Global urban growth between 1870 and 2100 from integrated high resolution mapped data and urban dynamic modeling

Xuecao Li,Gokul Iyer,Mohamad Hejazi,Chris Vernon,Marshall Wise,Wei Chen,Yuyu Zhou

doi:10.1038/s43247-021-00273-w

Abstract

Long term, global records of urban extent can help evaluate environmental impacts of anthropogenic activities. Remotely sensed observations can provide insights into historical urban dynamics, but only during the satellite era. Here, we develop a 1 km resolution global dataset of annual urban dynamics between 1870 and 2100 using an urban cellular automata model trained on satellite observations of urban extent between 1992 and 2013. Hindcast (1870–1990) and projected (2020–2100) urban dynamics under the five Shared Socioeconomic Pathways (SSPs) were modeled. We find that global urban growth under SSP5, the fossil-fuelled development scenario, was largest with a greater than 40-fold increase in urban extent since 1870. The high resolution dataset captures grid level urban sprawl over 200 years, which can provide insights into the urbanization life cycle of cities and help assess long-term environmental impacts of urbanization and human–environment interactions at a global scale.

Highlights

Long term, global records of urban extent can help evaluate environmental impacts of anthropogenic activities
The presented spatiotemporal dynamics of global urban extent are consistent with the findings reported in the “World Urbanization Prospects”[46]
It is worth to note that the harmonization between History Database of the Global Environment (HYDE) and nighttime light (NTL) derived urban extents was conducted based on observations in 1992, which led to a slightly abrupt change around 1990 in the overall trend

Summary

Introduction

Global records of urban extent can help evaluate environmental impacts of anthropogenic activities. The high resolution dataset captures grid level urban sprawl over 200 years, which can provide insights into the urbanization life cycle of cities and help assess long-term environmental impacts of urbanization and human–environment interactions at a global scale. Panel data analysis is a commonly used approach for estimating urban demand, which essentially characterizes a linear relationship between per capital urban area and socioeconomic variables (e.g., per capital GDP and the urbanization rate) of all spatial units[26] Such a relationship could be too simple to capture discrepancies of urban demand in different urbanization stages and differences of urban sprawl pathways in different regions[42,43,44]. We combined urban extent from observations, hindcast, and projection and generated the long-term dataset of urban extent from 1870 to 2100

Methods

Results

Conclusion