ZZ M2GD Modeling Workflow
The objectives depend on the uses of the modeling project:
- Understanding - Design a mathematical abstraction of the real world that helps to identify gaps in our mechanistic understanding of processes.
- Prediction - Design a mathematical abstraction of the real world that will be used to predict the response of the system to changes in inputs.
- Engineering - Design a mathematical abstraction of the real world with the goal of producing certain outputs given certain inputs.
Complexity vs Simplicity
- Mechanism vs. Empiricism
- Alternative Formulations (stochastic?)
An iterative process to assemble, build and manage the tools will be used to build the model and carry out the subsequent steps: Parameterization, Model Testing, and Model Analysis.
- Code, Compile, Test, Commit, Deploy
- Literature-based
- Empirically based
- Calibration (Tuning)
- Role in Scaling
- Test model vs. development data (verification)
- Test model vs. independent data (validation)
- Test model vs. new observations (extrapolation)
- Parameter Sensitivity and Uncertainty
- Formulation Uncertainty
- Conceptualization Uncertainty
Steps that must be done before and after running the model when the model is being used in any of the parameterizing, testing or analysis steps.
- Preparing input data
- Preparing parameter data
- Setting up model run
- Refining outputs (summarizing, visualizing, etc)