General overview of the methodology

This study aims to quantify terrestrial carbon losses from global impervious surface area (ISA) expansion (referred to as ISA-driven emissions) over 1993–2018. Four state-of-the-art global remote-sensing products of impervious surface or urban area were used. Our analysis focuses on the dynamics of impervious surface area and assumes 100% coverage of ISA for the identified ISA (or urban) pixels in the ISA (or urban) products (hereafter referred to as ISA products) at their original spatial resolutions by ignoring any potential sub-pixel, urban green areas (e.g., street trees, residential gardens, etc., see Supplementary Information ¹ for justification of our decision). The four ISA products were used to derive individual ISA expansion rates, whose average value was then obtained and examined. ESA CCI land cover products were used to obtain source land cover information for global ISA expansion on an annual time scale.

Carbon density maps of various carbon pools, consisting of living biomass (both above- and below-ground), surface litter, dead wood, and soil organic carbon (SOC), were then overlaid with the maps of ISA expansion derived from the different ISA products to quantify terrestrial carbon losses. Immediate carbon losses on conversion of vegetated lands to ISA were assumed, and hence the quantified carbon losses were committed emissions. The estimated ISA-driven emissions were further compared with carbon losses due to land conversion to settlements as reported in national greenhouse gas inventories (NGHGIs) for Annex I (AI) countries of the United Nations Framework Convention on Climate Change (UNFCCC). Finally, the temporal relative change rates in ISA-driven carbon emissions were attributed to different driving factors, with contrasting patterns of emissions dynamics being identified for AI and NAI (i.e., non-Annex I) country groups, respectively.