Scientific Grounding

This site is a teaching project, but it should still sit on solid scientific ground. The workflow combines three established ideas:

1. environmental data preparation,
2. spatial decision support,
3. binary optimization models used by quantum-inspired and hybrid solvers.

Environmental Data Preparation

The emulator does not work directly on raw geospatial layers. It needs a decision table. That is why harmonization remains part of the repository.

Analysis-ready geospatial systems such as the Open Data Cube emphasize consistent, well-organized Earth observation data because downstream analyses depend on stable spatial and temporal inputs. In this repo, the harmonizer plays a smaller but similar role: it helps convert environmental layers into comparable site features.

Spatial Decision Support

The site-selection example is deliberately aligned with systematic conservation planning. The classic framing from Margules and Pressey (2000) emphasizes that conservation decisions must account for representation, persistence, threats, and broader landscape context.

Tools such as Marxan and prioritizr show how this becomes computational: candidate planning units, biodiversity features, costs, constraints, penalties, and solver outputs. The ecological monitoring demo is not a replacement for those mature tools. It is a quantum-ready teaching analogue.

QUBO And Ising Models

A QUBO represents a decision problem using binary variables and pairwise terms. That form is attractive because many hard combinatorial problems can be mapped into binary quadratic or Ising-style models.

Useful technical starting points:

• Glover, Kochenberger, and Du provide a tutorial on formulating and using QUBO models.
• Lucas surveys Ising formulations for many NP problems.
• D-Wave Ocean documentation describes samplers that accept quadratic models and return variable assignments.

What "Quantum-Ready" Means

Quantum-ready means the problem has been expressed in a form that compatible quantum-inspired, annealing, or hybrid solvers can accept. It does not mean a quantum computer was used, and it does not imply quantum advantage.

The default workflow runs locally on classical hardware. That is intentional. For near-term quantum technology, Preskill's NISQ framing is a useful caution: current and near-term devices are scientifically important, but broad practical advantage should not be assumed.

What Counts As Evidence Here

The evidence in this teaching repo is not speedup. It is interpretability:

• Can we trace data from harmonized layers to decision-table columns?
• Can we explain what each binary variable means?
• Can we justify each reward and penalty?
• Can we compare emulator results with a classical baseline?
• Can we map selected sites back to geography?
• Can we rerun scenarios and see whether conclusions are stable?

That is the scholarly value of the emulator for EDS training. It gives researchers a reproducible way to learn quantum-ready optimization while keeping ecological interpretation and methodological caution at the center.