Read elevation and quality together

AW3D30 is a digital surface model derived from original ALOS PRISM observations, with approximately thirty-metre horizontal spacing that depends on latitude. “Surface” matters: a surface-height product should not automatically be treated as bare-ground terrain everywhere. Grid spacing describes horizontal sampling, not a promise of equally precise vertical heights.

Imagine an original planning exercise in which someone wants to identify a one-metre step beside a doorway from a thirty-metre grid. The proposed feature is much smaller than the grid spacing. Enlarging the displayed pixels makes them easier to see but does not add the missing measured detail. A different survey is needed for that particular question.

The AW3D30 distribution includes height, mask, source-stack and quality information. Masks identify conditions such as cloud, snow and complementary elevation sources used to fill gaps. The product’s regions can have different versions, and later processing can revise values. A filled cell is not necessarily a new observation made by ALOS at the time of publication.

For a fictional comparison, one cell is supported by several source scenes while a nearby cell was filled from another dataset. That difference does not prove either height is wrong, but it changes the provenance that an analyst should inspect. Keep the quality files with the height grid instead of discarding them as administrative clutter.

Keep the evidence beside the elevation grid

A disaster-response map needs a clear account of what was observed, when it was observed and how it was interpreted. An older elevation model can provide background context without establishing the latest state of a damaged road. A recent land observation can show a change without determining whether a structure is safe to enter.

In an original briefing exercise, a team has a pre-event surface model and a recent radar map with an unresolved quality flag near a bridge. A responsible summary identifies the dates and flag and requests corroborating information for the bridge. It does not turn incomplete map evidence into a confident access instruction. The useful output is a traceable statement of findings and limits.