1. NEW PERSPECTIVES ON ATMOSPHERIC MERCURY Daniel J. Jacob with Noelle E. Selin and Christopher D. Holmes Supported by NSF, EPA and Sarah Strode and Lyatt Jaegle (U. Washington)

2. RISING MERCURY IN THE BIOSPHERE 3000-yr record in Swiss bog Roos-Baraclough and Shotyk, ES&T 2003 Mercury in polar bear fur up 5-12X since 1890 Dietz et al., ES&T 2006 States with fish mercury advisories

3. GEOCHEMICAL CYCLE OF MERCURY (present) SOIL 1000 ATMOSPHERE 5.4 SURFACE OCEAN 10 wet &dry deposition evasion rivers 0.5 SEDIMENTS Natural (rocks, volcanoes) Anthropogenic (fossil fuels) evasion 0.2 burial 3.2 2.8 Inventories in Gg, fluxes in Gg yr -1 2.2 1.5 DEEP OCEAN 280 LAND SURFACE uplift 3.8 2.3 0.6 0.5 Selin et al. [2006], Strode et al. [2006]

4. SOURCE-DEPOSITION RELATIONSHIPS OF MERCURY CONTROLLED BY ATMOSPHERIC CHEMISTRY Hg(0) Hg(II) SURFACE RESERVOIRS (LAND, OCEAN) ATMOSPHERE Hg(II)Hg(0) hv, bio OH (80%), O 3 (20%), Br,…? aqueous hv? combustion H ~ 10 6 M atm -1 fast deposition H = 0.1 M atm -1 3.9 0.3 14 8 dry 4.7 wet 2.1 4.8 1.3 0.7 Tropospheric nventories in Gg, fluxes in Gg yr -1 evasion Selin et al. [2006],

5. ANTHROPOGENIC EMISSIONS OF MERCURY: recent shift from N.America/Europe to Asia 1990 Total: 1.88 Gg yr -1 2000 Total: 2.27 Gg yr -1 Pacyna and Pacyna, 2005

6. GEOS-Chem GLOBAL 3-D SIMULATION OF ATMOSPHERIC AND OCEANIC MERCURY Atmospheric chemical transport model (CTM) driven by assimilated meteorological data from the NASA Global Modeling and Assimilation Office (GMAO) with 3-6 h frequency, 1 o x1 o horizontal resolution Mercury simulation includes dynamic coupling of atmosphere and slab mixed layer ocean Hg(0), Hg(II), and inert particulate Hg in both atmosphere and ocean emissions, chemistry, deposition as described in previous slides horizontal resolution 4 o x5 o

7. SIMULATION OF TOTAL GASEOUS MERCURY (TGM) Land-based sites observed: 1.58 ± 0.19 ng m -3 model: 1.60 ± 0.10 ng m -3 …but not clear why NH cruise data are so high! Annual mean surface air concentrations and ship cruise data Selin et al., 2006 Circles are observations; background is model R 2 =0.51

8. SEASONAL VARIATION OF TGM AT NORTHERN MID-LATITUDES 12 sites Observed Model (standard) Model (OH oxidation only) Model (O 3 oxidation only) Selin et al., 2006 some evidence for photochemical redox chemistry of Hg

9. Hg DEPOSITION OVER U.S. : LOCAL VS. GLOBAL SOURCES Wet deposition fluxes, 2003-2004 (background=model,dots=measured) max in southeast U.S. from oxidation of global Hg pool 2 nd max in midwest from regional sources (mostly dry deposition in GEOS-Chem) 2/3 of Hg deposition over U.S. in model is dry, not wet! Simulated % contribution of North American sources to total Hg deposition U.S. mean: 20% Selin et al. [2006]

10. OBSERVATIONS AT OKINAWA (Asian outflow) Observed (Jaffe et al., 2005) Model CO Hg(0) RGM Reactive gaseous mercury (RGM) is gas-phase component of Hg(II) Correlation of Hg(0) with CO (r 2 = 0.84 in obs, 0.91 in model) indicates dominant Asian source (underestimated by 30% in model) RGM not correlated with Hg(0) either in model or observations; source from subsidence

11. DIURNAL CYCLE OF RGM AT OKINAWA Observations Model Large diurnal cycle at Okinawa implies a photochemical source (OH in the model – but Br would better explain early-morning rise) a very fast sink (uptake by sea salt aerosols in the model) Selin et al. [2006] von Glasow et al., GRL 2002 OH Br Production Rate Typical diurnal variation of OH and Br

12. PREDICTION OF INCREASING Hg(II) WITH ALTITUDE Hg(II) Hg(0) because Hg(II) loss is largely restricted to lower troposphere: dry deposition, precipitation, clouds Selin et al. [2006]

13. SIMULATED vs. OBSERVED RGM AT Mt. BATCHELOR, OREGON (2.7 km) Model Observed: day (upslope) night (subsidence) all Evidence for high RGM in subsiding air masses High RGM events in observations reproduced only timidly by model; These events are anticorrelated with Hg(0), reflecting Hg(0)  Hg(II) conversion Swartzendruber et al. [2006]

14. FAST Hg OXIDATION BY Br OBSERVED IN POLAR SPRING Hg HgBrHgBrX X = OH, Br… Br T Tropospheric BrO Sep 1997 GOME - SLIMCAT Hg(0) depletion events observed together with ozone depletion events Spitzbergen data [Sprovieri et al., 2005]Theys et al. [2004] High tropospheric BrO in polar spring O3O3 Hg(0)

15. ATOMIC Br UBIQUITOUS IN TROPOSPHERE from oxidation of bromocarbons, release by sea salt Br (%) of Br y (March noon 180 °W) Global CTM simulation [Yang et al., 2005] (10 13 cm -2 ) Simulated BrO concentrations consistent with few remote sensing data available (0.1-1 pptv) Fractionation as Br is highest in upper toposphere (strong hv) HOBr BrNO 3 HBr hv, OH BrBrO O3O3 hv Br y

16. GLOBAL OXIDATION OF Hg(0) BY ATOMIC Br Br concentrations [Yang et al., 2005] resulting Hg(0) lifetime yields global tropospheric lifetime for Hg(0)  Hg(II) of 200 days (160-510); Br could be a (the?) major Hg(0) oxidant! Holmes et al. [2006]

17. SOME INDIRECT EVIDENCE FOR Hg OXIDATION BY Br RGM O3O3 TGM RGM O3O3 TGM Mercury depletion event in Antarctic summer associated with elevated ozone [Temme et al., 2003]: subsidence of upper tropospheric air? Single-particle aircraft observations [Murphy et al., 2006]; elevated aerosol Hg(II) above tropopause associated with elevated Br

18. EFFECTS OF CLIMATE CHANGE ON MERCURY CYCLE Precipitation Winds Exchange with Surface reservoirs Ocean currents biota Chemistry