DECISION SUPPORT · RASUWA, NEPAL
From observation
to the next decision.
Compare the evidence. Choose where to look next. Deliver a review brief with its sources intact.
Window counts, not land-area coverage. No field-validated damage labels.
01 / REVIEW QUEUE
Where to look next
Change share is the fraction of valid 40 m tokens above the ordinary-change threshold. It is not damage probability.
03 / DECISION & DELIVERY
Your review list 6
Compare the places you chose, then deliver their evidence together. The six published leads are selected initially; your selection is not an automated response plan.
| Place | Change share | Observable | Your status | Selection |
|---|---|---|---|---|
| v003 | 13.3% | 45.1% | Not reviewed | |
| v025 | 13.2% | 90.6% | Not reviewed | |
| v024 | 10.7% | 94.6% | Not reviewed | |
| v027 | 8.1% | 94.4% | Not reviewed | |
| v022 | 7.3% | 86.7% | Not reviewed | |
| v002 | 7.1% | 60.2% | Not reviewed |
Includes scores, observation dates, source checksums, limitations, and your notes. No account required.
Review notes and selection are private to this browser until you share a download.
METHOD & ACCOUNTABILITY
Keep the evidence attached.
Frozen OlmoEarth v1 Base encodes the same locations before and after the event. Windows are ordered by the share of valid tokens above the pooled p99 of three ordinary transitions.
Optical comparison intervals are not fully time-matched. The method did not outperform a strong post-event NDWI baseline in the token-scale external-proxy evaluation. This optical workflow is separate from the radar fusion experiment.
Read the research and corrections ↗What comes next
Versioned administrative indicators, infrastructure context, and a retrospective Arab-region case would extend this workflow. They are proposed work, not results shown on this page.
efd1bb01899fPOOLED OPTICAL THRESHOLD0.316978Original observable windows ↗Original six review leads ↗