Towards a Global Synthesis of Methane Fluxes From Land Ecosystems

1. Towards a Global Synthesis of Methane Fluxes From Land Ecosystems Report from the ESF Exploratory Workshop held in Hamburg (Germany), 10-12 April 2012 Lars Kutzbach, Miguel Mahecha, Timo Vesala http://www.klimacampus.de/esfworkshopch4.html

2. Motivation for the exploratory workshop • Since the 1980s, hundreds of studies on CH4 emission from various land ecosystems have been published, mostly based on chamber methods. • In the 1990s, the first eddy covariance (EC) measurements of CH4 fluxes were published; however, few because of expensive and technically demanding instruments; recently, the number of EC studies is strongly increasing due to fast technical development of new trace gas analysers. • There is no global synthesis of these extensive datasets so far. • Neither, there are generally accepted best practices protocols for chamber and EC methodologies. • Recently, many different initiatives focusing on CH4 were initiated. • Stronger links between different networks appear necessary.

3. Now many initiatives on CH4: need for networking between networks ? GHG-Europe (EU) MethaneNet (NERC, UK) PAGE21 (EU) DEFROST (Scandinavia) PERGAMON (COST, EU) (KoPF, BMBF, Germany) ABBA (COST, EU) many smaller projects InGOS (EU) ICOS (EU) FLUXNET

4. General goals of the exploratory workshop • Initiation of a global synthesis activity that aims at • harmonizing the steadily increasing number of methane (CH4) measurements from terrestrial ecosystems worldwide • and evaluating them jointly • Cooperation of scientists with complementary experiences: field measurements, advanced empirical data analyses, remote sensing, mechanistic modeling • Setting the stage for following larger-scale international activities (e.g., open science workshop in Hyytiälä, September 2012)

5. Initiation of a CH4-FLUXNET • A global CH4-FLUXNET (or GHG-FLUXNET) could …. • provide the scientific community with independent empirical reference for: • Local terrestrial CH4 dynamics • Empirical relations e.g. temperature sensitivities • Up-scaled fluxes • … AND this initiative could avoid many difficulties by learning from (and being integrated in) the existing (CO2-)FLUXNET experience, e.g., • Data policy • PI involvement • Data standardization Reference for constraining diagnostic/predictive models

6. Outline of the workshop • 2.5 days • 8 plenary sessions: 6 topical scientific sessions, 14 scientific talks, 2 summarising discussion sessions • 3 time slots (in total 5 hours) for 3 parallel break-out sessions in which subgroups focussed on advancing the specific questions/deliverables

7. Outline of the workshop • 26 participants from 11 countries (10 European countries, US) • 10 females, 16 males • 15 young scientists, 11 experienced scientists (according to the ESF definition)

8. Outline of the workshop Tuesday, 10.04.12 • A. Plenary session: The need for a FLUXNET-CH4; state of the art of land-atmosphere CH4 flux research (talks Kutzbach, Mahecha, Scarascia-Mugnozza, Heimann) • B. Plenary session: CH4 flux data availability from different land ecosystems (talks Rinne, Oechel, Zona, Schrier) Wednesday, 11.04.12 • C. Plenary session: Towards standardized methodologies for CH4 flux studies(talks Riis Christiansen, Eugster, Dengel) • D. Plenary session: Setting up a FLUXNET-CH4 initiative(talks Papale, Vesala, Sachs, v. Huissteden)

9. Outline of the workshop Wednesday, 11.04.12 • E. Break-out sessions a) Next steps for the initiation of a FLUXNET-CH4 initiative b) Joint efforts on best practices for closed chamber CH4 flux studies c) Joint efforts on best practices for eddy covariance CH4 flux studies • F. Break-out sessions – Continuation • G. Plenary session: Representation of CH4 processes in climate-carbon cycle models(talks Heimann, v. Huissteden, Vesala)

10. Outline of the workshop Thursday, 12.04.12 • H. Plenary session: Global-scale analyses/synthesis of CH4 processes and fluxes (talks Matthews, Mahecha, Bloom, Bousquet) • I. Plenary session: Summary and general discussion of break-out sessions • J. Plenary session: General discussion on next steps to achieve the workshop deliverables, follow-up research activities and/or collaborative actions or other specific outputs • K. Break-out sessions – Continuation

11. Outcomes: Next steps for the initiation of a FLUXNET-CH4 initiative • The participants recognised the need for a „network of networks” for CH4 research within Europe but preferably also on the global scale: a „FLUXNET-CH4“. • This “network of networks” primarily needs a functional and consistent database and a well-maintained, functional and inspiring information exchange platform. • The participants suggested using existing infrastructure to efficiently build up these two basic prerequisites for a global CH4 synthesis activity.

12. Outcomes: Next steps for the initiation of a FLUXNET-CH4 initiative • As information exchange platform, the existing internet platform www.MethaneNet.org (currently funded by the NERC, UK) is suggested. • Recommendation to avoid redundant databases and to establish the needed new CH4 flux database as a part of the already existing European Eddy Fluxes Databases Cluster (www.europe-fluxdata.eu). • For long-term archiving, using the open access library PANGAEA (http://www.pangaea.de) should be considered.

13. Outcomes: Next steps for the initiation of a FLUXNET-CH4 initiative • Dario Papale agreed to expand the European Eddy Fluxes Databases Cluster so that it fits to the needs of a CH4 flux synthesis. • It was agreed that the new database should include chamber measurements. • Database structure and the required content of the metadata information needs to be discussed.  Much is already implemented by Dario, but further discussion and optimisation necessary.

14. a b Graphic: P. Schreiber Graphic: A. Hormann Outcomes: Best practices for closed chamber CH4 flux studies

15. Outcomes: Best practices for closed chamber CH4 flux studies • In a recent extensive CH4 flux chamber intercomparison experiment (18 types), the tested chambers differed widely in their performance to replicate a reference flux, suggesting that published estimates of trace-gas fluxes are often biased. • Systematic errors derive from saturation of headspace, lack of mixing and disturbance of the headspace. • However, these biases can be minimised by applying non-linear regression flux calculation methods, increasing the chamber volume to basal area ratio, sufficient headspace mixing and minimising manual gas sampling volumes.

16. Outcomes: Best practices for closed chamber CH4 flux studies • Chamber results are generally considered highly valuable for budget considerations, characterisation of spatial variability across scales, and validation of modelling results. • However, it is necessary to develop a strategy to correct or “post-calibrate” existing chamber data based on metadata information on the experimental set-up. ( New database!) • Uncertainty ranges due to different measurement designs have to be assigned to each chamber data set following a standardised approach. • Is it possible to recommend “the perfect chamber” for future research?  Yes.

17. Outcomes: Best practices for eddy covariance CH4 flux studies

18. Outcomes: Best practices for eddy covariance CH4 flux studies • Currently, many new CH4 analysers based on laser spectroscopy are made available on the market. • Large fluxes can be measured with any of the currently available optical CH4 analysers. Large fluxes can be measured also with weak pumps (even with high damping losses). • Small fluxes are still a challenge making careful instrument selection important. • Data processing is still an issue as CH4 analysers have different properties than CO2 analysers, and additional or modified correction methods may be needed.

19. Outcomes: Best practices for eddy covariance CH4 flux studies • Generally, most principles of flux processing established for EC measurements of CO2 are also valid for CH4. • Major methodological differences between EC measurements of CH4 and CO2 are related to lower flux magnitudes and less or no diurnal variability of CH4 fluxes as well as to the differences in the instrumentation needed to detect CH4 concentration quickly and and precisely enough for the EC approach. • Need to better understand the impact of the instrument design constraints (flow distortion, sensor heating) on the accuracy of the CH4 flux measurements.

20. Outcomes: Best practices for eddy covariance CH4 flux studies • Impact of advection and storage on the CH4 fluxes needs to be systematically researched. • The effect of density corrections according to the Webb- Pearman-Leuning (WPL) on EC CH4 flux estimates is not sufficiently investigated. Is density correction better done by the classical WPL term on the fluxes or a point-by-point dilution calculation on the rawdata? • The detection of CH4 ebullition fluxes by the EC methodology needs to be studied in depth.  Hot spots and hot moments!!

21. Outcomes: Initiation of several collaborative synthesis papers • E. Matthews: Meta-data analysis of the existing CH4 flux literature, development of a „methanocentric“ wetland classification

22. E. Matthews: What are wetlands and where are they? papyrus, Uganda Okavango, Botswana Zambezi River, Africa Sudd Swamp, Sudan Amazon, Brazil Pantanal, Brazil Lake Jipe, Tanzania Amazon, Peru N. Australia

23. Outcomes: Initiation of several collaborative synthesis papers • E. Matthews: Meta-data analysis of the existing CH4 flux literature, development of a „methanocentric“ wetland classification • M. Mahecha et al.: Using new statistical time series analysis methods to extract the temperature sensitivity of ecosystem CH4 fluxes • W. Eugster, J. Rinne, L. Kutzbach et al.: State of the art of CH4 flux measurements by eddy covariance methodology • J. Riis Christiansen, S. Petersen, L. Kutzbach et al.: State of the art of CH4 flux measurements by closed chamber methodology • Many more ideas for papers and projects put forward in Hyytiälä

24. Next steps • Larger‐scale, international open science workshop on • The importance of land-atmosphere fluxes of methane and nitrous oxide for the global greenhouse‐gas balance – The need for a FLUXNET-GHG • Hyytiälä, Finland 2‐7 September 2012 • (95 participants incl. US, Europe, China, Japan, etc.)

25. Next steps • Working on the database structure for CH4 fluxes within the European Fluxes Database Cluster (www.europe-fluxdata.eu) • Which metadata needs to be included? • What needs to be incorporated in addition to the existing CO2 flux database? • If you want to contribute ideas on these questions, please contact Dario Papale (darpap@unitus.it) or Lars Kutzbach (lars.kutzbach@zmaw.de). • Much is already implemented, but further discussion and optimisation needed.

26. Next steps • Working on collaborative synthesis papers • either in small groups using CH4 data from these groups (or from literature reviews) • and/or through contributing to the proposed new larger-scale CH4 database within the European Fluxes Database Cluster(www.europe-fluxdata.eu) • Register your site, explore the database structure and provide feedback to Dario Papale (darpap@unitus.it)

27. Further need for integration ! • Networking between different CH4 science communities still needs to be intensified • measurements – experiments – models • land – ocean – atmosphere • different climate zones • different continents • Other CH4 flux database initiatives ? European? Global?

28. …of field researchers and modellers… International teamwork needed German Slovakian Russian

29. Thank you for your attention ! More information on the ESF Exploratory workshop (e.g., programme, list of participants, presentations): http://www.klimacampus.de/esfworkshopch4.html Email: lars.kutzbach@zmaw.de Photo: H. Joosten

31. Objectives of a CH4-FLUXNET: To provide a solid basis for a global quantification of the natural spatial and temporal (seasonal to interannual) variability of natural CH4-effluxes

32. Objectives of a CH4-FLUXNET: To provide a solid basis for a global quantification of the natural spatial and temporal (seasonal to interannual) variability of natural CH4-effluxes •  e.g. via “upscaling” fluxes (even if this sounds now “too early” for CH4) Example from CO2-fluxes… to be adapted

33. Objectives of a CH4-FLUXNET: To provide a solid basis for a global quantification of the natural spatial and temporal (seasonal to interannual) variability of natural CH4-effluxes  e.g. via “upscaling” fluxes (even if this sounds now “too early” for CH4) Monitoring of relationships between drivers and fluxes at site level! “Machine Learning” of relationships between drivers and fluxes Applying algorithm to gridded variables: generating a “quasi observational” flux field Jung et al. (2010) Nature. 467: 951-954. Exploratory Workshop: Towards a global Synthesis of CH4 fluxes from land ecosystems

34. Objectives of a CH4-FLUXNET: To provide a solid basis for a global quantification of the natural spatial and temporal (seasonal to interannual) variability of natural CH4-effluxes Why this works? Monitoring of relationships between drivers and fluxes at site level! “Machine Learning” of relationships between drivers and fluxes Applying algorithm to gridded variables: generating a “quasi observational” flux field

35. Objectives of a CH4-FLUXNET: To provide a solid basis for a global quantification of the natural spatial and temporal (seasonal to interannual) variability of natural CH4-effluxes Complications in case of CH4 fluxes  Sampling of the climate space less dense! More drivers needed (wetland extend?)  How to deal with “regime shifts”?  How to deal with the much higher spatiotemporal resolution (heterogeneity!) Monitoring of relationships between drivers and fluxes at site level! “Machine Learning” of relationships between drivers and fluxes Applying algorithm to gridded variables: generating a “quasi observational” flux field

36. Objectives of a CH4-FLUXNET: • Help constraining terrestrial biosphere models • FLUXNET = becoming “the reference” for structural adjustments or parameter optimization of terrestrial biosphere models. • Are similar approaches thinkable for a CH4 initiative? (… At least for specific regions?) • M. Williams et al. (2009) Improving land surface models with FLUXNET data. Biogeosciences, 6: 1341-1359. Data could be either eddy-covariance, chamber, or upscaled data!

37. Expected deliverables • The seed for an international CH4-FLUXNET initiative which builds on and connects previous synthesis and networking activities.  • Open science (!) workshop in Hyytiälä, Finland 2-7 September 2012 • The importance of land-atmosphere fluxes of methane and nitrous oxide for the global greenhouse-gas balance – The need for a FLUXNET-GHG • ( Sigrid Dengel) • A strategy for the implementation of a functional and consistent database collecting CH4 flux measurements from the global land ecosystems applying different approaches, e.g. chamber and eddy covariance methods. • Papers on recommendations on best practices in chamber and eddy covariance applications for CH4 flux measurements and standardized protocols for data processing and quality control.

38. Expected deliverables • Initialization of a series of synthesis papers on different aspects of CH4 flux processes, for instance on the temperature sensitivity of CH4 • Strategy for the generation of scaled up CH4 fields, a “quasi-observational” space-time explicit picture, at least for the northern temperate to arctic latitudes.