A dedicated flask sampling strategy developed for Integrated Carbon Observation System (ICOS) stations based on CO2 and CO measurements and Stochastic Time-Inverted Lagrangian Transport (STILT) footprint modelling Ingeborg Levin, Ute Karstens, Markus Eritt, Fabian Maier, Sabrina Arnold, Daniel Rzesanke, Samuel Hammer, Michel Ramonet, Gabriela Vítková, Sebastien Conil, et al. To cite this version: Ingeborg Levin, Ute Karstens, Markus Eritt, Fabian Maier, Sabrina Arnold, et al.. A dedicated flask sampling strategy developed for Integrated Carbon Observation System (ICOS) stations based on CO2 and CO measurements and Stochastic Time-Inverted Lagrangian Transport (STILT) foot- print modelling. Atmospheric Chemistry and Physics, European Geosciences Union, 2020, 20 (18), pp.11161-11180. 10.5194/acp-20-11161-2020. hal-02977162 HAL Id: hal-02977162 https://hal.archives-ouvertes.fr/hal-02977162 Submitted on 26 Oct 2020 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Atmos. Chem. Phys., 20, 11161–11180, 2020 https://doi.org/10.5194/acp-20-11161-2020 © Author(s) 2020. This work is distributed under the Creative Commons Attribution 4.0 License. A dedicated flask sampling strategy developed for Integrated Carbon Observation System (ICOS) stations based on CO2 and CO measurements and Stochastic Time-Inverted Lagrangian Transport (STILT) footprint modelling Ingeborg Levin1, Ute Karstens2, Markus Eritt3, Fabian Maier1, Sabrina Arnold4, Daniel Rzesanke3, Samuel Hammer1, Michel Ramonet5, Gabriela Vítková6, Sebastien Conil7, Michal Heliasz8, Dagmar Kubistin4, and Matthias Lindauer4 1Institut für Umweltphysik, Heidelberg University, 69120 Heidelberg, Germany 2ICOS Carbon Portal, Lund University, 22362 Lund, Sweden 3Max Planck Institute for Biogeochemistry, ICOS Flask- und Kalibrierlabor, 07745 Jena, Germany 4Meteorologisches Observatorium Hohenpeißenberg, Deutscher Wetterdienst, 82383 Hohenpeißenberg, Germany 5Laboratoire des Sciences du Climat et de l’Environnement (LSCE), IPSL, CEA-CNRS-UVSQ, Université Paris-Saclay, 91191 Gif-sur-Yvette, France 6Global Change Research Institute of the Czech Academy of Sciences, 603 00 Brno, Czech Republic 7DRD/OPE, Andra, 55290 Bure, France 8Centre for Environmental and Climate Research, Lund University, 22362 Lund, Sweden Correspondence: Ingeborg Levin ([email protected]) Received: 27 February 2020 – Discussion started: 17 March 2020 Revised: 2 June 2020 – Accepted: 8 August 2020 – Published: 29 September 2020 Abstract. In situ CO2 and CO measurements from five flasks collected around midday will likely be sampled during Integrated Carbon Observation System (ICOS) atmosphere low ambient variability (<0:5 parts per million (ppm) stan- stations have been analysed together with footprint model dard deviation of 1 min values). Based on a first application runs from the regional Stochastic Time-Inverted Lagrangian at the Hohenpeißenberg ICOS site, such flask data are princi- Transport (STILT) model to develop a dedicated strategy for pally suitable for detecting CO2 concentration biases larger flask sampling with an automated sampler. Flask sampling than 0.1 ppm with a 1σ confidence level between flask and in in ICOS has three different purposes, namely (1) to provide situ observations from only five flask comparisons. In order an independent quality control for in situ observations, (2) to have a maximum chance to also sample ffCO2 emission to provide representative information on atmospheric com- areas, additional flasks are collected on all other days in the ponents currently not monitored in situ at the stations, and afternoon. To check if the ffCO2 component will indeed be 14 (3) to collect samples for CO2 analysis that are signifi- large in these samples, we use the continuous in situ CO ob- cantly influenced by fossil fuel CO2 (ffCO2) emission areas. servations. The CO deviation from an estimated background Based on the existing data and experimental results obtained value is determined the day after each flask sampling, and at the Heidelberg pilot station with a prototype flask sam- depending on this offset, an automated decision is made as 14 pler, we suggest that single flask samples are collected reg- to whether a flask shall be retained for CO2 analysis. It ularly every third day around noon or in the afternoon from turned out that, based on existing data, ffCO2 events of more the highest level of a tower station. Air samples shall be col- than 4–5 ppm that would allow ffCO2 estimates with an un- lected over 1 h, with equal temporal weighting, to obtain a certainty below 30 % were very rare at all stations studied, true hourly mean. At all stations studied, more than 50 % of particularly in summer (only zero to five events per month Published by Copernicus Publications on behalf of the European Geosciences Union. 11162 I. Levin et al.: A dedicated flask sampling strategy developed for ICOS stations from May to August). During the other seasons, events could terms of source and/or sink attribution, separation of the fos- be collected more frequently. The strategy developed in this sil fuel from the biogenic CO2 signal is, therefore, manda- 14 project is currently being implemented at the ICOS stations. tory. Precise CO2 measurements are, however, currently only possible in dedicated laboratories and on discrete sam- ples. In Europe the Integrated Carbon Observation System re- 1 Introduction search infrastructure (ICOS RI; https://www.icos-cp.eu/, last access: 20 September 2020) has been established to monitor Since the pioneering work by Charles David Keeling who, GHG concentrations and fluxes in the atmosphere, in various already in the 1950s, started continuous monitoring of at- ecosystems, and over the neighbouring ocean basins. ICOS mospheric carbon dioxide concentrations at the South Pole atmosphere has set up a pan-European network of preferen- and Mauna Loa (Brown and Keeling, 1965), global cover- tially tall tower stations located at least 50 km away from in- age of continuous greenhouse gas (GHG) observations has dustrialised and highly populated areas. The primary purpose considerably improved (https://gaw.kishou.go.jp, last access: is to monitor biogenic sources and sinks in Europe and moni- 20 September 2020). However, there still exist large observa- tor their behaviour under changing climatic conditions. In ad- tional gaps in remote marine and continental regions of the dition to continuous CO2, CH4, and CO observations, a sub- globe, which have partly been filled by regular flask sampling set of stations (Class 1 stations) perform 2-week integrated 14 and analysis in central laboratories. If frequently conducted, sampling of CO2 for C analysis. Class 1 stations are addi- data from flask sampling in the marine realm are often rep- tionally equipped with an automated flask sampler dedicated resentative of the large-scale distribution of GHGs in the at- to three major objectives. First, the collected flasks shall pro- mosphere and, thus, suitable for estimating large-scale flux vide an independent quality control (QC) for the continuous distributions by inverse modelling. The situation is more dif- in situ measurements of CO2, CH4, CO, and further species ficult when it comes to representative flask sampling at con- mole fractions. Second, flasks shall be collected for the anal- tinental sites because there the distribution of sources and ysis of additional trace components not measured in situ at sinks is much more heterogeneous and variable than over the the stations; finally, flasks with a potentially elevated fossil oceans. fuel CO2 component originating from anthropogenic sources 14 In the last few decades, observational networks have in the footprint of the stations shall be analysed for CO2. been extended to the continents in order to closely monitor Dedicated sampling strategies had to be developed for GHG concentrations and quantify terrestrial GHG sources ICOS which best meet these three objectives and which can and sinks. These heterogeneous terrestrial fluxes are often be accomplished in the framework of the infrastructure and less well implemented in models compared to ocean fluxes its available capabilities and resources. This includes techni- (Friedlingstein et al., 2019). As biogenic sources and sinks cal constraints at the stations but also analysis capacity at the are strongly influenced by regional climatic variability, only ICOS Central Analytical Laboratories, which are analysing continental observations can provide insight into the asso- all flask samples in ICOS. The ICOS flask sampling strat- ciated ecosystem processes (Ciais et al., 2005; Ramonet et egy might change in the future, e.g. when real-time GHGs or al., 2020). Besides monitoring the terrestrial biosphere, mea- footprint prediction tools become available. surements over continents are also conducted to observe an- In the current paper, we first give an introduction to the thropogenic emissions, in particular from fossil fuel burn- current ICOS atmosphere station network and then present ing and agriculture. Due to their proximity