1 / 57

The Virtual Tall Towers approach: A link to the global CO 2 flux network

The Virtual Tall Towers approach: A link to the global CO 2 flux network. Ken Davis, Natasha Miles, Scott Richardson, Weiguo Wang, Chuixiang Yi and colleagues The Pennsylvania State University Scott Denning, Joanne Skidmore and Marek Uliasz Colorado State University Peter Bakwin NOAA CMDL

dingbang-lio
Télécharger la présentation

The Virtual Tall Towers approach: A link to the global CO 2 flux network

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. The Virtual Tall Towers approach: A link to the global CO2 flux network Ken Davis, Natasha Miles, Scott Richardson, Weiguo Wang, Chuixiang Yi and colleagues The Pennsylvania State University Scott Denning, Joanne Skidmore and Marek Uliasz Colorado State University Peter Bakwin NOAA CMDL With support from: DoE Terrestrial Carbon Processes Program, DoE National Institutes for Global Environmental Change, NOAA

  2. Outline • Philosophy/hypotheses • Method • Virtual tall towers • Applications of existing VTT data • New research activities • Regional flux project: ChEAS • Gashound-based mixing ratio sensor • AmeriFlux gets calibrated!

  3. Philosophy/hypotheses • Flux networks and mixing ratio networks should be complementary (and co-located?). • Abundant, continuous terrestrial mixing ratio data will • enable regional, high-frequency inversions and • improve the accuracy of annual inversions. • A moderate accuracy and precision, high density network complements a sparser, high precision and accuracy network.

  4. Virtual tall towers method • Calibrate flux tower LI-CORs! Bakwin et al, 1995. Zhao et al, 1997. • Sub-sample for midday, well-mixed conditions in the atmospheric boundary layer. • Synoptic, seasonal and annual gradients resolved • Micrometeorological correction applied to remove surface layer – boundary layer offset. • Improves data quality, helpful for annual and high resolution inversions

  5. (One) motivation for virtual tall towers Estimated uncertainty of annual carbon budget from global inversion simulation using different North American observation network scenarios. Note that the assumed monthly random error for VTT sites was 2 ppm, about 5x greater than the expected level of random error. Systematic biases was assumed to be zero.

  6. Photo credit: UND Citation crew, COBRA WLEF tall tower (447m) CO2 flux measurements at: 30, 122 and 396 m CO2 mixing ratio measurements at: 11, 30, 76, 122, 244 and 396 m

  7. Diurnal cycle of CO2 in the ABL:midday vertical gradients are very small Bakwin et al, 1998

  8. ABL CO2 signals vs. the surface-layer - ABL bias at WLEF Davis et al, in prep

  9. Monthly average offsets between surface layer and mid-ABL CO2

  10. Monthly mean CO2 mixing ratios in the convective boundary layer (CBL) at WLEF sub-sampled for convective afternoon hours at WLEF, free troposphere (FT) from aircraft flights at Carr, CO, and marine boundary layer (MBL) at 44.4N. Also shown is cumulative NEE at WLEF. Yi et al, to be submitted.

  11. Observed synoptic cycle of CO2 mixing ratios and temperature (for reference) at the WLEF tower. (a) Hourly, continuous data for 396m, data subsampled for daytime, well-mixed conditions at 30m. (b) Difference between daytime-averaged 30m and 396m CO2 mixing ratio data (+), and a virtual tall tower correction (diamond) using simple assumptions and gradient functions. Bias is just under 0.2 ppm (monthly mean). Standard error (monthly mean) is less than 0.2 ppm. Davis et al, in prep. Synoptic variability in CO2

  12. Micrometeorological correction: VTT methodology • Vertical gradients are determined by fluxes, mixing depth, and vigor of mixing. Universal gradient functions relate these quantities to the vertical gradient (Wyngaard and Brost, 1984). • Gradient functions have been computed via large eddy simulation (Wyngaard and Moeng, 1984; 1989; Patton et al, 2002; in prep) and observed (Davis et al, in prep).

  13. z/h Field data(Davis et al, in prep) LES results (Moeng and Wyngaard, 1984) z/h gb gt

  14. LES results: Patton et al, in prep

  15. Summary of VTT method • VTT correction is modest (~2ppm max at WLEF). Maximum for large fluxes, and measurements close to the surface • Significant uncertainty exists re: proper gradient functions over forest • Gradient functions can be determined more precisely via further study of ABL turbulence

  16. Applications of existing tall tower, aircraft and vtt data • Davis et al, Global Change Biology, 2003. • Seasonal offset exists between CO2 fluxes and ABL mixing ratios measured at the WLEF tower. • Bakwin et al, in review, Tellus • Similar offset exists at 3 of 4 flux towers, correlated with marine BL mixing ratios, can be used with reanalysis data to estimate regional surface fluxes of CO2 • Hurwitz et al, in press, J. Atmospheric Sciences. • Cause of seasonal offset appears to be vertical mixing caused by synoptic passages. Tropospheric mixing ratios very similar to marine BL mixing ratios. • Helliker et al, to be submitted • Similarity between vertical mixing of water vapor and CO2 can be used to derive a regional, synoptic, ABL to free troposphere flux-gradient relationship. CBL mixing ratios can estimate regional CO2 fluxes with surprising long-term accuracy. • Butler et al, in prep • Seasonal flux anomaly in the spring of 1998 shows up across much of northeast and northcentral North America. Flux anomaly is correlated with a mixing ratio anomaly measured both at the flux towers (VTT method) and in the marine flask network.

  17. Net ecosystem-atmosphere exchange of CO2 in northern Wisconsin Davis et al, Global Change Biology, 2003.

  18. Common seasonal patterns across flux tower sites: link to marine flask network Results show that the sum of vertical and horizontal transport is related to the difference between the tower midday CO2 mixing ratio and the marine boundary layer CO2 mixing ratio. Bakwin et al, in review

  19. Cold frontal passage and CO2 advection (14 July, 1998) Hurwitz et al, in press, JAS

  20. Helliker, Berry et al: Boundary layer cuvette,orSynoptic flux-gradient relationship Utilizes similarity in vertical mixing of scalars between the ABL and the free troposphere. Tropospheric mixing occurs with synoptic events. Synoptic events are analogous to ABL eddies, thus the method is analogous to surface layer similarity and flux-gradient relationships.

  21. Synoptic flux-gradient vs. eddy-covariance fluxes of CO2, WLEF tower. Helliker et al, to be submitted.

  22. Early leaf-out, 1998, Wisconsin

  23. Impact on atmospheric [CO2]

  24. Spatial coherence of seasonal flux anomalies A similar pattern is seen at several flux towers in N. America and Europe. Three sites have high-quality [CO2] measurements + data at Fluxnet (NOBS, HF, WLEF). The spring 98 warm period and a later cloudy period appear at all 3 sites.

  25. Detection of the spring 98 anomaly via oceanic flasks? 2 Alaskan flask sites have slightly higher [CO2] in the spring of 98. Mace Head, Ireland shows a depression of [CO2] in the spring of 98. Potential exists to link flux towers with seasonal inverse studies. Butler et al, in prep.

  26. Summary of VTT applications • Seasonal and synoptic signals at terrestrial sites are large and robust. Vertical mixing dominates. • VTT data capture large-scale mixing ratio data (consistent relationships vs. marine ABL mixing ratios, FT mixing ratios). • Flux data capture large-area temporal patterns. • Budget and flux-gradient approaches have promise for deriving regional fluxes. • Analyses based on spatial gradients among VTT sites are lacking to date.

  27. New research • Regional tower network • AmeriFlux VTT network

  28. ChEAS regional flux experiment • Derive daytime and daily seasonal fluxes using regional atmospheric inversions and relatively cheap, abundant, in situ CO2 sensors. • Overarching goal – evaluate/merge multiple approaches of studying terrestrial fluxes of CO2. • Merge flux-tower based upscaling with downscaled inversion methodology. Regional integration and mechanistic interpretation. • Determine interannual variations in seasonal fluxes on a regional basis. Again, integrate with regional flux measurements/mechanistic interpretations. • If possible, derive net annual fluxes. Spatial resolution is limited by the magnitude of the annual signal.

  29. Methods for determining NEE of CO2Methods for bridging the gap Upscale via ecosystem models and networks of towers. Move towards regional inverse modeling

  30. Chequamegon Ecosystem-Atmosphere Study (ChEAS) flux towers        100 km

  31. NEE and Gross Fluxes: All Current Data

  32. Experimental design • Deploy a regional network of in situ CO2 sensors. • Calibrate secondary standards for in situ sensors. Four standards/site, midday mixing ratio range. • NOAA CMDL standards used to create secondary working standards. • Sensors mounted on communications towers at a similar height, ~75m above ground. • Differences in mixing ratio across the region reveal regional surface fluxes of CO2. • Compare results to upscaling via regional flux tower cluster.

  33. ChEAS regional flux experiment domain = LI-820 sampling from 75m above ground on a communications tower. = 40m Sylvania flux tower with high-quality standard gases. = 447m WLEF tower. LI-820, CMDL in situ and flask measurements.

  34. Expected regional mixing ratio differences (winter to summer)

  35. Influence functions for the WLEF tower (z=400m) for the June, July, August and September 2000 Simulation: RAMS v4.3 with two nested grids (Δx=100km and 20 km) + LPD (Lagrangian Particle Dispersion) model in a receptor-oriented mode. The 2nd finer grid covers the domain around the WLEF tower used for dispersion calculations. Concentration sampling: The influence functions, travel time and influence frequency are presented for selected 2 hour sampling periods during the day during August 2000. The 00-24 hour period represents the results for all sampling times during the month. All sampling times are local (GMT-6h). The influence frequency is derived in reference to the sampling period (i.e., how often the signal from a given source area is observed at the receptor during the sampling period). Travel distance is derived from the presented influence functions but averaged over 45o sectors and shown in polar coordinates. Tracers: 1. Passive tracer with a constant flux – the spatial distributions are the same as for the respiration flux including dependence on the soil temperature. 2. A-tracer (assimilation tracer) with a daytime flux driven by shortwave radiation Reference: Uliasz, M. and A. S. Denning, 2002:Deriving mesoscale surface fluxes of trace gases from concentration data. submitted to: J. Appl. Meteor. Download: http://biocycle.atmos.colostate.edu/~marek/research/publications.htm

  36. Influence function climatology

  37. CO2 Measurement System Schematic

  38. Deployment at Wittenberg

  39. Why use the Licor 820? • Fast time response is not required (feature of LI-6262, LI-7000) • ~3 ppm peak-to-peak noise for 1-sec samples is large… • LI-6262 ~ 0.3 peak-to-peak noise • But noise is random • And noise reduces to < 0.2 ppm for a 5-minute average

  40. LI-820 features • Single cell IRGA • Cost:$2,400 per sensor • LI-6262 ~ $10,000 • LI-7000 ~ $14,000 • Size: 22 x 15 x 8 cm • Weight: ~ 1 kg • Temperature controlled emitter and detector • Zero drift < 1 ppm in 24 hrs • Power: 12-30 VDC, 0.3 A 12 V (1.2 A warm-up) • Output: Analog Voltage or mA or Digital TTL • XML communication protocol

  41. Planned intercalibration procedures • Sample same air with all units for a limited period of time • Circulate high-pressure target tanks • Co-locate one sensor with the WLEF tower flask and in situ sensors

More Related