. 2017. Agriculture production as a major driver of the Earth system exceeding planetary boundaries. ABSTRACT. We explore the role of agriculture in destabilizing the Earth system at the planetary scale, through examining nine planetary boundaries, or "safe limits": land-system change, freshwater use, biogeochemical flows, biosphere integrity, climate change, ocean acidification, stratospheric ozone depletion, atmospheric aerosol loading, and introduction of novel entities. Two planetary boundaries have been fully transgressed, i.e., are at high risk, biosphere integrity and biogeochemical flows, and agriculture has been the major driver of the transgression. Three are in a zone of uncertainty i.e., at increasing risk, with agriculture the major driver of two of those, landsystem change and freshwater use, and a significant contributor to the third, climate change. Agriculture is also a significant or major contributor to change for many of those planetary boundaries still in the safe zone. To reduce the role of agriculture in transgressing planetary boundaries, many interventions will be needed, including those in broader food systems.

Hydrological change is a central part of global change 1-3 . Its drivers in the past need to be understood and quantified for accurate projection of disruptive future changes 4 . Here we analyse past hydro-climatic, agricultural and hydropower changes from twentieth century data for nine major Swedish drainage basins, and synthesize and compare these results with other regional 5-7 and global 2 assessments of hydrological change by irrigation and deforestation. Cross-regional comparison shows similar increases of evapotranspiration by non-irrigated agriculture and hydropower as for irrigated agriculture. In the Swedish basins, non-irrigated agriculture has also increased, whereas hydropower has decreased temporal runoff variability. A global indication of the regional results is a net total increase of evapotranspiration that is larger than a proposed associated planetary boundary 8 . This emphasizes the need for climate and Earth system models to account for different human uses of water as anthropogenic drivers of hydro-climatic change. The present study shows how these drivers and their effects can be distinguished and quantified for hydrological basins on different scales and in different world regions. This should encourage further exploration of greater basin variety for better understanding of anthropogenic hydro-climatic change.Drivers of freshwater changes are multiple and difficult to distinguish and quantify 9-12 . Global increase of evapotranspiration (ET) by irrigation and a parallel decrease by deforestation has been estimated by spatial compilation of crop and forest data 2 . Furthermore, spatial analysis of relatively short-term (30-year) data has shown effects of large dams and their reservoirs on atmospheric variables of convective available potential energy, specific humidity and surface evaporation extending over distances of around 100 km away from reservoir shorelines 13 . Such spatial results indicate ET changes by different land and water uses, but do not reveal the actual change dynamics and timing. Moreover, global averaging hides the spatial variation of ET and its regional change drivers and further effects on water change (through changed water loss by ET to the atmosphere 5,7,14 ) and climate change (through changed latent heat flux 6,15 and atmospheric circulation 16 ).Hydrological model interpretation of consistently long, twentieth century time series of relevant hydro-climatic and landwater use data has distinguished between climate and irrigation drivers of ET change in the drainage basins of the Aral Sea in Central Asia 5,6 and Mahanadi River in India 7 (Fig. 1a). Both of these basins represent areas with major irrigation developments that have driven increases of ET and associated water loss to the atmosphere. Direct temporal analysis of long-term hydro-climatic and land-water use data, accounting for the water balance constraints of hydrological basins, can thus complement spatial studies, particularly for the twentieth century, which has seen large land-water use changes, wh...

Flow regulation and irrigation alter local freshwater conditions, but their global effects are highly uncertain. We investigated these global effects from 1901 to 2008, using hydroclimatic observations in 100 large hydrological basins. Globally, we find consistent and dominant effects of increasing relative evapotranspiration from both activities, and decreasing temporal runoff variability from flow regulation. The evapotranspiration effect increases the long-term average human consumption of fresh water by 3563 ± 979 km(3)/year from 1901-1954 to 1955-2008. This increase raises a recent estimate of the current global water footprint of humanity by around 18%, to 10,688 ± 979 km(3)/year. The results highlight the global impact of local water-use activities and call for their relevant account in Earth system modeling.