| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
im
nekc
a Institute of Environmental Science & Research Ltd, P.O. Box 29181, Christchurch, New Zealand
b Landcare Research, Private Bag 3127, Hamilton, New Zealand
c Dep. of Environmental Sciences, Univ. of California, Riverside, CA 92507
* Corresponding author (Liping.pang{at}esr.cri.nz).
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Received 8 June 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
Abbreviations: BTC, breakthrough curve • cfu, colony-forming unit, a unit indicating the number of bacteria present in a water sample • DSE, dairy-shed effluent • PV, pore volume
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
Soils act as natural filters that can attenuate microbial contaminants, but they vary widely in their ability to remove them. The ability of New Zealand soils to remove microorganisms has been evaluated previously in a few studies. Wells (1973) rated the properties of New Zealand soils in relation to effluent disposal and concluded that young (2000–7000 yr of age), fine-grained tephric soils had the best characteristics for effluent disposal, while the gleyed soils were unsuitable because of their anaerobic conditions. There were no experimental data to support this conclusion, however. Childs et al. (1977) performed some lysimeter leaching experiments examining the ability of some New Zealand soils (peat, clay, silt loam, sandy loam, sand, and stony soil) to remove chemicals and fecal coliforms from sewage effluent. They found that concentrations of fecal coliforms derived from effluent were >100 colony-forming units (cfu) L–1 in leachate collected at 50-cm depth for all soils under both spray and flood irrigation, except for dune sand and a silt loam derived from volcanic ash under spray irrigation. In a field study of a border strip (flood) effluent irrigation scheme in Canterbury, New Zealand, Sinton et al. (2005) observed rapid leaching of fecal coliforms and F-RNA phages through alluvial gravel soils (15.7–39.2 m h–1), and the estimated reduction in the bacteria and phages was only log10(cmax/co) = 1.42 and 2.63, respectively (the log10 reduction in maximum concentration compared with original concentration), through a 16.8-m vadose zone. In recent years, lysimeter studies were performed to examine the leaching of fecal coliforms and bacteriophages through a range of key New Zealand soils commonly used for effluent disposal (Aislabie et al., 2001; McLeod et al., 2001, 2003, 2004). The results of these studies suggest that the vertical movement of bacteria and viruses varies significantly with soil type.
Although experimental data were collected in some of these studies, no modeling work was performed. Thus parameters (such as removal rates) that describe the fate and transport of microbes in New Zealand soils remains unknown, limiting the applicability of the data in the field. Although some overseas studies have derived parameters of microbial transport in undisturbed soils (Shelton et al., 2003; Guber et al., 2005; Frazier et al., 2002), the physicochemical properties of some New Zealand soils (e.g., allophanic soils) are quite unique. In addition, it is difficult to compare the results from different studies because the removal of microbes in soils changes according to the experimental conditions, especially the rate of irrigation (Guber et al., 2005).
Contaminant transport models are useful tools for the quantitative assessment of microbial transport in soils and the elucidation of the importance of factors that control microbial concentrations in receiving waters. By calibrating transport models with observed data, we can characterize microbial attenuation and transport in a specific soil system using descriptive parameters (e.g., removal, attachment, and detachment rates). These calibrated parameter values together with transport models can then be used by others (e.g., government agencies, regional authorities, researchers, and consultants) to manage and predict similar systems for various purposes (e.g., resource management, land-use planning, design of monitoring programs, and risk analysis).
New Zealand has a diverse landscape with a wide range of contrasting soil types occurring across short distances (McLeod et al., 2001). As dairy farming is widespread across the nation, it is important to evaluate the leaching of microorganisms in the different soils in New Zealand so that best farming practices can be implemented. The objectives of this study were (i) to better understand microbial transport in major New Zealand soils under DSE irrigation, and (ii) to establish values of important parameters for describing microbial transport and attenuation in these soils by applying an appropriate contaminant transport model to the data obtained from lysimeter experiments generated in this and previous studies (Aislabie et al., 2001; McLeod et al., 2001, 2003, 2004). These results will be of critical importance to regulatory agencies that need scientific information about microbial removal in different soils to enable them to effectively manage development and evaluate the risk of groundwater contamination.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
|
|
Lysimeters and Leaching Experiments
The methods used for obtaining intact cores of Lismore and Templeton soils and leaching experiments were the same as those used for the other soils by Aislabie et al. (2001) and McLeod et al. (2001, 2003, 2004). The description below applies for all 10 soils used for modeling in this study.
Undisturbed soil cores were hand carved in situ from the ground surface, and the lysimeter casings were progressively pressed down over the cores. The lysimeters were made of high-density polyethylene (HDPE) pipe, with an internal diameter of 46 cm and a length of 47 to 70 cm. A 1-cm internal annulus was filled with petroleum jelly to prevent water preferentially flowing at the soil-casing interface. The bottom of the lysimeter was welded to a 1-cm-thick HDPE base plate, with a sampling port installed in the center to allow the collection of leachate. The pore volume (PV) for each set of lysimeter cores was estimated from previous analyses of the total porosity of these soils at the same site. One PV is the amount of space in the soil core occupied by soil pores or cracks. This ranged from 45 to 67% of the total soil volume (Table 2).
In the laboratory, the lysimeters were initially irrigated for 5 d with tap water to field saturation with leachate emanating from the sampling port, then were allowed to drain for 7 d. Seven days is similar to the return period commonly used at effluent irrigation sites. Following this, each lysimeter was irrigated, at a constant rate of 5 mm h–1, with a 25-mm depth of DSE, containing native fecal coliforms (typically 105–109 cfu mL–1) and spiked with a tracer solution containing Salmonella bacteriophage (109 plaque-forming units [pfu] mL–1) and Br (2 g L–1). The lysimeters were then irrigated continuously with tap water at a rate of 5 mm h–1 using a drip-type rainfall simulator with drippers spaced on a 20-mm triangular grid, approximately 170 mm above the soil surface. Typically the chemical properties of the tap water were as follows: pH = 7.7, electrical conductivity = 21.2 mS m–1, and ionic strength = 0.003 mol L–1. Experiments were performed on triplicate intact cores for each soil type.
To obtain the background levels of Br and microbial tracers in the leachate, samples were first taken at the end of the wetting period. During the leaching experiments, soil leachate was collected into sterile bottles and subsampled. Subsamples for microbial assays were stored at 4°C and analyzed within 2 h of collection. Bromide tracer samples were analyzed within 1 wk of collection.
Batch tests were also performed in the current study to determine the inactivation rates of the microbial indicators in DSE under similar experimental conditions to those used in the lysimeter studies (background solution, pH, concentrations of microbial indicators, and Br). The reactors were incubated at 15°C in the dark for 14 to 17 d. Phages were sampled daily for 14 d, while fecal coliforms were sampled twice a day for the first 5 d and then once a day between Days 5 and 17.
Sample Analysis
The bacteriophage propagation and assay method used was detailed in McLeod et al. (2001). Essentially, Phage 28B was grown overnight on its host strain S. typhimurium Type 5 in tryptic soy broth at 37°C. The phages were isolated by chloroform lysis of the bacterial host, then passed through a 0.45-µm mixed cellulose ester-based membrane filter to remove cell debris. To obtain a clean virus preparation free of organic material, the filtrate was centrifuged at 25,000 x g (Sorval T21, Thermo Fischer Scientific, Waltham, MA) for 2 h at 4°C, the supernatant was poured off, and the phages were resuspended in 1 to 2 mL of phage storage buffer (Sambrook et al., 1989) and stored at 4°C until required. Phage stocks were enumerated using a soft agar overlay method. Leachate samples were mixed with the host-strain culture and poured onto nutrient agar plates. After incubation for 18 to 24 h at 37°C, well-formed, clear plaques were counted and reported as plaque-forming units per milliliter. Each reported phage concentration is the average of three replicates.
Fecal coliforms were determined in soil leachates using a membrane filtration technique (American Public Health Association, 1998). Samples were diluted in phosphate-buffered water (pH 7.0), then filtered according to standard procedures. The filters were placed on mFC agar (Difco, BD Diagnostics, Auckland, NZ) and blue colonies were counted after incubation for 24 h at 44.5°C. Bromide concentrations in the leachate samples were measured using an ion-selective electrode (Metrohm 6.0502.100, Herisau, Switzerland).
Modeling and Data Analysis
The Richards equation (Richards, 1931) was used to describe variably saturated one-dimensional vertical water flow in the soil lysimeters:
 | [1] |
is the volumetric water content [L3 L–3], t is time [T], x is the spatial coordinate [L], and K(h) is the unsaturated hydraulic conductivity function [L T–1]. The variable K(h) is described by the van Genuchten–Mualem hydraulic model (van Genuchten, 1980). The parameters used in the van Genuchten–Mualem hydraulic model (Table 3
) were obtained by analyzing measured water retention curves ( vs. h) using the RETC program (van Genuchten et al., 1991) or by using soil texture data and the Rosetta program (Schaap et al., 2001) if water retention data were absent.
|
 | [2] |
 | [3] |
= m + im is the volumetric moisture content [L3 L–3]; D is the dispersion coefficient [L2 T–1], which is the product of dispersivity 
[L] and the spatial coordinate x (i.e., D = x); q is the water flux [L T–1]; and 
is the first-order mass-exchange coefficient [T–1] governing the rate of solute exchange between the mobile and immobile water regions. The two-region mobile–immobile model (MIM) assumes that water flow and contaminant transport is limited to the mobile water region and that water in the immobile water region is stagnant, with a first-order diffusive exchange process between the two regions.
In our recent unpublished batch study, Br was adsorbed in soils that contain volcanic materials (Waihou allophanic soil, Atiamuri pumice soil, and Hamilton clay loam). Therefore, the distribution coefficient for linear adsorption, Kd [L3 L–1], was also considered for these soils. Assuming all adsorption sites equilibrate with the mobile water phase, Eq. [2] becomes
 | [4] |
b is the bulk density of the soil material [M L–3]. The Kd values determined from the unpublished batch study were used in the model.
We assumed that microbes were excluded from the immobile water region (thus no mass exchange between two regions, = 0), and thus their attachment and inactivation occurred only in the mobile water region. Under this assumption, the two-region MIM for describing transport of microbes at a constant flow is expressed as
 | [5] |
 | [6] |
The inactivation rate for the microbial indicators in the liquid phase (µ) was determined by fitting the experimental data of the inactivation batch tests with an exponential function, c = co exp(–µt), where co is the influent concentration, using the Solver function of the Microsoft Excel spreadsheet.
Since an equal flux is applicable for both Br and microbial tracers, q = (mvm)microbe = (mvm)Br, where v is the pore-water velocity, velocity enhancement of a microbial tracer can be measured from the ratio
 | [7] |
imnek et al., 2005). The HYDRUS model consisted of four to five layers for a lysimeter, each with variable inputs of soil hydraulic properties (using the van Genuchten–Mualem model) and bulk density. Values of , m (via im), and were optimized for the Br data, while , m (again via im), katt, and kdet were optimized for the microbial data. The measured inactivation rates in the liquid phase for the bacteria and phages, µ, were fixed in the model. Single values of , m, , katt, and kdet were assigned for all layers. The distinctive transport patterns of Br and the microbial tracers (Fig. 1
), as a result of size exclusion in microbial transport, meant that and m for the microbes had to be independently estimated from those for Br. HYDRUS-1D was modified for this study to simultaneously assign the same values of optimized parameters to all layers and to keep the mobile water content the same in all layers. The reason for using of the same parameter values for multiple layers was to reduce the number of parameters to be optimized. This treatment is reasonable, as all BTC data were obtained from one sampling point at the end of the lysimeters.
|
|
|
Results and Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
In contrast, the peak concentrations of Br BTCs arrived later than 1 PV in Waihou allophanic soil. This later breakthrough of Br indicates that Br is not conservative in allophanic soil. Results of a batch test (our recent unpublished data) showed that Br is adsorbed on Waihou allophanic soil, with a distribution coefficient (Kd) of 0.06 L kg–1 in the top 15 cm of soil and 0.33 L kg–1 in the subsoil (these values were used in the modeling). Sorption of Br in an allophanic soil could be explained by its soil surface charge. Amorphous allophanic clays have a high isoelectric point of pH 6.0 (Cooper and Morgan, 1979), and Waihou soils have a topsoil field pH that is typically less than that (Table 2), so the allophane has a net positive surface charge (McLeod et al., 2001) and thus has an affinity for anions. Sorption of Br in allophanic soils has already been reported (Close et al., 2003). Although Br was also adsorbed in Atiamuri pumice soil (Kd = 0.01 L kg–1 for topsoil and 0.03 L kg–1 for subsoil) and Hamilton clay loam (Kd = 0.00 for topsoil and 0.03 L kg–1 for subsoil), the degree of adsorption was not high enough to cause a shift in the concentration peak.
Model-simulated Br concentrations are compared with observed concentrations in Fig. 2 , selecting one lysimeter per soil type. The MIM-derived dispersivity for Br transport (Table 5 ) is on average 7% of the lysimeter length (SD = 6%, n = 29). Although there is a lot of variability in the data, dispersivity of Br transport appears to be inversely correlated with the measured macroporosity (Fig. 3 ). This is expected, as with an increase in macropores in the soils, convection flow would be more dominant and less dispersion would occur.
|
|
|
m/, (Table 5) accessible for Br transport is on average 0.45 (SD = 0.23, n = 29). The m/ values are lowest for Netherton clayey soil and Hamilton clay loam, implying that, in highly heterogeneous soils, only a small fraction of water is needed to leach solute contaminants. This suggests that the fraction of mobile water reflects the degree of soil heterogeneity. The fact that the pattern of model-predicted mobile water content agrees with measured macroporosity (Fig. 3) also supports this comment.
The mass exchange rate, , (Table 5) is much higher for allophanic and pumice soils than for other soils, indicating that Br transport is not at equilibrium in volcanic soils. In contrast, the value is close to zero for dune sand, suggesting that Br transport in uniform dune sand was under equilibrium conditions. The relatively symmetric shape of Br BTCs with little tailing for this soil also supports this conclusion. The relationship between and macroporosity is not clear in this study (Fig. 3).
Transport of Fecal Coliforms and Bacteriophages
Breakthrough Curves
The BTCs of microbial tracers show a very different pattern from those of Br (Fig. 1). In comparison with Br, the BTCs of microbial tracers commonly peak much earlier and are much less spread out. Many others (e.g., Shelton et al., 2003; Guber et al., 2005; Levy et al., 2007) have also observed an earlier breakthrough (or velocity enhancement) of microbes than conservative solutes in undisturbed soils. Except for Waitarere dune sand and phages in one lysimeter in Atiamuri pumice soil, all microbial BTCs finished ahead of the front of the Br BTCs. This suggests that microbes are transported through a macropore network and that the soil matrix has little impact on their transport. Similarly, others have also commented that microbial transport occurs primarily through interconnected large pores (Wollum and Cassel, 1978) and almost follows a piston flow (Germann et al., 1987). The effect of macropores on colloid transport in intact cores was also suggested in the review of DeNovio et al. (2004).
Figure 4 shows that MIM-simulated BTCs fit well the observed data. As it is superfluous to present all model-simulated BTCs here, we present only one lysimeter for one soil type in Fig. 4, which was typical (representative) of other BTCs.
|
Mobile–Immobile Water, Velocity Enhancement
In comparison with Br, the fraction of mobile water, m/ (Table 6
), is much smaller for microbial BTCs that exhibit velocity enhancement (i.e., excluding Waitarere sandy recent and Atiamuri pumice soil), on average only 0.19 (SD = 0.09, n = 29), in a narrow range of 0.08 to 0.24, except for Manawatu fine sandy loam (0.40–0.45). This is because microbes can only access a smaller range of larger pores due to size exclusion, unlike the Br solute, which can access a wider range of pore sizes. Slower velocity zones near the solid surfaces, which are accessible to Br and might be considered mobile for Br, would become inaccessible to microbes and would thus be treated as immobile. Water will then travel faster if going through a smaller porosity. Shelton et al. (2003) estimated that the porosity available for fecal coliform transport in undisturbed stony silt loam is 15% of the total water content, which is similar to our findings. In contrast, m/ is much larger for microbial BTCs that arrived later than Br (Table 6), on average 0.80 (SD = 0.10, n = 7). Like the results from Br, the pattern of predicted mobile water content agrees with the measured macroporosity (Fig. 3).
|
m-Br/m-microbe = vm-microbe/vm-Br ratios derived from MIM (Table 7
) suggest that the transport of the microbes was always faster (mostly two to four times) than that of Br, except in Waitarere dune sand and one lysimeter of Atiamuri pumice soil. This indicates that velocity enhancement is a common phenomenon in the structured soils investigated here because the microbes are excluded from smaller pores in the soil medium. The degree of velocity enhancement, as measured by m-Br/m-microbe, is in the order of: Waikiwi silt loam (4) > Lismore shallow silt loam, Waikoikoi silt loam, and Templeton silt loam (2–4) > Hamilton clay loam, Manawatu fine sandy loam, and Netherton clayey soil (1–2) > Waitarere sandy recent soil (<1). For a particular lysimeter, there was little difference in speed between bacteria and phages as reflected in their similar m-Br/m-microbe ratios (Table 7). Our finding on the faster velocity of microbes than Br is similar to the observation of Shelton et al. (2003), who conducted a lysimeter study. They found the average velocity of coliform bacteria leaching from 90-cm-long undisturbed stony soil was about seven times greater than the average pore velocity.
|
m-Br/m-microbe ratio, relates to the mean velocity or center of mass. If the relative speed is examined using the ratio of pore volume at peak concentration of Br to that of a microbial tracer (Table 4), which reflects macropore velocity (De Jonge et al., 2004), the relative speed for microbial transport is much greater than for Br. It is in the order of Hamilton clay loam (21–87) > Lismore shallow silt loam over gravels, Waikiwi silt loam, and Waikoikoi silt loam (6–14) > Templeton silt loam (4–5) > Manawatu fine sandy loam and Atiamuri pumice soil (1–2) > Waitarere sandy recent soil (<1). The ratio for heterogeneous Netherton clayey soil varies between 2 and 9.
The m-Br/m-microbe ratios for Waitarere dune sand were <1 (Table 7), suggesting that the transport velocity of microbes in Waitarere dune sand was slower than that of Br. This is probably because this soil has only minor pedological development, with a single-grain structure throughout. It is therefore more uniform than other soils, thus matrix flow plays an important role in microbial transport in the dune sand. The m-Br/m-microbe ratio was also <1 for one lysimeter of Atiamuri pumice soil, suggesting that transport of phages was slower than that of Br in one of the pumice soil lysimeters.
Dispersion
Dispersivity values derived from most microbial BTCs (Table 6) are much smaller than those from Br BTCs (average Br/microbe = 4.69, SD = 3.85, n = 35, Table 7) as a result of a narrower pore network involved in microbial transport. The lower dispersion in microbial transport compared with solute transport was also observed in the lysimeter study of Shelton et al. (2003) with an undisturbed silt loam. There are a few exceptions of greater values for microbes than for Br, but this is believed to be an artifact of the erratic data for microbial BTCs.
The dispersivity/transport distance ratio for microbe transport was less than that for Br, on average 5% (SD = 6%, n = 36, Table 6), mostly ranging from 1 to 3%. As for Br, the dispersivity of microbe transport appears to be inversely correlated with macroporosity (Fig. 3).
Leaching Vulnerability of Fecal Coliforms and Bacteriophages
The PV at maximum concentration (cmax) (Table 4), the normalized volume of water required for reaching maximum concentration in the leachate, reflects the leaching vulnerability of microbes through soils. The leaching vulnerability for microbes is in the order of: Netherton clayey soil and Hamilton clay loam (0.01) > Lismore shallow silt loam over gravels (0.03) > Waikiwi silt loam (0.06) > Templeton silt loam and Waikoikoi silt loam (0.08–0.12) > Manawatu fine sandy loam, Waitarere dune sand, and Atiamuri pumice soil (0.24–0.89) > Waihou allophanic soil (not leached). This sequence is consistent with a general trend of decreasing heterogeneity of soil structure. This suggests that heterogeneous soils, due to the presence of preferential flow paths, are more vulnerable to microbial leaching through soils. The small PV values at cmax (Table 4) suggest that, in structured heterogeneous soils, even a very small amount of water in the unsaturated soils could lead to a rapid and significant leaching of microbes through bypass flow. This result suggests that the leaching vulnerability of microbes into shallow water bodies could be the greatest in clayey soils, which often crack, followed by silt loam over gravels, silt loam, sandy loam, dune sand, pumice soil, and allophanic soil. Figure 3 shows that there is a positive relationship between macroporosity and leaching vulnerability of microbes.
Removal of Fecal Coliforms and Bacteriophages
Reduction in Concentration and Mass
The greatest reduction in phage concentration was observed in Waihou allophanic soil—no phages were detected in the leachate from this soil. As mentioned above, previous studies have suggested that allophanic clays have a net positive charge in the topsoil, which prompts bonding of negatively charged phages. In addition, allophane has a very large surface area, 700 to 900 m2 g–1 (Aislabie et al., 2001), further enhancing the bonding of phages with the soil medium. The second greatest reduction in phage concentrations occurred in the Atiamuri pumice soil [16–18 log m–1, where the unit log is log10(cmax/co), Table 4]. The presence of allophanic clay, even though just a small fraction in the pumice soil (Table 1), is believed to have played an important role in removal of phages in Waihou allophanic soil and Atiamuri pumice soil.
For most other soils, the reduction in fecal coliform and phage concentrations is about 2 to 3 log m–1 (Table 4), except for a greater removal of fecal coliforms in Manawatu fine sandy loam (9–10 log m–1). The clayey gley soil had the least reduction in microbial concentration (0.1–2 log m–1). This finding agrees with the inference of Wells (1973) that tephric soils have the best characteristics for effluent disposal, and soils with cracks, such as gley soils, are not suitable for effluent disposal.
Attachment and Removal Rates
The MIM-derived katt rates (Table 6) are the highest for the soils of volcanic origin (complete removal in Waihou allophanic soil, 11–12 d–1 for Atiamuri pumice soil and Hamilton clay loam). The high level of attachment associated with volcanic soils is attributed to their large surface areas, which are variably charged, as discussed above. Silt loams of greywacke origin (Templeton silt loam, Waikoikoi silt loam, and Waikiwi silt loam) have the second highest katt rates, ranging from 7 to 10 d–1. In these soils, soil particles are often coated with a thin layer of Fe and Mn oxides, which promote attachment of microbes. The lowest katt rates are found for the granular young soils (1–3 d–1 for Waitarere dune sand soil, 4–5 d–1 for Manawatu fine sandy loam). This is expected because granular soils are generally weaker adsorbents for microbes than clays and minerals (Sobsey et al., 1980; Moore et al., 1981). In addition, much less coating is present on the grain (largely silica) surfaces of these young soils. Heterogeneous soils have the most variable katt rates (1–16 d–1 for Netherton clayey gley loam, 5–26 d–1 for Lismore shallow silt loam), probably due to an uneven distribution of readily available attachment sites. The explanation for their possibly high katt rates is that clays and metal oxides have a strong affinity for microbes. In Lismore silt loam over gravels, >50% of the stones (derived from greywacke) show coatings. The relationship between katt and macroporosity is not clear (Fig. 3). These results suggest that attachment and thus removal of microbes is predominantly influenced by soil chemistry, particularly the lithologic origin.
The above patterns are also applicable for ktot rates determined from MIM (Table 6). The ktot rates derived from MIM are the highest in soils of volcanic origin (allophanic soil > 0.51 h–1, pumice soil 0.51 h–1, and clay loam containing tephra 0.48–0.52 h–1), second in silt loams of greywacke origin (0.30–0.43 h–1), weakest in granular young sandy soils (fine sandy loam 0.17–0.23 h–1, dune sand soil 0.07–0.15 h–1), and the most variable in heterogeneous soils (clayey gley loam 0.06–0.68 h–1, silt loam over gravels 0.23–1.09 h–1).
The kdet rate (mean 0.08 d–1, SD = 0.07, n = 36, Table 6) is on average only 1% of the katt rate (mean 8.28 d–1, SD = 4.95, n = 36). The very low kdet rates suggest that detachment of microbes in the structured soils investigated here was negligible and that microbial attachment could be considered to be irreversible. Other researchers have also commented that microbial attachment to porous media is primarily an irreversible process (Yao et al., 1971; Rajagopalan and Tien, 1976; Pieper et al., 1997).
Relative Contribution of Individual Processes
The inactivation rate derived from the batch incubation tests conducted at 15°C (a similar temperature to that used in the lysimeter experiments) was 0.298 d–1 for fecal coliforms and 0.214 d–1 for the phages. The contribution of inactivation to the total removal rate was, on average, 3.78% for phages (SD = 3.17%, n = 25) and 5.54% for fecal coliforms (SD = 5.34%, n = 11).
Although the contribution of inactivation to total removal could be examined, the individual contributions of straining and air–water interaction effects could not be quantified. The effects of these processes are combined in the total removal rates. We could only identify the possible presence of these processes.
Straining occurs when the ratio of the colloid to medium grain diameter, Dp/d, is >0.5% (Bradford et al., 2004), and will be significant when the ratio is >8% (McDowell-Boyer et al., 1986). Using these criteria, straining may be expected in the transport of fecal coliforms through all of the lysimeters (Dp/d > 0.69%, Table 2), and could be significant in clayey soil, clay loam, and silt loam (Dp/d 
10%). Based on these criteria, straining may have also occurred in the transport of phages in clayey soil and clay loam (Dp/d = 0.7–3%). It should be noted that the criteria for straining described above were developed from media with uniform grain sizes and that these criteria do not consider the effect of soil aggregates and macropores. Although the grain diameter may be small in aggregated soils, the effective diameter may be a lot larger because the grains are clumped and the microbes are not interacting with single grains. The large interconnected macropores developed in aggregated soils could allow the rapid transport of microbes with little straining. Therefore, the effect of soil aggregation and macropores on straining is unknown, especially for clayey soils, which naturally develop cracks within the soil. This could not be assessed in this study as no attachment profiles were measured during the experiments.
The air–water interface plays an important role in the enhanced removal of microbes under unsaturated conditions compared with saturated conditions (Chu et al., 2003; Torkzaban et al., 2006). Retention of microbes increases as water content decreases (Jin et al., 2000; Han et al., 2006; Torkzaban et al., 2006). Evaluating the effect of the air–water interaction is difficult in this study, however, as water contents were not measured.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
Breakthrough of both microbes and Br occurred at PV << 1, suggesting that rapid and significant leaching of microbial and solute contaminants could occur through bypass flow in structured soils with a small amount of water. An exception is for the allophanic soil, in which Br was adsorbed and the concentration peaked at >1 PV. Leaching vulnerability seems to relate to soil heterogeneity, being greatest in clayey and clay soils with cracks, followed by silt loam over gravels, silt loams, sandy soil, dune sand soil, pumice soil, and allophanic soil.
For both Br and the microbes, the general pattern of predicted mobile water content agrees with measured macroporosity, which is positively related to leaching vulnerability but negatively related to dispersivity. This suggests that soil structure plays the most important role in the transport of microbes and Br. The relationship between macroporosity and the mass exchange rate, however, was not clear in this study. The mass exchange rate between mobile and immobile water regions was greatest in volcanic soils and lowest in granular, young sandy soils.
Soil lithology has the greatest influence on the attenuation of microbes (attachment, removal, and reduction). The relationship between the attachment rate and macroporosity is not clear. The total removal rates (the sum of attachment rate and inactivation rate) derived from the MIM were highest in soils of volcanic origin (allophanic soil >0.51 h–1, pumice soil 0.51 h–1, and clay loam containing tephra 0.48–0.52 h–1), second highest in silt loams of greywacke origin (0.30–0.43 h–1), lowest in granular, young sandy soils (fine sandy loam 0.17–0.23 h–1, dune sand soil 0.07–0.15 h–1), and most variable in heterogeneous soils (clayey gley loam 0.06–0.68 h–1, silt loam over gravels 0.23–1.09 h–1).
The reduction in microbial concentration was the greatest in the allophanic soil (total removal) and second greatest in the pumice soil (16–18 log m–1). The least reduction in phage concentrations occurred in clayey gley soil (0.1–2 log m–1). For most other soils, fecal coliforms and phages had a concentration reduction of about 2 to 3 log m–1, except that the fine sandy loam had a greater concentration reduction for fecal coliforms (9–10 log m–1). These removal rates represent a reduction in microbial concentration on a natural log scale, and it needs to be divided by a factor of 2.3 to convert them to log10. The detachment rate of microbes was, on average, only 1% of the attachment rate, suggesting that detachment of microbes was negligible and attachment of microbes could thus be treated as irreversible.
The results of data analysis and modeling obtained in this study help our understanding of leaching vulnerability of microbes and the efficiency of effluent treatment for microbial removal in different soil media. We believe that this information could be applicable for soils under similar conditions, and that the results summarized from this study could provide useful information to regulatory agencies to more effectively manage land use and development and evaluate the risk of groundwater contamination.
These results, however, are applicable only for soils above the vadose zone (the zone between the soil root zone and the groundwater table). The ability of the vadose zone to remove microbes is expected to be much less effective. Thus, the reduction rates derived from soils cannot be extrapolated to vadose-zone media, i.e., for estimation of vertical separation distances. Further study on microbial removal in vadose zone media is needed.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionConclusionsREFERENCES |
|---|
imnek, and M.Th. van Genuchten. 2004. Straining and attachment of colloids in physically heterogeneous porous media. Vadose Zone J.
3
:384–394.imnek, J., M.Th. van Genuchten, and M. ejna. 2005. The HYDRUS-1D software package for simulating the one-dimensional movement of water, heat, and multiple solutes in variably-saturated media. Version 3.0. HYDRUS Softw. Ser. 1. Dep. of Environ. Sci., Univ. of California, Riverside.This article has been cited by other articles:
|
|
 |
|
  L. Pang Microbial Removal Rates in Subsurface Media Estimated From Published Studies of Field Experiments and Large Intact Soil Cores J. Environ. Qual., June 23, 2009; 38(4): 1531 - 1559. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
used in the van Genuchten–Mualem hydraulic model.
