Biodegradation of labile dissolved organic carbon under losing and ...

LIMNOLOGY and

OCEANOGRAPHY

Limnol. Oceanogr. 61, 2016, 1839–1852 C 2016 Association for the Sciences of Limnology and Oceanography V

doi: 10.1002/lno.10344

Biodegradation of labile dissolved organic carbon under losing and gaining streamflow conditions simulated in a laboratory flume Natalie De Falco,1 Fulvio Boano,2 Shai Arnon*1 1

Zuckerberg Institute for Water Research, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Israel 2 Department of Environment, Land and Infrastructure Engineering, Politecnico di Torino, Turin, Italy

Abstract Carbon in surface waters is widely recognized as a key element that influences nutrient cycling, metal availability, and water quality. Its degradation in streams occurs primarily by benthic microbial communities that colonize the underlying sediment, which is commonly termed the hyporheic zone (HZ). The biodegradation of a labile dissolved organic carbon (DOCL), exemplified by sodium benzoate, was studied in a novel laboratory flume system under a combination of different overlying water velocities, losing or gaining fluxes, and biofilm distribution (“surficial” or “homogeneous distribution”). The overall objective of this study was to evaluate the effect of different flow conditions on DOCL biodegradation in the HZ. The results showed that overlying velocity was the dominant factor affecting DOCL biodegradation, regardless of biofilm distribution. Gaining flow conditions also induced a slight increase in the biodegradation rates as compared to losing or neutral flow conditions, due to additional oxygen input from the upwelling water. The aerobic reactive zone under all flow conditions was limited to the upper section of the benthic biofilm (several millimeters), where the surficial biofilm showed the highest activity. Our results demonstrate the processes affecting DOCL biodegradation in the hyporheic zone and will help to implement future modeling of DOC transport in streams.

Dissolved organic carbon (DOC) is a fundamental component of aquatic ecosystems, being the basic building block of food webs and influencing the bioavailability of nutrients and metals (Porcal et al. 2009; Benstead and Leigh 2012). Despite the relatively low coverage area of streams relative to terrestrial and marine environments, their significant role in the global carbon cycle has just recently been acknowledged (Battin et al. 2009; Benstead and Leigh 2012; Regnier et al. 2013). The breakdown of organic matter and the recycling of key nutrients are driven by streambed heterotrophic microorganisms that are preferably organized in benthic biofilm communities (Bott and Kaplan 1985; Battin et al. 2003; Nogaro et al. 2013). Thus, consumption of DOC by streambed biofilms influences the biogeochemistry of the streambed by controlling the chemical conditions and the cascading events of redox-related reactions (Butturini et al. 2000; Sobczak and Findlay 2002; Miller et al. 2006). Regardless of the DOC source, its fate depends on in-stream condi-

*Correspondence: [email protected] Additional Supporting Information may be found in the online version of this article.

tions and the hydrologic interaction with the adjacent sedimentary environment, which is termed the hyporheic zone (HZ). The level and composition of DOC in streams regulate heterotrophic activity, including respiration and growth, which are ultimately represented as energy flow in the system (Vannote et al. 1980; Baker et al. 1999; Battin et al. 2008). While in the past DOC was considered as either labile or recalcitrant (i.e., bioavailable or not bioavailable, respectively), today there is a notion that DOC reactivity can be represented as a continuum (e.g., Koehler et al. 2012). The quality of the DOC depends on its source, in which autochthonous sources (i.e., algal) are considered, in general, more labile than allochthonous sources (Aitkenhead-Peterson et al. 2002). The definition of labile DOC (DOCL) does not always refer to the fraction of the organic carbon that has a rapid turnover time. For example, a recent study also suggested that DOC may be considered as refractory simply because it is too dilute to be an attractive source of energy for microorganisms (Arrieta et al. 2015). Zou et al. (2005) claimed that chemical and physical fractionation methods for estimating DOCL are indirect and lack a clear biological definition. In addition, in a review done on rivers, lakes and marine

1839

De Falco et al.

DOC uptake in losing and gaining streams

systems, Sondergaard and Middelboe (1995) discussed the meaning of DOCL and defined it as the fraction that can decompose in less than 2 weeks. DOCL plays a major role in various biogeochemical processes (e.g., denitrification), especially in the HZ and shallow groundwater due to the relative proximity to its sources (Baker et al. 1999, 2000; Zarnetske et al. 2011). Hyporheic exchange flow exposes the DOC to the benthic environment, and various studies have demonstrated how DOC composition can change in stream networks and along subsurface pathways (Miller et al. 2006; Creed et al. 2015; Helton et al. 2015) and how redox zonation occurs (Bardini et al. 2012; Briggs et al. 2015). Nevertheless, the complex hydrological interactions between the stream and the groundwater have still not been widely considered in studies on DOC metabolism. Stream and groundwater systems can be considered as a continuous environment where water exchange between the stream and the subsurface occurs via the HZ and may vary in time and space (Brunke and Gonser 1997; Boulton et al. 1998; Gomez et al. 2012; Byrne et al. 2014). The intensity and the direction of the exchange between streams and subsurface water follow complex patterns that depend mainly on the hydrologic conditions in both the stream and the groundwater (e.g., water levels), the sediment permeability, and the morphology of the sediment surrounding the stream channel (Dent et al. 2007; Cardenas 2009; Buffington and Tonina 2009). While hyporheic exchange usually refers to the flux of water that goes into the streambed and back to the stream after a variable period of time in the HZ (Boano et al. 2014), losing and gaining fluxes refer to the ultimate water balance between the stream and the groundwater. Under losing conditions, the net vertical flow is directed toward the groundwater, while under gaining conditions, upwelling groundwater flows toward the stream. It was also shown recently that as losing and gaining fluxes increase, the hyporheic exchange flux decreases, regardless of the direction of flow (Cardenas and Wilson 2007; Trauth et al. 2013; Fox et al. 2014). The proximity of the hyporheic zone to the surface and its high reactivity make it an ideal location for biogeochemical processes, in general (Krause et al. 2011), and particularly for DOC transformations (Kaplan et al. 2008; Wagner et al. 2014). These types of zones with disproportionally high reaction rates are usually termed “hotspots” (McClain et al. 2003). At the same time, it was widely acknowledged that losing and gaining flow conditions affect the stream and groundwater chemistry, depending on the flow direction and the chemistry of each environment (e.g., Findlay 1995; Krause et al. 2013). However, the outcome of how the hydrological interaction between the hyporheic exchange and the groundwater upwelling or downwelling regulates DOCL transformation in biogeochemically active streambeds is still difficult to foresee.

Several previous studies that did not take into account the interaction with groundwater have clearly demonstrated that increasing overlying velocities will enhance the mass transfer between the water and the streambed and will increase nutrient uptake (Leu et al. 1998; Thomas et al. 2000; Arnon et al. 2013). The interaction of hyporheic exchange processes and losing/gaining fluxes results in a reduction in the hyporheic fluxes and the size of the hyporheic flow cell (Trauth et al. 2013; Fox et al. 2014), with implications for biogeochemical processes. For example, Trauth et al. (2014) demonstrated that aerobic respiration and denitrification decrease with increasing gaining fluxes (upwelling groundwater) as compared to neutral conditions, while they remained about the same magnitude under losing conditions. Fox et al. (2014) demonstrated that hyporheic exchange flux is controlled by the interaction between the overlying water velocity, and the magnitude and vertical direction of the exchange flux between the stream and the groundwater. Thus, it is expected that such an interaction will control the active region of bacteria (and their proliferation) in the streambed due to the variable delivery of nutrients under different flow conditions. In this study, we investigated the complex interaction between different flow conditions and the biodegradation of DOCL by benthic biofilm. We used a novel recirculating flume system that enabled us to simulate losing and gaining flow conditions and different biofilm distributions. The central hypothesis was that DOCL biodegradation will be mainly influenced by overlying water velocity (regardless of the biomass distribution) and that the most efficient reduction in DOCL concentrations in the stream water will occur under neutral flow conditions. Moreover, we asserted that biomass distribution would play an important role in the degradation of DOCL under losing and gaining flow conditions.

Materials and methods Experimental system The effect of flow conditions and biofilm distribution in the HZ on DOCL biodegradation was studied using a 260cm-long and 29-cm-wide flume (Fig. 1). The flume was packed with natural silica sand (average diameter of 384 lm; hydraulic conductivity of 0.1195 cm s21) to manually form a sand bed with stationary bed forms. The length of the bed forms was 15 cm, and their maximum and minimum depths were 12 cm and 10 cm, respectively. The bed form height and wavelength did not vary along the width of the flume. Average water depth was measured from the water surface to the crest and was equal to 6.5 cm. The aforementioned physical characteristics of the sand bed are representative of many sandy streams as reported in the literature (Stofleth et al. 2007; Lewandowski et al. 2011; Harvey et al. 2013). The flow in the flume was driven by a variable-speed pump that circulates water in the main channel, and creates

1840

De Falco et al.


Fig. 1. Schematic illustration of the flume setup for studying DOC biodegradation processes in streams under various overlying water velocities and losing and gaining fluxes (not to scale). The numbers indicate the main channel (1), the sand bed (2), the drainage/injection system (3), the centrifugal pump for driving the flow in the channel (4), the flowmeter (5), the pump of the filtering loop (6), the filter (7), the UV disinfection system (8), the chiller (9), the spectro::lyser for measuring DOC concentrations (10), the reservoir for storing injected fresh water to the flume under gaining conditions (11), and the peristaltic pump for imposing losing or gaining flow conditions (12). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

a purely gravity-driven open channel flow. The discharge was measured with a magnetic flow meter placed in the return pipe (SITRANS F M, MAG 5100 W Siemens, Germany). The flume was equipped with a drainage system emplaced at the bottom of the sand bed, which enabled us to enforce a known losing or gaining flux (Fox et al. 2014) (Fig. 1). Briefly, the drainage system was made from 64 evenly distributed single pumping points. Gaining and losing stream flow conditions were achieved by injecting or pumping water through the bottom of the sand bed using a peristaltic pump (Master Flex model 7523-80, Cole Parmer, Canada). The losing flux (qL) and gaining flux (qG) were calculated as the water flux imposed by the peristaltic pump per unit of bed area (Fox et al. 2014). Under gaining conditions, the same volume of water that was added through the bottom of the bed was pumped out from the main channel, to maintain a constant volume of water in the system. Under losing conditions, the same amount of water that was pumped out from the bottom was added to the main channel by using distilled water. Neutral conditions referred to the situation when losing or gaining flow conditions were not applied. A side loop of water recirculation was used for filtering the surface water (25 lm), sterilizing it using UV light (UVCD925T5-55W, Aquasafe, Fyshwick, Australia), cooling the

water with a chiller (TR/TC 10, TECO Refrigeration Technologies, Ravenna, Italy), and measuring DOC concentration using an online spectrophotometer (spectro::lyser, S::CAN Messtechnik, Vienna, Austria) (Fig. 1). Biofilm growth Two different biofilms were consecutively developed in the flume and termed “surficial biofilm” and “homogeneously distributed biofilm” Initially, the surficial biofilm was developed by seeding microorganisms that were obtained by collecting pebbles from the Yarqon Stream, Israel. The surficial biofilm was representative of streambeds with stationary bed forms. The biofilm was scraped off the pebbles, filtered through 220-lm mesh to remove large grazers and particles, mixed with a mineral nutrient solution that contained the following constituents (mg in 400 L21): 400 Ca(NO3)2 4H2O, 248 KH2PO4, 500 MgSO4 7H2O, 316 NaHCO3, 45 EDTA Na2, 49.6 H3BO3, 278 MnCl2 4H2O, 60 NaMoO4 2H2O, 0.8 Cyanocobalamin (B12), 0.8 Thiamine HCl, 1140 NaSiO3 9H2O Na2SiO3, 0.8 Biotina, 1.2 CoCl2 6H20, 60 ZnSO4 7H2O, and 12 CuSO4 2H2O, and manually spread over the sand bed. The biofilm was fed every two days by adding 6.8 g of sodium benzoate (NaC7H5O2), and the relative amounts of potassium nitrate (KNO3) and potassium dihydrogen phosphate (KH2PO4)

1841

De Falco et al.


necessary to reach a C:N:P molar ratio of 106: 16: 1. Preliminary tests showed that there was no effect of the side loop, including the UV light, on the concentrations of sodium benzoate. The water temperature was maintained at 25 6 18C, and the flume remained under dark conditions to prevent the growth of photoautotrophic microorganisms for the entire duration of the growth period and the experiments. The biofilm grew under neutral conditions for three months, which were followed by six months of experiments to test the effect of flow conditions on DOCL biodegradation. In the second phase of the experiments, the homogeneously distributed biofilm was developed, following the same feeding strategy that was used to grow the surficial biofilm. The homogeneously distributed biofilm mimicked a case that may occur in migrating bed forms where the streambed is manually mixed at a rate that is faster than microbial growth. It was formed in the flume by manually mixing and homogenizing the streambed the day before each experiment was conducted. Sodium benzoate was used as the carbon source for biofilm growth and later for the reactive transport experiments (see next sections). Various types of DOCL were used in the literature for experiments on biodegradation processes, including acetate (e.g., Zarnetske et al. 2011), glucose and € sel 2010; arabinose (e.g., Leu et al. 1998; Augspurger and Ku Singer et al. 2011), and sodium benzoate (e.g., C¸inar and Leslie Grady 2001; Li et al. 2001), among other compounds. All of the above compounds are labile and possess typical chemical structures that are often observed in nature; thus, they are all suitable for evaluating the biodegradation of DOCL. The main advantage of using sodium benzoate is the ability to take continuous measurements at a high resolution by online spectrophotometry, which enables the degradation rates to be quantified with high accuracy. Biodegradation potential of sodium benzoate A batch degradation experiment was performed in autoclaved 500-mL Erlenmeyer flasks to assess the potential degradation of sodium benzoate as DOCL, applying four different treatments as follows: (1) 120 g of wet sand from the flume and 300 mL of deionized water; (2) 100 g of clean autoclaved dry sand and 300 mL of deionized water (control treatment); (3) 300 mL of sterile deionized water (control treatment), and (4) 300 mL of flume water. Flasks were incubated in a temperature controlled room (258C) with shaking at 120 rpm. Experiment was started immediately after spik21 ing the flasks with DOCL (480 lmol L21), NO2 3 (32 lmol L ) 21 32 and PO4 (16 lmol L ). Experiment was performed in triplicates and duplicates for the sand and water treatments, respectively. Samples were taken every 15 min for the first 3 hours, and then every 45 min for an additional 3 hours. DOC was measured by absorbance at 254 nm, using a Biomate 5 Spectrophotometer (Thermo Scientific, Cambridge, UK).

The effects of flow conditions on DOCL biodegradation The rates of DOCL biodegradation were evaluated by monitoring the DOC concentration in the bulk water under the combinations of two overlying water velocities, (either 4 cm s21 or 12 cm s21), and four subsurface fluxes (qL or qG fluxes of 20 or 50 cm d21). Under neutral flow conditions, rates were also measured under 6, 8, and 10 cm s21. The biodegradation rate was measured by spiking the flume water with a solution containing 6.8 g of sodium benzoate (NaC7H5O2), and the relative amounts of potassium nitrate (KNO3) and potassium dihydrogen phosphate (KH2PO4) necessary to reach a C: N: P molar ratio of 106 : 16 : 1, and following the rate of decline of the sodium benzoate concentrations. Each week started and ended with measurements of the DOCL biodegradation rate under neutral conditions to evaluate the variability within and between weeks. These consecutive additions allowed us to evaluate the effects, such as priming and potential microbial dynamics, that might have influenced the results. In addition to the neutral conditions, measurements under losing and gaining fluxes in the same weeks were alternated to ensure that the variance between the biodegradation rates of DOCL was not caused by the order of the flow conditions that were enforced in the system. The overlying water velocity was kept constant during each week while losing and gaining flow conditions were alternated. For the gaining flow conditions we used deionized water with oxygen concentrations at saturation. Conditioning of the system before each DOCL addition was identical and included the following steps: replacing all of the water in the channel and the flume tanks ( 400 L) with deionized water about 12–16 h before spiking with DOCL, setting up the flow conditions to resemble those planned for the next day; and filtering the water (25 lm). A new filter was also installed just before each spiking. The DOCL concentration, NO2 3 concentration, and turbidity were very low and similar at the starting point of each nutrient addition. Analytical techniques Equivalent DOC concentrations in the bulk water were measured using the spectro::lyser, a UV/Vis online spectrophotometer that uses a Xenon lamp as a light source, and records light absorption between 200 nm and 750 nm. The equivalent DOC concentrations were measured with the spectro::lyser using a built-in algorithm, which is further calibrated with the standard solutions of sodium benzoate to enhance the accuracy of the measurements. These equivalent DOC concentrations were compared to DOC measurements from the flume water (analyzed with a TOC analyzer of Analytik Jena multi N/C 2100S, Jena, Germany). A high correlation between the DOC measurements using the spectro::lyser and the TOC machine was found (R2 5 0.98, n 5 20, Supporting Information I). The fact that the only major DOC source

1842

De Falco et al.


in the flume was sodium benzoate suggests that the DOC measurements by the spectro::lyser reliably reflect the concentrations of sodium benzoate. Oxygen profiling The vertical distribution of oxygen concentrations in the benthic biofilm was measured using a microelectrode system during the DOCL biodegradation experiments. A Clark-type oxygen microelectrode, mounted on a micromanipulator with computerized depth control and data acquisition (Unisense, Aarhus, Denmark), was inserted from the bulk water into the sand bed to measure the change of oxygen concentrations in the middle of the stoss side of the bed form. The oxygen microelectrode had a tip diameter of 50 lm, a stirring sensitivity of < 2% and a 90% response time of < 5 s. The microelectrodes were calibrated every day following the manufacturer’s protocol to ensure that there were no drifts in the measurements. Biofilm sampling and analysis Samples were taken at the end of each set of experiments with constant overlying velocity and varying neutral/losing/ gaining flow conditions. Biofilm sampling (from the surficial biofilm and the homogeneously distributed biofilm) was done from three different bed forms that were located in the central part of the flume to ensure that there were no boundary effects due to the potential variability of flow patterns near the entrance and exit of the main channel. In each bed form, four sampling locations were located at distances of 0, 4, 8, and 13 cm along the bed form (A, B, C, and D, respectively; Supporting Information II). Location A is at the trough, B is on the stoss side, C is on the crest, and D is on the lee side of the bed form. Initially, the flow was stopped and the water level was lowered to the surface of the sand. Autoclaved plastic rings with a diameter of 2.5 cm and a height of 1 cm were used to collect the samples. The rings were inserted into the sand and the samples were scooped out from inside the ring (approximately 30 g of wet sand was collected from each spot). In each of the four locations (A, B, C, and D, Supporting Information II), three spots were chosen along the width of the flume from which the biofilm was collected and mixed to form a composite sample. After the surface sample was collected, the sand below was removed to a depth of 3 cm, and the same sampling procedure that was described above was repeated. Afterwards, the sand was removed again, and another sample was taken from a depth of 7 cm. After every sample collection, the rings were cleaned and sterilized by dipping them in a 75% alcohol solution. The biofilm samples were analyzed for biofilm thickness, extracellular polymeric substance (EPS) concentration and microbial abundance. Biofilm thickness was measured with confocal laser scanning microscopy (CLSM) by staining with the LIVE/DEADV Bacterial Viability Kit (BacLightTM Molecular Probes, Eugene, OR). Sand samples (0.5 g) were placed R

inside Eppendorf tubes. Water was replaced with phosphatebuffered saline (PBS) solution (10 mM) by using a pipette. A mixture of two nucleic acid-binding stains was prepared by mixing 1.5 lL of SYTO9 (5 lM) and 1.5 lL of propidium iodide (30 lM) in 1 mL of PBS. Then, 100 lL of the mixture was added to each Eppendorf tube and covered with aluminum foil. After 15 min, the staining solution was replaced with a PBS solution. Blanks were prepared from a sample from the flume without staining to check the autofluorescence of the biofilm, from sterile clean sand to evaluate the autofluorescence of the sand itself, and from sterile clean sand with the nucleic acid mixture to evaluate the adsorption of the stain to the sand. A 510 Meta CLSM (Zeiss-Meta, € ttingen, Germany) was used to visualize the biofilm using Go a dry objective (LCI Plan NeoFluor) at 20X magnification and a numerical aperture of 0.5. Biofilm thickness was measured by scanning 65 grains at each depth, and measuring the distance from the bottom of the biofilm to the top at scanning steps of 3.5lm. For EPS measurements, 30 g of wet sand was placed in 50mL sterile tubes. The sand was mixed with 22 mL of NaCl (10 mM) and 18 lL of formaldehyde (36%), and incubated with gentle shaking at 48C for 1 h. Afterwards, 0.4 mL of NaOH (1 M) was added to each tube and incubated with gentle shaking for an additional three hours at 48C. The samples were centrifuged at 13,000 g for 30 min at 48C, and the supernatant was then filtered through a 0.22-lm filter, and dialyzed against double-distilled water until the electrical conductivity (EC) reached 2.5 lS/cm (a total of 48 h). The concentration of total organic carbon (TOC) in the water was measured using a TOC analyzer (Analytik Jena Multi N/C 2100S, Germany) following the method of Liu and Fang (2002), and the concentrations of EPS were expressed as TOC. Microbial abundances, expressed as 16S gene copy numbers, were quantified by extracting DNA from 0.5 g sand samples using the Powermax Soil DNA extraction kit according to manufacturer’s instructions (MoBio, Carlsbad, CA, U.S.A.). Quantitative PCR (qPCR) amplification reactions were carried out in a C1000 thermocycler (Biorad, Hercules, CA, U.S.A.). The reaction mix for qPCR included the following: 12.5 lL of KAPA SYBR Fast Universal ReadyMix (KAPA Biosystems, Woburn, MA, U.S.A.); 100 nmol of each primer (341F: 50 -GCCTACGGGAGGCAGCAG 230 ; 518R: 50 ATTACCGCGGCTGCTGG-30 ), 1 lL of template (extracted DNA or standard plasmid), and DDW to constitute 25 lL. The reaction profile included an initial denaturation step at 958C, followed by 40 cycles of 958C for 30 s; 608C for 30 s; and 728C for 30 s. Acquisition was performed at the completion of each cycle, following a short (2 s) step at 788C to ensure primer dimer denaturation. The melting curve (72– 958C) showed only one peak for all qPCR reactions. A calibration curve was created by conducting a 10-fold dilution series ( 1032109 copies) of plasmids containing environmental copies of the 16S gene fragment. The calibration

1843

De Falco et al.


Table 1. A summary of the schedule and procedures that were used for conducting the experiments in the flume. The data in the table describes the experiments with the surficial biofilm. Items 2–5 in the table were repeated for the homogeneously distributed biofilm immediately after the surficial biofilm analysis was completed.

Activity 1 Growing phase

Duration 3 months

Procedure

Flow conditions

Feeding biofilm with C: N: P on a

Overlying water velocity,

2 weeks

Not relevant

V 5 12 cm s21

daily basis 2 Biodegradation of DOCL under differ-

Statistical methods for analysis

12–16 h before the experiment, water in the flume was replaced

ent overlying water

with freshwater and the over

velocities

lying water velocity was imposed. Filters were replaced

Simple regression and The overlying water velocities Wilcoxon test were imposed in the following order: 4, 6, 8, 10, and 12 cm s21

just before spiking. 3 Biodegradation of DOCL under differ-

4 weeks

12–16 h before the experiment, water in the flume was replaced

ent losing and

with freshwater and the overlying

gaining fluxes, and overlying water

water velocity and losing/gaining flux were imposed. Filters were

velocities

replaced just before spiking.

Every week, 6 spiking events were conducted in the follow

4-way ANOVA

ing order: Neutral-Losing-Gaining-LosingGaining-Neutral. (qL5 losing flux, qG5 gaining flux). 1st week: V 5 12 cm s21, qL 2 qG 50 cm d21 2nd week: V 5 12 cm s21, qL 2 qG 20 cm d21 3 week: V 5 4 cm s21, qL 2 qG rd

20 cm d21 4 4 Oxygen profiling

4 weeks (Same time as the experiments described in item 3 in

Profiling in the middle of the stoss

th

week: V 5 4 cm s21, qL 2 qG 50 cm d21

Same conditions as above

side one hour after spiking.

No statistical method was applied

this Table) 5 Biofilm sampling

One day of sampling fol- Sampling along three bed forms in lowed by 3 weeks of the central part of the flume analyses

channel. See Supporting Information II for more details on the sampling.

Analysis included: Friedman test - Biofilm thickness (confocal laser scanning microscopy) - EPS concentration (EPS extraction) - 16S gene copy (DNA extraction)

curve had R2 5 0.98, and the slope was between 23.0 and 23.9, corresponding to a PCR efficacy of 90–111%. Data analysis Linear regressions of the reduction in DOCL concentration over time in the flume was used to evaluate the rates of biodegradation. Since replacement water was added to the flume to preserve the overall water volume under losing and gaining fluxes, the rates of DOC decline were first corrected to remove the influence of the water replacement (i.e., dilution) and, afterwards, were converted to mass loss over time. Then, the values were divided by the bed area to yield a biodegradation rate per unit bed area, reported in the units of

lmol m22 min21.To evaluate whether the rates of DOCL biodegradation or biofilm distributions were significantly different, we used various types of statistical tests. We used parametric statistics wherever the data met the basic assumptions, and nonparametric methods in all other cases. Specifically, an analysis of variance (four-way ANOVA) was used to evaluate the interactions between different overlying water velocities, subsurface fluxes, and biofilm type after verifying that the data were normally distributed. A non-parametric analysis (Wilcoxon test) was performed to test the differences between the DOCL biodegradation rates according to the different type of biofilms under different overlying velocities. The Friedman test was used to evaluate the difference

1844

De Falco et al.


Fig. 2. Vertical distributions of biofilm thickness (A), 16S gene copy number (B) and EPS concentration (C) in the sand bed. Errors bars indicate the standard deviation of the measurements (n 5 65 for biofilm thickness, n 5 18 for the 16S gene copy number, and n 5 3 for EPS concentrations). between biofilm thickness, microbial abundance and EPS concentration at the different depths. All statistical tests were conducted using SPSS Statistics (SPSS, IBM Corp, Version 22.0. Armonk, NY, U.S.A.). The Tukey’s HSD test for the ANOVA post-hoc analysis was performed with JMP (JMPV, SAS Institute Inc., Cary, NC, U.S.A.). A summary of the schedule, order of experiments, procedures, and statistical tests that were used for conducting and analyzing the experiments in the flume is presented in Table 1. R

Results Biofilm analyses: vertical distribution of biomass in the streambed Biofilm thickness, microbial abundance and EPS concentrations were the highest at the top of the bed and became significantly smaller with depth for the surficial biofilm (Friedman ANOVA, df 5 2, p < 0.05) (Fig. 2). In contrast, biofilm thickness, microbial abundance and EPS concentrations were uniform (no significant difference) with depth for the homogeneously distributed biofilm (Friedman ANOVA, df 5 2, p > 0.05). Biofilm thickness, microbial abundance and EPS concentrations near the surface of the bed were significantly higher for the surficial biofilm than for the homogeneously distributed biofilm (Friedman ANOVA, df 5 1, p < 0.05). Biodegradation potential of DOCL Batch experiments for evaluating the biodegradation potential of DOCL were done in Erlenmeyer flasks. Biodegradation was significantly higher in the flume sand than in all other treatments (F4,8 5 132.63, p < 0.05, Fig. 3). Only a slight reduction in the DOCL concentration was observed in the flume water, while no significant change in the DOCL concentrations was observed for the sterilized water and sand (control treatments).

Biodegradation of DOCL under different overlying water velocities Rates of DOCL biodegradation were evaluated by monitoring the concentration in the bulk water under different overlying water velocities (Fig. 4). The DOCL biodegradation rate increased linearly with increasing overlying water velocity regardless of the biofilm type (R2 5 0.90 for all the data). Biofilm type did not have a significant effect on the absolute rates of biodegradation (n 5 5, p < 0.05). Biodegradation of DOCL under losing and gaining fluxes Pooling all the data together reveals that the biodegradation rates of DOCL under neutral conditions were significantly lower than under gaining conditions (Tukey HSD, p < 0.05, Fig. 5), but not significantly different than the rates under losing conditions (ANOVA, df 5 1, p > 0.05, Fig. 5). This is mainly caused by the results under the slower overlying water velocity. In addition, the DOCL concentrations in the water that was pumped out of the flume under losing conditions revealed that less than 1% of the DOCL leached out of the flume (under both 20- or 50-cm d21 losing flux). Overall, the biodegradation rate of DOCL by surficial biofilm was higher than the rate by the homogeneously distributed biofilm (ANOVA, df 5 1, p < 0.05, Fig. 5). No increase in the optical density (measured at 600 nm), used to detect microbial growth, was observed when the UV disinfection system was operating during the experiments. Supporting Information III includes a table summarizing the results from all the statistical tests that were used to analyze the data shown in Figs. 2-5. Oxygen profiles The oxygen profiles during the DOCL biodegradation experiments are displayed in Fig. 6. The oxygen concentrations in the bulk water and in the sand bed were at 100% saturation before spiking the flume with DOCL (data not

1845

De Falco et al.


Fig. 3. Potential DOCL biodegradation in water and sand from the flume as measured by incubation experiments. Errors bars indicate the standard deviations of triplicates (sterile flume sand and flume sand) and duplicates (deionized sterile water and flume water).

Fig. 5. The effect of losing or gaining fluxes (designated here as bot-

Fig. 4. DOCL biodegradation by surficial and homogeneously distrib-

tom fluxes, qbot) on the DOCL biodegradation rate by surficial and homogeneously distributed biofilms at overlying water velocities of 12 cm s21 (A), and of 4 cm s21 (B). Errors bars indicate the standard deviations of two independent experiments conducted during the same week.

uted biofilms under different overlying water velocities.

shown). About 90 min after spiking with DOCL, the oxygen concentration in the bulk water declined to approximately 50% and 70% saturation for water velocities of 4 cm s21 and 12 cm s21, respectively. Consecutive profiling revealed that this was also the moment after which no further temporal differences were observed for the profiles. Under losing conditions, oxygen concentrations decreased in the sand bed and reached a value of 0–10% saturation within a depth of 5 mm. In contrast, under gaining conditions, the oxygen concentrations slightly declined near the water-sand interface (< 1 mm) but quickly recovered and increased within

the bed to levels between 65% and 90% for water velocities of 4 cm s21 and 12 cm s21, respectively. The shapes of the oxygen profiles and the depth of activity under both gaining and losing fluxes were relatively similar for the two types of biofilms under an overlying water velocity of 12 cm s21. However, differences in the profiles were observed for the two types of biofilm at 4 cm s21. Under losing conditions, the oxygen concentrations decreased within the upper first millimeter of the surficial biofilm, while for the homogeneously distributed biofilm, the decrease occurred over 4 millimeters. Under gaining conditions, oxygen declined in the homogeneously distributed biofilm to 60%, while in the

1846

De Falco et al.


Fig. 6. Oxygen profiles in the upstream section of the bed form under different overlying velocities of 12 cm s21 (A), and of 4 cm s21 (B), under losing (50 cm d21) and gaining (50 cm d21) fluxes and different biofilm types. The dashed horizontal line represents the interface between the water and the sand. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

surficial biofilm, it only slightly decreased, and at greater depths, it increased to 90%.

Discussion Biodegradation of DOCL This study investigated the effects of overlying water velocity, losing and gaining fluxes, and biofilm distribution on the biodegradation of DOCL in a sandy streambed. Biodegradation mostly occurred in the aerobic fraction of the upper millimeters at the interface between the water and the sediment (Fig. 6). Comparing these observations to the notion presented in Sondergaard and Middelboe (1995), which discusses DOCL as the fraction that can be decomposed in less than two weeks or as the fraction that is consumed within surface biofilms, clearly indicates that the sodium benzoate can be considered as DOCL in our study. The reduction in DOCL concentrations was interpreted as biodegradation following the results shown in Fig. 3, where no change in DOCL concentrations was observed in either sterile water or clean sand as compared to the flume sand. There are two pathways available upon the reduction in DOCL, specifically either mineralization with the release of

CO2 and H2O, or microbial assimilation (i.e., incorporation into the biomass). Separation between mineralization and the release of CO2, and microbial assimilation is usually performed by employing incubation experiments (Berggren et al. 2010; Guillemette and del Giorgio 2011; Fasching et al. 2014). Fasching et al. (2014), for example, found that terrigenous dissolved organic matter is respired by microorganisms rather than incorporated into their biomass. Conversely, Berggren et al. (2010) found that DOCL (related to lowweight molecular compounds) was preferentially assimilated into biomass. Because we used here a single source of DOC, it is assumed that its use was divided between assimilation and energy production but at unknown ratios. The upper section of the biofilm near the water-sediment interface served as the hotspot for aerobic DOCL metabolism. Thus, the biodegradation rates of DOCL were strongly influenced by mass transfer processes at the interface between the water and the streambed, which were strongly linked to the overlying water velocity but less influenced by the losing and gaining fluxes. Whereas several studies have previously measured the effect of overlying water velocity on DOC uptake (e.g., Battin et al. 2003), this is, to our knowledge, the first study that explicitly established the link between

1847

De Falco et al.


overlying water velocity and gaining and losing flow conditions on DOCL biodegradation by stream biofilms. It was clearly shown that overlying water velocity was the most significant factor influencing the DOCL biodegradation rate (Figs. 4, 5), confirming the underlying hypothesis of this research. As in previous studies on nutrient dynamics, the DOCL biodegradation rate increased linearly with the increase in overlying velocity, despite the fact that hyporheic exchange increases with the square power of velocity (Packman et al. 2004; Arnon et al. 2013). This is probably due to the fact that the increase in surface water velocity also leads to higher seepage velocities and lower contact times between DOCL and the biofilm, thus partially compensating for the higher supply rate of DOCL due to hyporheic exchange. For example, using the same flume, sand and streambed morphology, Fox et al. (2014) found that increasing the overlying velocity from 4.3 cm s21 to 12.3 cm s21 resulted in an increase in the hyporheic exchange flux (water flux per streambed area) by one order of magnitude (from 1.83 cm d21 to 14.9 cm d21). It is also possible that the increased supply of DOCL led to a reduction of streambed permeability due to the biofilm growth and, consequently, to a less marked increase in the biodegradation rates. The losing and gaining fluxes had only a slight effect on DOCL biodegradation (Fig. 5), which was more pronounced under the slower overlying water velocity (4 cm s21). This is due to the relative importance of the losing/gaining fluxes vs. the hyporheic flux. Under a constant losing/gaining flux, the hyporheic flux becomes more important as the overlying water velocity increases (Fox et al. 2014). When the overlying water velocity increases, the hyporheic exchange becomes dominant up to a level where the losing/gaining flux does not practically affect solute exchange. This threshold velocity depends on the streambed conductivity and morphology. While the results suggest that overlying water velocity is the most important flow characterization for DOCL biodegradation in a sandy streambed, it is expected that for more coarse beds (e.g., gravel beds), losing and gaining fluxes may become a more important factor since the overall flux through the bed can be significantly higher than in sandy streambeds. Integrative information gained from the entire set of experiments revealed that the gaining conditions led to a slight increase in biodegradation rate. The gaining water in our experiments was saturated with oxygen (100%), and the resulting supply of oxygen to the streambed led to an increase in the size of the aerobic zone (Fig. 6) and accelerated the DOCL biodegradation rate. The experimental constraint that have forced us to either work with saturated (or strict anaerobic upwelling water) may led to enhanced respiration. Conversely, it facilitated the comparison between the different flow conditions and the evaluation of the governing mechanisms. In natural systems, the contribution of oxygen from groundwater systems to streams can vary over a

wide range of concentrations (e.g., Malard and Hervant 1999; Krause et al. 2013). The aforementioned studies suggest that concentrations around 50% saturation are more likely to occur in natural systems (as compared to the concentrations used in the experiments). In cases of upwelling groundwater with very low oxygen concentrations, enhanced DOCL biodegradation is not expected. Overall, aerobic respiration will be affected by the stoichiometry of carbon and oxygen, their mass fluxes and their physical mixing, and this type of biophysicochemical process coupling could only be fully examined by complex modeling scenarios (Trauth et al. 2014). Effect of biofilm distribution on biodegradation of DOCL DOCL biodegradation under neutral conditions was not statistically different between the different biofilms (Fig. 4). However, when all the rates under losing and gaining flow conditions were evaluated together, the surficial biofilm was characterized by slightly faster biodegradation rates. These results are partly in agreement with our hypothesis that the biofilm distribution would have an effect on the DOCL biodegradation. This implies that the interactions between the biofilm distribution and the flow conditions must be considered. The DOCL utilized in this study was consumed in the first upper millimeters of the bed (Fig. 6), thus giving the surficial biofilm an advantage due to its higher biomass near the surface (Fig. 2). In natural environments, biofilms tend to colonize different sediment depths according to their functional needs, as shown in previous works on the functional layering of microbial communities in the HZ (Hendricks 1996; Franken et al. 2001; Fischer et al. 2003). Structural layering of sediments and biomass can be smeared by bed form migration, which leads to mixing of the streambed and homogenization of the biomass. While our experiments were done only under stationary bed forms, we tried to account for the effect of moving bed forms through the manipulation of biomass distribution by manually mixing the streambed (homogeneously distributed biofilm, Fig. 2). The net exchange of solute between the stream and the HZ due to slow moving bed forms was found to be insignificant (Packman and Brooks 2001). However, the increase of bed form migration rates time scale of solute exchange without moving bed forms can significantly affect solute exchange, as was shown for oxygen distribution by Precht et al. (2004). Conversely, it was demonstrated by a modeling study that moving bed forms have a negligible effect on denitrification and on coupled nitrificationdenitrification (Kessler et al. 2015). DOC uptake in the hyporheic zone The experimental results from this study implicate the coupling between flow conditions, biomass distribution and DOC uptake in the hyporheic zone. Two major observations that are commonly seen in natural systems were clearly replicated in the experiments: the higher density of benthic

1848

De Falco et al.


Fig. 7. Schematic illustration of two different hypothetical cases showing flow patterns under neutral conditions in the hyporheic zone and the spatial distribution of DOC concentration portraying continuum reactivity (A) and DOCL (B). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

biofilm close to the interface with the water and the strong uptake of DOCL in the upper few millimeters of the streambed (e.g., Fischer et al. 2003; Revsbech et al. 2005). Despite the ubiquity of the aforementioned features, modeling of nutrient cycling in the hyporheic zone and in streams has rarely considered these structural aspects. To start with, most modeling studies commonly represent DOC uptake rates with a single reaction rate coefficient (e.g., Bardini et al. 2012) due to the lack of information on bacterial growth dynamics. In addition, biomass distribution is also not commonly included in models predicting nutrient dynamics in streams. Although including detailed information may lead to extremely high computational resources, which are not always feasible, it is argued that additional complexity should be introduced into existing models, as was done through variable reaction rates in different sections of the biofilm for denitrification (Kessler et al. 2012). The results of DOCL biodegradation in the HZ can follow several patterns, which can be reflected in the spatial distribution of microbial activity in sandy streambeds. Two disparate hypothetical cases are shown in Fig. 7 by illustrating the continuum reactivity of DOC (A) and DOCL (B) distributions and flow patterns under neutral conditions in the HZ of sandy streambeds. Hyporheic flow paths lead to increases in nutrient fluxes from the stream (DOC, N, O2, etc.) at the central section of the stoss side of the bed form, and the DOC concentration gradient follows the shape of a curved front (“half ball shape”) toward the low pressure regions at the lower parts of the bed form (Fig. 7A). In such a case, oxygen concentrations are expected to follow the same pattern as DOC (see, for example, Bardini et al. 2012; Kessler et al. 2012). In cases where DOCL is abundant and a thick biofilm develops at the surface, strong consumption will occur in

the upper few millimeters of the stream bed (e.g., Altmann et al. 2004; Revsbech et al. 2005, Fig. 6, 7B). In this case, the shape of the hyporheic flow paths will have little effect on DOCL consumption, which will occur along the entire bed form-water interface, and the interface between the aerobicanaerobic zone will have a different shape than in the complex DOC case (Fig. 7A). The pattern shown in A is commonly reported in modeling studies (such as those mentioned above); however, observations from other laboratories and field studies suggest that the pattern shown in B is also commonly found in aquatic systems such as lowland streams (Altmann et al. 2004; Revsbech et al. 2005). Different flow conditions and streambed heterogeneity are also expected to change the distributions of DOC concentrations that are schematically shown in Fig. 7. For example, losing or gaining fluxes are anticipated to compact the hyporheic flow zone, and to shrink accordingly the size of the zone in which DOC concentrations change in the HZ (Fig. 7A). Streambed heterogeneity, on the other hand, may also change the shape and flow patterns, as shown by Salehin et al. (2004). It is not expected, however, that the structure shown in Fig. 7B will significantly change except for a minor swelling or shrinking of the aerobic zone. The current level of understanding of the physical processes in the hyporheic zone allows us to start incorporating a simplified version of biomass distribution and variable reaction rates in homogeneous streambeds. This could be done initially by adopting spatially variable reaction rates to represent different microbial biomass abundances or community structures. Such work assists in comprehending the governing processes and will enhance future studies aimed at understanding the complex interactions between flow conditions and biogeochemical processes in streams.

1849

De Falco et al.


References Aitkenhead-Peterson, J. A., W. H. McDowell, and J. C. Neff. 2002. Sources, production, and regulation of allochthonous dissolved organic matter inputs to surface waters, p. 25–70. In S. E. G. Findlay and R. L. Sinsabaugh [eds.], Aquatic ecosystems: Interactivity of dissolved organic matter. Academic Press. Altmann, D., P. Stief, R. Amann, and D. De Beer. 2004. Distribution and activity of nitrifying bacteria in natural stream sediment versus laboratory sediment microcosms. Aquat. Microbiol. Ecol. 36: 73–81. doi:10.3354/ ame036073 Arnon, S., K. Yanuka, and A. Nejidat. 2013. Impact of overlying water velocity on ammonium uptake by benthic biofilms. Hydrol. Process. 27: 570–578. doi:10.1002/hyp.9239 Arrieta, J. M., E. Mayol, R. L. Hansman, G. J. Herndl, T. Dittmar, and C. M. Duarte. 2015. Dilution limits dissolved organic carbon utilization in the deep ocean. Science 348: 331–333. doi:10.1126/science.1258955 € sel. 2010. Flow velocity and priAugspurger, C., and K. Ku mary production influences carbon utilization in nascent epilithic stream biofilms. Aquat. Sci. 72: 237–243. doi: 10.1007/s00027-009-0126-y Baker, M. A., C. N. Dahm, and H. M. Valett. 1999. Acetate retention and metabolism in the hyporheic zone of a mountain stream. Limnol. Oceanogr. 44: 1530–1539. doi: 10.4319/lo.1999.44.6.1530 Baker, M. A., H. M. Valett, and C. N. Dahm. 2000. Organic carbon supply and metabolism in a shallow groundwater ecosystem. Ecology 81: 3133–3148. doi:10.1890/00129658(2000)081[3133:OCSAMI]2.0.CO;2 Bardini, L., F. Boano, M. B. Cardenas, R. Revelli, and L. Ridolfi. 2012. Nutrient cycling in bedform induced hyporheic zones. Geochim. Cosmochim. Acta 84: 47–61. doi: 10.1016/j.gca.2012.01.025 Battin, T. J., L. A. Kaplan, J. D. Newbold, C. M. E. Hansen, and J. Denis Newbold. 2003. Contributions of microbial biofilms to ecosystem processes in stream mesocosms. Nature 426: 439–442. doi:10.1038/nature02152 Battin, T. J., and others. 2008. Biophysical controls on organic carbon fluxes in fluvial networks. Nat. Geosci. 1: 95–100. doi:10.1038/ngeo602 Battin, T. J., S. Luyssaert, L. A. Kaplan, A. K. Aufdenkampe, A. Richter, and L. J. Tranvik. 2009. The boundless carbon cycle. Nat. Geosci. 2: 598–600. doi:10.1038/ngeo618 Benstead, J. P., and D. S. Leigh. 2012. An expanded role for river networks. Nat. Geosci. 5: 678–679. doi:10.1038/ ngeo1593 € m, and M. Jansson. Berggren, M., H. Laudon, M. Haei, L. Stro 2010. Efficient aquatic bacterial metabolism of dissolved low-molecular-weight compounds from terrestrial sources. ISME J. 4: 408–416. doi:10.1038/ismej.2009.120

Boano, F., J. W. Harvey, A. Marion, A. I. Packman, R. Revelli, € rman. 2014. Hyporheic flow and L. Ridolfi, and A. Wo transport processes: Mechanisms, models, and biogeochemical implications. Rev. Geophys. 52: 603–679. doi: 10.1002/2012RG000417 Bott, T. L., and L. A. Kaplan. 1985. Bacterial biomass, metabolic state, and activity in stream sediments: Relation to environmental variables and multiple assay comparisons. Appl. Environ. Microbiol. 50: 508–522. Boulton, A. J., S. Findlay, P. Marmonier, E. H. Stanley, and H. M. Valett. 1998. The functional significance of the hyporheic zone in streams and rivers. Annu. Rev. Ecol. Syst. 29: 59–81. doi:10.1146/annurev.ecolsys.29.1.59 Briggs, M. A., F. D. Day-lewis, J. P. Zarnetske, and J. W. Harvey. 2015. A physical explanation for the development of redox microzones in hyporheic flow. Geophys. Res. Lett. 42: 1–9. doi:10.1002/2015GL064200 Brunke, M., and T. Gonser. 1997. The ecological significance of exchange processes between rivers and groundwater. Freshwater Biol. 37: 1–33. doi:10.1046/j.1365-2427.1997. 00143.x Buffington, J. M., and D. Tonina. 2009. Hyporheic exchange in mountain rivers II: Effects of channel morphology on mechanics, scales, and rates of exchange. Geogr. Compass 3: 1038–1062. doi:10.1111/j.1749-8198.2009.00225.x Butturini, A., T. J. Battin, and F. Sabater. 2000. Nitrification in stream sediment biofilms: The role of ammonium concentration and DOC quality. Water Res. 34: 629–639. doi: 10.1016/S0043-1354(99)00171-2 Byrne, P., and others. 2014. Control of river stage on the reactive chemistry of the hyporheic zone. Hydrol. Process. 28: 4766–4779. doi:10.1002/hyp.9981 Cardenas, M. B. (2009), Stream-aquifer interactions and hyporheic exchange in gaining and losing sinuous streams. Water Resour. Res. 45, W06429, doi:W0642910.1029/ 2008wr007651. Cardenas, M. B., and J. L. Wilson. 2007. Exchange across a sediment–water interface with ambient groundwater discharge. J. Hydrol. 346: 69–80. doi:10.1016/j.jhydrol.2007. 08.019 € and C. P. Leslie Grady. 2001. Aerobic and anoxic C¸inar, O., biodegradation of benzoate: Stability of biodegradative capability under endogenous conditions. Water Res. 35: 1015–1021. doi:10.1016/S0043-1354(00)00340-7 Creed, I. F., and others. 2015. The river as a chemostat: Fresh perspectives on dissolved organic matter flowing down the river continuum. Can. J. Fish. Aquat. Sci. 1285: 127221285. Dent, C. L., N. B. Grimm, E. Martı, J. W. Edmonds, J. C. Henry, and J. R. Welter. 2007. Variability in surfacesubsurface hydrologic interactions and implications for nutrient retention in an arid-land stream. J. Geophys. Res.: Biogeosci. 112: 2005–2012. doi:10.1029/2007JG000467

1850

De Falco et al.


Fasching, C., B. Behounek, G. A. Singer, and T. J. Battin. 2014. Microbial degradation of terrigenous dissolved organic matter and potential consequences for carbon cycling in brown-water streams. Sci. Rep. 4: 4981. doi: 10.1038/srep04981 Findlay, S. 1995. Importance of surface-subsurface exchange in stream ecosystems: The hyporheic zone. Limnol. Oceanogr. 40: 159–164. Fischer, H., A. Sukhodolov, S. Wilczek, and C. Engelhardt. 2003. Effects of flow dynamics and sediment movement on microbial activity in a lowland River. River Res. Appl. 19: 473–482. doi:10.1002/rra.731 Fox, A., F. Boano, and S. Arnon. 2014. Impact of losing and gaining streamflow conditions on hyporheic exchange fluxes induced by dune-shaped bed forms. Water Resour. Res. 50: 1895–1907. doi:10.1002/2013WR014668 Franken, R. J. M., R. G. Storey, and D. D. Williams. 2001. Biological, chemical and physical characteristics of downwelling and upwelling zones in the hyporheic zone of a north-temperature stream. Hydrobiologia 444: 183– 195. doi:10.1023/A:1017598005228 Gomez, J. D., J. L. Wilson, and M. B. Cardenas (2012), Residence time distributions in sinuosity-driven hyporheic zones and their biogeochemical effects. Water Resour. Res. 48, W09533, doi:10.1029/2012WR012180 Guillemette, F., and P. A. del Giorgio. 2011. Reconstructing the various facets of dissolved organic carbon bioavailability in freshwater ecosystems. Limnol. Oceanogr. 56: 734– 748. doi:10.4319/lo.2011.56.2.0734 € hlke, M. A. Voytek, D. Scott, and C. R. Harvey, J. W., J. K. Bo Tobias. 2013. Hyporheic zone denitrification: Controls on effective reaction depth and contribution to whole-stream mass balance. Water Resour. Res. 49: 6298–6316. doi: 10.1002/wrcr.20492 Helton, A. M., M. S. Wright, E. S. Bernhardt, G. C. Poole, R. M. Cory, and J. A. Stanford. 2015. Dissolved organic carbon lability increases with water residence time in the alluvial aquifer of a river floodplain ecosystem. J. Geophys. Res.: Biogeosci. 120: 1–14. doi:10.1002/ 2014JG002832 Hendricks, S. P. 1996. Bacterial biomass, activity, and production within the hyporheic zone of a north-temperate stream. Arch. Hydrobiol. 136: 467–487. Kaplan, L. A., T. N. Wiegner, J. D. Newbold, P. H. Ostrom, and H. Gandhi. 2008. Untangling the complex issue of dissolved organic carbon uptake: A stable isotope approach. Freshw. Biol. 53: 855–864. doi:10.1111/j.13652427.2007.01941.x Kessler, A. J., R. N. Glud, M. Bayani Cardenas, M. Larsen, M. F. Bourke, and P. L. M. Cook. 2012. Quantifying denitrification in rippled permeable sands through combined flume experiments and modeling. Limnol. Oceanogr. 57: 1217–1232. doi:10.4319/lo.2012.57.4.1217

Kessler, A. J., M. B. Cardenas, and P. L. M. Cook. 2015. The negligible effect of bed form migration on denitrification in hyporheic zones of permeable sediments. J. Geophys. Res.: Biogeosci. 120: 1–11. doi:10.1002/2014JG002852. Received Koehler, B., E. Von Wachenfeldt, D. Kothawala, and L. J. Tranvik. 2012. Reactivity continuum of dissolved organic carbon decomposition in lake water. J. Geophys. Res.: Biogeosci. 117: 1–14. doi:10.1029/2011JG001793 Krause, S., and others. 2011. Inter-disciplinary perspectives on processes in the hyporheic zone. Ecohydrology 4: 481– 499. doi:10.1002/eco.176 Krause, S., C. Tecklenburg, M. Munz, and E. Naden. 2013. Streambed nitrogen cycling beyond the hyporheic zone: Flow controls on horizontal patterns and depth distribution of nitrate and dissolved oxygen in the upwelling groundwater of a lowland river. J. Geophys. Res.: Biogeosci. 118: 54–67. doi:10.1029/2012JG002122 Leu, H.-G. G., C.-D. D. Lee, C. F. Ouyang, and H.-T. T. Tseng. 1998. Effects of organic matter on the conversion rates of nitrogenous compounds in a channel reactor under various flow conditions. Water Res. 32: 891–899. doi:10.1016/S0043-1354(97)00224-8 € tzmann, and J. H. Lewandowski, J., L. Angermann, G. Nu Fleckenstein. 2011. A heat pulse technique for the determination of small-scale flow directions and flow velocities in the streambed of sand-bed streams. Hydrol. Process. 25: 3244–3255. doi:10.1002/hyp.8062 Li, L., I. Yolcubal, S. Sandrin, M. Q. Hu, and M. L. Brusseau. 2001. Biodegradation during contaminant transport in porous media: 3. Apparent condition-dependency of growth-related coefficients. J. Contam. Hydrol. 50: 209– 223. doi:10.1016/S0169-7722(01)00111-5 Liu, H., and H. H. P. Fang. 2002. Extraction of extracellular polymeric substances (EPS) of sludges. J. Biotechnol. 95: 249–256. doi:10.1016/S0168-1656(02)00025-1 Malard, F., and F. Hervant. 1999. Oxygen supply and adaptation of animals in groundwater. Freshw. Biol. 41: 1–30. doi:10.1046/j.1365-2427.1999.00379.x McClain, M. E., and others. 2003. Biogeochemical hot spots and hot moments at the interface of terrestrial and aquatic ecosystems. Ecosystems 6: 301–312. doi:10.1007/ s10021-003-0161-9 Miller, M. P., D. M. McKnight, R. M. Cory, M. W. Williams, and R. L. Runkel. 2006. Hyporheic exchange and fulvic acid redox reactions in an alpine stream/wetland ecosystem, Colorado front range. Environ. Sci. Technol. 40: 5943–5949. doi:10.1021/es060635j Nogaro, G., T. Datry, F. Mermillod-Blondin, A. Foulquier, and B. Montuelle. 2013. Influence of hyporheic zone characteristics on the structure and activity of microbial assemblages. Freshw. Biol. 58: 2567–2583. doi:10.1111/ fwb.12233

1851

De Falco et al.


Packman, A. I., and N. H. Brooks. 2001. Hyporheic exchange of solutes and colloids with moving bed forms. Water Resour. Res. 37: 2591–2605. doi:10.1029/2001WR000477 Packman, A., M. Salehin, and M. Zaramella. 2004. Hyporheic exchange with gravel beds: Basic hydrodynamic interactions and bedform-induced advective flows. J. Hydraul. Eng. 130: 647–656. doi:10.1061/(ASCE)0733-9429(2004)130:7(647) Porcal, P., J.-F. F. Koprivnjak, L. A. Molot, and P. J. Dillon. 2009. Humic substances-part 7: The biogeochemistry of dissolved organic carbon and its interactions with climate change. Environ. Sci. Technol. 16: 714–726. doi:10.1007/ s11356-009-0176-7 Precht, E., U. Franke, L. Polerecky, and M. Huettel. 2004. Oxygen dynamics in permeable sediments with wavedriven pore water exchange. Limnol. Oceanogr. 49: 693– 705. doi:10.4319/lo.2004.49.3.0693 Regnier, P., and others. 2013. Anthropogenic perturbation of the carbon fluxes from land to ocean. Nat. Geosci. 6: 597–607. doi:10.1038/ngeo1830 Revsbech, N. P., J. P. Jacobsen, and L. P. Nielsen. 2005. Nitrogen transformations in microenvironments of river beds and riparian zones. Ecol. Eng. 24: 447–455. doi: 10.1016/j.ecoleng.2005.02.002 Salehin, M., A. I. Packman, M. Paradis, M. Salehin, A. I. Packman, and M. Paradis. 2004. Hyporheic exchange with heterogeneous streambeds: Laboratory experiments and modeling. Water Resour. Res. 40: W11504. doi:10.1029/ 2003WR002567 Singer, G., K. Besemer, G. Hochedlinger, A. K. Chlup, and T. J. Battin. 2011. Monomeric carbohydrate uptake and structure-function coupling in stream biofilms. Aquat. Microbiol. Ecol. 62: 71–83. doi:10.3354/ame01454 Sobczak, W. V, and S. Findlay. 2002. Variation in bioavailabtlity of dissolved organic carbon among stream hyporheic flowpaths. Ecology 83: 3194–3209. doi:10.1890/ 0012-9658(2002)083[3194:VIBODO]2.0.CO;2 Sondergaard, M., and M. Middelboe. 1995. A cross-system analysis of labile dissolved organic-carbon. Mar. Ecol. Prog. Ser. 118: 283–294. Stofleth, J. M., F. Douglas Shields, and G. A. Fox. 2007. Hyporheic and total transient storage in small, sand-bed streams. Hydrol. Process. 22: 1885–1894. doi:10.1002/ hyp.6773 Thomas, F. I. M., C. D. Cornelisen, and J. M. Zande. 2000. Effects of water velocity and canopy morphology on ammonium uptake by seagrass communities. Ecology 81:

2704–2713. doi:10.1890/0012-9658(2000)081[2704:EOWVAC]2.0.CO;2 Trauth, N., C. Schmidt, U. Maier, M. Vieweg, and J. H. Fleckenstein. 2013. Coupled 3D stream flow and hyporheic flow model under varying stream and ambient groundwater flow conditions in a pool-riffle system. Water Resour. Res. 49: 5834–5850. doi:10.1002/ wrcr.20442 Trauth, N., C. Schmidt, M. Vieweg, U. Maier, and J. H. Fleckenstein. 2014. Hyporheic transport and biogeochemical reactions in pool-riffle systems under varying ambient groundwater flow conditions. J. Geophys. Res.: Biogeosci. 119: 910–928. doi:10.1002/2013JG002586 Vannote, R. L., G. W. Minshall, K. W. Cummins, J. R. Sedell, and C. E. Cushing. 1980. River continuum concept. Can. J. Fish. Aquat. Sci. 37: 130–137. doi:10.1139/f80-017 Wagner, K., M. M. Bengtsson, K. Besemer, A. Sieczko, N. R. Burns, E. R. Herberg, and T. J. Battin. 2014. Functional and structural responses of hyporheic biofilms to varying sources of dissolved organic matter. Appl. Environ. Microbiol. 80: 6004–6012. doi:10.1128/AEM.01128-14 Zarnetske, J. P., R. Haggerty, S. M. Wondzell, and M. A. Baker (2011), Labile dissolved organic carbon supply limits hyporheic denitrification. J. Geophys. Res., 116, G04036, doi:10.1029/2011JG001730 Zou, X. M., H. H. Ruan, Y. Fu, X. D. Yang, and L. Q. Sha. 2005. Estimating soil labile organic carbon and potential turnover rates using a sequential fumigation–incubation procedure. Soil. Biol. Biochem. 37: 1923–1928. doi: 10.1016/j.soilbio.2005.02.028 Acknowledgments This research was supported by a grant from the Israel Science Foundation (grant 694/11) to S. Arnon, and by a grant (Israel-Italy Joint Innovation Program for Scientific and Technological Cooperation in R&D) from the Israeli Ministry of Science, Technology and Space and the Italian Ministry of Foreign Affairs, Directorate General for Political and Security Affairs to S. Arnon and F. Boano (grant 146). We thank Dr. Tali Brunner and Uri Wieler for laboratory assistance, Dr. Elite Levine for assistance in statistical analysis, and Efrat Elliani-Russak for helpful discussions.

1852

Submitted 19 July 2015 Revised 28 February 2016; 20 April 2016 Accepted 25 April 2016 Associate editor: Anna Romanı