Planet makes available select imagery for major disaster events, including major earthquakes, floods, storms, wildfires, and human-made disasters. These are available here in a STAC Catalog under a CC-BY-NC license. See the Planet Disaster Data page for more information.
Planet satellite imagery of the 26 August 2026 outburst flood on the Bhote Koshi–Trishuli river corridor, published as a self-contained STAC catalog.
On 26 August 2026, an outburst flood swept the Bhote Koshi–Trishuli river corridor across the Nepal–China border, killing at least 160 people and leaving hundreds missing, including tourists and workers. Nepal Police reported at least 157 deaths, while Chinese state media reported three deaths and 265 people missing in Gyirong County, Tibet Autonomous Region. Preliminary expert assessment attributes the trigger to an ice and rock avalanche from a glacier in the upper catchment; the cause remains under investigation, and all casualty figures are provisional as of 26 August 2026 and were still rising at time of release.
29 scenes from three sensors in seven collections, split by the 26 August 2026 event date. Collections are keyed by sensor and acquisition date, so resolutions never mix within one collection.
Pre-event PlanetScope is a single north–south strip (satellite 254a) running from Trishuli
Bazaar (Bidur) north through Betrawati, Dhunche and Syabrubesi to the Rasuwagadhi border
crossing and on into Gyirong County, Tibet. Acquired three months before the flood, at the end
of the dry season.
Post-event PlanetScope, 26 August is two adjacent swaths acquired the morning of the flood
— a western swath (255f) at 05:01 and an eastern swath (251f) at 05:45. Together they span
a wider corridor than the baseline, reaching further west and east toward the Langtang valley.
Post-event SkySat, 27 August is two collects acquired the following morning, about a day after the flood: one over the Rasuwagadhi border crossing, one over Syabrubesi and the upper Trishuli.
Post-event Pelican, 27 August is three sequential frames acquired four hours after the SkySat collects, running north from Syabrubesi to Rasuwagadhi and overlapping the SkySat footprints. At ~0.55 m this is the sharpest imagery in the catalog and shows a bright sediment-laden river channel traceable along the corridor.
Post-event PlanetScope, 28 August is a four-frame strip (2544) down the corridor from
above Rasuwagadhi to Betrawati, plus a fifth frame (2520) offset south-west along the lower
Trishuli past Trishuli Bazaar. It reaches 27.711 N, the southernmost extent in the catalog.
Post-event SkySat, 31 August is two collects over the lower corridor, running south from around Betrawati to the Trishuli Bazaar (Bidur) area. This is the only sub-metre coverage of the downstream reach — the 27 August SkySat and Pelican collects both sit upstream — and it images the largest settlement captured anywhere in this catalog at that resolution.
Post-event Pelican, 1 September re-images 98% of the 27 August Pelican footprint six days after the flood. It is the only same-sensor, same-target sub-metre repeat in the catalog, five days apart, and the basis for change detection at ~0.56 m. Its northern frame over Rasuwagadhi is the clearest Pelican imagery here at 31% clear.
Cloud cover is significant throughout, and reported clear fractions are unreliable.
This is monsoon season in the Nepal Himalaya, so cloud is genuinely heavy. But the reported
clear fractions understate usable ground, in places severely. All three 27 August Pelican
scenes are reported as 0% clear, yet each frame contains substantial interpretable terrain,
including the clearest view of the flood channel in this catalog. Steep relief, deep shadow
and bright cloud appear to defeat the per-pixel classifier. Inspect the imagery or the mask
directly rather than filtering on clear_percent.
PlanetScope analytic products are not all the same processing level. Pre-event carries
analytic_sr (surface reflectance); both post-event PlanetScope collections carry analytic
(top-of-atmosphere radiance). Planet had not published surface-reflectance products for the
post-event scenes when this catalog was assembled. Do not difference pre- and post-event
analytic bands without applying an atmospheric correction first. The visual assets are
directly comparable.
Scene quality varies by collection. 13 of the 14 PlanetScope scenes acquired on 27 May and
26 August are flagged quality_category: test rather than standard in Planet's Data API —
only 20260527_053217_72_254a is standard. By contrast, all five 28 August PlanetScope
scenes, all four SkySat collects and all six Pelican scenes are standard.
Footprints do not fully coincide. The post-event PlanetScope swaths extend further west, east and south than the pre-event strip, so much post-event area has no matching baseline. The SkySat and Pelican collections cover focal points within the corridor, not its full length.
Mask assets differ by sensor. PlanetScope and Pelican carry udm2 (multi-class, with
cloud/shadow/haze/snow and confidence). SkySat carries udm, the older single-band
unusable-data mask. Code consuming masks must handle both.
Resolution spans nearly 7× (0.55–3.9 m). Read gsd per item; do not assume a uniform pixel
size across the catalog.
Each collection also carries items.parquet, a STAC GeoParquet index of its items; each phase
carries an aggregate items.parquet across its collections.
Planet Crisis Response imagery, licensed CC-BY-NC-4.0 — attribution required, non-commercial use only.
Imagery © Planet Labs PBC, made available through the Planet Crisis Response Program.
Part of the Planet disaster data collection on Source Cooperative.
| 2026-08-27 06:10 |
| ~0.55 m |
| Pansharpened |
udm2 |
post-event/planetscope-2026-08-28 | Post-event | PlanetScope | 5 | 2026-08-28 04:57 & 05:01 | ~3.8 m | TOA radiance | udm2 |
post-event/skysat-2026-08-31 | Post-event | SkySat | 2 | 2026-08-31 09:25 | ~0.64 m | Pansharpened | udm |
post-event/pelican-2026-09-01 | Post-event | Pelican | 3 | 2026-09-01 05:06 | ~0.56 m | Pansharpened | udm2 |