A Correction to this paper has been published: https://doi.org/10.1038/s41597-021-00851-9.

The eddy-covariance method represents the only direct way to measure the turbulent fluxes of momentum, temperature, trace gases, and particles between the land surface and the atmosphere. It is a direct measurement of the net carbon-dioxide budget and dry deposition. For that purpose, it is widely used in networks of long-term ecosystem observatories around the world and is the centerpiece of intensive field campaigns investigating biosphere–atmosphere exchange processes. The instrumentation typically consists of a 3-D sonic anemometer/thermometer and one or more additional gas analyzers that are able to measure the high-frequency fluctuations of the scalar to be transported. These instruments are mounted on a meteorological mast to sample the turbulent field under the assumption that eddies are carried along with the mean wind. Further prerequisites of the method are horizontal homogeneity, steady-state conditions and well-developed turbulence. For successful application of the method, a series of quality tests and flux corrections is required, which will be presented together with commonly used instrumentation and postprocessing software. Moreover, we will provide a historical overview and provide guidelines for site selection and setup and the necessary maintenance procedures.

Methane emissions of a grazing herd of Belgian Blue cattle were estimated per individual on the field by combining eddy covariance measurements with geolocation of the cattle and a footprint model. This method allows the measurement of outdoor non-invasive methane emissions but is complex and subject to methodological issues. Estimated emissions were 220 +/- 35 g CH4LU-1 day(-1) (grams of methane per livestock unit per day), where the uncertainty corresponds to the random error and does not include any possible systematic error. Cattle behavior was also monitored and presented a clear daily pattern of activity with more intense grazing after sunrise and before sunset. However, no significant methane emission pattern could be associated with it, the diurnal emission variation being lower than the measurement precision.

L21: As already said above, in view of the uncertainties and of the interannual variability, I think that the difference between the two sites is not significant (and a source of 82 gC m-2 over 4 years is by no means “considerable” as the uncertainty on the annual sequestration is about 50 gC m-2yr-1). I found thus misleading to present one site as a source and the second as a sink. Both sites do probably not depart significantly from equilibrium. As this assertion is central to the paper, it questions all the discussion and conclusions. Introduction L78-80, Independently of the fact that I disagree with the sentence, an introduction should finish with a presentation of the objectives of the papers in the form of scientific questions, not with a result.

The presence of cows on a pasture considerably modifies exchanges of biogenic volatile organic compounds (BVOCs). By regulating the biomass present, they can have an impact on the constitutive flux (exchanges from soil and grass that are not induced by leaf wounding or trampling by cows) but they can also cause direct emissions from exhalation and indirect emissions by leaf injury (grazing), trampling and wastes. In this study conducted on the ICOS pasture site of Dorinne (Belgium), we disentangled these different sources/sinks for three oxygenated BVOCs commonly exchanged on grasslands (methanol, acetaldehyde and acetone), using a combination of turbulent flux measurements, enclosure flux measurements, tools to detect the presence and activity of cows in the footprint of the turbulent flux measurements and a flux footprint model. Direct exhalation emissions were low, representing only 2.3% and 10% of the spring total flux of methanol and acetone respectively. Comparison of grazed and non-grazed enclosures pointed out that emissions following leaf wounding were significant for all studied BVOCs, decreased exponentially with time to become negligible after maximum five days. Cow indirect emissions at the pasture scale (turbulent flux measurements) where likely dominated by grazing and were shown to be a major component of the total diurnal flux for each of the three studied BVOCs. Comparison with a hay meadow also showed that the temporal dynamics of those BVOC emissions were very different according to the grass management type, calling for specific parametrization in up-scaling emission models.

The effects of atmospheric nitrogen deposition (Ndep) on carbon (C) sequestration in forests have often been assessed by relating differences in productivity to spatial variations of Ndep across a large geographic domain. These correlations generally suffer from covariation of other confounding variables related to climate and other growth-limiting factors, as well as large uncertainties in total (dry + wet) reactive nitrogen (Nr) deposition. We propose a methodology for untangling the effects of Ndep from those of meteorological variables, soil water retention capacity and stand age, using a mechanistic forest growth model in combination with eddy covariance CO2 exchange fluxes from a Europe-wide network of 22 forest flux towers. Total Nr deposition rates were estimated from local measurements as far as possible. The forest data were compared with data from natural or semi-natural, non-woody vegetation sites. The response of forest net ecosystem productivity to nitrogen deposition (dNEP ∕ dNdep) was estimated after accounting for the effects on gross primary productivity (GPP) of the co-correlates by means of a meta-modelling standardization procedure, which resulted in a reduction by a factor of about 2 of the uncorrected, apparent dGPP ∕ dNdep value. This model-enhanced analysis of the C and Ndep flux observations at the scale of the European network suggests a mean overall dNEP ∕ dNdep response of forest lifetime C sequestration to Ndep of the order of 40–50 g C per g N, which is slightly larger but not significantly different from the range of estimates published in the most recent reviews. Importantly, patterns of gross primary and net ecosystem productivity versus Ndep were non-linear, with no further growth responses at high Ndep levels (Ndep > 2.5–3 g N m−2 yr−1) but accompanied by increasingly large ecosystem N losses by leaching and gaseous emissions. The reduced increase in productivity per unit N deposited at high Ndep levels implies that the forecast increased Nr emissions and increased Ndep levels in large areas of Asia may not positively impact the continent's forest CO2 sink. The large level of unexplained variability in observed carbon sequestration efficiency (CSE) across sites further adds to the uncertainty in the dC∕dN response.

The impact of atmospheric reactive nitrogen (Nr) deposition on carbon (C) sequestration in soils and biomass of unfertilized, natural, semi-natural and forest ecosystems has been much debated. Many previous results of this dC∕dN response were based on changes in carbon stocks from periodical soil and ecosystem inventories, associated with estimates of Nr deposition obtained from large-scale chemical transport models. This study and a companion paper (Flechard et al., 2020) strive to reduce uncertainties of N effects on C sequestration by linking multi-annual gross and net ecosystem productivity estimates from 40 eddy covariance flux towers across Europe to local measurement-based estimates of dry and wet Nr deposition from a dedicated collocated monitoring network. To identify possible ecological drivers and processes affecting the interplay between C and Nr inputs and losses, these data were also combined with in situ flux measurements of NO, N2O and CH4 fluxes; soil NO3- leaching sampling; and results of soil incubation experiments for N and greenhouse gas (GHG) emissions, as well as surveys of available data from online databases and from the literature, together with forest ecosystem (BASFOR) modelling. Multi-year averages of net ecosystem productivity (NEP) in forests ranged from −70 to 826 g C m−2 yr−1 at total wet + dry inorganic Nr deposition rates (Ndep) of 0.3 to 4.3 g N m−2 yr−1 and from −4 to 361 g C m−2 yr−1 at Ndep rates of 0.1 to 3.1 g N m−2 yr−1 in short semi-natural vegetation (moorlands, wetlands and unfertilized extensively managed grasslands). The GHG budgets of the forests were strongly dominated by CO2 exchange, while CH4 and N2O exchange comprised a larger proportion of the GHG balance in short semi-natural vegetation. Uncertainties in elemental budgets were much larger for nitrogen than carbon, especially at sites with elevated Ndep where Nr leaching losses were also very large, and compounded by the lack of reliable data on organic nitrogen and N2 losses by denitrification. Nitrogen losses in the form of NO, N2O and especially NO3- were on average 27 % (range 6 %–54 %) of Ndep at sites with Ndep < 1 g N m−2 yr−1 versus 65 % (range 35 %–85 %) for Ndep > 3 g N m−2 yr−1. Such large levels of Nr loss likely indicate that different stages of N saturation occurred at a number of sites. The joint analysis of the C and N budgets provided further hints that N saturation could be detected in altered patterns of forest growth. Net ecosystem productivity increased with Nr deposition up to 2–2.5 g N m−2 yr−1, with large scatter associated with a wide range in carbon sequestration efficiency (CSE, defined as the NEP ∕ GPP ratio). At elevated Ndep levels (> 2.5 g N m−2 yr−1), where inorganic Nr losses were also increasingly large, NEP levelled off and then decreased. The apparent increase in NEP at low to intermediate Ndep levels was partly the result of geographical cross-correlations between Ndep and climate, indicating that the actual mean dC∕dN response at individual sites was significantly lower than would be suggested by a simple, straightforward regression of NEP vs. Ndep.

The understanding of biogenic volatile organic compound (BVOC) exchanges has become a key scientific issue because of their high reactivity and their impact in the atmosphere. However, so far, few studies have focused on BVOCs exchanged by agricultural species, and in particular by winter wheat, despite this species being the leading worldwide crop in terms of harvested area. This study for the first time investigated BVOC exchanges from winter wheat during most developmental stages of the plant. Fluxes were measured in Belgium at the ecosystem-scale using the disjunct eddy covariance by mass scanning technique, and a proton-transfer-reaction mass spectrometer for BVOC ambient mixing ratio measurements. As is usually observed for crops and grasses, the winter wheat field emitted mainly methanol, although bi-directional exchanges were observed. The second most exchanged compound was acetic acid which was captured during the entire growing season. Bi-directional exchanges of acetaldehyde and acetone were also reported. Terpene exchanges were 22 times smaller than oxygenated VOC (OVOC) exchanges. For all compounds, the exchanges were the most pronounced at the end of the growing season, i.e., under warm, dry and sunny conditions. Senescence-induced emissions were furthermore observed for methanol and acetaldehyde. For all investigated OVOCs, the exchanges very likely originated from both the soil and the plants. Despite their mixed origin, the MEGAN (Model of Emissions of Gases and Aerosols from Nature) v2.1 up-scaling model could adequately reproduce the methanol, acetaldehyde and acetone exchanges measured at this site during the mature and senescence phases of the plant, when the standard emission factor and the leaf age factor were adapted based on the measurements. In contrast, the model failed to reproduce the measured acetic acid exchanges. When the standard emission factor values currently assigned in MEGAN were applied, however, the exchanges were largely over-estimated for all compounds.

Using long-term measurements in a mature beech (Fagus sylvatica L.)-dominated forest located in east Belgium, this paper showed that spring and autumn temperature increases during the last two decades led to an earlier end and a shortening of the growing season. These phenological shifts impact negatively but not significantly the forest annual net ecosystem productivity. The mechanisms controlling temperate forest phenology are not fully understood nor are the impacts of climate change and the consequences for forest productivity. The aim of this paper is to contribute to the understanding of how temperate forest phenology and net ecosystem productivity (NEP) interplay and respond to temperature and its evolution. Indicators of leaf phenology and productivity dynamics at the start and the end of the growing season, as well as combinations of these indicators (length of the growing season), were derived from a long-term (1997–2014) dataset of eddy covariance and light transmission measurements taken over a mature beech-dominated temperate forest. The start and the end of the growing season were correlated to spring (and autumn, for the end) temperatures. Despite no trends in annual average temperatures being detected during the observation period, April and November temperatures significantly increased. As a result, an earlier but slower start and an earlier end, inducing a shorter length of the growing season, were observed over the studied period. The first shift positively impacts the mixed forest NEP but is mainly related to the presence of conifers in the subplot. The earlier end of the growing season, more related to beech phenology, negatively impacts the forest NEP. Overall, these two effects partially compensate each other, leading to a non-significant impact on NEP. Increasing temperatures over the 18-year studied period shortened the growing season length, without affecting significantly the mixed forest NEP. However, as beeches are only affected by the earlier end of the growing season, this suggests a phenologically driven beech productivity reduction in the forest.

Using the eddy covariance technique, half-hourly N2O fluxes were measured over a sugar beet crop (ICOS Station, Lonzee, BE) from fertilization to harvest. Several parameters of the data quality control tests were adapted to suit the characteristics of N2O, No u* filtering threshold could be seen for N2O fluxes; therefore, it was determined based on CO2 data. The uncertainty of N2O fluxes was assessed for several aspects of data treatment (total random uncertainty, spectral correction, u* filtering, gap-filling), which were combined to determine the uncertainty of the N2O budget. Between fertilization and harvest, the crop emitted 1.83 ( +/- 0.21) kg N2O-N ha(-1) corresponding to 1.2% of N supplies. Flux variability was characterized by three episodes of high emissions across the experiment, interspersed with lower background fluxes. These peak events were driven by soil moisture and temperature, dependent on the time-scale. Soil water content at 5 cm was identified as the single trigger for N2O emission peaks given sufficient N availability, while intraday oscillations were positively correlated to the variations in surface temperature rather than deeper soil temperatures. For the first time, an inhibiting and short-term effect of topsoil disturbance (seed-bed preparation) on N2O fluxes was recorded, which interrupted the peak that followed fertilization, and delayed the start of the next high emission episode. This observation, along with the synchronicity found between surface temperature and diel oscillations of N2O fluxes, supports the hypothesis of a N2O-producing microbial community located in the topmost soil layer. Given that a third of the overall N2O emissions during the measurement campaign occurred between fertilization and seed-bed preparation, further investigation into the timing of farming operations as mitigation strategies is needed. The contribution of N2O emissions to the net greenhouse gas balance (which comprises CO2 and N2O fluxes) was estimated at between 20 and 66%. Our results stress the importance of including nitrous oxide when measuring gas exchanges in fertilized crops, and to do so at high temporal resolution for improved estimates.

The eddy covariance (EC) technique has been widely used to quantify the net CO2 ecosystem exchange (NEE) of grasslands, which is an important component of grassland carbon and greenhouse gas budgets. In free range grazed pastures, NEE estimations are supposed to also include cattle respiration. However, cattle respiration measurement by an EC system is challenging as animals act as moving points emitting CO2 that are more or less captured by the EC tower depending on their presence in the footprint. Often it is supposed that, over the long term, cattle distribution in the pasture is homogeneous so that fluctuations due to moving sources are averaged and NEE estimates are reasonably representative of cattle respiration. In this study, we test this hypothesis by comparing daily cow respiration rate per livestock unit (LU) estimated by postulating a homogeneous cow repartition over the whole pasture with three other estimates based on animal localization data, animal scale carbon budget and confinement experiments. We applied these methods to an intensively managed free range grassland and showed that the NEE estimate based on a homogeneous cow repartition was systematically lower than the three other estimates. The bias was about 60 g C m–2 yr–1, which corresponded to around 40% of the annual NEE. The sign and the importance of this bias is site specific, as it depends on cow location habits in relation to the footprint of the EC measurements which highlight the importance of testing the hypothesis of homogeneity of cattle distribution on each site. Consequently, in order to allow estimating the validity of this hypothesis but also to improve inter site comparisons, we advocate to compute separately pasture NEE and grazer’s respiration. For the former we propose a method based on cattle presence detection using CH4 fluxes, elimination of data with cattle and gap filling on the basis of data without cattle. For the second we present and discuss three independent methods (animal localization with GPS, animal scale carbon budget, confinement experiments) to estimate the cattle respiration rate.

The Integrated Carbon Observation System Re-search Infrastructure aims to provide long-term, continuous observations of sources and sinks of greenhouse gases such as carbon dioxide, methane, nitrous oxide, and water vapour. At ICOS ecosystem stations, the principal technique for measurements of ecosystem-atmosphere exchange of GHGs is the eddy-covariance technique. The establishment and setup of an eddy-covariance tower have to be carefully reasoned to ensure high quality flux measurements being representative of the investigated ecosystem and comparable to measurements at other stations. To fulfill the requirements needed for flux determination with the eddy-covariance technique, variations in GHG concentrations have to be measured at high frequency, simultaneously with the wind velocity, in order to fully capture turbulent fluctuations. This requires the use of high-frequency gas analysers and ultrasonic anemometers. In addition, to analyse flux data with respect to environmental conditions but also to enable corrections in the post-processing procedures, it is necessary to measure additional abiotic variables in close vicinity to the flux measurements. Here we describe the standards the ICOS ecosystem station network has adopted for GHG flux measurements with respect to the setup of instrumentation on towers to maximize measurement precision and accuracy while allowing for flexibility in order to observe specific ecosystem features.

Research infrastructures play a key role in launching a new generation of integrated long-term, geographically distributed observation programmes designed to monitor climate change, better understand its impacts on global ecosystems, and evaluate possible mitigation and adaptation strategies. The pan-European Integrated Carbon Observation System combines carbon and greenhouse gas (GHG; CO2, CH4, N2O, H2O) observations within the atmosphere, terrestrial ecosystems and oceans. High-precision measurements are obtained using standardised methodologies, are centrally processed and openly available in a traceable and verifiable fashion in combination with detailed metadata. The Integrated Carbon Observation System ecosystem station network aims to sample climate and land-cover variability across Europe. In addition to GHG flux measurements, a large set of complementary data (including management practices, vegetation and soil characteristics) is collected to support the interpretation, spatial upscaling and modelling of observed ecosystem carbon and GHG dynamics. The applied sampling design was developed and formulated in protocols by the scientific community, representing a trade-off between an ideal dataset and practical feasibility. The use of open-access, high-quality and multi-level data products by different user communities is crucial for the Integrated Carbon Observation System in order to achieve its scientific potential and societal value.

Plants, under stressful conditions, can proceed to photosynthetic adjustments in order to acclimatize and alleviate the detrimental impacts on the photosynthetic apparatus. However, it is currently unclear how adjustment of photosynthetic processes under environmental constraints by plants influences CO2 gas exchange at the ecosystem-scale. Over a 2-year period, photosynthetic performance of a temperate grassland ecosystem was characterized by conducting frequent chlorophyll fluorescence (ChlF) measurements on three primary grassland species (Lolium perenne L., Taraxacum sp., and Trifolium repens L.). Ecosystem photosynthetic performance was estimated from measurements performed on the three dominant grassland species weighed based on their relative abundance. In addition, monitoring CO2 fluxes was performed by eddy covariance. The highest decrease in photosynthetic performance was detected in summer, when environmental constraints were combined. Dicot species (Taraxacum sp. and T. repens) presented the strongest capacity to up-regulate PSI and exhibited the highest electron transport efficiency under stressful environmental conditions compared with L. perenne. The decline in ecosystem photosynthetic performance did not lead to a reduction in gross primary productivity, likely because increased light energy was available under these conditions. The carbon amounts fixed at light saturation were not influenced by alterations in photosynthetic processes, suggesting photosynthesis was not impaired. Decreased photosynthetic performance was associated with high respiration flux, but both were influenced by temperature. Our study revealed variation in photosynthetic performance of a grassland ecosystem responded to environmental constraints, but alterations in photosynthetic processes appeared to exhibit a negligible influence on ecosystem CO2 fluxes.

Soil methanol (CH3OH) exchange is often considered as several orders of magnitude smaller than plant methanol exchange. However, for some ecosystems, it is significant in regard with plant exchange and worth thus better consideration. Our study sought to gain a better understanding of soil exchange. Methanol flux was measured at the ecosystem scale on a bare agricultural soil over two contrasted periods using the disjunct eddy covariance by mass scanning technique. A proton-transfer-reaction mass spectrometer was used for the methanol ambient mixing ratio measurements. Bi-directional exchange dynamics were observed. Methanol emission occurred under dry and warm conditions and correlated best with soil surface temperature, whereas methanol uptake occurred under wet and mild conditions and correlated well with the methanol ambient concentration. After having tested a physical adsorption-desorption model and by confronting our data with the literature, we propose that the exchange was ruled by both a physical adsorption/desorption mechanism and by a methanol source, which still needs to be identified. The soil emission decreased when the vegetation developed. The reasons for the decrease still need to be determined. Overall, the dynamics observed at our site were similar to those reported by other studies for both cropland and forest ecosystems. The mechanism proposed in our work can thus be possibly applied to other sites or ecosystems. In addition, the methanol exchange rate was in the upper range of the exchange rates reported by other soil studies, suggesting that cropland soils are more important methanol exchangers than those in other ecosystems and should therefore be further investigated.

Eddy covariance is increasingly used to monitor cattle emissions. However, the turbulent flux calculation method and the footprint models upon which calculations are based are insufficiently validated. In addition, available footprint models presume the source to be placed at soil height, which is obviously not the case for cattle. The present study uses a single known artificial point source placed at cow's muzzle height in order to assess the impact of the flux calculation method (averaging method, averaging period, quality filters) and of the footprint model on the emission estimates. The optimal calculation method and footprint model combination (running mean, 15 min averaging periods, no application of the Foken and Wichura (1996) stationarity filter, and the use of the Kormann and Meixner (2001) footprint function) led to estimated emissions between 90 and 113% of the true emission, leading to the conclusion that the use of eddy-covariance for point-source emission estimation is feasible provided an adequate calculation method is selected.

Grassland carbon budgets are known to be greatly dependent on management. In particular, grazing is known to directly affect CO2 exchange through consumption by plants, cattle respiration, natural fertilisation through excreta, and soil compaction. This study investigates the impact of two grazing methods on the net ecosystem exchange (NEE) dynamics and carbon balance, by measuring CO2 fluxes using eddy covariance in two adjacent pastures located in southern Belgium during a complete grazing season. Rotational (RG) grazing consists of an alternation of rest periods and short high stock density grazing periods. Continuous grazing (CG) consists of uninterrupted grazing with variable stocking rates. To our knowledge, this is the first study to assess the impact of these grazing methods on total net ecosystem exchange and CO2 exchange dynamics using eddy covariance. The results showed that NEE dynamics were greatly impacted by the grazing method. Following grazing events on the RG parcel, net CO2 uptake on the RG parcel was reduced compared to the CG parcel. During the following rest periods, this phenomenon progressively shifted towards a higher assimilation for the RG treatment. This behaviour was attributed to sharp biomass changes in the RG treatment and therefore sharp changes in plant photosynthetic capacity. We found that differences in gross primary productivity at high radiation were strongly correlated to differences in standing biomass. In terms of carbon budgets, no significant difference was observed between the two treatments, neither in cumulative NEE, or in terms of estimated biomass production. The results of our study suggest that we should not expect major benefits in terms of CO2 uptake from rotational grazing management when compared to continuous grazing management in intensively managed temperate pastures.

This study presents twenty years of Net Ecosystem Productivity estimations obtained using eddy covariance in a mixed forest, dominated by beech with sparse conifers, at the Vielsalm station, in the Belgian Ardennes. First the quality and reliability of the data set is discussed. An uncertainty analysis showed that if, on one hand, the site heterogeneity and set-up changes may strongly affect yearly NEP estimates, questioning thus the total carbon budget relevance, on the other hand, robust inter-annual anomalies may be obtained as long as a site dedicated data treatment is carefully applied. A validation of the anomalies by comparison with a growth index derived from tree ring measurements is given. The resulting anomalies (range: [-206; +123] g C m(-2) yr(-1), standard deviation: 93g C-2 yr(-1)) being larger than their own uncertainty (similar to 30 g C m(-2) yr(-1)), an inter-annual variability analysis is possible. This analysis shows that the sources of NEP inter-annual variability at the Vielsalm station are multiple but the most prominent causes are biotic processes driven by carry-over effects of preceding meteorological events. The lowest observed NEP, in 2000, resulted from a bark beetle attack probably prompted by an early frost event in 1998. Besides, the robust lagged correlation between NEP anomalies and mean vapor pressure deficit during the preceding vegetation season also suggests a carry-over effect of water limitation during the previous year on the beech NEP. Mechanisms driving this carry-over effect are supposedly linked to tree physiology, which is confirmed by a dependency of canopy photosynthetic capacity to previous year water limitation. Some hypotheses, involving biomass allocation and bud formation, are proposed to explain its lagged impact on canopy photosynthetic capacity. Other causes of NEP inter-annual variability are the radiation during the current vegetation season and the temperature at the end of the winter but the latter variable rather indicates an effect on the conifers interspersed in the plot. Overall, the photosynthetic capacity combined with these two factors explained about 75% of NEP inter-annual variability.