Connecting Air Pollution, Climate Change, Energy and Health

J. Jason West Department of Environmental Sciences & Engineering University of North Carolina, Chapel Hill

Beijing – January 14, 2013 Some stations reported >500 µg m-3 24-hr avg. PM2.5 (>20 times the WHO guideline)

• Air pollution is underappreciated for global health. • Air pollution and its health impacts are changing globally, and will change in ways interrelated with climate change. • Air pollution science offers new possibilities: new measurement methods measuring more chemical components, cheap sensors that can be widely deployed, satellites, and models. • There is a need for the communities of air pollution science and air pollution health effects science to work together better.

Modeling Global Air Pollution & Health Chemical Mechanism

Meteorology

Concentrations

Emissions

Health Function Health Impacts

Typical horizontal resolutions: Global model – 5 to 0.5 degrees Regional model (continental) – 50 to 10 km

Incidence Data

Health impact function ∆Mort = y0 x AF x Pop ∆Mort = y0 x (1-exp-βΔX) x Pop Exposed Population

Baseline mortality rate

(≈ CPD) - Respiratory diseases (RESP) (inc. COPD – chronic obstructive pulmonary disease)

- Cardiovascular diseases (inc. IHD – ischemic heart disease, STROKE – cerebrovascular disease)

ΔX = Change in concentration β = Concentrationresponse factor

- Lung Cancer (LC) 7

Population and Baseline Mortality Rates Total Population, persons (Landscan 2011 at 30”x30” gridded to 0.67°x0.5°)

Baseline Mortality Rates, deaths per year per 100,000 (GBD 2010, country level, AllAges > gridded to 0.67°x0.5°)

IHD

Stroke

COPD

LC

6,946 million

Population 25+, persons (Landscan 2011 at 30”x30” gridded to 0.67°x0.5°)

3,839 million

8

-20% Global Anthrop. Methane Emissions: 30,200 avoided premature deaths in 2030 due to reduced ozone

West et al., PNAS, 2006

Ozone from N. American and European emissions causes more deaths outside of those regions than within

Avoided deaths (hundreds) from 20% regional ozone precursor reductions, based on HTAP simulations, Anenberg et al. (EST, 2009)

Global mortality burden – ACCMIP ensemble Ozone-related mortality 470,000 (95% CI: 140,000 - 900,000)

PM2.5-related mortality(*) 2.1 million (95% CI: 1.3 - 3.0 million)

(*) PM2.5 calculated as a sum of species (dark blue) PM2.5 as reported by 4 models (dark green) Light-colored bars - low-concentration threshold (5.8 µg m-3)

Silva et al. (ERL, 2013)

11

Global Burden: Ozone-related mortality Global and regional mortality per year Regions

Total deaths

North America Europe Former Soviet Union Middle East India East Asia Southeast Asia South America Africa Australia

34,400 32,800 10,600 16,200 118,000 203,000 33,300 6,970 17,300 469

Deaths per million people (*) 121 96 66 68 212 230 119 38 73 29

472,000

149

Global

Respiratory mortality , deaths yr -1 (1000 km2)-1, multi-model mean in each grid cell , 14 models

(*) Exposed population (age 30 and older)

Silva et al. (ERL 2013)

12

Global Burden: PM2.5-related mortality Global and regional mortality per year Regions North America Europe Former Soviet Union Middle East India East Asia Southeast Asia South America Africa Australia

Global

Deaths Total per deaths million people (*) 43,000 152 154,000 448 128,000

793

88,700 397,000 11,049,000 158,000 16,800 77,500 1,250

371 715 1,191 564 92 327 78

2,110,000

665

CPD+LC mortality , deaths yr -1 (1000 km2)-1, multi-model mean in each grid cell , 6 models

(*) Exposed population (age 30 and older)

Silva et al. (ERL 2013)

13

Global burden of disease of air pollution (2013) 5.5 million deaths per year (95% CI: 5.1 – 5.9 million)

~10% of all deaths globally!

Global Deaths per Year Ambient PM2.5 pollution:

2.9 (2.7 – 3.1) million

Household air pollution from solid fuels: 2.9 (2.5 – 3.3) million Ambient ozone pollution:

0.22 (0.16 – 0.27) million Forouanzafar et al., 2015

GBD: Global Deaths per Year (2013) Ambient PM2.5 pollution: 2.9 (2.7 – 3.1) million -> 7th most important risk factor for deaths, 5th in East Asia. Household air pollution from solid fuels: 2.9 (2.5 – 3.3) million -> 8th most important risk factor for deaths, 2nd in South Asia. Ambient ozone pollution:

0.22 (0.16 – 0.27) million

GBD: US Deaths per Year (2013) Ambient PM2.5 pollution:

79,000 (70,000 – 89,000) ~3% of all US deaths! Household air pollution from solid fuels: 0 Ambient ozone pollution:

12,000 (4,000 – 21,000)

Other US estimates Ambient PM2.5 pollution: 130,000

66,000 (39,000 – 85,000) (Punger & West, 2013) (Fann et al., 2012)

Ambient ozone pollution: 4,700

21,000 (6,000 – 34,000) (Punger & West, 2013) (Fann et al., 2012)

Ozone-related mortality (sectors zeroed-out) Contributions of each sector to total ozone respiratory mortality, fraction of total burden in each cell

Energy

Global total: 65,200 deaths/year

Land Transportation

Global total: 80,900 deaths/year

Industry

Global total: 45,600 deaths/year

Residential & Commercial

Global total: 53,700 deaths/year

Silva et al. (EHP, submitted)

17

PM2.5-related mortality (sectors zeroed-out) Contributions of each sector to total PM2.5 mortality (IHD+Stroke+COPD+LC), fraction of total burden in each cell

Energy

Industry

Global total: 290,000 deaths/year

Land Transportation

Global total: 212,000 deaths/year

Global total: 323,000 deaths/year

Residential & Commercial

Global total: 675,000 deaths/year

Silva et al. (EHP, submitted)

18

Future ozone-related mortality - ACCMIP

Global Respiratory Premature Ozone Mortality: 2030, 2050 and 2100 vs. 2000 conc. - Uncertainty for the ensemble mean is a 95% CI including uncertainty in RR and across models. Silva et al. (ACPD, 2016)

19

Future PM2.5-related mortality - ACCMIP

Global CPD+LC Premature PM2.5 Mortality: 2030, 2050 and 2100 vs. 2000 conc.

-Uncertainty for the ensemble mean is a 95% CI including uncertainty in the RR and across models. – Silva et al. (ACPD 2016)

20

Future PM2.5-related mortality - ACCMIP IHD+Stroke+ COPD+LC Premature PM2.5 Mortality

RCP2.6

6 models

RCP8.5

6 models

IHD+Stroke+COPD+LC Premature PM2.5 Mortality – 2100 (two scenarios)

-1000

-100

-10

-1

-0.1

-0.01

0.01

0.1

1

10

100

1000

deaths yr-1 (1000 km2)-1

Silva et al. (ACPD 2016)

21

Global air pollution burden on mortality Ozone

PM2.5 IHD+Stroke+COPD+LC mortality (million deaths/year)

Respiratory mortality (million deaths/year)

RCPs

Concentrations

Population

Baseline MR

2000 minus 1850

Present-day

Present-day

2000 minus 1850

Future

Future

Future minus 1850

Future

Future Silva et al. (ACPD 2016)

22

Climate Change – What’s certain 1) Greenhouse gases cause warming by absorbing infrared radiation. John Tyndall (1820-1893) – Conducted lab experiments in 1860s and discovered that CO2, CH4, H2O are greenhouse gases, but N2, O2, Ar are not.

Climate Change – What’s certain 2) Greenhouse gas concentrations are increasing, caused by human emissions.

Climate Change - What’s certain 3) Global temperatures are increasing

NASA GISS

Climate Change Attribution: High Confidence

“It is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century.” IPCC, 2013

Climate Change – Uncertain Future

Year 2100 under no climate policy: 4 (3-5.5) °C = 7 (5-10) °F.

IPCC, 2013

Paris climate agreement: historical backdrop • 1992 – United Nations Framework Convention on Climate change • 1997 – Kyoto Protocol • 2001 – US pulls out of the Kyoto Protocol • 2010 – Copenhagen pledges • Summer 2015 – US & China announce pledges • 2015 – Paris Accords

Paris Accords (Dec. 2015) • Not legally binding. • Ambitious targets to reduce emissions by 2030 by most nations globally. • Plan to meet every 5 years to renew pledges. • Aspirational goal of limiting temperature change to 1.5°C – implies ZERO emissions by roughly 2070. ** Paris requires countries to follow through on their commitments. In US, that is the Clean Power Plan.

Consequences of Paris agreement

Fawcett et al., 2015

Impact of RCP8.5 Climate Change on Global Air Pollution Mortality: ACCMIP Models OZONE

Million deaths yr -1

Mean (95% CI)

(thousands deaths . yr-1)

2030 11 (-30, 86)

2100 127 (-193, 1,070)

PM2.5 Million deaths yr -1

Mean (95% CI)

(thousand deaths . yr-1)

2030 56 (-34 , 164)

2100 215 (-76 , 595) Silva et al., in prep.

Co-benefits of GHG Mitigation for Air Quality 1) Immediate and Local Air pollutants Sources & Policies GHGs

Air pollution

Climate Change

Human Health

2) Long-Term and Global

Objective: Analyze global co-benefits for air quality and human health to 2100 via both mechanisms.

Approach Years

Emissions GCAM

Meteorology GFDL AM3

Name

2000

2000

2000

2000

GCAM Reference

RCP8.5

REF

RCP4.5

RCP4.5

RCP4.5

GCAM Reference

RCP4.5

eREFm45

2030, 2050, 2100

• Use the GCAM reference for emissions rather than RCP8.5, for consistency with RCP4.5. • Simulations conducted in MOZART-4. - 2° x 2.5° horizontal resolution. - 5 meteorology years for each case. - Fixed methane concentrations. - Compares well with ACCMIP RCP4.5.

Co-benefits – PM2.5 Concentration

Global population-weighted, annual average PM2.5 West et al. NCC 2013

Co-benefits – Global Premature Mortality Projection of global population and baseline mortality rates from International Futures.

PM2.5 co-benefits (CPD + lung cancer mortality) 2030: 0.4±0.2 million yr-1 2050: 1.1±0.5 2100: 1.5±0.6

Ozone co-benefits (respiratory mortality) 2030: 0.09±0.06 2050: 0.2±0.1 2100: 0.7±0.5 West et al. NCC 2013

Co-benefits – Valuation of Avoided Mortality

Red: High valuation (2030 global mean $3.6 million) Blue: Low valuation (2030 global mean $1.2 million) Green: Median and range of global C price (13 models) West et al. NCC 2013

Downscaling Co-benefits to USA (2050) RCP4.5 - REF PM2.5 (annual avg.) US mean = 0.47 µg/m3

(a)

Ozone (May-Oct MDA8) US mean = 3.55 ppbv

(b)

Zhang et al. ACPD, 2016

Downscaling Co-benefits to USA (2050) PM2.5 0.35 µg/m3

Ozone 0.86 ppb

Domestic

Most PM2.5 cobenefits from domestic reductions.

Foreign

0.12 µg/m3

2.69 ppb

Most ozone co-benefits from foreign and methane reductions.

Zhang et al. ACPD, 2016

Domestic vs. Foreign Co-benefits: PM2.5 Domestic (20800 deaths/yr)

Foreign (4600 deaths/yr)

Ø Domestic GHG mitigation accounts for 85% of the total avoided PM2.5 mortality.

Zhang et al. in prep

Domestic vs. Foreign Co-benefits: O3 Domestic (4600 deaths/yr)

Foreign (7600 deaths/yr)

Ø Foreign countries’ GHG mitigation accounts for 62% of the total avoided deaths of O3.

Zhang et al. in prep

US Co-benefits in 2050 • Avoided premature deaths from GHG mitigation: 24500 (CI: 17800-31100) from PM2.5, and 12200 (CI: 5400-18900) from O3. • Avoided heat stress mortality from RCP4.5 relative to RCP8.5: 2340 (CI: 1370-3320) (Ying Li). • Monetized co-benefits in 2050 are $74 (46-101) per ton CO2 reduced at low VSL, $220 (140-304) at high VSL. • Foreign GHG mitigation accounts for 62% of the total avoided deaths from O3, and 15% for PM2.5. • Previous regional or national co-benefits studies may underestimate the full co-benefits of coordinated global actions. • U.S. can gain significantly greater co-benefits, especially for ozone, by collaborating with other countries to combat climate change. Zhang et al. in preparation

Thank you Contributions from: Students/Postdocs: Raquel Silva, Yuqiang Zhang, Yasuyuki Akita, Zac Adelman, Meridith Fry, Susan Anenberg, CiaoKai Liang Collaborators: Steve Smith, Vaishali Naik, Larry Horowitz, Drew Shindell Jean-Francois Lamarque, Jared Bowden, Arlene Fiore, ACCMIP modelers, HTAP modelers Funding Sources: • EPA STAR Grant #834285 • NIEHS Grant #1 R21 ES022600-01 • EPA Office of Air Quality Planning and Standards • Portugal Foundation for Science and Technology Fellowship • EPA STAR Fellowship • UNC Dissertation Completion Fellowship • US Department of Energy, Office of Science • NOAA GFDL for computing resources

UNC Climate Health and Air Quality Lab www.unc.edu/~jjwest

834285

Projecting Baseline Health

2010 from WHO GBD Future from International Futures (mortality rates are at country level)

12-km CMAQ

Global model 2.8º resolution

Bias in US Deaths from PM2.5 80,000

CMAQ Model Output (12km-grid)

Deaths/Year

60,000

40,000

30~40% lower than

20,000

0

12km grid estimate 0

100

200 300 Grid Resolution (km)

400

Punger and West (AQAH, 2013)

46

US Bias is different for Satellite PM2.5! 12

Satellite

PWC (µg/m3)

10 8

CMAQ

6 4 2 0

0

100

200 300 Grid Resolution (km)

400

500

Akita et al. in prep.

Representative Concentration Pathways (RCPs) CO2 concentration (ppmv)

Sulfur emissions (TgSO2.yr-1)

1000

130

800

90

600

50

400 200

2000

2030

2050

2100

10

2000

2030

2050

NOx emissions (TgNO2.yr-1)

CH4 emissions (TgCH4.yr-1)

160

1000

120

700

80

400

40

100

2000

2030

2050

2100

2000

2030

2050

2100

2100

Source: RCP Database - http://tntcat.iiasa.ac.at:8787/RcpDb/dsd?Action=htmlpage&page=compare 48

Future ozone-related mortality (II) Respiratory premature ozone mortality - 2030, 2050, 2100 vs. 2000

RCP2.6

11 models

RCP8.5

13 models

Respiratory premature ozone mortality - 2100 (two scenarios)

deaths yr-1 (1000 km2)-1 -1000

-100

-10

-1

-0.1

-0.01

0.01

0.1

1

10

100

1000

Silva et al. (ACPD 2016)

49

Our co-benefits approach: advantages • First co-benefits study to use a global atmospheric model. – Capture effects of long-range transport and methane.

• First to estimate co-benefits by two mechanisms. • Use consistent future scenarios built on RCP4.5: emissions, population, economics (valuation). • By embedding the US study within a prior global study, we capture US co-benefits from foreign reductions.

Results – PM2.5 Concentration Annual average PM2.5

Total change RCP4.5 - REF

Meteorology eREFm45 - REF

Emissions RCP4.5 – eREFm45

2050

2100

West et al. NCC 2013

Results – Ozone Concentration

Global population-weighted, max. 6 month average of 1 hr. daily max ozone

West et al. NCC 2013

Results – Ozone Concentration Max. 6 month average of 1 hr. daily max ozone Total change RCP4.5 - REF

Meteorology eREFm45 - REF

Total co-benefit

#2 Meteorology

Emissions RCP4.5 - eREFm45

2050

#1 Emissions

2100

West et al. NCC 2013

Results – Global Premature Mortality

West et al. NCC 2013

Co-benefits: conclusions • Global abatement of GHG emissions brings substantial air quality and human health co-benefits. • Global GHG mitigation (RCP4.5 relative to REF) causes 0.5±0.2 million avoided deaths in 2030, 1.3±0.5 in 2050, and 2.2±0.8 in 2100 • Global average monetized co-benefits are $50-380 / ton CO2 – Greater than previous estimates – Greater than abatement costs in 2030 and 2050.

• The direct co-benefits from air pollutant emission reductions exceed those via slowing climate change. West et al. NCC 2013