Skip to content

Ticket 033: Continuous Spatiotemporal Wind Grid

Status: implemented.

Goal

Model wind as a deterministic function of position, altitude, and elapsed time beyond the current constant, altitude-layered, and scenario wind-change models.

Current Gap

This ticket is complete. Wind can be constant, altitude-banded, changed at scenario event times, or supplied through a deterministic spatiotemporal grid asset.

Scope

  • Add a deterministic spatiotemporal wind provider interface.
  • Define a versioned wind-grid input format.
  • Add interpolation rules for latitude, longitude, altitude, and elapsed time.
  • Add bounds and missing-data diagnostics.
  • Keep existing ConstantWindProvider, LayeredWindProvider, and scenario wind_change behavior compatible.
  • Add estimator, scenario, CLI, and fixture coverage.

Acceptance Criteria

  • Users can run deterministic estimates against an offline wind grid.
  • Every wind sample is reproducible for a given input grid and route state.
  • Unsupported or incomplete grid data fails explicitly.

Delivered

  • SpatiotemporalWindProvider in estimator/environment/wind.py — quadrilinear interpolation across time, altitude, lat, and lon axes; flat tuple storage with stride-based indexing; clamping at domain boundaries
  • _lerp and _interp_index module-level helpers using bisect.bisect_right
  • wind_provider_id updated to return "spatiotemporal_grid" for SpatiotemporalWindProvider
  • adapters/wind_grid.py — load_wind_grid and WindGridLoadError for YAML/JSON grid files
  • Load-time validation: axis monotonicity, minimum 2 entries per axis, exact values shape check
  • assets.wind_grid_file field on MissionAssets schema
  • CLI loads wind grid from assets.wind_grid_file; --wind-layer flags take precedence
  • WIND_GRID_SCHEMA_VERSION = "wind-grid.v1" and wind_grid provenance in EnvelopeInputs
  • SpatiotemporalWindProvider exported from estimator
  • examples/wind/pipeline_wind_grid.yaml example grid
  • Golden fixture scenario tests/fixtures/golden/spatiotemporal_wind/
  • 22 new tests in tests/test_wind_spatiotemporal.py

Integrated Surfaces

  • Mission YAML uses assets.wind_grid_file to reference offline wind grids.
  • Wind examples live under examples/wind/.
  • The estimate and scenario CLI commands load wind-grid assets through the same mission YAML path used by library callers.
  • Wind-grid behavior composes with vehicle YAML, mission route actions, terrain assets, geofence checks, landing-zone checks, scenario wind changes, result envelopes, Markdown reports, and golden fixtures.
  • Scenario initial wind settings override mission wind-grid assets when set; otherwise scenarios can inherit assets.wind_grid_file from their referenced mission. wind_change events can still replace the active wind provider at deterministic timeline triggers.

Out of Scope

  • Live weather API calls from core estimation.
  • Probabilistic weather ensembles.
  • Meteorological forecast validation.