It is widely accepted that the main driver of the observed decline in biological diversity is increasing human pressure on Earth’s ecosystems. However, the spatial patterns of change in human pressure and their relation to conservation efforts are less well known. We developed a spatially and temporally explicit map of global change in human pressure over two decades between 1990 and 2010 at a resolution of 10 km².  We evaluated 22 spatial data sets representing different components of human pressure  and used them to compile a Temporal Human Pressure Index (THPI) based on 3 data sets: human population density, land transformation, and electrical power infrastructure. We investigated how the THPI within protected areas correlate to International Union for Conservation of Nature (IUCN) management categories and the Human Development Index (HDI), as well as how the THPI was correlated to accumulative pressure using the original Human footprint. Since the early 90’s, human pressure increased 64% in terrestrial areas; the largest increases were in Southeast Asia. Protected areas also exhibited overall increases in human pressure, the degree of which varied with location and IUCN management category. Only wilderness areas and natural monuments (management categories Ib and III) exhibited decreases in pressure. Protected areas not assigned any category exhibited the greatest increases. High HDI values and greater mean elevation correlated with greater reductions in pressure across protected areas, while increasing age of the protected area correlated with increases in pressure. Our analysis is an initial step toward mapping changes in human pressure on the natural world over time. That only 3 data sets could be included in our spatio-temporal global pressure map, highlights the challenge to measuring pressure changes over time.

Data from 1) The inter-calibrated stable night lights  version 4, 2) The Gridded Population of the World (GPW) version 3, and 3) The History Database of the Global Environment (HYDE) 3.1 were spatially aggregated to a resolution of 5.0 arc minutes (approximately 10km² at Equator), the original resolution of the HYDE 3.1 cropland data. This aggregation caused some loss of resolution for the other 2 data sets (approximately 2.8 km² for stable nightlights and 5 km² for human population density). For each terrestrial pixel, we calculated the difference between values in the first and last year. This was done separately for the 3 layers. We transformed human population density to the square root. Data transformation of variables is a standard procedure for spatial pressure mapping, and it allows comparison between different data types and distributions . We chose square-root transformation because it accounted for the expected declining impacts per person in densely populated areas but still had a range distribution similar to the original data (Supporting Information). The result was 3 maps displaying the change in absolute values of human population density, stable nightlights, and land use respectively.

These maps had very different data ranges (human population, – 8,532 to 11,423 people per pixel; stable nightlights, -62 to 63 on an arbitrary scale; cropland, -86 -70% change). To account for these inherent differences, values for each layer were standardized on a scale of -1 to 1, which allowed us to summarize these 3 different components of pressure. We used the same weighting as the original human footprint (Sanderson et al. 2002), giving equal weight to stable nightlights and human population while weighting land-use change at 0.8 (for justification see Supporting Information). Finally, we combined the 3 layers by adding the values within each pixel and standardizing the resulting score on a scale of -100 to 100, where positive values mean increased human pressure and negative values mean decreased human pressure. This final product forms our temporal human pressure index (THPI), which measures changes in human pressure for the 3 data sets over 15 years from 1990 to 2010.