Microbial Transport, Retention, and Inactivation in Streams: A

Article pubs.acs.org/est

Microbial Transport, Retention, and Inactivation in Streams: A Combined Experimental and Stochastic Modeling Approach Jennifer D. Drummond,*,†,‡ Robert J. Davies-Colley,‡ Rebecca Stott,‡ James P. Sukias,‡ John W. Nagels,‡ Alice Sharp,‡,§ and Aaron I. Packman† †

Department of Civil and Environmental Engineering, Northwestern University, Evanston, Illinois 60208, United States NIWA (National Institute of Water & Atmospheric Research Ltd.), Hamilton, New Zealand § University of Waikato, Hamilton, New Zealand ‡

S Supporting Information *

ABSTRACT: Long-term survival of pathogenic microorganisms in streams enables long-distance disease transmission. In order to manage water-borne diseases more effectively we need to better predict how microbes behave in freshwater systems, particularly how they are transported downstream in rivers. Microbes continuously immobilize and resuspend during downstream transport owing to a variety of processes including gravitational settling, attachment to in-stream structures such as submerged macrophytes, and hyporheic exchange and filtration within underlying sediments. We developed a stochastic model to describe these microbial transport and retention processes in rivers that also accounts for microbial inactivation. We used the model to assess the transport, retention, and inactivation of Escherichia coli in a small stream and the underlying streambed sediments as measured from multitracer injection experiments. The results demonstrate that the combination of laboratory experiments on sediment cores, stream reach-scale tracer experiments, and multiscale stochastic modeling improves assessment of microbial transport in streams. This study (1) demonstrates new observations of microbial dynamics in streams with improved data quality than prior studies, (2) advances a stochastic modeling framework to include microbial inactivation processes that we observed to be important in these streams, and (3) synthesizes new and existing data to evaluate seasonal dynamics. sediments, microbes are transported along porewater flow paths, and retained in the sediments by a combination of filtration, gravitational settling, and straining.17−19 Hyporheic exchange also delivers organic carbon and nutrients into the streambed, which stimulates growth of microorganisms in benthic and hyporheic biofilms that trap stream-borne microbes.20,21 Microbes trapped within the biofilm matrix can be remobilized back into the water column by biofilm sloughing or bulk erosion.22 Slow release of microbes from hyporheic porewater back to the water column can contribute significantly to overall downstream transport and disease transmission risks.23,24 The repeated deposition and resuspension of microbes combined with long periods of immobilization between resuspension events leads to a wide distribution of in-stream residence times. Reversible storage of microbes occur at both the stream reach-scale (e.g., via hyporheic exchange) and locally within sediments (e.g reversible filtration).17,18 A

1.0. INTRODUCTION Water-borne diseases present a prevalent health burden worldwide. In areas with substantial rangeland and dairy agriculture, the transmission of waterborne zoonotic diseases can be a serious public health issue.1−4 Diffuse sources are of particular concern in areas with large numbers of rangeland animals.5−7 Zoonotic fecal pathogens and indicator microbes enter water bodies in a variety of ways, notably via overland flow and direct inputs from animals. These pathogens are transported from agricultural lands to downstream rivers, lakes, and coastal waters, causing hazards to recreation water users, drinking water supplies, and food supplies via shellfish consumption and irrigated agriculture.1,4,8−11 Monthly monitoring provides an overview of microbial water quality. However, for lotic systems such as streams, transport of microbes is dominated by storm events that are often missed by monthly sampling.12−15 In particular, improved predictions of how and when microbes are transported are needed to protect downstream water users. A wide range of processes affect the transport and retention of microbes in rivers. Microbes are transferred into the underlying sediments by hyporheic exchange.16 Within the © 2015 American Chemical Society

Received: Revised: Accepted: Published: 7825

August 4, 2014 June 2, 2015 June 3, 2015 June 3, 2015 DOI: 10.1021/acs.est.5b01414 Environ. Sci. Technol. 2015, 49, 7825−7833

Article

Environmental Science & Technology

2.0. EXPERIMENTAL METHODS 2.1. Field Tracer Injection Experiment. The tracer injection was conducted in the Toenepi stream, an agricultural (mainly dairy pasture) stream on the North Island of New Zealand. The Toenepi stream has been described in detail previously.12,15,42 Injection and sampling sites matched those used in prior studies.42 In-stream sampling included background measurements taken 9 m upstream of the injection site (Site 1), and at 3 downstream sampling locations (Sites 2, 3, and 4) 63, 130, and 200 m downstream of the injection site, respectively. The injection was conducted in Autumn (April 2012) when the Toenepi streamflow was relatively low, ∼13 L s−1 at the time of the injection. A mixture of a conservative solute (rhodamine WT), fine sediment (clay), and microbes (E. coli) was injected continuously into the stream. The injection commenced in the late afternoon at 17:00 NZDT to minimize inactivation of E. coli by solar radiation. Methods for the injection and in-stream sampling are reported in the SI. We also compare the results with those obtained from a previously published injection experiment42 within the same study reach in Spring (October 2012) to evaluate seasonal variations in solute and fine particle transport. 2.2. Laboratory Column Filtration Experiment. A column filtration experiment was performed on a core of Toenepi streambed sediment collected upstream of the injection site. The streambed was cored using polycarbonate tubes (13.25 cm long × 6 cm diameter) sharpened on one end and a specially designed coring device, as described previously.42 The sediment core was promptly transported to the laboratory and stored in the dark at 4οC until processed (within 48 h). The column experiment was conducted to observe the transport and retention of a conservative solute tracer (sodium chloride), microbes (E. coli), and fluorescent fine particles (Dayglo Fluorescent AX Pigments-Aurora pink, Cleveland, OH) in the core. These particles have a diameter of 1−10 μm, averaging ∼4 μm, with a similar density (1.36 g cm−3) as natural stream particulate organic matter (seston).43 The top 2.4 cm of the sediment core was extruded into a clean polycarbonate tube (6 cm diameter). The extruded subcore was then capped with 250 μm filters, foam caps, and inlet and outlet tubing. Flow through the column was directed upward. The column was flushed with DI water for 3 h before the injection at a flow rate of 3.33 mL min−1 to remove mobile material from porewater. The injection was then performed at this same flow rate. The experiment was performed in the dark at 15 °C, the average stream temperature during the field injection study. The focus of this experiment was to examine the long-term storage and breakthrough behavior of the injected tracers through the sediment column. The injectate containing solute tracer, particles, and E. coli was pumped into the column for 60 min. Preparation of the injectate is detailed in the SI. Immediately following the injection, DI water was pumped into the column for another 6 h to observe the transport and remobilization of the salt tracer, fine particles, and E. coli from the column. The effluent was sampled 30 min before the injection, every 5 min during the injection, then every 10 min for 80 min, every 15 min for 60 min and every 20 min for 160 min. At the end of the experiment, the core was subsectioned at 1 and 1.4 cm depths for analysis of spatial distributions of immobilized particles and E. coli following methods previously described.42 E. coli was

transport model is needed that can link microbial deposition and resuspension processes at multiple scales with wide residence time distributions of storage. Transport models have been developed to estimate retention and downstream propagation of microbes in streams. However, current models typically underestimate both microbial deposition rates and retention time scales. This is due in part to the assumption that bacteria in the water column only deposit by means of gravitational settling when attached to suspended sediment.25−28 Although bacteria readily attach to sediment in the water column, ∼90% attaches to fine sediments with diameters less than 12 μm.28 Therefore, the Stokes’ settling velocity of both freely suspended and particle associated bacteria is low and the leading mechanism of transport into and out of streambed sediments is hyporheic exchange. Immobilization within stream channels is underestimated when hyporheic exchange is not considered.16,29 Recent models have assumed first-order removal and resuspension rates of microbes between the water column and sediments, which imply an exponential residence time distribution of microbes within stream storage areas.30−32However, residence time distributions in streams vary greatly, and recent studies have shown that they will often follow a power-law distribution.33−36 Such long residence time distributions cannot be represented well with first-order models. Both assumptions underestimate exchange and storage of microbes within sediments. Bacteria such as Escherichia coli also have the potential to both grow and inactivate in streams.37,38 Inactivation can be measured directly or estimated as a function of temperature and sunlight.39−41 Inactivation is usually assumed to be negligible and not dependent on the spatial distribution of E. coli in the stream. However, transport of bacteria between the stream and streambed is expected to influence inactivation because exposure to sunlight is normally the dominant control on inactivation,40,41 and this only occurs in the water column. The main objectives of this study were to (1) obtain new observations of microbial dynamics in streams and (2) advance the modeling framework needed to represent the many distinct processes that influence microbial storage and export in river reaches. We extended an existing transport model for solutes and fine particles at baseflow that includes hyporheic exchange, porewater transport and reversible filtration within the streambed sediments18,42 to also account for microbial inactivation in both the water column and sediments. The key processes governing microbial dynamics in streams are shown in the conceptual model in the graphical abstract and Supporting Information (SI). We then used the model to analyze E. coli transport, retention, and inactivation in a small pastoral stream that is subject to considerable inputs of fecal bacteria from livestock. We parametrized the model using results from an in-stream injection of a conservative solute, fine particles, and E. coli, and a supporting column filtration experiment. The multiscale model enables information on solute, particle, and E. coli retention in the column experiment to be related to net downstream transport at the stream-reach scale. We were able to obtain better quality data of microbial transport and retention in streams with improved observations above in-stream background concentrations at both the localand reach-scale than prior studies and advanced the modeling framework to include microbial inactivation processes that we observed to be important in these streams. Finally, we synthesized new and existing data on microbial transport in streams to evaluate seasonal dynamics. 7826

DOI: 10.1021/acs.est.5b01414 Environ. Sci. Technol. 2015, 49, 7825−7833

Article


where C is in-stream concentration, t is the elapsed time, M(t) is the memory function and U and K are, respectively, the velocity and dispersion coefficients that describe motion in the stream. The memory function represents the fraction of solutes, particles, or microbes that are immobilized at time t and are still immobile at a later time (t + dt). The memory function is normally written in Laplace space to simplify the expressions, where the Laplace transform, L{f}(u), of a function f(t) is equal 18,45 −ut to ∫ ∞ 0 e f(t)dt.

analyzed by the Colilert Quanti- Tray/2000 (IDEXX) MPN method, salt tracer was measured by conductivity, fluorescent particles were measured by a combination of flow cytometry and microscopy. All analytical methods are detailed in the SI. 2.3. Calculation of Inactivation Coefficients. The inactivation rate of E. coli in the dark was determined by fitting a first-order expression to observed die-off. The E. coli concentration in a streamwater sample obtained during the injection experiment was monitored over 300 h while it was stirred continuously in the dark at 15 °C. The resulting dark inactivation rate coefficient was kH = 0.12 day−1 = 1.4 × 10−6 s−1, corresponding to 11% loss of E. coli per day. This is in broad agreement with values reported in the literature,44 which show a wide range, but average ∼0.3 day−1. We calculated an average first-order sunlight inactivation rate during the injection using an existing equation for E. coli42 and hourly measurements of insolation at a NIWA meteorological station within the Toenepi catchment. This is a bulk estimation of a sunlight inactivation rate that has commonly been applied in the study region, but does not take into account local variations in water clarity and depth. The resulting average sunlight inactivation rate constant, k0 = 0.8 day−1 = 9.2 × 10−6 s−1.

M̃ (u) = ut ̅

∫0

t

⎡ ∂C(x , t ′) M(t − t ′)⎢ −U ∂x ⎣

+K

∂ 2C(x , t ′) ⎤ ⎥dt ′ ∂x 2 ⎦

1 − ψĩ (u)

(2)

where M̃ (u) is the memory function in Laplace space, u is the Laplace variable, t ̅ is the average travel time in the reach, defined as the stream reach length divided by the mean stream velocity, and ψ̃ i(u) is the residence time probability distribution. This model can be used to represent transport of solutes, fine particles, and microbes with suitable residence time distributions, denoted by subscript i = S, P, and M, respectively. Memory function is able to incorporate Individual processes describing microbial retention, release, and survival can be integrated into this modeling framework within the memory function, which is a key advantage to using this model framework over a first-order model. Sections 3.1 and 3.2 describe how the residence time probability distributions (ψi) are determined for in-stream and porewater transport. 3.1. In-Stream Modeling. The overall residence time distribution in the stream (ψi) is defined by the residence time distribution in the mobile region (water column) (ψ0), the probability of immobilization (Λi [T−1]), and the residence time distribution in the immobile region (φi). Here, we assume that a single distribution ψ0 characterizes the transport of solutes, fine particles, and microbes, since these materials should be transported very similarly in the water column. We take this as an exponential distribution ψ0(t) = e−t for convenience, but the exact form of this distribution does not significantly affect the results.We assume that delivery of fine particles and microbes to the streambed is controlled purely by advective hyporheic exchange and that gravitational settling is negligible because the Stokes’ settling velocity of fine particles is very low, especially organic particles and microbial cells that have low specific gravity. In this case, hyporheic exchange of solute, fine particles, and microbes is similar, and ΛS ≈ΛP ≈ΛM. The residence time distribution for solutes, φS, is based on the time solutes are retained within the streambed by hyporheic exchange or in the water column in dead zones. Solute residence time distributions have often been found to follow a heavy-tailed power law, where φS (t) ∼ t−βS, for 0 < βS < 1.47 “Heavy tailed” here means that the residence time distribution has an infinite mean and variance, which is true for power laws with slope 0 < βS < 1. Power-law exponents closer to 0 have less steep slopes and greater residence times. The residence time distribution for particles, φP, is dependent on the residence time distribution for solutes (φS), the probability that fine particles immobilize (i.e., filter, attach, or deposit) within the immobile region (ΛHP [T−1]), and the residence time distribution of fine particles deposited within the immobile region(φHP). Thus, the model accounts for both fine particle transport into and out of regions of storage, notably the hyporheic zone, and also deposition and resuspension within these regions. The residence time distribution of fine particles is

3.0. STOCHASTIC MOBILE-IMMOBILE MODEL FOR MICROBIAL TRANSPORT IN RIVERS We previously developed a stochastic mobile-immobile model framework for solutes and fine particles in rivers.18,42 Here we extend the model to simulate microbial transport with inactivation both in the water column and in storage areas, and develop capability to explicitly link observations of solute, particle, and microbial dynamics in streambed sediments to reach-scale transport. This multiscale model can incorporate the wide range of residence times of microbes within streams that result from repeated deposition and resuspension with long periods of immobilization between events. Stochastic transport theory suggests that long-term retention depends only on the slowest release process, which can be determined from the model parameters describing transport and retention.48−50 Therefore, this multiscale framework allows us to compare the local- and reach- scale microbial immobilization parameters to assess if the reach-scale retention of microbes primarily reflects the time scale of remobilization within the sediment bed. A conceptual model of the microbial processes included within this mobile-immobile model framework, as well as a more detailed description of the model, are included in the SI. Here we provide a brief review of the key equations and parameters and details on the model updates to account for microbial inactivation. The mobile−immobile model framework is convenient for transport in rivers as the water column can be considered mobile and material retained in streambed sediments or instream structures, such as macrophytes, is effectively immobile.45,46 Advection and dispersion within the water column is convolved with a memory function that describes advective hyporheic exchange and immobilization (eq 1).18,45 ∂C(x , t ) = ∂t

ψĩ (u)

(1) 7827


Article

Environmental Science & Technology also represented as a heavy-tailed power-law distribution φHP (t) ∼t−βHP, for 0 < βHP < 1. To represent inactivation in the water column, the residence time distribution of microbes within the mobile region is subject to the first-order inactivation rate constant for sunlight, k0, yielding ψ0M(t) = e−te−k0t. We also take the residence time distribution of microbes in storage regions as a heavy-tailed power-law residence time distribution subject to the first-order rate for dark inactivation, kH, yielding φHM(t) ∼ t−βHM e−kHt, with 0 < βHM < 1. 3.2. Streambed Sediment Modeling. Within the sediments, porewater flow represents the mobile domain, and storage occurs by either retention in stagnant porewater, for example, in dead-end pores or regions with very low permeability, or immobilization of fine particles and microbes due to filtration. Transport through streambed sediment is dependent on the residence time distribution of solute caused by heterogeneity along hyporheic flow paths (φCS), the probability of immobilization (ΛCj), and the distribution of residence times in the immobile region of the sediments (φCj), where subscript C denotes sediment column and j = P for particles and M for microbes. φ̃ CS can be exponential or powerlaw depending on sediment properties. Fine particles and microbes are immobilized by filtration and subsequently resuspend from the surfaces of bed sediment grains. The probability of particle immobilization, ΛCP, reflects filtration within sediment porewaters. The residence time distribution of fine particles within sediments is often found to be power-law, φCP(t) ∼ t−βCP where the exponent is 0 < βCP < 1. Following prior studies of reversible filtration of organic particles and microbes,17,18 we represent reversible filtration of E. coli as a mixture of power-law and uniform residence time distributions, with both subject to exponential inactivation: φCM (t ) ∼ c × t −βCMe−kHt + (1 − c) × Te−kHt

Figure 1. Observations and simulations of conservative solute, fine particles, and E. coli in the effluent from a column of Toenepi sediment. Lines represent best model fits. For E. coli, the dashed line is the power-law distribution and the solid line is the combination of the power-law and uniform distributions.

tailing that later levels off to a uniform distribution (Figure 1). Similar tailing behavior was previously observed in column filtration experiments with Cryptosporidium parvum oocysts.17The uniform distribution tail cannot continue indefinitely because of the finite number of E. coli cells injected into the column. Therefore, the E. coli concentration is expected to decrease at a later time (beyond the period of observation). Alternatively, it is possible that E. coli could be reproducing within the sediments, which could yield an ongoing release of E. coli to the column effluent. We fit the sediment column data using the stochastic mobileimmobile subsurface model. The model parameters obtained from the column experiment are summarized in Table 1. In Figure 1, we show the model output for E. coli with a pure power-law residence time distribution (c = 1 in eq 3) and also with a mixed power-law and uniform distribution (c = 0.992 in eq 3). Both E. coli models use βCM = 0.99, which is close to an exponential distribution (βCM = 1). Thus, E. coli shows nearexponential removal in the sediment column until later time scales when a uniform distribution is observed. From the breakthrough curves, we evaluated the amount of each tracer that was retained within the sediment column at the end of the 7 h experiment. Although the solute exhibited anomalous transport, 100% of the injected solute was eluted by the end of the experiment. In contrast, the majority of the tracer particles (63.4%) and most of the E. coli (93.5%) injected into the column were retained. However, some of the loss observed between the injected and eluted E. coli is due to inactivation (die-off) during the experiment. The subcoring results indicated that 1.1 × 106 tracer particles were retained in the core, of which 55% of these were retained within the first 1 cm of the column. The total E. coli retained within the core was 1.8 × 104 MPN, with 77% retained with the first 1 cm of the column. These results indicate that most of the fine particles and E. coli are immobilized in the streambed very near to the sediment-water interface. 4.2. Solute, Particle, And Microbe Transport in Toenepi Stream. Observed in-stream breakthrough curves for the conservative solute (rhodamine), tracer particles, and E. coli at Sites 2, 3, and 4 are shown in Figure 2. Solute, tracer particles and E. coli all show power-law tailing behavior, which appears as a straight line in log10−log10 space. Therefore, the delays in solute due to hyporheic exchange and flow around instream structures lead to a wide range of retention times in the

(3)

where c is the fraction of power-law behavior, βCM is the powerlaw exponent for microbes, kH is the dark inactivation rate [T−1], T represents the truncation time for the distribution, i.e., the maximum observation time for the breakthrough curve. The local- and reach-scale microbial immobilization parameters (i.e., ΛCM vs ΛHM, and βCM vs βHM) can then be compared to assess the importance of local-scale processes on reach-scale processes.

4.0. MULTISCALE SIMULATIONS OF MICROBIAL TRANSPORT 4.1. Transport and Retention within Streambed Sediment. The concentration vs time plots (breakthrough curves) of all tracers in the column effluent are shown in Figure 1. Effluent concentrations are normalized by injectate concentrations in order to directly compare the breakthrough curves of each tracer. Tracer particles and E. coli were retained within the sediment to a greater extent than sodium chloride (Figure 1). Further, E. coli was filtered to a greater extent than tracer particles. The solute breakthrough curve exhibited power-law tailing, represented by a straight line in log10−log10 space at late times (Figure 1). This anomalous solute transport (transport that does not follow the predictions of classic advection−diffusion-dispersion theory) demonstrates the wide range of time scales of porewater transport in the heterogeneous Toenepi sediment. The tracer particles also show power-law tailing, whereas E. coli shows initial power-law 7828


Article

Environmental Science & Technology Table 1. Model Parameters for the Laboratory Column Experiment solute parameters L (cm) 2.5

v (m s−1)

D (m s−2)

−5

−5

2.0 x 10

1.7 × 10

E. coli parameters

particle parameters βCS 0.58

ΛCP (s−1) −3

3.0 × 10

βCP 0.20

ΛCM (s−1) −2

5.0 × 10

βCM

c

T (h)

0.99

0.992

15

Figure 3. In-stream breakthrough curves and stochastic model first for the conservative solute and tracer particles at Site 4 in log10−log10 space. Concentrations are normalized by the respective average concentrations at the injection point. Breakthrough curves plotted in linear space are shown in the SI.

are summarized in Table 2. The mean stream velocity v = 0.022 m s−1 and dispersion coefficient D = 0.092 m s−2 were determined from the solute data. The probability of immobilization of solute, ΛS, was 2.0 × 10−2 s−1, indicating that 2% of the solute mass exchanges with storage regions per second. The solute residence time distribution was a power-law with exponent βS = 0.58, the same exponent found for the sediment column experiment. The model best represented the tracer particle data with a probability of immobilization of ΛHP = 3.0 × 10−3 s−1 and a power-law residence time distribution with exponent βHP = 0.20, which also match the results of the sediment column experiment. Stochastic transport theory suggests that long-term retention depends only on the slowest release process.48−50 The fact that the power-law slope for the solute residence time distribution in the stream Matches that observed in the sediment column indicates that hyporheic exchange and slow porewater transport controls overall solute retention in the stream. Similarly, the particle immobilization probability and residence time distribution in the stream match those observed in the sediment column experiment, indicating that the slow remobilization due to reversible filtration of particles from the streambed controls long-term particle retention in the stream. Observations and simulations of E. coli breakthrough curves in the stream are presented in Figure 4. For the simulations, we assume that retention of E. coli in the stream is controlled by hyporheic exchange followed by reversible filtration. Therefore, we simulate E. coli transport at the stream-reach scale using the multiscale model with the probability of immobilization from the stream, ΛS, and power-law remobilization with exponent, βS, based on the solute data. We then apply the probability of immobilization and power-law remobilization within the streambed based on the column data, ΛHM=ΛCM = 5.0 × 10−2 s−1and βHM = βCM = 0.99 (eq 3). We then applied the dark inactivation rate (kH = 1.4 × 10−6 s−1) in the immobile region, and the sunlight inactivation rate (k0 = 9.2 × 10−6 s−1) in the mobile region. The model simulations using these parameters underestimated in-stream retention, as can be seen in Figure 4 with higher in-stream concentrations predicted by the model in the tail of the breakthrough curve compared to data. Increasing the immobile inactivation parameter to kH = 4.0 × 10−3 s−1, yielded an improved fit that better represented the tail of the breakthrough curve. Thus, it appears that the inactivation of E. coli in the streambed was underestimated using the laboratory derived rate. Alternatively, some additional (nonmodeled) process could have caused additional removal of E. coli within the stream.

concentrations are normalized by the respective average concentrations at the injection site. The peak of the particle breakthrough curve is less than the solute peak, reflecting net removal of particles from suspension in the stream. This indicates greater retention of tracer particles within the stream reach compared to the conservative solute, as expected. The best-fit model parameters for in-stream transport and retention

5.0. COMPARISON OF AUTUMN AND SPRING IN-STREAM EXPERIMENTS To evaluate seasonal dynamics of solute, fine particles, and E. coli transport in the Toenepi stream, we compared the results presented here from April 2012 (autumn) with those from another injection experiment conducted in the same reach in October 2012 (spring).42 The flow was lower in April (13 L s−1) than in October (35 L s−1), and the stream reach also

Figure 2. Breakthough curves for a conservative solute (rhodamine), tracer particles, and E. coli in the Toenepi stream, plotted in log10− log10 space. These same breakthrough curves plotted in linear space are shown in the SI.

stream. Fine particles are expected to have an even wider residence distribution than solute since they are subject to longterm retention by immobilization in the streambed, as indicated by the sediment column results (Section 4.1). The stochastic mobile-immobile model was used to fit the instream breakthrough curves at Site 4 using Site 2 as the input boundary condition. The best-fit simulations for solute and fine particles at Site 4 are shown in Figure 3. All in-stream

7829


Article


Table 2. Stochastic Mobile-Immobile Model Parameters for Solute, Fine Particle and E. coli Breakthrough Curves at Site 4a solute parameters v (m s−1) −1

42

October (Spring, 35 L s ) previous study April (Autumn, 13 L s‑1) this study

D (m s−2)

0.052 0.022 E. coli parameters ΛHM (s‑1) 5.0 × 10−2

0.095

particle parameters

ΛS (s−1) 6.0 × 10

−2

βS 0.72

ΛHP (s−1) −2

βHP

8.0 × 10

0.85

3.0 × 10−3

0.20

0.092

2.0 × 10−2

0.58

βHM 0.99

k0 (s−1) 9.2 × 10−6

kH (s−1) measured: 1.4 × 10−6 fit: 4.0 × 10−3

The model parameters for solute and fine particles from a previous study42 at Site 4 within the same stream reach during an October experiment are also included (See Section 5.0).

a

concentration than the clay used in April. Therefore, the similarity in particle retention between October and April despite the ∼3× difference in streamflow rate may result from either different particle transport mechanisms associated with differences in stream conditions, or from using different tracer particles. Further experiments are needed to evaluate transport mechanisms of fine particles having varying properties. Very few particle injections have been conducted in streams and rivers, and the available knowledgebase is insufficient to assess the effects of particle composition and surface properties on transport dynamics. Figure 4. E. coli breakthrough curves in the Toenepi stream at Site 4 in log10−log10 space. The dark line represents the model prediction and the dashed gray line is the model simulation with an increased dark inactivation rate. Breakthrough curves plotted in linear space are shown in the SI.

6.0. DISCUSSION We used a multiscale stochastic mobile-immobile model to integrate laboratory measurements of solute, particle, and microbial transport and retention in streambed sediments with field observations of transport at the stream-reach scale. By linking multiple scales through parametrization of the stochastic model, we determined that hyporheic exchange controlled solute and fine particle transport and retention in the stream. The new observations combined with the advancements of the stochastic model to incorporate long retention times and microbial inactivation within the water column and within storage demonstrated E. coli net removal at the reach scale, with greater overall removal in the stream than at the laboratory scale. Furthermore, a comparison of stochastic model parameters of multiple in-stream experiments (this study and a previously published experiment42) showed the increased retention of solute, particles, and microbes during the lowerflow (Autumn) experiment. Solute, fine particles and microbes were delivered into streambed sediments by hyporheic exchange and transported advectively through porewaters. Power-law tailing of solute was observed even within a short (2.4 cm) column of streambed sediments, indicating that the heterogeneous Toenepi sediment produced anomalous porewater transport. The same power-law tailing was also observed at the stream-reach scale. Stochastic transport theory suggests that long-term retention depends only on the slowest release process.48−50 Therefore, the asymptotic power-law tail, which characterizes overall retention, observed in reach-scale breakthrough curves should be controlled by the heaviest power-law tail observed at any scale. The fact that the reach-scale power-law tail matches the column-scale power-law tail indicates that sediment heterogeneity controls long-term solute retention in the Toenepi Stream. Fine particle transport at the reach scale was similarly controlled by hyporheic exchange and sediment heterogeneity, as both the column and reach-scale observations were described by the same model parameters (ΛHP = ΛCP and βHP = βCP). In

contained more macrophytes in April. Therefore, we expected to find increased solute retention in April owing to greater delays in transport through submerged macrophytes and slower hyporheic exchange flow. Similarly, we expected greater particle and microbial retention in April owing to greater opportunity for immobilization in sediments and on macrophyte surfaces. Total retention of injected solute, tracer particles, and E. coli are compared in Table 3, and best-fit model parameters for Table 3. Total Retention of Solute, Tracer Particles and E. coli in the Toenepi Stream in April vs. October, 2012 solute April (Autumn, 13 L s−1) October (Spring, 35 L s−1)

33.8% 20.9%

tracer particles 64.6% (clay turbidity) 62.0% (fluorescent fine particles)

E. coli 67.1% 50.3%

solute and fine particles are compared in Table 2. Solute and E. coli retention was appreciably higher in April, whereas retention of different tracer particles happened to be very similar between the two experiments. The stream velocity was nearly 3-fold greater in October, but the dispersion coefficients were similar. The probability of immobilization of both solutes and particles was higher in October when there was a higher flow rate. However, the power-law exponent was closer to 1 for both solutes and fine particles in October, indicating a narrower distribution of retention times in the stream in October compared with April. The similar retention of tracer particles might be artifactual given that different tracer particles were used in April and October. The fluorescent fine particles used in October were detected at a lower concentration and had a lower background 7830


Article


presented here. As a result, it is possible that our simulations underestimate sunlight inactivation. Inactivation of E. coli immobilized at the streambed surface or within submerged vegetation is particularly likely to be underestimated since neither storage zone, in contrast to the hyporheic zone, is likely to be fully dark. The model better matched the data when the dark inactivation rate constant was increased, which lead to fewer viable E. coli cells returning to the stream from the streambed. Alternate methods may be more accurate for calculating dark inactivation rates,51,52 however, without further supporting data−such as direct observations of retention and inactivation in multiple areas of the stream, we cannot conclude which region (i.e., stream, subsurface, or both) controls the difference between the model prediction and the data. These processes could readily be included in the stochastic model framework presented here, if suitable data were available to parametrize these variations within the study reach. At the stream reach scale we observed net removal of E. coli, indicating that the combination of uptake by vegetation, sedimentation, sunlight inactivation, and protozoan grazing collectively outweighed growth. However, the possibility of E. coli replication in the sediments must be considered as E. coli is a robust facultative organism known to form biofilms under diverse environmental conditions. If E. coli grows in the sediments, then the streambed can become persistently colonized by E. coli, making the sediments an ongoing source of this bacterium. In fact, we previously observed long-term (months) retention of fine particles and continuous presence of E. coli in Toenepi streambed sediments.15,42 Based on these prior observations and the results presented here it appears likely that spatial variability in hyporheic exchange and filtration within streambed sediments controls overall immobilization and resuspension of fine particles and bacteria at the streamreach scale, with long-term storage deeper in the streambed or in sediment patches that represent hot spots for deposition. The net role of inactivation vs growth of E. coli can be incorporated into the model, but data are generally not available to uniquely parametrize growth and death processes in the field. Better observations of ecological interactions between enteric organisms and aquatic ecosystems are needed. Further, rapid resuspension of microbes occurs during periods of high streamflow that scour streambed sediments.15,53 The extent of microbial resuspension is dependent on the frequency of high flow events; microbial colonization of benthic biofilms, macrophytes, and sediments; and potential for erosion or remobilization from these reservoirs.12,13 Temporal variability (nonstationarity) in environmental processes (e.g., time-varying solar irradiance) and in E. coli behavior (e.g., biofilm formation that increases the probability of longer term retention) is not currently represented in the model, but these types of processes can be incorporated into the stochastic model framework. This work demonstrated how a combination of laboratory and in-stream measurements combined with modeling can advance our understanding of the transport and retention of fine particles and microbes in rivers. Power-law tailing was observed in both the laboratory sediment column and stream reach-scale results, indicating that solutes, particles, and bacteria were all slowly remobilized over a long period of time. Such a wide range of residence times cannot be adequately represented by models that assume first-order immobilization/resuspension behavior or exponential residence time distributions.31−33 Our model incorporates hyporheic exchange that produces longterm retention even of conservative solutes, while immobiliza-

particular, since the power-law residence time distribution for particles in the sediment column matched the one that described the in-stream breakthrough curve (βCP = βHP = 0.20), we conclude that particle remobilization at the streamreach scale was controlled by reversible filtration in the streambed. Submerged vegetation and in-stream dead zones can also capture large numbers of particles and microbes, but are less likely to retain them for long periods of time because of the greater potential for remobilization from in-stream locations relative to the sediment bed. Previously, we observed that while substantial numbers of fine particles attach to vegetation, over time scales of months these particles are remobilized and migrate into streambed sediments, where they are retained for much longer periods of time.42 Our findings for fine particle transport in streams also apply broadly to microbes, but microbes are also subject to unique biological processes. E. coli breakthrough curves showed increased retention compared to fine particles both in the sediment column and in the Toenepi stream. E. coli exhibited an increased probability of immobilization and longer time scales of retention within the sediment column compared to fine particles (βCM = 0.99, βCP = 0.20). Further, direct measurements of material deposited within sediment cores showed that E. coli was retained within the streambed sediments to a greater extent than fine particles. The majority of both fine particles (55%) and E. coli (77%) were retained within the first centimeter of the sediment column. Similarly rapid immobilization within streambed sediments was observed in previous field studies in the Toenepi Stream: following a reach-scale injection, 88% of injected fine particles and 75% of E. coli were found within the top 3 cm of bed sediments.42 Thus, both fine particles and microbes delivered into streambed sediments by hyporheic exchange deposit quickly and are retained near the sediment−water interface. However, within the sediments, fine particles showed only power-law tailing while E. coli showed a combination of power-law and uniform tailing. Fine particles will be released indefinitely but at everdecreasing concentrations. E. coli are released at higher concentrations when the tail is controlled by the uniform residence time distribution, but the time period of release must be shorter than fine particles (i.e., not infinite) as long as the E. coli are not reproducing in the sediments. The transport and retention of E. coli at the reach scale was predicted using in-stream solute data to characterize in-stream transport and hyporheic exchange, immobilization parameters from column experiments, and the rate constants for inactivation in sunlight and in the dark. The model successfully predicted (without fitting) the overall shape of the in-stream breakthrough curve for E. coli. This represents the first model able to predict reach-scale microbial transport from a combination of measured reach-scale hydrodynamic transport parameters and upscaling of measured local-scale microbial immobilization, remobilization, and inactivation rates. The model underestimated the magnitude of in-stream retention, implying that either the rate of inactivation of E. coli in retention areas was underestimated or another process (unrepresented in the model) removed E. coli from the stream. Factors that affect photoinactivation such as depth of the water column, streamflow, and water column optics (e.g., UV-light screening by chromophoric dissolved organic matter) are also likely to be important controls of microbial densities throughout the day. However, we did not have the local observations needed to include these effects in the simulations 7831


Article


(3) Pitkänen, T. Review of Campylobacter spp. in drinking and environmental waters. J. Microbiol. Methods 2013, 95 (1), 39−47 DOI: 10.1016/j.mimet.2013.06.008. (4) Scallan, E.; Hoekstra, R. M.; Angulo, F. J.; Tauxe, R. V.; Widdowson, M. A.; Roy, S. L.; Jones, J. L.; Griffin, P. M. Foodborne illness acquired in the United StatesMajor pathogens. Emerging Infect. Dis. 2011, 17 (1), 7−15 DOI: 10.3201/eid1701.P11101. (5) Cooley, M.; Carychao, D.; Crawford-Miksza, L.; Jay, M. T.; Myers, C.; Rose, C.; Keys, C.; Farrar, J.; Mandrell, R. E. Incidence and tracking of Escherichia coli O157:H7 in a major produce production region in California. PLoS One 2007, 2 (11), e1159 DOI: 10.1371/ journal.pone.0001159. (6) Gorski, L.; Parker, C. T.; Liang, A.; Cooley, M. B.; Jay-Russell, M. T.; Gordus, A. G.; Atwill, E. R.; Mandrell, R. E. Prevalence, distribution, and diversity of Salmonella enterica in a major produce region of California. Appl. Environ. Microbiol. 2011, 77 (8), 2734−2748 DOI: 10.1128/AEM.02321-10. (7) Till, D. G.; McBride, G. B.; Bail, A.; Taylor, K.; Pyle, E. Large scale freshwater microbiological study: Rationale, results and risks. J. Water Health 2008, 6 (4), 444−460 DOI: 10.2166/wh.2008.071. (8) Wilkes, G.; Edge, T. A.; Gannon, V. P. J.; Jokinen, C.; Lyautey, E.; Neumann, N. F.; Ruecker, N.; Scott, A.; Sunohara, M.; Topp, E.; Lapen, D. R. Associations among pathogenic bacteria, parasites, and environmental and land use factors in multiple mixed-use watersheds. Water Res. 2011, 45 (18), 5807−5825 DOI: 10.1016/ j.watres.2011.06.021. (9) Bradford, S. A.; Morales, V. L.; Zhang, W.; Harvey, R. W.; Packman, A. I.; Mohanram, A.; Welty, C. Transport and fate of microbial pathogens in agricultural settings. Crit. Rev. Env. Sci. Technol. 2013, 43 (8), 775−893 DOI: 10.1080/10653389.2012.710449. (10) Emerging Issues in Water and Infectious Disease; WHO Library Cataloguing-in-Publication Data: Geneva, Switzerland, 2003; http:// www.who.int/water_sanitation_health/emerging/emerging.pdf. (11) Waterborne Zoonoses: Identification, Causes, and Control; Cotruvo, J. A.; Dufour, A.; Rees, G.;Bartram, J.; Carr, R.; Cliver, D. O.; Craun, G. F.; Fayer, R.; Gannon, V. P. J., Eds.; World Health Organization and IWA Publishing: London, 2004. (12) Davies-Colley, R.; Lydiard, E.; Nagels, J. Stormflow-dominated loads of faecal pollution from an intensively dairy-farmed catchment. Water Sci. Technol. 2008, 57 (10), 1519−1523 DOI: 10.2166/ wst.2008.257. (13) Jaimeson, R.; Joy, D. M.; Lee, H.; Kostaschuk, R.; Gordon, R. Transport and deposition of sediment-associated Escherichia coli in natural streams. Water Res. 2005, 39, 2665−2675 DOI: 10.1016/ j.watres.2005.04.040. (14) McKergow, L. A.; Davies-Colley, R. J. Stormflow dynamics and loads of Escherichia coli in a large mixed land use catchment. Hydrol. Process. 2005, 24, 276−289 DOI: 10.1002/hyp.7480. (15) Stott, R.; Davies-Colley, R.; Nagels, J.; Donnison, A.; Ross, C.; Muirhead, R. Differential behavior of Escherichia coli and Campylobacter spp. in a stream draining dairy pasture. J. Water Health 2011, 09 (1), 59−69 DOI: 10.2166/wh.2010.061. (16) Searcy, K. E.; Packman, A. I.; Atwill, E. R.; Harter, T. Deposition of Cryptosporidium oocysts in streambeds. Appl. Environ. Microbiol. 2006, 72 (3), 1810−1816 DOI: 10.1128/AEM.72.3.1810-1816.2006. (17) Cortis, A.; Harter, T.; Hou, L.; Atwill, E. R.; Packman, A. I.; Green, P. G. Transport of Cryptosporidum parvum in porous media: Long-term elution experiments and continuous time random walk filtration modeling. Water Resour. Res. 2006, 42, W12S13 DOI: 10.1029/2006WR004897. (18) Drummond, J. D.; Aubeneau, A. F.; Packman, A. I. Stochastic modeling of fine particulate organic carbon dynamics in rivers. Water Resour. Res. 2014a, 50 (5), 4341−4356 DOI: 10.1002/ 2013WR014665. (19) Harter, T.; Wagner, S.; Atwill, E. R. Colloid transport and filtration of Cryptosporidium parvum in sandy soils and aquifer sediments. Environ. Sci. Technol. 2000, 34, 62−70 DOI: 10.1021/ es990132w.

tion and resuspension in streambed sediments extend the residence time of both particles and E. coli within the stream. The model also accounts for different rates of inactivation of bacteria within the mobile (sunlit) and immobile (dark) regions. Thus, the stochastic mobile-immobile model provides a framework to connect immobilization and resuspension of fine particles and microbes in streambed sediments and the water column over a wide range of spatial and temporal scales. Additional experimental observations are needed to improve the ability to predict pathogen transmission in streams, particularly to parametrize microbial transport, retention, growth, and inactivation in sediments, macrophyte stands, biofilms, and other important retention areas. Because our model framework provides a means to properly upscale local observations, it can incorporate detailed measurements of transport, retention, growth and inactivation processes at multiple scales and can thereby support improved management of waterborne diseases by providing better capability to estimate delivery of viable pathogens from distributed sources to locations of water use.

■

ASSOCIATED CONTENT

S Supporting Information *

(1) Methods for the injection and in-stream sampling, (2) details of injectate preparation, (3) details of all analytical methods, (4) a conceptual model of the microbial processes included within the mobile-immobile model framework, (5) a more detailed description of the model, and (6) breakthrough curves plotted in linear space. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b01414.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Funding for this work was provided through the “Clean Water − Productive Land” Research programme, New Zealand Ministry of Business Innovation and Employment contract C10X1006, U.S. National Science Foundation grants EAR1215898 and EAR-1344280, and an Environmental Protection Agency STAR Graduate Fellowship to J. Drummond. We thank Karen Thompson for guidance on flow cytometry and Margaret Bellingham for providing hydrological and solar irradiance information on Toenepi Stream.

■

REFERENCES

(1) Dufour, A.;Wade, T. J.; Kay, D. Epidemiological studies on swimmer health effects associated with potential exposure to zoonotic pathogens in bathing beach waterA review. In Animal Waste, Water Quality and Human Health; Dufour, A., Bartram, J., Bos., R., Gannon, V., Eds.; World Health Organization/IWA Publishing: London, U.K., 2012; pp 415−428. (2) Jokinen, C. C.; Edge, T. A.; Koning, W.; Laing, C. R.; Lapen, D. R.; Miller, J.; Mutschall, S.; Scott, A.; Taboada, E. N.; Thomas, J. E.; Topp, E.; Wilkes, G.; Gannon, V. P. Spatial and temporal drivers of zoonotic pathogen contamination of an agricultural watershed. J. Environ. Qual. 2012, 41 (1), 242−252 DOI: 10.2134/jeq2011.0203. 7832


Article


and the Human Health; Bartram, J., Ed.; WHO/IWA Publishing: London, U.K., 2003; pp 84−85. (38) Chekabab, S. M.; Paquin-Veilette, J.; Dozois, C. M.; Harel, J. The Ecological habitat and transmission of Escherichia coli O157:H7. FEMS Microbiol. Letters 2013, 341 (1), 1−30 DOI: 10.1111/15746968.12078. (39) Blaustein, R. A.; Pachepsky, Y.; Hill, R. L.; Shelton, D. R.; Whelan, G. Escherichia coli survival in waters: Temperature dependence. Water Res. 2013, 47 (2), 569−578 DOI: 10.1016/ j.watres.2012.10.027. (40) Sinton, L. W.; Finlay, R. K.; Lynch, P. A. Sunlight inactivation of fecal bacteriophages and bacteria in sewage-polluted seawater. Appl. Environ. Microbial. 1999, 65 (8), 3605−3613. (41) Sinton, L. W.; Hall, C. H.; Lynch, P. A.; Davies-Colley, R. J. Sunlight inactivation of fecal indicator bacteria and bacteriophages from waste stabilization pond effluent in fresh and saline waters. Appl. Environ. Microbial. 2002, 68 (3), 1122−1131 DOI: 10.1128/ AEM.68.3.1122−1131.2002. (42) Drummond, J. D.; Davies-Colley, R. J.; Stott, R.; Sukias, J. P.; Nagels, J. W.; Sharp, A.; Packman, A. I. Retention and remobilization dynamics of fine particles and microorganisms in pastoral streams. Water Res. 2014b, 66, 459−472 DOI: 10.1016/j.watres.2014.08.025. (43) Cushing, C. E.; Minshall, G. W.; Newbold, J. D. Transport dynamics of fine particulate organic matter in two Idaho streams. Limnol. Oceanogr. 1993, 38 (6), 1101−1115 DOI: 10.4319/ lo.1993.38.6.1101. (44) Hipsey, M. R.; Antenucci, J. P.; Brookes, J. D. A generic, process-based model of microbial pollution in aquatic systems. Water Resour. Res. 2008, 44, W07408 DOI: 10.1029/2007WR006395. (45) Boano, F.; Packman, A. I.; Cortis, A.; Revelli, R.; Ridolfi, L. L. A continuous time random walk approach to the stream transport of solutes. Water Resour. Res. 2007, 43 (10), W10425 DOI: 10.1029/ 2007WR006062. (46) Schumer, R.; Meerschaert, M. M.; Baeumer, B. Fractional advection-dispersion equations for modeling transport at the Earth surface. J. Geophys. Res. 2009, 114, F00A07 DOI: 10.1029/ 2008JF001246. (47) Boano, F.; Harvey, J. W.; Marion, A.; Packman, A. I.; Revelli, R.; Ridolfi, L.; Wörman, A. Hyporheic flow and transport processes. Rev. Geophys.. 2014, 52(4); DOI: 10.1002/2012RG000417. (48) Schumer, R.; Benson, D. A.; Meerschaert, M. M.; Baeumer, B. Fractal mobile/immobile solute transport. Water Resour. Res. 2003, 39 (10), 1296 DOI: 10.1029/2003WR002141.. (49) Benson, D. A.; Meerschaert, M. M. A simple and efficient random walk solution of multi-rate mobile/immobile mass transport equations. Adv. Water Resour. 2009, 32, 532−539 DOI: 10.1016/ j.advwatres.2009.01.002. (50) Metzler, R.; Klafter, J. The random walk’s guide to anomalous diffusion: A fractional dynamics approach. Phys. Rep. 2000, 339, 1−77 DOI: 10.1016/S0370-1573(00)00070-3. (51) Gronewold, A. D.; Myers, L.; Swall, J. L.; Noble, R. T. Addressing uncertainty in fecal indicator bacteria dark inactivation rates. Water Res. 2011, 45 (2), 652−664 DOI: 10.1016/ j.watres.2010.08.029. (52) Gronewold, A. D.; Qian, S. S.; Wolpert, R. L.; Reckhow, K. H. Calibrating and validating bacterial water quality models: A Bayesian approach. Water Res. 2009, 42 (13), 2688−2698 DOI: 10.1016/ j.watres.2009.02.034. (53) Muirhead, R. W.; Davies-Colley, R. J.; Donnison, A. M.; Nagels, J. W. Faecal bacteria yields in artificial flood events: Quantifying instream stores. Water Res. 2004, 38, 1215−1224 DOI: 10.1016/ j.watres.2003.12.010.

(20) Arnon, S.; Marx, L. P.; Searcy, K. E.; Packman, A. I. Effects of overlying velocity, particle size, and biofilm growth on streamsubsurface exchange of particles. Hydrol. Process. 2010, 24, 108−114 DOI: 10.1002/hyp.7490. (21) Battin, T. J.; Kaplan, L. A.; Findlay, S.; Hopkinson, C. S.; Marti, E.; Packman, A. I.; Newbold, J. D.; Sabater, F. Biophysical controls on organic carbon fluxes in fluvial networks. Nat. Geosci. 2008, 1, 95−100 DOI: 10.1038/ngeo101. (22) Boulêtreau, S.; Garabétian, F.; Sauvage, S.; Sánchez-Pérez, J.-M. Assessing the importance of a self-generated detachment process in river biofilm models. Freshwater Biol. 2006, 51, 901−912 DOI: 10.1111/j.1365-2427.2006.01541.x. (23) Ginn, T. R.; Wood, B. D.; Nelson, K. E.; Scheibe, T. D.; Murphy, E. M.; Clement, T. P. Processes in microbial transport in the natural subsurface. Adv. Water Resour. 2002, 25, 1017−1042. (24) Litton, R. M.; Ahn, J. H.; Sercu, B.; Holden, P. A.; Sedlak, D. L.; Grant, S. B. Evaluation of chemical, molecular, and traditional markers of fecal contamination in an effluent dominated urban stream. Environ. Sci. Technol. 2010, 44, 7369−7375 DOI: 10.1021/es101092g. (25) Bai, S.; Lung, W. Modeling Sediment impact on the transport of fecal bacteria. Water Res. 2005, 39, 5232−5240 DOI: 10.1016/ j.watres.2005.10.013. (26) Droppo, I. G.; Krishnappan, B. G.; Liss, S. N.; Marvin, C.; Biberhofer, J. Modelling sediment-microbial dynamics in the South Nation River, Ontario, Canada: Towards the prediction of aquatic and human health risk. Water Res. 2011,45, 3797−3809. DOI 10.1016/ j.watres.2011.04.032. (27) Gao, G.; Falconer, R. A.; Lin, B. Numerical modelling of sediment-bacteria interaction processes in surface waters. Water Res. 2011, 45, 1951−1960 DOI: 10.1016/j.watres.2010.12.030. (28) Walters, E.; Graml, M.; Behle, C.; Müller, E.; Horn, H. Influence of particle association and suspended solids on UV Inactivation of fecal indicator bacteria in an urban river. Water, Air, Soil Pollut. 2014b, 225, 1822 DOI: 10.1007/s11270-013-1822-8.. (29) Ren, J.; Packman, A. I. Effects of particle size and background water composition on stream-subsurface exchange of colloids. J. Environ. Eng. 2002, 128 (7), 624−634 DOI: 10.1021/es034852l. (30) Cho, K. H.; Pachepsky, Y. A.; Kim, J. H.; Guber, A. K.; Shelton, D. R.; Rowland, R. Release of Escherichia coli from bottom sediment in a first-order creek: Experiment and reach-specific modeling. J. Hydrol. 2010, 391, 3220332 DOI: 10.1016/j.hydrol.2010.07.033.. (31) Grant, S. B.; Litton-Mueller, R. M.; Ahn, J. H. Measuring and modelling the flux of fecal bacteria across the sediment-water interface in the turbulent stream. Water Resour. Res. 2011, 47, W05517 DOI: 10.1029/2010WR009460. (32) Yakirevech, A.; Pachepsky, Y. A.; Guber, A. K.; Gish, T. J.; Shelton, D. R.; Cho, K. H. Modelling transport of Escherichia coli in a creek during and after artificial high-flow events: Three-year study and analysis. Water Res. 2013, 47 (8), 2676−2688 DOI: 10.1016/ j.watres.2013.02.011. (33) Haggerty, R.; Wondzell, S. M.; Johnson, M. A. Power-law residence time distribution in the hyporheic zone of a 2nd-order mountain stream. Geophys. Res. Lett. 2002, 29 (13), 1640 DOI: 10.1029/2002GL014743. (34) Cardenas, M. B. Potential contribution of topography-driven regional groundwater flow to fractal stream chemistry: Residence time distribution analysis of Tóth flow. Geophys. Res. Lett. 2007, 34, L05403 DOI: 10.1029/2006GL029126. (35) Wörman, A.; Wachniew, P. Reach scale and evaluation methods as limitations for transient storage properties in streams and rivers. Water Resour. Res. 2007, 43 (10), W10405 DOI: 10.1029/ 2006WR005808. (36) Stonedahl, S. H.; Harvey, J. W.; Detty, J.; Aubeneau, A.; Packman, A. I. Physical controls and predictability of stream hyporheic flow evaluated with a multiscale model. Water Resour. Res. 2012, 48 (10), W10513 DOI: 10.1029/2011WR011582. (37) Leclerc, H. Relationships between common water bacteria & pathogens in drinking-water. In Heterotrophic Plate Counts and Drinking-Water Safety: The Significance of HPCs for Water Quality 7833


Microbial Transport, Retention, and Inactivation in Streams: A

Recommend Documents