| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a School of Engineering
b Department of Environmental Biology, University of Guelph, Guelph, Ontario, Canada N2G 2W1
c Department of Geography, University of Guelph, Guelph, Ontario, Canada N2G 2W1
d Department of Engineering, Nova Scotia Agricultural College, Truro, Nova Scotia, Canada B2N 5E3
* Corresponding author (jamiesor{at}uoguelph.ca)
Received for publication May 25, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: E. coli NAR, Escherichia coli resistant to nalidixic acid
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The resuspension of bacteria-laden bed sediments, and the subsequent degradation of in-stream water quality, has been illustrated in a number of field studies (McDonald et al., 1982; Sherer et al., 1988; Nagels et al., 2002). These studies have been able to isolate the impacts of bed sediment bacteria reservoirs by creating artificial flooding events or by mechanically agitating (raking) bed sediments. Previous studies, however, have not attempted to quantitatively link the timing and magnitude of entrained bed sediment bacteria reservoirs with hydraulic conditions and bed sediment characteristics.
As well, it is important to note that bacteria are typically associated with fine, or cohesive, sediment particles in aquatic environments (Gannon et al., 1983; Auer and Niehaus, 1993). Cohesive sediment is typically defined as sediment particles less than 60 µm in diameter (Partheniades, 1977). When salt is added to a suspension of particles in this size range the previously dispersed particles will aggregate to form larger flocs. The ions in solution will suppress interparticle electrochemical repulsive forces and allow attractive London–van der Waals forces to dominate (Mehta et al., 1989). In general, this size distinction would primarily include silt- and clay-sized materials. Due the differences in the respective processes that govern their movement, cohesive and noncohesive transport equations have evolved independently of each other (Mehta and Lee, 1993). For noncohesive particles the submerged weight of the particle is the primary force resisting movement in a flowing water system. There are three primary forces resisting the movement of cohesive sediment particles: (i) gravity, (ii) mechanical friction between particles, and (iii) physiochemical interparticle attraction (Mehta et al., 1989). Characterizing the influence of interparticle attractive mechanisms is difficult as the process is influenced by a number of factors including particle size, mineralogical composition, organic matter, temperature, pH, ionic strength, and flow conditions (Mehta et al., 1989; Droppo and Ongley, 1994).
Most in-stream water quality models treat microorganisms as free-floating colloids with a neutral buoyancy, despite the general consensus that bacteria are associated with sediments in stream environments. Wilkinson et al. (1995) and Tian et al. (2002) have both tried to incorporate sediment associations within in-stream models. In both models, bacteria resuspension and deposition rates are simulated as simple empirical functions of discharge.
The pattern and magnitude of bacteria resuspension in alluvial streams should be related to the transport characteristics of the sediments to which they are attached. In this study, a tracer bacteria was used to investigate the resuspension and persistence of sediment-associated bacteria in a small alluvial stream. The study was conducted in Swan Creek, located within the Grand River watershed of Ontario, Canada. The primary objectives of this experiment were to examine: (i) hydraulic conditions necessary to resuspend bed sediment reservoirs of bacteria, and (ii) bacteria resuspension rates during entrainment events. The experiment was also designed to provide information on the persistence of enteric bacteria in bed sediments and to study the redistribution of enteric bacteria within alluvial stream reaches.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The experiment was conducted during May and June of 2003 in Swan Creek, a tributary of the Grand River located approximately 20 km northwest of the city of Guelph (Fig. 1) . The Grand River watershed is the largest in southern Ontario. A 1.1-m2 section of the bed was seeded with E. coli NAR on 13 May 2003 following a procedure described in a previous study (Jamieson et al., 2004). The seeding procedure was designed to produce bed sediment E. coli levels similar to those observed in intensively farmed watersheds (Jamieson et al., 2003). The stream bed inoculum suspension was grown by transferring a loopful of cells from an E. coli NAR plate culture to four 1000-mL Erlenmeyer flasks, each containing 500 mL of tryptic soy broth (TSB) supplemented with nalidixic acid (100 mg L–1). The flasks were incubated at 37°C with shaking at 200 rpm for 14 to 16 h. Cells were harvested by centrifugation at 5000 x g for 20 min, washed twice with sterile physiological saline (0.85% NaCl), and resuspended in saline to a density of 1010 mL–1. This procedure yielded approximately 2 x 1012 bacterial cells. The density of cells within the concentrated suspension was estimated spectrophotometrically by measuring the absorbance at 550 nm.
|
To seed the bed, a circular steel mixing chamber (1.2 m in diameter), open at both ends, was first driven into the stream bed. The E. coli NAR inoculum suspension was added to the water inside the mixing chamber and the sediment was raked to suspend approximately the top 2 cm of the sediment into the water column. The mixing chamber was covered and the sediment allowed to settle. It was anticipated that the tracer bacteria would attach to the suspended particles and then settle to the bottom of the mixing chamber. After 24 h, the mixing chamber was carefully lifted out of the stream, with minimal disturbance. Before the mixing chamber was removed, both the water and bed sediments from within the chamber were sampled and analyzed for E. coli NAR. Escherichia coli NAR was enumerated in serially diluted water samples using the membrane filtration technique (American Public Health Association, 2000). Enumeration of sediment samples was accomplished by placing 1 g of fresh sediment in 99 mL of sterile phosphate-buffered dilution water. Serial dilutions of the sediment and water solution were then prepared and analyzed using the membrane filtration technique. The medium used for enumerating E. coli NAR in all water and sediment samples was mTec agar supplemented with nalidixic acid (100 mg L–1).
The mixing chamber was removed from the stream on 14 May 2003. Water sampling stations were established at distances of 10, 100, 500, and 1700 m downstream of the source cell. Water samples were collected at each station at varying time intervals following removal of the mixing chamber. During each sampling event composite samples were collected at each station. The stream cross section at the downstream sampling locations was divided into four sections. A depth-integrated sample was taken within the water column within each section using a DH-48 sediment sampler. The four depth-integrated samples were then composited together and split between two 1000-mL sterile plastic bottles. Analysis of the water samples included E. coli NAR and total suspended solids (TSS) concentrations. Three grab samples were also collected during the course of the experiment and analyzed by an accredited lab (University of Guelph Laboratory Services Division, Guelph, ON) for dissolved phosphorus (P), ammonia nitrogen (NH3–N), nitrate nitrogen (NO3–N), electrical conductivity (EC), and pH. All water and sediment samples were placed in ice-filled coolers for transport. Bacterial analyses were performed within 6 h of sample collection. As well, a Model 3700 autosampler (ISCO, Lincoln, NE) was installed at the 100-m sampling location before an anticipated high rainfall event on 23 May and programmed to collect discrete 500-mL samples at 1-h intervals for a 24-h period.
At time intervals ranging from 1 to 7 d, tracer bacteria levels in the bed sediments within the source cell were determined. Bed sediment samples were collected from within the source cell using a sterilized metal spoon. The depth of sediment sampled was approximately 2 cm. Three small subsamples were collected and composited during each sampling event. The location of each sample was noted to ensure it was not sampled more than once. Previous work had shown that this method of composite bed sediment sampling was adequate for characterizing source cell tracer bacteria concentrations (Jamieson et al., 2004). As the bed sampling procedure may have resulted in small releases of tracer bacteria to the water column, bed sediment sampling was always performed after the downstream water samples had been collected. Enumeration of E. coli NAR in the bed sediment samples followed the procedure outlined previously. As well, bed sediment samples were collected at 14 locations throughout the 1700-m study area and analyzed for the presence of E. coli NAR on two occasions: 1 and 3 wk after the introduction of the tracer bacteria. Bed sediment samples were collected adjacent to the source cell at the time of bed seeding and analyzed for grain size distribution, total P, NH3–N, NO3–N, EC, pH, cation exchange capacity (CEC), and organic matter content (loss on ignition, LOI).
Stream discharge was determined continuously at distances of 100 and 1700 m downstream of the source cell. At each location, WL15 water-level loggers (Global Water Instrumentation, Gold River, CA) were placed 5 cm above the stream bed. Water depths were continuously recorded on an hourly basis. At each location, discharge was measured on several occasions using the velocity–area method (Canadian General Standards Board, 1991). Water depth measurements were then converted to flow estimates by constructing a stage–discharge relation for each site.
A topographic survey was performed before the experiment. Channel characteristics that were measured included cross-sectional geometry, longitudinal profile along the thalweg, and water surface slope. Cross-section surveys were performed approximately every 100 m or as needed to capture all major channel characteristics such as riffles, pools, and meander bends. The water surface slope was determined at three locations within the study reach. The size distribution of bed material was assessed at multiple locations within the study reach. The analysis focused on characterizing fine-grained bed material within the study reach, as bacteria are primarily associated with these sediments. Bulk samples were collected at 14 locations throughout the 1700-m study reach that had been visually identified as sediment deposition areas. The particle size distribution of each sample was determined by a mechanical sieve analysis.
Water temperature was measured at two locations within the study reach using HOBO Water Temp Pro temperature sensors (Onset Computer Corporation, Pocasset, MA). One of the locations was well shaded while the other received little shade. The sensors were secured at a mid-depth position within the channel and programmed to collect temperature measurements on an hourly basis. Daily rainfall totals were measured using a digital recording rain gauge that was installed in an open area near the source cell.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The median grain size within deposition zones varied from 0.07 to 0.30 mm, with the finest sediments found in the large deposition zone located at 800 m. The physicochemical characteristics of fine sediments collected adjacent to the source cell are presented in Table 1. The sediments at this location possessed a median grain size of 0.11 mm and an organic matter content of 8%. Average daily flows in Swan Creek and daily rainfall depths recorded adjacent to the reach during the experiment are presented in Fig. 3. The average stream flow during the experiment was 0.25 m3 s –1. Maximum and minimum average daily flows recorded during the experiment were 0.7 and 0.07 m3 s–1, respectively. The peak instantaneous flow recorded was 0.8 m3 s–1. Several rainfall events resulted in the generation of storm hydrographs, the largest occurring approximately 9 d after the source cell was seeded with the tracer bacteria. The bankfull discharge for Swan Creek, computed using the cross-sectional geometry of the stream near the source cell, is 2.4 m3 s–1. Swan Creek stream water samples were slightly alkaline and possessed a mean NO3–N concentration of 0.95 mg L–1 (Table 1).
|
|
|
Bed sediment samples were collected from 14 local deposition areas downstream of the source cell after the first storm event at 100 h and again after the large storm event at 225 h. The tracer bacteria were not recovered in any bed sediment samples collected downstream of the source cell, which would indicate that the majority of E. coli NAR leaving the source cell were traveling through the study reach without depositing.
Persistence of the Tracer Bacteria in the Water Column
The presence and concentration of E. coli NAR in water samples collected at the four sampling stations are presented in Fig. 3 . Only samples possessing detectable levels of E. coli NAR are shown in the figure. A summary of the maximum concentrations observed in the water column during resuspension events is provided in Table 2. The passage of the initial pulse of nonsettled E. coli NAR, after the removal of the mixing chamber, was observed at all four sites. As expected, the peak concentration of the initial pulse decreases in the downstream direction (Table 2) due to dispersion, and possibly settling losses. Concentrations of the tracer bacteria declined to nondetectable levels at 100, 500, and 1700 m approximately 24 h after the mixing chamber was removed. The tracer bacteria were recovered in water samples at 10 m for the first 75 h of the experiment. The tracer bacteria were also recovered at 10 and 100 m during a small storm event at 100 h.
|
Figure 3 clearly illustrates that the presence of the tracer bacterium within the water column is directly linked to stormflow events. Escherichia coli NAR was recovered at all sampling stations during the rising limb of storm hydrographs at 225, 550, and 600 h after the source cell was seeded. The appearance of the tracer bacterium coincided with increases in total suspended solids, indicating that the E. coli NAR that were being resuspended were sediment-associated. As the concentrations of E. coli NAR do not increase in the downstream direction it appears that the tracer bacteria are originating primarily from the source cell. The tracer bacterium were not recovered in any water sample after the 600-h storm.
Critical Flows and Resuspension Rates
The data were further analyzed to determine the hydraulic conditions necessary to cause resuspension of sediment-associated tracer bacteria. A relationship between flow and bed shear stress was developed using the cross-sectional geometry of the stream at the source cell. Manning's equation was used to express flow (Q) as a function of hydraulic geometry:
 | [1] |
 | [2] |
b is the bed shear stress (N m–2) and y is the specific weight of water (N m–3).
Equations [1] and [2] can then be combined to express b as a function of Q:
 | [3] |
This relationship was used to construct a continuous record of b at the source cell for the entire study period. An ISCO 3700 autosampler was deployed at the 100-m sampling location before the start of the large storm event that occurred at 225 h. Concentrations of E. coli NAR in the water column at 100 m during this stormflow event are presented in Fig. 4
. Also provided in Fig. 4 are total suspended solids concentrations, Q, and b at the source cell during the storm.
|
b necessary to first mobilize the tracer bacteria from the stream bed. The critical values were assumed to correspond with the first appearance of the tracer bacteria on the rising limb on the storm hydrograph (Fig. 4). The critical bed shear stress, c, and critical flow, Qc, observed during this event were 1.7 N m–2 and 0.37 m3 s–1, respectively. Values of c and Qc were also estimated for storms that occurred at 550 and 600 h (Table 3). The hydraulic conditions required to entrain sediment-associated E. coli NAR were similar for the three storms.
|
 | [4] |
b.
The results from this experiment have illustrated some important characteristics of bacteria resuspension in alluvial streams. It appears that a finite supply of bacteria are available for resuspension within each stormflow event and that this supply is depleted on the rising limb of the storm hydrograph. This is illustrated clearly in Fig. 4, as the value of b corresponding to cessation of resuspension (2.3 N m–2) is higher than the value of c determined from the rising limb of the hydrograph (1.7 N m–2). This phenomena has been noted by other researchers who have studied bacteria resuspension in channels downstream of dam releases (Wilkinson et al., 1995; Nagels et al., 2002). It should be noted, however, that a constant water surface slope, measured during steady flow conditions, was used to compute values of b. During storm hydrographs, the slope of the water surface is not constant, and will be steeper on the rising limb than on the recession limb (Meirovich et al., 1998). This means that values of b on the rising limb would be larger than b values for equivalent discharges on the falling limb of the hydrograph. However, the effects of variable water surface slopes are typically deemed negligible in perennial, nonflashy streams where stage increases are slow (Meirovich et al., 1998), as in this study, where water depths increased by only 0.2 m during a 24-h period.
It was also observed that the supply of bacteria available for resuspension regenerates between storm events. As this study involved the use of a tracer bacteria, the transport and deposition of bacteria from upstream sources within the study reach can be eliminated as the regeneration mechanism. One explanation could be that fresh, nutrient-rich sediments are deposited onto the surface of the source cell on the falling limb of the storm hydrograph. Bacteria in the bed would then migrate toward and colonize the fresh surface sediments, thus regenerating the supply of bacteria available for movement during the next stormflow event. Motile bacteria can respond to chemical gradients and move toward nutrient-rich solid surfaces (Marshall, 1985). Vaituzis and Doetsch (1969) reported that Pseudomonas aeruginosa could move at a velocity of 56 µm s–1.
This study provided evidence that bacteria resuspension in streams possessing cohesive beds does not occur until a critical flow condition has been surpassed. It was interesting to note that c values for each of the three storms listed in Table 3 are quite similar. The bed sediments present in Swan Creek are fine-grained and cohesive in nature. Methods used to predict conditions necessary for initiation of motion for noncohesive sediments, such as the Shield's Curve (Chang, 1988), are therefore not applicable. For example, using the Shield's Curve, a c of only 0.15 N m–2 would be required to initiate movement of the median grain size (0.1 mm) of bed sediments present near the source cell. The results from Swan Creek illustrated that c was an order of magnitude greater than indicated by Shield's Curve.
Unfortunately the specification of c for cohesive bed sediments is a difficult task. Cohesive bed shear strength is influenced by organic matter, the chemical environment, consolidation, and temperature (Mehta et al., 1989). In general, there is very little guidance available to predict c for cohesive sediments. Critical bed shear conditions for cohesive beds have primarily been determined through flume studies and have ranged from 0.5 to 1.5 N m–2 (Partheniades, 1965).
These observations can provide some guidance toward the approach used to simulate bacteria resuspension in water quality models. The model presented by Wilkinson et al. (1995) basically assumes that each step increase in flow results in the entrainment of a fraction of the bed sediment bacteria population. The model does not include the identification of a critical condition for resuspension and therefore would probably not be applicable to stream beds possessing cohesive properties. The model of resuspension proposed by Tian et al. (2002) involves the identification of a critical flow that must be surpassed in order for entrainment to occur and would thus be more representative of cohesive conditions. The equation used by Tian et al. (2002) to predict resuspension rates is similar in form to those commonly used to predict erosion of cohesive sediments. Based on the results of this study, and others, their model could be improved by specifying that resuspension only occurs on the rising limb of storm hydrographs and thus incorporating the concept of a finite supply of bacteria.
Current bacteria transport models are somewhat simplified because they do not explicitly simulate sediment movement. The two factors limiting the development of more physically based modeling approaches are the: (i) lack of robust methods for predicting sediment transport, and (ii) lack of information with respect to the partitioning of bacteria between adsorbed and nonadsorbed phases in stream beds. Until these knowledge deficiencies are addressed, simulating bacteria resuspension directly as a function of hydraulic conditions (Q, v, b) will remain the only practical approach.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In streams possessing a mixture of cohesive and noncohesive sediments, the transport properties of the cohesive sediments will largely control the movement of sediment-associated bacteria. The theories that have been developed to describe cohesive sediment transport appear to be applicable to describing sediment-associated bacteria transport in the stream examined in this study. The critical shear stress for E. coli NAR resuspension in Swan Creek ranged from 1.5 to 1.7 N m–2, which is comparable with literature values for critical shear stresses for erosion of cohesive sediments. It was also noted that resuspension of E. coli NAR only occurred on the rising limb of storm hydrographs, implying that a finite supply of sediment-associated microorganisms are available for resuspension during individual storm events.
Future studies, using the experimental procedure discussed in the paper, should be conducted in additional streams to further investigate the influence of bed sediment characteristics, soil and water chemistry, hydraulics, and seasonal conditions on bacterial persistence and movement.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  A. K. Knox, R. A. Dahlgren, K. W. Tate, and E. R. Atwill Efficacy of Natural Wetlands to Retain Nutrient, Sediment and Microbial Pollutants J. Environ. Qual., August 8, 2008; 37(5): 1837 - 1846. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Koirala, R. W. Gentry, E. Perfect, J. S. Schwartz, and G. S. Sayler Temporal Variation and Persistence of Bacteria in Streams J. Environ. Qual., June 23, 2008; 37(4): 1559 - 1566. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Stoeckel and V. J. Harwood Performance, Design, and Analysis in Microbial Source Tracking Studies Appl. Envir. Microbiol., April 15, 2007; 73(8): 2405 - 2415. [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kay, A. C. Edwards, R. C. Ferrier, C. Francis, C. Kay, L. Rushby, J. Watkins, A. T. McDonald, M. Wyer, J. Crowther, et al. Catchment microbial dynamics: the emergence of a research agenda Progress in Physical Geography, February 1, 2007; 31(1): 59 - 76. [Abstract] [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 |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
