Evaluating emissions from top-down observations T. Ryerson – NOAA Chemical Sciences Division Motivation Most inventories compiled from bottom-up estimates, where reported data from many individual sources are summed an enormous undertaking regulatory / legal “baggage” independent assessments are needed to evaluate accuracy Emissions routinely change over all time scales hours - power plant loads - 30% days - urban mobile fleet mix - 50% years - response to control strategies - up to 80% Accurate inventory input is critical for models (AQ, Carbon Cycle, Climate). Evaluation methods and selected examples: Power plant NO x, petrochemical VOCs, ship BC, oil & gas CH 4 (biogenics covered in other presentations today) difficult, and slow, to update }

Method accuracy: ratios are within ±5% of CEMS Peischl et al., JGR (submitted) absolute emissions within ±30% Ryerson et al., JGR (1998), Science (2001) - apply top-down approach to wider variety of sources with reasonable confidence W.A. Parish emission estimates compared NO x /CO 2 SO 2 /CO 2 observed slope0.24 ± ±0.3 CEMS inventory Top-down methods for inventory assessment – power plant NO x, SO 2, CO, CO 2 Emission ratios derived from field data “snapshots” agree with continuous emission monitoring systems (CEMS) at power plants W.A. Parish power plant plume 1-Hz aircraft data

NOAA WP-3D aircraft data flight track scaled by measured CO NOAA Tall Tower Network data 457m level, KWKT tower, Waco, TX Additional data, analysis: A. Andrews, NOAA GMD Flexpart model: A. Stohl, NILU Top-down methods for inventory assessment – urban CO Urban CO emissions overestimated in US national inventory

NOAA WP-3D aircraft data flight track scaled by measured CO NOAA Tall Tower Network data 457m level, KWKT tower, Waco, TX Additional data, analysis: A. Andrews, NOAA GMD Flexpart model: A. Stohl, NILU Top-down methods for inventory assessment – urban CO Urban CO emissions overestimated in US national inventory

Top-down methods for inventory assessment – urban CO Urban CO emissions overestimated in US national inventory NOAA WP-3D aircraft data flight track scaled by measured CO NOAA Tall Tower Network data 457m level, KWKT tower, Waco, TX Additional data, analysis: A. Andrews, NOAA GMD Flexpart model: A. Stohl, NILU field project inventory “snapshots” very similar to data from longer-term monitoring top-down assessments are critical for both carbon cycle… - determining biospheric CO 2 uptake using tall-tower data requires accurate knowledge of urban CO/CO 2 signatures

Top-down methods for inventory assessment – petrochemical industry Errors in petrochemical VOC inventories persist in Houston Ethene flux from the Freeport petrochemical complex Ryerson et al., JGR (2003); de Gouw et al., ES&T (2009); Washenfelder et al., JGR (submitted). Measurement TCEQ inventory Top-down methods point to large (factor of 30) underestimates of alkene emissions from petrochemical industries in Houston, TX in 2000 and 2006

Top-down methods for inventory assessment – petrochemical industry Errors in petrochemical VOC inventories persist in Houston These alkenes dominate the most extreme (>150 ppbv) ozone exceedances Ethene flux from the Freeport petrochemical complex Measurement TCEQ inventory Top-down methods point to large (factor of 30) underestimates of alkene emissions from petrochemical industries in Houston, TX in 2000 and 2006 Ryerson et al., JGR (2003); de Gouw et al., ES&T (2009); Washenfelder et al., JGR (submitted).

Models realistically simulate observations of ozone and other photoproducts only after reactive VOC emissions are made consistent with top-down assessments State of Texas changed its emissions control strategy as a result of these findings 3D model and analysis: Grell, Bao, McKeen - NOAA GSD and CSD Gaussian plume model WRF-Chem 3D model Gaussian model and analysis: M. Trainer - NOAA CSD Ozone (ppbv) longitude (deg) Sweeny Freeport Measured O 3 Predicted O 3 - Reported VOC emissions Predicted O 3 - Corrected VOC emissions Reported VOC emissions Increased alkene emissions Top-down methods for inventory assessment – petrochemical industry Errors in petrochemical VOC inventories persist in Houston

Top-down methods for inventory assessment - shipping Lack et al., GRL (2008). Soot emission factors vary by a factor of 3 between marine engine types Soot emission dependencies on engine load and fuel quality are not well established Observed variability suggests using a single value in large-scale models will hamper the understanding of regionality of soot forcings and exposure

EDGAR 3.2 (2000 reporting year) US/Canadian Arctic total: 0.4 Tg CH 4 /year Top-down methods for inventory assessment – oil & gas extraction 32 million cubic feet per day ~0.3 Tg CH 4 per year) NASA DC-8 data from ARCTAS April 2008 Prudhoe Bay plume Prudhoe Bay includes the world’s largest gas-handling facility, and is the single largest CH 4 source in the EDGAR oil production sector

EDGAR inventories for 2000 and 2005 show 1000-fold differences for Prudhoe Bay Emissions unlikely did not not change substantially between these two years Working with EDGAR inventory teams to resolve these issues; a direct comparison is on hold Top-down methods for inventory assessment – oil & gas extraction

Tabulated inventories can be substantially in error, by factors of 2 to 10 or more, for many important chemical species from many of the largest source types. - continuing need for top-down, independent assessment of inventories Bottom-up inventories appear to be accurate to ±25% or better only if directly measured monitoring data are routinely available (e.g., point source CEMS) Errors are sufficiently large to confound development of scientifically sound control strategies based on anthropogenic emissions reductions. - relevant to air quality, carbon cycle, and climate issues Top-down methods for inventory assessment Conclusions