Research / IcebergsICESat-2 · Labrador Sea · 55–65°N
Iceberg Drift Forecasting
Automated calving-front detection using ICESat-2 photon-counting lidar, paired with a Kalman filter to project 30-day drift corridors and covariance growth. Data refreshes weekly via GitHub Actions.
| km² tracked total | 4,782 |
|---|---|
| active tracks | 3 |
| mean drift km/day | 12.4 |
| last refreshed | 2025-05-10 |
Active tracks ICESat-2 passes
| Track | Area km² | Freeboard m | Confidence | Last position |
|---|---|---|---|---|
| LB-2024-001 | 8.3 | 12.1 | 0.91 | 56.10, -50.70 |
| LB-2024-002 | 3.1 | 8.4 | 0.84 | 61.30, -47.20 |
| LB-2024-003 | 15.7 | 18.2 | 0.96 | 54.10, -52.70 |
Drift forecasts Kalman covariance, not promises
Each row is a forecast horizon: the filter's mean position and the 1σ error ellipse (major × minor axes, in km). The ellipse grows with the horizon — that growth is the honest part of the forecast.
| Track | Horizon | Mean position | σ ellipse km |
|---|---|---|---|
| LB-2024-001 | +1d | 55.90, -50.40 | 8 × 5 |
| LB-2024-001 | +7d | 55.30, -49.10 | 22 × 14 |
| LB-2024-001 | +14d | 54.80, -47.80 | 41 × 26 |
| LB-2024-001 | +30d | 53.90, -45.20 | 78 × 49 |
| LB-2024-002 | +1d | 61.10, -46.90 | 6 × 4 |
| LB-2024-002 | +7d | 60.60, -45.90 | 18 × 11 |
| LB-2024-002 | +14d | 60.10, -44.80 | 34 × 21 |
| LB-2024-002 | +30d | 59.20, -42.50 | 65 × 40 |
| LB-2024-003 | +1d | 53.90, -52.30 | 11 × 7 |
| LB-2024-003 | +7d | 53.30, -51.00 | 29 × 18 |
| LB-2024-003 | +14d | 52.80, -49.60 | 54 × 34 |
| LB-2024-003 | +30d | 51.90, -46.90 | 103 × 65 |
Pipeline
- source — github.com/roshantaneja/icespy
- data — mock beacon JSON until the real pipeline publishes releases
- sister site — Rainwater for the Maasai (tropical scarcity to this page's polar excess)