The Integrated Carbon Observation System in Europe

,

are the primary cause of the ongoing climate warming (IPCC 2019).The atmospheric buildup of CO 2 would have been about twice as large had approximately half of the carbon emitted to the atmosphere not been sequestered by ocean and land sinks, leading to the rate of warming being reduced (Friedlingstein et al. 2020).However, the size, nature, and stability of these sinks are uncertain, which, together with the uncertainty of the speed of release of the heat stored in the ocean surfaces, leads to large uncertainties in the projected global climate warming with different GHG mitigation scenarios (Ma et al. 2020;Rhein et al. 2013).Improving the quantification and reducing the uncertainty of these projections is important to support policy-making and the size and timing of reductions in global emissions.There are uncertainties in the emission sources, but the largest cause for uncertainty in the global carbon budget estimates are likely due to the lack of understanding of land and ocean sinks (Friedlingstein et al. 2020).Understanding of these sinks, sources, and the related processes can only be achieved with research based on spatially and temporally comprehensive and precise data.This is ever more important now that a specific goal has been set at limiting average global surface temperature increases to well below 2°C above preindustrial levels: the Paris Agreement (United Nations 2015a) was ratified by 189 countries to guide the actions to combat climate change.
Anthropogenic emissions of GHGs to the atmosphere are superimposed on with the much larger natural GHG exchange fluxes between the atmosphere and the terrestrial ecosystems and ocean, which are further affected by ongoing climate warming.Quantifying the anthropogenic perturbation therefore depends on quantifying both natural and anthropogenic emissions and sinks and understanding the drivers of feedback mechanisms over both.
The Integrated Carbon Observation System (ICOS), which currently includes over 140 stations, was designed as the European in situ observation and information system to support science and society in their efforts to mitigate climate change.ICOS is motivated by understanding the sources, sinks, and cycling of greenhouse gases in the atmospherebiosphere-hydrosphere continuum.The European Commission, Belgium, Finland, France, Germany, Italy, the Netherlands, Norway, Sweden, and Switzerland committed to this mission when the ICOS European Research Infrastructure Consortium (ERIC) was established in 2015.
Key aspects of climate science addressed by ICOS have been elaborated on in earlier articles, with Schulze et al. (2009) emphasizing the importance of N 2 O and CH 4 in the European greenhouse gas budget, Peters et al. (2010) quantifying European net terrestrial CO 2 exchange, Gielen et al. (2017) briefly summarizing the different components of the network, Franz et al. (2018) giving an overview on ICOS ecosystem observations, Steinhoff et al. (2019) describing the ocean network, and Levin et al. (2020) addressing the atmospheric network.This article provides a comprehensive overview of the ICOS Research Infrastructure (RI), including a historical overview, describing the structure, operations, and financial sustainability of the ICOS RI, elaborating present and future scientific questions, and discussing lessons learned and challenges addressed by ICOS.

The rationale and path into Integrated Carbon Observation System
Even though the connection between human actions and climate change had been made by the end of the twentieth century (IPCC 1992), many important questions were still open, such as how much CO 2 from fossil fuel burning remains in the atmosphere and how much was taken up by oceans and terrestrial ecosystems (Keeling 1978).A major obstacle in answering these questions was limited data availability and the use of different observational methods, units, and scales by different countries and sites.This required global harmonization of observations, first started in the atmosphere by the World Meteorological Organization Global Atmosphere
Another obstacle was how to draw conclusions from various pieces of data and information.This called for a framework how to systematically provide scientific knowledge in global scale, giving birth to the Intergovernmental Panel on Climate Change (IPCC), established in the end of 1980s.Eyes turned next to land, where various methods had been developed to understand highly diverse and complex terrestrial ecosystems.This posed challenges to compare the results, and the Global Climate Observing System (GCOS) was established to harmonize terrestrial observations and to define a set of Essential Climate Variables (GCOS 1994(GCOS , 2016;;WMO 2009).Quantifying relatively small long-term trends in CO 2 and other GHG concentration and fluxes against a background of much larger short-term variations caused by the "natural" carbon cycle requires highly precise and accurate observations.To decrease uncertainties by improving the quality of observations, and to draw general conclusions, research-and investigator-based European ecosystem networks, with foci on CO 2 , energy, and water exchange, emerged in the 1990s with the support of the European Commission funding programs (EuroFlux, CarboEurope IP, and GHG Europe).During 1998-2002, the EuroFlux network covered 30 stations mainly in forest ecosystems across Europe (Janssens et al. 2003), which later developed into the network of ecosystem stations within ICOS.
At the beginning of the 1990s, the Global Ocean Observing System (GOOS) was established to coordinate and harmonize ocean observations together with GCOS.The scientific community undertook the task to provide open access to global ocean surface CO 2 data via the Surface Ocean CO 2 Atlas (SOCAT; Pfeil et al. 2013).These data are essential to estimate ocean carbon budget and acidification.As a community effort, SOCAT depends heavily on voluntary data submission and secondary quality control, and the Ocean Carbon Data System of the National Oceanic and Atmospheric Administration and ICOS support SOCAT and contribute significantly to its data operations and development.
The development of observation networks had been fragmented into various projects in Europe (see Fig. 1 in Franz et al. 2018).By the beginning of the 2000s, it was possible to estimate the European terrestrial carbon budget by either using the few ecosystem network data available (e.g., Papale and Valentini 2003) or by methods using atmospheric network data, but these provided dissimilar and highly uncertain results ( Ja nssens et a l.2003).The results suggested that increase in ecosystem representation and data would reduce the uncertainty in the bottom-up approach and that including more atmospheric stations would improve the accuracy of topdown estimates.T h e E U -f u n d e d CarboEurope Integrative Project (2004-08), was a

E859
major step toward integrated studies, harmonized observations, and data flows, covering atmosphere and ecosystem sciences (Schulze et al. 2009).In parallel, the CarboOcean IP conducted over 2005-09 developed systematic ocean carbon observations and analysis across Europe.The observations collected in the context of these projects were an important example to demonstrate how a large and coordinated network could provide a unique dataset valuable for the modeling activities to estimate continental-scale GHG fluxes (e.g., Luyssaert et al. 2010;Schulze et al. 2009;Vetter et al. 2008).
European countries have been at the forefront of setting up the Paris Agreement to reduce emissions.Implementation of climate change mitigation is done by individual nations, but to effectively curb the increase of GHG concentrations in the future, a comprehensive strategy of emission reductions and natural sink conservation must be designed collectively.The success of the scientific projects showing capability of the scientific community to provide quantitative information at a European scale paved way for the political will to develop ICOS-an observation system that will narrow down future uncertainties and provide observational evidence of the current state of the carbon cycle perturbation.Throughout the development of ICOS, the policy-makers, funders, and scientists have been in constant dialogue to improve the scientific foundation of decision-making and obtain the political and financial commitments across European countries.
The ICOS foundation required negotiating the concept for such as system, with clear purpose and governance as well as financial structure and responsibilities of each participants, in which the countries could then commit.This was the purpose of ICOS preparatory phase project in 2008-13 (FP7 project 211574, see also appendix A).
A user-centric approach drove the development of a centralized data provision hub for all ICOS data, the Carbon Portal.The problem of different types of observations in atmosphere, ecosystem, and ocean stations was addressed by centralizing the quality control and data processing in three respective Thematic Centres with specific experience and knowledge.To allow measurements of required precision, the Central Analytical Laboratories was designed to provide calibration gases to atmospheric and ocean monitoring stations.The process for scientific development was planned on the interactions between these components and the monitoring station assemblies (MSAs), which include all station principal investigators (PIs).
The financial challenges were tackled by acquiring commitments from various countries interested to build a national network of ICOS stations or propose a central facility.The host countries provide the majority of the financial support by direct governmental grants (ICOS Ecosystem Thematic Centre is hosted by Italy, France, and Belgium; Atmosphere Thematic Centre by France and Finland; Ocean Thematic Centre by Norway and the United Kingdom; the Central Analytical Laboratories by Germany; and the Carbon Portal by Sweden and the Netherlands).The stations are maintained by individual countries, and each country also contributes to the general costs for the upkeep of the RI.The principles for sharing the financial responsibilities were written in the ICOS financial rules.
With scientific, technical, and financial concepts in place, the last challenge was how to coordinate such an infrastructure across many countries.The solution was to establish a legal entity designed to manage Research Infrastructures and recognized in all European countries, called ICOS European Research Infrastructure Consortium (ERIC; https://ec.europa.eu/info/research-and-innovation/strategy/european-research-infrastructures/eric_en), hosted by Finland and France with participation from all member countries), to coordinate the whole research infrastructure and to report to and consult with the ministerial stakeholders.
The mission of ICOS is to harmonize European carbon and GHG observations, ensure the related long-term financial commitments, provide easy access to well-documented and reproducible high-quality data and related protocols and tools for scientific studies, and to deliver GHG-related products to stakeholders in society and policy.The first five years of ICOS from 2015 to 2019 focused on establishing an operational infrastructure, and as an acknowledgment of successful implementation, ICOS ERIC was included in the European Strategy Forum on Research Infrastructures' strategy as a landmark infrastructure (ESFRI 2016).Since becoming operational in 2015, the Czech Republic, Denmark, the United Kingdom, and Spain have joined ICOS, and Poland has announced to join ICOS, considerably expanding the network; in addition, negotiations are currently under way with Estonia, Greece, Hungary, Ireland, Portugal, and Romania.The second phase of ICOS, described in the ICOS strategy published in 2019 (ICOS 2019), and its associated implementation plan, will place emphasis on the use of data and on enhancing the network's capability to analyze anthropogenic impacts on the carbon cycle.We foresee the new European Green Deal launched by the EU Commission at the end of 2019 (https://ec.europa.eu/info/strategy/priorities-2019-2024/european-green-deal_en),designed to make Europe the first net-zero continent, will further strengthen ICOS' role in the forthcoming years.

Description of ICOS observations and data
ICOS provides the core network of highly accurate, long-term European in situ observations of carbon and GHGs (see appendix B for full list of observed variables).The terrestrial network of over 100 stations ranges from Sweden (latitude 68°N in WGS84 coordinates) to the Mediterranean Sea (latitude 36°N), from the United Kingdom (longitude 3°W) to Finland (longitude 30°E), from the lowlands near sea level to alpine regions (2168 m MSL, Italy), with stations also outside of Europe (e.g., in French Guyana, Greenland, and the Democratic Republic of the Congo).The marine network of over 20 stations and vessels extends from polar areas to the equator and from coasts to open ocean (Fig. 1).
The atmospheric network of tall towers, mountain, and coastal stations covers large parts of Europe with continuous measurements of CO 2 and CH 4 mole fractions.When coupled with an atmospheric model, these data provide an integrated view of all natural and anthropogenic fluxes.In fully equipped ICOS stations, meteorological variables, N 2 O and 222 Rn are observed to link concentration variations to atmospheric mixing.In addition, N 2 O, SF 6 , H 2 , and for CO 2 source apportionment CO, 13 C-CO 2 , 14 C-CO 2 , 18 O-CO 2 , and the O 2 :N 2 ratio, are analyzed in air sampled by automated flask samplers at the most extensively equipped stations, called class 1 stations (ICOS 2017a; Levin et al. 2020;appendix B).
GHG fluxes in different terrestrial ecosystems (forests, croplands, grasslands, mires, wetlands, shrublands, lakes, Mediterranean savannas, urban sites) are observed at comprehensively equipped stations to quantify the exchange of carbon, GHGs, and energy between the atmosphere and the ecosystems (Franz et al. 2018), by using the eddy covariance technique (Rebmann et al. 2018).Biosphere-atmosphere exchange measurements at flux towers represent the only direct method to provide detailed data at ecosystem scale, and they are valuable also for different user communities: e.g., sensible and latent heat fluxes are important to understand the water cycle and to improve weather forecasts, and turbulence data are used in studying boundary layer physics.In a subset of stations, fluxes of CH 4 and N 2 O are observed (Nemitz et al. 2018).Complementary data comprise, e.g., soil organic carbon content, green area index, litterfall, aboveground biomass, records of disturbances, and vegetation properties such as leaf nutrients and phenological status, as well as management activities (Arrouays et al. 2018;Gielen et al. 2018;Hufkens et al. 2018;Loustau et al. 2018;Op de Beeck et al. 2018;Pavelka et al. 2018;Saunders et al. 2018).These various observations are used to estimate the contributions of different components of the ecosystem, such as soil or vegetation, to the seasonal and interannual variability of the carbon and GHG budget of the whole ecosystem, as well as when upscaling carbon fluxes to regional and global scales (Jung et al. 2020).
The ocean observations are conducted either on fixed platforms (e.g., moorings and surface buoys) or on ships operating predominantly in the North Atlantic, Nordic, Baltic, and

E861
Mediterranean Seas, but occasionally also in polar regions and the equatorial Atlantic.Partial pressure of sea surface CO 2 is used in conjunction with other parameters (temperature, salinity, mixed layer depth) and satellite remote sensing products including wind fields and chlorophyll to calculate oceanic uptake of CO 2 (ICOS 2017b(ICOS , 2020b;;Steinhoff et al. 2019).Other carbon cycle parameters (pH, alkalinity, dissolved inorganic carbon) and related properties such as nutrients and oxygen are used to investigate ocean transports and controls over carbon uptake.This latter work involves collaboration across various elements of the European RI landscape and components of the Global Ocean Observing System (GOOS) including Euro-Argo (European Consortium for Operating Argo Floats), EMSO (the European Multidisciplinary Seafloor and Water Column Observatory), and GO-SHIP (the Global Ocean Ship-Based Hydrographic Investigations Program).While the main policy framework that ICOS contributes to is the Paris Agreement, Agenda 2030 (United Nations 2015b) and its Sustainable Development Goal 14.3 are also supported with monitoring of ocean acidification.
The variables and related costs for all ICOS observations are detailed in the ICOS Handbook, released every 2 years (ICOS 2020a).For example, to build one fully equipped atmosphere or ecosystem station (class 1) costs between 0.5 and 1 million euros (not including personnel costs), whereas costs for stations having only a subset of observations (class 2) can be between 0.1 and 0.3 million euros.Maintenance requirements and collection of ancillary data are also provided, with the most significant component being person-power, ranging from 0.3 to 4 full-time equivalents per annum depending on the type of the station.
Quality assurance and quality control (QA/QC) Within the CarboEurope IP project (Schulze et al. 2010) the challenges to achieve highly compatible atmospheric data became obvious.Large efforts were undertaken to assess the compatibility of atmospheric measurements done by different laboratories at different observational sites.Yet, results from these exercises repeatedly yielded evidence that the WMO compatibility goals were not met by all participants and for all tracers.Biases between laboratories could sometimes be of the same order of magnitude as the atmospheric signals that should be captured, and it was not possible to define a network data compatibility.This was motivating the ICOS concept with highly standardized measurement approaches at the observatories (including aspects such as instrumentation, procedures to account for atmospheric humidity, and calibration procedures) and the establishment of central facilities that assess the adequate performance of all installed analyzers and assure transparent data processing [Atmosphere Thematic Centre (ATC)], as well as the consistency of sample measurement results and reference gas assignments that are used within the monitoring network [Central Analytical Laboratories (CAL)].To have the ability to make a defendable uncertainty assessment that is required for observational data (WMO 2020b), the following QA/QC approaches are applied that cover all levels of the observational system (stations as well as central facilities).To minimize uncertainties in both observations provided by single stations and studies using data from multiple stations, several steps are taken.The instruments themselves have strict requirements, they are systematically calibrated, their setup is based on stringent protocols, and the data are processed by the Thematic Centres with proven and standardized methodologies (Hazan et al. 2016;El Yazidi et al. 2018;Vitale et al. 2020).
Scientists in the ICOS atmosphere community, coordinated via the Atmospheric MSA, have defined and approved protocols for instrumentation setup and sampling strategies (ICOS 2017a; Levin et al. 2020) to ensure that atmospheric measurements comply with the compatibility goals set by the WMO for measurements of major GHGs and associated tracers (WMO 2020b).Stringent network compatibility within ICOS and with other networks is key when using the observations in concert with atmospheric transport models to quantify GHG sources and sinks.Calibration gases are prepared and calibrated centrally for the network

E862
by the Flask and Calibration Laboratory (FCL) of the CAL, which maintains tight links to the WMO Central Calibration Laboratory to ensure the traceability of ICOS data to internationally accepted WMO calibration scales by one unique path.To assess the accuracy of the implementation of these scales, the FCL maintains several ongoing round robin exercises with the NOAA laboratories as quality control.The FCL is also responsible for flask analyses except for 14 C-CO 2 , which is analyzed by the Central Radiocarbon Laboratory of the CAL.The precision and stability of all GHG analyzers are tested at the Atmosphere Thematic Centre prior to deployment (Yver Kwok et al. 2015).A comprehensive overview of the optimization of the quality management as part of the labeling process of atmosphere stations is given in Yver- Kwok et al. (2021).For quality control of the continuous in situ measurements, automated QC figures are generated on a daily level by the ATC that summarize the statistics of the measurement precision (repeatability and target gas bias), which provide a basis to quantify the measurement uncertainty.Additional auditing is made with traveling instrumentation (ICOS Mobile Laboratory) operated at selected stations for a couple of weeks, and by ongoing comparison of flask results with in situ observations at class 1 stations (Levin et al. 2020).
To ensure high quality of observations in diverse ecosystems, with various drivers influencing the carbon and GHG fluxes, the observation methods need careful attention.Since diverse observation methods were established for different climate regions and ecosystem types in the past decades, a community-driven effort was necessary to define key and ancillary components to be observed in each ecosystem type to analyze carbon and GHG fluxes.Also, much effort has been put into defining the specifications and methodology of observations by the community, together with the Ecosystem MSA and the Ecosystem Thematic Centre.Both optimal sets of variables and practical feasibility were considered when harmonizing the observations, which resulted in a compromise suitable for high-quality and long-term continental-scale observation system (Franz et al. 2018).Over 100 scientists' efforts were acknowledged in a set of publications describing the ecosystem measurement protocols in 2018 (see special issue of International Agrophysics, 2018, Vol. 32, No. 4).Starting from the protocols, more practical and detailed instruction documents were prepared and published by the ETC (http://www.icos-etc.eu/documents/instructions)that are revised and updated regularly, following the newest developments and knowledge.
For the ocean observations, the major challenges are the complexity of the carbonate system, the often remote location of stations, and suitability of different observing methods for different types of stations.Tailored solutions are needed in order for each station to deliver the best possible data, and the ICOS ocean community, supported by the Ocean Thematic Centre, has adopted and adapted existing and proven best-practice guidelines and protocols (Dickson et al. 2007) for observations made by different types of stations (Steinhoff et al. 2019).ICOS is the first multinational entity within the marine community that has standardized CO 2 observations (Steinhoff et al. 2019).The Fixed Ocean Stations' maintenance and calibration are done during the visit by research vessels, ideally several times per year, whereas observations on Ships of Opportunity (SOOP) are calibrated even more frequently.Inclusion of marine towers with direct flux observations is currently under development (Steinhoff et al. 2019).CO 2 observations are calibrated with standards traceable to the WMO calibration scales (ICOS 2020b).Data quality, control, and uniformity is also supported by a customized QuinCe tool developed by the OTC.
To guarantee the quality of observations in all three network components (atmosphere, ecosystem, and ocean), it is necessary for a station to pass an ICOS station certification process.Here, the station characteristics are evaluated, its compliance with measurement protocols and standards is analyzed, and data transfer and quality are evaluated by the respective Thematic Centres during a test measurement period of a few months.After successfully completing this process, which typically takes 2-3 years, the station receives the E863 ICOS certificate.This means the station meets the high standards of the ICOS network and continuously provides ICOS data.Of the over 140 stations in the ICOS network, over 60 stations have been certified by the end of 2020.
Open data access ICOS has addressed the major challenge of data access and simplification of data use (Fig. 2) thanks to the PI and Central Facilities' work that agreed on a continuous data submission and adoption of an open data license (Creative Commons Attribution 4.0 International, which also allows commercial use).Additionally, the services provided to make data distribution easier and assignment of digital object identifier (DOI) to datasets are major advances to improve and promote open data.
To serve various user needs, different levels of the data are openly accessible with different levels of processing and quality check.Much attention has been paid to the metadata that follow the specifications defined by the Carbon Portal in collaboration with the Thematic Centres, also considering existing international standards.All steps of data flows were designed based on the Findable, Accessible, Interoperable, Reusable (FAIR) principles (Wilkinson et al. 2016), giving the user sufficient tools to interpret the data (ICOS 2015).
Different levels of data are stored throughout the process, from raw sensor data (level 0), to the automatically calibrated near-real-time data (level 1; available within 24 h of the measurement) to the final, quality-checked data (level 2).All the data are passed on to the Carbon Portal, which provides free and open access to ICOS data.The data are minted with Persistent Identifiers to provide unique identification and citation of the datasets and their contributors (ICOS 2019).The Carbon Portal offers tailor-made tools and services (Fig. 3) and distributes products (level 3) that are created by the scientific community based on ICOS data and possibly from other data sources (Fig. 4).The Carbon Portal has started to develop and provide tools for online analysis of data and model results (see, e.g., ICOS 2020c).These enable transparent analyses of data by station PIs, interactive collaboration with the data users, and utilization of cloud services as virtual working environments.
Major scientific questions and glimpse to the future Many major scientific questions have guided the development of systematic, continental observations of carbon and GHG budgets.Scientists have been able to answer how much

E864
of emitted CO 2 from fossil fuels have accumulated in the atmosphere, oceans, and terrestrial ecosystems (Friedlingstein et al. 2020).Many advancements have been made in defining how terrestrial ecosystems are affected by and how they feed back to climate change, e.g., by changes in evapotranspiration or albedo.
With the ICOS network reaching maturity via station certification, the compilation of the European carbon and GHG budget, which was previously possible as one-time effort (Schulze et al. 2009), can soon be produced annually at high spatial resolution and with reduced uncertainty.This is a significant step forward in assessing changes and trends on the continental scale.Advancements have been made in providing detailed information on the dominantly studied ecosystems, e.g., forests, grasslands, and croplands, while we still have only rudimentary understanding of some other ecosystem types, e.g., lakes, rivers, peatlands, Mediterranean savannas, and Arctic tundra (Baldocchi 2014;Schulze et al. 2010), or on urban systems.Mitigation capacity of urban areas as well as their adaptation capacity will need much deeper attention as the urban population is continuously growing and urban areas represent the major sources of GHGs in Europe and in most of the continents (Calfapietra et al. 2015).

E865
ICOS data are widely used in publications from various scientific fields.The number of ICOS-related publications per year have increased from 30 in 2012 to roughly 200 in 2020, and the citations from 600 to 11,000, respectively (https://www.icos-cp.eu/science-and-impact/societyimpact/references).The publications are associated to almost 60 categories, with the two largest being meteorology and atmospheric sciences (37% of all publications) and environmental sciences (34%) (ICOS 2021).The cross-domain integration in ICOS allows us to comprehensively address the biogeochemical fluxes of carbon and GHGs and to identify and study existing gaps in knowledge.A recent example are the 17 publications, based on data from more than 100 stations, following the drought in Europe in 2018.The drought was analyzed from how it was detectable in the atmospheric station network and how it affected ecosystem processes and GHG budgets, to regional assessments of its influences on ecosystem carbon exchange, and relations to major crops (see special issue of Philosophical Transactions of the Royal Society, 2020, Vol.B375, No. 1810).ICOS made this rapid scientific response possible by building the foundation for fast action, by harmonizing observations and centralized data processing, by analyzing the data in near-real time to detect anomalies in drivers and ecosystem responses, by facilitating networking of scientists, and by providing virtual solutions for joint work.The results show that the drought affected more the productivity of crops and grasslands than forests, which protected themselves by reducing their evaporation and growth, leading to decreased uptake of carbon dioxide (see special issue of Philosophical Transactions of the Royal Society, 2020, Vol.B375, No. 1810).In general, carbon sinks decreased by 18% in a study covering 56 ecosystem sites (Graf et al. 2020).The dry conditions even turned some mires from sinks into sources (Rinne et al. 2020).In some parts of Europe, the winter of 2018 was wet, leaving a lot of soil moisture in the ground, while spring was sunny and came early-this caused the vegetation to grow in spring more than average.In some places, this early spring growth was enough to offset the reduction of carbon uptake later in summer (e.g., Smith et al. 2020).Currently, there is a joint effort of similar magnitude under preparation analyzing the warm winter of 2019/20.The above mentioned topics are also reflected in the biennial ICOS Science Conference that brings scientists from different disciplines together to discuss science as well as, e.g., methodological improvements and societal relevance of long-term observations of climate-related variables.Now, with the Paris Agreement having clear processes to guide the nations with climate change mitigation, the pressure is increasing to provide robust information to support the review of the impact of these actions (Article 14.1 in the Paris Agreement).ICOS is actively engaged with GCOS to provide observations of Essential Climate Variables and to draft a suitable indicator representing terrestrial ecosystems.ICOS provides data and participates to the development of Global Carbon Budget to reduce the uncertainty of the global estimates and to build a solid foundation for some of the global data sources, such as SOCAT and FLUXNET, the global network of gas flux observations between ecosystems and the atmosphere (Papale 2020).ICOS is in active dialogue with the United Nations Framework Convention on Climate Change (UNFCCC) Subsidiary Body for Scientific and Technological Advice to facilitate discussion between science and policy.ICOS is currently focusing on providing the needed information at national and regional levels with the separation of natural and anthropogenic fluxes.For example, VERIFY (H2020 project 776810) aims to improve national GHG inventories with top-down (atmospheric inversions) and bottom-up (inventories made with complementary methods and data than used by governmental authorities) scientific approaches (Petrescu et al. 2021).
The capability to disentangle the natural cycle and the anthropogenic disturbance has made progress, and consensus exists that the required next step is to link tightly in situ and remote sensing observations and modeling to more accurately quantify anthropogenic CO 2 emissions (Copernicus 2015(Copernicus , 2019)).The calibration and verification of satellite products and models within this system aim to rely on the in situ ICOS network, including potential atmospheric vertical profiling of GHGs using AirCore (Karion et al. 2010)

Fig. 1 .
Fig. 1.Map of ICOS stations.The dots represent fixed stations in different domains (ocean, ecosystem, atmosphere) and lines represent the Ships of Opportunities.Up-to-date details (e.g., station class, contact info, data) from each station can be found at www.icos-cp.eu/observations/station-network.

Fig. 2 .
Fig. 2. Schematic figure of the carbon cycle and related data collection process and user access to all the data via the Carbon Portal.The color-coding links the areas of the biogeochemical carbon cycle to the respective stations and Thematic Centres.The green color indicates the exchange of carbon, GHGs, and energy between the atmosphere and ecosystems (vertical arrows); the red color indicates the atmospheric gas concentrations, chemistry, and transport processes (horizontal arrows); and the blue color indicates the ocean-atmosphere gas exchange (vertical arrows), observed within the ICOS stations of respective domains (dots in the lower part and also in Fig. 1).The observations are centrally processed within the Thematic Centres and the data stored to ICOS repository, with Carbon Portal serving as a one-stop shop for all ICOS data products.

Fig. 3 .
Fig. 3.An example of ICOS tool to analyze the potential impact of natural and anthropogenic emissions to the CO 2 concentrations in the atmosphere, based on model simulation of Lagrangian transport model Stochastic Time Inverted Lagrangian Transport (STILT) together with emissionsector and fuel-type-specific emissions from a prerelease of the EDGARv4.3 inventory provided by the European Commission, Joint Research Centre (JRC)/Netherlands Environmental Assessment Agency (PBL).Figure shows results for the ICOS class 1 Jungfraujoch station in Switzerland, including biospheric and anthropogenic carbon emissions.(a) Modeled footprint and wind directions influencing the measurement tower signal.(b) Selected towers, location of atmospheric tower in Europe, and variables that are available for interactive visualization.(c) Time series of a selected variable, including measured and modeled concentrations.

Fig. 4 .
Fig. 4. Example of a regional-scale atmospheric inversion result, estimating the net ecosystem exchange (NEE) based on atmospheric observations from ICOS and other stations.The presented example is part of a multimodel ensemble of atmospheric inversions, available at Carbon Portal (Thompson et al. 2019), that was used to estimate the effect of the 2018 drought on net ecosystem exchange over Europe (Thompson et al. 2020).
and collaboration with the Total Carbon