Research funded by NSF

1. Global budget and radiative forcing of black carbon aerosol: constraints from pole-to-pole (HIPPO) observations across the Pacific Qiaoqiao Wang, Daniel J. Jacob, J. Ryan Spackman, Anne E. Perring, Joshua P. Schwarz, NobuhiroMoteki, Eloïse A. Marais, Cui Ge, Jun Wang, Steven R.H. Barrett Research funded by NSF AGU talk on Dec 10, 2013

2. BC exported to the free troposphereis a major component of BC direct radiative forcing Integral contribution To BC forcing Global mean BC profile (Oslo CTM) • • • Export to free troposphere deep convection BC forcing efficiency • 50% from BC > 5 km • • scavenging • • frontal lifting • • • • • • • • • • • • • • • • • • • • • • • • • • BC source region (combustion) Ocean Samset and Myhre [2011]

3. Multimodelintercomparison and comparison to observations Multimodel intercomparisons and comparisons to observations Free tropospheric BC in AeroCom models is ~10x too high This has major implications for IPCC radiative forcing estimates TC4 (Costa Rica, summer) Observed Models Pressure, hPa • Large overestimate must reflect • model errors in scavenging BC, ng kg-1 HIPPO over Pacific (Jan) Pressure, hPa obs models 20S-20N obs models 60-80N BC, ng kg-1 BC, ng kg-1 Koch et al. [2009], Schwarz et al. [2010]

4. GEOS-Chemaerosol scavenging scheme Previous application to Arctic spring (ARCTAS) IN+CCN Dealing with freezing/frozen clouds is key uncertainty CCN Anvil precipitation Cloud updraftscavenging Large scale precipitation CCN+IN, impaction • Below-cloud scavenging (accumulation mode aerosol), • different for rain and snow entrainment • BC has 1-day time scale for conversion from hydrophobic (IN but not CCN) to hydrophilic (CCN but not IN) detrainment • Scheme evaluated with aerosol observations worldwide • 210Pb tropospheric lifetime of 8.6 days (consistent with best estimate of 9 days) • BC tropospheric lifetime of 4.2 days (vs. 6.8 ± 1.8 days in AeroCom models)

5. GEOS-Chem BC simulation: source regions and outflow BC source (2009): 4.9 Tg a-1 fuel + 1.6 Tg a-1 open fires Observations (circles) and model (background) Normalized mean bias (NMB) in range of -10% to -30% NMB= -27% Wang et al., accepted surface networks AERONET BC AAOD NMB= -28% NMB= -32% Aircraft profiles in continental/outflow regions HIPPO (US) Asian outflow (A-FORCE) US (HIPPO) observed model Arctic (ARCTAS) NMB= -12%

6. Comparison to HIPPO BC observations across the Pacific Observed Model PDF • Model doesn’t capture low tail, is too high at N mid-latitudes • Mean column bias is +48% • Still much better than the AeroCom models Wang et al., accepted

7. Zonal mean BC in GEOS-Chem Direct Radiative Forcing due to BC • A four-stream broadband radiative transfer model for DRF estimates • Global BC DRF=0.19 W m-2 (AAOD=0.0017) • Uncertainty range based on atmospheric distribution • AAOD: 0.0014-0.0026 • DRF: 0.17- 0.31 W m-2

8. BC top-of-atmosphere direct radiative forcing (DRF) Absorbing aerosol optical depth (AAOD) Mass absorption coefficient Forcing efficiency DRF = Emissions X Lifetime X X Global load • Our best estimate of 0.19 W m-2 is at the low end of literature and of IPCC AR5 recommendation of 0.40 (0.05-0.8) W m-2 for fuel-only • Models that cannot reproduce observations in the free troposphere should not be trusted for DRF estimates Wang et al., accepted

9. Zonal mean BC in Conclusions • Observed BC concentrations across the Pacific range is very low, implying much more efficient scavenging than is usually implemented in models. • The model with updated scavenging is able to reproduce the observed seasonality and latitudinal, and overall agrees with the HIPPO data within a factor of 2 • The simulation yields global mean BC AAOD of 0.0017 and DRF of 0.19 W m-2, reflecting low BC concentrations over the oceans and in the upper troposphere • Previous estimates of DRF are biased high because of excessive BC concentrations over oceans and in the free troposphere