhydrol

Phytoremediation: modeling plant uptake and contaminant transport in the soil– plant– atmosphere continuum Ying Ouyang* Department of Water Resources, St Johns River Water Management District, Groundwater Programs, P.O. Box 1429, Palatka, FL 32178-1429, USA Received 2 October 2001; revised 8 May 2002; accepted 14 May 2002

Abstract Phytoremediation is an emerging technology that uses plants and their associated rhizospheric microorganisms to remove, degrade, detoxify, or contain contaminants located in the soil, sediments, groundwater, surface water, and even the atmosphere. This study investigates phytoremediation of 1,4-dioxane from a contaminated sandy soil by a poplar cutting, which is associated with water flow in the soil as well as water movement and 1,4-dioxane translocation in the xylem and phloem systems. An existing one-dimensional mathematical model for coupled transport of water, heat, and solutes in the soil –plant – atmosphere continuum (CTSPAC) is modified for the purpose of this study. The model is calibrated with the laboratory experimental measurements prior to its applications. A simulation scenario is then performed to investigate phytoremediation of 1,4-dioxane by a poplar cutting in response to daily water flow and 1,4-dioxane transport for a simulation period of 7 days. Simulation shows that 1,4-dioxane concentration is high in leaves and low in roots with the stem in between. However, 1,4-dioxane mass in the stem (60%) is higher than that of leaves (28%) and roots (12%). This occurs because the stem volume used in this study is larger than those of leaves and roots. The simulation further reveals that about 30% of the soil 1,4-dioxane is removed within 7 days, resulting mainly from root uptake. A plot of the 1,4-dioxane concentrations in plant compartments as a function of time shows that the highest concentration in leaves is about 2600 mg/cm3 and the lowest concentration in roots is about 350 mg/cm3 at the end of the simulation. Results indicate that leaves are an important compartment for 1,4-dioxane accumulation and transpiration. This study suggests that the modified CTSPAC model could be a useful tool for phytoremediation estimations. q 2002 Elsevier Science B.V. All rights reserved. Keywords: Contaminant; Modeling; Phytoremediation; Water flow

1. Introduction Phytoremediation is an emerging technology that uses plants and their associated rhizospheric microorganisms to remove, degrade, or contain contaminants located in the soil, sediments, groundwater, surface water, and even the atmosphere (Chappell, * Tel.: þ1-386-312-2320; fax: þ 1-386-329-4125/4329. E-mail address: [email protected] (Y. Ouyang).

1997). Scientists have found that plants can be used to treat most classes of contaminants, including petroleum hydrocarbons, chlorinated solvents, pesticides, metals, radionuclides, explosives, and excess nutrients. These plants can be herbs, shrubs, and trees, and they can concentrate organics and heavy metals at levels much greater than normal (Brown, 1995; Ma et al., 2000). Despite plants having this useful trait, nothing was known until 1948 when the Italian researchers first

0022-1694/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 1 6 9 4 ( 0 2 ) 0 0 1 1 6 - 6

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

reported nickel hyperaccumulation in the Italian serpentine plant Alyssum bertolonii. The anomaly was all but forgotten until 1977, when the plant scientist Robert Brooks, of Massey University in New Zealand, reported similar findings (Brown, 1995). In recent years, numerous efforts have been devoted to investigating the feasibility of phytoremediation technology for removal of contaminants from soil and aquifer systems. An elaborate review of phytoremediation of contaminated soil and groundwater can be found elsewhere (Pivetz, 2001; Trapp and Karlson, 2001). In general, the phytoremediation technology involves the use of the following five mechanisms (Nyer and Gatliff, 1996). (1) Uptake of nutrients. Many surface and ground resources in the world have developed problems due to high nitrate and phosphorus concentrations. If the root system of a plant is in contact with water contaminated with nutrients, the plant will remove the nutrients from the water, resulting in faster growing plants and cleaner water. (2) Uptake of non-essential metals and organics. It has long been recognized that plants can take up nonessential metals and organics into their tissues and can be used to remove metals and organics from groundwater and soil. (3) Creating an environment of diverse microbial population. The next area that has been found to be useful is the actual root system that the plants set up in the soil. It has been reported that the root system is an excellent location to grow a diverse group of microorganisms for microbial degradation. (4) Water pumping action. The uptake of water by trees can substantially influence the local hydraulics of a shallow aquifer, thus controlling the migration of a contaminant plume. (5) Volatilization and stabilization. The uptake of organics can be volatilized by transpiration from the leaf surface into the atmosphere, and the uptake of metals can be stabilized into persistently non-bioavailable forms. Some examples about the use of this technology are given below. Blaylock et al. (1997) used Indian mustard to demonstrate the capacity of plants to accumulate high concentrations of lead (Pb) when grown in a Pbcontaminated soil. Their study results showed that the accumulation of Pb in the shoots of Brassica juncea could be enhanced through the application of synthetic chelates to the soil, facilitating high biomass

67

accumulation as well as Pb uptake. More recently, Ma et al. (2000) used a plant to remove arsenic from the contaminated field soil. These authors found that the highest arsenic concentration in aboveground biomass was 7500 ppm, which was 200 times greater than in the soil. Thompson et al. (1998) examined the potential for using hybrid poplar trees to remediate sites contaminated with the high explosive 2,4,6-trinitrotoluene (TNT). These authors performed laboratory experiments to assess the uptake of TNT from both hydroponic and soil experiments. Their results showed that TNT was strongly bound and transformed by root tissues and that it only translocated slightly to the leaves of the poplar cuttings. The translocation of TNT was found to be similar to that reported for other plant species, with up to 75% of the TNT uptake remaining in root tissues. TNT was transformed by the tree to 4-amino-2, 6-dinitrotoluene (4-ADNT), 2amino-4, 6-dinitrotoluene (2-ADNT), and to a number of unidentified compounds that are more polar than TNT. Aitchison et al. (2000) studied the phytovolatilization of 1,4-dioxane by hybrid poplar cuttings in hydroponic and soil experiments. These authors found that a removal of about 54% of 1,4-dioxane was achieved within 9 days in their hydroponic experiments. The removal corresponded to a transpiration stream concentration factor of about 0.72. Only about 8.8% of the initial spiked 1,4-dioxane remained in the soil after 15 days, compared to about 72% remaining in sterilized, unplanted soil. In both hydroponic and soil experiments, 76 – 83% of 1,4-dioxane taken up by the poplars was transpired from leaf surfaces to the atmosphere, where it was readily dispersed and photodegraded. It has long been recognized that phreatophytic trees (plants known for fast growth and high water usage rates) are effective at a rooting depth of 0.7 m and more. The Miami Conservancy District (1991) reported that a single willow tree transpires more than 18,900 liters of water in one summer day. At a site in southwestern Ohio, cottonwood trees demonstrated considerable pumping capacity, even in a relatively humid environment. An ideal situation was available where two 15-m-tall cottonwood trees could be isolated and evaluated. Monitoring wells were placed around the cottonwood trees and monitored for

68

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Fig. 1. Schematic diagram showing (a) processes involved in chemical transport/uptake in the soil–plant–atmosphere continuum and (b) a compartment model for chemical transport within a plant system. The numbers in circles are the plant compartment numbers. The diagram is redrawn after Lindstrom et al. (1990).

an entire season. Experiment results showed that the pumping rate for each cottonwood tree ranged between 189 and 1323 liters a day (Nyer and Gatliff, 1996). A thorough literature search revealed that very few mathematical models have been developed for phytoremediations during the last decade due to the complexity of the soil – plant – atmosphere continuum. Trapp and McFarlane (1995) developed the PLANTX model to simulate uptake of xenobiotic contaminants into plants. The model describes (1) the dynamic uptake from soil, solution, and atmosphere and (2) metabolism and accumulation of anthropogenic chemicals in roots, stem, leaves, and fruits. It accounts

for diffusion exchange in soil water and air pores to roots; transfer into roots with the transpiration stream; translocation into stems and leaves via the transpiration stream; partitioning into the stem; transport into fruit via the assimilation stream; diffusive exchange between air and leaves via stomata and cuticle; and metabolism and dilution by growth. This model is a very useful tool for predicting plant contamination. In the late 1980s, Boersma et al. (1988a,b and 1991) and Lindstrom et al. (1990) developed a onedimensional mathematical model for coupled transport of water, heat, and solutes in the soil –plant – atmosphere continuum (CTSPAC) in the vadose zone

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

as controlled by atmospheric, soil physical, and biological conditions. CTSPAC consists of two submodels: the soil and plant sub-models. The soil submodel describes coupled transport of water, heat, and chemicals in a soil of the vadose zone, and the plant sub-model describes root uptake, xylem and phloem transport, and accumulation of chemicals by plants. This CTSPAC model has provided a fundamental understanding of coupled transport of water, heat, and chemicals through the soil – plant – atmosphere continuum (Boersma et al., 1991; Lindstrom et al., 1991). However, no effort has been devoted to investigating the phytoremediation of contaminants using the model. Although the phytoremediation technology has shown significant potential for application (Trapp and McFarlane, 1995; Burken and Schnoor, 1996; Voudrias and Assaf, 1996; Thompson et al., 1998; Aitchison et al., 2000; Ma et al., 2000; Nedunuri et al., 2000; Trapp and Karlson, 2001), the plant physiological, microbiological, hydrological, and environmental controls upon its applications are still poorly understood. The goals of this study are (1) to calibrate the CTSPAC model for simulating the phytoremediation of 1,4-dioxane by the poplar cuttings using the laboratory experimental data and (2) to apply the model to investigate daily uptake and transport of 1,4-dioxane associated with water flow in the soil – plant – atmosphere continuum.

2. Description of CTSPAC The CTSPAC model consists of coupling a soil sub-model to a plant sub-model. The soil sub-model has three time-dependent equations for vertical simultaneous flow and transport of water, solutes, and heat through the vadose zone (Fig. 1a). Water movement is modeled using the Richards equation associated with the effects of daily cycles of surface water infiltration and evaporation, root water uptake, and leaf transpiration. Soil heat flux is described by heat conduction in the solid, liquid, and air phases; by heat convection in the liquid phase; and by the transport of latent heat. Chemical transport is described by the mechanisms of convection, dis-

69

persion/diffusion, sorption, degradation, and root uptake. The plant sub-model is based on compartmentalization of the plant into local regions of similar tissue structure and function (Fig. 1b). The plant is represented as a generic plant with three leaf clusters, each simulating a cluster of geometrically similar leaves. Properties of root compartments can be varied to simulate root density. Plant compartments have “xylem” and “phloem” regions, each containing the actual physiological and anatomical structures and the functions of xylem and phloem. Water moves from the soil to the atmosphere through the compartments representing roots, stems, and leaves. Solute transport in the phloem is modeled as a pressure-driven flow according to the Mu¨nch hypothesis. The model also accounts for water movement between xylem and phloem compartments in roots, stems, and leaves. Water movement is driven by water potential gradients, which is induced by differences in sugar concentrations in the phloem compartments. Although an elaborate description of the model is beyond the scope of this study, a moderate overview of the governing equations, boundary conditions, and mathematical functions pertaining to this study is presented below. 2.1. Soil sub-model The major features of the soil sub-model are (1) simultaneous transport of water, heat, and solutes in the soil slab, (2) dynamic coupling boundary conditions at both the atmosphere –soil and vadose zone – groundwater interfaces, (3) introduction of chemicals by rain, surface air, groundwater, and initially distributed sources in the soil layers—these sources may operate singly or in any combination, and (4) balance rules for mass, momentum, and heat. The equation for water movement through the vadose zone used by the soil sub-model is given as follows:   ›ð rw uÞ › ð rw uVl Þ ›ðrwv ðe 2 uÞVwv Þ þ þ VREVs ›t ›z ›z ¼ qws ðz; tÞAPR ðzÞ ð1Þ where VREVs is the representative elementary soil

70

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

volume (cm3); rw and rwv, the densities of water and water vapor (g/cm3), respectively; u, the volumetric water content (cm3/cm3); t, the time (h); Vl and Vwv, the velocities of liquid and vapor phase water (cm/h), respectively; z, the soil depth (cm); e , the soil porosity (cm3/cm3); qws ðz; tÞ; the water flux due to root extraction (g water/cm2 h); and APR ðzÞ; the effective soil –root contact area (cm2). The water flux across the soil –root interface is given as:     L qws ðz; tÞ ¼ 2rw Kðz; uÞ cs s ; z; t 2 cr ð0; z; tÞ ð2Þ 2

conduction in the vapor phase and the transport of latent heat (cal/cm2 h). The heat fluxes in the solid, liquid, and air phases in Eq. (5) are defined as follows: (a) solid phase Hss ¼ 2lsolid

›T ›z

ð6Þ

(b) liquid phase Hsl ¼ 2lw

›T þ c w r w Vl T ›z

ð7Þ

and Vl and Vwv are postulated by:   1 ›u ›T 2 DTl þ K s ð uÞ Vl ¼ 2Dul u ›z ›z Vwv ¼ 2Duwv

›u ›T 2 DTwv ›z ›z

(c) air phase ð3Þ Hsv ¼ 2LDatm atort

ð8Þ

ð4Þ

where Kðz; uÞ is the effective soil water conductivity (cm/h); cs and cr, the soil and root water potentials (cm), respectively; Ls, the thickness of the soil sheath around each of the rootlets (cm); Dul and Duwv ; the liquid and vapor phase water diffusivities (cm2/h), respectively; DTl and DTwv ; the liquid and vapor phase water diffusivities (cm2/h) due to soil temperature; T, the temperature (K); and Ks(u ), the soil hydraulic conductivity (cm/h). The equation for heat flux through the unsaturated soil is defined by:  › VREVs ½ð1 2 e ÞCsolid rsolid T þ ðe 2 uÞCair rair T ›t  › ½ð1 2 e ÞHss þ uHsl þ ðe 2 uÞHsv  þuCw rw T þ ›z ¼ APR qws ðzÞCw T

›rwv ›T 2 lair ›z ›z

ð5Þ

where lsolid, lw, and lair are the thermal conductivities (cal/cm h K) in the solid, water, and air phases, respectively; Datm, the molecular diffusion coefficient in air (cm2/h); and atort, the soil tortuosity factor (dimensionless). The equation for solute transport and fate is postulated by: ( VREVs

› ½ðu þ ðe 2 uÞHc ÞCl þ ð1 2 e ÞS ›t

› ½uqcl þ ðe 2 uÞqcv  þ½u þ ðe 2 uÞHc LCl þ ›z ( ¼

APR ðzÞqws ðz; tÞCl ðz; tÞ

for qw # 0

APR ðzÞqws ðz; tÞCPR ðz; tÞ for qw . 0 þAPR qrs þ VREVs Qso ðz; tÞ

where Csolid is the specific heat of soil particles (cal/cm h K); rsolid and rair, the densities of soil particles and air (g/cm3), respectively; Cair and Cw, the specific heats of air and water (cal/cm h K), respectively; Hss, the heat conduction through soil particles (cal/cm2 h); Hsl, the heat conduction and convection in the liquid phase (cal/cm2 h); and Hsv, the heat

)

ð9Þ

where Cl is the concentration of solute in liquid phase (mg/cm 3); Hc , the Henry’s law constant (cm 3 water/cm3 air); S, the average concentration of solute in the sorbed phase (mg/cm3); L, the cumulative firstorder loss coefficient (1/h); qcl and qcv, the solute

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

fluxes (mg/cm 3 h) in liquid and vapor phases, respectively; CPR, the solute concentration inside the plant root (mg/cm3); qrt, the solute diffusive flux through the soil near the root – soil interface (mg/cm2 h), then through root membranes, and finally through plant cells into the xylem vessels; and Qso the sources of solute (mg/cm3 h). The solute fluxes in the liquid and air phases and root – soil interface in Eq. (9) are defined as follows: (a) liquid phase qcl ¼ 2Dcl

›C l þ Vl Cl ›z

ð10Þ

›C v ›z

ð11Þ

(b) air phase qcl ¼ 2Dcv

(c) root – soil interface qrt ¼ 2

Dc ðz; uÞ ðC ðz; tÞ 2 CPR ðz; tÞÞ DXmb ðzÞ l

ð12Þ

where Dcl and Dcv are the diffusion coefficients of chemicals in the water and air phases (cm2/h), respectively, and Dc, the effective molecular diffusion coefficient across the root cell membrane (cm2/h).

71

gradients of total water potential through roots, stem, and leaves in both xylem and phloem, (3) chemical transport in phloem driven by a gradient of positive pressure, and (4) water vapor flow from intercellular spaces to the atmosphere for evapotranspiration, and control of water vapor loss and carbon dioxide uptake through stomata. Fig. 1b shows the compartments for root, stem, and leaf of a generic plant. These compartments contain the actual physiological and anatomical structures and functions of xylem and phloem. Each compartment should be visualized as a transport unit. Water moves from the soil to the atmosphere through the compartments. Leaves are represented by individual xylem and phloem parts, which contain the mesophyll, intercellular air space, stomatal cavities, and stomatal pore regions. Water also moves from xylem to phloem compartments and vice versa in roots, stem, and leaves. This transport is driven by water potential gradients that are affected by the sugar concentrations in these compartments. As sugar accumulates in the leaf, the total water potential decreases and water flows from xylem to phloem. In the root, water flows from phloem to xylem when sugar is unloaded and stored (Nobel, 1983). Transport in the phloem is modeled as a pressure-driven flow according to the Mu¨nch hypothesis (Geoschl et al., 1976). Sugar concentrations at points along the pathway affect the gradients that drive phloem transport because of the interrelationship between osmotic potential and turgor pressure, given by:

2.2. Plant sub-model Plants are more complicated geometrically, physiologically, and biologically than the soils of the vadose zone. While it is relatively easy to represent solute transport in the vadose zone by continuous space-time models, it is extremely difficult, if not impossible, to represent transport and fate processes in plants using the same concept (Lindstrom et al., 1990). In CTSPAC, compartmentalization of the plant into local regions of similar tissue structure and function is used. Compartments are chosen to account for important flow processes including (1) water uptake by roots, (2) water transport driven by

ct ¼ c p þ cp

ð13Þ

where ct is the total plant water potential (cm), cp, the osmotic potential (cm), and cp, the turgor pressure (cm). The equation for water movement in plant tissues between compartments is defined by:   Qij ¼ 2Aij Lij cjt 2 cit ðk~Þ

ð14Þ

where Qij is the volume flow rate between adjacent

72

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

compartments i and j (cm3/h), Lij , the transport coefficient between compartments i and j (cm/h MPa), and k~; the unit vector for positive vertical flow direction downward. By applying the continuity principle, water flows into and out of compartment i can be given as: ði21Þi 0 ¼ Qxy †ð2k^Þ þ Qðiþ1Þi †ðk^Þ þ Qijxyph †ð2^ıÞ xy

ð15Þ

The first two terms on the right side of Eq. (15) account for xylem water transport into and out of compartment i in a vertical flow direction. The third term accounts for water transport between xylem and phloem; ^ı is the unit vector for lateral flow, with positive toward the right direction. The transport and storage of a solute in plants is described by: d½V2i ð1 þ B2i ÞC2i  dt ð16Þ ¼

Di A ðC 2 Ci Þ 2 Qi C2i 2 li M2i Dxi i 2i

where Vi is sugar molar volume in compartment i (cm3); Bi, the sorption coefficient for compartment i describing the immobilization of the solute by reversible sorption to cell walls or large molecules in compartment i (dimensionless); Ci, the concentration of sugar in compartment i (mole/l); Di, the diffusion coefficient across membrane along the flow path i (cm2/h); Ai, the contact area between compartment i and the adjacent compartment (cm2); Qi, the water and water vapor flow rate (cm3/h); li, the rate constant for all other firstorder loss processes in compartment i describing immobilization of solute by incorporation into structural materials or loss of solute due to metabolism (1/h); and Mi ; the solute mass in compartment i (mg). 2.3. Atmospheric conditions The soil and plant sub-models are both coupled to atmospheric conditions, which are the highly non-linear and dynamic top boundary conditions. These boundary conditions include a daily cycle of soil temperature that is determined by the energy balance at the soil surface, a daily cycle of leaf evapotranspiration, and daily variations of air

Fig. 2. Comparison of the predicted (line) and measured (point) 1,4-dioxane mass in soil and plant. The measured data were obtained from Aitchison et al. (2000).

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

temperature and relative humidity. Some of the important top boundary conditions are given below. The boundary conditions for water movement at the atmosphere –soil interface are represented by the following non-linear equation:

rw Dul þ ðe 2 uÞrwv Duv



#

(" ð1 2 e Þlsolid þ u0 lw þ cw rw DTl T0 !#)

" ¼ ðrw cw Trw frain Þ 2 LDatm



›T 2 ›z

! z¼0

þrw cw Ks ðu0 ÞT0

  drsat ðT Þ T 2 Ta ¼ rw frain ðtÞ 2 Dpatm wv 0 ha 0 dT dz hðu0 ; T0 Þ 2 ha dz

! z¼0

›u 2 ›z

cw rw Dul T0 þ ðe 0 2 u0 ÞLrwv Duv

þ ðe 2 u0 Þ lair þ Lrwv DTv

þrw Ks ðu0 Þ



#

"

"



  ›T þ rw DTl þ ðe 2 uÞrwv DTv 2 ›z z¼0

þrsat wv ðTÞ

atmosphere – soil interface are described by:

þ

 ›u 2 ›z z¼0

73

hðu0 ; T0 Þ 2 ha dz

þrsat wv ðTa Þ ð17Þ

!# 2

T 0 2 Ta dz

lpair

!

!

" þ ð1 2 e Þð1 2 asoil Þ þ u0 ð1 2 awater Þ !

where u0 is the water content at the atmosphere – soil interface (cm3/cm3); Dul ; the diffusion coefficient of water in liquid phase (cm2/h); e , the soil porosity at the atmosphere – soil interface (cm3/cm3); Duv ; the diffusion coefficient of water in vapor phase (cm2/h); DTl ; the thermal diffusion coefficient of water in liquid phase (cm2/h); DTv ; the thermal diffusion coefficient of water in vapor phase (cm2/h); c, the soil water potential (cm); Frain, the rainfall rate (cm/h); Dpatm, the boundary layer wind speed dependent coefficient of dispersion of water vapor (cm2/h); rsat wv ðTa Þ the density of water vapor at saturation (g cm3), which is a function of atmospheric temperature Ta; ua, the relative humidity in the atmosphere (dimensionless); T0, the temperature at the atmosphere – soil interface (K); dz; the thickness of the boundary layer at the atmosphere – soil interface (cm); and hðu0 ; T0 Þ the relative humidity at the atmosphere – soil interface (dimensionless). The boundary conditions for heat flux at the

drsat T 2 Ta wv ðTa Þ ha 0 dTa dz

þ e0 2

up0

!# 12

apwater

!4 qswr þ s

Tap

"

!#

£ e soil ð1 2 e 0 Þ þ " þe air sTa4

e water up0

þ e air e 0 2

qffiffiffiffiffiffiffiffiffiffiffi# 0:605 þ 0:048 eair water ðTa Þ

up0

ð18Þ

where lsolid, lw, and la are the thermal conductivities (cal/cm K h) in solid, water, and vapor phases, respectively; s , the Stefan – Boltzman constant (cal/cm2 K4h); eair water ; the saturated vapor pressure of the air (mm Hg); aair, awater, and asoil, the albedos of air, water, and soil, respectively (dimensionless); qswr, the heat flux into or out of soil by short wave radiation (cal/cm2 h); and e air, e water, and e soil, the emissivities of air, water, and soil, respectively (dimensionless). All other variables in Eq. (18) are defined previously. The top boundary conditions for chemicals at the

74

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Table 1 Major input parameter values for simulations and 1,4-dioxane properties Parameter Atmospheric conditions Air temperature (K) Relative humidity (dimensionless) Wind speed (km/h) Rainfall rate (cm/h) Density of air (g/cm3) Specific heat of air (cal/g/k) Thickness of the boundary layer at the air– soil interface (cm) Emissivity of air above the soil The short wave reflectivity of air or albedo (dimensionless) Molecular diffusion coefficient of the vapor phase (cm2/h) Thermal conductivity of air at 20 8C (cal/g/k/h) Plant compartments Contact area between compartments (cm2) Plant compartment water conductivity (cm/h) Reflection coefficients for compartments (the ease with which solute crosses the membrane) (dimensionless) Volume of poplar cutting (cm3) Roots Stem Leaves 1,4-Dioxane sorption coefficient and transpiration to the atmosphere for compartments (1/h) Parameter for stomatal control function Thickness of soil sheath around each rootlet (cm) Average diameter of rootlet (cm) Number of rootlets Simulation controls Simulation time step (h) Simulation period (h) Simulation vertical soil distance (cm) Soil properties Soil type Soil porosity (cm3/cm3) Hydraulic conductivity (cm/h) Thermal conductivity of sand (cal/g/k/h) Thermal conductivity of clay (cal/g/k/h) Thermal conductivity of water (cal/g/k/h)

Value

Reference

Daily cycle Daily cycle Daily cycle 0 1.1 £ 1023 0.24 0.1

Lindstrom et al. (1990) Lindstrom et al. (1990) Lindstrom et al. (1990) Assumed West (1986) West (1986) Assumed

0.90

West (1986)

0.05

West (1986)

360 0.216

West (1986) West (1986)

Ranging from 0 to 848 Ranging from 0 to 10000 Ranging from 0 to 1

Lindstrom et al. (1990) Lindstrom et al. (1990) Calibrated

20 36 9 Ranging from 0 to 0.01

Assumed Assumed Assumed Calibrated

Ranging from 1 to 20.7 0.2

Lindstrom et al. (1990) Lindstrom et al. (1990)

0.0158 2769

Lindstrom et al. (1990) Lindstrom et al. (1990)

0.1 168 85

Sandy 0.45 0.1 75.6 25.2 4.93

Assumed Assumed Assumed

Assumed Assumed Hillel (1982) West (1986) West (1986) West (1986) (continued on next page)

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

75

Table 1 (continued) Parameter

Value

Reference

Thermal conductivity of organic matter (cal/g/k/h) Temperature at groundwater table (8C) Effective soil–root interface membrane diffusivity (cm2/h) Initial soil water content (cm3/cm3) Initial soil temperature (8C) Initial 1,4-dioxane concentration (mg/cm3)

2.16 20 5.6 £ 1027 0.3 20 50

Assumed Lindstrom et al., 1990 diffusivity (cm2/h) Assumed Assumed Assumed

1,4-Dioxane properties Molecular formula Molecular weight (g) Water solubility Boiling point (8C) Vapor pressure (kPa) log Kow

C4H8O2 88.1 Miscible 101.1 at 100 kPa 5.05 at 25 8C 20.27

Howard (1990) Howard (1990) Howard (1990) Howard (1990) Boublik et al. (1984) Hansch and Leo (1985)

atmosphere – soil interface are defined as follows:   ›Cl ½u0 Dcl þ ðe 2 u0 ÞDcv  2 þ ðuVl0 ÞCl0 ›z z¼0   p Hc Cl0 2 Cair ¼ Crain frain 2 Dc ð19Þ dz where Dcl and Dcv are the diffusion coefficients of chemicals in liquid and vapor phases (cm2/h), respectively; Vl0, the velocity of water flow at the atmosphere – soil interface (cm/h); Cl0, the concentration of chemical at the atmosphere – soil interface (mg/cm3); Crain, the concentration of chemical in rain (mg/cm3); and Cair, the concentration of chemical in air (mg/cm3). The stomatal aperture adjusts to leaf water potential and carbon dioxide concentration. However, the precise nature of the stomatal opening and closing mechanism remains unclear at this time. From the report by Lindstrom et al. (1990) the stomatal opening area, AK, is given as: AK ¼ AOK †SF

ð20Þ

with the stomatal control function of transpiration as: 2 þ bLk 1  L  L 2 L 1 þ bk exp gk ðck þ cTHrk Þ 1 þ bk SF ¼ 2 þ bLk 1   L 2 L 1 þ bk exp gLk cTHrk 1 þ bk

ð21Þ

where AOK is the 100% open stomatal area (cm2); bLk

West (1986)

(dimensionless) and gLk (1/MPa), the shaping factors; and cTHrk, the threshold water potential (cm).

3. Model calibration The CTSPAC model must be calibrated prior to its applications for predicting phytoremediation of 1,4dioxane in the soil – plant – atmosphere continuum. In this study, author attempted to calibrate the plant submodel under the constant soil water content and soil temperature using the soil experimental data reported by Aitchison et al. (2000). These authors studied the phytoremediation of 1,4-dioxane by hybrid poplar trees in both hydroponic and soil experiments. In the hydroponic experiments, hybrid poplar cuttings (Populus deltoiddes X niggra, DN 34, Imperial Carolina) removed 23 mg/l 1,4-dioxane rapidly. Within 9 days, a removal of about 54% was achieved. In the soil experiments, a sandy soil was obtained from the perimeter of a wastewater treatment lagoon located near Salisbury, North Carolina, and the poplar cuttings with a length of 20 cm were used. After hydroponic growth, 15 rooted cuttings were transferred to prepared 278-ml flasks. Between 180 to 200 g of air-dried soil was added to each reactor, and deionized water was added to achieve 80% of field capacity. The 15 reactors were weighed and placed in the plant growth chamber for a 3 –5 day acclimation period. At the end of the acclimation period, the

76

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Fig. 3. 1,4-dioxane accumulation in plant compartments (a) and removal from soil (b). Initral soil 1,4-dioxane mass is 1.8 £ 106mg.

reactors were modified and spiked with 14C-1,4dioxane prior to growing poplar cuttings. Four unplanted controls were also spiked at the same time as the 15 planted reactors. Mercuric chloride was added to poison microorganisms and thus to sterilize the soil. The reactors were placed in a plant growth chamber with the temperature varied from 24 to 29 8C and the water content at about 80% of field capacity. Samples from soil, roots, stem, petioles, and leaves were measured using a liquid scintillation counter. Only about 19% of the initial 1,4-dioxane spike (10 mg/kg) remained in planted soil after 15 days, compared with about 72% remaining in sterilized, unplanted soil. A comparison of measured and predicted 1,4dioxane masses in soil, roots, stem, and leaves (including petioles) for the soil experiment is shown in Fig. 2. With a linear regression equation of

YPrediction ¼ 1:0628XMeasurement and R2 ¼ 0:9977; author concludes that a good agreement is obtained between the model predictions and the experimental measurements. It should be pointed out that no measured data for phytoremediation of contaminants under changing soil water content, temperature, and atmospheric conditions are available for simultaneous calibration of the soil and plant sub-models.

4. Simulations To obtain a better understanding of the role of phytoremediation of contaminants in a soil –plant – atmosphere continuum, a simulation scenario is performed to investigate the daily phytoremediation of 1,4-dioxane by a poplar cutting from a contaminated sandy soil in response to daily cycles of water

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

77

Fig. 4. Simulated soil water content (a) and 1,4-dioxane concentration (b). The maximum root length is 20 cm.

flow and heat flux. A soil depth of 85 cm with the poplar cutting of 65 cm3 in volume is selected as the modeled domain. This poplar cutting is assumed to have an averaged root volume of 20 cm3, which is equally divided into five sectors; a stem volume of 36 cm3 as one sector; and three leaf clusters with a total volume of 9 cm3, which is equally divided into three sectors. Each sector is further divided into two compartments (one for xylem and the other for phloem), resulting in 18 plant compartments. The conceptual diagram of the modeled system used for this scenario is similar to Fig. 1b. Water and 1,4dioxane enter the root compartments from the surrounding soil through root uptake. They then transport and/or accumulate in the stem and leaf compartments in the xylem and phloem systems. Table 1 lists the major input parameter values for the atmospheric conditions, the soil physical properties, simulation controls, and the poplar cutting character-

istics used in this study. In addition, the 1,4-dioxane properties are also included in the table. These parameter values are obtained either from published experimental measurements or from theoretical calculations. The simulation starts at 0 h (midnight) and lasts for 168 h (7 days). Simulation results are presented in Figs. 3– 7 and are discussed below. 4.1. 1,4-Dioxane mass in plant and soil Accumulation of 1,4-dioxane mass in roots, stem, and leaves is shown in Fig. 3a. 1,4-Dioxane mass in roots, stem, and leaves increases consecutively with time from 0 mg at the beginning of the simulation to 8538, 44650, and 20680 mg, respectively, at the end of the simulation (168 h). That is, about 60% of 1,4-dioxane in the poplar cutting is stored in the stem, 28% in the leaves, and 12% in the roots in 7 days. This kind of distribution

78

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Fig. 5. Water potential in root (a), stem (b), and leaf (c) compartments of the xylem system as a function of time.

pattern is similar to the one reported by Aitchison et al. (2000). 1,4-Dioxane enters the roots from the soil and fluxes from the xylem system of the roots upward to the stem and finally to the leaves, where some transpires into the atmosphere and some transports downward along with photosynthesis products (e.g., sugars) from the leaves to the stem and finally to the roots in the phloem system. Meanwhile, 1,4-dioxane could also enter the phloem compartments from the xylem compartments by diffusion exchanges. Changes of 1,4-dioxane mass in soil show that the 1,4-dioxane mass decreases with time from 1.8 £ 106 mg initially to 1.25 £ 106 mg at the end of the simulation (Fig. 3b). This mass is equivalent to about 30% of soil 1,4-dioxane being removed,

resulting mainly from root uptake since 1,4-dioxane is a recalcitrant compound. 4.2. Water flow and 1,4-dioxane transport in soil Changes in soil water content and 1,4-dioxane concentration during 168 h of simulation are shown in Fig. 4. The simulation starts with an initial water content of 0.3 cm3/cm3, a temperature of 20 8C, and a 1,4-dioxane concentration of 50 mg/cm3 throughout the entire soil profile. The saturated soil water content at the bottom of the soil profile (groundwater table) is assumed to be 0.45 cm3/cm3. Soil water content above the soil depth of 30 cm decreases slightly with time from 24, 72, 120, to 168 h due to surface water evaporation and increases at the same time intervals

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

79

Fig. 6. A plot of 1,4-dioxane concentration in plant compartments at times of 24m 72 and 168 hours.

below this depth due to water supplied from the groundwater (Fig. 4a). Overall, the changes in soil water content are small, with a variation of ^ 0.02 cm3/cm3. Such a small variation would have no effects on plant growth and minimum effects on soil temperature. Changes in 1,4-dioxane concentrations with time as a function of soil depth (Fig. 4b) show a depletion of 1,4-dioxane above the soil depth of 20 cm (the maximum rooting depth) due to the root uptake. The 1,4-dioxane concentration in the root zone decreases consecutively from 0 to 168 h, with the lowest

concentration at 0 cm and the highest concentration at 20 cm in this zone. This occurs because no 1,4dioxane is supplied from the soil surface but it is supplied from below 20 cm. Fig. 4b also shows a decrease in 1,4-dioxane concentration with time below the rooting depth due to the vertical transport and diffusion of 1,4-dioxane into the groundwater. The dramatic changes in 1,4-dioxane concentration from the soil depth of 80 – 85 cm occur because the initial soil 1,4-dioxane concentration is set up to be 50 mg/cm3 above 80 cm depth and 0 mg/cm3 below this depth.

80

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Fig. 7. 1,4-dioxane concentration in Xylem and phloem of the stem compartments.

4.3. Water flux and 1,4-dioxane translocation in plant In CTSPAC, water movement in the xylem and the phloem is driven by water potential gradients. Water moves from high water potential (a less negative value) to low water potential (a more negative value). Fig. 5 shows the plant water potential as a function of time in root, stem, and leaf compartments of the xylem system. In general, the water potential is high in root compartments and low in leaf compartments, which creates a potential gradient for water flux upward from roots through the stem to the leaves. As shown in Fig. 5a, water potential is high in root compartment one and starts to decrease as the compartment number increases. The decrease in water potential with compartment number indicates that water moves from root compartment one upward to compartment seven (see Fig. 1b for the layout of the compartments). The water then moves from the root (compartment seven) to the stem (compartment nine, Fig. 5b) and finally to the leaves (compartments 13, 15, and 17, Fig. 5c) due to the water potential gradient. In addition, Fig. 5c shows a daily cycle of water potential in leaf compartments. Water potential in these compartments decreases (more negative) during the day and increases (less negative) during the night, resulting from the daily cycle of leaf water transpiration. A plot of 1,4-dioxane concentrations in plant compartments at times of 24, 72, and 168 h is given in Fig. 6. 1,4-Dioxane concentration is high in the leaf

compartments and decreases from leaves through the stem to the roots. The highest concentration, in the leaves, is about 2600 mg/cm3, while the lowest concentration, in the roots, is about 350 mg/cm3. Results suggest that leaves are the important compartments for 1,4-dioxane accumulation. It should be pointed out that although the total 1,4-dioxane mass in the stem is much more than that in the leaves (Fig. 3), this might not indicate that the stem is a favorable location for 1,4-dioxane accumulation. Rather, author attributes this to a larger stem volume used in this study (36 cm3 for stem, 20 cm3 for leaves, and 9 cm3 for roots). It should also be noted that although the root volume is larger than that of the leaves, the opposite is true for the 1,4-dioxane mass accumulation. That is, more 1,4-dioxane accumulates in the leaves than in the roots. Results indicate that 1,4dioxane storage rate in the roots is much less than in the leaves. This finding is consistent with that of Aitchison et al. (2000). While the simulation reveals that 1,4-dioxane concentration in the xylem and phloem systems of the roots and the leaves are about the same, there is a distinct difference between the xylem and the phloem systems of the stem (Fig. 7). 1,4-Dioxane concentration in the xylem system of the stem remains almost constant throughout the entire simulation period, whereas 1,4-dioxane concentration in the phloem system of the stem increases exponentially with time. At the end of the simulation (7 days), the 1,4-dioxane concentration is 1036 mg/cm3 in the phloem but only 57 mg/cm3 in the xylem. The former is about 18 times greater than the latter. A possible explanation of this phenomenon could be that 1,4-dioxane is not incorporated into the xylem tissue during water flux and 1,4-dioxane transport processes since it is not a nutrient needed for the poplar tree. Instead, some 1,4dioxane transports along with the photosynthesis products from the leaves to the stem of the phloem system and accumulates there.

5. Summary and conclusions In this study, a simulation scenario was chosen to demonstrate the phytoremediation of 1,4-dioxane by the poplar cutting associated with water flow, dioxane transport, and uptake and accumulation in

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

the soil –plant –atmosphere continuum for 7 days, using the modified CTSPAC model. A soil depth of 85 cm with a poplar cutting of 65 cm3 in volume was selected as the modeled domain. Of which, the root volume was 20 cm3, the stem was 36 cm3, and the leaf cluster was 9 cm3. The model was calibrated prior to its applications for predicting the uptake and accumulation of 1,4-dioxane in roots, stem, and leaves, using the laboratory experimental measurements. A good agreement was obtained between the model predictions and the experimental measurements. Simulation showed that 1,4-dioxane concentration was high in leaves and low in roots, with the stem in between. However, 1,4-dioxane mass in the stem (60%) was higher than in the leaves (28%) and the roots (12%). This occurs because the stem volume (36 cm3) used in this study was larger than that of the leaves (9 cm3) and the roots (20 cm3). The simulation further revealed that about 30% of the soil 1,4dioxane was removed within 7 days, resulting mainly from root uptake. Plant water potential is a driving force for water flux through the xylem system. Water potential was high in the root compartments and low in the leaf compartments, with the stem compartment in between. This potential gradient induced water flow from the roots upward to the leaves. A daily cycle of leaf water potential—driven by the daily cycle of leaf water transpiration—was observed. A plot of 1,4-dioxane concentration in plant compartments as a function of time showed that the highest concentration, in leaves, was about 2600 mg/cm3 and the lowest concentration, in roots, was about 350 mg/cm3. Results suggest that leaves are the important compartments for 1,4-dioxane accumulation. It should be pointed out that phytoremediation of contaminants in a growing-plant system is not included in CTSPAC and therefore was not considered in this study. In addition, although CTSPAC is complex, it is a one-dimensional model that limits its applications to solve field-scale problems. Further study should focus on simplifying CTSPAC and extending it to a threedimensional model for coupled transport and uptake of water, contaminants, and heat in a growing-plant system. A potential approach to tackle this issue would be the integration of the modified CTSPAC model with the GIS cell-based modeling tools.

81

References Aitchison, E.W., Kelley, S.L., Schnoor, J.L., 2000. Phytoremediation of 1,4-dioxane by hybrid Poplar trees. Water Environment Research 72, 313 –321. Blaylock, M., Salt, D.E., Raskin, I., 1997. Enhanced accumulation of Pb in Indian Mustard by soil-applied chelating agents. Environmental Science and Technology 31, 860–865. Boersma, L., McFarlane, C., McCoy, E.L., 1988a. Uptake of organic chemicals by plants: a theoretical model. Soil Science 146, 403–417. Boersma, L., McFarlane, C., McCoy, E.L., 1988. Model of couple transport of water and solutes in plants. Spec. Report 818. Agric. Exp. Sta., Oregon State University, Corvallis, Oregon, USA, p. 109. Boersma, L., Lindstrom, F.T., Childs, S.W., 1991. Model for steady state coupled transport in xylem and phloem. Agronomy Journal 83, 401–415. Boublik, T., Fried, V., Hala, F., 1984. The Vapor Pressures of Pure Substances, Elsevier Science, Amsterdam, p. 17. Brown, K.S., 1995. The green clean. Bioscience 45, 579– 582. Burken, J.G., Schnoor, J.L., 1996. Phytoremediation: plant uptake of atrazine and role of root exudates. Journal of Environmental Engineering 122, 958–963. Chappell, J., 1997. Phytoremediation of TCE Using Populus. Status report prepared for the US EPA Technology Innovation Office under a National Network of Environmental Management Studies Fellowship; http://clu-in.com/phytoTCE.htm. Geoschl, J.D., Magnuson, C.E., DeMichele, D.W., Sharpe, P.J.H., 1976. Concentration-dependent unloading as a necessary assumption for a closed form mathematical model of osmotically driven pressure flow in phloem. Plant Physiology 58, 556 –562. Hansch, C., Leo, A.J., 1985. Medchem project Issue 26. Pomona College, Claremont, Calif. Hillel, D., 1982. Introduction to Soil Physics, Academic Press, New York. Howard, P.H., 1990. Solvents, Handbook of Fate and Exposure Data for Organic Chemicals, Lewis Publishers, Chelsea, p. 2. Lindstrom, F.T., Boersma, L., Yingjajaval, S., 1990. CTSPAC: mathematical model for coupled transport of water, solutes, and heat in the soil–plant–atmosphere continuum. Mathematical theory and transport concepts. Agric. Exp. Sta., Oregon State University, Corvallis, Oregon, USA, Station bulletin 676. Lindstrom, F.T., Boersma, L., McFarlane, C., 1991. Mathematical model of plant uptake and translocation of organic chemicals: development of the model. Journal of Environmental Quality 20, 129–136. Ma, L.Q., Tu, C., Kennelley, E.D., Komar, K.M., 2000. Phytoremediation of arsenic contaminated soils and wastes. Annual meetings abstracts, American Society of Agronomy, Minneapolis, Minn. November, 5–9. Miami Conservation District, 1991. Aquifer Update 1, 1. Nedunuri, K., Govindaraju, V., Chen, R.S., 2000. Evaluation of phytoremediation for field-scale degradation of total petroleum hydrocarbons. Journal of Environmental Engineering 126, 483 –490.

82

Y. Ouyang / Journal of Hydrology 266 (2002) 66–82

Nobel, P.S., 1983. Biophysical Plant Physiology and Ecology, Freeman, San Francisco. Nyer, E.K., Gatliff, E.G., 1996. Phytoremediation. GWMR, 58– 62. Pivetz, B.C., 2001. Phytoremediation of contaminated soil and ground water at hazardous waste sites. Office of Research and Development, US-EPA, EPA/540/S-01/500. Trapp, S., Karlson, U., 2001. Aspects of phytoremediation of organic pollutants. Journal of Soils and Sediments 1, 37 –43. Trapp, S., McFarlane, J.C., 1995. Plant Contamination: Modeling and Simulation of Organic Chemical Processes, Lewis Publishers, Boca Raton.

Thompson, P.L., Ramer, L.A., Schnooer, J.L., 1998. Uptake and transformation of TNT by hybrid poplar trees. Environmental Science and Technology 32, 975– 980. Voudrias, E.A., Assaf, K.S., 1996. Theoretical evaluation of dissolution and biochemical reduction of TNT for phytoremediation of contaminated sediments. Journal of Contaminant Hydrology 23, 245 –261. West, R.C. (Ed.), 1986. Handbook of Chemistry and Physics, 67th ed., CRC Press, Boca Raton.