Greenhouse Gas Reduction policy models.

[Jay Forrest]

Hi Paige!

I am a thesis student in Studies of the Future at the University of Houston
-- Clear Lake. I have an undergraduate degree in chemical engineering and
have a pretty strong background in biology and environmental science.

Last spring I took a course on Future Environmental Issues and, as a project
for the class, built a system dynamic model of the global carbon cycle. I
can sympathize with the desire for a useful model, but the state of current
knowledge of the carbon cycle, much less other greenhouse gases, is such
that the output of any such model is highly speculative. Based on my
experience, putting what is "known" about carbon results in significant
"closure" problems. Tweaking the model into a stable (no human emissions)
state can be done without substantial problems, but the addition of
emissions highlights the fact that "knowledge" does not simulate "reality".
I wrote a pretty comprehensive paper on the topic which I will gladly
provide at your request.

Rather than reinventing the wheel, and perhaps becoming more expert in the
failings of "knowledge" than you would like, I would suggest contacting the
folks at the University of South Australia and obtaining a copy of the IMAGE
model for greenhouse gases. The model is impressive, including carbon
dioxide, carbon monoxide, methane, nitrous oxide and chlorofluorocarbons.
It includes regional and global variables and is carefully tweaked to
simulate actual data. To reproduce this model on a functional basis would
be a major undertaking. The university site is on the favorite sites page
of my web site under "Modeling".

In addition to the people suggested in other messages, Dan Bernstein is on
this site and has done good work on carbon also.

Good luck!
Jay Forrest


Pteragenesis
Chaos, Complexity, System Dynamics, Studies of the Future
pteragen@neosoft.com
http://www.neosoft.com/~pteragen
(713)493-5022 (voice) (713)493-6715 (fax)

[Erling.Moxnes@snf.no (Erling Mox]

Reply to Paige Brown -- Paigeb@wri.org

Hatlebakk and Moxnes (1992) ("Misperceptions and mismanagement of the
greenhouse effect? The simulation model") proposes a model to be used for
experimental investigations of peoples understanding of the greenhouse
problem. If this looks interesting for your purpose, I will forward a copy
of the report.

Erling Moxnes
Erling.Moxnes@snf.no

[Jay Forrest]

Hi Tom!

Wow! What an e-mail. I have copied it into my archives as a reference
document. You refer to several games/simulations I have not heard of. Good
to have info on them.

I agree on IMAGE. It is my impression it would take 6 months to a year to
learn how to really use it/make it sing.

My main point to Paige was that building a new model from scratch is a very
time involving and to some extent arbitrary activity and would require
learning more about the ocean and biota than might be appropriate (or within
their ability). With a policy focus, perhaps they should build on an
existing model rather than starting from scratch!

My wife is copying my paper today and I should have it in the mail tomorrow!
Thanks for the help!
Jay Forrest



Pteragenesis
Chaos, Complexity, System Dynamics, Studies of the Future
pteragen@neosoft.com
http://www.neosoft.com/~pteragen
(713)493-5022 (voice) (713)493-6715 (fax)

[Jim Hines]

Id say that Tom Fiddamans response to Paige sets the standard for
responding to queries.

Jim Hines
jimhines@interserv.com

[Paige Brown]

World Resouces Institute is considering using a simulation model as a
vehicle for engaging stakeholders in the climate change debate in a
policy design exercise. Potential participants could include
representatives from environmental advocacy groups, utilities, and
petroleum companies. The ideal model would allow users to employ
different policy responses to slow greenhouse gas emissions, and
different levels of responses, to examine the impact on emissions, the
economy and perhaps various industries.

We are considering building such a model ourselves, but first want to
examine what other models exist and if we could possibly use an
existing model. So, I would appreciate any information or leads that
anyone has regarding the existence of such a model.

Thank you in advance for any help,

Paige Brown -- Paigeb@wri.org
202/662-3081

[forda@mail.wsu.edu (Andy Ford)]

Reply to Paige Brown of the World Resources Institute
on Question of GreenHouse Gas Emissions

Hi Page, this is Andy Ford from WSU with some papers on greenhouse gas
modeling:
First, Fossil2, a System Dynamics model of the US energy system, has been
used to study US policy on greenhouse gas emissions. Roger Naill has two
articles on this topic in the System Dynamics Review (Winter 1992, Summer
1992). Also, Thomas Fiddaman presented a plenary paper on your topic at
the 1995 International Meeting in Tokyo. His paper, entitled "Forumulation
Experiments with a Simple Climate/Economy Model," appears in the
Proceedings of the Toyko Conference.

Nail is with AES, Inc of Arlington Virginia. Fiddaman is a graduate
student in the MIT System Dyanmics program.

----------------------------------------------------------
Andy Ford
Program in Environmental Science and Regional Planning
Washington State University
Pullman, WA 99164-4430
(509) 335-7846
FordA@mail.wsu.edu
----------------------------------------------------------

[tomfid@MIT.EDU (Tom Fiddaman)]

(The following is a slight rewrite of my direct response to Paige Browns
query. I will try to place some of the models and papers mentioned below on
a web page soon).

Dear Paige,

John Sterman forwarded your query about greenhouse gas policy models to me.
My dissertation research involves the creation of an integrated
climate/energy/economy model that challenges some of the traditional
economic assumptions prevalent in the existing models. As part of my work,
I have replicated several of the major models and become intimately
familiar with others. Ive also done quite a bit of work on the development
of simulation games in general and on climate in particular.

Three years ago, I worked with Bert de Vries at the Dutch national
environmental institute (RIVM), which created the IMAGE 1 & 2 and TARGETS
models. We created a fairly detailed background plan for a model-supported
greenhouse policy exercise; Id be happy to send you a copy. We never
actually created the policy exercise, but the model became the basis for
the TARGETS/IMAGE energy system module (called TIME; many more acronyms to
come, unfortunately), described below.

Bert and Dennis Meadows went on to create a very simple simulation game
called SusClim. Participants represent national planners in two regions,
one rich and one poor. They manage their economy and energy system by
making decisions about investment in goods production, energy efficiency,
and carbon and non-carbon based energy production. The two regions must
allocate resources (by negotiating transfer or trade agreements) to
collectively avoid climate change. The game takes place over 5 to 10
rounds, representing 50 to 100 years, and requires about 2 hours. Decisions
are made using a game board as a visual negotiating space and allocation
tool, and then are entered into a computer, which generates a detailed
report for the next period.

The game is quite robust and generates an interesting discussion; I had
good results running it with a group of NGOs, academics, and policy makers
in Costa Rica. However, I think it would not suit your purposes for two
reasons. First, the roles in the model dont really conform to the types of
stakeholders youre interested in involving. For example, the model doesnt
generate the kind of economic data (e.g. profits) that oil companies would
presumably like to work with. Second, the model suffers from a problem that
so far afflicts all games - its difficult to get people to manage a
complex negotiation in the short run without losing sight of long run
dynamic issues. Nevertheless, you should definitely take a look at SusClim.
A report is published by RIVM, and copies of the game will soon be
available.

Ted Parson at the Kennedy School of Govt created another very interesting
simulation exercise, in which I participated last summer at IIASA.
Participants formed teams representing four regions to negotiate climate
change agreements in two periods, 2005 and 2020, spread across 2 days. The
purpose of the exercise was to promote creative thinking about climate
change abatement options and implementation, so team objectives were quite
unstructured and there was no formal scoring system or underlying model.
However, we did have two models at our disposal for planning purposes,
MiniCAM and IMAGE 2.0. While my group did explore the models somewhat, we
found that they were not tightly coupled with the negotiation process
enough to be compelling. To make a model really useful to participants in a
simulation exercise, I think one needs to be designed around the other. Ted
also previously created a one-period, multi-region negotiation exercise
about climate change based on a very simple payoff-matrix model.

Models

An excellent place to start looking at the variety of models available is
CIESINs Thematic Guide to integrated models, on the web at:

http://sedac.ciesin.org/mva/

In particular, take a look at:

http://sedac.ciesin.org/mva/iamcc.tg/TGsec4.html
http://sedac.ciesin.org/mva/mvas.home/compare.html

Many of the models on their pages are available, and I have replicated a
few of them. My comments on the models I know best follow:

TIME, TARGETS, IMAGE - The IMAGE model is primarily an impacts model, with
detailed atmospheric chemistry and a spatially disaggregated climate and
biosphere. It generates very compelling images of regional changes in
agriculture, sea level, etc. Unfortunately, its too big to run
interactively, but RIVM has created libraries of scenario outcomes which
can be explored through a nice interface. It would be hard to build a
simulation around IMAGE, but it could be a nice supplement to other models.
The TIME and TARGETS models, on the other hand, have a lot more detail
about the energy system and the economy - about what you would need to
create specific reports for utility or petroleum interests. TIME is
available in ithink, a simulation package for which it is easy to generate
a gaming interface, and I have translated it to Vensim, which provides even
more extensive interface capabilities. Both packages run on Macintosh &
Windows systems. The TARGETS model was developed in RIVMs own workstation
model-visualization language, M, but has a web-based interface. You should
contact Bert de Vries for more details about these models.

DICE & RICE - Nordhaus DICE model is probably the most widely available
integrated model. Its very simple - only a few dozen equations, so it can
be readily explained. Its also the model which best embodies neoclassical
economic assumptions which some of your participants would likely find
highly suspect. A model at this level of complexity would lend itself to an
exercise that focused on the evolution of the climate problem over time,
rather than on a single detailed negotiation. Ive replicated DICE model in
Vensim, and would be happy to provide you with the code or a run-time
version; Im sure you could also get it in its original form from Nordhaus.
The RICE model is a regionalized version of DICE. Again, its simplicity is
a great asset, but if you want to involve participants from developing
regions I think they would be incensed by the models implicit assignment
of a greater weight to individuals in richer regions when calculating
global welfare.

Yohe/CONNECTICUT - This model adds an explicit energy sector, with carbon
and non-carbon fuels, to the DICE model. Thus its quite similar to the
model that underlies SusClim. While the model is implemented in GAMS, it
could be readily translated to another language that supports an
interactive interface.

MiniCAM - The MiniCAM model is a general equilibrium energy-economy model
coupled to climate and impact modules. It has a lot of regional and
sectoral detail, which would be helpful if it were to be coupled to a
negotiation exercise with multiple regions. However, the level of detail
would be daunting unless you somehow restrict participants to a fairly
simple set of policies. There are several other detailed general
equilibrium models available, including OECD-Green/MIT-EPPA (too big to
use, I think) and MERGE (much simpler).

ICAM - The ICAM model is one of several designed to focus on the
propagation and analysis of uncertainty. The model generates distributions
of outcomes, rather than single numbers, and would certainly help
participants develop a healthy appreciation for the uncertainty around the
climate issue. Its developers at Carnegie Mellon have done a lot of
interesting work looking at the implications of uncertainty for different
stakeholders. You might visit their web pages at
http://miami.epp.cmu.edu:80/global_change/. The model has considerable
regional detail and four energy sources (coal, oil, gas, renewable), but I
think its rather weak dynamically. The software in which its implemented,
DEMOS, has a graphical interface, but I dont think it would support
interactive gaming very well, and the model is too slow anyway. However, it
might be possible to create a simplified version; Im about half way
through the process of translating the model to Vensim, but I probably
wont finish.

My own work focuses on the implications of adding behavioral decision rules
(rather than optimization), induced technological progress, explicit
capital stocks, and other kinds of feedback complexity to the
energy-economy system. My model is similar in its basic structure to the
CONNECTICUT model, but incorporates very different assumptions. The major
lesson is that, contrary to what some models suggest, it is not advisable
to wait ten or twenty years, and then take action only if climate change
appears to be a problem. It is possible to explore a much wider spectrum of
assumptions about whether the economy is near an equilibrium or optimal
state, and to test some ideas which lie outside the economic paradigm. My
model will be fully documented and available in a run-time version in a few
months. I hope to create a gaming interface for it as well, though that
will have to wait until my dissertation is finished in January. Id be
happy to provide you with a copy of the model when its ready.

Two fellow PhD students at MIT, Liz Krahmer and Scott Rockart, developed a
model of carbon sequestration in Costa Rica, which looks at the trade-offs
between carbon sequestration and other land uses and at the long-term
future of Costa Ricas energy system. While the model is not really mature,
its already quite interesting. It suggests to me that it might be easier
(especially from a modeling standpoint) to develop an engaging simulation
exercise with realistic roles around a regional aspect of the climate
problem.

I hope this information is helpful to you in your search for existing models.

- Tom Fiddaman

______________________________________________________________
Thomas Fiddaman, PhD Candidate
MIT System Dynamics Group
E60-355, 30 Memorial Drive, Cambridge, MA 02142
MIT: 617-253-3958 home: 603-497-2273 email: tomfid@mit.edu
______________________________________________________________