global transport and cycling of mercury and persistent...

Global Transport and Cycling of Mercury and Persistent Organic Pollutants: Science and Policy Perspectives

Noelle E. Selin Assistant Professor of Engineering Systems and Atmospheric Chemistry Massachusetts Institute of Technology 14 September 2012 selin@mit.edu http://mit.edu/selin http://mit.edu/selingroup

Transport and Fate of Pollutants: Science and Policy at Multiple Scales

Anthropogenic sources are both intentional and unintentional (byproducts)

Atmospheric transport and deposition leads to effects in wildlife, humans (uncertainties about atmospheric chemistry, processes)

Particular concern in the Arctic environment due to contamination of traditional foods

Global treaty negotiations on mercury began June 2010 (finish 2013? ; Stockholm Convention on POPs signed in 2001

Ongoing efforts to regulate emissions

Policy Context: International

Policy Context: U.S.

Guiding Questions

o How can we better understand the transport and fate of toxic atmospheric constituents, in ways relevant to policy? n  What are the pathways by which pollutant

emissions impact people? n  What are the feedbacks and interactions

between pollutants and society? n  How can we inform more effective policies?

Framework for assessing pollution and impacts

Environmental transport and transformations

Policy decisions and societal adaptations

Integrated policy-relevant assessment

Human activities creating pollution

Environmental and human impacts

Global Earth System Modeling

Economics and Social Science

Presentation Outline

o  Two in-depth examples using modeling tools to inform policy: n  Global challenges of legacy emissions: the

case of mercury n  Present and future Arctic transport of polycyclic

aromatic hydrocarbons (PAHs)

GEOS-Chem: Modeling Transport and Fate of Persistent Pollutants

Global, 3D tropospheric chemistry model, 4x5 degree resolution, assimilated meteorology

[Bey et al., 2001]

Mercury simulation includes land-atmosphere-ocean coupling (Selin et al., 2007, 2008; Strode et al., 2007; Holmes et al., 2010; Soerensen et al., 2010) POPs simulation includes atmospheric processes (so far…for PAHs, Friedman and Selin, 2012)

Understanding the Present and Future Global Biogeochemical Cycle of Mercury

[Selin, Ann. Rev. Env. Res., 2009]

Natural Processes

Human activities and policy interactions

Technical, policy and ecosystem uncertainty

Human and environmentimpacts

Soil lifetime: Role of Historic Mercury

Atmospheric Reactions of Mercury

[OH,  O3]  Br  

in-­‐cloud  photoreduc6on?  

Hg0   Hg2+  

Wet  and    Dry  Deposi6on  

1.6 ng m-3 1-100 pg m-3

Atmospheric lifetime 0.5-2 y Insoluble

Soluble

Measurements: TGM=Total Gaseous Mercury, RGM=Reactive Gaseous Mercury

Hg(P)

North American Contribution to Mercury Deposition

[Selin & Jacob, Atmos. Env. 2008]

Up to 60% of deposition in Midwest/Northeast U.S. is from domestic sources

Southeast has highest wet deposition in the U.S., but mostly from non-US sources: this is due to rainout of mercury from higher altitudes in summertime

Policy implications: Reducing deposition in both Midwest and Southeast will require policy actions on multiple political scales (national and global)

[Selin et al., GBC 2008]

Contribution to U.S. Deposition

22% International

32% Natural

20% U.S.

25% Historical

Present vs. Historical Sources of Mercury

•  Factor of 3 enrichment on average since pre-industrial times (constrained by sediment core records), but spatial variation

•  Historical legacy continues to affect ecosystems through deposition

From Deposition to Fish Methylmercury

[Engstrom, 2007]

Freshwater Deposition and Source Attribution

24.21 µg m-2 y-1 34.08 µg m-2 y-1

Pre-industrial + Historical

International Anthropogenic

N. American Anthropogenic

11%

23%

66% 59%

9%

32%

Northeast U.S. Southeast U.S.

Lake, River, Watershed, and Aquatic food web models [Knightes et al., 2009]

Policy and Timescale Analysis

How do sources affect fish methylmercury, and on what timescales?

[Selin et al., Environ. Health Persp., 2010]

Freshwater Ecosystem Timescale Analysis Fi

sh M

eHg

(ppm

)

Same deposition,but different ecosystem dynamics lead to very different source attributions (and concentrations) over time (watershed role)

Regional differences in deposition sources lead to different attributions in similar ecosystems

Note difference in scale!

Each ecosystem driven by present-day deposition for 40 years Policy experiment: All Hg is “historical” at t=0. How is anthropogenic signal reflected in fish, and on what timescale?

[Selin et al., EHP, 2010]

Ecos

yste

m B

Ec

osys

tem

A

Local Exposure from Freshwater Fish

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Stratified Seepage Stratified Seepage

MeH

g i

nta

ke (

ug

/kg

/d

ay)

Northeast Southeast

6.4

WHO intake threshold

EPA Reference Dose

North American anthropogenic

Historical+Natural International anthropogenic

Ecosystem A Ecosystem A Ecosystem B Ecosystem B

2 x 100 g fish meals/week (60 kg person) @ t=40 y

[Selin et al., EHP, 2010]

Population-Wide Hg Exposure from Marine Fish

No mechanistic link (yet) from oceanic Hg concentration to fish methylmercury

Historical exposure could continue to increase, complicating policy decision-making

Different challenges on different scales (local to global)

“current emissions” scenario 14-box ocean model: Sunderland and Mason, 2007 [Selin et al., EHP, 2010]

Adaptation and mitigation necessary? (Learning lessons from other issue areas)

How to explain recent trends in Hg?

r = 0.86 Lifetime = 0.67 d

Slemr et al. (2011): 20% declines in atmospheric Mercury at Cape Point, Mace Head since 1995. How to explain this? Several possibilities: -Change in emission

-ocean? -Change in oxidation

What sources influence concentrations?

Mace Head Cape Point

[Song and Selin, in prep]

A combination of factors is most plausible

Evaluation with 12-box model of global mercury cycle Coming soon: probabilistic analysis…. [Song and Selin, in prep]

GEOS-Chem POPs Simulation

Polycyclic Aromatic Hydrocarbons (PAHs)

Phenanthrene (PHE)

Pyrene (PYR)

Benzo[a]pyrene (BaP)

GAS-PHASE

PARTICLE-PHASE

r = 0.86 Lifetime = 0.67 d

Annual average benzo[a]pyrene vs. observations, mean 2005-2009

Emissions from Zhang and Tao [2009], GEOS-Chem at 4°x5°; includes gas-particle partitioning (to BC/OC), gas-phase oxidation by OH; wet/dry deposition; (particle-phase oxidation) [Friedman and Selin, ES&T, 2012]

Model Performance

0.0001  

0.001  

0.01  

0.1  

1  

10  

100  

0.0001   0.001   0.01   0.1   1   10   100  

Simulated

 (ng  m

-­‐3)  

Observed  (ng  m-­‐3)  

PHE  PYR  BaP  

Mean  total  (gas  +  parCculate)  concentraCons  at  background  locaCons  (2005-­‐2009)    PHE:  r  =  0.64  PYR:  r  =  0.72    BaP:  r  =  0.74    Atmospheric  lifeCmes  (hours):  PHE:  4  PYR:  3  BaP:  6  

q We capture mid-latitude and Arctic concentrations (roughly) q Our atmospheric lifetime is much lower than the “threshold”

criteria for regional/global action (2.8 days) q Simulation sensitive to: temperature sensitivity of partitioning,

on-particle oxidation [Friedman and Selin, ES&T, 2012]

We capture episodic transport to Spitsbergen

0  

0.005  

0.01  

0.015  

0.02  

0.025  

0.03  

0.035  

0.04  

1   53   105   157   209   261   313   365  

Day  of  year,  2007  

Modeled  -­‐  full  emissions  DetecCon  limit  Measured  

Concen

traC

on  (n

g  m

-­‐3)  

r  =  0.70  

BaP  

[Friedman and Selin, ES&T, 2012]

We capture episodic transport to Spitsbergen

0  

0.005  

0.01  

0.015  

0.02  

0.025  

0.03  

0.035  

0.04  

1   53   105   157   209   261   313   365  

Day  of  year,  2007  

Modeled  -­‐  full  emissions  DetecCon  limit  Measured  Measured  -­‐  interpolated  

Concen

traC

on  (n

g  m

-­‐3)  

r  =  0.70  

BaP  

[Friedman and Selin, ES&T, 2012]

European and Russian sources contribute the most to [BaP] at Spitsbergen

0  

0.005  

0.01  

0.015  

0.02  

0.025  

0.03  

0.035  

0.04  

1   53   105   157   209   261   313   365  

Day  of  year,  2007  

Modeled  -­‐  full  emissions  Modeled  -­‐  without  European  emissions  Modeled  -­‐  without  Russian  emissions  DetecCon  limit  Measured  Measured  -­‐  interpolated  

Concen

traC

on  (n

g  m

-­‐3)  

r  =  0.70    

BaP  

[Friedman and Selin, ES&T, 2012]

0  

0.002  

0.004  

0.006  

0.008  

0.01  

0.012  

0.014  

25   30   35   40   45   50   55   60  

Day  of  year,  2007  

Modeled  -­‐  full  emissions  Modeled  -­‐  without  European  emissions  Modeled  -­‐  without  Russian  emissions  DetecCon  limit  Measured  Measured  -­‐  interpolated  

Concen

traC

on  (n

g  m

-­‐3)  

European and Russian sources contribute the most to [BaP] at Spitsbergen

è  ~80  %  

[Friedman and Selin, ES&T, 2012]

Future climate will have the greatest impact on the number of BaP high pollution events

0  

100  

200  

300  

400  

500  

600  

PHE   PYR   BaP  

No.  event  >  present-­‐day  m

ean  +  1SD  

PAH  

Present  Future  

(1997-­‐2005)  (2047-­‐2055)  

[Friedman and Selin, in prep]

Future climate will have the greatest impact on atmospheric BaP concentrations

-­‐35  

-­‐25  

-­‐15  

-­‐5  

5  

15  

25  

35  

0  

100  

200  

300  

400  

500  

600  

PHE   PYR   BaP  

%  Change  in  m

ean  concentraCon  No.  event  >  present-­‐day  m

ean  +  1SD  

PAH  

Present  Future  Change  in  mean  concentraCon  

(2047-­‐2055)  (1997-­‐2005)  

[Friedman and Selin, in prep]

Future climate will have the greatest impact on BaP health risks

-­‐35  

-­‐25  

-­‐15  

-­‐5  

5  

15  

25  

35  

0  

100  

200  

300  

400  

500  

600  

PHE   PYR   BaP  

%  Change  in  m

ean  concentraCon  No.  event  >  present-­‐day  m

ean  +  1SD  

PAH  

Present  Future  Change  in  mean  concentraCon  

1  in  10  billion  (small)   6  in  100  billion  (small)  

3  in  100  million  (bigger)  

WHO:  LifeCme  lung  cancer  risk  =  BaP-­‐eq  (ng  m-­‐3)  ×  UR  where  UR  =  unit  risk  =  8.7  ×  10-­‐5,  or  8.7  cases  per  100,000  people  exposed  to  1  ng  m-­‐3  for  70  years  

(1997-­‐2005)  (2047-­‐2055)  

[Friedman and Selin, in prep]

What about Policy?

Policy interfaces imperfectly with scientific timescales (example: Mercury)

[Selin, JEM, 2011]

Improving Knowledge of Policy Responses to Pollutant Impacts

Negotiations for global mercury treaty began June 2010

Research questions: q How does scientific information

inform global environmental policies?

q What are best practices for scientists and engineers to have an impact?

Methods: Development of a “Mercury Game”– a negotiation simulation with dual goals: q  Understand the ways in which science is used in global policy q  Teach scientists and engineers about the process and how to

participate

Play the game! http://mit.edu/mercurygame

With Leah Stokes, Lawrence Susskind

The Selin Group 2012

•  Postdocs: –  Carey Friedman (PhD, URI): Transport and fate of persistent organic pollutants –  Tammy Thompson (PhD, U. Texas): Regional-to-global atmospheric chemistry

modeling

•  Graduate Students: –  Rebecca Saari, Eng Sys 3rd yr: Air pollution health impacts –  Ellen Czaika, Eng Sys 3rd yr: Sustainability decision-making –  Shaojie Song, EAPS, 2nd yr: Mercury –  Colin Pike-Thackray, EAPS, 2nd yr (1/2012): POPs –  Amanda Giang, Technology/Policy 2nd yr: Mercury –  Leah Stokes, Urban Studies/Planning 3rd yr: Mercury science-policy (primary

advisor: L. Susskind)

Acknowledgments: NSF: Atmospheric Chemistry Program, “CAREER: Understanding Chemistry, Transport and Fate of Mercury and Persistent Organic Pollutants through Global Atmospheric Modeling,”; MIT Research Support Committee Ferry fund; U.S. EPA: Science to Achieve Results (STAR) Program, "Air Pollution, Health and Economic Impacts of Global Change Policy and Future Technologies: An Integrated Model Analysis,”; MIT Joint Program on the Science and Policy of Global Change