Simultaneous heterotrophic and sulfur-oxidizing autotrophic denitrification process for drinking water treatment: control of sulfate production

A long-term performance of a packed-bed bioreactor containing sulfur and limestone was evaluated for the denitrification of drinking water. Autotrophic denitrification rate was limited by the slow dissolution rate of sulfur and limestone. Dissolution of limestone for alkalinity supplementation increased hardness due to release of Ca²⁺. Sulfate production is the main disadvantage of the sulfur autotrophic denitrification process. The effluent sulfate concentration was reduced to values below drinking water guidelines by stimulating the simultaneous heterotrophic and autotrophic denitrification with methanol supplementation. Complete removal of 75 mg/L NO₃-N with effluent sulfate concentration of around 225 mg/L was achieved when methanol was supplemented at methanol/NO₃-N ratio of 1.67 (mg/mg), which was much lower than the theoretical value of 2.47 for heterotrophic denitrification. Batch studies showed that sulfur-based autotrophic NO₂-N reduction rate was around three times lower than the reduction rate of NO₃-N, which led to NO₂-N accumulation at high loadings.     

Small-scale, hydrogen-oxidizing-denitrifying bioreactor for treatment of nitrate-contaminated drinking water

Nitrate removal by hydrogen-coupled denitrification was examined using flow-through, packed-bed bioreactors to develop a small-scale, cost effective system for treating nitrate-contaminated drinking-water supplies. Nitrate removal was accomplished using a Rhodocyclus sp., strain HOD 5, isolated from a sole-source drinking-water aquifer. The autotrophic capacity of the purple non-sulfur photosynthetic bacterium made it particularly adept for this purpose. Initial tests used a commercial bioreactor filled with glass beads and countercurrent, non-sterile flow of an autotrophic, air-saturated, growth medium and hydrogen gas. Complete removal of 2 mM nitrate was achieved for more than 300 days of operation at a 2-h retention time. A low-cost hydrogen generator/bioreactor system was then constructed from readily available materials as a water treatment approach using the Rhodocyclus strain. After initial tests with the growth medium, the constructed system was tested using nitrate-amended drinking water obtained from fractured granite and sandstone aquifers, with moderate and low TDS loads, respectively. Incomplete nitrate removal was evident in both water types, with high-nitrite concentrations in the bioreactor output, due to a pH increase, which inhibited nitrite reduction. This was rectified by including carbon dioxide in the hydrogen stream. Additionally, complete nitrate removal was accomplished with wastewater-impacted surface water, with a concurrent decrease in dissolved organic carbon. The results of this study using three chemically distinct water supplies demonstrate that hydrogen-coupled denitrification can serve as the basis for small-scale remediation and that pilot-scale testing might be the next logical step.     

Effects of pH and precipitation on autohydrogenotrophic denitrification using the hollow-fiber membrane-biofilm reactor

Experiments carried out in a hollow-fiber, membrane-biofilm reactor (HFMBR) showed that the optimum pH for autotrophic denitrification was in the range 7.7-8.6, with the maximum efficiency at 8.4. Increasing the pH above 8.6 caused a significant decrease in nitrate removal rate and a dramatic increase in nitrite accumulation. The pH rose by 1.2 units when a large buffer was not added, suggesting that some field applications may require pH control. Precipitation of Ca(2+) occurred in every experiment. Precipitation was the largest sink for carbonate, and it also offset alkalinity production by denitrification. Although the alkalinity increased in most cases, systems with a high carbonate buffer and high pH accentuated precipitation, and the net change in alkalinity was negative. The long-term success of field applications of the HFMBR may depend upon the interactions among calcium concentration, total carbonate concentration, pH, and alkalinity changes.     

Systematic evaluation of nitrate and perchlorate bioreduction kinetics in groundwater using a hydrogen-based membrane biofilm reactor

To evaluate the simultaneous reduction kinetics of the oxidized compounds, we treated nitrate-contaminated groundwater (∼9.4 mg-N/L) containing low concentrations of perchlorate (∼12.5 μg/L) and saturated with dissolved oxygen (∼8 mg/L) in a hydrogen-based membrane biofilm reactor (MBfR). We systematically increased the hydrogen availability and simultaneously varied the surface loading of the oxidized compounds on the biofilm in order to provide a comprehensive, quantitative data set with which to evaluate the relationship between electron donor (H₂) availability, surface loading of the electron acceptors (oxidized compounds), and simultaneous bioreduction of the electron acceptors. Increasing the H₂ pressure delivered more H₂ gas, and the total H₂ flux increased linearly from ∼0.04 mg/cm²-d for 0.5 psig (0.034 atm) to 0.13 mg/cm²-d for 9.5 psig (0.65 atm). This increased rate of H₂ delivery allowed for continued reduction of the acceptors as their surface loading increased. The electron acceptors had a clear hydrogen-utilization order when the availability of hydrogen was limited: oxygen, nitrate, nitrite, and then perchlorate. Spiking the influent with perchlorate or nitrate allowed us to identify the maximum surface loadings that still achieved more than 99.5% reduction of both oxidized contaminants: 0.21 mg NO₃-N/cm²-d and 3.4 μg ClO₄/cm²-d. Both maximum values appear to be controlled by factors other than hydrogen availability.     

Combined bioelectrochemical and sulfur autotrophic denitrification for drinking water treatment

A combined bioelectrochemical and sulfur autotrophic denitrification process for drinking water treatment was put forward and investigated extensively in this paper. In this new process, the bioelectrochemical denitrification was carried out in the upper part of the reactor while sulfur denitrification in the lower part. The H+ produced in Sulfur Part could be consumed by hydrogen denitrification in Bioelectrochemical Part. Therefore, the limestone for pH adjustment in Sulfur Part was not necessary in this combined process, which avoided the problem of hardness increase. The sulfate accumulation in this combined reactor was less than that of the sulfur limestone autotrophic denitrification system. The effluent from two parts was kept neutral at optimum operation conditions. When the influent nitrate was 30 mg-N/L, the reactor could be operated efficiently at the hydraulic retention time ranged from 1.9 to 5h (corresponding minimum current was 16-3 mA), i.e. the effluent NO3(-)-N removal ranged from 90% to 100% without nitrite accumulation and the effluent sulfate concentration was lower than 170 mg/L. The maximum volume-loading rate of the reactor was 0.381 kg NO3(-)-N/(m3d). The biomass and scanning electron microscope micrographs of Sulfur Part were also analyzed.     

Use of formic acid as reducing agent for application in catalytic reduction of nitrate in water

The reduction of nitrate in nitrogen using bimetallic palladium tin catalysts and hydrogen is an interesting process for water treatment. The aim of the present study is to use formic acid (FA) as a reducing agent and a pH buffer in order to substitute the mixture of hydrogen and carbon dioxide. The catalytic performances of a palladium tin catalyst supported on silica were evaluated in the presence of FA, as a function of the initial acid concentration and of the gas phase (N(2), CO(2), or H(2)). Results were compared to those obtained with hydrogen in the presence of carbon dioxide. Similar mechanisms seem to explain the identical catalytic performances observed with these two reducing agents.     

A study of nitrate in Tehran ground water and a method of its removal

This study investigates the concentration of nitrate in groundwater in the province of Tehran. It also proposes the combination of ion exchange and biological denitrification as an effective way of removing nitrate from drinking water. The concentration of nitrate was measured in more than 200 deep and semi‐deep wells spread throughout urban and rural areas in Tehran. The result proved the presence of nitrate in a wide area in the province, especially in the southern part of the city of Tehran, and areas next to industrial or agricultural centers. The results show that there is a linear relationship between R (the proportion of nitrate ion concentration to total nitrate and sulfate ion concentrations in the influent) and the amount of each nitrate and sulfate ion derived from total capacity of resins. The last step in this study included denitrification of the spent regenerant by SBR and BPBR. At high concentrations of salt and nitrate, BPBR performed well while SBR showed an average performance.     

A review of the impact of climate change on future nitrate concentrations in groundwater of the UK

This paper reviews the potential impacts of climate change on nitrate concentrations in groundwater of the UK using a Source-Pathway-Receptor framework. Changes in temperature, precipitation quantity and distribution, and atmospheric carbon dioxide concentrations will affect the agricultural nitrate source term through changes in both soil processes and agricultural productivity. Non-agricultural source terms, such as urban areas and atmospheric deposition, are also expected to be affected. The implications for the rate of nitrate leaching to groundwater as a result of these changes are not yet fully understood but predictions suggest that leaching rate may increase under future climate scenarios. Climate change will affect the hydrological cycle with changes to recharge, groundwater levels and resources and flow processes. These changes will impact on concentrations of nitrate in abstracted water and other receptors, such as surface water and groundwater-fed wetlands. The implications for nitrate leaching to groundwater as a result of climate changes are not yet well enough understood to be able to make useful predictions without more site-specific data. The few studies which address the whole cycle show likely changes in nitrate leaching ranging from limited increases to a possible doubling of aquifer concentrations by 2100. These changes may be masked by nitrate reductions from improved agricultural practices, but a range of adaption measures need to be identified. Future impact may also be driven by economic responses to climate change.     

Present limitations and future prospects of stable isotope methods for nitrate source identification in surface- and groundwater

Nitrate (NO₃⁻) contamination of surface- and groundwater is an environmental problem in many regions of the world with intensive agriculture and high population densities. Knowledge of the sources of NO₃⁻ contamination in water is important for better management of water quality. Stable nitrogen (δ¹⁵N) and oxygen (δ¹⁸O) isotope data of NO₃⁻ have been frequently used to identify NO₃⁻ sources in water. This review summarizes typical δ¹⁵N- and δ¹⁸O-NO₃⁻ ranges of known NO₃⁻ sources, interprets constraints and future outlooks to quantify NO₃⁻ sources, and describes three analytical techniques (“ion-exchange method”, “bacterial denitrification method”, and “cadmium reduction method”) for δ¹⁵N- and δ¹⁸O-NO₃⁻ determination. Isotopic data can provide evidence for the presence of dominant NO₃⁻ sources. However, quantification, including uncertainty assessment, is lacking when multiple NO₃⁻ sources are present. Moreover, fractionation processes are often ignored, but may largely constrain the accuracy of NO₃⁻ source identification. These problems can be overcome if (1) NO₃⁻ isotopic data are combined with co-migrating discriminators of NO₃⁻ sources (e.g. ¹¹B), which are not affected by transformation processes, (2) contributions of different NO₃⁻ sources can be quantified via linear mixing models (e.g. SIAR), and (3) precise, accurate and high throughput isotope analytical techniques become available.     

Simultaneous removal of perchlorate and nitrate from drinking water using the ion exchange membrane bioreactor concept

This work evaluates the feasibility of the ion exchange membrane bioreactor (IEMB) concept for the simultaneous removal of perchlorate and nitrate from drinking water, when nitrate is present in the ppm range and perchlorate in the ppb range. The IEMB concept combines Donnan dialysis and simultaneous biological degradation of both pollutants. Membrane transport studies showed that Donnan dialysis is suitable for obtaining water with concentrations of perchlorate and nitrate below the recommended levels. However, the pollutants were accumulated in a receiving stream, thus requiring additional treatment before disposal. On the other hand, the IEMB process operated with hydraulic retention times ranging from 1.4 to 8.3h in the water compartment, proved to remove effectively perchlorate and nitrate while preserving the water composition with respect to other ions, thus avoiding secondary contamination of the treated water. For a polluted water stream containing 100 ppb of ClO(4)(-) and 60 ppm of NO(3)(-), the concentrations of both ions in the treated stream were kept below the recommended levels of 4 ppb for ClO(4)(-) and 25 ppm for NO(3)(-). The IEMB system was operated under ethanol limitation, but even under these conditions, an increase of the perchlorate and nitrate concentrations in the treated water was not observed for up to 6 days.     

Kinetics of nitrate and perchlorate removal and biofilm stratification in an ion exchange membrane bioreactor

The biological degradation of nitrate and perchlorate was investigated in an ion exchange membrane bioreactor (IEMB) using a mixed anoxic microbial culture and ethanol as the carbon source. In this process, a membrane-supported biofilm reduces nitrate and perchlorate delivered through an anion exchange membrane from a polluted water stream, containing 60 mg/L of NO₃⁻ and 100 μg/L of ClO₄⁻. Under ammonia limiting conditions, the perchlorate reduction rate decreased by 10%, whereas the nitrate reduction rate was unaffected. Though nitrate and perchlorate accumulated in the bioreactor, their concentrations in the treated water (2.8 ± 0.5 mg/L of NO₃⁻ and 7.0 ± 0.8 μg/L of ClO₄⁻, respectively) were always below the drinking water regulatory levels, due to Donnan dialysis control of the ionic transport in the system.Kinetic parameters determined for the mixed microbial culture in suspension showed that the nitrate reduction rate was 35 times higher than the maximum perchlorate reduction rate. It was found that perchlorate reduction was inhibited by nitrate, since after nitrate depletion perchlorate reduction rate increased by 77%. The biofilm developed in the IEMB was cryosectioned and the microbial population was analyzed by fluorescence in situ hybridization (FISH). The results obtained seem to indicate that the kinetic advantage of nitrate reduction favored accumulation of denitrifiers near the membrane, whereas per(chlorate) reducing bacteria were mainly positioned at the biofilm outer surface, contacting the biomedium. As a consequence of the biofilm stratification, the reduction of perchlorate and nitrate occur sequentially in space allowing for the removal of both ions in the IEMB.     

Iron release from corroded iron pipes in drinking water distribution systems: effect of dissolved oxygen

Iron release from corroded iron pipes is the principal cause of "colored water" problems in drinking water distribution systems. The corrosion scales present in corroded iron pipes restrict the flow of water, and can also deteriorate the water quality. This research was focused on understanding the effect of dissolved oxygen (DO), a key water quality parameter, on iron release from the old corroded iron pipes. Corrosion scales from 70-year-old galvanized iron pipe were characterized as porous deposits of Fe(III) phases (goethite (alpha-FeOOH), magnetite (Fe(3)O(4)), and maghemite (alpha-Fe(2)O(3))) with a shell-like, dense layer near the top of the scales. High concentrations of readily soluble Fe(II) content was present inside the scales. Iron release from these corroded pipes was investigated for both flow and stagnant water conditions. Our studies confirmed that iron was released to bulk water primarily in the ferrous form. When DO was present in water, higher amounts of iron release was observed during stagnation in comparison to flowing water conditions. Additionally, it was found that increasing the DO concentration in water during stagnation reduced the amount of iron release. Our studies substantiate that increasing the concentration of oxidants in water and maintaining flowing conditions can reduce the amount of iron release from corroded iron pipes. Based on our studies, it is proposed that iron is released from corroded iron pipes by dissolution of corrosion scales, and that the microstructure and composition of corrosion scales are important parameters that can influence the amount of iron released from such systems.     

Effects of dissolved oxygen on formation of corrosion products and concomitant oxygen and nitrate reduction in zero-valent iron systems with or without aqueous Fe2+

Batch tests were conducted in zero-valent iron (ZVI or Fe0) systems to investigate oxygen consumption and the effect of dissolved oxygen (DO) on formation of iron corrosion products, nitrate reduction, the reactivity of Fe0, the role Fe2+ (aq) played, and the fate of Fe2+. The study indicates that without augmenting Fe2+ (aq), neither nitrate nor DO could be removed efficiently by Fe0. In the presence of Fe2+ (aq), nitrate and DO could be reduced concomitantly with limited interference with each other. Unlike nitrate reduction, DO removal by Fe0 did not consume Fe2+ (aq). A two-layer structure, with an inner layer of magnetite and an outer layer of lepidocrocite, may be formed in the presence of DO. When DO depleted, the outer lepidocrocite layer was transformed to magnetite. The inner layer of magnetite, even in a substantial thickness, might not impede the Fe0 reactivity as much as the thin interfacial layer between the oxide coating and liquid. Surface-bound Fe2+ may greatly enhance the electron transfer from the Fe0 core to the solid-liquid interface, and thus improve the performance of the Fe0 process.     

Comparison of pulsed and continuous addition of H2 gas via membranes for stimulating PCE biodegradation in soil columns

Column experiments were performed to investigate a technology for remediating aquifers contaminated with chlorinated solvents. The technology involves installation of hollow-fiber membranes in the subsurface to supply hydrogen gas (H2) to groundwater to support biological reductive dechlorination in situ. Three laboratory-scale columns [control (N2 only), continuous H2, and pulsed H2] were packed with aquifer material from a trichloroethene (TCE)-contaminated wetland in Minnesota and supplied with perchloroethene (PCE)-contaminated synthetic groundwater. The main goals of the research were: (1) evaluate the long-term performance of the H2 supply system and (2) compare the effects of pulsed (4 h on, 20 h off) versus continuous H2 supply (lumen partial pressure approximately 1.2 atm) on PCE dechlorination and production of by-products (i.e. methane and acetate). The silicone-coated fiberglass membranes employed in these experiments were robust, delivering H2 steadily over the entire 349-day experiment. Methane production decreased when H2 was added in a pulsed manner. Nevertheless, the percentage of added H2 used to support methanogenesis was similar in both H2-fed columns (92-93%). For much of the experiment, PCE dechlorination (observed end product = dichloroethene) in the continuous and pulsed H2 columns was comparable, and enhanced in comparison to the natural attenuation observed in the control column. Dechlorination began to decline in the pulsed H2 column after 210 days, however, while dechlorination in the continuous H2 column was sustained. Acetate was detected only in the continuous H2 column, at concentrations of up to 36 microM. The results of this research suggest that in situ stimulation of PCE dechlorination by direct H2 addition requires the continuous application of H2 at high partial pressures, favoring the production of bioavailable organic matter such as acetate to provide a carbon source, electron donor, or both for dechlorinators. Unfortunately, this strategy has proven to be inefficient, with the bulk of the added H2 used to support methanogenesis.     

Potential nitrate removal in a coastal freshwater sediment (Haringvliet Lake, The Netherlands) and response to salinization

Nitrogen transformations and their response to salinization were studied in bottom sediment of a coastal freshwater lake (Haringvliet Lake, The Netherlands). The lake was formed as the result of a river impoundment along the south-western coast of the Netherlands, and is currently targeted for restoration of estuarine conditions. Nitrate porewater profiles indicate complete removal of NO(3)(-) within the upper few millimeters of sediment. Rapid NO(3)(-) consumption is consistent with the high potential rates of nitrate reduction (up to 200 nmol N cm(-3) h(-1)) measured with flow-through reactors (FTRs) on intact sediment slices. Acetylene-block FTR experiments indicate that complete denitrification accounts for approximately half of the nitrate reducing activity. The remaining NO(3)(-) reduction is due to incomplete denitrification and alternative reaction pathways, most likely dissimilatory nitrate reduction to NH(4)(+) (DNRA). Results of FTR experiments further indicate that increasing bottom water salinity may lead to a transient release of NH(4)(+) and dissolved organic carbon from the sediment, and enhance the rates of nitrate reduction and nitrite production. Increased salinity may thus, at least temporarily, increase the efflux of NH(4)(+) from the sediment to the surface water. This work shows that salinity affects the relative importance of denitrification compared to alternative nitrate reduction pathways, limiting the ability of denitrification to remove bioavailable nitrogen from aquatic ecosystems.     

The role of climate on inter-annual variation in stream nitrate fluxes and concentrations

In recent decades, temporal variations in nitrate fluxes and concentrations in temperate rivers have resulted from the interaction of anthropogenic and climatic factors. The effect of climatic drivers remains unclear, while the relative importance of the drivers seems to be highly site dependent. This paper focuses on 2-6 year variations called meso-scale variations, and analyses the climatic drivers of these variations in a study site characterized by high N inputs from intensive animal farming systems and shallow aquifers with impervious bedrock in a temperate climate. Three approaches are developed: 1) an analysis of long-term records of nitrate fluxes and nitrate concentrations in 30 coastal rivers of Western France, which were well-marked by meso-scale cycles in the fluxes and concentration with a slight hysteresis; 2) a test of the climatic control using a lumped two-box model, which demonstrates that hydrological assumptions are sufficient to explain these meso-scale cycles; and 3) a model of nitrate fluxes and concentrations in two contrasted catchments subjected to recent mitigation measures, which analyses nitrate fluxes and concentrations in relation to N stored in groundwater. In coastal rivers, hydrological drivers (i.e., effective rainfall), and particularly the dynamics of the water table and rather stable nitrate concentration, explain the meso-scale cyclic patterns. In the headwater catchment, agricultural and hydrological drivers can interact according to their settings. The requirements to better distinguish the effect of climate and human changes in integrated water management are addressed: long-term monitoring, coupling the analysis and the modelling of large sets of catchments incorporating different sizes, land uses and environmental factors.     

Zero-valent iron reduction of nitrate in the presence of ultraviolet light,organic matter and hydrogen peroxide

This paper describes the use of metallic iron (Fe(0)) powder for nitrate removal in a well-mixed batch reactor. Important variables explored include Fe(0) dosage (1-3g/L), UV light intensity (64-128 W), and the presence of propanol (20 mg/L as DOC) and H(2)O(2) (100-200 mg/L). Accumulation of ferrous ions released from the Fe(0) surface can be expressed by an S-curve, which involves lag growth phase, exponential phase, rate-declining phase, and saturation phase. The removal of nitrate increases with increasing Fe(0) dosage; however, the removal makes no difference as the Fe(0) dosage is greater than 2 g/L. UV irradiation retards the dissolution of ferrous ion and the removal of nitrate. The species of propanol, which has a functional group of -OH, plays a role of organic inhibitor for Fe(0) corrosion. The presence of H(2)O(2) appears to inactivate all reactions as the Fe(0) of 10 microm was used; the final H(2)O(2) remains intact throughout the entire reaction period, and there were no removal of nitrate and no dissolution of ferrous ion. Surprisingly, with the use of a larger Fe(0) particle size of 150 microm, the H(2)O(2) was seen to decompose rapidly through Fenton reaction. Nevertheless, the rate of ferrous accumulation or nitrate removal is slow.     

Effect of salinity and inorganic nitrogen concentrations on nitrification and denitrification rates in intertidal sediments and rocky biofilms of the Douro River estuary, Portugal

The regulatory effects of salinity and inorganic nitrogen compounds on nitrification and denitrification were studied in intertidal sandy sediments and rocky biofilms in the Douro River estuary, Portugal, over a 12-month period. Nitrification and denitrification rates were measured in slurries of field samples and enrichment experiments using the difluoromethane and the acetylene inhibition techniques, respectively. Salinity did not regulate denitrification in either environment, suggesting that halotolerant bacteria dominated the denitrifier communities. However, nitrification rates were stimulated when salinity increased from 0 to 15 practical salinity units. NO3- addition experiments revealed that NO3- availability stimulates denitrification rates in sandy sediments, but not in rocky biofilms; however, in rocky biofilms a positive and linear relationship was observed between denitrification rates and water column NO3- concentrations (r=0.92) during the monthly surveys. The N2O:N2 ratios increased rapidly when NO3- increased from 63 to 363 microM; however, results from monthly surveys showed that environmental parameters other than NO3- availability may be important in controlling the variation in N2O production via denitrification. Ammonium additions to sandy sediments stimulated nitrification rates by 35% for the 20 microM NH4+ addition, but NH4+ appeared to inhibit nitrification at high concentration addition (200 microM NH4+). In contrast, rocky biofilm nitrification was stimulated by 65% when 200 microM NH4+ was added.     

Effect of organic enrichment and thermal regime on denitrification and dissimilatory nitrate reduction to ammonium (DNRA) in hypolimnetic sediments of two lowland lakes

We analyzed benthic fluxes of inorganic nitrogen, denitrification and dissimilatory nitrate reduction to ammonium (DNRA) rates in hypolimnetic sediments of lowland lakes. Two neighbouring mesotrophic (Ca' Stanga; CS) and hypertrophic (Lago Verde; LV) lakes, which originated from sand and gravel mining, were considered. Lakes are affected by high nitrate loads (0.2-0.7 mM) and different organic loads. Oxygen consumption, dissolved inorganic carbon, methane and nitrogen fluxes, denitrification and DNRA were measured under summer thermal stratification and late winter overturn.Hypolimnetic sediments of CS were a net sink of dissolved inorganic nitrogen (−3.5 to −4.7 mmol m⁻² d⁻¹) in both seasons due to high nitrate consumption. On the contrary, LV sediments turned from being a net sink during winter overturn (−3.5 mmol m⁻² d⁻¹) to a net source of dissolved inorganic nitrogen under summer conditions (8.1 mmol m⁻² d⁻¹), when significant ammonium regeneration was measured at the water-sediment interface. Benthic denitrification (0.7-4.1 mmol m⁻² d⁻¹) accounted for up to 84-97% of total NO₃⁻ reduction and from 2 to 30% of carbon mineralization. It was mainly fuelled by water column nitrate. In CS, denitrification rates were similar in winter and in summer, while in LV summer rates were 4 times lower. DNRA rates were generally low in both lakes (0.07-0.12 mmol m⁻² d⁻¹). An appreciable contribution of DNRA was only detected in the more reducing sediments of LV in summer (15% of total NO₃⁻ reduction), while during the same period only 3% of reduced NO₃⁻ was recycled into ammonium in CS.Under summer stratification benthic denitrification was mainly nitrate-limited due to nitrate depletion in hypolimnetic waters and parallel oxygen depletion, hampering nitrification. Organic enrichment and reducing conditions in the hypolimnetic sediment shifted nitrate reduction towards more pronounced DNRA, which resulted in the inorganic nitrogen recycling and retention within the bottom waters. The prevalence of DNRA could favour the accumulation of mineral nitrogen with detrimental effects on ecosystem processes and water quality.