Environmental Modeling and GIS

Intro

• Top Down: Dynamical Systems, ie STELLA, stock and flows, LinReg

• Bottom Up: Cellular Automata and Agent-based, state vectors

• Models don't need to be "reality based", examples: Daisyworld, gravity

• Why? Because we can! Geocomputation capacity... Also increased pressure and awareness

• Integration of EM with Policy, example Klamath Basin

• Good data make good models

Modelling

• Data intensive, cross-disciplinary, dynamic, complex

• In a fuzzy world between natural science and social science

• Microworlds

  • Ontology: contents, spatial structures, temporal structures, physics, logic (conceptualisation of a knowledge domain)

  • Internal consistency more important than capturing all of reality

  • Cellular Automata, Predator-Prey

  • Open vs Closed systems

  • Landscape evolution example, SS thickness

  • Start with the rules and relations (data model of reality), then instatiate the geocomputation (CBMOR)

  • Write down everything, throw away as much as possible, sensitivity analysis

  • Glacier regions example and hierarchies discussion

  • Choice of variables for purpose, ie policy vs research

  • Physics vs physics-like

Complexity and Systems Thinking

• Relationships between entities, ie feedback loops

• Chaos, self-organization, emergence, etc

• Codifying the "sum is greater than the whole of parts"

• Simple countable, large uncountable disorganized, organized complexity

• Approaches

  • Top down - Systems analysis, decomposition into subsystems

  • Bottom up - Self organization, "surprise" from simple rules

Geocomputation

• Visualization as a substitute for formal analysis

• Object-orientation vs procedural (top down)

• Loosely coupled vs Tightly Coupled

• Distributed real-time analysis and modeling via internet and XML (TerraML)

• "Computational Theory of Complex SpatioTemporal Processes"