Climate Engineering with Aerosols -- Predictable Consequences? - - PowerPoint PPT Presentation

Climate Engineering with Aerosols -- Predictable Consequences?

• Introduction: what is out there?
• Engineering through sulfate aerosols: new results
• Why might we try it?

– The impact of global warming on global food production

David S. Battisti, University of Washington

Collaborators: Kelly McCusker, Cecilia Bitz and Phil Rasch

• 1. Introduction: what is out there?
• Govindasamy and Caldiera (2000)

– Use NCAR’s Atmosphere GCM (CCM) coupled to a slab

• cean
• AGCM resolution: L19, T31

– Experiments:

• Control: 280ppm
• Double CO2
• Engineered world: reduce Solar constant by 1.8%

– Conclusion: Global and regional temperature and precipitation changes simulated with doubled CO2 are nearly neutralized by an appropriate reductions in the solar constant.

• 1. Introduction: what is out there?
• Govindasamy and Caldiera (cont)
• Problems

– Reducing insolation at the TOA is expensive. More feasible to reduce shortwave within the troposphere (different spatial forcing) – Regional cancellation due to enhanced sea ice in engineered world

• No ocean dynamic feedback or no sea ice dynamics -- both of which greatly affect

high latitude climate (Seager et al 2001)

Double CO2 Double CO2 & -1.8% solar

Δ Annual Surface Temperature

• 1. Introduction: what is out there?
• Rasch et al (2008)

– Use NCAR’s AGCM coupled to a slab ocean

• AGCM resolution: L52, 1.9º x 2.5º ; interactive chemistry
• No ocean current feedbacks; no sea ice dynamics

– Experiments:

• Control (circa 1950 boundary conditions: 350ppm of CO2, etc.)
• Double CO2
• 2Tg of S into stratosphere
• Double CO2 plus 2Tg of S into stratosphere

– Details of SO2 injections

• Injected into tropics (10ºN to 10ºS)
• 2km layer centered at 25km
• SO2 oxidized to SO4
• Particles volcanic in size (0.43 µm effective radius)

• 1. Introduction: what is out there?
• Rasch et al (cont)

– Results: engineering to a less warm (0.7ºC) planet than doubling CO2 (2.1ºC) also reduces the regional changes in temperature and precipitation

• Problems

– No dynamic ocean feedback -- which greatly affects tropical precipitation patterns. – No dynamic ocean feedback and no sea ice dynamics -- both of which greatly affect the high latitude climate

Δ Annual Surface Temperature Δ Annual Precipitation

• 1. Introduction: what is out there?
• Robock et al (2008)

– Uses GISS ModelE climate model

• Low resolution atmosphere (L23, 4º x 5º) and ocean (L13, 4º x 5º)

– Experiments:

• Control: a 1999 run
• A 40 yr integration with the A1B emission (starting at 1999)
• Engineered world: Various injection scenarios of SO2 into the

stratosphere

– Concludes:

• Aerosols introduced into either the tropics or the northern polar

regions have far field climate affects

• The net effect of engineering by stratospheric aerosols (against

the A1B emission scenario) would include reduced rainfall in India and throughout the maritime continent (1.5B people) – Problems:

• Ocean model too crude to represent tropical ocean dynamics
• Ocean and atmosphere model too crude for ice-edge dynamics

Hypotheses

• 1. Regional changes in temperature and

precipitation due to increasing CO2 and climate engineering (via aerosols or any other process) do not depend critically on tropical atmosphere-

• cean dynamics and the processes that are

fundamental for determining sea ice extent

• 2. There are regions of the world where climate

engineering will produce knowable changes in temperature or precipitation that are comparable to those due to a doubling of CO2

• 2. Engineering through sulfate aerosols: new results

Today: testing (1) and exploring (2)

• Does sea ice dynamics matter?

– Model: the same model used in Rasch et al:

• NCAR’s CAM (but T42 2.8º x 2.8º, L26) coupled to a slab ocean
• Now include sea ice dynamics

– Experiments: the same as in Rasch et al:

• Control (circa 1950 boundary conditions: 350ppm of CO2, etc.)
• Double CO2
• 2Tg of S into stratosphere (forcing prescribed from Rasch et al)
• Double CO2 with 2Tg of S into stratosphere (forcing prescribed from

Rasch et al)

• One additional experiment: Double CO2 with enough aerosol to nearly

cancel the global temperature change

– All experiments run to equilibrium. Results presented for annual and seasonal averages (30 years)

• 2. Engineering through sulfate aerosols: new results

The impact of sea ice dynamics

Change in Annual Average Sea Ice Concentration

“2 Tg of S“ minus “control (350ppm)”

Sea Ice extent and concentration is enhanced by the inclusion of sea ice dynamics (SH: moves farther into westerlies)

no ice dynamics with ice dynamics

• 0.2

+0.2

• 0.4

+0.4

The impact of sea ice dynamics

Change in Annual Average Surface Temperature

“2 Tg of S“ minus “control (350ppm)”

Sea Ice dynamics amplifies the cooling that is induced by the injection of aerosols but warms the central arctic (~2ºC changes)

• 6º

+6º no ice dynamics with ice dynamics ºC

Engineering of a double CO2 world w/ sea ice dynamics

2 x CO2 minus control 2 x CO2 & aerosols minus control 2 x CO2 & more aerosols minus control

Change in Annual Average Temperature Not bad. Regional temp changes <2 ºC

Global Ave ΔT (ºC)

2xCO2 - cnt 2.7 º 2xCO2 &aerosols

• cnt

0.8 º 2xCO2 & more aerosols - cnt 0.1 º

ºC ºC

Engineering of a double CO2 world w/ sea ice dynamics

Change in Annual Average Precipitation (%) Not bad, but annual deficits

• f 15% in

some places (e.g., SW US)

2 x CO2 minus control 2 x CO2 & aerosols minus control 2 x CO2 & more aerosols minus control % change wrt control

• Does dynamic ocean feedback matter?

– Model: the same model, but now add ocean dynamics

• NCAR’s CAM (T42 2.8º x 2.8º, L26), including sea ice dynamics
• Now include ocean dynamics (“high resolution”)

– Experiments: as in Rasch et al:

• Control (circa 1950 boundary conditions: 350ppm of CO2, etc.)
• Double CO2 (ramping)
• Double CO2 with aerosols (ramping to 2 Tg S at time of CO2 doubling)

– Expect

• Large differences in the tropical Pacific atmosphere and ocean (and

teleconnections there from)

• Large differences in high latitudes

Engineering a 2xCO2 world: impact of ocean & ice dynamics

Engineering a 2xCO2 world: impact of ocean & ice dynamics

Change in Annual Average Precipitation

2xCO2 & aeros minus control (ocn ice dyn) 2 x CO2 & aerosols minus control (slab) Difference due to ocean & ice dynamics

Switches the sign

• f the large (~40%)

precipitation changes in the central Pacific

2xCO2 - control

% change wrt control

(Global Ave ΔT = 0.8ºC) (Global Ave ΔT = 0.3ºC)

Difference in Annual Average Surface Temperature due to dynamic ocean feedback

Engineering a 2xCO2 world: impact of ocean & ice dynamics

2xCO2 - control (slab)

• Enhanced cooling in

the southern hemisphere (amplifies dynamic sea ice response)

• Enhanced cooling in

the N. Atlantic (1/3 reduction in ocean heat flux convergence)

• Warming (reduced

cooling) in the arctic ºC

0 ºC 4

• 4

6ºC 5ºC

Engineering a 2xCO2 world: impact of ocean & ice dynamics

Difference in Seasonal Average Precipitation (%) due to dynamic ocean feedback There is an large increase in precipitation in the central Pacific due to the inclusion of ocean

• dynamics. In winter, these changes affect the

northern hemisphere temperature

June-August December - February % change wrt control

Engineering a 2xCO2 world: impact of ocean & ice dynamics

The winter temperature differences in Australia, China and northern Canada are likely teleconnected from the tropical Pacific

ºC

Difference in December-February Average Temperature due to dynamic ocean feedback

Summary of Lessons Learned

• Greatly reduces the large-scale, regional climate

changes due to a doubling of CO2

• The inclusion of ice and ocean dynamics contributes

significantly to the regional climate response:

– Ice dynamics amplifies the high latitude cooling that results from injecting aerosols into the stratosphere by ~ 1-3ºC – Ocean dynamics also amplifies the aerosol-induced cooling throughout the mid and high latitudes in the Southern Hemisphere and in the North Atlantic, but it warms the arctic (3ºC impact). – Ocean dynamics greatly shapes the response of the tropical Pacific atmosphere-ocean system to climate engineering, and hence affects the wintertime climate response throughout the northern hemisphere (2ºC impact).

Insertion of aerosols into the stratosphere of one climate model

Summary of Lessons Learned

• Hypothesis 1 appears to be false: ice and ocean dynamics may

significantly shape the regional temperature (+/-3ºC) and precipitation (10-40%) changes due to engineering against doubling CO2

• Unfortunately, these processes are also on the top of the list of

major deficiencies in the climate models: – Clouds (Achilles heal of the low cloud solution) – Simulations of climatological sea ice and its natural variability (uncertainty in sea ice dynamics, errors in atmosphere &

• cean pbl parameterizations, etc.)

– The tropical Pacific climate: deficiencies in the paramerizations of clouds and ocean and atmosphere PBL physics lead to gross errors in the simulated tropical Pacific climate and the response of the atmosphere and ocean to prescribed forcing

Summary of Lessons Learned

• One model = one projection. What is needed to

reduce uncertainty? An effort on a scale similar to the IPCC:

– A large international R&D effort to simulate the response of the climate system to anticipated engineering solutions (not to mention the R&D on the deployment strategies, etc) – An IPCC-like mechanism to articulate and evaluate standardized experiments across model platforms, etc. – A large increase in qualified human resources

• What is the best we will be able to do?

Insertion of aerosols into the stratosphere of one climate model

Summary Comments

• Opinion: Assuming perfect deployment, errors in the projection
• f the regional response to climate engineering against 2xCO2

will be +/- 2ºC and 20% of annual precipitation (same uncertainty in GH projections)

• Opinion: There will be unwanted surprises.

– Coal -> Energy -> Acid rain – CFCs -> Many useful applications -> Ozone Hole – Lead in Gasoline -> high power engines -> lead poisoning

• Opinion: Success in quantifying the regional climate responses

may necessitate a new international policy framework

– Who can deploy? Under what condions? Who determines the

• bjectives? ….

• Sea level rise: for low laying populated areas
• Precipitation/snow pack: in stressed areas
• Temperature: for global food supply

Why would we consider climate engineering solutions?

• Sea level rise: for low laying populated areas
• Precipitation/snow pack: in stressed areas
• Temperature: for global food supply
• Sea level rise: for low laying populated areas
• Precipitation/snow pack: in stressed areas
• Temperature: for global food supply

Projected JJA Average Surface Temperature Change: “2080-2099” minus “1980-1999”

Average of 21 climate models forced by Scenario A1B. Multiply by ~1.2 for A2 and ~0.66 for B1

Extreme Heat in Western Europe in 2003: JJA temperature 3.6°C above normal

• Italy:

36% drop in maize yields

• France:

30% decrease in maize and fodder production 25% decline in fruit harvests 21% reduction in wheat yields

Why? heat stress, increased respiration, reduced grain filling, increased spikelet sterilization, …

By 2100, years of similar temperature stress on agriculture will be the norm throughout the tropics and subtropics due to the summer average temperature changes.

Refs: Battisti and Naylor 2009; UNEP 2007; Easterling 2007; Earth Policy Institute 2006; Eurosurveillence 2005

• Reduced yields of wheat, rice, maize and soybeans

throughout the tropics/subtropics (equatorward of 35º)

– Approximately -10% per 1ºC warming –

• Est. reduction of 30-40% by 2100 in India, Africa, Middle East,

Central America etc.

• Reduced nutritional content (especially protein in wheat

and rice)

• Increased disease transmission rates
• Loss of water stored in snow pack and glaciers (e.g.,

Sierra, Himalaya)

– Also reduced duration of river flow

Impacts of Climate Change on Food Security

Impacts of increased temperature (only):

Impacts of Climate Change on Food Security

• Decreased precipitation throughout the subtropics
• Increased carbon dioxide and plants

– Small (0-10%) yield increase for C3 crops for a doubling of CO2 – Effects on plant pathology (reduced protein content and resilience to disease)

• Negative effects on soil BGC (fertility/water capacity)
• Sea level rise: salinization and loss of arable land
• Changes in pest and pathogens (yet unknown)

Other impacts of climate change on agriculture

Where do the Food Insecure live?

The food insecure

• depend heavily on

agriculture for food and income

• live in regions where

agriculture will be most stressed by global warming

• live in countries that

have the greatest population growth rates

960 M people are malnourished today

• 95% are in the tropics/subtropics

Estimates: 200-400M more people at risk

• f hunger by 2080 due to climate change

Lobell et al (2008)

Summary of Lessons Learned

• Greatly reduces the large-scale, regional climate changes due

to a doubling of CO2

• The inclusion of ice and ocean dynamics contributes to the

regional climate response in non-trivial ways. Unfortunately, these processes are also are on the top of the list of major deficiencies in the climate models

• One model = one projection. What is needed to reduce

uncertainty?

– A large international R&D effort to simulate the response of the climate system to anticipated engineering solutions (not to mention the R&D on the deployment strategies, etc) – An IPCC-like mechanism to articulate and evaluate standardized experiments across model platforms, etc. – A large increase in qualified human resources

• Success in quantifying the regional climate responses will

necessitate a new international policy challenges and framework

Insertion of aerosols into the stratosphere of one climate model

Impacts of Climate Change on Food Security

Increasing temperature over the next 50 years will cause decreases in yield:

• Decrease in grain filling
• Decrease in spikelet fertility
• Increased water stress
• Increased respiration

Important for all crops, but especially for wheat, rice and soybeans (nb, these are the C3 crops that would otherwise benefit from increased CO2) and maize

Lobell 2007

Wheat Yield in Yaqui Valley, MX

Jan-Mar Night Temp (°C)

Projections of Growing Season Temperature

Projections use 22 climate models (IPCC AR4) forced by A1B Emission scenario. Variability taken from

• bservations

Observed JJA Temp (1900-2007)

2003

France

Projections of Growing Season Temperature

The Sahel

2003

France

Projections of Growing Season Temperature

By the end of the 21st Century it will be much hotter everywhere In most of the tropics/subtropics, the seasonal average temperature will very likely exceed the warmest year on record

Battisti and Naylor 2008

World Food Facts

• Average ~2700 cal/person
• Total Calories

– from plants: 84% (54% from cereals; rice and wheat are ~1/2) – from animals: 16% – from fish: 1%

• 40% of the world’s food comes from 17% of cropland that is irrigated
• Rates of Change

– Cereal yield increase in Green Revolution (~1960-80): 2%/year – Cereal Demand increasing 2%/year – World Cereal Production Peaked in 1975 – Soil loss 15Mha/yr (~1% of arable land per year) – 1/4 of arable land degraded in past 50 years