Large disagreements in estimates of urban land across scales and their implications

Abstract

Improvements in high-resolution satellite remote sensing and computational advancements have sped up the development of global datasets that delineate urban land, crucial for understanding climate risks in our increasingly urbanizing world. Here, we analyze urban land cover patterns across spatiotemporal scales from several such current-generation products. While all the datasets show a rapidly urbanizing world, with global urban land nearly tripling between 1985 and 2015, there are substantial discrepancies in urban land area estimates among the products influenced by scale, differing urban definitions, and methodologies. We discuss the implications of these discrepancies for several use cases, including for monitoring urban climate hazards and for modeling urbanization-induced impacts on weather and climate from regional to global scales. Our results demonstrate the importance of choosing fit-for-purpose datasets for examining specific aspects of historical, present, and future urbanization with implications for sustainable development, resource allocation, and quantification of climate impacts.

Fig. 1. Country level urban land and its variability across datasets.

a shows mean urban percentage based on eight global estimates of urban land by country. b shows the overall urban percentage from these eight datasets for the ten countries with the highest mean urban area (increasing downwards). c shows the coefficient of variation (standard deviation divided by mean) among those urban estimates. For (a, c), the respective values are annotated for some of the bigger countries for context. The legend value ranges exclude the upper bound.

Fig. 2. Present-day urban estimates across datasets from regional to grid scale.

a Shows the urban percentage across eight datasets for four selected regions in the world. The extent and location of these regions is shown in the inset. b Shows the coefficient of variation (standard deviation divided by mean) among those urban estimates for 0.9° × 1.25° grids. The legend value ranges exclude the upper bound.

Fig. 3. Present-day urban extent across datasets for select cases.

The spatial extents of urban pixels for the eight datasets for a the Delhi Metropolitan Area in India and b the Shanghai Metropolitan Area in China. Latitudes and longitudes are shown in the corners.

Fig. 4. Urban percentage and its long-term changes across datasets.

Urban percentage from 12 global data products for a World, b Africa, c Asia, d Europe, e North America, f Oceania, and g South America. Long-term changes are shown for datasets that span multiple successive years.

Fig. 5. Country level urban percentage and its long-term changes across datasets.

Urban percentage from 12 global data products for a China, b United States, c India, d Russia, e Brazil, f Japan, g Germany, h Indonesia, i France, and j Mexico. Long-term changes are shown for datasets that span multiple successive years.

Fig. 6. Impacts of variability in urban estimates on observational assessments.

a Shows the absolute coefficient of variation in calculated surface urban heat island intensity during 2018–2022 summer for around 10,000 global urban clusters from eight urban land cover datasets. b Shows estimated change in urban land in flood plains for the world and all continents between 1985 and 2015 from three long-term urban estimates.

Fig. 7. Differences with model representations of urbanization.

a Shows global urban percentage across eight urban land cover datasets as well as two additional estimates of urban areas used in weather and climate models over Europe. LCZ 1 to 10 refer to the ten built up local climate zones. b Shows linear regressions between grid-wise urban percentage in the GISA dataset for the year 2001 versus the total urban percentage from medium density, high density, and tall building districts classes of the Jackson et al. (2010) dataset for the world and each continent. The line of best fit, coefficient of determination (r²), mean bias error (MBE), mean percentage error (MPE), and sample size (n) are provided for each case.

Fig. 8. Future projections of urban land from various datasets.

Projected percentage of global urban area from 2020 to 2100 for various Shared Socioeconomic Pathways (SSPs) based on multiple km-scale estimates for a World, b Africa, c Asia, d Europe, e North America, f Oceania, and g South America.