CDIAC AmeriFlux Data Management

1. CDIAC AmeriFlux Data Management Tom Boden (Bai Yang) Carbon Dioxide Information Analysis Center (CDIAC) Oak Ridge National Laboratory (ORNL) Presented September 27, 2010 at the CDIAC User Working Group Meeting Oak Ridge National Laboratory OAK RIDGE NATIONAL LABORATORY U. S. DEPARTMENT OF ENERGY

2. AmeriFlux Network ~140 sites in 5 countries; 92 active sites, 48 inactive sites Participation Requirements Make year-round core measurements using the eddy-covariance technique Submit data to CDIAC within 1 year of collection Participate in AmeriFlux Science Meetings and synthesis/modeling activities DOE supports the AmeriFlux infrastructure including the chairperson, calibration laboratory, and a permanent data archive (CDIAC since 1997).

3. F L U X N E T - a network of networks Global Network Non-network sites Regional Networks CARBOEURO- FLUX AmeriFlux Canadian CP AsiaFlux OzNet Oceania Americas Europe Asia - Japan Tower Sites Architecture of Global / Regional Flux Networks

4. Primary FY2010 CDIAC AmeriFlux Team Misha Krassovski – data system - (0.9 FTE) Bai Yang – micrometeorologist [QA/QC] - (0.8 FTE) Barbara Jackson – SAS programming - (0.7 FTE) Ben Norton – web & DB maintenance - (0.5 FTE) Tom Boden – coordination - (0.3 FTE) Lianhong Gu – A-Ci analysis – (0.04 FTE)

5. AmeriFlux Data Processing & Products Site Investigator & Team CDIAC FLUXNET Site Proc. Flux/Met Data 30 min Averages Quality Flags Gap-filled NEP/Re QA/QC Network-wide Database Standard Files Site Proc. Biological Data BADM CADM Level 0 Level 1 Level 2 Level 3 & 4

6. Level 1 to Level 2 Data Processing Scheme Level 1 data received from AmeriFlux site PIs • Flux, Meteorological, and Ecological Data Files • Stored “as received” from investigator • File format, parameter names, and units unchanged L1 files available from FTP area • Add value through automated QA/QC • Range checks, missing values, check solar elevation and time stamp • Basic Statistics (mean, variance, range) • Gap Analysis: determine frequency and duration of gaps in record • Spike detection, stationarity, diurnal and seasonal pattern • Cross checks/correlations (Rg vs. PAR) • Graphical Analysis: time series, property vs property, frequency | | 5 days | | Perform QA/QC • Incorporate into network-wide database • Convert into standard AmeriFlux units, reporting intervals, and naming conventions • Generate core parameters (e.g., VPD) where needed Calculate additional parameters, add to network database L2 data products available via Web interface and from FTP area • Web access to network-wide database • Querying capabilities – time, measurement parameter, and site • Graphics generation and data retrieval capabilities • FTP access to standard files • Standard files with 40 core variables (csv, NetCDF, tar) • Metadata files (html & txt formats) Incorporate into Web data interface & produce standardized data files & metadata reports

7. Progress in gap-filling meteorological data • Completed. We have completed a prototype data set of gap-filled meteorological data for the majority of active AmeriFlux sites. • Ongoing. This data set has been evaluated against other data sets filled by different techniques. The evaluation will continue if necessary and the sensitivity of ecosystem models to the gap-filled data sets will be estimated. • Ongoing. We are improving our gap-filling methods by including reanalysis data sets (NARR, DayMet, etc) and radar precipitation data. Present focus on filling long gaps and improving data reliability for precipitation, radiation, soil moisture and soil temperature. • Goal. We hope to finalize the gap-filled met-data sets before the next AmeriFlux annual meeting (Feb. 2011).

8. AmeriFlux Data System Site & PI Information, Pubs, & Instruments 12/2008 http://ameriflux.ornl.gov/edit http://ameriflux.ornl.gov Metadata (e.g., Fc corrections, instrument calibrations FY 2011 Reported Variables Flux & Met/Bio 1/2009 3/2009 CDIAC MS SQL 2005 RDBMS Flux & Meteorological Data Commenced 8/2009 Users Worldwide PHP Interface ORNL AmeriFlux & FLUXNET Data Team PHP Interface Leaf Aci 3/2010 Biological Data BADM 11/2009 FY2010 Priorities Data extraction/e-mail notification Load remaining L2 data, begin loading L4 data Web migration Inclusion of L2 & L4 data in the Earth System Grid Biological data editing functionality http://leafweb.ornl.gov Gu EDO methodology & code Web interface/Sharepoint for data submission and dissemination of results

9. Examples of CDIAC AmeriFlux Contributions to Synthesis & Modeling Activities • La Thuile, Italy workshop and resulting data sets • NACP Interim Site Synthesis • C-LAMP

10. La Thuile, Italy FLUXNET WorkshopFebruary 18-22, 2007 Mandatory data: NEE/FC CO2/SFC U*/TAU Rg/PAR Ta H2O/RH/VPD • Participation and data response • ~60 participants, good representation from regional networks and major programs (LBA, TCOS) • 921 site yrs of data from 240 sites worldwide (33 countries) • 318 site yrs of data from 77 AmeriFlux sites • Resulted in ~60 proposed papers • http://www.fluxdata.org • Workshop highlighted the importance of biological/ecological data and consistent ancillary information (ecosystem classifications)

11. C-LAMP • The Carbon-Land Model Intercomparison Project (C-LAMP) consists of an experimental protocol, model evaluation metrics, a prototype diagnostics package, model output standards, and a database of model simulation results on the Earth System Grid (ESG). See http://www.climatemodeling.org/c-lamp • In the first set of runs, over 16ky of simulation and 50TB of output were generated using the Climate Science End Station INCITE allocation at ORNL. • CLM3.1 (Community Land Model) combined with CASA´ (Carnegie-Ames-Stanford Approach) and CN (carbon & nitrogen) biogeochemistry modules were evaluated against best-available satellite- and ground-based measurements, and new runs are being used to evaluate CLM4 performance. • C-LAMP will serve as a benchmarking prototype for Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) in conjunction with ESG. Randerson, J. T., F. M. Hoffman, P. E. Thornton, N. M. Mahowald, K. Lindsay, Y. H. Lee, C. D. Nevison, S. C. Doney, G. B. Bonan, R. Stöckli, C. C. Covey, S. W. Running, and I. Y. Fung. “Systematic Assessment of Terrestrial Biogeochemistry in Coupled Climate-Carbon Models.” Global Change Biology, in press.

12. AmeriFlux Data Are An Important Component of C-LAMP • C-LAMP diagnostics include model comparisons with MODIS (Moderate Resolution Imaging Spectroradiometer) leaf area index (LAI) and net primary production (NPP), Globalview amplitude and phase, AmeriFlux Level 4 energy and carbon fluxes, Free Air CO2 Enrichment (FACE) site measurements, and various estimates of carbon stocks and transient dynamics. • Shown at right is a time series comparison of net ecosystem exchange (NEE), incoming shortwave, latent and sensible heat, gross primary production (GPP), and ecosystem respiration from CLM3.1-CN against measurements from the Morgan Monroe site. • C-LAMP diagnostics must be run before any model changes to the CLM biogeophysics or biogeochemistry can be made.

13. Future Directions & Emphasis? • Gap-filling meteorological data • Biological data • QA/QC of existing submissions • Improved parameterizations for models • Expanding LeafWeb • Amassing key measurements (e.g., soil respiration) • Flux re-analysis product within the ORNL SFA framework • FY 2011 DOE BER “AmeriFlux Center” solicitation • Collaborations with NEON