📚 Training Ideas & Test Cases
Sensor calibration experiments, how-we-got-there journey logs, and ideas being developed into active missions. These are the test beds that taught us what the tools can actually do.
🧪 Completed Test Cases
Newell Eddy — ICESat-2 Depth Ladder (ATL03 / ATL24 / ATL29)
Newell Eddy is a known 165 ft freighter sitting ~7.3 km west of Spectacle Reef. The goal was to test how far we could push ICESat-2 photon penetration using ATL03 (standard), ATL24 (shallow water classified), and the experimental ATL29 deep-window search mode — and to understand the limits before committing these tools to the Piper Tier A box.
| Product | Track | Photons | Max subsurface reach | Result |
|---|---|---|---|---|
| ATL03 | 2020-06-14 (best track) | 53,604 | 38.4 m (p95 23m) | Partial — track hits Newell geometry; clear days missed in 2024 |
| ATL03 (2024 clear days) | Sept/Oct 2024 — 231 CMR granules | 0 at pin | — | Track geometry misses Newell; still 231 granules to scan |
| ATL24 | Same granules | 0 classified | — | Shallow-water classifier needs photons present first |
| ATL29 | Applied to 2020-06-14 | 887 subsurface | ~40m window | ATL29 widens search window only — does not create new photons |
| ATL03 (Spectacle Tier A) | 2024-10-04 granule | 16 | — | Sparse at exact pin; wider reef gets 260 photons |
| ATL03 (Cedarville cal.) | Any clear pass | 1,092 | ~77m p95 | Stack works when track intersects — shallow target confirmed reachable |
Status: Calibration complete. Data lives in the active mission. Newell remains as a depth-baseline cross-check — not a primary target.
1959 Wyman Ice Drift — Free-Drift vs High-Windage Calibration
Two drift charts generated to test CESAROPS ice-drift proxy physics for the Wyman person-on-ice hypothesis. Goal: understand if a thin floe + person could reach Spectacle Light from the ditch in a physically plausible time.
| Mode | Windage | Result @ 2h | Verdict |
|---|---|---|---|
| Free-drift companion | 0.03 (Nansen 3%) | 1.22 nm @ ~046° backward from light | ✅ Matches ditch scale — preferred proxy |
| High-windage sail | 0.10 (upper bound) | ~12 nm / 6h — overshoots | Upper bound only; multi-hour integration not physical for 1nm hop |
| Forward from ditch (NE 2h) | 0.03 | 0.22 nm to light | Short closing motion consistent with Wyman's note |
💡 Ideas to Develop
Newell ATL03 Clear-Day Geometry Hunt — 231 Remaining Granules
231 CMR granules in newell_cmr_search.json were catalogued but not downloaded.
The goal is to find a clear-day (low WVHT) track that actually intersects the Newell Eddy pin
geometry at full 50m depth penetration. The 2020-06-14 track hit but only reached 38m.
A clear-day match would confirm the depth ladder for the Spectacle Tier A extrapolation.
Aircraft-Class Windage Entry for Drift Engine
The drift engine models ship / lifeboat / life_ring / body / debris / cargo but has no
aircraft_debris class. For a downed Piper specifically, debris windage would be
very low (aluminum panels + dense engine). Adding an aircraft-class entry would make the
drift engine useful for general downed-plane over-water cases, not just the Wyman seabed search.
Crib / Lighthouse Structure Exclude Mask (GeoJSON)
The limestone crib and charted light structure at Spectacle Reef Light (45.77333, −84.13667) will register as false-positive anomalies on any bathymetric hillshade or photon density map. A GeoJSON exclude polygon around the crib should be emitted before any BAG/hillshade anomaly promotion step so the stone structure is never promoted as a candidate hit.