The global urbanization rate is accelerating; however, data limitations have far prevented robust estimations of either global urban expansion or its effects on terrestrial net primary productivity (NPP). Here, using a high resolution dataset of global land use/cover (GlobeLand30), we show that global urban areas expanded by an average of 5694 km² per year between 2000 and 2010. The rapid urban expansion in the past decade has in turn reduced global terrestrial NPP, with a net loss of 22.4 Tg Carbon per year (Tg C year⁻¹). Although small compared to total terrestrial NPP and fossil fuel carbon emissions worldwide, the urbanization-induced decrease in NPP offset 30% of the climate-driven increase (73.6 Tg C year⁻¹) over the same period. Our findings highlight the urgent need for global strategies to address urban expansion, enhance natural carbon sinks, and increase agricultural productivity.

Global land-use and land-cover change (LUCC) data are crucial for modeling a wide range of environmental conditions. So far, access to high-resolution LUCC products at a global scale for public use is difficult because of data and technical issues. This article presents a Future Land-Use Simulation (FLUS) system to simulate global LUCC in relation to human–environment interactions, which is built and verified by using remote sensing data. IMAGE has been widely used in environmental studies despite its relatively coarse spatial resolution of 30 arc-min, which is about 55 km at the equator. Recently, an improved model has been developed to simulate global LUCC with a 5-min resolution (about 10 km at the equator). We found that even the 10-km resolution, however, still produced major distortions in land-use patterns, leading urban land areas to be underestimated by 19.77 percent at the global scale and global land change relating to urban growth to be underestimated by 60 to 97 percent, compared with the 1-km resolution model proposed through this article. These distortions occurred because a large percentage of small areas of urban land was merged into other land-use classes. During land-use change simulation, a majority of small urban clusters were also lost using the IMAGE product. Responding to these deficiencies, the 1-km FLUS product developed in this study is able to provide the spatial detail necessary to identify spatial heterogeneous land-use patterns at a global scale. We argue that this new global land-use product has strong potential in radically reducing uncertainty in global environmental modeling.