Sensors
- Maxar WorldView Legion: Panchromatic resolution of approximately 29 cm and multispectral at around 1.2 m. Up to 15 revisits per day over a target at mid-latitudes when the full constellation is operational, making it the fastest very-high-resolution optical option for post-event tasking.
- Airbus Pléiades Neo: 30 cm panchromatic, 1.2 m multispectral, four-satellite constellation offering same-day stereo acquisition. Useful for 3-D roof-collapse detection and for generating post-event digital surface models to compare against pre-event baselines.
- Sentinel-1 IW (SAR, C-band): 10 m ground range resolution in Interferometric Wide swath mode, 250 km swath, 6-day repeat at the equator with both satellites active. Penetrates cloud cover that blocks optical collection entirely, making it the workhorse sensor for the first 24-72 hours after a storm.
- Planet SkySat: 50 cm panchromatic, 0.9 m multispectral. A fleet of around 20 satellites allows rapid tasking, though individual daily revisit per site is lower than WorldView Legion. Useful for wide-area screening before committing to more expensive tasking.
What the damage grades actually mean
UNOSAT and the Copernicus Emergency Management Service both use a four-point scale derived from the UN-SPIDER and FEMA methodologies: destroyed (structure collapsed or burnt-out shell), severely damaged (major structural failure, roof gone, walls partially standing), moderately damaged (partial roof or wall loss, structure standing), and no visible damage. The grading is applied per building footprint, not per block or neighbourhood.
The critical word is 'visible'. Both UNOSAT and Copernicus EMS analysts are reading roof condition from nadir or near-nadir imagery. A building whose walls have sheared but whose roof slab remains largely intact will be under-graded. Surge damage to ground-floor interiors is invisible from above. The damage map is a triage instrument, not a structural engineer's report, and experienced humanitarian buyers treat it as such.
The resolution floor below which grading breaks down
Published UNOSAT validation work, and the Copernicus EMS technical specifications, converge on a practical resolution floor of around 0.5 m for reliable per-building grading in dense urban fabric. At 1 m resolution you can detect destroyed structures with reasonable confidence but the moderate-damage class becomes ambiguous because the roof-texture cues that signal partial loss are blurred. At 3 m, which is the native resolution of some commercial multispectral products, per-building grading is not credible; you are mapping neighbourhood-level damage proxies at best.
This matters for sensor selection. Sentinel-2 at 10 m is useful for burn-scar and flood extent on sibling pages in this library, but it contributes nothing to per-building cyclone damage grading. Sentinel-1 at 10 m is similarly too coarse for building-level optical grading, though it serves a different and important role in the workflow described below.
SAR intensity change does the work when clouds do not move
Tropical cyclones arrive with cloud decks that often persist for 48 to 96 hours after the storm passes. Optical sensors are useless in that window. Sentinel-1 C-band SAR, and commercial X-band systems such as ICEYE or Capella, image through cloud and rain. The method is intensity change detection: a pre-event SAR image is compared pixel-by-pixel with a post-event acquisition, and areas of significant backscatter change are flagged as candidate damage.
The physics behind it is straightforward. Intact buildings produce strong, repeatable double-bounce returns to the radar. Collapsed structures scatter energy in irregular directions, reducing the return. The change is detectable at Sentinel-1's 10 m resolution for clusters of collapsed buildings, but isolated single-storey structures may fall below the detection threshold. Analysts typically use SAR change maps to prioritise where to task optical sensors once cloud clears, rather than as a standalone damage product.
Speed is the variable that determines operational value
The Copernicus EMS activation workflow targets a first rapid damage assessment within 24 hours of a request. UNOSAT's crisis mapping operations aim for similar timelines. In practice, the latency is dominated by two factors: cloud clearance over the affected area, and the time required for a commercial satellite to be repositioned and tasked if no suitable pre-event archive image exists over the target.
Pre-event archive depth matters enormously. Maxar's WorldView archive and Airbus's Pléiades archive both hold imagery over most populated coastal areas, but the vintage of the most recent clear pre-event image varies widely. A pre-event image from 18 months before the storm, in a rapidly developing coastal town, introduces change-detection noise that is not storm damage. Analysts must assess archive recency before committing to a change-detection approach.
The International Charter on Space and Major Disasters (covered separately in this library) is the primary mechanism by which affected governments access priority tasking from multiple operators simultaneously. Activation is free for qualifying disasters and typically mobilises Sentinel-1, Pléiades, and WorldView assets within hours of a formal request.
Honest limits of the combined optical-SAR approach
Three limits are worth naming plainly. First, informal and self-built housing stock, which is often the most vulnerable, is the hardest to grade reliably. Corrugated-iron roofs on timber frames produce ambiguous spectral signatures even at 30 cm resolution; a roof that has been lifted and replaced at a slight angle may read as undamaged.
Second, wind damage and surge damage are not separable from optical imagery alone. A building that was structurally sound but inundated by a 4-metre surge will appear intact from above. Integrating storm-surge inundation layers (covered in the SAR coastal inundation page in this library) with the damage map is necessary to flag probable interior damage in the surge zone.
Third, damage maps degrade in accuracy at the urban periphery where pre-event building footprint databases are incomplete or absent. Without a reliable footprint layer, analysts must digitise buildings from the pre-event image, which adds hours and introduces its own errors.
Putting the workflow together for a government buyer
A practical post-cyclone workflow runs in two phases. Phase one, starting within hours of the storm: Sentinel-1 IW acquisitions are pulled from the Copernicus Dataspace, pre- and post-event images are co-registered and differenced, and a SAR intensity change layer is delivered as a GIS file to the national disaster management authority. This layer is coarse but cloud-proof and fast.
Phase two begins when optical windows open. Commercial tasking is placed over the highest-priority areas identified in phase one. Pléiades Neo or WorldView Legion imagery is orthorectified, building footprints are matched against the pre-event database, and analysts apply the UNOSAT four-point grading scheme building by building. The output is a vector layer with damage grades per footprint, exportable to standard GIS formats and compatible with OCHA's humanitarian data standards.
Satellize runs this two-phase workflow on open Sentinel-1 data combined with commercial optical tasking placed on client licence. The Tonga crop-estimation programme gave us direct experience of the cloud and archive challenges that characterise small-island Pacific operations, which is where some of the world's most exposed cyclone-vulnerable populations live. Government buyers who want to discuss a standing activation agreement before the next storm season can request a scoping call with our disaster-response team.
Typical figures
| Optical resolution (VHR tasking) | 29-50 cm panchromatic (WorldView Legion, Pléiades Neo, SkySat) |
| SAR resolution (Sentinel-1 IW) | 10 m ground range, 250 km swath |
| Minimum building size for reliable grading | ~50 m² footprint at 30 cm optical resolution; larger at coarser resolutions |
| Sentinel-1 revisit | 6 days at equator (both satellites); some regions 3-4 days due to orbit overlap |
| VHR optical revisit | Up to 15 times/day (WorldView Legion full constellation); 1-2 times/day typical for Pléiades Neo |
| Phase-one SAR product latency | 6-24 hours post-acquisition, subject to data downlink and processing queue |
| Phase-two optical product latency | 24-72 hours after cloud clearance, depending on tasking queue and analyst capacity |
| Damage grade classes | Destroyed / Severely damaged / Moderately damaged / No visible damage (UNOSAT/Copernicus EMS standard) |
| Archive depth (commercial optical) | Maxar archive from ~2008; Pléiades from 2012; recency of clear pre-event image varies by location |
| Delivery formats | GeoPackage, Shapefile, GeoTIFF, KMZ; OCHA-compatible attribute schema on request |
Analytics Satellize can run
| SAR intensity change layer | Pre/post-event backscatter differencing on co-registered Sentinel-1 IW GRD scenes | GeoTIFF and vector hotspot layer, delivered within hours of post-event acquisition |
| Per-building damage grade map | UNOSAT four-point visual interpretation protocol applied to VHR optical imagery against pre-event footprint database | Vector polygon layer with damage-grade attribute per building footprint, GeoPackage or Shapefile |
| Post-event digital surface model | Stereo photogrammetry from Pléiades Neo along-track stereo pairs; differenced against pre-event DSM where available | 1 m DSM GeoTIFF with height-change layer highlighting collapsed volumes |
| Neighbourhood-level damage density summary | Spatial aggregation of per-building grades to administrative unit boundaries; proportional damage index calculated | Tabular report and choropleth GIS layer by admin-2 or admin-3 unit, compatible with OCHA reporting templates |
| Surge-zone damage flag | Intersection of per-building damage layer with SAR-derived inundation extent to flag probable interior damage in structurally intact buildings | Annotated vector layer with surge-exposure attribute added to damage grade records |
| Damage progression monitoring | Repeat VHR optical tasking at 3-day and 7-day intervals to track demolition, debris clearance and early reconstruction | Time-series GIS layers with change flags; summary statistics for situation reports |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.