This video demonstrates how to derive the urban footprint from a land-cover raster in QGIS. First, we download the municipality polygon from OpenStreetMap with the QuickOSM plugin. Then, we create a buffer to accommodate for urban expansion beyond the municipality boundary. Next, we download the ESA WorldCover raster from TerraScope. After reclassifying, we apply a circular neighbourhood to a Boolean raster with built-up pixels using r.neigbors from GRASS. This results in the urban footprint.
Thanks to colleague Dr. William Veerbeek to figure out this method.
Reference to the method: Angel, S., Sheppard, S., Civco, D. L., Buckley, R., Chabaeva, A., Gitlin, L., ... & Perlin, M. (2005). The dynamics of global urban expansion (p. 205). Washington, DC: World Bank, Transport and Urban Development Department.
Explanation of parameters:
To have a circular neighbourhood of 1 km², we need a diameter of 1.128 km. The pixel size of WorldCover is 10 m, which means 112.8 pixels. Rounded to an odd number that is 113 pixels. If more than 10% built-up is considered urban, then the threshold of the sum of built-up in the circle is 985.
Open data: (c) OpenStreetMap Contributors, ESA WorldCover
Plugins: QuickOSM
QGIS version: 3.40
TerraScope: https://terrascope.be/en
0:00 Introduction
0:54 Download municipality boundary from OpenStreetMap with QuickOSM
2:38 Download ESA WorldCover raster from TerraScope
5:04 Reproject land-cover raster
6:14 Expand muncipilaty boundary with a buffer
6:50 Clip land-cover with buffer polygon
7:40 Create a boolean raster with built-up vs other areas
9:00 Neighborhood analysis with circular window using r.neighbors
10:21 Determine threshold value with histogram
11:05 Use raster calculator to calculate urban footprint
#QGIS #urbanplanning