linking regional aerosol emission changes with multiple ... · objective 4: iterate...

Photo: NASA

Linking Regional Aerosol Emission Changeswith Multiple Impact Measures through

Direct and Cloud-Related Forcing Estimates

Tami C. Bond, Yanju Chen, and Kevin Hade University of Illinois at Urbana-Champaign

Xin-Zhong Liang & Hao He University of Maryland

David G. Streets, Ekbordin Winijkul, and Fang Yan Argonne National Laboratory

Praveen Amar, Danielle Meitiv Clean Air Task Force

Project Design

2

The initial project concept remains:Find a dose-response curve for the atmosphere.

Forcing

Emission

As shown in May 2012, Nov 2014 Project Design Bond et al.: Aerosol Emission Changes & Metrics 3

The project period kicked off by finishing “Bounding-BC”

Bond et al. JGR 2013 Figure by D. Fahey

Project Design Bond et al.: Aerosol Emission Changes & Metrics 4

 

 

 

 

Bounding-BC fingered the big uncertaintyin BC-rich sources

� BC � direct forcing ~ bounded � BC � cloud forcing

~ large uncertainties – especially in ice/mixed � OC + SO4 � direct forcing

~ small for BC-rich sources � OC + SO4 � cloud forcing

~ large and probably negative

It’s the indirect effects of co-emitted species that cause big questions about immediate forcing

Review from May 2012, Nov 2014 Project Design Bond et al.: Aerosol Emission Changes & Metrics 5

Despite all the scientific complexity,

6

So you got [some scientific thing]

right. Who cares? Tell me if I should

turn this off!

Can you wait 6 months? I have

to run my model…

Bond et al.: Aerosol Emission Changes & Metrics

policy discussions need simplicity

Project Design

This simplicity is distilled into climate “metrics”

Normal people think: A metric is something you can measure, and report

The climate policy community says: A metric is a well-defined calculation that can be used to equate impact of a mass emission of some species to a mass emission of the big bear, CO2


Climate metrics for short-lived species can be calculatedfrom a single measure

Absolute global warming potential Global warming potential

Global temperature potential Specific forcing pulse

For short-lived species (τ<4 mo),ENF Emission-Normalized Forcing is the

only model output required to calculate any of these metrics.forcingENF = Other considerations affect the values of emission metrics, emission but they all come from models of the carbon cycle or Earth’s heat capacity, NOT from models of aerosols


Thought process for project design

For use in climate metrics, we need emission-normalized forcing for: -black & organic carbon - direct & cloud-related forcing

And the aerosol-cloud interactions aren’t any good if the aerosol size and number isn’t right.

Also, we should check with policy-makers if we are making this too complicated to understand.The cloud-related forcing

isn’t any good if the clouds aren’t right.


Project objectives reflected those considerations


And the aerosol-cloud interactions aren’t any good if the aerosol size and number isn’t right.

Objective 2: Employ an ensemble of parameterizations in regional-scale models to identify best estimates and uncertainties for direct and cloud-related forcing

Also, we should check with policy-makers if we are making this too complicated to understand.




Objective 1: Develop size-resolved, speciated emission inventories of aerosols and aerosol precursors





Objective 3: Determine functional relationships that express changes in direct and cloud radiative forcing as a function of emission changes in particular locations






Objective 3: Determine functional relationships that express changes in direct and cloud radiative forcing as a function of emission changes in particular locations


Objective 4: Iterate emission-to-forcing measures as communication tools between decision makers and climate scientists



Size-resolved inventories

Objective 1: Get the size right

(David Streets, Ekbordin Winijkul – Argonne)

Size-resolved inventories Bond et al.: Aerosol Emission Changes & Metrics 14

Global size-resolved emission inventory has been produced.

Size-resolved global emission inventory of primary particulate matter (PM) from energy-related combustion sources E. Winijkul, F. Yan, Z. Lu, D. G. Streets, T. C. Bond, Y. Zhao Atmos Env, 28 August 2015


This work includes uncertainty and illustrative reductionscenarios

Residential: Industrial: Switching from Baghouses on cement kilnssolid fuel to LPG

Winijkul et al., Atmos Env., 2015


The connections still need work.

Objective 3: Determine functional relationships that

changes in particular locations


express changes in direct and cloud radiative forcing as a function of emission

Disconnect: Most models are not ready to accept spatially-dependent,

size-resolved emissions.


Teaser: We are now using aerosol-resolved models toestimate plume-exit composition and CCN.

10-2 10-1 100 0

1

2

3

4

5 x 1015

Supersaturation [%]

EI C

CN

[kg -1

]

OC BC Inorganic BC & OC BC & Inorg OC & Inorg

There’s a limit on CCN emission.

Mena, Fierce, Bond & Riemer, in prep for ACP


Regional Cloud Modeling

Objective 2: Get the Clouds Right

(Hao He, Xin-Zhong Liang – Univ of Maryland) Regional cloud modeling Bond et al.: Aerosol Emission Changes & Metrics 19

eeee ooooo oooo666666

Aerosol(106)

Cloud(1010)

n

We used CWRF with an ensemble model to choose one combination of cloud-aerosol-radiation.

cst ( 3)

ccs ( 4)

ccb ( 5)

cci ( 2)

bin ( 3)

rel ( 6)

rei ( 7)

rer ( 1)

liquid ( 7)

ice (16)

rain ( 3)

graup ( 1) snow ( 1)

liquid ( 4)

ice (17)

rain ( 2)

graup (1)

snow ( 1)

emis ( 3)

binding ( 3)

overlap ( 3)

types (13)

bgd (1)

sul (2)

ssa (2)

dst (2)

cab (2)

tot (3)

bgd (1)

sul (2)

ssa (2)

optics (1152)

vertical profiles

(8)

(72)

(16) dst (2)indirect effects (18) ccn (6)

cab (2)

cloud radius (3)

cover (360)

optics (137088)

water (3)

radius (42)

(336)

(408)

geometry (9)

mosaic (9)

(4000 8)

(333333333 6)

Cloud

Radiation

AAAAAAAAA AAAA AAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAA AAAAAA AAAAAAAA AAAAAAAeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr rrrrrrrrrrroooooooooooooooooooooooooooooooooooooooooooooooo ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssoooooooooooooooo oooooooooooooo ooooooooooooooooooooooooooooooooooooooooooooooooooooooooo lllllllllllllllllllllllllllll (((((((((((((((((((( 1111100000000000000000000 6666666666666666 66666666666666666666666666666666666666666666666 )))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) Aerosol

(1010) (106)

(102)

1018

configurations

earth orbit

surface characteristics radiative gases

n C

However, when clouds were right, aerosol wasn’t, so we used offline aerosol fields (not ideal).

Regional cloud modeling Bond et al.: Aerosol Emission Changes & Metrics 20

Comparison between modeled and observed fluxes(average over Continental US)Error bars are std dev of all grid boxes

(MISR)

Regional cloud modeling Bond et al.: Aerosol Emission Changes & Metrics 21

Biases appear lower (but not perfect) withGOCART fields

22

TOA

Flu

x C

loud

rad.

forc

ing

Bond et al.: Aerosol Emission Changes & Metrics Regional cloud modeling

ISCCP and MISR ISCCP and GOCART

The connections still need work.

Objective : Determine functional relationships that express changes in direct and cloud radiative forcing as a function of emission changes in particular locations

Disconnect: Model components are

getting more rigid– difficult to switch in

components that work best

Objective 2: Employ an ensemble of parameterizations in regional-scale models to identify best estimates and uncertainties for direct and cloud-related forcing Size-resolved inventories Bond et al.: Aerosol Emission Changes & Metrics 23

Emission-to-forcing measures

Objective 3: Model Interpretation for Policy Relevance

(Yanju Chen– Univ of Illinois) Emission-to-forcing Bond et al.: Aerosol Emission Changes & Metrics 24

 

 

25

Basic approach: Global model, regional reductions

Reported previously: � Used Community Atmosphere Model (CAM), version

modified for polar transport � Determined that 30°x30° is optimum aggregation region

��

�� Emission

region

Extensive assistance from H. Wang, P. Rasch at PNNL

Emission-to-forcing Bond et al.: Aerosol Emission Changes & Metrics 25

We developed a “linearity diagnostic” R

50% present-day

100% present-day

emission R =

F100 − F50

F100 − F0

emission

0 emission

R� 0.5: Forcing is linear in emission. R< 0.5: Small emission change from present-day

produces less forcing change than one would expect


Direct forcing is linear, as expected

Not much to see here. Just checking.

Greater response than expected

Lower response

than expected


Indirect forcing is nonlinear and lower than expected

Greater response than expected

Lower response

than expected

Cause: Saturation; CCN not activating

Not a new finding, but we didn’t know magnitude. Asian OC CRF is about 40% less negative than simple proportion would predict.


Policy-relevant metrics

Objective 4: Tell the Story

Praveen Amar, Danielle Meitiv– Clean Air Task Force Kevin Hade, Tami Bond – University of Illinois

Policy-relevant metrics Bond et al.: Aerosol Emission Changes & Metrics 29

 

    

 

    

Common metrics have common failings

� Global Warming Potential• Integrated forcing per emission• Doesn’t take Earth system inertia into account• Doesn’t communicate immediacy• Requires choice of time horizon

� Global Temperature Potential• Temperature change at single year in future • Doesn’t communicate trajectory experience

• Doesn’t communicate immediacy• Requires choice of time

 

 

 

Two sets of surveys conducted

� Original goal: Survey state decisionmakers

� Survey 1: 35 policy-oriented people and scientists

� Survey 2: Eight city managers (where the climate action is now)

2 reports have been communicated to EPA


 

 

 

Main messages had nothing to do with metrics

� Non-specialists need simple ways to communicate black carbon’s effects to non-specialists• Even terms like “radiative forcing” and “feedback” are

not as straightforward as you think. � People want to hear about certainty, not uncertainty.


Scientists understand the importance of GWP timehorizon…

images: smh.com.au, dalje.com

…but policymakers don’t care


 

 

 

The challenge

�Communicate immediacy …without minimizing importance of CO2

�Communicate timing �Make it relevant

Also, scenarios & emissions should have parallel

treatment?!

Our solution… which WILL be redacted from this presentation…

involved a simple energy-balance model that was calibrated against MAGICC


TP = (Tscen,i −Tref ,i )i∑

generational Integrated Temperature Perturbation

gI fsceenn ,,ii rree ffee ,,ii i

scen i refe i

Super simple. Sum of temp changes across 1 generation.

Following Peters et al. 2011 Can be used for both emission metrics & scenario


Avoided gITP for SLCF at different rates and CO2

In G1, SLCF avoids as much temperature as RCP 8.5�RCP 2.5 Don’t forget you need INVESTMENTS to achieve both

Avoi

ded


linking regional aerosol emission changes with multiple ... · objective 4: iterate...

Documents