Earth System Science Data (Sep 2020)

Earth transformed: detailed mapping of global human modification from 1990 to 2017

  • D. M. Theobald,
  • D. M. Theobald,
  • C. Kennedy,
  • B. Chen,
  • J. Oakleaf,
  • S. Baruch-Mordo,
  • J. Kiesecker

Journal volume & issue
Vol. 12
pp. 1953 – 1972


Read online

Data on the extent, patterns, and trends of human land use are critically important to support global and national priorities for conservation and sustainable development. To inform these issues, we created a series of detailed global datasets for 1990, 2000, 2010, and 2015 to evaluate temporal and spatial trends of land use modification of terrestrial lands (excluding Antarctica). We found that the expansion of and increase in human modification between 1990 and 2015 resulted in 1.6 M km2 of natural land lost. The percent change between 1990 and 2015 was 15.2 % or 0.6 % annually – about 178 km2 daily or over 12 ha min−1. Worrisomely, we found that the global rate of loss has increased over the past 25 years. The greatest loss of natural lands from 1990 to 2015 occurred in Oceania, Asia, and Europe, and the biomes with the greatest loss were mangroves, tropical and subtropical moist broadleaf forests, and tropical and subtropical dry broadleaf forests. We also created a contemporary (∼2017) estimate of human modification that included additional stressors and found that globally 14.6 % or 18.5 M km2 (±0.0013) of lands have been modified – an area greater than Russia. Our novel datasets are detailed (0.09 km2 resolution), temporal (1990–2015), recent (∼2017), comprehensive (11 change stressors, 14 current), robust (using an established framework and incorporating classification errors and parameter uncertainty), and strongly validated. We believe these datasets support an improved understanding of the profound transformation wrought by human activities and provide foundational data on the amount, patterns, and rates of landscape change to inform planning and decision-making for environmental mitigation, protection, and restoration. The datasets generated from this work are available at (Theobald et al., 2020).