1 Aerosols and Climate Peter J. Adams and Neil DonahueCenter for Atmospheric Particle Studies (CAPS) Carnegie Mellon University CEDM Annual Meeting May 25, 2017

2 Outline Aerosols 101 Their Climate ForcingsClimate Sensitivity and Aerosols Why Aerosol Forcings are Uncertain Health Effects and Hidden Warming Black carbon (soot) as short-term climate mitigation

3 Recall Definitions Aerosol = collection of particles (solid or liquid), suspended in a gas (=atmosphere for us) Particulate matter (PM) = aerosol Two National Ambient Air Quality Standards (NAAQS) PM2.5: the mass concentration of particles whose aerodynamic diameter is <= 2.5 microns PM10: … same but <= 10 microns PM2.5 is more likely to penetrate deeper in the lungs than particles larger than 2.5 microns

4 Composition Is ComplexInorganic ions: sulfate, nitrate, ammonium, others Organic carbon Elemental carbon (“soot”) Trace metals: Pb, V, Fe, Ni, Cu, etc. Crustal/mineral dust species: Si, Ca, Al, Fe, Ti, etc Sea spray: NaCl (and some organics) Biological particles: pollen, spores, etc. Fly ash Biggest contributors to PM2.5 (fine) mass conc. Health effects? Coarse PM, key natural sources, obvious in satellite imagery

5 Size Ranges Atmospheric particles vary greatly in size< 1 nm, you are probably a gas molecule > 10 mm, you won’t be in atmosphere very long (gravitational settling) factor of 104 in diameter, 1012 in volume from a cluster of a few molecules … to a cloud droplet Dp (diameter) 1 nm 10 nm 100 nm 1 mm 10 mm

6 Terminology *for regulatory purposes, 2.5 mm is cutoff between “fine” and “coarse” Natural, wind-blown sea spray, mineral dust Anthropogenic aerosol forcing “Aitken” mode / ultrafine mode “Fine” mode / accumulation mode Nucleation mode “Coarse” mode Dp (diameter) 1 nm 10 nm 100 nm 1 mm* 10 mm Sub-micron Super-micron

7 Key Processes: Particle FormationMaking big particles out of bigger “particles” (the Earth) 1) wind-blown emission (primary particles) high winds lift dust and sea spray into atmosphere 2) mechanically generated (primary particles) construction, abrasion of tires on roadway etc. 3) regional nucleation (secondary particles) high (super-saturated) levels of H2SO4(g) … plus “other stuff” (ammonia, amines, oxidized organics … still a research question) gas molecules cluster until a stable particle is formed 4) near-source nucleation (“primary” particles) at point of emission (e.g. tailpipe of car), emissions are gases but rapid cooling causes nucleation and growth of particles near source sometimes in source: e.g. combustion soot Making small particles out of gas molecules

8 Particle Evolution Condensation (and evaporation) Coagulationrecall gas-phase reactions that produce: H2SO4, HNO3, semi-volatile organic compounds these tend to condense onto particles, causing them to grow (1-20 nm/h is typical during day) Coagulation particles collide and stick due to Brownian (“random walk”) diffusive motion mostly a mechanism for ultrafine and nucleation mode particles to add to larger particles It should be clear that particles are mixtures of lots of stuff, not all from same source

9 Particle Removal Dry deposition Wet depositionmostly for super-micron particles (tD is ~1 day) Wet deposition mostly for Dp > 100 nm (tD is 4-8 days) *In some sense, coagulation is also “removal” (decreases N conc., M conc. remains the same)

10 Super-micron vs Sub-micronNucleation mode “Ultrafine” mode “Fine” mode / accumulation mode “Coarse” mode Freshly nucleated particles are smallest (by definition) Windblown and mechanically generated particles are bigger Dp (diameter) 1 nm 10 nm 100 nm 1 mm* 10 mm Near-source nucleation (from combustion) are a little bigger Sub-micron Super-micron

11 Super-micron vs Sub-micronThis drives a big difference in composition between sub-micron / super-micron Super-micron: wind-blown soil, sea spray, mechanically generated dust, tire wear, etc. Sub-micron: primary combustion particles and all secondary material produced by gas-phase chemistry inorganics: sulfate, nitrate, ammonium organics: POA, SOA (primary/secondary organic aerosol) black carbon (elemental carbon / soot)

12 Life Cycle of Super-micron (Boring)Nucleation mode “Ultrafine” mode “Fine” mode / accumulation mode “Coarse” mode Windblown and mechanically generated particles are bigger Particle emitted Negligible evolution due to condensation/coagulation Removed usually by dry deposition Dp (diameter) 1 nm 10 nm 100 nm 1 mm* 10 mm Sub-micron Super-micron

13 Life Cycle of Sub-Micron Particle (Interesting)Nucleation mode “Ultrafine” mode “Fine” mode / accumulation mode “Coarse” mode Freshly nucleated particles are smallest (by definition) Particles grow by condensation (1-20 nm/h) Ultrafine/nucleation mode particles lost by coagulation on timescale of hours Both processes drive mass to “accumulation” mode (slow dry deposition) Eventual removal by precipitation condensation Dp (diameter) 1 nm 10 nm 100 nm 1 mm* 10 mm Near-source nucleation (from combustion) are a little bigger Sub-micron Super-micron

14 Aerosols Scattering SunlightDust and smoke over Australia (Terra)

15 Aerosols Absorbing Sunlight (Direct)photo courtesy of Jay Apt Kuwaiti oil fires

16 Clouds and Climate Atmosphere 342 77 67 30 168 Earthfluxes are W m-2 Width of arrow proportional to flux Atmosphere 342 77 67 (reflected) (absorbed) 30 168 Earth 23% of incoming sunlight reflected by atmosphere (mostly by clouds) Without cloud reflection, the Earth would be ~15º C warmer

17 Aerosol Cloud Reflectivity EffectClean air mass Lower CCN concentration Higher Transmittance Lower reflectivity (albedo) Better chance of precipitation? Shorter cloud lifetime? Polluted air mass Higher CCN concentration Lower Transmittance Higher reflectivity Less precipitation? Longer cloud lifetime?

18 Cloud Optics: Surface AreaFor a given amount of liquid water (or ice): More pollution/CCN → More cloud droplets → More surface area → More scattering → Brighter cloud → Cooler Earth Darker clean cloud (Few CCN) Brighter polluted cloud (More CCN) photo courtesy of Amy Sage

19 Cloud Condensation Nuclei (CCN)In a particle-free atmosphere, a strong supersaturation (~400% relative humidity) is required to nucleate new liquid droplets Instead, cloud water condenses onto pre-existing particles: cloud condensation nuclei (CCN) Clear Sky (RH < 100%) Cloudy Sky (RH > 100%) Activation: water condenses on CCN to form cloud droplets Other particles (aerosols) Cloud droplets (~10 mm) CCN (~100 nm)

20 Aerosol Activation “Activation” = formation of cloud dropletinvolves a competition between solute and surface tension effects Depends on number concentration above “critical diameter” Number Diameter

21 Aerosols and Climate: Semi-Direct EffectAtmospheric heating from absorptive aerosol Warmer air  lower RH  less cloud formation Atmospheric Heating High RH Low RH

22 How direct is direct? Direct effect: scattering/absorbing sunlightSemi-direct effect: aerosol absorption heats atmospheric layer warmer air → lower relative humidity → less/no cloud forcing or feedback? Indirect effect: modifying cloud properties “brightness (first) effect” “lifetime (second) effect”

23 Other Forcings BC on snow/ice surfaces BC as ice nucleiMakes surface darker and more absorbing Relatively small global forcing May be regionally important BC as ice nuclei Ice nuclei are important triggers of precipitation May change cloud cover, affect hydrological cycles Almost nothing very concrete known about this effect

24 Radiative Forcing (W m-2)Long-lived GHGs: +3 W/m2 (+/- 20%) Aerosol effects: mostly cooling, highly uncertain Radiative Forcing (W m-2)

25 Climate Change Uncertainty“Climate sensitivity” is a key parameter Key parameter is l,“climate sensitivity” 0.3 to 1 °C per W/m2 °C for doubling of CO2 Implicit assumption that l is same for all kinds of forcings (absorbing aerosols are known exception) Boundary between “forcing” and “feedback” is fuzzy sometimes In climate models, representation of cloud feedback is largest source of uncertainty In retrospective studies, knowledge of aerosol forcing is lacking global average temperature change global average radiative forcing

26 Aerosols and Climate UncertaintyGHG forcing High sensitivity Aerosol (haze) + GHG forcing Low sensitivity 20th century T increase

27 Climate Models: Sensitivity / AerosolsFigure from Kiehl et al., GRL v34 doi: /2007GL

28 Climate Models: Sensitivity / AerosolsFigure from Kiehl et al., GRL v34 doi: /2007GL

29 intra-hemispheric mixing Mean aerosol residenceChallenges Need to characterize particle mass/number concentration size distribution: ~10 nm to 10 mm chemical composition: >hundreds compounds mixing state interactions with clouds (sub-grid) separate anthropogenic from natural Highly variable in space and time: intra-hemispheric mixing Mean aerosol residence Mean CO2 residence NH/SH mixing century decadal annual daily monthly hourly

30 Aerosol Variability

31 Uncertainty in pre-industrial/natural baseline matters

32 20th Century Forcings

33 Temperature Reconstruction

34 Temperature Reconstruction

35 Temperature Reconstruction

36 Temperature Reconstruction

37 Temperature Reconstruction

38 PM and Life Expectancy: 1980

39 PM and Life Expectancy: 2000

40 Global Distribution of PM2.5

41 Global Mortality from PM2.5

42 Hidden Warming from China’s AerosolsChina is ~25% of world emissions for anthropogenic aerosol precursors IPCC AR5 Table 7.1 Global aerosol forcing is -0.8 W/m2 (best guess) … so China is -0.2 W/m2 -2 W/m2 (coolest) … so China is -. 5 W/m2 Climate sensitivity 1.5 to 4.5 deg C for 2xCO2 (~4 W/m2) 0.3 to 1.0 K/(W/m2) Multiplying by climate sensitivity… Best guess: China is -0.2 W/m2 x 0.6 K/W/m2 = deg C Worst case: so China is -. 5 W/m2 x 1.0 K/W/m2 = -0.5 deg C

43 Won’t Climate Response Take a Long Time? … NoInstantaneous global desulfurization (red) Surface air T Y2000 levels of sulfate and GHGs held constant (blue)

44 Won’t Climate Response Take a Long Time? … NoSurface layer of ocean adjusts quickly (0.8 K warming in one decade) More gradual response as deep ocean equilibrates Surface air T

45 Black Carbon as Climate Mitigation?Warming Cooling Black carbon sources contain cooling/scattering aerosol contribute to cloud condensation nuclei Sunlight Absorption Cloud Brightening Cloud Burnoff Co-emitted Reflectors Snow/Ice Darkening How much cooling goes with the warming? Kuwaiti oil fires (photo courtesy of Jay Apt)

46 Timing BC mitigation would lead to fast benefitsSource: UNEP report, 2011

47 Climate Effects BC 0.05 to 0.55 W/m2 (IPCC) but maybe ~1 W/m2?Not easily separable from cooling aerosols This is cut and pasted as PARTS of figure 2.21 in the AR4, I selectively chose what I thought as relevant. The true figure is very tall, and includes short lived gases and is very hard to have at an appropriate resolution (See to the right) Source: IPCC AR4

48 Co-Emitted Species: Early DebateResponse from Joyce Penner (selected quotes) he obtains a warming from the combined direct and indirect effects of f.f. BC + OM that is inadequately documented warming may disagree with the results of published models by the aerosol-climate community. Penner et al. [2003] … obtain a forcing for f.f. BC + OM that is not significantly different from zero Jacobson [2002] has an inferred forcing for f.f. BC + OM that is 0.5 W/m2

49 Co-Emitted Species Most recent comprehensive analysis (Bond et al., 2012) shows co-emitted cooling could completely offset BC warming Source: Bond et al., 2012

50 Warming more likely than notCo-Emitted Species Different BC sources have different warming potential A, H, RC all different models Warming more likely than not Look into A,H,RC technical difference Source: Kopp and Mauzerall, 2010

51 Net Warming/Cooling By SourceDiesel: + ~.15 W/m2 “high confidence in net positive total climate forcing is possible only for black-carbon source categories with low co-emitted species, such as diesel engines.” (Bond et al. 2013) Biofuel cooking: + but small Residential coal: + but small Biomass burning: net cooling ~.2 W/m2 from right sources → 0.06 to .2 K

52 BC Controls Reduce CDNCIn global annual average, 50% FF: CDNC reduced by 4.6% 50% CARB: CDNC reduced by 8.7% CDNC [cm-3 ] Smax [%] Reff [μm] 195.6 0.26 8.27 In these simulations, we linked the CCN distributions to cloud droplet number concentrations using the activation parameterization of Thanos Nenes. What you see here is the ratio of CDNC in the half-fossil fuel case compared to the base case. For that scenario, CDNC decreased by almost 5% in the global average and you can see the more detailed map here. When we cut all carbonaceous emissions, we get a stronger effect: almost a 9% reduction in cloud drop number. These are significant reductions in the AIE that, by themselves, would lead to some warming. Ratio of cloud droplet number (CDNC) 50% FF Base case

53 Black Carbon “Warming and Drying”Aerosol Forcing (W m-2) Governs global mean T -0.5 W m-2 +2.0 W m-2 Mention that GHG’s do “Warming and Wetting” Reduces evaporation -2.5 W m-2 Ramanathan et al., Science 294, , 2001.

54 Summary Estimating aerosol climate forcing is toughSimilar/related to cloud feedback problem Estimates of aerosol forcing and climate sensitivity are correlated Strong aerosol cooling is bad news Implies high climate sensitivity, lots of hidden warming (e.g. China) Mitigating black carbon is worthwhile but expectations for climate benefits should be limited