Nature: Floods and droughts cause perhaps the most human suffering of all climate-related events; a major goal is to understand how humans alter the incidence and severity of these events by changing the terrestrial water cycle. Here we use over 1,500 estimates of annual evapo-transpiration and a database of global land-cover change¹ to project alterations of global scale terrestrial evapo-transpiration (TET) from current anthropogenic land-cover change. Geographic modelling reveals that land-cover change reduces annual TET by approximately 3,500 km³ yr⁻¹ (5%) and that the largest changes in evapo-transpiration are associated with wetlands and reservoirs. Land surface model simulations support these evapo-transpiration changes, and project increased runoff (7.6%) as a result of land-cover changes. Next we create a synthesis of the major anthropogenic impacts on annual runoff and find that the net result is an increase in annual runoff, although this is uncertain. The results demonstrate that land-cover change alters annual global runoff to a similar or greater extent than other major drivers, affirming the important role of land-cover change in the Earth System^{2, 3, 4}. Last, we identify which major anthropogenic drivers to runoff change have a mean global change statistic that masks large regional increases and decreases: land-cover change, changes in meteorological forcing, and direct CO₂ effects on plants.

Main

Humans have altered a large proportion, approximately 41% (ref. 1), of the Earth’s surface (Fig. 1a), replacing natural vegetation such as forests and wetlands with anthropogenic land cover (ALC) such as croplands and built-up land. Of all ALCs, grazing land covers the largest area, taking place on roughly 1/4 of the land surface. Land-cover change alters the water cycle through direct changes to the timing and magnitude of evapotranspiration (ET), because land-cover change alters available energy, available water, photosynthesis rates, nutrient levels and surface roughness at the land surface.

Here the GIS method is used, unless otherwise noted. a, Distribution of present extent of ALC change, totalling 41% of land surface. ALC is expressed as a percentage of cell at a 5 min resolution. Land area is set to 1.34×10¹⁴ m², to represent land areas not covered with permanent ice. b, Distribution of potential vegetation ET (m yr⁻¹) at 5 min resolution, totalling to 66,000 km³ yr⁻¹ and within the range published values (63,000–85,000 km³ yr⁻¹; Supplementary Table S1, (S1.19)). To create the map of potential ET, we added the ET layers for each PLC in a GIS. c, Estimate of current levels of appropriated ET (m yr⁻¹), the volume of changed ET in each cell that incorporates the area of land-cover change in each cell, which has been converted to m yr⁻¹ through division by cell area, at a 5 min resolution. ALCs here are not censored. Appropriated ET here is 41% of TET. d, Estimate of percentage change in TET, ((anthropogenic TET–potential TET)/potential TET), calculated at 5 min, then plotted at 30 min resolution, for visibility purposes. e, Estimate of percentage change in TET, from ORCHIDEE LSM simulation of ((current TET–potential TET)/potential TET), at a 1° resolution.

There are two key indicators of human impact on ET at the global scale: change and appropriation⁵. Here appropriation is defined as a measure of the amount of the flux, or mass, touched (but not necessarily quantitatively altered) by human activity. The most commonly cited statistic of human impact on the global water cycle⁶, that humans appropriate 23% of total TET was calculated through a conversion of appropriated net primary productivity (HANPP) to TET by a scale relation with total terrestrial net primary productivity (TNPP): (HANPP×(TET/TNPP))/(TET)); this estimate, however, is in error because TET cancels out of this equation.

Until recently, robust studies on how land cover changes TET have been scarce. Many studies^{7, 8, 9, 10, 11, 12} have examined the effect of one or two types of human-dominated land cover, and, as a result, underestimate the area of land surface that humans have altered and the types of land-cover changes that impact ET (refs 1, 13, 14, 15).

Here, we use a novel approach to estimate appropriated and changed TET, using a database of over 1,500 ET observations for discrete land-cover types. The data are analysed in a Geographic Information System (GIS) at 5 min resolution, facilitating analysis of the impact of ET by fine-scale individual land-cover changes. The changed TET is then validated against an independent method of estimating change in TET, using a land surface model (LSM) simulation. For all estimates of ET change and appropriation, we use the same land-cover change database that includes the major suite of ALC changes¹. We calculate appropriation of TET using a raster analysis of land-cover maps and fields of ET for discrete land covers (Fig. 1b,c). GIS results indicate that humans are appropriating 41% of TET, a level approximately twice as large as the original estimate⁶. It is also larger than other recent estimates¹⁶, which use a different definition of HANPP. Although the exact correlation between percentage occupied land and percentage appropriation of ET is probably fortuitous, the numbers do reflect that humans tend to occupy areas of higher ET and avoid areas of lower ET (Fig. 1b), such as polar and desert areas.

We next estimate how humans have changed current TET by comparing ET in ALC with ET in potential land cover (PLC), defined as the land cover that would exist in the present climate, without anthropogenic alteration. Results indicate that the current extent human land-cover change has reduced global annual average TET by ~ 5% (~3,500 km³ yr⁻¹; Table 1 and Fig. 1d), coherent with other recent estimates^{11, 12, 14}, and approximately equivalent in volume to annual global water withdrawals (~3,200 km³ yr⁻¹; Fig. 2). Of all ALC change activities, conversion of PLCs to non-irrigated cropland causes the largest total volume change in TET, reducing the TET volume by approximately 2,300 km³ yr⁻¹, and thus accounting for over half the volume in the TET reduction of 5%, even though the area affected is much less than grazing¹ (Fig. 3b).

Source and Full article>>