Discussion on urban areas boundaries

[pzwsk]

This issue is to discuss the value add of the use of the GHSL Urban Center as an alternative boundary for urban areas.

ThinkHazard traditionally used the ADM boundaries and names provided by GAUL.

[matamadio]

Issue objective unclear:

• GAUL is discontinued and no longer the base for WB ADM layer > new official ADM layer in use
• There is no ADM boundary set covering urban areas
• GHSL is currently the only consistent dataset describing functional urban extents

A few alternative datasets that come to mind to describe urban/built-up footprint:

• WSF provides built-up footprint at 10m (tif), but does not define the continuos extent of an urban area
• GUF (Global Urban Footprint) from DLR is focused on individual urban areas, but again raster tiles and only covering 2010-2013.

Where GHSL is genuinely unique is in combining:

• A consistent built-up footprint with a functional urban classification (DEGURBA — distinguishing urban centres, urban clusters, rural areas)
• Integration with population data in the same framework
• UN endorsement of its urban definitions for official statistics

So if "urban footprint" means a functional definition of urban areas (i.e., accounting for density thresholds, contiguity, and population), then GHSL/DEGURBA is largely in a class of its own for global, consistent, policy-relevant urban delineation.

[matamadio]

Keeping open looking for refinement of urban areas / adding missing ones.