Printer Friendly

Modelling water and chemical transport in large undisturbed soil cores using HYDRUS-2D.

Introduction

Pollution of groundwater through leaching of surface-applied chemicals continues to be a worldwide topic of environmental concern (Pang et al. 2000; Magesan et al. 2003; Hassan et al. 2005; Phillips and Burton 2005; Starr et al. 2005). This is particularly the case where fertilisers, solid wastes (biosolids), or liquid wastes (effluent from sewerage treatment plants and intensive livestock industries) are applied to soil types having coarse-textured or well-structured characteristics, which are located in areas of moderate to high rainfall, have shallow watertables, and/or receive water supplement through surface irrigation.

Chemical transport models are capable of identifying potential environmental risks associated with leaching of surface-applied chemicals, thereby assisting in the development of best-management practice guidelines and recommendations. Numerous chemical transport models of varying complexity are currently available (CHEMFLO[TM]-2000, LEACHM, UNSATCHEM, and SWIMv2); however, recently there has been interest in the HYDRUS-2D software (Simunek et al. 1999) to model water and chemical transport in variably saturated media. Although HYDRUS-2D has been used extensively to model soil-water flow (e.g. Rassam et al. 2003; Schmalz et al. 2003; Hassan et al. 2005), published literature on its use to simulate chemical transport involving non-reactive and reactive chemicals (Pang et al. 2000; Close et al. 2003; Magesan et al. 2003; Sarmah et al. 2005; Starr et al. 2005), and/or chemicals undergoing transformation during leaching (e.g. nitrogen), are relatively scarce.

Previous work by Phillips (2002a, 2002b) and Phillips and Burton (2005) has generated considerable information on fertiliser (KCl) and wastewater (piggery effluent) leaching in large, undisturbed cores of a coarse-textured Podosol soil, and a well-structured Vertosol clay soil, respectively. The aim of this study was to demonstrate the capability of HYDRUS-2D to simulate the observed soil-water movement, transport of non-reactive (chloride [[Cl.sup.-]]) and reactive (potassium [[K.sup.+]]) chemicals, and nitrogen transformation (ammonium [N[H.sub.4.sup.+]] to nitrate [N[O.sub.3.sup.-]]) data reported by Phillips (2002a, 2002b) and Phillips and Burton (2005).

Materials and methods

Column leaching study 1: piggery wastewater applied to a Vertosol

This study investigated nutrient leaching losses from a Vertosol (Isbell 1996) following repeated surface applications of piggery wastewater. Detailed information on this leaching study was presented by Phillips (2002a, 2002b), and is only briefly provided here. Undisturbed soil cores (0.30 m i.d. by 0.55 m long) were collected from a piggery wastewater re-use area near Toowoomba, south-east Queensland, Australia. The Vertosol was a well-structured clay soil, with cracking evident at the surface. The collected core comprised 2 main layers: a 0.20-m A horizon underlain by a 0.35-m B horizon. Relevant chemical and physical properties of this soil are provided in Table 1.

Each soil core received 4 individual wastewater applications over the duration of the experiment (Day 1, 8, 19, 40, and 82), and this provided about 1500 kg [C.sup.l-]/ha and 1235 kg N[H.sub.4.sup.+]/ha. Deionised water was applied to the soil surface following each wastewater addition using a calibrated rainfall simulator. The average delivery rate of water equated to a surface flux density of 3.7 cm/h, and the total depth of water plus effluent applied over the experiment was approximately 0.92 m. Although the delivery rate of water exceeded the measured [K.sub.sat] of the 0-0.20m layer (Table 2), no surface ponding of water was observed throughout the application periods. This was because the average pore-water velocity in the surface 0-0.10 m was approximately 100 cm/h due to the presence of numerous surface cracks (Phillips 2002b). Leachate was collected in aliquots of about 500 mL, and analysed by flow injection analysis (LaChat Flow Injection Analysis QuickChem 8000) for [Cl.sup.-] (QuickChem method 10-117-07-2-B), N[H.sub.4.sup.+]-N (QuickChem method 10-107-06-4-D), and N[O.sub.3.sup.-]-N (QuickChem method 10-1070-4-1-H). Hereafter, N[H.sub.4.sup.+]-N, and N[O.sub.3.sup.-]-N will be referred to as N[H.sub.4.sup.+] and N[O.sub.3.sup.-], respectively.

The predominantly permanent negative surface charge of the Vertosol (Table 1) meant that [Cl.sup.-] behaved as a non-reactive ion, and N[H.sub.4.sup.+] behaved as a reactive ion, during leaching. Furthermore, despite the very high N[H.sub.4.sup.+] concentration in the applied wastewater (0.042mmol/[cm.sup.3]), negligible concentrations of this cation were detected either adsorbed by the soil colloids or in the leachate because it was converted to N[O.sub.3.sup.-] within, and subsequently leached from, the soil core (Phillips 2002b). The conversion of N[H.sub.4.sup.+] to N[O.sub.3.sup.-] during the study provided an excellent database for evaluating the capability of HYDRUS to simultaneously predict the transport, adsorption, and transformation of surface-applied N[H.sub.4.sup.+].

Column leaching study 2: KCl applied to a Podosol

This study investigated KCl leaching losses from a coarse-textured Podosol (Isbell 1996). Detailed information on this leaching study was presented by Phillips and Burton (2005), and is only briefly described here. Four undisturbed soil cores (0.30m i.d. by 0.85m long) were collected from a second rotation exotic pine forest plantation on Bribie Island, Queensland, Australia. The collected core comprised 3 main layers: a 0.30-m A11 horizon underlain by a 0.20-m A12 horizon, underlain by a 0.35-m A13 horizon. Relevant chemical and physical properties of this soil are provided in Table 1.

Potassium chloride (KCl) was applied to the soil surface in 100 mL of solution containing either (equivalent on a surface area basis): 0 kg [K.sup.+]/ha (as KCl) + 50 kg P/ha (as DAP); 50 kg [K.sup.+]/ha + 50 kg P/ha; 100 kg [K.sup.+]/ha + 50 kg P/ha; and 300 kg [K.sup.+]/ha + 50 kg P/ha (treatment K0, K50, Kl00, and K300, respectively). Each soil core was intermittently leached with deionised water using a calibrated rainfall simulator, which supplied water to the soil surface at a constant flux density of 2.1 cm/h. Over the duration of the experiment, each soil core underwent 10 leaching events, and received approximately 0.65m of irrigation. The experiment was terminated after the fertiliser pulse had emerged from the base of the soil core, and the subsequent EC of the leachate remained relatively low and constant (~50 [micro]S/cm). Leachate was collected in aliquots of about 500 mL, and analysed for [K.sup.+] using atomic absorption spectroscopy (Varian Spectra AA), and [Cl.sup.-] using flow injection analysis (LaChat Flow Injection Analysis QuickChem 8000, method 10-117-07-2-B).

Model input requirements

HYDRUS-2D (hereafter referred to as HYDRUS) is an interactive software system for simulating 1-dimensional and 2-dimensional water and chemical transport in uniform or multi-layered unsaturated, partially saturated, and fully saturated soil systems. Water movement is modelled using Richards' equation, and chemical transport is modelled using convection-dispersion type equations. The transport equations include non-linear, non-equilibrium reactions between the solid and liquid phases for reactive chemicals, and first-order degradation reactions for chemicals undergoing transformations such as nitrification of N[H.sub.4.sup.+] to N[O.sub.3.sup.-]. Detailed description of the governing equations in HYDRUS and their application can be found in Simunek and van Genuchten (1999), Simunek et al. (1999), and Rassam et al. (2003).

HYDRUS requires information on the hydraulic properties of each soil material, the sorption properties of each soil material for each chemical modelled, and the transformation properties of each chemical studied. All information required for the HYDRUS outputs reported here was obtained from the respective nitrogen transformation, sorption isotherm, and column leaching studies. The values of the parameters used to simulate the data of Phillips (2002a, 2002b) and Phillips and Burton (2005) are presented in Table 2.

Column leaching study 1: piggery wastewater applied to a Vertosol

Two distinct layers were distinguished in the Vertosol core (0-0.20 and 0.20-0.55 m), and each layer was assumed to have uniform physical and chemical properties. Simulations were nm over a 2088-h period using a grid spacing of 1 cm. The amounts of measured irrigation (water and effluent) and calculated evaporation were used as a time-variable boundary for the soil surface. Chemicals ([Cl.sup.-] and N[H.sub.4.sup.+]) were introduced to the model domain through the amount and concentration of effluent on each of the application days. The maximum and absolute minimum soil potential allowed at the soil surface (hCritA) was set at 0 and 10 000 cm, respectively. The hydraulic parameters of each layer (residual soil-water content [[theta].sub.r] saturated soil-water content [[theta].sub.s], and coefficients [alpha] and n) were estimated by fitting soil-water retention data (Klute 1986) using the RETC software (van Genuchten et al. 1991). The N[H.sub.4.sup.+] sorption parameters (the distribution coefficient [k.sub.d] ([cm.sup.3]/g), and coefficients [beta] and [eta]) were estimated by fitting sorption isotherm data to the general non-linear sorption equation:

(1) S = ([k.sub.d] [c.sup.[beta]])/(1 + ([eta][c.sup.[beta]])

where S is the sorbed ion concentration (mmol/g) and c is the solution ion concentration (mmol/[cm.sup.3]). The Langmuir-type equation ([beta]= 1) provided the best-fit to the N[H.sub.4.sup.+] sorption isotherm data ([r.sup.2] = 0.99; Fig. 1). The first-order transformation rate constant for N[H.sub.4.sup.+] transformation to N[O.sub.3.sup.-] was estimated by fitting the data of Phillips (2002a) to:

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] is the rate of N[H.sub.4.sup.+] transformation (mmol/[cm.sup.3]/h), [Sink.sub.liquid] is a first-order transformation rate constant (1/h), and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] is the chemical concentration in solution (mmol/[cm.sup.3]) (Cho 1971; Simunek et al. 1999).

Column leaching study 2: KCl applied to a Podosol

Three distinct layers were distinguished in the Podosol core (0-0.30, 0.30-0.50, and 0.50-0.85m), and each layer was assumed to have uniform physical and chemical properties. Simulations were run over a 696-h period using a grid spacing of 1 cm. The amounts of measured irrigation and calculated evaporation were used as a time-variable boundary for the soil surface. Chemicals ([Cl.sup.-] and [K.sup.+]) were introduced to the model domain through the amount and concentration of effluent on the day of application. The absolute minimum soil potential allowed at the soil surface (hCritA) was set at 10 000 cm. The maximum soil potential allowed at the soil surface was set to zero as [K.sub.sat] values (Table 2) were greater than the applied water flux density, and no surface water ponding was observed throughout the leaching study. The hydraulic parameters of each layer ([[theta].sub.r], [[theta].sub.s], [alpha], and n) were estimated by fitting soil-water retention data (Klute 1986) using the RETC software (van Genuchten et al. 1991). The [K.sup.+] sorption parameters ([k.sub.d], [beta], and [eta]) were estimated by fitting sorption isotherm data to the general non-linear sorption equation (see above), and the Langmuir-type equation ([beta] = 1) provided the best-fit to the [K.sup.+] sorption isotherm data ([r.sup.2] = 0.97; Fig. 1).

Evaluating model performance

The performance of HYDRUS to simulate cumulative drainage and cumulative loss of applied [K.sup.+], N[H.sub.4.sup.+]/N[O.sub.3.sup.-], and [Cl.sup.-] from the soil cores was assessed against experimental values using linear regression analysis, and the following goodness-of-fit parameters (Close et al. 2003; Sarmah et al. 2005):

Residual sum of squares (RSS) = [n.summation over i=1][([P.sub.i] - [O.sub.i]).sup.2]

where [O.sub.i] and [P.sub.i] are the observed and predicted values in sample i. The ideal value of RSS would be zero.

Coefficient of residual mass (CRM) = [[summation].sup.n.sub.i=1]([O.sub.i] - [P.sub.i])/[[summation].sup.n.sub.i=1]([O.sub.i])

This relationship compares the mass of water or ion leached from the soil core irrespective of its distribution in the leachate, and a CRM value close to zero indicates a good prediction of the mass of water or ion in the leachate.

Coefficient of determination (CD) = [[summation].sup.n.sub.i=1][([O.sub.i] - [P.sub.i]).sup.2]/[[summation].sup.n.sub.i=1][([P.sub.i] - [O.sub.i]).sup.2]

where [O.sub.m] is the mean of the observed data, and the other terms are as defined above. When CD = 1, the predicted data corresponds perfectly with the measured data.

Coefficient of model efficiency (CME) = [[summation].sup.n.sub.i=1][([O.sub.i] - [O.sub.m]).sup.2] - [[summation].sup.n.sub.i=1][([P.sub.i] - [O.sub.m]).sup.2]/[[summation].sup.n.sub.i=1][([O.sub.i] -[O.sub.m]).sup.2]

The CME is an indicator of the overall agreement between the measured and predicted data, and values of CME can vary from 1, where there is a perfect agreement between the measured and predicted data, to -[infinity]. A negative value of CME means that model predictions are no better than predictions using a constant equal to the mean measured value.

Results and discussion

Column leaching study 1: piggery wastewater applied to a Vertosol

The cumulative volumes of drainage measured from the Vertosol soil core, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 2a. HYDRUS predicted the cumulative volumes of water that drained from the soil core very closely, which was consistent with a CME value approaching unity, and a CRM value approaching zero (CME = 0.976 and CRM = 0.039; Table 3). Furthermore, the mean coefficient of determination ([r.sup.2]) using linear regression analysis was high ([r.sup.2] = 0.987), and the slope of the linear relationship between measured and predicted volumes was 1.067 (Table 3). This close agreement between measured and predicted drainage volumes shows HYDRUS accurately simulated water movement in these large, undisturbed soil cores that exhibited non-homogeneous textural and hydraulic properties, under intermittent leaching conditions. Although surface cracking was observed for the Vertosol, the lack of connectivity to greater depths, coupled with the absence of strong texture contrasts within the profile for both soil types studies, may be a contributing factor to the close relationship between measured and predicted values observed in this study.

The cumulative amounts of [Cl.sup.-] leached from the soil core as a function of cumulative drainage, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 2b. HYDRUS predicted the cumulative amounts of [Cl.sup.-] leached from the Vertosol very closely (CME = 0.997 and CRM = 0.039), with a linear slope of 1.048 and an [r.sup.2] = 0.985 when comparing measured and predicted amounts (Table 3). These data suggest that the movement of applied [Cl.sup.-] within the Vertosol was well described by the convection-dispersion equation for non-reactive chemicals.

The cumulative amounts of N[O.sub.3.sup.-] leached from the soil core as a function of cumulative drainage, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 2c. Cumulative amounts of N[H.sub.4.sup.+] were not presented because measured concentrations of this cation in the leachate were negligible (< [10.sup.-7] mmol/[cm.sup.3]; Phillips 2002b), and HYDRUS also predicted similar leachate N[H.sub.4.sup.+] concentrations.

HYDRUS predicted the cumulative amounts of N[O.sub.3.sup.-] leached from the Vertosol very closely (CME = 0.924 and CRM=-.028), with a slope of 1.146 and an [r.sup.2] = 0.960 (Table 3). HYDRUS simulates coupled N[H.sub.4.sup.+]-N[O.sub.3.sup.-] leaching firstly by removing the amount of applied N[H.sub.4.sup.+] sorbed by the soil (as predicted by the general non-linear sorption equation; Fig. 1), and then transforming solution N[H.sub.4.sup.+] to N[O.sub.3.sup.-] using a first-order reaction for which the rate of transformation is directly proportional to its solution concentration (Cho 1971; Simunek et al. 1999). Analysis of the nitrogen transformation data by Phillips (2002a) found [Sink.sub.liquid] = 0.003 (1/h) provided a good estimate of the rate constant for N[H.sub.4.sup.+] transformation into N[O.sub.3.sup.-] within the Vertosol soil core. Since the experimental data showed that little of the applied N[H.sub.4.sup.+] was retained within the soil core either adsorbed to the soil colloids or in the soil solution (Phillips 2002b), much of the adsorbed N[H.sub.4.sup.+] predicted by HYDRUS was allowed to be available for nitrification. Simulations with HYDRUS found that setting the parameter [Sink.sub.soil] equal to 0.00125 (1/h) provided the best fit to the measured N[O.sub.3.sup.-] data. This analogy assumes that much of the applied N[H.sub.4.sup.+] initially adsorbed by the soil colloids (largely via cation exchange reactions; Phillips 2002a) was subsequently desorbed (via cation exchange and changes in soil solution chemistry), transformed to N[O.sub.3.sup.-], and leached from the soil core.

[FIGURE 1 OMITTED]

Column leaching study 2: KCl applied to a Podosol

The cumulative volumes of drainage measured from the Podosol soil cores, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 3. HYDRUS predicted the cumulative volumes of water draining from each core very closely, which was reflected in the high values of CME (0.867-0.951) and corresponding low CRM (-0.062 to 0.133) (Table 3). Furthermore, the mean coefficient of determination ([r.sup.2]) using linear regression analysis was high ([r.sup.2] > 0.980), and the slope of the linear relationship ranged between 1.085 and 1.291 (Table 3). Cumulative drainage from each soil core was predicted using the same hydraulic input parameters (Table 2), which suggests that the structural integrity of each undisturbed soil core was very similar. This may not be unexpected given that the pine plantation area from where this very sandy Podosol was obtained had not been significantly disturbed for many years, the individual layers of the soil profile did not display strong texture contrasts, and no evidence of preferential pathways were evident within the soil profile at the time of sampling.

[FIGURE 3 OMITTED]

The cumulative amounts of [Cl.sup.-] leached from each Podosol soil core as a function of cumulative drainage, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 4a-c. HYDRUS predicted the cumulative amounts of [Cl.sup.-] leached from each of the 3 KCl treatments very closely. For the K50, K100, and K300 treatments, values of CME and CRM were close to unity (CME >0.97) and zero, respectively, with a slope from linear regression analysis ranging between 0.994 and 1.135, and [r.sup.2] values >0.99 (Table 3). These data demonstrate that the movement of applied [Cl.sup.-] within these soil cores was well-described by the convection--dispersion equation for non-reactive chemicals, and further supports the contention that water movement in each of the 4 undisturbed soil cores was very similar.

[FIGURE 4 OMITTED]

The cumulative amounts of [K.sup.+] leached from each core as a function of cumulative drainage, and those predicted by HYDRUS, over the duration of the experiment are presented in Fig. 4d-f In contrast to the cumulative drainage and [Cl.sup.-] data (Figs 3 and 4a-c), agreement between measured and predicted [K.sup.+] values was very poor, particularly for the lower applied [K.sup.+] treatments (K50 CME=-1.452 and K100 CME=0.274). Although the measured and predicted concentrations displayed closer agreement for K300 (CME = 0.678), the correlation between the 2 sets of data remains unacceptable.

The reason for the discrepancy between the measured and predicted cumulative [K.sup.+] data (i.e. HYDRUS predicted longer retention times within the soil core and greater leaching losses relative to the measured values) is unclear, given the close agreement for water and [Cl.sup.-] transport. Possible reasons for the discrepancy may include: (1) the measured leachate data was in error, (2) the fitted sorption isotherm parameters ([k.sub.d], [beta], and [eta]) were in error, and/or (3) applied [K.sup.+] was behaving in the soil in a manner not well-described by HYDRUS.

There was no evidence to suggest the measured [K.sup.+] concentrations in the leachate were in error as all analyses were done in accordance with standard analytical and instrumental procedures, and internal standards were incorporated for quality assurance. The coefficients for describing [K.sup.+] sorption were obtained by fitting the Langmuir form of the general non-linear equation (Simunek et al. 1999) to the sorption isotherm data. The Langmuir equation provided an excellent ([r.sup.2] =0.97; Fig. 1) fit to the data, so the values assigned to the sorption parameters were considered to be appropriate for use in HYDRUS. The Podosol is classified texturally as a sand, with quartz as the dominant mineralogical component, and exhibits a net negative surface charge (Table 1). Consequently, this soil was unlikely to contain sorption surfaces that would require invoking non-equilibrium sorption conditions within HYDRUS, which would in fact exacerbate the retardation of [K.sup.+] movement.

To obtain the best fit to the measured data, the parameter F (a dimensionless fraction of the sorption sites subject to instantaneous sorption) was required to be adjusted between 1 and 0.5, with the optimum value being F = 0.75 (Table 2). By setting F = 0.75 only 75% of the potential [K.sup.+] sorption sites were available (Simunek et al. 1999), which resulted in HYDRUS predicting greater leaching of applied [K.sup.+], and a better fit to the measured data. By having to set F < 1 for the column study suggests the batch technique may have overestimated [K.sup.+] sorption. Thus, while the sorption parameters ([k.sub.d], [beta], and [eta]) described [K.sup.+] sorption in the batch tests very well (Fig. 1), these parameters may not reflect the sorption process occurring within the large undisturbed soil core (e.g. Bond and Phillips 1990). Furthermore, unpublished work by the author found HYDRUS successfully predicted [K.sup.+] absorption in this Podosol during unsteady, unsaturated soil-water flow in horizontal columns using data by Phillips (2004). The absorption experiment was undertaken using a surface flux density of 0.398 cm/h, and soil from the <0.5mm size-fraction, and these conditions would have been more conducive to maximising [K.sup.+] contact with the soil colloids due to slower pore-water velocities and greater reactive surface area, relative to the large undisturbed soil cores. If the applied [K.sup.+] was not directly available to the total sorption sites, then this would support the use of an F value less than unity (i.e. F = 0.75) when modelling [K.sup.+] leaching in the undisturbed Podosol cores. However, because simulations with F = 0.75 still did not adequately predict [K.sup.+] leaching from the Podosol soil core, this mechanism was not the only reason for the discrepancy between measured and predicted data.

Another mechanism that may have affected [K.sup.+] behaviour concerns the equilibrium [K.sup.+] solution activity ratio. Phillips et al. (1988a, 1988b) showed that [K.sup.+] leaching with no net change in the adsorbed fraction occurs once the equilibrium [K.sup.+] solution activity ratio has been reached. This condition would result in faster [K.sup.+] transport than predicted using the soil retardation factor, and, consequently, HYDRUS would predict greater retention of applied [K.sup.+] within the soil core. This effect would be more pronounced at the lower applied [K.sup.+] concentrations as in these treatments, the equilibrium [K.sup.+] solution activity ratio would be achieved much sooner, resulting in greater discrepancy between the measured and predicted values (compare Fig. 4d-f).

Conclusions

The water and chemical transport model HYDRUS predicted very closely the behaviour of water and non-reactive chemicals in large, undisturbed cores of 2 contrasting soil types. The absence of strong texture heterogeneities within the profile of both soil types may be a contributing factor to the close relationship between measured and predicted values observed in this study. For the well-structured Vertosol, HYDRUS also simulated the transformation, and subsequent leaching, of applied N[H.sub.4.sup.+] to N[O.sub.3.sup.-]. In contrast, agreement between measured and predicted [K.sup.+] leaching behaviour in a coarse-textured Podosol was poor. Possible reasons were forwarded to explain this discrepancy between measured and predicted values; however, no obvious cause or mechanism could be confirmed. The ability of HYDRUS to accurately simulate water and non-reactive chemical transport in soil agrees with previous studies, and suggests it could be an additional tool for persons responsible for evaluating the environmental risks associated with land disposal of nitrogen-rich wastewater. However, its use in predicting reactive chemical transport in undisturbed soil cores and under field conditions warrants further investigation.

References

Bond WJ, Phillips IR (1990) Cation exchange isotherms obtained with batch and miscible-displacement techniques. Soil Science Society of America Journal 54, 722-728.

Cho CM (1971) Convective transport of ammonium with nitrification in soil. Canadian Journal of Soil Science 51, 339-350.

Close ME, Pang L, Magesan GN, Lee R, Green SR (2003) Field study of pesticide leaching in an allophonic soil in New Zealand. 2: Comparison of simulations from four leaching models. Australian Journal of Soil Research 41,825-846. doi: 10.1071/SR02081

van Genuchten MTh, Leij FJ, Yates SR (1991) The RETC code for quantifying the hydraulic functions of unsaturated soils. US Environmental Protection Agency, R. S. Kerr Environmental Research Laboratory, Office of Research and Development, Ada, OK.

Hassan G, Reneau RB, Hagedorn C, Saluta M (2005) Modeling water flow behavior where highly treated effluent is applied to soil at varying rates and dosing frequencies. Soil Science 170, 692-706. doi: 10.1097/01.ss.0000185911.10836.07

Isbell RF (1996) 'The Australian Soil Classification.' (CSIRO Publishing: Collingwood, Vic.)

Klute A (1986) Water retention: Laboratory methods. In 'Methods of soil analysis. Part 1. Physical and mineralogical methods'. (Ed. A Klute) pp. 635-562. (American Society of Agronomy: Madison, WI)

Magesan GN, Vogeler I, Clothier BE, Green SR, Lee R (2003) Solute movement through an allophonic soil. Journal of Environmental 2Quality 32, 2325-2333.

Pang L, Close ME, Watt JPC, Vincent KW (2000) Simulation of picloram, atrazine, and simazine leaching through two New Zealand soils and into groundwater using HYDRUS-2D. Journal of Contaminant Hydrology 44, 19-46. doi: 10.1016/S0169-7722(00)00091-7

Phillips IR (2002a) Phosphorus sorption and nitrogen transformations in soils treated with piggery wastewater. Australian Journal of Soil Research 40, 335-349. doi: 10.1071/SR01040

Phillips IR (2002b) Nutrient leaching in large undisturbed soil cores following surface applications of piggery wastewater. Australian Journal of Soil Research 40, 515-532. doi: 10.1071/SR01040

Phillips IR (2004) Measurement and prediction of potassium chloride movement in an unsaturated sand. Communications in Soil Science and Plant Analysis 35, 1663-1679.

Phillips IR, Burton ED (2005) Nutrient leaching in undisturbed cores of an acidic sandy Podosol following simultaneous potassium chloride and di-ammonium phosphate application. Nutrient Cycling in Agroecosystems 73, 1-14. doi: 10. I007/s10705-005-6080-8

Phillips IR, Black AS, Cameron KC (1988a) Effects of cation exchange on the distribution and movement of cations in soils with variable charge. I. Effects of lime on potassium and magnesium exchange equilibria. Fertilizer Research 17, 21-30. doi: 10.1007/ BF01050454

Phillips IR, Black AS, Cameron KC (1988b) Effects of cation exchange on the distribution and movement of cations in soils with variable charge. II. Effects of lime and phosphate on potassium and magnesium leaching. Fertilizer Research 17, 31-46. doi: 10.1007/BF01050455

Rassam D, Simunek J, van Genuchten MTh (2003) 'Modelling variably saturated flow with HYDRUS-2D.' (ND Consult: Brisbane, Qld)

Sarmah AK, Close ME, Pang L, Lee R, Green SR (2005) Field study of 2pesticide leaching in a Himatangi sand (Manawatu) and a Kiripaka bouldery clay loam (Northland). 2. Simulation using LEACHM, HYDRUS-1D, GLEAMS, and SPASMO models. Australian Journal of Soil Research 43, 471-489. doi: 10.1071/SR04040

Schmalz B, Lennartz B, van Genuchten MTh (2003) Analysis of unsaturated water flow in a large sand tank. Soil Science 168, 3-14. doi: 10.1097/00010694-200301000-00002

Simunek J, van Genuchten MTh (1999) Using the HYDRUS-1D and HYDRUS-2D codes for estimating unsaturated soil hydraulic and solute transport parameters. In 'Characterization and measurement of the hydraulic properties of unsaturated porous media'. (Eds MTh van Genuchten, FJ Leij, L Wu) pp. 1523-1536. (University of California: Riverside, CA)

Simunek J, Sejna M, van Genuchten MTh (1999) 'Simulating water flow and solute transport in two-dimensional variably saturated media.' (International Ground Water Modeling Center: Golden, CO)

Starr JL, Sadeghi AM, Pachepsky YA (2005) Monitoring and modelling lateral transport through a large in situ chamber. Soil Science Society of America Journal 69, 1871-1880. doi: 10.2136/sssaj2004.0162

Manuscript received 8 August 2005, accepted 6 December 2005

I. R. Phillips

School of Environmental Engineering, Griffith University, Nathan, Qld 4111, Australia.

Email: I.Phillips@griffith.edu.au
Table 1. Selected physical and chemical properties of each soil
(Phillips 2002a, 2002b; Phillips and Burton 2005)

Parameter Vertosol

 0-0.20 m 0.20-0.55 m

[[rho].sub.b] (g/[cm.sup.3]) 1.1 1.3
pH(water) 4.95 5.53
EC ([micro]S/cm) 363 272
2 M KCl N[H.sub.4.sup.+] -N
 (mg/kg) 31.61 2.56
2 M KCl N[0.sub.3.sup.-] -N
 (mg/kg) 11.38 7.39
Organic C (%) 2.41 0.98
Hydrous Fe oxide (mg/kg) 9200 9100
Hydrous Al oxide (mg/kg) 2400 3200
Exchang. ions ([cmol.sub.c]/kg)
 [Ca.sup.2+] 19.71 12.88
 [Mg.sup.2+] 5.02 6.80
 [K.sup.+] 2.50 0.77
 [Na.sup.+] 0.52 0.47
 [Al.sup.3+] 0.02 <0.01
ECEC ([cmol.sub.c]/kg) 27.77 20.94

Parameter Podosol

 0-0.30 m 0.30-0.50 m 0.50-0.85 m

[[rho].sub.b] (g/[cm.sup.3]) 1.0 1.6 1.6
pH(water) 3.34 3.45 3.59
EC ([micro]S/cm) 118 187 14
2 M KCl N[H.sub.4.sup.+] -N
 (mg/kg) 7.18 0.39 0.22
2 M KCl N[0.sub.3.sup.-] -N
 (mg/kg) 0.86 0.87 0.62
Organic C (%) 1.48 0.27 0.05
Hydrous Fe oxide (mg/kg) 190 96 50
Hydrous Al oxide (mg/kg) 220 121 40
Exchang. ions ([cmol.sub.c]/kg)
 [Ca.sup.2+] 1.123 0.194 0.005
 [Mg.sup.2+] 0.239 0.152 0.023
 [K.sup.+] 0.012 0.005 0.001
 [Na.sup.+] 0.008 0.004 0.001
 [Al.sup.3+] 0.735 0.556 0.140
ECEC ([cmol.sub.c]/kg) 2.117 0.911 0.170

Table 2. Input parameters used for running HYDRUS for each soil

[[theta].sub.r], residual soil-water content; [[theta.sub.s],
saturated soil-water content; [alpha] and n, fitting parameters for van
Genuchten model; [K.sub.sat], saturated soil hydraulic conductivity;
[D.sub.L], longitudinal dispersivity; [D.sub.T], transverse
dispersivity; [k.sub.d], [eta], and [beta], fitting parameters for
the non-linear sorption equation; F, dimensionless fraction of the
sorption sites subject to instantaneous sorption; [Sink.sub.soil],
first-order transformation rate constant for solid phase
in the nitrification chain reaction; [Sink.sub.liquid], first-order
transformation rate constant for dissolved phase in the nitrification
chain reaction

 [[theta]. [[theta].
Depth Ion sub.r] sub.s] [alpha]
(m) (l/cm)
 ([cm.sup.3]/[cm.sup.3])

 Vertosol

0-0.20 [Cl.sup.-] 0.24 0.48 0.026
0.20-0.55 [Cl.sup.-] 0.25 0.47 0.037
0-0.20 N[H.sub.4.sup.+] 0.24 0.48 0.026
0.20-0.55 N[H.sub.4.sup.+] 0.25 0.47 0.037

 Podosol

0-0.30 [Cl.sup.-] 0.04 0.47 0.037
0.30-0.50 [Cl.sup.-] 0.03 0.46 0.029
0.50-0.85 [Cl.sup.-] 0.02 0.45 0.037
0-0.30 [K.sup.+] 0.04 0.47 0.037
0.30-0.50 [K.sup.+] 0.03 0.46 0.029
0.50-0.85 [K.sup.+] 0.02 0.45 0.037

 [K.sub.
Depth Ion n sat] [D.sub.L] [D.sub.T]
(m) (cm/h)
 ([cm.sup.2]/h)

 Vertosol

0-0.20 [Cl.sup.-] 1.569 1.46 20 1
0.20-0.55 [Cl.sup.-] 2.027 3.62 20 1
0-0.20 N[H.sub.4.sup.+] 1.569 1.46 20 1
0.20-0.55 N[H.sub.4.sup.+] 2.027 3.62 20 1

 Podosol

0-0.30 [Cl.sup.-] 2.145 25 30 1
0.30-0.50 [Cl.sup.-] 2.916 35 30 1
0.50-0.85 [Cl.sup.-] 3.312 35 30 1
0-0.30 [K.sup.+] 2.145 25 30 1
0.30-0.50 [K.sup.+] 2.916 35 30 1
0.50-0.85 [K.sup.+] 3.312 35 30 1

Depth Ion [k.sub.d] [eta] [beta] F
(m) ([cm.sup.3]/g)

 Vertosol

0-0.20 [Cl.sup.-] -- -- -- 1
0.20-0.55 [Cl.sup.-] -- -- -- 1
0-0.20 N[H.sub.4.sup.+] 34.6 532.6 1 1
0.20-0.55 N[H.sub.4.sup.+] 32.4 315.9 1 1

 Podosol

0-0.30 [Cl.sup.-] -- -- -- 1
0.30-0.50 [Cl.sup.-] -- -- -- 1
0.50-0.85 [Cl.sup.-] -- -- -- 1
0-0.30 [K.sup.+] 0.88 38.8 1 0.75
0.30-0.50 [K.sup.+] 0.68 64.6 1 0.75
0.50-0.85 [K.sup.+] 0.49 110.0 1 0.75

Depth Ion [Sink.sub.soil] [Sink.sub.liquid]
(m) (l/h)

 Vertosol

0-0.20 [Cl.sup.-] -- --
0.20-0.55 [Cl.sup.-] -- --
0-0.20 N[H.sub.4.sup.+] 0.0012 0.003
0.20-0.55 N[H.sub.4.sup.+] 0.0012 0.003

 Podosol

0-0.30 [Cl.sup.-] -- --
0.30-0.50 [Cl.sup.-] -- --
0.50-0.85 [Cl.sup.-] -- --
0-0.30 [K.sup.+] -- --
0.30-0.50 [K.sup.+] -- --
0.50-0.85 [K.sup.+] -- --

Table 3. Linear regression analysis and coefficient of modelling
efficiency (CME) for measured (M) and predicted (P) cumulative
drainage, cumulative [Cl.sup.-] and cumulative N[O.sub.3.sup.-] for
the Vertosol, and cumulative drainage, cumulative [Cl.sup.-], and
cumulative [K.sup.+] for the Podosol soil cores

Parameter Treatment Linear regression [r.sup.2]
 equation

Vertosol

Drainage Effluent P = 1.067 M - 3.05 0.987
[Cl.sup.-] Effluent P = 1.048 M - 7.17 0.985
N[O.sub.3.sup.-] Effluent P = 1.146 M - 7.45 0.960

Podosol

Drainage K0 P = 1.174 M - 2.33 0.993
[Cl.sup.-] K0 n.d. n.d.
[K.sup.+] K0 n.d. n.d.
Drainage K50 P = 1.100 M - 4.45 0.982
[Cl.sup.-] K50 P = 1.135 M - 0.536 0.994
[K.sup.+] K50 P = 1.046 M - 1.045 0.712
Drainage K100 P = 1.291 M - 3.73 0.981
[Cl.sup.-] K100 P = 1.051 M - 0.117 0.995
[K.sup.+] K100 P = 1.814 M - 1.037 0.844
Drainage K300 P = 1.085 M - 3.64 0.987
[Cl.sup.-] K300 P = 0.994 M - 2.077 0.992
[K.sup.+] K300 P = 1.297 m - 2.866 0.879

Parameter CME CRM CD RSS

Vertosol

Drainage 0.976 0.039 0.863 150
[Cl.sup.-] 0.977 0.039 0.894 1603
N[O.sub.3.sup.-] 0.924 -0.028 0.730 4490

Podosol

Drainage 0.951 -0.045 0.715 61
[Cl.sup.-] n.d. n.d. n.d. n.d.
[K.sup.+] n.d. n.d. n.d. n.d.
Drainage 0.897 0.133 0.767 141
[Cl.sup.-] 0.972 -0.025 0.770 3
[K.sup.+] -1.452 0.543 0.282 17
Drainage 0.867 -0.062 0.583 13
[Cl.sup.-] 0.987 -0.040 0.897 8
[K.sup.+] 0.274 0.183 0.399 8
Drainage 0.942 0.104 0.815 86
[Cl.sup.-] 0.975 0.087 0.988 95
[K.sup.+] 0.678 0.023 0.522 147

n.d., Not determined.
COPYRIGHT 2006 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Author:Phillips, I.R.
Publication:Australian Journal of Soil Research
Geographic Code:8AUST
Date:May 1, 2006
Words:6060
Previous Article:Surface and sub-surface salinity in and around acid sulfate soil scalds in the coastal floodplains of New South Wales, Australia.
Next Article:Palaeochannels in Northern New South Wales: inversion of electromagnetic induction data to infer hydrologically relevant stratigraphy.
Topics:

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters