JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 6 July 2006
Published in J Environ Qual 35:1204-1212 (2006)
DOI: 10.2134/jeq2005.0286
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Agricola
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Related Collections
Right arrow Ground Water Quality
Right arrow Laboratory Column Studies
Right arrow Colloids
Right arrow Microorganisms
Right arrow Solute Transport Models

TECHNICAL REPORTS

Ground Water Quality

Distance and Flow Effects on Microsphere Transport in a Large Gravel Column

Murray E. Close*, Liping Pang, Mark J. Flintoft and Lester W. Sinton

Institute of Environmental Science & Research Ltd., P.O. Box 29181, Christchurch, New Zealand

* Corresponding author (murray.close{at}esr.cri.nz)

Received for publication July 26, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Consumption of microbially contaminated ground water can cause adverse health effects and the processes involved in pathogen transport in aquifers need to be understood. The influences of distance, flow velocity, and colloid size on colloid transport were examined in homogenous pea-gravel media using an 8-m column and three sizes (1, 5, and 10 µm) of microspheres. Experiments were conducted at three flow rates by simultaneously injecting microspheres with a conservative tracer, bromide. Observed concentrations were simulated with CXTFIT and analyzed with filtration theory. The results demonstrate that colloid concentration is strongly log-linearly related to transport distance (as suggested by filtration theory) in coarse gravels, similar to our previous field studies. In contrast, the log-linear relationship is often reported to be invalid in fine porous media. The observed log-linear relationship is possibly because straining is negligible in the coarse gravels investigated. This has implications in predicting setback distances for land disposal of effluent, and suggests that setback distances in gravel aquifers can be estimated using constant spatial removal rates (f). There was an inverse relationship between transport distance and colloidal concentration, but not with temporal attachment rate (katt) and collision coefficient ({alpha}). Increases in flow velocity result in increasing colloidal recovery, katt and {alpha} but decreasing f. Increases in sphere size result in decreasing colloidal recovery with increasing katt, f, {alpha}, and velocity enhancement. Diffusion is the dominant collision mechanism for 1-µm spheres (81–88%), while settling dominates for 5- and 10-µm spheres (>87%), and interception is very small for all spheres investigated.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
MANY WATERBORNE DISEASE OUTBREAKS are caused by the consumption of ground water contaminated by pathogenic bacteria and protozoa. This has resulted in an increasing need in our understanding of the processes and the influencing factors that are involved in pathogen transport in aquifers. Such knowledge is a fundamental requirement to accurately model and predict pathogen transport in ground water. A good understanding of the relationship of microbial concentrations to transport distance has particular implications in accurately predicting setback distances for land disposal of effluent and solid wastes.

Filtration is one of the most important processes that determine the fate and transport of microbes in aquifer systems. Our understanding of microbial filtration in porous media has largely developed from classical filtration theory (Yao et al., 1971). It assumes that microbial concentrations are log-linearly related to distance and collision efficiency (or removal rate coefficient) is a constant, independent of transport distance and flow velocity. However, recent studies have shown deviations from this log-linear relationship, with a common finding that attachment rate coefficients (katt) or collision efficiency ({alpha}) decrease with distance (Albinger et al., 1994; Redman et al., 2001; Li et al., 2004). In the study of Albinger et al. (1994), the collision efficiency of bacteria decreased 10-fold over a bed depth of only 1 cm. It has been found that katt or {alpha} is also related to pore-water velocity. Harter et al. (2000) found that at a high flow velocity, {alpha} is significantly greater and velocity enhancement (ratio of colloid velocity to bromide velocity) of bacteria is less pronounced. Hendry et al. (1999) found that katt is velocity independent for flow velocity at 0.7 to 3.1 m d–1 but reduced considerably for a flow velocity of 0.1 m d–1, and the peak concentrations of bacteria decreased with decreasing velocity. Camesano and Logan (1998) found that {alpha} was constant and independent of low velocity for immotile bacteria, but for motile bacteria, {alpha} was two orders of magnitude higher at high velocities than at low flow velocities.

However, the previously observed depth-dependant filtration phenomenon are all derived from studies of fine media in small columns (mostly a few centimetres), in which straining is significant and flow velocities are low (mostly <5 m d–1). For fine materials, depth-dependant katt or {alpha} values may reflect the effect of straining, which decreases with depth (Bradford et al., 2003, 2004). In an evaluation of field data from 11 microbial tracer experiments carried over distances from tens to hundreds of meters, Pang et al. (2005) have demonstrated that the log-linear dependency of microbial concentrations on transport distance is approximately valid in gravel aquifers for observation wells (excluding injection well data). One of the reasons for the valid log-linear relationship between microbial concentrations and transport distances observed for coarse gravel aquifers could be the negligible straining and high flow velocities (a few meters to a few hundred meters per day). Other possibilities could include distributed surface properties on either the colloids or the aquifer media.

In this study we look at a system that should have minimal straining and examine the relationship between microbial concentrations and distance. The objectives are to determine whether the log-linear relationship between microbial concentrations and transport distance is truly valid for gravel media at negligible straining and high flow velocities at different flow velocities and particle sizes and to examine the relationships between the filtration parameters for different flow rates and particle sizes.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The column (made of a PVC pipe) was 8 m long with a 30-cm internal diameter and had sampling points at 1-m intervals. The column was inclined upward at a 30 degree angle. Pea-gravel of irregular but well-rounded shape, 5 to 7 mm in diameter, was filled from the top of the inclined column and compacted to give a bulk density of 1.74 g cm–3 and porosity of 0.35. The partially de-aerated tapwater (pH = 7.9; electrical conductivity = 15 mS m–1), sourced from the Canterbury alluvial gravel ground water, was then introduced from the bottom of the column from a constant header-tank. The upward water movement helped to remove any entrapped air and to minimize preferential flow. There was no change in pH or electrical conductivity during passage through the column. Negatively charged polystyrene latex spheres (Interfacial Dynamics Corporation, Tualatin, OR) were used. Their physical properties are listed in Table 1.


View this table:
[in this window]
[in a new window]
  		Table1. Physical properties of microspheres used in the experiments.

 
The study was designed to examine, under controlled laboratory conditions, whether the log-linear relationship is truly valid for gravel media at negligible straining and high flow velocities. To minimize straining, we used coarse pea gravel in our experiments. The flow velocities applied are in the range of those typically observed for gravel aquifers. The uniform sizes of pea gravels allow application of filtration theory, which assumes a single particle grain size. Gravel aquifer material with a natural particle size distribution is not used in our experiments as Pang et al. (2005) have shown that the use of collision efficiency, estimated from a single particle size, is misleading for poorly sorted gravel media that have a wide particle size distribution.

Three different sized microspheres were used to study the dependence of filtration processes on transport distance and flow velocity. Microspheres have been widely used in study of filtration and straining processes (McCaulou et al., 1995; Bradford et al., 2003; Tufenkji et al., 2004) and ground water tracer studies (Harvey et al., 1989; Bales et al., 1997; Harvey and Harms, 2002; Auckenthaler et al., 2002). They are often used as surrogates of bacteria and protozoa (Dai and Hozalski, 2003) because (i) they are harmless to humans and the environment so that the hazards associated with the use of pathogenic organisms can be avoided, (ii) the sizes and surface characteristics of microspheres are comparable to those of bacteria and protozoa, and (iii) they are not subject to complicating effects of die-off and growth, so that filtration processes can be explicitly examined.

Experiments were conducted at three different pore-water velocities, 13, 29, and 55 m d–1, hereafter referred to as low, intermediate, and high flow velocities, respectively. In each experiment, a solution containing a conservative tracer, bromide (Br), together with microspheres, was injected into the column and monitored at 1-m intervals down the column. The mass injected at each flow velocity was kept approximately constant, and comprised about 0.5 g Br, 1.0 x 1011 1-µm spheres, about 3 x 109 5-µm spheres, and 4.2 x 108 10-µm spheres. The time of injection varied with the flow velocity so that the volumes injected were constant. Two series of experiments were performed, one in 2002 (denoted as Experiment 02) and one in 2004 (denoted as Experiment 04). In Experiment 04, 1-, 5-, and 10-µm spheres were used, while in the Experiment 02, only 5- and 10-µm spheres were used. All microspheres were counted using a Abakus particle counter (Klotz, Bad Liebenzell, Germany). The fluorescent microspheres were also analyzed using a RF-1501 spectrofluorimeter (Shimadzu, Kyoto, Japan). The Br samples were analyzed using an ion selective electrode.


   DATA ANALYSIS AND MODELING
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Transport of microspheres in saturated pea gravels was described by an advection–dispersion model that was coupled with first-order attachment–detachment kinetics (Hornberger et al., 1992; McCaulou et al., 1994):

Formula 1[1]
where c is the microsphere concentration in the solution (g m–3), S is the microsphere concentration attached to the gravels (g g–1), t is the time since injection (h), D is the dispersion coefficient of the microsphere (m2 h–1), x is the distance along the column from the column inlet (m), v is the mean pore-water velocity of the microsphere (m h–1), {rho}b is the bulk density of the gravels (g m–3), {theta} is the porosity of the gravels (m3 m–3), and katt and kdet are the first-order rate coefficients for microsphere attaching/detaching onto/from the gravels (h–1), respectively.

As our results show that the microspheres were not retarded compared to Br and showed little tailing in their concentration breakthrough curves (BTC), attachment was considered to be irreversible (i.e., kdet = 0). Thus, we used a reduced form of Eq. [1] that has the same form as the common advection–dispersion equation (ADE) with a first-order reduction term. Consequently, CXTFIT (Toride et al., 1995) was used to simulate the microsphere breakthrough curves at each distance.

According to classic filtration theory, reduction of colloid concentration is log-linearly related to transport distance, which is expressed by a spatial removal rate coefficient (Yao et al., 1971):

Formula 2[2]
where co is the input concentration (g m–3). In this study, we experimentally derived the f factor from the slope of the log(cmax/co) vs. distance plot, where cmax are the peak concentrations (g m–3) obtained from multiple sampling locations. In addition, the f factor was also directly calculated from individual cmax/co and x values using Eq. [2].

In colloid filtration theory, temporal attachment rate coefficients are related to the physical properties of aquifer media and colloids (Yao et al., 1971; Bales et al., 1991; McCaulou et al., 1994):

Formula 3[3]
where d is the average diameter of aquifer media grain (m), {alpha} is the collision efficiency (–), which is defined as the ratio of spheres that attach to soil grains to spheres that collide with the soil grains, and {eta} is the collector efficiency (–) representing the fraction of spheres that collide with the collector. The collector efficiency is a function of the physical properties of both the porous media and the spheres. The relationship between temporal and spatial removal rate coefficients is given by (Logan et al., 1995):

Formula 4[4]

The collector efficiency {eta} is the sum of collector efficiencies due to diffusion, interception, and settling (sedimentation), as expressed below (Harvey and Garabedian, 1991):

Formula 5[5]
and:

Formula 6[6]
where {eta}D is the collection due to Brownian diffusion (–), {eta}I is the collection due to interception (–), {eta}G is the collection due to settling (–), kB of 1.38 x 10–23 (kg m2 s–2 K–1) is the Boltzmann constant, T is the absolute temperature (K), µ is the water viscosity (kg m–1 s–1), dp is the colloidal particle diameter (m), d is the average diameter of the porous media grain (m), v is the pore-water velocity (m s–1), {rho}p is the colloidal density (kg m–3), {rho} is the water density (kg m–3), g is the gravitational constant (m s–2), and As is a porosity-dependent parameter (–).

For a spherical collector, As has been given as:

Formula 7[7]


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
A large amount of experimental data and modeling data has been generated in this study, and only selected examples of breakthrough curves are presented in this paper. Figure 1 gives an example of observed and model-simulated concentrations for the 1- and 10-µm microspheres and Br (at 2, 4, 6, and 8 m) for the intermediate flow velocity. Separate attachment coefficients were fitted to each breakthrough curve. The observed breakthrough curves of Br and microspheres are reasonably symmetrical with little tailing. This indicates that the pea gravels used are homogenous and attachment of microspheres is predominantly irreversible. Thus, the one-site kinetic model, CXTFIT, generally well described the observed sphere concentrations. Results derived from samples analyzed from the spectrofluorimeter and the particle counter for 1-µm microspheres are very similar (an example is given in Fig. 2). For the 10-µm spheres in Experiment 04, spectrofluorimetric analysis was much more sensitive than the particle counter so those data were used in the data analysis.


Figure 1
View larger version (17K):
[in this window]
[in a new window]
  		Fig. 1. Observed (solid symbols) and model-simulated (dotted line) concentrations at 2, 4, 6, and 8 m under intermediate flow velocity for Br and 1- and 10-µm spheres. The term c is the concentration in solution and co is the injected concentration.

 

Figure 2
View larger version (13K):
[in this window]
[in a new window]
  		Fig. 2. Concentrations of 1-µm spheres at 3- and 8-m distances determined from spectrofluorimeter and particle counter for the intermediate flow velocity. The term c is the concentration in solution and co is the injected concentration.

 
Velocity Enhancement
There was no clear relationship between transport distance or flow rate on velocity enhancement (Table 2). There is greater velocity enhancement at lower flow rates for the 5- and 10-µm spheres in Experiment 02. However, in Experiment 04, although the results for the 5-µm spheres are consistent with the earlier results, there are no trends and little, if any, velocity enhancement for the 1- and 10-µm spheres (Tables 2 and 3). In Experiment 04, the 5-µm spheres generally show the highest ratios of vsphere/vBr at all flow velocities. This is possibly because the 5-µm spheres have the lowest surface charge density (Table 1), and thus the observed differences would be related to charge rather than to a size effect. Different 10-µm spheres were used for Experiments 02 and 04 (Table 1) and this is possibly the reason for the different behavior observed for the 10-µm spheres between the two experiments. Another possible reason is the change in sensitivity in the analytical method from particle counting to spectrofluorimetric analysis. For Experiment 02, the 10-µm spheres show a generally greater vsphere/vBr ratio than for the 5-µm spheres (Table 3), which is consistent with pore-size exclusion concept. The observed behavior is probably a combination of the effects of charge and size of the microspheres.


View this table:
[in this window]
[in a new window]
  		Table 2. Impact of low (L), intermediate (I), and high (H) flow rate on velocity enhancement and filtration parameter values.{dagger}

 

View this table:
[in this window]
[in a new window]
  		Table 3. Impact of colloidal size on velocity enhancement and filtration parameter values.{dagger}

 
Relative Concentrations (cmax/co)
Figure 3 demonstrates that cmax/co of the spheres decreases log-linearly with increasing transport distance [note that –log(cmax/co) is plotted in Fig. 3], as suggested by the filtration theory (Eq. [2]). This is consistent with the results of Pang et al. (2005) for the field data obtained from gravel aquifers. This finding is consistent with the idea that, when straining is negligible, the log-linear relationship between colloidal concentration and distance may still be valid. Straining will be significant when the ratio of the colloid to media grain diameter is greater than 8% (McDowell-Boyer et al., 1986) and should only occur when the ratio is greater than 0.5% (Bradford et al., 2004). The size of the largest microsphere (10 µm) is 0.2% of the smallest pea gravels (5 mm), which is below the 0.5% threshold. Lahav and Tropp (1980) also report a log-linear relationship between microsphere concentrations and transport distance in saturated soil columns and support the log-linear relationship. Straining is considered to be also negligible in their study comparing the size of soil media (sandy soils) and microspheres (0.12 and 0.21 µm) that they used.


Figure 3
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between transport distance and concentration of microspheres for (a) Experiment 04 and (b) Experiment 02. The term F or P denotes whether analysis was by fluorimetry or particle counting, respectively; all analyses for Experiment 02 used particle counting. The term cmax is the maximum observed concentration in solution and co is the injected concentration.

 
The magnitude of cmax/co reflects relative recovery of spheres. Summarizing the median values of cmax/co (0.14 for 1 µm, 0.05 for 5 µm, and 0.03 for 10 µm; 0.05 for low flow velocity, 0.06 for intermediate flow velocity, and 0.09 for high flow velocity, n = 30), there is a general pattern of decrease in recovery (i.e., greater retention) with increasing size of spheres and decreasing flow rate. This trend is expected as large spheres are more readily removed than small ones, and at low flow velocities, there is a longer residence time for the spheres to attach onto the media surfaces. Lahav and Tropp (1980) also found retention of microspheres decreased with increasing flow rate.

Spatial Removal Rate (f)
Table 4 compares f values estimated from the three independent methods [medians are given for the values derived from individual log(cmax/co)/distance and model-derived katt data]. It shows that f values estimated from the slopes of the log(cmax/co) vs. distance plots are consistently smaller than those estimated from f values calculated from individual log(cmax/co)/distance data (both calculated using Eq. [2]). The f values converted from the model-derived katt values (Eq. [4]) are most variable as the results are affected by inter-parameter relationships involved during the inverse modeling. There will be some error in all measurements of both co and cmax. The values of f derived from the slope of the log(cmax/co) vs. x plots are not affected by any variations in co, as variations in co affect the intercept but not the slope. However it will affect the calculations of f from individual log(cmax/co) values. These factors will tend to result in the f values derived from the slope being less than the f values derived from the individual measurements of log(cmax/co), as seen in Table 4. We consider that the f values estimated from the slopes of the log(cmax/co) vs. distance plots are the most reliable.


View this table:
[in this window]
[in a new window]
  		Table 4. Spatial removal rate (f) estimated from three independent methods.{dagger}

 
The f values determined from the slopes of the log(cmax/co) vs. distance plots (Fig. 3) are relatively constant for 1-µm spheres (varying 0.27–0.35 m–1) and seem to be independent of flow velocity. Both the analytical data from the particle counting and fluorimetric analyses for the 1-µm spheres are shown in Fig. 3 to show the consistency of the derived f factors. The f values for 5- and 10-µm spheres are similar and are somewhat greater than those for 1-µm spheres. This is expected because there is a factor of 5 between 1- and 5-µm spheres but only a factor of 2 between 5- and 10-µm spheres. A greater removal is expected for larger particles than small particles. There seems to be an inverse relationship between f and flow velocity for 5-µm spheres and Experiment 02 10-µm spheres. A possible explanation is that less residence time is available for collision of spheres to the gravels at the high flow velocity, and hence less removal. Kretzschmar et al. (1994) also observed an inverse relationship between f and flow velocity. Compared to Experiment 04, the f values determined from the Experiment 02 are very similar for the 5-µm spheres but higher for the 10-µm spheres. This is possibly because the 10-µm spheres used for the two experiments had slightly different properties (Table 1).

The f values determined in this study (predominantly 0.1 m–1) for uniform pea gravels are one order of magnitude greater than those for poorly sorted uncontaminated alluvial gravels (0.01 m–1) reported by Pang et al. (2005). This is because the mean grain size of pea gravels (D50 = 6 mm) is much smaller than that of the actual aquifer gravels (D50 = 18.3 mm). According to Eq. [4] and [5], the smaller the porous media, the greater the filtration coefficient would be. Some very high f values are seen in column experiments with sand materials, for example, 12 m–1 in Toran and Palumbo (1992) and 5.2 m–1 in Higgo et al. (1993).

Attachment Rate Coefficients (katt) and Collision Efficiency ({alpha})
The plots of katt versus transport distance are inconsistent and show a lot of variability (Fig. 4). Similar patterns are also observed for {alpha} and f converted from katt (graphs not shown). Of the 15 experiments, about a third of the plots show a negative relationship, three plots show a positive relationship, and the remainder show no relationship, with most plots showing significant variability. The positive slopes tend to be at the lower flow velocities while the negative slopes are observed at the higher flow velocities (Fig. 4).


Figure 4
View larger version (23K):
[in this window]
[in a new window]
  		Fig. 4. Impact of transport distance on attachment rate coefficient (katt) for (a) Experiment 04 and (b) Experiment 02. The term F or P denotes whether analysis was by fluorimetry or particle counting, respectively; all analyses for Experiment 02 used particle counting.

 
Our data show that katt and {alpha} are clearly related to flow velocity (Table 2). At the same transport distance, the greatest katt and {alpha} values are predominately observed at the high flow velocity for all spheres. There is a tendency of increasing katt and {alpha} with flow rate, especially for the 10-µm spheres. This result is in agreement with finding of Harter et al. (2000) who observed that at high flow rates, {alpha} is significantly greater than at low flow rates. This relationship can be explained by Eq. [4].

For Experiment 02 at the same transport distance, katt and {alpha} values are consistently greater for the 10-µm spheres than for 5-µm spheres (Table 3). This is expected as the larger the particles are, the more filtration processes would occur. For Experiment 04, the greatest {alpha} values are often related to the 5-µm spheres, particularly at low and intermediate flow velocities. Like the pattern for f values, the lowest {alpha} and katt values are mostly associated with the 1-µm spheres at the same distance at all flows, and most {alpha} and katt values are not significantly different between 5- and 10-µm spheres.

Contribution of Settling, Interception, and Diffusion to Filtration
The {eta} values estimated for the 1-µm spheres are in the magnitude of 0.001 at the low flow velocity and 0.0001 at intermediate and high flow velocities. For the 5-µm spheres, it is consistently in the magnitude of 0.001 at all flow velocities. For the 10-µm spheres, it is in the magnitude of 0.01 at the low flow velocity and 0.001 at both intermediate- and high-flow velocities. Results obtained from Experiment 02 and Experiment 04 are very similar.

Deriving from ratios of {eta}G/{eta}, {eta}I/{eta}, and {eta}D/{eta}, Table 5 summarizes the contribution of settling, interception, and diffusion to filtration. The results suggest that diffusion is the dominant collision mechanism for 1-µm spheres (81–88%), while it is small for 5-µm spheres (5–8%) and very small for 10-µm spheres (1–2%). Settling (sedimentation) dominates the filtration process for 5-µm (87–94%) and 10-µm spheres (93–98%). For all spheres, filtration by interception is very small (0–1% for 1 µm and 1–5% for 5 and 10 µm). The contribution of diffusion and interception to filtration increases with increasing flow velocity for all spheres. In contrast, the contribution of settling to filtration decreases with increasing flow velocity for all spheres. These contributions are independent of transport distances but would be specific to the system being considered in this study.


View this table:
[in this window]
[in a new window]
  		Table 5. Contributions of different filtration mechanisms.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Consistent with our field experimental results, our column experimental results have also demonstrated that colloid concentration is strongly log-linearly related to transport distance in coarse gravels. Therefore, setback distances in gravel aquifers could be estimated based on a constant spatial removal rate (f). The most reliable approach in estimation of spatial removal rate is to estimate it from the slope of a log(cmax/co) ~ x plot. Although it can be also converted from model-derived temporal removal rates (attachment rates), the results often contain some uncertainties generated from inverse modeling. The f values determined for the 1-µm spheres are relatively constant and not sensitive to changes in flow velocity. The f values for 5- and 10-µm spheres are similar and are somewhat greater than those for 1-µm spheres, and seem to have an inverse relationship with flow velocity. These experimental results, showing a log-linear relationship between colloid concentration and transport distance, suggest that filtration theory can be valid in systems where straining is negligible, although other possible explanations have not been explored.

Our experimental results show that, for the system studied here, the attachment rates and collision coefficients are not constant and tend to increase with flow velocity and colloid size. Their relationship with transport distance is inconsistent. There is an overall pattern of a greater retention with increasing colloid size and decreasing flow velocity. Our results suggest that diffusion is the dominant collision mechanism for 1-µm spheres (81–88%), while settling dominates the filtration process for 5-µm (87–94%) and 10-µm spheres (93–98%). Filtration by interception is negligible for all spheres investigated. With increasing flow velocity, the relative contribution from diffusion becomes more important while from settling becomes less important. Velocity enhancement compared to the conservative Br tracer was observed in some experiments, even for this uniform pea-sized gravel media.


   ACKNOWLEDGMENTS
 
This study was funded by the New Zealand Foundation for Research, Science, and Technology (Contract CO3X0303).


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DATA ANALYSIS AND MODELING
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Agricola
Right arrow Articles by Close, M. E.
Right arrow Articles by Sinton, L. W.
Related Collections
Right arrow Ground Water Quality
Right arrow Laboratory Column Studies
Right arrow Colloids
Right arrow Microorganisms
Right arrow Solute Transport Models

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal