Production of hydrogen peroxide in an intra-meander hyporheic zone at East River, Colorado

The traditionally held assumption that photo-dependent processes are the predominant source of H2O2 in natural waters has been recently questioned by an increrasing body of evidence showing the ubiquitiousness of H2O2 in dark water bodies and in groundwater. In this study, we conducted field measurement of H2O2 in an intra-meander hyporheic zone and in surface water at East River, CO. On-site detection using a sensitive chemiluminescence method suggests H2O2 concentrations in groundwater ranging from 6 nM (at the most reduced region) to ~ 80 nM (in a locally oxygen-rich area) along the intra-meander transect with a maxima of 186 nM detected in the surface water in an early afternoon, lagging the maximum solar irradiance by ∼ 1.5 h. Our results suggest that the dark profile of H2O2 in the hyporheic zone is closely correlated to local redox gradients, indicating that interactions between various redox sensitive elements could play an essential role. Due to its transient nature, the widespread presence of H2O2 in the hyporheic zone indicates the existence of a sustained balance between H2O2 production and consumption, which potentially involves a relatively rapid succession of various biogeochemically important processes (such as organic matter turnover, metal cycling and contaminant mobilization). More importantly, this study confirmed the occurrence of reactive oxygen species at a subsurface redox transition zone and further support our understanding of redox boundaries on reactive oxygen species generation and as key locations of biogeochemical activity.

Reactive oxygen species (ROS), such as superoxide (O 2 ·− ), hydrogen peroxide (H 2 O 2 ) and hydroxyl radical (HO · ), are oxygen-containing reactive molecules with considerable environmental importance due to their high reactivity in mediating redox transformations 1,2 . Extensive studies have shown that ROS are involved in various biogeochemical and ecologically significant processes, including nutrient cycling 3,4 , contaminant transformation 5 , oxidative stress and intracellular signalling in organisms [6][7][8] , as well as the establishment of microbial symbioses 9,10 . In addition, recent increasing recognition of the significant roles of ROS in soil carbon remineralisation 11,12 and metal bioavailability 3 urges a more comprehensive understanding of ROS generation and transformation in natural environments.
The dominant source of ROS in natural aquatic environments has long been attributed to photo-dependent processes. For example, photo-oxidation of chromophoric dissolved organic matter (CDOM) is generally considered the predominant source of H 2 O 2 (the longest-lived ROS) in surface waters, with a number of studies reporting the occurrence of H 2 O 2 in a wide range of natural waters, including oceans [13][14][15][16] , lakes [17][18][19] , coastal and estuarine waters 20,21 , freshwater [22][23][24] , and geothermal waters 25,26 . However, this commonly held assumption has been recently questioned by an increasing body of evidence demonstrating ROS production under dark conditions, either of biological and/or chemical origins [27][28][29][30] . To be more specific, while Zhang et al. 27 reported the widespread occurrence of ROS (O 2 ·− and H 2 O 2 ) in dark brackish and freshwater environments and attributed such dark profiles (at comparable concentrations as in sunlit waters) to biological contributions, Page et al. 28 and Tong et al. 31 detected extensive abiotic ROS flux from oxygenation of reduced humic acids and subsurface sediments under dark conditions, respectively. Further, our recent field study revealed light-independent generation of H 2 O 2 in groundwater of an alluvial aquifer adjacent to the Colorado River near Rifle, CO and demonstrated that fluctuating redox conditions created at oxic/anoxic interfaces are likely to be the key locations for ROS

Materials and methods
Reagents. Analytical grade chemicals were purchased from Sigma-Aldrich or VWR International (or as otherwise stated) and used without further refinement. Acridinium ester (AE, 10-methyl-9-(p-formylphenyl) acridinium carboxylate trifluoromethanesulfonate) for measurement of H 2 O 2 was purchased from Cayman Chemical (Michigan, USA). AE reagent of 5.0 µM at pH 3.0 was prepared each sampling day by dilution of a refrigerated AE stock solution (250 µM, pH 3.0) by adding 1.0 mM phosphate buffer (pH 3.0). H 2 O 2 (30%, ultra-high purity) was purchased from BDH Chemicals and calibrated spectrophotometrically using a UV-Vis spectrophotometer 41 . 0.1 M carbonate buffer of pH 11.3 was prepared weekly and other stock solutions were refrigerated at 4 °C in the dark.
Field information and experimental measurement. Field site and water sampling. Field measurements of H 2 O 2 were conducted in middle July 2016 at the East River floodplain, located in a high elevation catchment in southwestern Colorado (between 38.8° to 38.9° N and 106.8° to 106.9° W) with rolling-to-mountainous topography. The floodplain includes multiple river meanders that extend over a distance of 11 km (Fig. 1). The site encompasses the drainages of the East River and Copper Creek with the stream flow fed predominantly by snowmelt in late spring to early summer and mid-to late-summer monsoonal rainfall triggering rapid but punctuated increases in flow. The area experiences an average temperature ranging from a low of − 8 °C in January to a high of 35 °C in July. More site information can be found in the recent study by Dwivedi et al. 32 Because of the inherent spatial heterogeneity in a floodplain environment, we investigated a typical transect of Meander C of the East River site that is approximately parallel to the hyporheic flow field, labeled as MCP on Fig. 1. Specifically, the MCP transect includes five piezometric observation wells for geochemical samples (Wells MCP1 to MCP5) with these wells and transect accordingly referred to as Meander C Piezometric (MCP) wells and MCP transect (Fig. 1). Stainless steel drive-point piezometers were installed to achieve a sampling depth of 1.85 m  www.nature.com/scientificreports/ below ground surface. Groundwater was pumped from each piezometer using a peristaltic pump and sampled after purging a volume of water equivalent to three times the water standing in each piezometer. H 2 O 2 detection was performed both in groundwater and in the surface river water at meander C. Groundwater was peristaltically pumped through opaque flexible tubing into an amber high density polyethylene (HDPE) bottle with no head space and was analysed within 4-6 min after collection. Filtration of the groundwater by 0.22 µm filters made no difference on the final concentration of H 2 O 2 , therefore no filtration was done to minimize the sampling time. Unfiltered groundwater samples were collected for pH, electrical conductivity (EC) (data published in Dwivedi et al. 32 , not included in this study), oxidation-reduction potential (ORP) and dissolved oxygen in the field. Filtered (0.45 m PVDF syringe filters) samples were collected for in-field ferrous ion (Fe(II)) measurement and laboratory analysis of metals and dissolved organic and inorganic carbon. Samples for elemental analysis and carbon measurements were collected and acidified to pH 3 with HCl (high purity, 30% w/v; Sigma) and kept cool or refrigerated until shipped to the laboratory in Berkeley, CA. Results and discussion H 2 O 2 profiles measured at the Meander C transect. The East River floodplain, as shown in Fig. 1, includes multiple active river meanders, fresh oxbow lakes and older oxbow deposits, within which pronounced redox gradients have been observed during preliminary work 32,46 . Among all the active meanders at the site, the one designated meander C in Fig. 1 is typical in many river meanders with hyporheic flowpaths cutting across large sections of the meander, leading to the potential of substantial sinuosity-driven hyporheic exchange and formation of sharp redox gradients that favor the generation of ROS 47 . Therefore, H 2 O 2 measurements were first conducted at transect of Meander C that is approximately parallel to the hyporheic flow field where seasonal fluxes are predicted to occur 32,36 (labeled as MCP in Fig. 1). In order to obtain an estimation of the differences in H 2 O 2 content in the river water and in the hyporheic zone, measurement of H 2 O 2 , oxidation-reduction potential (ORP), dissolved oxygen (DO), Fe(II) and dissolved organic carbon (DOC) was also conducted in surface river water adjacent to the bank, with results shown in Fig. 2. First of all, despite the prevailing reducing conditions at MCP wells, concentrations of H 2 O 2 were significantly above the detection limit (1.0 nM) 30 in all waters sampled across the Meander C transect with a maximum of 88 nM found in the surface water and a minimum of 6.4 nM detected in the aphotic zone located in the middle of the intra-meander (MCP4, likely the most reduced region across the transect). Our detected H 2 O 2 profile in groundwater agrees with another field study conducted at Sand Ridge site, Illinois, which reported persistent detection of H 2 O 2 at nanomolar levels in groundwater over a 2-year period 48 . Figure 2 clearly demonstrates that H 2 O 2 pattern closely follows the local redox gradient, with relatively abundant H 2 O 2 generated in areas with more DO available and less H 2 O 2 occurring in regions under severely reducing conditions, as indicated by low ORP and high Fe(II) concentrations (Fig. 2b). As suggested in our previous study 30 , considering the slow groundwater flux velocity along the meander (< 10 -5 m/s) 32 , H 2 O 2 observed in the hyporheic zone was presumably actively produced in the vicinity of the sampling location rather than from H 2 O 2 transported downstream from the river water. To be more specific, a rough calculation from the data reported in Moffett (Fig. 2a). At the same time, the 78 nM H 2 O 2 observed in the river water near MCP1 will decrease to ~ 1 nM after three days. For the same meander, the groundwater velocities reported in Dwivedi et al. 32 are in the range of 1.25 × 10 -5 to 5.09 × 10 -6 m/s in the top 2 m, which suggested that the residence time of groundwater along the MCP transect was on the order of 1 to 2 months. Therefore, peroxide in river water would be all consumed before reaching the sampling points and H 2 O 2 transport within 2 days of the sampling time will not contribute to our data at any significant level. Moreover, fresh water has been reported to have higher H 2 O 2 decay rates, with a range of 0.11-8.9 h −1 reported by Bond et al. 40 based on previous investigations, suggesting that our estimation for H 2 O 2 decay here is highly conservative, as H 2 O 2 consumption in hyporheic zone will be much faster due to greater microbial activity. The absence of detectable dissolved oxygen at MCP4 (lower than the detection limit of 20 µg/L) did not negate the occurrence of H 2 O 2 completely, though the distinguishable rebound of ORP value at this location (Fig. 2a) suggested a slight variation in the "bulk" redox condition of local water, possibly resulting from the existence of other oxidants (nitrate, iron, and sulfate et al.). Similar findings were evident in our previous study at the Rifle, CO field site where measureable H 2 O 2 was discovered in regions of the aquifer characterized as strongly O 2 -deficient 30 . Recent studies by Zhang et al. 29 and Borda et al. 50 proposed that H 2 O 2 production resulted from the combination of two HO · formed by the oxidation of water at sulfur-deficient sites on pyrite surfaces (Eq. 1). www.nature.com/scientificreports/ These studies were undertaken under anoxic conditions, demonstrating that O 2 is not necessarily a prerequisite for ROS formation under such circumstances. While speculative at this stage, such a mechanism could account for the observed H 2 O 2 at MCP4 considering that the East River valley in the study area is comprised of Mancos Shale bedrock formed by an agglomeration sequence of heterogeneous marine black shale with regions of elevated metal, metalloid, and pyrite content 46,51 . This is further supported in a recent study by Dwivedi et al. 32 revealing that pyrite is one of the major reactive minerals in river sediments collected from the MCP transect.
It is not unexpected that Fe(II) profile observed in the hyporheic zone at Meander C is roughly in an opposite pattern to that of DO (or at least for results from wells MCP1 to MCP3), as shown in Fig. 2b. Oxygen was consumed rapidly when moving from the river bank to the center of the meander, with a concomitant increase in Fe(II) concentration, suggesting that when oxygen content decreased to a certain level, other electron acceptors that are available, such as nitrate, iron, and sulfate, could also be reduced. Dwivedi et al. 32 also reported that the redox zonation along the MCP transect was highly dynamic and predominantly driven by groundwater velocities resulting from river-stage fluctuations. The presence of elevated Fe(II) at well MCP3 indicated that the groundwater along the meander centerline was probably segregated and not actively receiving oxygenated recharge. The oxidation of reduced transition metals, such as Fe(II) (in both complexed and uncomplexed forms), by oxygen has long been reported to result in the formation of O 2 ·− as a by-product and subsequent formation of H 2 O 2 by its disproportionation or further reaction with reduced metals or organics (Eqs. 2-4) [52][53][54][55] . While the di-radical reaction (Eq. 3) is less important considering the very low concentration of O 2 ·− , its further interactions with reduced metals (Eq. 4) and/or organics (such as hydroquinones) 55,56 could play a significant role in H 2 O 2 generation. Therefore, constant hyporheic exchange could facilitate the generation of ROS by the extensive intrusion of oxygen-rich river water to the porous, nutrient-rich sediment, as shown in a recent study by Murphy et al. 11 , in which the transformation of H 2 O 2 was observed when iron(II) monosulfide was exposed to oxic conditions caused by tidal cycling.
It is also noteworthy from Fig. 2 that even though the concentrations of both DO and Fe(II) are more elevated at MCP3 than that at MCP2, a lower H 2 O 2 content was detected in waters sampled at MCP3, indicating that the magnitude of H 2 O 2 presence at such locations is likely to be co-mediated by mechanisms other than simply the interaction between DO and Fe(II). In fact, the co-occurrence of H 2 O 2 and Fe(II) presents the possibility of Fenton-type reactions occurring, which have been generally considered to promote H 2 O 2 decomposition and generation of more reactive oxidants, including HO · (an indiscriminate oxidant that reacts rapidly with organic carbon and other recalcitrant compounds, Eq. 5) and high valent Fe species [57][58][59] . Similar opposite patterns of Fe(II) and dissolved oxygen was also observed in a field study at Illinois by Barcelona, et al. between 1984Barcelona, et al. between to 1986 , which indicated that such chemical gradients were very stable during their 2-year observation period regardless of the changes in reginal water levels at a rate of ~ 0.4 m/year. Futher, the organic carbon content presented across this hyporheic transect, as shown in Fig. 2c with an opposite pattern to that of H 2 O 2 , could also contribute to the evolution of ROS. Page et al. 12 observed that more HO · was generated in more reduced, iron and organic carbon rich waters and proposed that such dark HO · formation came from the oxidation of ferrous iron (through dark Fenton reactions) and reduced organic carbon by oxygen. This is in agreement with our observation that less H 2 O 2 was presented in regions with higher organic carbon content, indicating that more H 2 O 2 were involved in reactions with soil carbon to produce other more reactive species that can further contribute to carbon cycling and pollutant degradation 56,60 . Recently Tong et al. 31 reported an abundant generation of HO · (with H 2 O 2 acting as the intermediate) on oxygenation of various subsurface sediments sampled from a variety of typical redox-fluctuating subsurface environments and found that Fe(II)-containing minerals, particularly phyllosilicates, are the predominant contributor to HO · formation. Consequently, it is clear that these various forms of Fe(II) species not only contribute to H 2 O 2 formation, but are also involved in subsequent transformation of H 2 O 2 to HO · , potentially resulting in mineralization of organic carbon to CO 2 at rates that are much higher than those estimated for oxygenation and photolysis of surface waters and of the same order of magnitude as that produced from soil respiration 31,61-64 . Since soils contain more than three times as much carbon as either the atmosphere or terrestrial vegetation, changes in this major pool of organic carbon could have huge impacts on the global carbon budget [65][66][67] , despite the fact that processes affecting carbon migration and transformations in the subsurface domain remain the least understood component of the global carbon cycle.
The extremely low concentration of Fe(II) in surface water near either MCP1 or MCP5 in Fig. 2b indicates that it is unlikely to be an important participant in generation and/or consumption of H 2 O 2 in such oxygenated system. Rather, solar radiation-driven photochemical and phototrophic processes are generally considered the predominant source of H 2 O 2 in such surface sunlit waters. Nevertheless, despite the aforementioned various possible chemical mechanisms for H 2 O 2 generation in the intra-meander hyporheic zone, biological production is another essential contributor, and could become the dominant source under certain conditions. For example,  69 reported a wide range of production (3-259 nM h -1 ) and decay rates (0.02-8.87 h -1 ), and suggested that microbes are the predominant contributor to both H 2 O 2 generation and consumption in the system. Although investigations on the biotic community that is actively contributing to the observed H 2 O 2 patterns in both groundwater and surface water at this transect is beyond the scope of this study, it is worth mentioning that there is increasing evidence demonstrating the involvement of various groups of microorganisms in extracellular ROS production, including fungi, various species of phytoplankton and microalgae, symbionts in corals, and heterotrophic bacteria 6,27,70 .

Stability of hyporheic H 2 O 2 concentrations compared to surface water.
In order to reach a better understanding on whether hyporheic ROS profiles were driven by changes in surface water ROS concentrations, measurments were also conducted in river water close to the bank at different times of the day, with results shown in Fig. 3. It is apparent in Fig. 3a that H 2 O 2 in the surface river water exhibited a pronounced diel variation, in line with the importance of light-mediated pathways (photochemical and/or phototrophic) on ROS generation and consumption 27,[71][72][73] . Take the river water adjacent to MCP5 as an example, H 2 O 2 concentrations increased rapidly from the early morning (~ 20 nM at 6:30 a.m.) to 1:30 pm (a maximum detected concentration of 186 nM) and then decreased shortly thereafter, with a much lower value of 88 nM obtained at 3:30 p.m. (Fig. 3a). Though more sampling shortly before or after 1:30 p.m. at this location would presumably provide more detailed information regarding the maximal levels that H 2 O 2 concentrations could reach at local surface water, our detected value of 186 nM at 1330 h is also a faithful representation, considering that the highest daytime concentration of H 2 O 2 resulting from photochemical formation was normally reached 1.5 h after the maximum solar irradiance 22 . Extensive measurements of both H 2 O 2 and solar radiation intensity in the surface water from Water of Leith (Dunedin, New Zealand) over three years by Rusak et al. 23 suggested that concentrations of H 2 O 2 in this surface water stream fluctuated roughly concomitant with the intensity of solar UVB radiation with H 2 O 2 contents higher in summer than in winter as a result of the large summertime increases in solar radiation at sea level in southern New Zealand, particularly with regard to the photochemically important UVB band 74 . Another repeat measurement of H 2 O 2 in groundwater at MCP1 was conducted on July 18th with 57 nM H 2 O 2 found in the morning (8:15 a.m.), as shown in Fig. 3b, which is quite close to the concentration of 59 nM detected at the same location, but in the late afternoon (4:50 p.m.) on July 15th. Though the temporal and spatial patterns of H 2 O 2 in the groundwater are expected to be affected by a variety of processes, including trace metal mediated redox reactions, the oxidation of reduced organic moieties present in dissolved or soil organic matter 30,55 and the biologically-mediated generation and decomposition of H 2 O 2 , such nearly constant concentrations observed at MCP1 could also indicate there might be a stable mechanism governing H 2 O 2 patterns at this location. Indeed, it has been previously shown that the redox gradients along the MCP transect at East River watershed are relatively stable and unresponsive to minor water-level fluctuations (but do respond to storm events and shift on seasonal time scales instead of diel cycles) 32 . Therefore, in this study, we suggest that there is a relatively stable balance between H 2 O 2 production and consumption at this location with our detected profiles of H 2 O 2 here more likely a result of a sustained balance between various sources and sinks in the vicinity (i.e., steady-state concentrations). While we cannot completely rule out the possibility that our observed H 2 O 2 in groundwater at MCP1 was originated from H 2 O 2 initially present in the river water 70 as MCP1 sits very close to the river bank and is strongly influenced by surface water recharge, considering the very low concentration of H 2 O 2 (~ 14 nM) observed in surface river water adjacent to MCP1 at 7:30 a.m. on the same day (Fig. 3a), this is unlikely to be the major www.nature.com/scientificreports/ contributor of the much elevated H 2 O 2 (57 nM) detected in groundwater at MCP1 after less than 1 h, indicating that the H 2 O 2 profiles observed here were a result of active production in the vicinity of the sampling locations. In order to better probe the pervasiveness of H 2 O 2 in East River sediments, the measurment of H 2 O 2 was also carried out at several other locations where piezometric observation wells have been previously installed. Figure 3b presents the various concentrations of H 2 O 2 (range from 11 to 57 nM) detected in groundwater at a few other locations of the East River aquifer. As shown in Fig. 1, Meander O is a cutoff meander in which the old riverbed channel is seasonally inundated by water during spring snowmelt and does not contribute substantially to the main flow of the river during dry periods 46 . However, the progressive evolution of meander cut-off and oxbow lake formation is expected to dramatically alter the redox gradients within the meander soils/sediments, changing biogeochemical functioning while creating a new biogeochemically distinct feature of an oxbow lake/ deposit. Measurements in groundwater close to the new river channel of Meander O (the "neck") found detectable levels of H 2 O 2 with 11 nM and 38 nM observed at locations MO3 and MO4, respectively (Fig. 3b). The slightly lower concentrations of H 2 O 2 at Meander O might be correlated to the lower concentrations of extractable metals found in such seasonally inundated locations than regions that remain unsaturated for most of the year 46 . Alternatively, it's worth mentioning that MO4 located in a relatively permanent wetland section of the oxbow in the former river channel with high content of aged organic carbon 75  Environmental implications. As a central interface connecting the stream and groundwater ecosystems, hyporheic zone has been recently identified as a disproportionately important component of the watershed reactor due to its transitional chemical and biological characteristics and distinct ecological functions 33,76 . The oscillating redox conditions, constant biogeochemical exchanges and intensive microbial activities occurring at various spatial and temporal scales in the hyporheic flow create ideal conditions for ROS production. This study represented the field measurement of H 2 O 2 both in an intra-meander hyporheic zone and in surface water at East River floodplain, CO. Collectively, our results demonstrate that H 2 O 2 concentrations in the intra-meander transect closely follow the distribution of local redox gradients, ranging from 6 nM detected at the most reduced region to ~ 80 nM at locally oxygen-rich area. There are many possible mechanisms contributing to such H 2 O 2 patterns, including dark biological production and consumption, and metal-mediated redox interactions between various important elements driven by fluctuating redox conditions. Measurements in surface water revealed typical diel changes of H 2 O 2 concentrations with a maxima of 186 nM detected at 1:30 pm, lagging the maximum solar irradiance by ∼ 1.5 h, indicating the essential contribution from sunlight-driven photochemical and phototrophic processes. A recent field study on the organic carbon in sediments of this floodplain reported that 23-34% of sediment organic carbon is derived from shale, another large fraction of the Earth's total carbon stocks, demonstrating continuous disturbing of the persistence of aged carbon and its active cycling into riverine systems 75 .
Our observation of the pervasiveness of H 2 O 2 across the whole intrameander transect suggests that the entire floodplain area defined by the meandering river has the potential to sustain substantial ROS concentrations. For instance, a recent estimate on the volume of groundwater associated with subterranean estuary emission suggested that approximately 6.0 × 10 10 kg of water exchanged between oxic and anoxic conditions per day for the entire South Carolina coastline, which implied a potential ROS flux of up to 1.5 × 10 7 mol day -1 , comparable in magnitude to photochemical sources of ROS in surface waters 11 . Although the generation and consumption of ROS will largely depend on the local redox matrix and microbial activity, such sustained ROS behaviour under rapidly oscillating redox conditions represents a previously under-appreciated source of oxidation that drives biogeochemical processes, such as metal bioavailability and carbon cycling. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2022