Program: Urban Expansion
Solly Angel in The Atlantic
The New Urban Peripheries, 1990–2014:
This chapter was written by: Shlomo Angel, Yang Liu, Alex M. Blei, Patrick Lamson-Hall, Nicolás Galarza Sanchez, and Sara Arango-Franco
Introduction:
Cities grow in population and wealth; as they grow, they both densify and expand. As cities expand, they convert new areas on their rural peripheries to urban use and create new urban peripheries in the process, newly settled areas or newly annexed villages and towns that were for-merly on the outskirts of cities and are now contiguous with or engulfed by the city’s expanding urban footprint. These new urban peripheries are the focus of this chapter. Previous research, supported by a rigor-ous theoretical framework, has established an important empirical regularity that characterizes new urban peripheries: They typically have lower average densities. More generally, it has been shown in numerous studies that population densities decline as distance from the city centre increases and, hence, we can expect that as cities expand outwards and away from their older centres, newly built areas will be further away from areas built earlier and will thus have lower average densities as well. Unfortunately, beyond this important finding, very little is known about urban peripheries and particularly about the new urban peripheries, areas built or incorporated into the urban footprint in recent decades. In this chapter, which is discursive and non-technical in nature, we report on a number of recent findings that characterize the new urban peripheries the world over– in our case, areas added to cities between 1990 and 2014. The report is based on our analysis of sat-ellite imagery and census data in a stratified global sample of 200 cities, a 4.7 per cent sample of the universe of all 4,231 cities and metropolitan areas that had 100,000 people or more in 2010. We also report on a few additional findings from a pilot study now underway in ten cities in ten different world subregions.
Intra-urban land use maps for a global sample
This article was written by Shlomo Angel, Patrick Lamson-Hall, Alejandro M. Blei, Eric Mackres and Brookie Guzder-Williams
Rationale:
Intra-urban land use maps provide important information to urban planners and policymakers, but these maps are costly, time consuming to generate and are often unavailable in developing countries where most urban growth is now occurring. This paper reports on machine learning methods to automate the production of land use maps from cloud-free mosaics of Sentinel-2 imagery. We have trained a novel neural network architecture to produced 5 meter resolution land use maps for a global stratified sample of 200 cities. The sample includes all world regions, 78 countries, and a range of population sizes. The model architecture is roughly 1 to 2 orders of magnitude smaller than similar architectures such as UNet (Ronneberger, Fischer, & Brox, 2015) and DeeplabV3+ (Chen, Zhu, Papandreou, Schroff, & Adam, 2018), significantly lowering the cost and computational requirements to produce maps. We are in the process of generating land use maps for all 4,000 + cities and metropolitan areas in the world with populations exceeding 100,000. The resulting product will be the first, regularly updated, freely available, global intra-urban land use maps at 5 meter resolution.
We present a 4-tier land use taxonomy which at its root distinguishes open-space from built-up area. At the second tier, it subdivides the built-up category into nonresidential and residential areas. The third tier distinguishes formal from informal residential land use, and the fourth tier further subdivides formal and informal residential land uses into more detailed categories. Accuracy scores at tier-1 and tier-2 were 86% and 79% respectively. Tiers 3 and 4 had an accuracy scores of 75% and 71% respectively. Additionally, we train a roads-only model and compare its output to the Atlas of Urban Expansion’s Arterial Roads dataset and Open Street Map. As an example use case, we train an Informal Settlement Classifier, correctly classifying 87% of the settlements.