Transcript: Energy & Climate: It's All About the Systems

THE CNA CORPORATION
 
ENERGY AND CLIMATE,
IT’S ALL ABOUT SYSTEMS
 
WELCOME AND MODERATOR:
STEVE WEHRENBERG,
U.S. COAST GUARD
 
SPEAKERS:
PETER SCHULTZ,
DIRECTOR,
U.S. CLIMATE CHANGE SCIENCE PROGRAM OFFICE
 
JOHN SHILLING,
CHAIR, BOARD OF TRUSTEES,
MILLENIUM INSTITUTE
 
ANDREA BASSI,
SENIOR MODELER,
MILLENIUM INSTITUTE
 
MONDAY, JULY 21, 2008
 
 
 
 
 
 
 
Transcript by
Federal News Service
Washington, D.C.
 
 
 
 
 
 
STEVE WEHRENBERG:  My name is Steve Wehrenberg.  I’ll be your
moderator this evening.  Welcome all of you to our – if we counted correctly – 27th
conversation.  Our goal here has always been to inform and to clarify the issues we face
regarding energy, energy security, and the consequences of our energy use.  We’ve
addressed in these 27 sessions energy sources, the sustainability of energy sources,
energy uses, the need to conserve energy, make most efficient use of the energy we have,
and, of course, the wildly divergent solutions that have been offered for responding to our
challenges.
 
 In these 27 events, we have concluded that there is no single solution that’s going
to win the day.  And, in fact, every solution that we’ve seen has unintended consequences
and that would include even the solutions that we have put into play.  It seems that
everything is indeed connected to everything else, just as we thought might be the case. 
And it also seems to be impossible to solve any single part of the problems that we face.
 
 I saw Mr. Gore earlier this week issue his challenge to the nation and to our
forthcoming political leadership.  Regardless of your politics, it’s very hard to argue with
this one particular statement: quote, “I am convinced that one reason we’ve seemed
paralyzed in the face of these crises is our tendency to offer solutions to each crisis
separately, without taking the others into account.  And these proposals have not only
been ineffective, they almost always make the other crises even worse.”
 
 I don’t think anybody would argue with that.  Certainly we wouldn’t here.  For
some time now, we have wanted to provide our participants in this conversation with a
way of looking at the entire system at once so that the best possible trade offs can be
made and the most sustainable good derived.  Tonight marks the first in what I believe
will be a number of conversations over the next year or so aimed at further expanding our
perspectives – as Mitzi so often puts it, to expand our apertures – so that we can see and
understand more all at once.  That seems to be what our situation asks of us today.
 
 You have access – if you haven’t seen them already, you will have access – to the
biographies, curriculum vitae of our speakers and I’ll keep these introductions short.  We
do have a two-pronged session tonight: Dr. Peter Schultz, the director of the U.S. Climate
Change Science Program Office, will take us up to about 10,000 feet or so – or 100 miles
or so or however high you plan to get over there, Peter – to help us see some of the
connections among many of the issues that we face and that we have discussed so far in
our conversations.
 
 Then Dr. Jed Shilling, board of trustees, chair, and Andrea Bassi, candidate and
senior modeler at the Millennium Institute, will demonstrate their internationally
renowned and much sought after Threshold 21 model, the U.S. version of which has
recently been updated, released, and will be available to everyone I am sure – yes,
absolutely.  And that particular model has a necessary focus on energy at the intersection
of our environmental, social, and economic spheres.
 
 The U.S. Climate Change Science Program began as a presidential initiative in
1989, was codified by Congress in the Global Change Research Act of 1990, which
directed a coordinated, interagency research program, receives oversight from OMB, the
Office of Science and Technology Policy, and the Council of Environment Quality. 
We’ll provide links on our website to that site and to the many publications and reports
that have been created by the U.S. CCSP.  
 
 Dr. Schultz joined the CCSP in 2004 after several years at the National
Academies where he directed studies of global environmental variability and change.  He
also worked at the NOAA Climate Analysis Center.  He does not speak for the
administration; he speaks for science.  Those are not necessarily not related.  (Laughter.)
 
 The Millennium Institute has over 20 years of experience at integrating economic,
social, and environmental sectors into a single dynamic feedback model that represents
the critical high-impact variables within and across sectors.  Notice I said the critical
high-impact variables across sectors, not everything.  It is not Prego spaghetti sauce.  I
warn you in advance that the model, used extensively, though not exclusively for
developing nations as systems, is not a predictive model, but more a way to explore the
relationships among the most important variables affecting a problem.  It does not
prescribe, but enables the right conversation.
 
 Dr. Shilling and Mr. Bassi have the requisite background, credentials, and
experience to help us explore the long-term sustainability of our nation, our society, and
of our species.  As is usually the case, there will be ample time for questions and
comments.  Mikes are set up so that you can sort of self-queue for that.  I do ask that
tonight, due to the somewhat mutually reciprocal nature of these two presentations – they
are much integrated – that you hold your general questions and comments until after both
presenters have laid their cards on the table.  If, of course, there is something you
absolutely don’t understand and can’t understand without asking a question, then please
feel free to ask, but if you’ll hold our conversation until both speakers have had a chance
to speak, I would appreciate that.
 
 On a personal note, this session has been a long time in the making for me.  I am
particularly excited to be presenting you with the T-21 model as it is a wonderful
example of a modeling discipline called system dynamics.  For those of you who are
familiar with it, you would of course immediately realize that what we have is a much
needed way to organize what we know and then to have a meaningful discussion about it. 
That’s probably the strength of this type of modeling discipline.
 
 I’ll give you a quick comparison which I think may peak your interest just a little
bit.  We’re all familiar with spreadsheets.  Spreadsheets have cells; cells have numbers or
labels or formulas in them somewhere.  Of course, you don’t see the formulas; they’re
transparent to us.  What you see is the result of the formulas in that particular cell.  In
other words, you have complete visibility of the data, but you have no visibility without
some significant sleuth-work about the relationships among the variables.  
 
 System dynamics is a discipline that turns that inside-out, making the
relationships visible and clearly available to you while putting the data behind a curtain
where it can be lifted if you need to take a look at it.  So I think you will find it quite
interesting, if you’re not familiar with it already.  As a modeling approach, I think it’s not
at its best making predictions, but it is certainly useful at helping understand what’s going
on, and that seems to be what we really need.  Without further ado, Dr. Schultz.
 
 PETER SCHULTZ:  All right.  Thanks, Steve.  Can everybody hear me in the
back?  Good.  I think I’m going to move over here just because this is where I’ll be
showing stuff.  Okay, let’s see, let’s get this going.  It was queued up and I de-queued it.
 
MR. WEHRENBERG:  Technology, the great time saver.  
 
MR. SCHULTZ:  All right.  Okay, so for those of you that saw that polar bear
shaving video to begin with, you might have a sense that that was a hoax, that because we
know that most polar bears are left-handed – (laughter) – and so wouldn’t – okay, so this
is what I’m going to talk about: “Climate Change: Taking the Binders Off and the
Necessities of Integration.”
 
Before I get into that, I’m going to spend two minutes with just an unpaid
advertisement to what we do in the U.S. Climate Change Science Program.  This year
we’re releasing something like 21 assessments of the state of the science related to
climate variability and change.  These are covers from nine of the reports.  I’ll draw from
some of these in my talk.  This one here at the end is titled “Scientific Assessment of the
Effects of Global Change on the United States.”  It’s a soup-to-nuts analysis of what we
know about the impact of climate change on the nation.  Let’s see.
 
I’ll just give a couple of highlights from one of them.  This one is an assessment
product 4.5.  And what’s really sad is I know these by their numbers, so my brain has
been fried.  This one is focused on the effects of climate change on both energy supply
and energy demand.  And some of the conclusions from this are that increases in extreme
weather events are expected to have direct impacts on energy supply systems.  The
reductions in – and increased variability of water supply, particularly in the U.S. west
associated with some hydrologic shifts there, snow pattern changes, which I’ll actually
talk about in a couple of minutes, are expected to have significant implications for
hydropower and thermal power plant cooling.  Facility siding is expected to be affected
by changing conditions.  And renewables are also expected to be affected.  And there are
many other conclusions at a much greater level of detail than what’s provided here.
 
On energy demand, this report concludes fundamentally, as others have, that we
expect to decrease energy consumption for space heating, but an increasing for space
cooling and refrigeration.  We expect an increase in electricity peak demand in most
regions except for the northwest.  And we also expect changes in the balance of energy
use among the delivery forms.  So electricity for AC over natural gas for heating.  And
then, of course, changes in energy consumption in key climate-sensitive sectors.
 
So therewith ends the advertisement for some of the products that we have – the
Climate Change Science Program.  You can learn more about these assessments at
climatescience.gov.
 
Okay, this says, “Fersteinser believes in strictly minding his own business,” and
Fersteinser here has blinders on.  And this is how we often operate, with blinders on.  It’s
comfortable.  I know that I tend to operate that way.  And my title before being the
director of the Climate Change Science Program office was the director of science
integration.  I was supposed to be focused on integration, but yet it’s comfortable to have
the blinders on, and this talk is about taking the blinders off.
 
Okay, so the outline of what I’m going to be talking about is the necessity of
scientific integration, integration on impacts analysis, the effects of climate change on
various things that we care about, integrated analyses of how we can adapt to those
things, how we can better prepare for the threats that are out there, as well as some of the
opportunities that will probably also be out there, and how we can have a more integrated
analysis of the options for mitigating or preventing climate change.  And the graphics
down here have a specific purpose here.  So much of this is about the glue, the bridges,
the relationships and the wiring diagrams that are needed to make all of this happen.
 
So let’s ride the way-back machine.  Back in 1924, Fourier hypothesized that the
greenhouse effect might exist.  Svante Arrhenius in 1896 made the first quantitative
estimate of warming from CO2 doubling, and he said it might be about 4.5 – 4 to 5
degrees C, which is not too far off from the state-of-the-art estimates, which are now
around 3 degrees C for a doubling of CO2, and that estimate was first put in 1979 by my
beloved institution, the National Academy of Sciences, where I worked for about eight
years.
 
But Arrhenius was able to make this jump by thinking about outside of the box
that Fourier was thinking about.  He was thinking about just the natural greenhouse
effect.  Arrhenius was thinking, well, what happens if we keep burning fossil fuels, keep
burning the coal that was making London so black at that time?  So he added in some
factors and got to this conclusion.
 
Okay, so what have we done in the last 20-or-so years since the 1979 – well, I
guess it’s more than 20 years – since the Charney Report?  One of the things we’ve done
is we’ve added in the ocean into the analyses.  So we have ocean-atmosphere coupled
models.  And one of the things that this allows us to do is to look at the time lag of the
warming.  So if we double CO2, it doesn’t instantly warm to that level because the oceans
buffer things.  So one of the things that we know is that if we were to stabilize the
concentration of CO2, just freeze it at its current level of 380-some PPM right now, we
would still get a warming that approaches 1 degree Fahrenheit over the long-term –
additional warming, even though we had frozen the level of CO2 in the atmosphere.  And
that has to do with the fact that the oceans are still coming up in the equilibrium with the
atmosphere.  So that comes through this more integrated perspective.
 
These lines on here are different assumptions about what the climate sensitivity is,
but they all basically say as we go out from the year 2000 out to the year 2400, we expect
to see warming if greenhouse gas emissions were frozen immediately.  This is no
warming, and this is 1 C warming.  So right in here is about 1 degree Fahrenheit
warming.
 
Okay, another thing that we note from this coupled integrated perspective is that
we expect to get this cold-ocean, warm-land pattern, so by 2020, 2029, we’re looking at
this kind of pattern.  By 2090, 2099, according to one set of model estimates, this is the
type of pattern that we’d be looking for where the oceans are warming some, but not as
much as over land, and this says something about the types of impacts that we would
expect, but we only get to that because we’ve taken this more integrated perspective.  
 
Okay, then in the last 10, 12 years, we begin to couple in the land surface, so we
now have ocean, atmosphere, land coupled models.  One of the things that we get is a
feedback so that as the surface climate changes, you get a change in the geography, in the
biogeography, so ecosystems shift, generally pole-ward.  And each of these color changes
you can’t possibly, in the back, can’t possibly read what these things say, but they say
things like this dark green is forest cover gain; this light green is shrub, woodland cover
gain; this light blue is herbaceous cover gain.  So it all represents shifts in the type of
ecosystem distribution across Europe, and that of course has local effects on the climate. 
It feeds back, so you change the type of cover on the land surface, and it actually changes
the microclimate.  
 
It does other things because as you change the type of ecosystem, some of the
carbon that’s sequestered in the ground might be released, or some of it might be stored. 
So you get feedbacks with carbon dioxide itself.  And what this confusing diagram, and
not all of my diagrams will be this kind of nonsensical, but let me just explain this very
quickly.  Here, we’re at 1850 going out to 2100; 2000 is right here.  These are different
model runs.  They’re different state-of-the-art coupled ocean, atmosphere, land chemistry
models, so chemistry has been added into this integration.  And the different models vary
in their predictions of future CO2 levels that – and these differences are just the
differences that are due to adding in that dynamic vegetation that we were just talking
about.  And the upshot of this is that it’s highly uncertain what the effect of ecosystems
will be in terms of the future levels of CO2.  
 
So we no longer have this safe, comfortable feeling of knowing that if we emit a
certain amount of CO2, the concentration of CO2 in the atmosphere will be X in the future
because now through this integrated inclusion of land surface processes, ecosystems
processes, and dynamic chemistry, we now realize, gosh, we actually know less than we
did before, or that we thought we knew before.  And this says something about the
certainty that we should place on our projections for the future.  Of course, there’s a lot
that we do know about what will happen, but this kind of integrated perspective really has
colored our view of what we know about the future.   And in some cases, yes, the
uncertainty, or the apparent uncertainty has grown.  In other cases, the apparent
uncertainty has decreased.  
 
Okay, so that was one set of views on the evolution and the science in terms of
that increasing integration that’s happened over time.  Now, I’ll say something about the
projected impacts, not that impacts isn’t science, but it’s not kind of the core of the sort of
the basic research.  This is about the types of impacts that we’re actually seeing right now
and observing right now.  One example if from the Western United States, right now in
the Western United States from about 1950 through about the present, one thing that
we’re seeing is a change in the snow-water equivalent on April 1.  April 1 is kind of the
measure that Western water resource managers use to determine what the extent of the
snow pack is that they’re going to have to work with in the regions that are snow-
supplied for the water.  
 
 And what they’re seeing is that the snow-water equivalent, so the amount of
water in the snow pack, is declining over time. And the red circles, the larger the red
circle is, the more that it’s declined over time.  The blue circles mean that the snow pack
has actually increased from 1950 through the present.  So what we’re seeing is a
declining April 1 snow pack and going along with that, we’re seeing that the runoff peak
is coming earlier.  And the darker the red circle is, the earlier in the season that it’s
coming.  So it’s running off, it’s melting earlier.  So that has significant implications for
hydropower production, among other things, we mentioned that a couple minutes ago.  
 
So it’s that kind of integrated analysis taking us from the snow pack to the runoff
peak, and then to the effects on hydropower, agriculture, urban water supply, flood
control, ecosystems, and a water resource manager has to represent all of these things
simultaneously.  There are levels of priority that are placed on each.  They have rules for
their decision making, but in order to do this, they have to have this integrated
perspective.  They don’t have the luxury of having the blinders on when they make these
decisions.  And these decisions are being affected right now by climate change, not in the
short term, but in terms of the types of infrastructure that’s being built.  
 
Okay, another example of integrated analysis of climate impacts is out in the
West, for those of you that have been out West recently, you may have seen these slopes
where you see these pretty red hues in the canopy.  And this is not deciduous trees in the
fall; these are trees dying.  This is a pine bark beetle that only its mother could love that
he or she, I don’t know if that’s a he or a she, I don’t even know if that’s how you
determine whether that was a he or a she – (laughter) – and its progeny, I guess.  Or
maybe that’s the childhood picture, I don’t know.  
 
And if you look closely at this, you can see the little bubbles that they make
where the sap comes out of the pine that they burrow under the bark of, and they just,
they kill these trees.  And it’s happening all across the West.  And the prevalence of these
pine bark beetles, these spruce bark beetles, is increasing because they are controlled in
part, their distribution is controlled in part by the wintertime temperatures.  And because
it’s warming, their populations are able to increase, in part as a response to that.  
 
Other things that are happening out in the west are that through the fire prevention
measures, the fuel loads have been increasing.  And a third thing that’s happening out
there is, particularly in the southwest, they’re in a pretty deep and sustained drought in
parts of the region.  So all three of these factors conspire, but before I get to the punch
line here, this is a map of all of the bark beetle infestations and all of the big red and blue
splotches that you can see from this survey in 1996 represent where the bark beetles were
mapped, well, not a lot.  But in 2003, the bark beetle populations just exploded.  I took
these slides from a talk that I gave in Steamboat in January, which by the way is not a bad
place to give a talk in January if you’re a skier.  
 
Okay, and what all of this leads to is an increase in fire threat.  And one of the
things that some agencies in our program do is they convene twice a year.  They convene
meetings with fire managers to tell the fire managers what they know about the current
conditions, the current climatic conditions, and what the short-term projections, what the
month to sort of half a year projections are for climate going out into the future.  And
they work with fire managers developing integrated solutions to the increasing fire threat
to the extent that that’s possible, particularly in the face of the declining budgets that the
forest service and others have to cope with fires.  
 
Okay, an integrated adaptation, adaptation is something that really has not
received as much attention on the national scene relative to mitigation, the prevention of
climate change, the reduction of CO2 and other greenhouse gases.  And it’s something
that, in my personal view, is a significant deficiency in the national dialogue.  Climate
will change.  We showed the slide that even if we were to freeze the level of greenhouse
gases in the atmosphere, at present, we’re still committed to future climate change.  We
must adapt.  We must cope with these changes.  
 
So this is an adaptation strategy from, I don’t know when it was, but it’s
electromagnetic bathing fluids that can deal with neuralgia, cholera, rheumatism,
paralysis, hip disease, measles, female complaints, necrosis, chronic abscesses, mercurial
eruptions, I don’t know what that is, epilepsy, and scarlet fever.  So this was, well, I
guess back then it was integrated adaptation analysis, but really not the type of thing that
we’re talking about here.  
 
So this is the Beaver family – isn’t that what they were called?  Cleaver family,
thank you, I didn’t watch enough TV as a kid.  I told my mom.  So here they are, they’re
out walking, and one of the things, just a simple, this is a very simplistic presentation I’ll
give you here for the next 90 seconds, a very simplistic kind of adaptive approach is to
build communities that promote foot transportation.  What does that do?  
 
You reduce CO2 emissions, you increase physical activity, so the Cleaver family
becomes more fit.  You decrease air pollution, you increase social capital.  So the Cleaver
family is out there and you can see that they’re just promenading for their neighbors. 
And they stop off and talk to the Smiths and the Jones’ and talk about their nice patent
leather shoes, and so the social capital in the community goes up.   And Mrs. Cleaver’s
depression goes down.  Roadway accidents go down.  Her osteoporosis that’s been
nagging her for years goes down because she’s getting more exercise, and by the way the
infrastructure costs go down.  
 
So this is a caricature, but it’s the type of systems thinking that we need that
integrates adaptive planning, adaptive community planning and mitigation planning.  And
much of, as I said at the outset, much of the thinking has been on mitigation, less so on
adaptation.  So some of these other climate change synergies that in terms of heat wave
planning, heat wave adaptation, one of the things that the Centers for Disease Control
promotes are these heat wave plans, including buddy systems.   What does that promote? 
Well, it promotes social capital in addition to doing the types of things that help to
prevent morbidity, mortality associated with heat waves.  You decrease vehicular traffic. 
You decrease car crashes.  You decrease air pollution, all of these side benefits in
addition to directly responding to climate change.  You increase fuel efficiency, decrease
air pollution, increase locally grown food where you’re decreasing the greenhouse gas
emissions. 
 
Just a personal confession, about five months ago I stopped eating red meat just
because I saw what the literature was saying about the greenhouse gas emissions that are
associated with eating beef and pork relative to chicken and veggies.  And I’m not
preaching about that, it’s simply a decision that I made because I want to put my mouth
where my mouth is, I guess, but it is something to consider. 
 
Energy-efficient buildings, well, yeah, that increases the greenhouse gas emission
burden, also decreases operating costs in many cases.  Alternative energy sources
promote business opportunities.  The United States must be a leader in this area.  We are
leading in some aspects of this, but I think the business opportunities are highly
underexploited.  
 
Okay, so now segueing into mitigation, two overarching conclusions from the
intergovernmental panel on climate change, the IPCC, one is that a portfolio of
adaptation and mitigation measures can further diminish the risks associated with climate
change.  And the second is that there is substantial economic potential for the mitigation
of global greenhouse gas emissions over the coming decades that could offset the
projected growth of global emissions or reduce emissions below current levels, but how?
 
Okay, so in this cartoon here, there is a very thin child here struggling with this
guy in a business suit, and there’s an ear of corn here.  And this guy is saying, excuse me,
I’m going to need this to run my car.  So 10 years ago, there was a lot of discussion about
corn-based ethanol.  We’ve moved out with policies on corn-based ethanol.  Was there
the requisite systems analysis?  Was it the requisite whole-system integrated perspective
brought the barrenness?  I don’t know because certainly the lifecycle energy costs of
corn-based ethanol are very high, sometimes exceeding the greenhouse gas savings.  The
food production displacement is quite significant, rising food costs, we know about these
things.  
 
And a third thing that gets a lot less attention is the deep sequestering of carbon. 
So there’s a lot of land in the United States that lies fallow; farmers are actually paid to
have that land lie fallow.  Now there will be increasing pressure to plant on that land, and
when that happens and when that land is plowed, then you have some of that carbon that
was formerly sequestered there go back up into the atmosphere.  A lot of these concerns
don’t exist for other forms of ethanol production, cellulosic ethanol, for example, but we
need that whole systems analysis of these issues instead of just jumping into solutions
without thinking about what the widespread ramifications are.  
 
Okay, when we think about climate change, one of the terms that exists in our
community is this cascade of uncertainty that cascades from scenarios of energy
production down to emissions scenarios, then to understand what the concentration will
be in the atmosphere, you have a carbon-cycle model which then tells you what the CO2
level will be in the atmosphere actually, which then you feed into a climate model, which
tells you what the change in climate is expected to be in terms of temperature,
precipitation, and other variables.  And you have impact models that we’ve talked about,
and then you look at the economic and sociologic consequences.  And we also have
models for these things and they can feed back up.  
 
And all the way down through this cascade, there is uncertainty, and we need to
track that uncertainty and we need to explain it to the decision-makers so that they can
make better decisions about wherever they lie in this whole cascade.  One tool for doing a
widespread or a wide range of analyses about mitigation options is a class of things called
integrated assessment models.  And I want to spend a minute or two describing what
these integrated assessment models are.  This here is a picture that you can’t possibly, in
the back, can’t possibly read what all these words are, but I’ll characterize some of these
things.   
 
This is a picture of the MIT’s integrated global systems model, and it includes an
atmosphere, ocean, land surface model here, so this is kind of the natural science model
part, kind of along the lines of what I was walking you through in the beginning of my
presentation.  It includes all of those factors.  It also includes urban air pollution
processes, which are also important.  Not only does the warming have direct health
effects, it also has indirect health effects by promoting the formation of ozone, which we
know has respiratory implications, and this model is able to generate some of those
photo-chemically produced species. 
 
So it has the chemistry in it.  It has modules that address agricultural impacts,
forestry impacts, bio-energy impacts, ecosystem productivity, hydrology, water
resources.  It looks at energy demand as a function of climate change, land-use change,
looks at sea-level change, health effects, and feeds that into an analysis of human activity. 
And embedded within this is an economic model, a computer general equilibrium model
and other components of the energy system, both a bottom-up and a top-down energy
look.
 
And some of the outputs that you can get from an integrated assessment model
like this are analyses of GDP growth, energy use, policy costs, ag, and health impacts, or
you can look at some of the physical dimensions of the system, like projected changes in
sea-ice cover or sea-level rise.  Or you can look at some of the biological changes in the
system.
 
As Steve said at the outset, models like this are not best in a predictive mode. 
There are too many levers on these things, too many sources of uncertainty on these
things.  But what they do allow us to do is to put all of our assumptions on a common
framework, and to look at them.  And if we say, let’s assume this level of GDP growth,
that level of exogenous – oh, population would be exogenous in the model.
 
This range of parameter assumptions, what do we get when we drive the model
out into the future?  When we take policy option A, what are the environmental
ramifications, what are the economic ramifications, and what – according to the model’s
assumptions, what kinds of fuel types could get us to the various sorts of greenhouse gas
emissions trajectories that we’re shooting for in some of these policy scenarios?  So these
models can look at these issues.
 
And I’ll give one example from three of these models.  This is from a report that
we issued several months ago.  These are three of the state-of-the art integrated
assessment models in the United States.  The IDSM model, which I – ooh – which I just
described; we saw that picture.  The MERGE model is another one, and the MiniCAM
model.
 
So what these are are models that are run from the year 2000 to the year 2100,
2000, 2100, et cetera.  And what this is on the Y axis is exojoules per year.  So it’s the
primary energy production in the United States to get to a stabilization of 550 parts per
million of CO2 equivalents in the atmosphere.  So what does it take in terms of energy
production types, fuel switching to stabilize the atmospheric concentration?  And so these
different color bars are different types of energy sources.  Yellow is non-biomass
renewables; orange is nuclear.  The hachured stuff is things that involve carbon capture in
storage.  The gray is simply conservation, energy reduction.
 
And you can see that these three state-of-the-art models give very different
answers, but there are also commonalities among them that allow them to, within their
parameter assumptions, to achieve the stabilization of 550 parts per million.  One is that
energy conservation is absolutely essential.  All of them indicate that that’s mandatory. 
That’s the gray.  There are different assumptions about the feasibility of nuclear.  IGSM,
their parameter assumptions did not allow for the development of nuclear in a way that
MERGE and MiniCAM did.
 
But we need a national dialogue.  Given the projected impacts on society and on
ecosystems, we have not had the dialogue that we probably need with the stakes as large
as they are on nuclear.  I’m agnostic, personally, on nuclear.  I need to be informed by I
think many of you in the crowd who know a lot more about this than I do.  I don’t know
which way to go.  But these types of models allow us to look at what happens if we
include things like that, how does it help us get toward these various targets.
 
Carbon capture and storage:  If you’re going aggressively after a stabilization
target, it’s pretty clear, it’s pretty consistent among these models that you need some
form of carbon capture and storage given their assumptions, which is where you have a
power plant, and somehow you’re scrubbing out, you’re taking out the CO2, and you’re
somehow taking it out of the atmosphere system.  You’re either burying it underground
geologically.  There are other ways to do it biologically.  You’re pumping it into the
ocean.  You’re taking it out of the system.
 
So these tools allow us to look at what-if scenarios that would get us to these
different kinds of stabilization analysis.  And Jed and Andrea will talk more about the
types of – in a way, it’s games that you can play with these, given some plausible
assumptions going into them.
 
And finally, I’ll just conclude with this thought:  We need a national adaptation
and mitigation strategy.  We don’t have one right now.  We have an approach or we have
a set of approaches with respect to mitigation.  We do not have a national adaptation
strategy.  Billions of dollars and thousands, probably millions, of lives will be lost if we
do not develop a national adaptation strategy.
 
The Lieberman-Warner bill would put billions of dollars a year toward adaptation,
but there is no strategy for spending that money if that were actually done.  And that
money – the way that that would work is it would come from the auction proceeds from a
cap-and-trade mechanism.  Now, will that play out?  I have no idea.  But what I do know
is that we are not thinking strategically about adaptation.  We have not looked at the
whole-system perspective with respect to adaptation.
 
I was just talking to – I was just briefing some developers, actually, who called
me last week who wanted a briefing on our scientific assessment.  I don’t know why
these developers wanted this, but I said, sure, I’ll talk to you.  And they said, you know,
there is just a simple technology that we want to explore, which is a porous roadway
surface, which we know that we’re expecting to see greater incidents of heavy
precipitation events, so you get pooling of the water on the roadway surface, you get
greater accidents.  It’s a public health threat.
 
But there is this porous asphalt that has recently been developed that can help to
ameliorate that.  And they say, we can’t do this because the governments, the state and
local governments, after we put this stuff in will cap it, because that’s the cheap thing to
do.  They’ll just put a simple asphalt cap on it, even though we’ve spent a significant
amount of money putting in this porous roadway.
 
So the system is not talking.  And that’s just one tiny example.  I could go through
countless other examples of insurance and on down the line where we have not done our
job here inside the Beltway getting our act together to help the nation adapt to the threats
of climate change, nor have we developed an integrated mitigation strategy.
 
Technology is not the only solution to climate change.  It’s a very, very important
part, but we haven’t looked at all of the pieces.  And as Steve said at the outset, there’s no
silver bullet, and we haven’t looked at really how all of those pieces come together, and
we need more discussions of the type that Mitzi and Steve and others have brought
together here.  So I thank you for your time, and look forward to answering any
questions.  (Applause.)
 
MR. WEHRENBERG:  Thank you, Peter.  We’ll get the – we’ll get Jed and
Andrea set up.  I just want to reflect on one thing you said there, Peter, while we’re doing
this, and that is, you said something about needing a dialogue and needing that national
dialogue.  And I have to say that – and that you assert that there haven’t been those
dialogues.  We may think that there have been those dialogues, but I suspect that what we
have really seen serial monologues, not dialogues.  (Laughter.)  And we definitely need
more conversation, thus events like this we believe contribute to that conversation.  So go
out and have dialogues – (laughter) – not right this minute, but – 
 
JED SHILLING:  Well, you’ve got a minute till we get this hooked up.  Okay,
I’m going to stay over here, but I’d like to thank Steve and Mitzi very much for inviting
us and having us here and offer the greetings our president, Hans Herren, who
unfortunately could not make it.  He has been in Rome working with FAO on some
global agricultural problems and trying to address and solve some of these problems
systemically.  But it’s a real pleasure to be here, and I appreciate Peter’s presentation,
which was very useful and valuable and gives us a good introduction to our system.
 
Now, just very briefly, the Millennium Institute is a not-for-profit that’s based
here in Arlington, and it was established in 1983 to promote holistic long-term strategic
planning.  The founder, Gerry Barney, is sitting out here in the audience.  He keeps track
of this and did a wonderful job setting it up.  The goals are to develop analytic tools to
support strategic planning at various levels, and to build capacity among the users so that
we pass out the tool so lots of other people can use it.  We don’t want to keep it locked in
our door, and we want to work with stakeholders and bring them together to help arrive at
agreements about the best feasible solutions.
 
Very rarely in the world have I seen the truly optimal solutions that some models
talk about, including general equilibrium models, ever occur.  I’ve really never seen an
economy in equilibrium like the models described, so it’s important to bring people
together to talk about and reach feasible solutions.  And the Threshold 21 model was
developed as history goes back to the Global 2000 Report that Gerry Barney did for
President Carter that came out in 1980, and it tried to look at some of the long-term
issues, building on the information from various departments of the U.S. government.  
 
And one of the problems Gerry recognized very early is that many of these
models that were used individually didn’t connect; they didn’t even follow the similar
assumptions.  And he was anxious to go on further, but unfortunately President Carter
was replaced by President Reagan who didn’t want to follow up on these kinds of issues. 
So Gerry went out and founded the Millennium Institute and started looking at and
examining models to see what he could find that would address these issues, and in fact,
there were none in those days.
 
And he then discovered system dynamics at MIT and with the help of Weishuang
Qu, our chief modeler who is here, developed some initial applications for a number of
other countries around the world.  And this has been expanded into sustainable
development models which have been applied on a broad number of issues in small and
large countries.  And the real contribution, as we will get into more details later, is
linking economic, social, and environmental factors into a single coherent framework so
we can see the interactions among all of them on a consistent basis, not just simply
feeding back and forth.
 
And our applications we’ll be talking about today are on the USA and energy
security, which began with a model version of the model.  We adapt the basic model to
lots of different circumstances.  At the New America Foundation last year, it was picked
by the American Society for Peak Oil, and we presented and discussed it at their
convention.  And this model, we focused a lot on these energies, and Andrea will present
the model.  Unfortunately, it hasn’t been updated since the end of last year, so we don’t
pick up everything that’s happened since, but we’re looking for support to expand it, to a
global energy model and do a broader global view, not just a countrywide view, although
it does take the global markets into account.
 
And in the meantime, we’ve been working in a number of other countries from
Jamaica to Mauritius in helping them do their own country models.  And we have
assisted over 45 countries prepare their models.  We’ve worked in a wide variety of
circumstances, and adapt and customize the model to individual circumstances.  It’s not a
one-size-fits-all framework; it’s very adaptable.  We basically understand that the
linkages between the three main sectors have to be there, but it has to be adapted, and
we’ve dealt with a number of islands, including several in Denmark.
 
We have addressed poverty issues, energy issues, including working with the
NCEP, where we have in fact addressed some carbon capping and carbon trading issues
and things like that.  We’ve worked in the North America model from – and China and
other of these countries.  So you can read the list of countries.  I don’t have to go through
them.  But it’s a very wide variety of circumstances we’ve looked at, and focused on
environmental factors, or education factors, or incorporating the potential of natural
disasters down the pike to see what would happen if a hurricane or a flood or a draught
hit a country.
 
Now, we can’t predict exactly when one of those things are going to happen, but
we know there are going to be hurricanes hitting Jamaica every several years or floods in
Mozambique or things like that.  But it’s descriptive, as what several people have said. 
This is not a precise projection model.  And in fact, based on recent experiences, I don’t
have a lot of faith in projection models.  If you look at the recent experience of these
detailed, highly sophisticated financial projection models and what they said about the
financial sector, I don’t trust them.  I’d just rather have models to describe likely options
and we can make our best decisions.
 
Here’s an area around the world where we have worked in a number of different
circumstances, including working with General Motors in looking at the expansion of
transportation markets in developing countries.
 
Now, very briefly, the Threshold 21 approach is based on system dynamics
methodology.  We try to build in a wide range of existing sector model and analysis.  We
don’t claim to know everything about every one of these sectors.  We bring in the best
information, and as Andrea will explain for energy, we brought in the energy sector
models from a number of major research centers and academics.  We bring in models
from agriculture, from demographics, from population, from other areas.  And we
incorporate them into this single framework.
 
And we try to do it in the way that most reflects the real-world relationships that
we observe, not a theoretical idea of what’s supposed to happen, but based on history and
observation effects, what relations really obtain and how they work, and look at the cross-
sector linkages so we can see what happens.  If you do something in the economic sector,
what’s the impact going to be on population or health, or the environment.  If you do
something environment, what’s the impact going to be on – so we try to bring these all
together.
 
And as I said, it’s composed of the three main pillars: the economic sector, which
is based on a social accounting matrix and market balances, and it in fact has been
reviewed by the World Bank and pronounced entirely satisfactory; the social sector is
built on dynamics of population health – HIV/AIDS - in a lot of the countries we work in,
is very important, and education; and the environmental factors we adapt it to specific
countries because nearly every country in the world has an environment that’s very
different from every other country, although sometimes they share environments within a
continental region or something like that.
 
But we adapt it to the data and issues of our client in the countries we’re working
on.  It doesn’t address everything, but as Steve said, it addresses the key issues and how
they link together, and highlights intersectoral feedbacks to show what the overall
implications are.  We calibrate it against history as part of the reality check to say, okay,
if we start this projection – and Andrea will show you this – in 1990, let’s see how
closely the model projects – because nearly everything is endogenous – what actually
happened.
 
And I know of very few models that will dare to try to run a projection against
history to see how accurate they were when we have real facts to judge them on.  And we
generate long-term scenarios as Peter and others have said.  You’ve got to look out 50
years or beyond to see what the impacts of many of these policies are going to be, even
though that is longer than the horizon of most politicians who are looking only as far as
the next election. 
 
But it’s also transparent and easy to use.  As you will see, this model, you will be
able to download soon from the web or even use it on the web when we get a little more
technology done.  So it’s something that anybody can get in there and use and run.  You
don’t have to hire an entire department from MIT to come and run a certain set of
scenarios.  You can test a lot of them yourselves, and we work with a number of clients to
adapt and integrate and incorporate more and more specific issues and sectors that make
it more adaptable to your particular concerns.
 
Here’s the basic structure which Andrea will go into in more detail.  And we do a
little cartoon here too.  It’s why do you really need to take a systemic view?  Because it’s
important that you look beyond the next couple of steps when you start something
because it may have some unintended consequences.  (Laughter.)  So the system issues
are being more and more recognized.  We’re looking at the rising energy prices related to
limited resources that has impacts on energy security, dependence on imports.  Shifts to
renewables have their own costs.  We’ve seen the problems with the corn ethanol, and
we’re looking at other ones.  We’ve seen what’s happening with rising food prices,
pollution, greenhouse gases, economic slowdown, other challenges of global warming.
 
People are now beginning to see these things are not separate issues, but they are
linked together, and they have to be analyzed in a single framework.  So in the energy
conversations, we have seen a number of specific sectors addressed.  I can’t list them all
because I haven’t made all of them.  But we’ve looked at various different sources of
energy: agriculture, water conservation, security issues, waste management.
 
And what we’ve learned is there’s no silver-bullet solution.  No policy, no
activity, is going to solve all of the problems. It’s probably going to create a few while it
solves some other ones.  So we need a systematic approach to understand and fit how the
– see how well these things fit together.  And I would love to be able to put all of those
individual projections together in a single framework and, say, is putting huge amounts of
money into nuclear really going to solve the problems, or is it going to generate some
other problems we have to deal with?  How much renewable can we really do with what
we’ll do?  And Andrea will show you a few of these results where we looked at some of
the policies that have been actively pursued recently.  But this helps to design and
disseminate more coherent and effective policies over the long term.
 
And T-21 is simply another tool that’s quite comprehensive, helps you to do it.  It
wont’ get into the great detail that some of these other specific sector models have.  But it
will take the key elements out of those and integrate them into a broader framework.  So
thank you very much for your attention, and let me turn it over to Andrea, who is a
master at putting all of these things together into a single modeling framework and
generating some results for you.  So Andrea.
 
(Applause.)
 
ANDREA M. BASSI:  Well, thank you very much, Jed.  I’ve never had such a
good introduction before.  Thank you all for being here.  And I have this high job today
to introduce you to an integrated model in 10, 15 minutes.  So I hope you will excuse me
because I’ll show bits and pieces here and there to show you what are actually the
capabilities of the tool and what we can look at.  What is the value-added we can get out
of the model?
 
So I’ll just jump straight into the structure of the model.  This is what Jed showed
very briefly.  For this specific application I’m going to show today, we have energy as the
core sphere, as we call it, of our system, and energy is connected to society, economy,
and the environment.  We break out these three main sectors – society, economy, and the
environment – to look at, for instance, population, which in this case is endogenously
calculated.
 
We look at labor force that can be used for production, for instance.  We look at
health, we look for infrastructure that is used both for providing social services, and as
investment in the economy to generate GDP, value-added.  We look at education issues. 
We also take into account the government accounts – so financial transfers, household
accounts, banks, the rest of the world.  And for what concerns the environment, we look
at land and location, into, for instance, forest, agricultural land, settlement land, so the
land that can be used actually to produce biofuels and so on.  We look at water
requirements, both safe drinking water, or water for energy production, again, biofuels or
tar sands.  This could be something we can add later on.
 
We look at emissions, greenhouse gas, and other energy-related emissions,
minerals, the effect of energy production and consumption on climate change.  This is
something we borrow from existing studies.  In fact, this is what we do.  We select the
best studies available out there and we take them and integrate them into our framework. 
So we’re not inventing much.  We are providing value-added by looking at an integrated
picture in this case.
 
So if we break out those sectors to look at what are the connections, which is
more or less what we have seen throughout the energy conversations, since I’m an Italian,
would you expect to see spaghetti diagram of course, like this one.  (Laughter.)  But it
looks very complex.  What you see are not all of the feedback loops or connections that
are included in the model, but what we are aiming at.  So we are trying to build on step
by step – using a model approach to try to get a comprehensive tool that can help us with
these complex systems.
 
So in this case, we can see that energy is connected to, for instance, labor.  Let me
go back.  I don’t get any way of showing that.  But you see that energy is connected to
labor.  We need energy for many different reasons.  Labor is an input that allow us to
produce energy.  At the same time, as I said, energy is used for water – well, water is
used for energy production.  At the same time, land is used, and land is also an input for
energy production.  With biofuels, energy produce – energy consumption produce
emissions.  We use energy to actually strike minerals.  And if you look at a lifecycle
approach, these minerals, or steel, aluminum, and other derivatives and so on allow us to
actually build infrastructure to produce energy.  At the same time, for what concerns the
economy, we need energy to produce value-added, and value-added is needed to
investments to actually produce more energy, and so on.
 
So we have a number of many important connections that can be identified.  And,
again, we don’t look at the details.  We try to identify what are the main causal
connections that allow us to define what are the main forces driving the system, and that
will become clear later on.
 
So just to give an idea of what are the characteristics of the tool and model that
will be used, we tend to use a descriptive, rather than a prescriptive approach.  So this
means that we try to understand what is the underlying structure of the system we’re
looking at or we have to deal with.  We always start from a set of issues.  In a way, every
model should be built around the core set of issues because no model can represent
everything; otherwise, we will aim representing reality as is, which is an ever-ending
process, and modelers don’t really want to do that.
 
And if you use a prescriptive approach instead, we apply assumptions that are
theories, economic theories and so on to actually get the results we want.  Now, in trying
to go through a two-ways approach here because descriptive models are also based on
causality, and prescriptive models use most of the time correlation.  So what we see in
causality is how do real systems interact?  So if there is a causal relation, we are able to
identify feedbacks.  We are able to identify what is the rationale behind these connection,
what is the strength of the connection through no linearity, representation of delays. 
Well, when we look at correlation, we to see from a statistics point of view if the
relationship is positive, negative.  How strong is the impact that one variable can have on
another without looking at the actual causal relation.
 
So a very simple example:  When ice cream sales go up, also debt-rate goes up. 
So this is a very strong statistical correlation, but there is no causal relation between those
two variables.  So in a way, we can say that when the temperature goes up, then ice
cream sales go up, but also with high temperature, there is a higher probability that
people with problems have problems or – really in sort of a late age can suffer more from
this problem.
 
So, again, we have in this case, correlation, but not causality.  And we really want
to focus on causality.  And when we represent feedback loops, non linearity, and delays,
when we run the simulation, correlation is actually the output of our simulation.  So if we
look at causal relations at input and still we get correlation as an input to check the
validity of our model and assumptions about the structure of the system.
 
Then at the same time, we want to look at policy evaluation more than
optimization.  This doesn’t mean that policy optimization is not good.  Actually, both are
needed.  In fact, these tools are somehow complementary.  Policy optimization allow us
to understand what is the goal we are aiming at, as Peter mentioned earlier.  So what is
the best set of policies we could put in place if you want to reach a specific goal, given a
specific condition and under a specific set of assumptions?
 
While with descriptive models, the look at causal relations, we really want to
evaluate what is the impact of policies that are being discussed currently to see whether
these policies can allow us to reach the goals we intend to reach.  So in this way, we are
trying to represent whether there is any policy resistance effect, if there is any side effect
that will arise over time.  So we are trying to understand how the different components of
the system interact with each other to find out whether there will be synergies or elements
of policy resistance.
 
Now, since we are talking today about mainly an energy model, there are many
similarities that you will see between T-21 and conventional energy models.  They all
look at energy demand, supply, technological development, energy prices, pollution,
what could be trade and investments depending on the scale of the models.  The
difference, the main difference here is that with T-21, we try to use as many endogenous
inputs or endogenous variables as we can.  So while conventional models, or CG models,
will use an exogenous assumption for GDP or for population, we replace this exogenous
assumption by using an endogenous formulation to calculate key variables in the system.
 
So once we start from a set of issues, we want to understand what are the causes
and the effects of those issues, and we tried to make sure that we represent endogenously
the main drivers that may actually lead to these issues to come up.  Then again, we look
at different types of energy.  For supply, we have fossil fuels, we have nuclear energy,
renewables.  We look at electricity of course as a secondary form.  We have different
sectors for demand: residential, commercial, industrial, transportation, and then we look
at final energy types or uses depending on the country we are looking at.  In this case, we
are talking mainly about the USA North American model.
 
Now, this is to give you a general overview of how we actually model, what is the
rationale behind the model we build.  This is somehow a very simplified representation of
the point of view that economists have on the economy.  So we have GDP that influence
investments, and investments accumulating to capital.  Capital is one of the major factors
influencing GDP.  And capital also requires employment to be run, and employment is
one of the – it’s actually one of the main factors influencing GDP.  So we have capital
labor plus total factor productivity.  So the value-added of our study is then to identify
what total factor productivity means.  And we include education, energy prices, and many
other factors.
 
So starting from GDP, looking at the causes, what defines what, we see that GDP
is one of the drivers for energy demand.  At the same time, energy demand is one of the
drivers for energy production, which we try to satisfy.  Energy production, on the other
hand, generates emissions, and production is – can be obtained by using, in the case of
fossil fuels, energy resource, which we know are not an infinite number.  At the same
time, energy production together with demand defines energy prices, and energy prices
has an impact on energy demand.  And energy prices finally have an impact on GDP.
 
So we are seeing here a couple of feedback loops already.  We have a reinforcing
loop of the economy, the higher GDP, the higher capital and employment, and the higher
GDP again.  The higher GDP and the more demand we can have, the more production
can be.  If production is available, energy prices will actually be fairly low, so GDP can
grow faster and higher.  But the more we produce, the more we deplete resources – fossil
fuel in these cases.  So this triggers an initial balancing loop that has a ripple effect on
energy prices in the economy.
 
But, again, if you look at renewable energy, which is something that is emerging
now very strongly, if you increase renewable energy, we have more production, and we
have a different impact on emissions.  So renewable energy, if you look at, again, a
lifecycle approach, they generate emissions.  But if you look at the running costs and the
running impacts, they don’t general many emissions.  Actually, in some case, it’s not at
all.
 
So what we are seeing here is the fact that with energy research declining and
with energy price increases, we actually go through an energy transition process that is
actually related to the use of renewable energy, and different energy sources, which, once
again, is the final question that optimization tools are asked to solve.  What is the best
supply mix that we need to put in place to reach our goals of emission concentrations and
so on?  But the approach is a lot more dynamic.  So we’re actually looking at causal
relations to find out what the consequences may be in this case.  So we are not assuming
what GDP would be.  GDP would be a consequence of our decisions or our policies, and
so on.  So we can appreciate in a much more integrated way what are the consequences of
the policies being discussed.
 
So this is – to give you some examples on what the outputs of the model can be,
with our model, or the approach – more than a model, we don’t really want to get
answers; we want to offer a tool for you to understand better how systems work, and to
run analysis, scenarios, different set of assumptions that you think are reasonable.  There
are way too many assumptions on the amount of oil available out there, how much is
actually recoverable, or what is the impact that energy prices can have on the economy. 
So we allow the users to actually run some of those assumptions and get the results to run
their own analysis afterwards.
 
So in this case, by using one basic simulation only, we can change the amount of
resource available or technological development and get a number of different scenarios
for our production.  At the same time, those projections are connected to oil prices.  And
I’m sure you can hardly see models projecting these curves for oil prices over time. 
Now, first of all, we have to say that we look at historical trends and we try to look at
historical trends starting from 1980 in this case and running to 2050.  We intentionally
decided to skip looking at the years 1980, 1985 because there were too many unknowns
of what concerns the energy sector or the economic sector.  We prefer to look at the
longer-term trends given that actually the oil crises were not a structural crisis, but it was
a temporary reduction in production, so not something that has a very longer-term
impact, at least on research. 
 
And what we see here is that by simulating different assumptions on peak oil, we
can get two types of reactions over the longer term.  The first one, when oil price spikes,
go up very high to above $225 per barrel is when actually a transition takes place.  But
then there are a number of other factors that define the second peak in prices, which is the
fact that reducing availability at the board level would make so that the U.S. would use
domestically produced oil, which is more expensive than the average price of Saudi
Arabian or Middle Eastern oil, for instance.  So we go to a transition that is not very
straightforward.  There are a number of different effects that can come up and influence
our analysis.
 
Now, what we did in July 2007, last year, we tried to simulate the impacts of a
crisis similar to the one in 1980.  So we were just wondering, what if, actually, oil prices
will go up again because we have a reduction in supply?  Now, what happens is that oil
price will shoot up, then technology can improve to a certain extent.  Certainly, demand
will go down when price goes around $200 per barrel.  And then, since this is not a
longer-term repercussion, we have oil prices adjusting, coming down, adjusting a little
bit, and the energy transition will happen earlier in time.
 
Then, we have prices that over the longer term, after 2030, would go down,
because we assume there is a full transition in this case.  Or maybe not full but almost full
transition beyond oil with unconventional or conventional oil – with unconventional oil,
biofuels, other sources, we assume there is a massive switch to other energy sources. 
Again, it’s an assumption, so we can change that.
 
Now, in October 2007, working for the Association for the Study of Peak Oil and
Gas, we were asked to look at different scenario.  So peak oil happening in 2011, I
assume most of your are familiar with the concept of peak oil, which is just oil
production declining, which doesn’t mean that we would run out of oil from one day to
another, which is represented by the red line.  Then, we simulated the Shell case,
represented by the blue line, and one of the EIA cases, which uses the medium-high
estimates for recoverable – (inaudible) – provided by USGS.
 
So in this case, still we compare with the historical data.  We start simulating in
1980.  And we want to look at what are the impacts of these different scenarios?  The
first one, pretty straightforward, is the impact on prices.  Now, if you look at the red line,
it might look a little bit unrealistic.  But the assumptions for that scenario were that
suddenly from night to day realized that we don’t have as much oil as expected.  So there
is a big deal of speculation ongoing.  Prices will jump very quickly.  There will be a
shock on demand that would be reduced.  And there will be a longer-term adjustments.
 
And the prices will keep increasing over the longer term, different from the other
scenario that showed, because we assume in this case, it’s not possible, given known
technology, to actually go to a full transition.  So in this case, it will cost more to produce
oil, because depletion goes on.  And we will not be able to actually reduce demand to a
high extent as in the previous case.
 
Now, if you look at the Shell and EIA scenarios, we see a more gradual increase. 
Now, this is generated using the same model, just applying different scenarios.  So we
can see the behavior adaptation and what kind of responses the market can provide, given
different assumptions.  So one of the good uses of this model would actually to bring
experts together, simulate different assumptions, compare the impact of the assumptions
they want to simulate have on different sectors, and then start the dialogue.  Actually, if
you look at the next graphs – for example, oil consumption and GDP, which would come
next – when we simulate assumptions of these different groups, we actually see similar
results for what concerns macro-variables.  So results are comparable with the energy –
oil and energy outlook and so on.
 
So in this case, for oil consumption, we see that for ASPO we have a forced
transition because of low availability, which is highly unexpected.  In the other two cases,
we have more of a market approach that allows to actually shift based on the most
profitable energy sources and the more acceptable or accessible one.
 
For what concerns GDP, we have a recession for a couple of years for the ASPO
case.  We have Shell that shows some stagnation for a while and then GDP increasing. 
And for EIA, we have – we can show that mainly the problem with supply over the
longer term doesn’t pose a major threat for the economy because there is enough time to
actually act and implement mitigation policies, knowing that this will happen beforehand. 
And some of these results may not be very reasonable if you think about reality, but they
are the result of assumptions that are used to build those models.  So they are consistent
with the assumption, if you look at the structure of what they use.
 
Now, of course, if you look at different scenarios, as Jed said, we merely were
looking at policies.  We want to see what is the impact of different policies that are
currently being discussed.  We can look at a number of different policies at the same
time.
 
So with T-21, customers to North America, even if you assume a scenario for
peak oil in 2011, 2012, we can examine or analyze what is the impact of renewable
portfolio standards or the corporate average fuel economy standards.  Or, we can look at
restriction on emissions, cap-and-trade, carbon tax policies, and what happens if we go on
with subsidies for ethanol?  Under different scenarios, those may be more or less efficient
and effective policies, of course.
 
And then, we can look at some of these policies in isolation like a maximum push
for a renewable solid.  What does that mean, or a low-carbon emission scenario with
renewable standards, efficiency, and so on.  And we’ll simulate one of these scenarios
later?
 
Again, for different polices, we can look at one or two or three different cases.  So
we support policy formulation and evaluation so we don’t have to stick with one single
proposal.  But we can actually simulate in a few seconds different cases.  The same for
renewable portfolio standards – with the current study we are doing with the National
Commission on Energy Policy for cap-and-trade cases.
 
Again, if we compare – and this is another value-added of the study – if we
compare the effect of policies currently being discussed with different scenarios, we can
understand what is the impact of something unexpected happens.  So if implementing
CAFE standards is a policy that is intended to reduce emissions over the longer term, if
we actually run into unexpected early peak oil, that policy is not going to have much of
an impact.  So we have to think of policies that might actually have some effects on a
very different scale.
 
Now, something else very interesting about the model is the dynamic
representation of the energy transition in a way.  In this case, we are looking at the
emissions, CO2 tons per year.  And we are simply looking at three different cases,
assuming that there is peak oil happening in 2011, which is being chosen because it
shows pretty well what happens in these cases.  We can see that in the ASPO case, when
there is no investment in renewable, and coal is mainly used to satisfy electricity demand
– the red line – we have a fairly high – still decreasing but fairly high level of emissions.
 
Now, if we implement the renewable portfolio standard case, this policy for 20
percent by 2020, what we see – this is the green line – is that we have lower emissions
until 2020.  But then the fact that we actually use renewable energy and not coal would
make so that coal would become cheaper.  So this is a direct feedback.  And energy-
intensive industries in the U.S. like aluminum, steel, and so on will actually profit from
this.  So there might be a positive effect of taking back some of those – (inaudible) – you
have industries in the U.S., supporting employment and so on.  But there is the negative
effect on emissions.  So again, policies can have unintended consequences or synergies
that are somehow unexpected at this time.
 
Then what happens with this provision for a new – (inaudible) – standards, which
is this one, as you see.  There is a massive increase until 2020 and then a slower increase
later on.  What happens is that since we reduce the growth in renewable energy
production, then coal price will go up again and emissions will go down over the longer
term.  So you actually are proposing something that is good for those industries.  After a
couple of years, they start consuming more.  They put in place some investments.  And
then, price go up again and they run into problems.  So there are all those transitions that
can be taken into account.
 
And if you look at the violet line that accounts for CAFE standards, renewable
portfolio standards, plus conservation, what we see towards the end of the simulation,
between 2004, 2050 is that those polices actually reduce the impact of oil on the
economy, because of reduced consumption of oil.  Again, we go through this transition
with coal.  But what is really important is that by reducing energy consumption and
dependence on oil, GDP can grow faster.  And over the longer term, energy consumption
will go up as well.  So we are fixing a problem today and the solution will become a
problem tomorrow.  So that’s something that we can take into consideration in these
different cases.
 
Now, let me switch over to the user interface so we can go through the simulation
of the model.  Finally, you will be able to see – T-21 in a nutshell, why – to learn how to
simulate and run custom scenarios.  And as Jed said, this interface is available online. 
It’s about 10 megabytes, I think, so you can load it and store it on your laptop.  We will
have an online interface fully working soon, probably a month from now.  But in the
meantime, you can play around with this one, since this should be seen as a game. 
Hopefully, it will be somehow fun as well.
 
So for this interface, we have three main sections.  The first one, when you can
review the structure of the model, the second one where you can analyze and compare the
results of the baseline simulation, and the final one, which is I think the most fun, where
you can actually simulate your own scenarios.
 
If you review the structure of the model, you can go through a conceptual
overview, which I showed earlier.  So you can see what are the sectors?  What are the
links between society and the rest of the three spheres?  Then add the economy to make it
a little bit more complex and the environment finally to appreciate the full view of what
the model can show.
 
Otherwise, if you want to have a more in-depth look at how does the model work
and what is the value-added of some of the specific sectors, we have what we call causal
loop diagrams that show what is the causal relation among selected variables.  In this
case, I’m sure many of you are familiar with the rebound effect that is the result of an
increased in energy efficiency.  So if you increase CAFE standards, what the rebound
effect or feedback loops says, over the longer term, the effect will free up resources to
houses, for instance, because energy efficiency is increasing, will make so they can
actually spend more or they can drive more.  They can afford to drive more and spend
more in gasoline, for instance.  So the intended effect that you have in reducing
consumption would be actually smaller than what you can actually see.
 
And this is simply what we call a balancing loop.  So you implement an action
and the system actually balances off the effect of this action.  And there are a number of
other examples for oil demand and supply, the relationship between energy and biofuels
and so on, which I will show later in a dynamic way.
 
So if we go back, for those who are actually interested in looking at the details of
the model, we have made available some boxes that allow you to look at equations.  And
let’s take for instance total energy demand.  In this case, we use very simple equations. 
We disaggregate variables as much as possible to make sure that every single variable
represents reality.  So every variable should have a real counterpart.  So it makes it easer
to understand how this works.  And we use extended variable names, which doesn’t mean
that the model works better or it’s better; but it makes so that the model is more
transparent.
 
So for what concerns total energy demand, we have nuclear production, which
you assume is consumed, then renewable energy, coal, natural gas, and oil.  Of course, I
guess, almost nobody would like to go through the equations and look at how actually the
model works, because we have many variables.  Another way of doing so is to check on
the causes tree, where instead of looking at the equation, you can actually see graphically
what variables are used to calculate the one you are looking at.
 
I don’t know how many of you can read.  But still, we are looking at total energy
demand in quads in this case.  And this is defined using, again, nuclear production,
renewable, coal, natural gas, and oil demand.  Again, if you want to look, for instance, at
how natural gas demand is defined, you can just select it, click on causes, and you see the
natural gas demand is defined by using electricity that is generated by using gas and then
commercial, industrial, residential, transportation, and so on, demand for natural gas.  So
you can actually go back and forth and navigate through the model to see why things
happen and how things are introduced and calculated.
 
Now, let’s keep in mind total energy production in quads, a very aggregated
variable here.  For those who really want to go into the structure of the model, this is
what we create when we open up the software and add variables and so on.  What you’re
looking at is the population sector.  And here, you see the stock of people, which is
divided into 82 age cohorts, which is how we calculate the population pyramid.
 
We look at labor force or population entering school and so on.  And we represent
this by looking at births as an inflow, deaths as an outflow.  There is another flow, which
is net migration.  So we actually try to understand how the real system works.  And then,
this is a general representation that can customize to different countries.  Different
countries have different effects defining birth rates, or death rates, and so on.  So we want
to be consistent.
 
Something interesting here is that you can double-click on any single variable in
the model and then select the causal strip and actually see the behavior of the variables
that are used to calculate the one you are interested in.  So at the top, we have total
energy demand in quads.  And then, after that one, we have all the baseline projections
for the variables that are used to calculate total energy demand.  And again, you can look
at this for all variables.  So if you implement the change, run your own custom
simulation, and you want to understand why a certain change happened, you can always
go back and select a variable you change and see where this one is having some impacts.
 
So in a way, there is no way that you cannot understand why the model generates
some specific results.  You’re always able to understand what happened.  Sometimes
there is some mistakes that we are happy to work and to correct.  In some other cases,
you get insights because there are some unintended consequences that emerge from the
behavior in the model.  So there is always something to learn in this sense, even if you
work with this model for quite a long time.
 
Then, the second section is where we actually look at results of the model respect
to the baseline scenario.  We have a number of different graphs and tools to do so.  We
have, for instance, population pyramid, which we propose here every five years.  This is
1985.  You can look at 2005 and go on.  So this helped us understand what may be the
issues, for instance, of Social Security, Medicare, or needs for employment.  Our
economy should grow to make sure that employment or unemployment rate is kept under
control and so on.  So this allows us to look what 15, 20 years ahead on the major trends
and how they will change, which I think is very useful.
 
Then, instead of looking at the regular time graphs, we can also see energy
demand shares – for instance, we would use commodity graphs to see what is the
contribution of the different energy sources to demand and supply?  In this case, we see
that in the baseline ASPO scenario, when there is no investment in renewable energy, no
major technological development, coal will be the most important energy source,
supplying energy for the future in the U.S.
 
Now, going back to something more interactive, in this case, what you see here is
a visualization – a dynamic visualization of the causal loop diagram I showed earlier
about the connections between GDP, energy demand, energy production, prices, and the
common way economists look at the economy with GDP that influences investment;
investment accumulates into capital and has an impact on GDP and employment and so
on.
 
So with this representation we can make explicit what are the connections within
the model and run simulations.  Now, I run a very fast one just to show the baseline
scenario.  You see the lines coming up slowly.  We have a red line, which is representing
a baseline simulation and a blue one, which probably you can hardly see, which
represents historical data.  Sometimes, arrows are blinking.  That means that there is a
major change happening.  And the size of the arrows change according to the changing
inputs.  So here, you can see there is a problem with energy price – let’s get rid of this –
with energy price, which is due to the fact that we assume there is peak oil happening in
2012 and it’s unexpected.  So this is just a simple way to get rid of the complexity in the
model and try to show things in a simpler way.
 
Now, something interesting as well is what we call the cheese slicer, which was
developed originally by Charlie Hall, professor for SUNY-ESF.  And what we see here is
our GDP is allocated at what are the impacts of peak oil on the differing sectors.  So we
have peak oil in 2012 now and you see with prices increasing, there is a lot more
expenditure that goes into energy.  And over the longer term, we see that investment –
discretionary investment – basically goes down to zero.  And energy acquisition and
maintenance for the energy sectors are the ones that require more money because the
energy return investment goes down and we have less profitable energy sources that we
can use.  So there are different ways of packaging the results of the model to reach
different audiences, ensure interesting results in this sense.
 
So now, I can tell you if you want to simulate this, it’s really funny.  If you go
very fast and you see the effects of peak oil bumping here and there.  So I hope you will
have some time to simulate different scenarios.
 
MR. WEHRENBERG:  I can’t wait.
 
(Laughter.)
 
MR. BASSI:  Now, the most interesting part, let’s see if we want to simulate new
scenarios, I’ll show very briefly using a Leiden diagram, the general one, some of the
scenarios I showed in the PowerPoint presentation.  Here we selected a number of
policies and assumptions that are customized to different audiences, different projects. 
So the whole modeling approach, the model, the interface, and the policy variables, can
very much customize depending on the issues that we need to look at.
 
So from what concerns the policies for now, we have a carbon tax, a tax on
gasoline, and a tax on income.  So this way, we can look for instance at what is the
tradeoff between increasing a gasoline tax and reducing income tax for the lowest income
classes.  Then, we can look at CAFE standards or electrification of rail.  In this case, this
is a prepackaged simulation that relates to energy security and national security.  We
have identified those 35,000 miles of rail that are strategically important.  And we looked
to what happens if we actually shift to electrified rail in this case.  What are the
implications for energy consumption and so on?
 
Otherwise, we can look at renewable portfolio policies as one aggregate or we can
actually select different renewable energy sources and waste and try to simulate different
assumptions of what actually is feasible and what are the consequences in terms of cost,
emissions, and so on.  Then, of course, we have a number of assumptions.  As you can
imagine, many of those relations are largely unknown.  So we allow you – so we shift the
burden to you, actually – to define what is the effect of energy prices on GDP?  In this
case, we use studies that were actually proposed and that came out in the ’80s, because
that’s where we start the simulation.  And then, the energy, the elasticity of GDP to
energy prices is actually dynamic.
 
Or, we allow you to simulate different energy technology scenarios.  Again, the
model calculates energy technology based on the most up-to-date information and based
on the investment flow that we calculate in the energy sphere.  But at the same time, one
of the major shortfalls of energy models was they underestimated technology
development.  Now, in this case, we allow you to actually simulate what happens if a
major technological breakthrough takes place, gets into commercial stage, and so on, to
see what is actually the gain that we can get out of that improvement.  And for
technology here, we have energy efficiency, we have exploration of fossil fuels; we have
recovery as well, so much oil can actually be taken out of reservoirs and so on.  
 
Then, we have assumptions on energy consumption, on the mileage per vehicle,
which is somehow connected to cultural factors, human development, and so on.  Transit-
oriented development is something that might be related to this.  And of course, you can
decide to change the initial amount of oil research and resources, or you can decide to
add a new discovery that was not taken into account in the first place.  What happens if
we find another Prudhoe Bay?  What happens if we find another Canterell field in
Mexico, and so on?  And then, we have some scenarios for the energy-return investment,
to see what is actually the contribution of, for instance, ethanol to transportation, first and
second generation and so on.  
 
So what I’m going to propose now, very briefly, is a simulation of the CAFE
standards, which have actually been approved in the H.R. 6 last year, so this is – one of
the analyses we run with Roscoe Bartlett in Congress.  So what you see here is actually
one of the variables in the model, which represents legislation.  This is what – the rule
that was in place for our concerns, the average fuel efficiency of passenger vehicles.  And
we have two ways of modifying this value:  Either we click on the graph – and right now
we are indicating about 44 miles per gallon in 2020, 45, and we assume that there is a
slow increase afterwards, which is what the legislation proposes – or we can actually type
in the exact number based on the legislation.  So we can be very accurate in this sense.
 
Now, then I would simulate the electrification of rail to see what is the impact of
electrifying these 35,000 miles of rail, strategic importance.  Let’s see what happens,
okay.  Now, if we simulate the model, what we see is a new simulation coming up.  Now,
our new simulation is the red one.  The reference case is represented by the green line;
it’s still quite difficult to see.  But what we would expect in this case is, of course, energy
demand to decline because we are increasing fuel efficiency of vehicles.  At the same
time, we’re reducing consumption of diesel oil and fuel, generally, for transportation.  So
in fact, we see energy demand is decreasing.  
 
At the same time, what we would expect is energy prices to be lowered – let’s
stay with this window for now – and energy prices are slightly lower, still assuming that
we have peak oil happen in 2011.  That can be changed later on, if you want to simulate
new scenarios.  And then we see the GDP’s growing faster, investments are proceeding in
a better way, capital accumulates more than in any other case.  But still, if you look at
emissions down here, we see that emissions over the longer term are increasing.  So this
is actually the simulation of one of the scenarios I showed earlier.  GDP grows, even if
we consume less oil; what happens is that without investing in renewable energy, we are
actually using coal to supply our increasing electricity needs.
 
So if you are aiming at reducing emissions, this policy doesn’t really work.  Still,
we increase energy security and we may increase other parameters, which is good, but
again, there is no silver bullet that can provide the answer.  So in this case, we can allow
you to see a different mix of policies and scenarios to see whether the policies are
effective, what is the cost, energy cost, financial cost, and so on associated to it, and see
what are the benefits at the same time.  
 
Now, one last simulation and then I’ll leave the floor for questions.  What I
mentioned about peak oil, which is a very much debated topic – what we can do is either
change the amount of resource available, but what we actually introduced here is one
specific study proposed by Matthew Simmons.  He says we have been overproducing
fields in Saudi Arabia and this makes so that we actually disrupt the availability of oil. 
So the more we pumped in the past, the more we destroy the geological formation of the
reservoir and we can extract less.  So this is one example in which we can borrow from
existing studies and research and put that assumption into the model.  
 
So in this case, we have this parameter that shows actually what is the amount of
oil that can be recoverable, with respect to our expectations, if we simply change that
variable and we say that we can actually recover more, what you expect is a completely
different reaction of energy prices, completely different impact on GDP, and different
impacts on emissions, of course, and land allocation and use, et cetera, et cetera.  So what
you see here in the price box here in the top-right corner is that price, in this case, will
grow slowly to reach more or less the same longer-term growth that we show in the peak
oil case.
 
Now, this window shows that you can actually look at the details of different
results, view the data, copy and paste into Excel, and do your own analysis.  You can also
load every single scenario we simulate; you can decide to save it and then load it back
after a month, two months, and actually check what were the differences of the policies
you had simulated in the first place to keep building on your initial analysis.
 
Now, before I conclude, since we mentioned some of the connections between
energy and other sectors, and we didn’t really talk much about that but we showed
different cases and shocks and so on.  This is, again, a visualization that looks, in this
case, at ethanol.  So we have here GDP, which is calculated endogenous in the model –
GDP is formed, is calculated using agricultural services industry production – having
agricultural production, we can calculate what is corn production by also analyzing how
much land is used or actually allocated for corn production.  Then corn can be used in
three ways:  This can be used for domestic food consumption, which is based on
population, mainly; or it can be exported, still for food, international aid, and so on; or
this can be used for ethanol production.  Once we estimate how much can be used for
ethanol production, we can see what is actual – the ethanol production, in terms of
gallons, and see what is the contribution of ethanol to motor gasoline consumption and
whether that can help reducing prices or the vulnerability of the economy to prices.
 
Now, what turns out here as an important result is that the net energy gain is
usually very, very small, especially for, you know, ethanol produced out of corn,
especially the first generation of ethanol.  And then, what is really interesting is to look at
what is the impact on water if we are able to produce what is projected by USDA.  How
much is available of this water; how much should be diverted and taken from other
functions?  In which areas are we going to suffer for water shortages in which other areas
we can actually produce sustainably?  So there’s a number of other questions that we
would like to address later on, for which we would really like to have your input in
understanding what you think is important, what you think is relevant, what you think are
the main drivers that will influence the future, in terms of the economy, society, and
environment, and the energy sector, of course, so that we can keep building on what we
have so far and try to make it more complete and closer to reality – still a simplified
reality, but that can be useful for policy formulation and evaluation analysis. 
 
So thank you very much.  I hope you will have some good questions.
 
(Applause.)
 
MR. WEHRENBERG:  Come on back up, Peter.  You’re part of this crew at this
point, so –
 
The floor is open for your questions.  I know you’re overwhelmed; I can tell by
the fact that you’re not jumping up to mob the microphones right this second.  
 
Well, thank goodness.
 
Q:  A question for Dr. Schultz –
 
MR. WEHRENBERG:  Will you identify yourself please?
 
Q:  Yeah.  I’m Dick Lawrence from ASPO-USA.  And we, by the way, worked
with the Millennium Institute on the model and helped them fund the first iteration of it. 
But a question for Dr. Schultz:  There is, in the IPCC scenarios, my understanding of this
– and this is secondhand, I haven’t gone into it myself – but the IPCC scenarios seem to
be unrealistically optimistic about the availability of fossil fuels, particularly oil and gas
and even coal.  And I’ve been unable to manage or find a way for there ever to be a
meeting of the minds between those people who believe are bumping into some ceilings
in production of fossil fuels and those people on the IPCC panels who are forecasting
these huge and, in my opinion, quite unrealistic assumptions about availability of fossil
fuels in the future.  Is there any way, that you see, that we could bring about a meeting of
the minds?
 
And I guess the second thing I want to say is this in no way says that the
probability of climate change goes away because if oil and gas become limited
availability, then certainly there’s a huge pressure to consume more coal.  But I’d like to
hear your opinion on that.
 
MR. SCHULTZ:  Yeah, our office, the Climate Change Science Program Office,
hopes to facilitate the work and engagement of the United States in the IPCC, activity
facilitates U.S. government review of it.  And we hope to select the authors for IPCC and
fund a variety of other things associated with the IPCC.  So anybody that has questions
about engagement with IPCC, contact me and I’ll help put you in touch with the right
people to engage in those dialogues.  And this is actually a good time to be thinking about
the scenarios because IPCC is rethinking their scenarios, at this point.  
 
Q:  Considering many things have exceeded their worst-case scenarios.
 
MR. SCHULTZ:  Right.  I mean, the one scenario, the most extreme scenario
that’s kind of part of their baseline, is the A1-FI.  And now, reality has emissions
exceeding that most extreme emissions scenario.  How long will that be sustained?  I
don’t know, but it certainly says something about where their whole family of scenarios
is with respect to reality.  So contact me.
 
Q:  Don Aerbach (ph), Agricultural Research Service, retired.  A couple of short
questions, one on the model:  Do you have anything that estimates the actual cost to
extract and deliver a barrel of oil at this time?
 
MR. BASSI:  Okay, well, thanks for your question.  It’s very difficult to answer. 
We don’t know what is the actual cost.  We are basically looking at the demand-and-
supply effect to find out what the price may be.  I tried to model cost, but my opinion, for
the study that I’ve seen, there is no reason why the cost should justify this high price.  So
from what I can read, the average cost now for production is still around $40 per barrel
less, depending on the areas.  Saudi Arabia is much lower.  But we are not attempting on
modeling costs right now, in that sense.  But we prefer using a demand-supply balance
approach to look at prices.
 
MR. SHILLING:  Another factor on the cost is it varies tremendously with
different producers.  In Saudi Arabia, it’s quite low; in Nigeria and Venezuela, it’s
higher.  In the United States, it’s even higher.  And the market tends to pick the highest
price, plus the speculation that’s going on in the background.  So it’s hard to figure out
the exact cost of doing any average; it’s not like the typical economic model, which
assumes that the costs equalize across all the producers.
 
Q:  Okay, and in the model, when you have a typical output for the model, what
kind of confidence interval do you have on that estimate?
 
MR. BASSI:  Well, we ran analyses on other models that we did about 10 years
ago, started about 10 years ago until four or five years ago.  If you look at the macro-
variables, such as GDP, population, there could be energy demand for some models, we
have an average error for the few that we projected, comparing with the actual historical
data, that is within 3 to 4 percent over the whole period.  So this is fairly high confidence.  
 
Then, if you look at detailed variables for what concerns the projections, of
course, the only confidence we can get is that we are able to represent historical behavior
for 15, 20 years.  But I wouldn’t bet too much on the fact that the model can spot one
single data point in time for next year or the next five years, because we actually aim at
reproducing or capturing the longer-term trend, not the short-term fluctuations of the
market.  So this confidence in longer-term trend, but of course it’s a model.  So no model
is perfect, and we can try to understand how the system works by looking at history.  So
we are limited in that sense.
 
MR. WEHRENBERG:  So don’t make your immediate investment decisions
based on this model, or any model, for that matter, but use it to understand the
relationships between – (off mike).
 
Q:  David Thomas of CENTEC.  I have another question about the model.  You
talked about the rebound effect; rebound effect is a voluntary-type action.  The major
issue that we’re addressing right now that has never been addressed before is the
population’s realization of the mass of global impact, and therefore that can drive
behaviors as opposed to simply an economic model.  Everything I saw looked like a
mathematical and economic model, and did not show any cultural awareness and
voluntary action.  So my question is does the societal section of your model address that,
and if so, how?
 
MR. BASSI:  Well, we look at the impacts of prices on consumption.  We tend to
use a behavioral formulation more than an economic formulation that shows the optimum
level for consumption or demand in expenditures and so on.  In some cases, we use
exogenous inputs such as energy conservation can be an exogenous scenario, or when we
are able to find what the causes may be, we can try to identify the strength of the
relationship.  For the specific USA model, we used exogenous assumptions and we used
a McKinsey study.  It was the carbon effect of climate change?  I think that was the study
that was looking at conservation and voluntary effects.  So again, we start from existing
studies and we try to incorporate them.  In some cases, they actually show the relations
within the system; in some other cases, they just focus on assumption and we try to
represent them the best way we can.
 
MR. SHILLING:  But that’s a very good question because most economic models
assume perfectly rational individuals making the choices.  And I’ve met a few of them,
but not very many in my career.  And one thing about this kind of modeling is if there’s
historical or academic research studies that show tipping points in behavior.  And people
were talking for a long time that $4 a gallon might be a tipping point when $3 a gallon
turned out not to be.  And they can say, well, if oil or gasoline reaches a certain price,
you’re going to start to see a very significant change in behavior.  We can incorporate
that into the model relatively easily and say when oil reaches a certain price, there’s
going to be significant changes in numbers of miles driven and things like that.  We can’t
predict that’s going to be the case, but we can say if that occurs at a various point, here’s
what the kind of reactions are going to be in the rest of the economy.  
 
And so you can think about it that way, so it helps you understand what the
relationships will be.  We’re not going to predict exactly when that’s going to happen, but
I think that we’re becoming more concerned now that there are a number of these tipping
points that are going to happen, and we need to look at what’s going to be the result.  I
mean, one clear one is people are stopping – are beginning to think they don’t want to
live 30 or 40 miles from their job; they want to live three or four miles from their job. 
And that’s going to have a major impact on the real estate sector, on transportation, on a
number of other things, so we can incorporate that in and look at the results.  We’re not
going to tell you when it’s going to happen or where, but it gives you a better way of
analyzing what’s likely to happen.  You may want to encourage it, you may not want to
encourage it, depending on the unexpected or the indirect results.
 
Q:  Phil Grossweiler (sp), I’m a science and technology fellow working for
Congresswoman Heather Wilson.  Two questions:  One is can you compare your model
with the basic MIT global energy model, in terms of the modeling approach?  Do you use
the same simulation language, numbers of variables included?  You know, just a general
perspective on that.  And the second question is will your simulation be able to, sometime
soon, answer the question, you know, should we or should we not be drilling in ANWR
or opening out a continental shelf, and help with a policy debate like that.
 
MR. WEHRENBERG:  I hope yes.
 
(Laughter.)
 
MR. SHILLING:  What was his first question?
 
Q:  Just a general comparison, you know, with the MIT global energy model.
 
MR. BASSI:  Yeah, and I can answer the first question on the MIT model.  As
Peter said, the MIT model is CG model, so it’s – the methodology’s different and the
rationale behind the model or the methodology is very different.  We can compare the
number of variables, the areas that are covered.  The first comment I would make is that
that model is made for climate change analysis; our model is not made for or conceived
for climate change analysis, so the simulation time is different.  That model runs until
2100; it looks at many different areas, such as solar radiation, such as – (inaudible) – 
volcanic activity and so on, which we don’t look at because we have a shorter timeframe
and we focus on issues or factors, and feedback groups or forces, that can influence the
behavior of this system with this timeframe.  So those are models that have different
purposes.
 
MR. SHILLING:  And that also it’s a global model, so our model is looking more
at what policy changes are going to have an impact in a country where you apply
policies, which is the way you have to deal with policies.  Unfortunately, you can’t do
policies globally, or at least we haven’t figured out how to do that yet.  
 
In terms of things like drilling in ANWR or places like that, if we had information
about how much it would cost, how long it would take, how much additional oil it would
produce, and what the side effects would be on the environment, on other things, we
could incorporate that in the model and see if it’s going to have an effect, and when that
impact on prices is going to occur.  And it would be nice to look at the same information
about the hundreds of thousands of acres of land, that are currently leased by the oil
industry and not being drilled on, to see what the impacts are going to be.  So those are
the kind of things you can incorporate in as you want to go forward.
 
MR. BASSI:  Yes, just an additional comment.  During the presentation of the
interface, it showed that we can actually increase the amount of research available and
resources.  So if you want to add all of it, it can come out with specific fields that we are
not expecting to be a variable right now but we know is out there.  You can actually say
how much can be developed, how much is technologically constrained, for the moment,
and then see what the impact of all of it over the years because there will be a delay
impacting when this side will be available, and investment connected to drilling activity,
recovering and actually shipping oil, what impact this may have.  
 
Of course, this depends on different scenarios.  You want to simulate on world oil
availability or the sensitivity of GDP with respect to oil prices, but these analyses can be
done, yes.
 
MR. SHILLING:  Most Alaskan oil now goes to Japan.
 
Q:  My name is Rod Adams and I write for Atomic Insights.  On the slide that
had the three different scenarios, three different models, with the stabilization of CO2
concentration, it looked like each one of those had very large segments for carbon capture
and storage.  Can you help us understand when there’s going to be a demonstration of
that actually working?
 
MR. SCHULTZ:  In the government right now, as we’ve configured the wiring
diagram, there’s the Climate Change Science Program and our sister program, the
Climate Change Technology Program, and I’m much less aware of the investments
they’re making.  I know they’re making significant investments in the planning and the
research behind that.  But as far as the actual implementation, I know there have been sort
of two steps forward and one step back, or maybe more steps back than that.  So I can put
you in touch with the people who can give you those answers, but I cannot myself give
you that answer.
 
Q:  I’m Robert Abreath (ph) of Sennerling (ph) Research Corporation.  First, let
me congratulate you on the concept that you’re pursuing here and especially on how
transparent you’ve made the model.  I’m in the field of validating models, and usually I
have people –
 
(Scattered applause.)
 
MR. :  I agree.
 
(Laughter.)
 
Q:  Usually I have people doing everything they can to keep me from seeing the
internal logic, so I really like that.  However, what struck me as you went through this
entire presentation is that you have a very large number of causality assumptions or
hypotheses, especially when you’re talking cross-sector interrelationships.  And I’m sure
that – reasonably sure that each time you had such a causality hypothesis, that there were
competing hypotheses.  What was the process that you followed to choose the ones that
you actually incorporated, and how have you validated those choices?
 
MR. BASSI:  Yes, thank you, very interesting question.  It goes down to the roots
of the modeling process.  Usually – well, let me say first the main structure of the T-21
model – so society, economy, and environment – was developed over the last 20 years. 
That’s been conceived by Dr. Qu here and a number of other experts, and tested for about
25 countries or more.  Millennium Institute has been collaborating with about 45
countries so far.  So most of these feedback loops or causal relations have been taken
from existing studies, well-known literature, and then applied to these different countries
to check the validity and their strength, in a way.  
 
For what concerned my specific job in modeling energy, of course we start from
existing studies.  In many cases they show correlation, and then we have to test and
validate those correlations with respect to causality, in a way.  So as a model, usually we
start with a single framework, and it’s very, very easy to make it more and more
complex.  It’s a much tougher job to create a very simple – the simplest model you can
get to actually produce insightful results.  So we start with a simple framework; it gets
more complicated than true simulation, in comparison with historical data, for different
countries or different areas.  
 
We are able to estimate whether these causal relationships are strong, are
relevant, are dominant for different time-stops or time periods.  And then, we can decide
whether to keep them or remove them.  It is very important to keep the focus on what the
issues are and understanding what are the causes that can actually define the issues, and
identify which causal relations or feedback groups are very relevant and valid for the
United States.  So it’s really a step-by-step process of looking at every single connection
by comparing it to existing studies and actual data that can show whether there is a
connection or not.
 
Q:  Good answer.  Thank you.
 
MR. SHILLING:  Yeah, that’s a key issue.  And a couple more things:  One, in
doing the validation of the model, when we do the historic projections and project over
the past 15 or 20 years with the model, to see how close it tracks actual history helps give
some validation that the causal relations we have found fit together.  And in some cases,
we found errors and data where the model showed – in population data, for example, the
country wasn’t taking adequate account of migration or things like that in the data they
had, so it helped identify the data.
 
The other thing is sometimes we will find a significant gap between what the
model is projecting and what actual history is, and then we go back and look and say why
did that gap happen.  We just had a case in Jamaica where the team that we were working
with in Jamaica went back and investigated why industrial production was lower than the
model was projecting, based on capital investment and on demand and things like that. 
And they determined that there had been, during the period there was the gap, some
significant direct foreign disinvestment.  And several firms, which were textile exporters,
had closed and moved elsewhere because they couldn’t run the three shifts they needed to
run.  The reason they couldn’t run them is the crime rate was so high that nobody would
come to work on the night shift unless they had armed escort, which they didn’t have.  So
once the government learned that, they began to address the crime issue differently and to
try to solve it.  So we’re not perfect in all of our fits, but by doing the model as
transparently – and to track history, we have an added way of checking the validity of the
various causal relations and variables that we have on the model.  
 
Now, any of them can change in the future, which is true of any model, and we
try to do the best we can and get as much expert support as we can to do that.  But that’s
sort of how we work and try to validate it as best as we can.
 
MR. SCHULTZ:  And there’s a research-wide effort that is happening under the
auspices of something called the Energy Modeling Forum.  There’s a Snowmass meeting
that’s happening right now, and I don’t know if you can hear the shrieks and the beatings
that are happening from Colorado, but it’s going on right now.  And those guys beat each
other up really hard over the different assumptions that they put into their models, and
they challenge each other, and that’s a very healthy thing for the modeling community to
be doing, to be challenging the assumptions that each other make and to kind of ferret out
what some of the weaknesses of the different approaches are.  And that’s something that
Department of Energy funds and so it’s a good thing for models to be participating in
that; it’s a very healthy exercise.
 
Q:  Hi.  I’m Greta Maseti (ph) from the Defense Science Forum.  And I apologize
if you asked – if you’ve mentioned this.  I was late coming in already.  This question is
directed to the climate change gentleman. 
 
I used to work for NASA and actually, I worked for the climate change program. 
And about 15 years ago we were talking about the need to dump all of the modeling data
into a federal and a state GIS so that resource managers, economic planners, everybody –
what’s that – did that ever happen?  What’s the status of it?  I mean, I know the models
aren’t quite ready for primetime, but has there been any progress to get this?  There’s a
lot of government employees in this room and government contractors, and these are the
kinds of people that need – and State has most of the jurisdiction over zoning and land
management, water resource development, so it really needs to get – what’s the status on
that?
 
MR. SCHULTZ:  Really good question, and I’ll give you a first – I’ll have my
company hat on first for the first part of my answer and then I’ll speak – I’ll take my
company hat off for the second part of my answer.  First part of the answer is that, in
terms of the climate model repository, there’s a really nice one.  The Program for Climate
Model Data Inter-Comparison, which is a DOE-funded thing, is a place where you can
go.  Any of you can go to the PCMDI and see all of the state-of-the-art climate models
that were used in IPCC.  And you can interrogate those models and pull them apart, and
it’s a one-stop shop for not just U.S. climate models, but models across the board
internationally.
 
As far as observations go, the story is not as pretty.  If you want NASA
observations, you go to the NASA data archives.  If you want NOAA observations, you
go to the NOAA archives.  If you want USGS observations, you go to the USGS
archives.  And there’s something in the United States that’s emerged in this
administration, called the Group on Earth observations.  It’s part of the global Earth-
observation system of systems.  And it’s a great idea; it still has not achieved the promise
of what we need, which is a place for everybody in this room to go, at least a point of
contact, to see where you can get the information.  
 
We have not achieved that level of integration, and so that’s the part that – I’m
speaking with my company hat off now – and it’s something that I feel very strongly. 
We need to do a better job of providing single points of contact for people who are not
experts in the field, to get their arms around and as – I mean, really one of the points of
this whole discussion is that the interesting stuff is at the intersections and the seams and
the cracks and the joints.  That’s, a lot of times, where you find the interesting nuggets
that have not been seized upon by others.  And to do that, you need to be able to bridge
across communities, and so you need to allow people who are experts in one field but not
experts in another field to get together and see each other’s data.  And I really feel
strongly that the federal government has not done everything that it can to help facilitate
that.  So that’s a self-criticism.
 
Q:  I’m Art Devons (sp).  I’m retired from the Atomic Industrial Forum, which is
now the Nuclear Energy Institute.  I have two questions.  The first is have you all invest –
I read recently that Disneyland is coming up with a new Tomorrowland and somehow, I
wonder if you’ve got involved with Disneyland in figuring out what the Tomorrowland is
going to be like.  Did you know that?  Has anybody got an answer to that?
 
MR. WEHRENBERG:  Well, it depends on what happened in Yesterdayland.
 
(Laughter.)
 
Q:  Second question is to Dr. Schultz.  I am a little sorry to hear that he’s an
agnostic on nuclear energy.  My own opinion is, 100 years from now, if we don’t blow
ourselves up in the meantime and destroy the world, that’s going to be our primary source
of energy.  You got anything to say about that?
 
MR. SCHULTZ:  I’m not an expert in that area and I need to be educated more. 
I’d like to learn more about what the risks are associated with it, what some of the
tradeoffs are and how it racks up in comparison to other things.  I simply am not educated
enough about it.  I can’t speak from a position of an expert.  I know that there are
probably dozens in this room who know more about it than me, so – I want to learn and I
appreciate your perspective.
 
Q:  Well, I suggest that we crank that end of the model.  Thank you.
 
Q:  Philip McCauley, Masenti (ph).  When you were describing this model, it
sounded to me like it was a fairly national-scale description of what you were – what
would happen as a result of various assumptions.  How do you handle feedback loops that
are of an international nature?
 
MR. BASSI:  Well, the USA model, from the very beginning, included some
words – global components, such as positive reduction of oil, gas, coal, and the
calculation of global prices, for instance, global demand, disaggregated in India and
China because we thought, of course, they were fairly dominant.  With the new version,
we have Canada and Mexico as well, disaggregated for demand and supply because in
the U.S., they are very important for what concerns energy trade.  So there are some
international components within the energy sector and also with economic sector, such as
exchanges with the rest of the world and so on.  So we take into account what we think
are the important global feedback loops.  There is a lot more of them; that’s why we
would like to go into, in modeling a global model, a global energy system or energy
sector.  But, yeah, we try to represent something that might be relevant to the U.S., in
terms of scenario analysis more than actually looking at the impacts that policies here can
have on other environments, like global environments.
 
MR. SHILLING:  And what we would like to be able to do, going forward, is to
do six or seven regional models that would – like the North America model, that produce
a global model but with enough differentiation by different sectors to see how various oil
production assumptions, policies and things, would affect things globally.  And we’d like
to see the world look at that as it’s moving toward Copenhagen, but we’re not sure that
any of the politicians want to get any kind of answers that we might produce.
 
MR. WEHRENBERG:  I thought you were going to go the other direction on that
question, by the way.  And just as an aside, the William Institute’s been asked to look at a
more local level; you know, looking at very local, actionable issues surrounding energy
and the environment.  So it works both directions – works in both directions.
 
Q:  Andrea, my question, I think, starts to get to what Steve was beginning –
sorry, Patrick Murphy, Homeland Security.  My question is similar, I think, to where
Steve was taking a follow-up with this, but more – rather than just local, into a more
granular view of the GDP.  And, for example, the feedback loop between energy and
GDP is obvious, but it’s not very direct.  You can get into transportation; you can get into
different modes of production, raw materials and into services, as part of the GDP.  And I
guess one of the things I’m wondering is have you looked into – or especially, are there
any tipping points from negative feedbacks where perhaps the cost of transportation
rockets so fast, when there’s a shortage of oil or oil shock, that has a serious and rapid
impact on production in general, that then causes a more rapid degeneration of GDP, if
that’s even been considered.
 
MR. BASSI:  That is very, very interesting.  We didn’t model that yet, but I think
it is a very good indication of the way forward.  Right now, what we are looking at is the
impact of energy prices or the energy scenarios on transportation issues or concerns,
portability of transportation, but not at the link between transportation and GDP.  That
has to do, I think, from a modeling point of view, with the desegregation of GDP into
more sectors.  But that’s certainly the way forward.  It has micro-impacts that may
actually be the leverage – you know, represent the leverage point.  It makes the system
not really collapse, but go different ways.  Those would be very, very relevant, yes. 
Thank you.
 
Q:  Hi, Nicole Cosman (sp) from Booz Allen Hamilton.  I was wondering if the
panelists could talk about the potential synergies between the T-21 model and the climate
models, which are very land-use, oceanographic, and atmospheric-based.  So you have
the physical sciences, where the – and the T-21 model seems to be energy-centric and
that deals with how people, you know, use energy, and you went through all the
variables.  I just see that, you know, since this talks about integration, there seems to
almost be a big gap between these models, and especially – like the T-21 model didn’t
integrate any cap-and-trade policy scenarios, which seems to be, especially on an
international scale, the likely scenario.  And I was just wondering if you could comment
on that.
 
MR. SCHULTZ:  I can start out on that.  There’s a whole family of these
integrated assessment models, and they all have varying strengths and weaknesses.  Some
are strong on the physical components, the ecosystem components, the societal
components, the energy components.  The three models that I showed results from here
actually have fairly complex energy sectors, both bottom-up and top-down.  They’re
regional models; they’re fairly disaggregated, and they also include fairly sophisticated –
or actually, the most sophisticated representations of the environment that we know that
are capable of being reproduced in the reduced form, that’s required computationally.  
 
But it’s certainly a direction for our program to increase, to the extent that we
can, the integration that, you know, I think that you are expressing by implication in your
question.  And it’s something that we need to move out in.  There are computational
limitations to doing that, to develop models that include the high-fidelity of these climate
models that are run on the world’s most powerful supercomputers.  I mean, the only other
things that are run on these supercomputers are bomb tests and a couple of other
applications, but it’s probably the most demanding civilian computational application. 
And to balance that with the need for running multiple scenarios, and sometimes an
optimizing mode in which you have to iterate.  And so there are those constraints, but
there’s a lot of work toward integrating all of those things.  And I really sort of welcome
the opportunity to talk about this whole range of models that exist, from the very high-
end, very high-resolution, down to the global.
 
MR. BASSI:  Yeah, that’s a very good answer.  I would like to add that, from the
point of view of T-21 for instance, we look at cap-and-trade policies for the U.S. that
account also for international offsets, international trade provisions, and so on.  So we are
working now with the National Commission on Energy Policy, the output – you know,
and the final report will be ready in about a couple of weeks.  We look at the Bingaman-
Specter bill, the Lieberman-Warner, and Senator Boxer’s and so on, so we use this model
approach and we add pieces as we go.  So we are a not-for-profit organization with no
core funding at this time, so we actually have to deal with our research based on projects
that we are asked to deliver.  But there is something in that area.
 
And then, I would like to add that these models can be integrated.  We did some
analysis and T-21 is mainly not energy-related.  The USA and North America models are
energy-focused, but other models, like the Mali, Jamaica, Malawi, Mozambique, models
for developing countries, are very much focused on the economy and poverty reduction
and so on.  And in the case of Mali, for instance, we used the output of the IPCC studies
and other models to see what the impact of climate change on precipitation may be in
those areas.  And in Mali, if there is a season with low precipitation, GDP goes down by
10 percent.  In the following year, if there is high precipitation, GDP goes up by 10
percent.  So we took the output of these studies and used them as an input for our models. 
Still, considering that our aim or the objective of our studies are not those of looking at
climate change developments, but looking at national development, in this case, so we
still have sort of a national perspective.  So the objectives, the boundaries, of the models
are different, but they can somehow be integrated if the issues allow to do that, or if there
is an interest to do that.
 
In fact, from the point of view – this last point – of integrating different
methodologies for the national commissioned project, I’m integrated NEMS, which is a
completely different model, the national energy modeling system of the EU, with our
approach to look at specific energy-intensive industries.  So it’s feasible if there are the
right reasons for doing that, in terms of the timeframe, in terms of the issues to be
analyzed, and compatibility of the outputs analysis.  
 
Q:  Thank you.
 
MR. SHILLING:  And just to add a brief point, building on what Steve said
earlier, no model can totally represent everything.  I mean, the world does that for us and
we just have to let it.  But these models are very complementary and focus on different
things, and what I think T-21 adds is one – the more complete integration of economic,
social, and environmental factors into a single framework with a feedback loop.  And the
other thing is the ease of use and transparency, where it runs on your laptop rather than
on the supercomputer, so that you can use it more directly and try a lot of different
assumptions, and things like that.
 
MR. SCHULTZ:  And there is a whole class of models that do run on laptops,
that are integrated assessment models, that do have those loops.  But one of the things
that I was really impressed by in this presentation was this transparency, and that’s
something that you don’t see a lot in these integrated assessment models, that are
primarily out of the research community, where you have a grad student that, you know,
has to spend three years learning how to use the model before, you know, they can
actually get results.  So yeah, I think that’s a real strength of what they’ve presented here.
 
Q:  I’d just sort of like to summarize an answer to a question about three
questions ago.  It was four years ago, in the ASCO international conference in Berlin, that
I made a proposal for a world model of energy flow, global energy flow.  And now, the
Millennium Institute has made a huge, huge step in that direction.  But I’d also
emphasize, in my opinion, we’re only halfway there.  And Jed alluded to dividing the
world into several regions, and then those regions representing the rest of the world, and
the real interactions of regions of the world’s suppliers and consumers of energy of all
forms, and that’s where we want to go.  And if anybody here has the connections or
brilliant ideas about how to get the rest of this funded, I’d love to hear from you or talk to
the folks up front.
 
MR. WEHRENBERG:  Yeah, we’re thinking about passing a hat, as a matter of
fact, to make sure that we can get this supported. 
 
Looks like we’ve run out of questions, steam, and we’ve almost run out of space
at the same time.  I got 10 seconds before they throw us out, so the timing is excellent. 
I’ll just offer a couple of things:  One, thank you very much, our panelists, for presenting
what I think is perhaps one of our more integrated views of the challenges that we face. 
And I really do appreciate that – hold your applause, hold your applause – which implies
that I want you to applaud in just a moment by the way – no, no, no, not yet, not yet. 
Peter’s a little slow.  
 
A couple of things:  I do want to mention that our next session will be September
15th.  I’m not going to announce who we’re going to have because we’ve got a couple of
alternatives in play, and if one works out you’ll certainly want to be here – and, now that
I think of it, if the other one works out, you’ll certainly want to be here, so plan on
September 15th regardless of who we end up with.  It’ll be posted on our website as soon
as possible.  We will put the presentations from this evening up on our website within
this week, and also links to the Millennium Institute’s website, where you can actually
download this model.  And I’ll offer you a challenge, as a matter of fact:  We talked sort
of tongue-in-cheek about the game; well, this is a game, in a sense, you know.  This is
Sim World, in a sense, and I challenge you to go play with Sim World.  And if you’re not
comfortable with it and not comfortable with the technology, get your kids or your
grandkids to play with Sim World.  So that would be just fine.  As a matter of fact, I’ll
offer a challenge:  Any of you who can create the most good; okay, so just go with that,
just create the most good.  That’s all we would ask.
 
MR. :  (Off mike.)
 
MR. WEHRENBERG:  Oh good, there are some brochures.  I don’t imagine you
brought CDs for everyone in the room, but – no, okay, but there are some brochures on
the model with contact information, et cetera.  If you will finally be so kind as to recycle
your nametags on the way out the door, that’s our part of reuse and recycle.  
 
Again, if you’ll please thank our panelists, we will see you in September.
 
(Applause.)
 
(END)

File Attachment: 

Schultz Transcript.pdf