Socio-ecological predictors of global patterns in human appropriation of net primary production

Abstract

Human impacts on natural systems are often analysed using a statistical model based on the 50-year old IPAT concept, where impact (I) is a function of population (P), affluence (A), and technology (T). Varied results have accrued, but problems remain: ecological predictors are not part of the anthropocentric IPAT concept or statistical model; vastly different countries are often treated as statistical replicates; alternative hypotheses are rarely compared; and most studies evaluated only CO2 emissions as impacts. Here we compare alternative mixed-effect models to predict human appropriation of net primary production (HANPP) as an indicator of global human impacts at relatively fine scale (5 arc-min grid). Predictors represent anthropocentric (P, A and T) and/or ecological (plant biomass and climate) effects, with countries (N = 168) and/or anthropogenic biomes (anthromes; N = 19) as random effects. The most efficient models predict location and amount of HANPP well (R2 = 0.91 and 0.63, respectively) and use all predictors listed above. In both cases, ecological predictors and population have greatest effects on HANPP, consistent with a general, ecological predator–prey relationship modified by socio-ecological conditions. Global human impacts on terrestrial ecosystems (measured as HANPP) depend on both ecological and socioeconomic factors revealed here. Understanding these relationships is a necessary step toward mitigating human impacts on land.