Search

International audience

Amazonian hydroelectric reservoirs produce abundant carbon dioxide and methane from large quantities of flooded biomass that decompose anaerobically underwater. Emissions are extreme the first years after impounding and progressively decrease with time. To date, only water-to-air fluxes have been considered in these estimates. Here, we investigate in two Amazonian reservoirs (Balbina and Petit Saut) the fate of above water standing dead trees, by combining a qualitative analysis of wood state and density through time and a quantitative analysis of the biomass initially flooded. Dead wood was much more decomposed in the Balbina reservoir 23 years after flooding than in the Petit Saut reservoir 10 years after flooding. Termites apparently played a major role in wood decomposition, occurring mainly above water, and resulting in a complete conversion of this carbon biomass into CO 2 and CH 4 at a timescale much shorter than reservoir operation. The analysis of pre-impounding wood biomass reveals that above-water decomposition in Amazonian reservoirs is a large, previously unrecognized source of carbon emissions to the atmosphere, representing 26–45% of the total reservoir flux integrated over 100 years. Accounting for both below- and above-water fluxes, we could estimate that each km 2 of Amazonian forest converted to reservoir would emit over 140 Gg CO 2 -eq in 100 years. Hydropower plants in the Amazon should thus generate 0.25–0.4 MW h per km 2 flooded area to produce lower greenhouse gas emissions than gas power plants. They also have the disadvantage to emit most of their greenhouse gases the earliest years of operation.

At the interface between continent and ocean, estuaries receive particles, and especially particulate organic matter (POM) originating from these two reservoirs, but also produce POM, through autochthonous primary production. The origin and composition of surface POM in the Gironde Estuary (SW France) and the environmental forcing of its variability was investigated using the data set produced by the French Coastal Monitoring Network SOMLIT (Service d'Observation en Milieu LITtoral; monthly like sampling during years 2007–2009). This estuary is considered as a model of macrotidal turbid estuaries. Using elemental and isotopic composition of the POM, we estimated that, at the inner estuary space scale and inter-annual time scale, surface particulate organic carbon (POC) was composed of terrestrial POM originated from the turbidity maximum (96.4%; refractory POC) and flood events (1.6%; labile and refractory POC), and of riverine (0.1%), estuarine (0.8%) and marine (1.1%) phytoplankton, i.e. that POC was 98% and 2% of terrestrial and phytoplankton origin, respectively. However, there was a clear spatial gradient: the phytoplankton contribution increases from ca. 1% in the upper and middle estuary to 8.5% in the lower estuary, where light condition is more favourable to plankton growth. The low contribution of phytoplankton to the POC is a characteristic of the Gironde estuary and contrast with other large temperate estuaries. Statistical analysis indicates that salinity, river flow and SPM concentration, and thus associated hydro-dynamic and sedimentary processes, were the only environmental forcings to the composition of surface POC in this system, at intra- and inter-annual time scale. In contrast, temperature and nutrient concentrations, and thus associated processes, do not force this composition of POC. By combining POC fluxes entering the inner estuary (literature data), POC loss as dissolved organic carbon and CO 2 and as sediment trapping within the inner estuary (literature data), and our estimate of the composition of POC flux at the mouth of the estuary (96% and 4% of terrestrial and phytoplankton origin), a first-order net export of POC originating from the Gironde to the continental shelf was estimated: it amounts 48,150 tC yr −1 , and is composed of 46,200 tC yr −1 of terrestrial material and of 1950 tC yr −1 of estuarine phytoplankton. POC exported by the Gironde Estuary is thus poorly bioavailable for shelf pelagic and benthic food webs.

Degassing of terrestrially-respired CO 2 from streams and small rivers appears to be a significant component in watershed carbon budgets. Here we propose an original approach to quantify CO 2 degassing in small headwater bodies using pCO 2 , DIC (or total alkalinity, TA) and δ 13 C-DIC data in stream waters that avoids the difficulty of measuring or choosing a gas transfer velocity. Our inversed model applies to acidic, non-buffered (humic-type) waters and relies on two main assumptions, i.e., the stable isotopic composition of DIC in groundwater seeping to surface water (CO 2 from respired soil organic carbon and HCO 3 - from weathering) and on kinetic fractionation at the water–air interface ( 12 CO 2 degases to the atmosphere more rapidly than 13 CO 2 ). We first consider both the soil organic matter isotopic composition and the isotopic fractionation of CO 2 in the soil, to derive the δ 13 C–CO 2 in that soil and groundwater. From the HCO 3 - concentrations in streams, we estimate the relative contribution of silicate and carbonate weathering (the latter being minor in these waters) to the HCO 3 - and its associated isotopic composition. Model calculations start from the δ 13 C-DIC value computed by the aforementioned method and consist of two interlocked iterative procedures. The first procedure simulates the decrease in pCO 2 and the increase in δ 13 C-DIC that occur along the stream watercourse during degassing, starting from an assumed initial soil pCO 2 and ending at the in situ pCO 2 or δ 13 C-DIC. The second iteration procedure consists of adjusting the initial soil pCO 2 until pCO 2 and δ 13 C-DIC simultaneously reach the in situ measured values. After convergence is obtained, the model computes a theoretical concentration of DIC, [DIC] ex. , that has been lost as CO 2 to the atmosphere from the headwater to the sampling point in the river. [DIC] ex. can be multiplied by the river discharge to derive the quantity of carbon degassed from the river surface. The model was tested on seasonal field datasets from three small rivers draining sandy podsols in southern France and gave annual areal degassing rates comparable to those reported in other studies, though somewhat larger (upper half range in two rivers, ∼10 times the average in one stream). Part of this discrepancy might have been caused by an intense degassing in the vicinity of groundwater seeps, which was accounted for our integrative method but not by classical methods based on stream water pCO 2 and gas transfer velocity. The sensitivity of the model results on the assumption of the importance of carbonate weathering might also explain part of this high degassing rate. The model reproduced consistent values and seasonal trends of soil pCO 2 (maximal in summer) and gas transfer velocity (maximal at high water flow). We discuss the sensitivity of the model to the different parameters and assumptions and propose some improvements including groundwater sampling, for better constraining the computed degassing rates.

Fatty acid (FA) composition and stable isotope (δ 13 C, δ 15 N) signatures of four aquatic plants, plankton, sediment, soil and suspended particulate organic matter (SPOM) collected from open floodplain lakes (Varzea) and rivers of the central Brazilian Amazon basin were gathered during high and low water stages in 2009. SPOM from Varzea had a major contribution of autochthonous material from phytoplankton and C 3 aquatic plants. As shown from stable isotope composition of SPOM (δ 13 C −31.3 ± 3.2‰; δ 15 N 3.6 ± 1.5‰), the C 4 aquatic phanerogam (δ 13 C −13.1 ± 0.5‰; δ 15 N 4.1 ± 1.7‰) contribution appeared to be weak, although these plants were the most abundant macrophyte in the Varzea. During low water season, increasing concentration of 18:3ω3 was recorded in the SPOM of lakes. This FA, abundant mainly in the Varzea plants (up to 49% of total FAs), was due to the accumulation of their detritus in the ecosystem. This dry season, when connectivity with the river mainstem was restricted, was also characterized by a high concentration in the SPOM of the cyanobacteria marker 16:1ω7 (up to 21% of total FAs). The FA compositions of SPOM from the Amazon River also exhibited significant seasonal differences, in particular a higher concentration of 16:1ω7 and 18:3ω3 during the dry season. This suggests a seasonal contribution of autochthonous material produced in Varzea to the Amazon River SPOM.

Estuarine turbidity maxima (ETMs) are sites of intense mineralisation of land-derived particulate organic matter (OM), which occurs under oxic/suboxic oscillating conditions owing to repetitive sedimentation and resuspension cycles at tidal and neap-spring time scales. To investigate the biogeochemical processes involved in OM mineralisation in ETMs, an experimental set up was developed to simulate in vitro oxic/anoxic oscillations in turbid waters and to follow the short timescale changes in oxygen, carbon, nitrogen, and manganese concentration and speciation. We present here the results of a 27-day experiment (three oxic periods and two anoxic periods) with an estuarine fluid mud from the Gironde estuary. Time courses of chemical species throughout the experiment evidenced the occurrence of four distinct characteristic periods with very different properties. Steady oxic conditions were characterised by oxygen consumption rates between 10 and 40 mmol L

The study of the fate and composition of organic matter in aquatic systems is of interest since organic matter is believed to play a key role in biogeochemical cycles. Fluorescent dissolved organic material (FDOM) represents a major fraction of dissolved organic matter (DOM) but is under explored, particularly in estuarine environments. In this study, 3D fluorescence excitation-emission matrix (EEM) spectroscopy was combined with tangential ultrafiltration (UF) to investigate the changes in FDOM properties during the mixing of fluvial and marine waters in two estuaries on the French Atlantic coast. Water samples were sequentially filtered through three membranes of decreasing molecular weight (MW) cut off points (3 kDa, 1 kDa and 500 Da). Four ultrafiltration fractions were obtained and were analysed using EEM spectroscopy, as well as for their DOC content. Satisfactory fluorescence and DOC mass balances were obtained after the sequential UF fractionation despite the variable DOC concentration and ionic strength of the estuarine samples, showing the reliability of the UF systems. DOC distribution in the different size classes was observed to vary between the upstream and downstream parts of the estuaries and to be different in the two estuaries. The value of combining UF and EEM spectroscopy was shown, since the coupling allowed observation and isolation of some fluorophores which could not be distinguished in the spectra of bulk samples. Thus, the β fluorophore, observed mainly in marine samples until now, and which is associated with material of autochthonous biological origin, was present in the lowest size fraction (MW < 500 Da) from all water samples, including freshwater ones. The γ fluorophore, characteristic of protein-like compounds, was observed in the small size fractions of some fresh and brackish water samples, but it was not visible in the spectra of the corresponding bulk water samples. The differences in fluorescence properties of the size fractions of DOM were further evaluated by determining three fluorescence indices: the fluorescence index (f450/f500), the humification index (HIX) and the index of recent autochthonous contribution (BIX). The origin and degree of transformation of DOM can be assessed through the calculation of these indices. We showed that the most humified OM was associated with different size fractions, depending on water sample origin, and was characterised by intermediate MW (higher for fresh than marine water). We also showed that DOM distribution and characteristics differ between the Gironde and Seine estuaries.

In estuaries, the gas transfer velocity (k) is driven by a combination of two major physical drivers, wind and water current. The k values for CO2 in the macrotidal Gironde Estuary were obtained from 159 simultaneous pCO2 and floating chamber flux measurements. Values of k increased with wind speed and were significantly greater when water currents and wind were in opposing directions. At low wind speeds ( 0.2 g L �1 ) had a significant role in attenuating turbulence and therefore gas exchange. This result has important consequences for modeling water oxygenation in estuarine turbidity maxima. For seven low turbidity estuaries previously described in the literature, the slope of the linear regression between k and wind speed correlates very well with the estuary surface area due to a fetch effect. In the Gironde Estuary, this slope follows the same trend at low turbidity (TSS < 0.2 g L �1 ), but is on average significantly lower than in other large estuaries and

Silva et al. propose a new method for quantifying benthic net community production (NCP) of tidal flats under submerged condition, based on the monitoring of water pCO2 in a transparent benthic chamber around high tide. I demonstrate here with theoretical considerations that this method is inappropriate for coastal environments, because it allows only the quantification of the change in the dissolved CO2 which, at classical seawater pH, is only ∼10% of the change of the dissolved inorganic carbon (DIC). Total Alkalinity and/or DIC must be measured at the beginning and end of incubations in order to compute NCP in coastal environments. However, I also demonstrate that when pH is below 7, more than 95% of the DIC change occurs in the CO2 pool. The method proposed by Silva et al. is thus valuable for freshwater environments with acidic, low alkalinity waters, where monitoring the water pCO2 in a vial or chamber provides alone a very close approximation of the planktonic or benthic net community production.

Tropical hydroelectric reservoirs contribute significantly to atmospheric CH4 and CO2 emissions. To evaluate the contribution of the mineralization of the flooded soils and biomass to these atmospheric gas emissions, field and laboratory experiments were conducted. Cores were retrieved inform the littoral zone of the Petit Saut Reservoir (French Guiana), flooded 10 a prior to sampling, and different soils and plant material in the tropical forest surrounding the reservoir. All the samples were flooded and incubated in anoxic conditions in the dark at 30 C. The potential CH4 and CO2 production rates were determined. Soils and plant material from the tropical forest were incubated over one year and the production measurements were performed at a frequency of 1–5 months. Methane and CO2 production rates of soils and littoral sediments were linearly correlated to the organic C (OC) content of the slurries. The slopes of the relationships were 2:6 � 10 2 � 5:6 �

The hypothesis of nutrient-rich pore-waters seeping at low tide through sediments to channel waters, which drain tidal flats during ebb, was evaluated in the Arcachon lagoon. The back of the bay is affected by freshwater inputs and underground freshwater discharges. The upper part of tidal flat consists of permeable sandy sediments, which are covered by a muddy sediment layer on the lower part. Permeable sediments outcrop in the bed of channel web. Surface water chemistry and early diagenesis processes in sediment were estimated by collecting channel web waters and cores on a tidal flat and in channels at different seasons and time scales. Waters from tidal creeks are under-oxygenated, and enriched in reduced solutes. Muddy sediments showed evidences of strong organic matter mineralization and bioturbation. Underlying permeable sandy sediments revealed as well evidences of an enrichment of inorganic nutrients, and dilution with fresh continental groundwater. During ebb, tidal creek waters stem from mudflats by seeping of anoxic pore-waters, and from permeable sediments by advection of reduced waters. A rough estimation shows that the yearly contribution of this tidal pump of pore-waters for dissolved inorganic phosphorus (DIP) and ammonia inputs is of the same order of magnitude than river inputs for the studied part of the bay. Extrapolated to the whole Arcachon lagoon, pore-water discharge at low tide supplies to water column at least 556 kmol yr(-1) and 18300 kmol yr(-1) of DIP and NH4+, respectively. Tidal drainage at low tide represents therefore a minimal contribution of recycled nutrient of 55% for DIP and 15% for dissolved inorganic nitrogen to the lagoon. (c) 2008 Elsevier B.V. All rights reserved.

An equilibrator system connected to an infrared photo acoustic gas analyzer was used in order to measure directly in situ the concentrations of dissolved CO2 and CH4 in waters of a tropical reservoir (Petit Saut, French Guiana). The performance of the system was tested both on a vertical profile in the stratified water body of the reservoir and in the surface waters of the river downstream the dam. Results agreed with conventional GC analysis at F15% in a wide range of concentrations (CO2:50–400 Amol l 1 and CH4:0.5–350 Amol l 1 corresponding to gas partial pressures of respectively 1700–13,000 and 12–8800 Aatm). The time needed for in situ measurements was equivalent to water sampling, time for GC analysis in the laboratory being suppressed. The continuous monitoring of gas concentrations for 24 h in the reservoir surface waters revealed rapid changes in concentrations highly significant in the daily emission budget. The system opens new perspectives for the monitoring of gas concentrations in highly dynamic systems like tropical reservoirs. D 2005 Elsevier B.V. All rights reserved.

Vertical profiles of heavy metals (Cd, Cu, Cr, Fe, Mn, Ni and Pb) in the particulate and dissolved phases and redox sensitive parameters (dissolved oxygen, nitrate, nitrite and ammonium) were analyzed for the first time in the continuum Maximum Turbidity Zone (MTZ)–Fluid Mud (FM; 50 g l 1

A cross-system comparison of organic carbon origin and behaviour in nine European estuaries is presented. The study sites display a very large range of hydrological and environmental conditions. The watershed of the respective estuaries were characterized by plotting the total organic carbon (TOC) in the rivers versus the inhabitants/discharge ratio, This allows to distinguish four types of watershed with regard to anthropogenic forcing and organic carbon levels: polluted by sewage inputs (Scheldt and to a much lesser extent, Ems, Sado and Thames), decontaminated (Elbe and Rhine), pristine (Gironde and Douro) and eutrophized (Loire and Scheldt). In the estuarine zone, dissolved organic carbon (DOC) almost always decreased linearly with increasing salinity. Exceptions were: the Scheldt, where a net consumption of sewage-derived DOC was observed, the Gironde, where a net production of DOC occurred in the maximum turbidity zone (MTZ) and the Sado and Ems, where DOC was supplied from large intertidal areas. By contrast, a large fraction of the riverine particulate organic carbon (POC) was mineralized in all the estuaries, except the Douro, where residence time of waters is only a few days. A fraction of POC appeared however refractory and accumulated in the MTZs, where terrestrial soil-derived material dominates (Elbe, Ems, Loire, Gironde and Sado). In the marine regions of most estuaries, autochthonous POC was present during spring and summer. The analysis of all river and estuarine data allows estimation of the loss of continental POC occurring in each estuary. It decreases in the following order: Scheldt much greater than Thames > Ems = Sado = Loire > Gironde > Elbe > Rhine > Douro, which almost corresponds to the anthropogenic pressure in the respective watersheds. Two major variables appear to control the intensity of this mineralization: the origin of the POC, the lability increasing with pollution, and the residence time of particles in the estuarine zone.

In the Gironde Estuary, most part of phytoplanktonic material carried by the rivers is mineralised in the maximum turbidity zone (MTZ). In order to follow the degradation of the phytoplanktonic material into the MTZ, we developed an in vitro approach based on the monitoring of phytoplanktonic pigments. Algal material from two chlorophytes (Scenedesmus suspicatus Chaudat and Chlamydomonas sp.) was incubated in the dark during 28 days into water samples from the Gironde estuary MTZ, at variable suspended solid concentrations (SPM) as well as in a sterilised turbid sample. First order decay constants of chlorophylls a and b and lutein increased by a factor 3 to 5 between SPM of 0 and 3 g l −1 . The production of pheophytin a in the presence of particles and the lack of degradation in the sterilised turbid sample confirmed the effect of attached bacteria on the particles. To cite this article: E. Lemaire et al., C. R. Geoscience 334 (2002) 251–258.

We present results of one year observations in highly heterotrophic and oxygen-depleted rivers of the polluted Scheldt basin. Monthly measurements revealed a high variability for dissolved inorganic carbon and nitrogen, with the following strong parallelism: highest alkalinity and NH 4 + were associated with lowest NO 3 − and oxygen and vice-versa . In river water incubations, nitrification lowered the alkalinity whereas denitrification raised it; in an anoxic, NO 3 − -free incubation an increase of alkalinity was observed, partially due to ammonification. A stoichiometric analysis, taking into account the amount of protons produced or consumed by each process involving nitrogen, revealed that monthly variations of NO 3 − and NH 4 + with ammonification, nitrification and denitrification could explain the 28 and 62% alkalinity variations at all stations, except one. The remaining part of the alkalinity variations was attributed to other anaerobic processes (Mn-, Fe- and SO 4 -reductions). This trend seems to be the result of the whole catchment metabolism (riverine waters and sediments, sewage networks and agricultural soils). The observed HCO 3 − concentrations in the Scheldt basin were 2–10 times higher than the representative concentrations reported in pristine basins and used in chemical weathering models. This suggests the existence of an anthropogenic source, originating from organic matter decomposition. We conclude that in highly polluted basins, nitrogen transformations strongly influence the acid–base properties of water.

The transport of excess atmospheric CO2 (defined as the fraction of dissolved inorganic carbon (DIC) that can escape as CO2 to the atmosphere due to water-air equili- bration), by the five rivers entering the Scheldt estuary is investigated. Excess CO2 originates from both respiration in the soil and the river and represents 10 % of the DIC and 6 % of the total carbon input into the estuary. The ventilation of this CO2 in the estuary is however a minor contribution (10 %) to the total estuarine emission to the atmosphere, compared to heterotrophic activity and acidification due to nitrification within the estuarine zone. © 2000 Academie des sciences / Editions scientifiques et medicales Elsevier SAS carbon dioxide / respiration / soils / rivers / ventilation / estuaries Resume - Exces de gaz carbonique atmospherique apporte par les fleuves a l'estuaire de l'Escaut. L'apport d'exces de CO2 (defini comme la fraction du carbone inorganique dissous (CID) qui peut s'echapper vers l'atmosphere par equilibration entre l'eau et l'air), par les cinq rivieres alimentant l'estuaire de l'Escaut est etudie. L'exces de CO2, produit par la respiration dans les sols et les rivieres, represente 10 % du CID et 6 % des apports totaux de carbone a l'estuaire. La ventilation de ce CO2 dans l'estuaire ne represente cependant qu'une faible part (10 %) de l'emission totale de CO2 vers l'atmosphere par l'estuaire, en comparaison avec l'activite heterotrophe et avec l'acidifica- tion due a la nitrification. © 2000 Academie des sciences / Editions scientifiques et medi- cales Elsevier SAS gaz carbonique / respiration / sols / rivieres / ventilation / estuaires