Engineering the Climate
Physics Departmental Seminar
University of Toronto
25 September 2008
David Keith
(keith@ucalgary.ca; www.ucalgary.ca/~keith)
Director, Energy and Environmental Systems Group
Institute for Sustainable Energy, Environment and Economy
University of Calgary
1
Morgan & Keith (1995)
Forest et al (2002)
2
Forest et al, expert prior (Morgan & Keith)
Sanderson et al (2007)
Forest et al, uniform
P>8=3%
P>8=7%
P>8=15%
4
Inertia
Turn wheel (e.g., enact policy)

Low emissions infrastructure is built at some rate after some time delay

Emissions reductions grow as integral of infrastructure build rate.

Concentration reductions (from BAU) grow as integral of emissions
reductions.

Reduction in T grow on several times scales, but delays from ocean
thermal inertia can be significant.
Climate reacts
6
Emissions are rising faster than expected
Skeptics argued that this “unrealistic” scenario
was included only to make the problem look
worse
2005
Raupach, et al., PNAS, 2007.
2006
This is where we need to be heading
7
New York Times
May 24th 1953
8
Human actions that
change climate
Climate
System
Climate impact
on human welfare
Human actions that
change climate
Mitigation
Climate
System
Climate impact
on human welfare
Geoengineering
Adaptation
Stratosphere
Lower atmosphere
Agung
El Chichon
Pinatubo
11
Temperatures after Mt. Pinatubo
Soden et al., 2002
If the radiative forcing from Mt Pinatubo
were sustained, temperature changes may
have been 10 times greater
(thermal inertia of ocean)
USGS
Putting sulfur in the stratosphere
Of order 1-2 Mt-S per year offsets the radiative forcing of 2×CO2
(~2-4% of current global S emissions)
~3 gram sulfur in the stratosphere roughly offsets 1 ton carbon in the
atmosphere (S:C ~ 1:300,000)
Assuming the NAS 1992 number of 20 $/kg  30 billion per year.
Methods:
1. Naval guns
2. Aircraft
3. Tethered balloon with a hose
13
Models suggest
the compensation
is quite good
2 x CO2
2 x CO2
and
1.8% reduction in
solar intensity
-6
-4
-2
0
2
4
6
ºC
Caldeira et al., in prep, 2007
Experiments by Phil Rasch, Paul Crutzen, Danielle Coleman
NCAR Community Atmosphere Model
Middle atmosphere configuration
• Model top at about 80km
• 52 layers
• 2x2.5 Degree Horizontal resolution
• Finite Volume solution for dynamics
with desirable properties for
transport
Photochemistry includes only
that relevant to oxidation of
DMS and SO2 –> SO4
Injection of SO2
• at 25km
• from 10N - 10S
• 1 Tg S/yr assuming a small (or
background) aerosol size
distribution
Pinatubo 10-30 Tg S
Rasch et al: Annual Average Surface Temperature
Geo-SO4/2xCO2
(1Tg Bkg)- Control
Geo-SO4/2xCO2
(2Tg Bkg)- Control
Engineered scattering systems
Alternative scattering systems
• Oxides
– H2SO4 or Al2O3
• Metallic particles (10-103 × lower mass)
– Disks, micro-balloons or gratings
• Resonant (104-106 × lower mass ??)
– Encapsulated organic dyes
What you might get:
• Much lower mass
• Spectral selectivity
18
19
Photophoresis
Uneven illumination
Sun light
Temperature gradient across particle
Net force toward cool side
net force
warm
cool
20
Gravito-Photophoresis
Sunlight warms particle evenly
Particles more likely to rebound hot
from bottom of particle
Sun light
Net upward force
net force
a=0.7
a=0.9
21
Force independent of pressure in middle of atmosphere
Force depends on T × p
1
T
F  a
p
4
T
…but T depends 1/p
S S

4
Solar input
 T
4
T E
Outgoing longwave
input from earth
3
T
   T (T  T )  V a
p
2
T
4
Radiative cooling
Conduction
So force ~ altitude independent until radiative  conductive heat loss @
about 100 km
22
What limits the size of particles that can be levitated?
The mass-specific force instantaneous force proportional to 1/r
But, net gravito-photophoretic force declines as r3 for particle radii below
~1 m because the orienting torque becomes overwhelmed by
Brownian motion as r0.
1
0.5
0.2
0.1
0.05
0.02
0.1
0.2
0.5
1
2
5
10
23
Gravity is not the only way to break symmetry
Magnetic or electrostatic torques can greatly exceed gravitational
torques for small particles in the upper atmosphere.
Consider the 1 m radius sphere (Rohatschek) in which the center
of mass is displaced 0.1 m from the geometric center.
A similar magnetite sphere with magnetization of 105 J T-1 m-3
would feel magnetic torques that exceeded gravitational torque
by a factor of ~8,000 at the typical terrestrial magnetic field
strength of 0.5×10-4 T
Similarly, a sphere of barium titanate13, a common ferroelectric,
with residual charge of 2×10-3 C m-2 would experience a torque
750 times the gravitational torque in the typical atmospheric
electric field of 100 V/m.
24
Conceptual design: A levitated disk
Radius ~10 m
50 nm
Al2O3
Al
BaTiO3
Magnetite (Fe3O4)
~500 X 500 nm
Electric field
100-200 V/m
Magnetic field
-4
10 T
Lifting force
Poleward force
25
26
Photophoretic levitation of nano-engineered scatterers
for climate engineering
1. Long atmospheric lifetimes
 Lower cost and impact of replenishment
 Can afford more elaborately engineered scatters
2. Particles above the stratosphere
 less ozone impact.
3. The ability to concentrate scattering particles near the poles
 Concentrate climate engineering where it’s needed most.
4. Non-spherical scattering particle designs
 Minimal forward scattering.
 Advanced designs that are spectrally selective.
27
28
29
Is climate control impossible?
Chaos = extreme sensitivity to initial conditions
×
One might assume: Weather is chaotic  control is impossible
Improved observations
Improved models
Improved analysis/forecast systems
See Ross Hoffman, “Controlling the global weather”,
Bulletin of the American Metrological Society February 2002 : 241-248
t2
X-29 NASA-DFRC
Not so!
Control of chaotic systems requires four things
1. A model (initial conditions  future state).
2. Observations.
3. An appropriate lever.
4. Feedback.
t1
A bigger lever  Smaller
perturbations needed to
achieve a given degree of
weather control
30
Radiative Forcing
Geoengineering
instead of mitigation
2000
CO2 Concentration
Albedo modification
2050
2100
31
Geoengineering to take
the edge of the heat
Radiative Forcing
Geoengineering
instead of mitigation
2000
CO2 Concentration
Albedo modification
2050
2100
2000
2050
2100
32
Warning: Moral Hazard
Knowledge that geoengineering is possible

Climate impacts look less fearsome

A weaker commitment to cutting emissions now
33
Value of knowing more about climate engineering
Assumptions:
1.
The prior probability that climate engineering will reduce climate risk.
2.
The cost of research to narrow the uncertainty about the
effectiveness of climate engineering.
3.
The probability of big climate impacts for CO2 above ~500 ppm.
Summary: you need to be very sure that climate engineering will never
work, or think that the climate risk is very small to conclude that
research is not justified.
35
Questions & Opinions
Opinions
1. We need a serious research program
– Impacts, methods and implications
– International
– Need not be large $$ to make enormous progress.
2. Current understanding of climate systems suggests that intelligently
executed climate engineering would reduce climate risks.
3. Geoengineering should be treated as a means of managing the worst
impacts of climate change, not as a substitute for emissions controls.
4. The science community should expect to loose control.
Questions
1. How can we best avoid the geoengineering  mitigation trade off?
2. Should we work toward a treaty? Norms? An alternate mechanism?
36
37
www.ucalgary.ca/~keith/Geoengineering.html
www.ucalgary.ca/~keith/Bibliography.html
www.ucalgary.ca/~keith/AirCapture.html
Username: carbon
Password: graphite
38
Warning: Slippery Slope
“Interest in CO2 may generate or reinforce a
lasting interest in national or international means
of climate and weather modification; once
generated, that interest may flourish independent
of whatever is done about CO2.”
1982 US National Academy study, Changing Climate.
39
40
Air capture  centralized control of diffuse emissions
Mt CO2/yr
50
150 largest coal power plants emit ~10% of global CO2
40
The next 1000 largest point sources account for the next ~30%
The next 7000 account for the next ~ 15%
30
Distributed and mobile sources account for nearly half of all emissions.
20
10
150
1000
2000
Number of Point Sources
Source: Kurt Zenz House, Alliance Bernstein
104
105
Building air capture on conventional process technologies
2005 spray tower
A
A
But, this was with ~100 µm drops
and we now know we can make ~20 µm drops
at low capital and energy cost.
2008 packed tower
Process design overview
Temp swing  deca
Solubility swing  anhydrous