2008/07/15 09:10 Steve Carpenter, "Scenario Thinking to Solve Complex Environmental Problems", ISSS Madison 2008

ISSS Madison 2008, 52nd Annual Meeting of the International Society for the Systems Sciences

This digest was created in real-time during the meeting, based on the speaker's presentation(s) and comments from the audience. The content should not be viewed as an official transcript of the meeting, but only as an interpretation by a single individual. Lapses, grammatical errors, and typing mistakes may not have been corrected. Questions about content should be directed to the originator. The digest has been made available for purposes of scholarship, posted on the ISSS web site by David Ing.

Gary Metcalf

Today, moving up to ecological systems, the earth

Steve Carpenter, research into inland water

  • Chair of a section of the Millenium Ecosystem Assessment


Scenarios works

Maybe not the only way, but an example of the way it works


  • Uncertainties
  • Scenarios
  • Summary, gaps, needs

Recognized uncertainties, imaginable outcomes, unasked questions

Even when we know what questions to ask, or forecasts are wrong

  • Evaluation of world-ending disasters

Some computable, e.g. super volcanoes

  • Some completely non-computable:  no experience with robots taking over the earth
  • Global warming physically computable, but social changes aren't

Computing probability of extreme events, surprisingly large

  • Gaussian:  if you know everything, it's pretty well
  • Tails thin, probabilities drop like a rock
  • If you don't know everything, then you have to use Student-T, which has fatter tails
  • If you've only seen one event in history, the distribution is almost flat
  • Probabilities are high when low degrees of freedom

Often, we don't know what questions to ask, we've failed to perceive what we know

  • e.g. dust bowl of great plains, 1930
  • Farmed in a model that worked in east
  • Thousands of farmers used wrong methods
  • Risk not computed, because questions not asked

All possible futures

  • Have models and observations
  • Helps in recognized uncertainties
  • But then there's so many unasked questions
  • We can try to colonize the space of unasked quesitons with imaginable outcomes
  • This is the task taken on by scenario thinking:  increasing the space of imaginable outcomes

Scenarios and process

Scenarios begin from available perspectives in the world:

  • Heterogeneous:  Some may be more competent, some may be louder
  • Need to sample and integrate them
  • Each perspective brings a small part
  • Averages may bring a blurry picture, a serious limitation to go forward
  • Better to struggle for a more precise view
  • Integrated view may be strange, but it's worth it to get there

To obtain scenarios, sample perspectives, cluster the perspectives and condense to a few scenarios

  • Scenarios have to be 2 to 5, and have to be even, so politicians won't try to just take the middle one

Millennium Ecosystem Assessment

  • 95 experts, 25 countries
  • Met on every continent
  • Figure out the future of life support in 2050
  • Traced out 4 stories from 2000 through to 2050

Quick summary:  Cork, Ptereset, et al. 2008

Poor / rich, and rural / urban

1. Global Orchestration: globalized

2. Order from strength: globalization fails

3. Adapting mosaic: heterogenous development path, rich and poor more intermingled, experiments, failures and successes, and people learned

4. Technogarden: globalized, centralized, massive investment

Combined information from qualitative assessment, quantitative data

  • Ran out of time, harmonization was less than perfect

How to use scenarios?

Example:  the U.S. need to spend $1.6 trillion to replace failing infrastructure

  • Could just rebuild the same stuff, same way
  • Or, could upgrade infrastructure and bring up to current standards, modern
  • We could completely rethink infrastructure for a changing world, e.g. challenges of epidemics that the transportation system needs to deal with, and then re-engineering around challenges
  • Scenarios gives framework, consistent logic and models to evaluate responses
  • Decisions robust across scenarios are the best choices


  • Playful thinking is serious business, means protected space
  • Need positive stories, better than negative visions
  • Need blunt warnings of dangerous paths
  • Need shadow networks to link groups outside of global politics, a source of resilience
  • Collective thinking by diverse cross sections of people is powerful


  • Few people trained for collaborative thinking:  lots of great people but finding and bringing them in is hard, could modify university training to make people more facile
  • Need institutions to look at forward-looking processes, need find to ways to get around the blockages
  • Have examples of structured processes to bring together multiple process, science and art
  • Need information for scenarios:  e.g. state of environmental monitoring, spotty, and getting worse in many places in the world
  • Need tools for networking, mapping, telecommunications


  • Education:  good at educating experts, but need to combine expertise with skills to collaborate with others, and skilled public
  • Shadow networks, free of internal governance
  • Need ongoing assessments of planetary life support, have IPCC (focused on climate) that we could build on, but needs to be ongoing

Each of us should think about planning for the future


Who's in the room, structured dialogue?  Generalists, subject experts?

  • Depends on scope
  • If regional, analyze the system first, and get political actors, economic actor, get a feel
  • For global scenarios, not really going to get 6.3 billion people together
  • Have seen global scenarios, a lot done already
  • Then did a telephone survey of leaders, heads of states of countries, presidents of environmental NGOs, all househouse names
  • Asked about resilience of the world, hopes and fears
  • Expertise: regional teams, try to bring in 6 or 8 people from demography, economic, ecology, range of ages, range of countries
  • Resiliance Alliance has brought in a lot of people
  • In addition to experts, there are team / fun people
  • In Millenium Assessment, had to represent world the best we could, much harder to run, less control over personalities

Gaps and needs, training collaborative skills.  Root process methodologies that you like?

  • Approach to pick people that I think are good at it, then learn by doing
  • It's been pretty much seat of the pants
  • Online workbook, resilience workbook, and we use that to train young people to participate
  • Mostly, we just jump in and do it
  • I run a course in scenario thinking with practical parts, and find that students are great at this
  • There must be formal ways of doing this teaching that I don't know about

Structured processes, science and art.  Projective scenarios and backcasting scenarios.  Probabilistic and possibilistic normative.

  • There's a large literature on forecasting, backcasting, normative and non-normative.
  • Approach that I've taken is to focus on peoples' views of resilence, hopes and fears, and find possible pathways to the future
  • Not setting normative to the future
  • After the scenarios have been set, that NGOs often taken them to develop normative goals
  • In projecting environmental futures, have low confidence in available models extrapolating from the past to the future
  • The parameters are out of historical ranges
  • Also, uncertainty estimates are wildly overstated:  tails are compressed far more than they should be
  • Those models are likely to be wrong, although that doesn't say they're useless, because they have good principles, e.g. conservation of mass -- is there enough phosphorous on the earth to do that?
  • Forecasting is weak, and don't do backcasting

What kind of systemic tools are you using?  Systems dynamics?  Dynamics for social networks?

  • Different experts from teams bring their own systems views
  • Typical systems dynamics types of models