1. (c) 2013 California Institute of Technology. Government sponsorship acknowledged. National Aeronautics and Space Administration GEOSS Future Products Workshop Silver Spring, MD USA 26-28 March 2013 Gary Geller Jet Propulsion Laboratory California Institute of Technology The Model Web for Biodiversity and Ecosystems SBAs Consultative infrastructures for decision makers and researchers

2. Overview  The problem  Ecology is problematic  Example Model Web  Growth phases  Wrap-up 2

3. NASA Biodiversity and Ecological Forecasting Team Meeting College Park MD ~2008 3

4. What do decision makers really want? 4

5. 1.What has / will change? 2.Why? 3.What are the impacts? 4.Which management option is best?

6. Specific Examples  BCC: What are the ecological and social implications of complete “build-out” in Boulder according to the existing regional zoning plan?  USGEO: What are the consequences of alternative land management practices on biological diversity in the context of climate change?  NEON: How do changes in the availability and distribution of the nation’s water affect ecological systems?  DS: How will the boreal forest shift as temperature and precipitation change at high latitudes? What will be the impacts on animal migration patterns and invasive species? 6 How will social and economic factors change forest cover of Peruvian Amazon by 2020? What effect will various management options have on the critical values of my park? What do we need to do to restore the landscape of an overgrazed African ecosystem? How will climate change and socioeconomic factors affect infection rates of vector-borne diseases? How will increased development and forest fragmentation effect the abundance of a forest-dependent bird species? What areas in Burma should be targeted to protect the critically endangered Gurney's Pitta? How much will sea level rise? How will water availability be affected, by annual and seasonal measures? Do we need to build more reservoirs? How will urban temperatures in SF be changed? How will fire risk and intensity be affected? How will snowfall be affected? How will flood risk be affected?

7. How many of these questions can be easily answered, or explored, now? Can a resource manager easily get information on these topics? 7

8. The Modeler’s Condition Its hard to make predictions, particularly about the future 8 We need to make things easier

9. Vision 9

10. Ecology vs other Disciplines  Meteorology: “its just physics” Translates easily into equations and models Straightforward observations (eg, balloons) Spatial scale of 100s of km—sparse observations Temporal scale of variation is days  Physical oceanography: more difficult Still physics Observations harder (eg submersibles) Spatial scale of 1s or 10s of km—more obs Temporal scale of variation is much longer—more obs

11. Ecology vs Other Disciplines  Ecology: very difficult Not physics—theory alone insufficient Science is complicated Species don't behave like predictable fluids Each species is different Interactions hard to predict Observations required is huge Observations difficult to get …

12. 12

13. GCMs Physical oceanography Primary producers Global Climate Where? How much? Temperature Salinity Currents Nutrients Met data 13 Model as a Service

14. GCMs Physical oceanography Small fish Primary producers Global Climate Where? How many? Where? How much? Temperature Salinity Currents Nutrients Met data 14 Model as a Service

15. GCMs Physical oceanography Small fish Primary producers Global Climate Big fish Where? How many? Where? How many? Where? How much? Temperature Salinity Currents Nutrients Temperature Salinity Currents Nutrients Met data 15 Model as a Service

16. GCMs Physical oceanography Small fish Primary producers Global Climate Big fish Where? How many? Where? How many? Where? How much? Temperature Salinity Currents Nutrients Temperature Salinity Currents Nutrients Met data 16 Model as a Service Model Web

17. GCMs Physical oceanography Small fish Primary producers Global Climate Big fish Where? How many? Where? How many? Where? How much? Temperature Salinity Currents Nutrients Temperature Salinity Currents Nutrients Met data 17 Consultative Infrastructure

18. GCMs Physical oceanography Small fish Primary producers Global Climate Big fish Where? How many? Where? How many? Where? How much? Temperature Salinity Currents Nutrients Temperature Salinity Currents Nutrients Met data 18 Keystone Models

19. Websites  Critical component  Connect people to information  Audience-specific 19

20. Growth Phases (Ecological Perspective) 20

21. Getting Started  Keystone models  Small, planned focused webs  Start simple  Minimize barriers to entry  Resist complicated metadatsets  Let some standards emerge  Don’t require new technologies 21 } “Bootstrapping”

22. Cowboy-Fascist Continuum  Complete freedom  Informal, little planning or regulation  Little/no automation  Minimal barrier to join 22 CowboyFascist  Complete control  Large investment in planning and regulating  Full automation  Higher barrier to join

23. Growth Phase  Individual models added  Small, planned webs added  Good models/webs become popular  Continual feedback and improvement  Gradual increase in automation and control 23 CowboyFascist

24. Mature Phase  Growth continues  Some parts highly automated  Other parts simple, raw  Have a consultative infrastructure 24 CowboyFascist …

25. Numerical Weather Prediction  1904First proposed  1922Detailed  ~1950First performed  1958Became useful  1970sBecame good  NowGlobal infrastructure Pag e 25 skepticism

26. Wrap Up  WWW for models  Remember Stefano’s Basic Principles  Actively encourage participation  Facilitate organic growth  Keep entry barriers low  Shift culture towards expectation of sharing  Keep it simple, flexible, scalable 26