Constraining fossil fuel CO emissions by the joint assimilation of atmospheric CO and CO measurements Sourish Basu, John Miller, Scott Lehman A T M O S P N D H A E R C I I C N A A E D M C O I N L I S A T N R O A I T T A O I N N U E . S C R . D E E M P M A O R C T M F E N T O GMD Annual Conference Boulder, May
What is the issue? dC dt = F oce + F bio + F fos 2010 2010 terrestrial flux fossil fuel 0.60 0.45 0.30 0.15 0.00 0.15 0.30 0.45 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 ◮ Almost all atmospheric CO inversions assume CO (ff) “perfectly” known, solve for natural fluxes
What is the issue? dC dt = F oce + F bio + F fos April 2010 April 2010 terrestrial flux fossil fuel spread 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 ◮ Almost all atmospheric CO inversions assume CO (ff) “perfectly” known, solve for natural fluxes ◮ Only true annually, for global and (some) national totals ◮ Usually not up to date, EDGAR yr old, Vulcan yr old
CO is a tracer for CO (ff) ∆ C ff − � (i.e., zero CO ) = − . � ∆ C for ppm CO (ff) Scaling in = fossil fuel, ocean and land disequilibrium, nuclear and fossil fuel only cosmogenic production
NWR CMA 2010 fossil fuel Miller 0 1 2 3 4 5
NWR NWR CMA CMA 2010 fossil fuel 2006 fossil fuel Miller Miller/Vulcan 0 1 2 3 4 5 0 1 2 3 4 5
Prior flux uncertainties ◮ US CO (ff): . ± . Pg CO ◮ Fossil fuel: . × inter-prior spread, km hybrid, month ◮ Land biosphere: . × respiration per grid cell, km (e), month ◮ Ocean: × abs(net flux), km (e), month ◮ Ocean disequilibrium: . × abs(net flux), regional, month ◮ Land disequilibrium: . × abs(net flux), regional, month Our measurements are CO and CO · ∆ CO
Model-observation mismatch of CO · ∆ CO Park Falls, Wisconsin, USA (46.0°N, 90.3°W, 472 masl) Observations Prior 20 Posterior CO · ¹ CO mixing ratio 15 10 5 0 Mar 2010 Jun 2010 Sep 2010 Dec 2010 Mar 2011 . × � · ppm Average prior mismatch = − . × � · ppm Average posterior mismatch =
Model-observation mismatch of CO · ∆ CO Residual of CO · ¹ CO 4 0 4 2 2 Argyle, AMT Maine, USA Boulder Atmospheric BAO Observatory, Colorado, USA Te inversion is doing what it is supposed to do Cape May, CMA New jersey, United States Drake DRP Passage Indiana Flux INX Experiment, USA Park Falls, LEF Wisconsin, USA Mt. Wilson MWO Observatory, USA Worcester, NHA Massachusetts, United States NWR Niwot Ridge, Colorado, USA Beech SCT Island, South Carolina, USA Shangdianzi, Peoples SDZ Republic of China Tae-ahn Peninsula, TAP Republic of Korea WGC Walnut Grove, California, USA Optimized Prior Moody, WKT Texas, USA
Adjustments to fluxes/optimized emissions ◮ Te CMA–NWR gradient is consistent with more CO (ff) emission inland NWR CMA 2010 fossil fuel Miller 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Adjustments to fluxes/optimized emissions ◮ Te CMA–NWR gradient is consistent with more CO (ff) emission inland ◮ Te inversion increases CO (ff) emission inland NWR NWR CMA CMA 2010 fossil fuel 2010 fossil fuel Miller adjustments 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8
Seasonal vs annual CO (ff) adjustments ◮ Adjustments at the monthly scale are larger than adjustments at the annual scale ◮ Spatial patterns of the two 2010 fossil fuel adjustments adjustments can be different 1.5 1.2 0.9 0.6 0.3 0.0 0.3 0.6 0.9 1.2 1.5 Jan 2010 Jul 2010 fossil fuel fossil fuel adjustments adjustments 1.5 1.2 0.9 0.6 0.3 0.0 0.3 0.6 0.9 1.2 1.5 1.5 1.2 0.9 0.6 0.3 0.0 0.3 0.6 0.9 1.2 1.5
Points to take home ◮ Fossil fuel CO “well known” at national/yearly scales, not at regional/monthly scales ◮ Errors in CO (ff) emission estimates cause errors in NEE estimates
Points to take home ◮ Fossil fuel CO “well known” at national/yearly scales, not at regional/monthly scales ◮ Errors in CO (ff) emission estimates cause errors in NEE estimates ◮ CO is a good tracer for CO (ff), can disentangle CO (total) from CO (ff) ◮ Even with ∼ times lower measurement density, CO measurements in a CO + CO inversion shifts emission of CO (ff) inland, as expected
Points to take home ◮ Fossil fuel CO “well known” at national/yearly scales, not at regional/monthly scales ◮ Errors in CO (ff) emission estimates cause errors in NEE estimates ◮ CO is a good tracer for CO (ff), can disentangle CO (total) from CO (ff) ◮ Even with ∼ times lower measurement density, CO measurements in a CO + CO inversion shifts emission of CO (ff) inland, as expected ◮ Very much a work in progress, not yet the optimal framework for utilizing CO measurements
Mass balance dC dt = F oce + F bio + F fos d dt ( C · ∆ atm ) =∆ fos F fos + ∆ atm ( F oce + F bio ) + ∆ oce F oce → atm + ∆ bio F bio → atm + α ( F nuc + F cosmo ) tracers transported fluxes estimated
Uncertainties ◮ Ocean disequilibrium: . × ◮ US CO (ff): . ± . Pg CO abs(net flux), regional, month ◮ Land biosphere: . × ◮ Land disequilibrium: . × respiration per grid cell, km abs(net flux), regional, month (e), month ◮ Ocean: × abs(net flux), km (e), month ◮ Fossil fuel: . × inter-prior spread, km hybrid, month
Posterior correlation between CO (ff) and CO (nat) Posterior correlation between CO (nat) and CO (ff) from 01/01/10 to 01/01/11 0.0 Posterior correlation 0.1 0.2 0.3 Traditional No 12C 0.4 CO2 (FF) Eastern Central Western Russia G8 China United Canada Brazil EU27 India States US US US
Recommend
More recommend
