Methodology
Design Considerations
We are actively working through several design questions:
Constraint Specification
How should managers specify their constraints? Options range from simple forms (seed count, labor hours, budget) to more complex spatial constraints (must be within 500m of roads, avoid sensitive areas).
Strategy Definition
Should users define candidate strategies manually, or should the system generate candidates from constraints? Manual definition is more transparent; automatic generation may surface non-obvious options.
Output Metrics
What metrics best support decision-making? Occupancy at year 30? But this loses temporal nuance. Full trajectories? But these may be overwhelming. We are exploring summary statistics that preserve decision-relevant information.
Uncertainty Presentation
Each strategy will be simulated many times to characterize uncertainty. How do we present the range of outcomes in a way that supports rather than paralyzes decision-making?
Architecture
From the perspective of Vegetation Modeling, each scenario is a completely independent simulation. The vegetation model doesn’t know or care that scenarios are being compared.
Resource Optimization’s job is:
- Cataloging — keep track of which simulations correspond to which strategies
- Triggering — drive the vegetation model to run appropriate scenarios
- Collecting — gather outputs from completed simulations
- Structuring — present comparisons in a decision-relevant format
Example Workflow
A manager might say: “I have 10 Joshua tree seedlings, limited labor, and seedlings need to be within 500m of roads for watering access. Compare passive recovery vs. seeding high-severity areas vs. planting seedlings.”
Resource Optimization would:
- Translate these constraints into three simulation specifications
- Drive josh to run each scenario (e.g., 1000 replicates each)
- Collect outputs (spatiotemporal rasters of predicted vegetation)
- Structure comparison: “Strategy A: 15% occupancy at year 30, high variance. Strategy B: 35% occupancy, moderate variance. Strategy C: 40% occupancy but 3× labor cost.”
The manager then decides which tradeoffs are acceptable given their values and context.
We Want Input
If you have experience with post-fire resource allocation decisions or ideas about how to structure scenario comparisons, please contact us.