Axis-aware observable scoring: score observables defined on structured axes

[jc-macdonald]

Julia parallel of trade-study (Python) #39. Extend the scoring framework to handle observables that are functions of structured axes (trait, spatial, age).

Tasks

• Define AxisMetadata type: name, bounds, grid, axis_type
• Extend StructuredObservable protocol to carry axis metadata
• Implement axis-aware reduction: integrate over subdomains, slice at coordinates
• Ensure scoring rules handle function-valued observables
• Test with TRIDENT-style observables

Cross-references

• trade-study #39 (Python parallel)
• TradeStudy.jl N-independence diagnostic: representation-sensitivity scoring rule #23 (N-independence diagnostic — an axis-aware score)
• OpSystem.jl TrialResult and ResultsTable types #2 (dynamic axis management — source of axis declarations)
• OpEngine.jl TrialResult and ResultsTable types #2 (adaptive grids — axis resolution affects scoring)