Large scale expansion of crop production for bioenergy would lead to a large increase in the amount of transpired water that is used for human purposes, perhaps equaling the present amount used by the end of the twenty-first century [29]. Globally, commercial bioenergy production is projected to consume 18-46% of the current agricultural use of water by the year 2050 [29]. Increased bioenergy production will strain water resources in all continents but the challenges will be most pronounced in Asia and Africa [30]. By 2075, water use for the production of bioenergy could push some important agricultural countries, including the United States and Argentina, from a condition of no water stress into a condition of incipient national water stress [29]. U. S. agriculture uses both blue water (water from aquifers and surface supplies) and green water (water stored in soil transpired by plants) [31]. U. S. agriculture is the second largest consumer of blue water [32,33] and it, along with forestry, are the major industries using green water. The supply of blue water is dependent on effective precipitation (EP), which is defined as the part of rainfall that reaches streams or recharges groundwater. The future biofuels production industry will create major new demands on the quantity of water used by agriculture and production forestry in the United States. New research and new tools are needed to account for these water demands as the nation implements sustainable biofuels production, because in many parts of the United States the agricultural sector already face water shortages. In the west, agricultural withdrawals account for 65-85% of total water withdrawals [33]. In the east, irrigation supplies are under pressure from competing uses, especially in periods of drought. Although overall water withdrawals in the United States have decreased since 1980 and irrigation efficiency improvements are still possible, the amount of both green and blue water needed for a biofuels-based energy supply is much greater than for the historic, fossil fuels based economy.

Comparisons of perennial bioenergy crops (both woody and herbaceous) have shown higher evapotranspiration (ET) and less EP than either annual crops or natural ecosystems. Simulations using the Environmental Policy Integrated Climate (EPIC) model showed the potential for 12-30% more ET from switchgrass than either corn or winter wheat in the Midwest [34]. Water balance measurements for pine plantations in the southeast showed 30% higher ET than natural pine forests [35]. At this point, watershed research has not shown that these increases in field scale ET will affect streamflow discharge for areas of perennial bioenergy crops expected in most watersheds. Modeling studies have examined the effects of conversion to perennial bioenergy crops, especially switchgrass, on water quantity and quality.

Simulations using the Soil and Water Assessment Tool (SWAT) in Minnesota showed only a small decrease in streamflow when 27% of the watershed was put into switchgrass instead of conventional crops [36]. Conversely, applying SWAT to the Delaware river watershed in Kansas and assuming 43% of the watershed was converted to switchgrass estimated that surface runoff would decline by 55% with large reductions in edge of field erosion, sediment yield, and nitrogen export. Reduction in nitrogen export depended on the fertilizer level assumed for the switchgrass [37]. Modeled results from the United Kingdom showed that EP was lower for both Miscanthus x giganteus (M. x giganteus) and short-rotation coppice (SRC) willow (Salix spp) [26]. The decrease was greater for SRC willow than for M. x giganteus, generally reflecting the lower water use efficiency in C3 plants (willow) than in C4 plants (M. x giganteus). On a watershed basis, the effects on EP will depend on both the fraction of the watershed in perennial biofuels crops and the precipitation. Modeled decreases in EP were greatest in areas below about 600 mm annual precipitation [26].

Vanloocke et al. [38] used a plant growth and ET model to estimate that M. x giganteus grown in the Midwest United States would use more water than the agro-ecosystems it might replace. The model showed that substantial increases in water evaporated to the atmosphere and potential decrease in EP to streams and groundwater would only occur when the M. x giganteus fraction cover for a watershed exceeded 25% in dry regions and 50% in nearly all of the rest of the Midwest. Conversion to woody biomass may also lead to reduced streamflow compared to non-forested watersheds. Farley et al. 2005[39] estimated that in arid areas where streamflow was 10% or less of rainfall, afforestation could eliminate most streamflow and that in areas where streamflow was 30% of rainfall, streamflow could be cut in half by afforestation of shrublands or grasslands.

Direct measurements of water quantity and quality effects of perennial biomass grasses are rare. McIsaac et al. [40] showed that nitrate leaching was very low beneath unfertilized M. x giganteus and switchgrass primarily because drainage water was reduced, especially under M. x giganteus. They estimated that in the tile-drained Midwest, M. x giganteus could reduce streamflow by as much as 32% compared to conventional corn/soybean rota­tions. Experimental studies of bioenergy feedstock crops generally give results that show the complex interactions among fertilizer regimes, crops, and development of perennial crop­ping systems. For instance, in north Alabama, coppiced sweet gum had much lower nitrogen and phophorus losses in runoff than corn, but switchgrass had phosphorus losses similar to corn and only in the final two years of a five-year trial were NO3-N losses lower [23]. In addition, there were high erosion losses for sweet gum unless it was grown with a cover crop.

Although runoff and erosion should be less from fields in perennial bioenergy crops than in annual crops, meaningful comparisons will depend on management of both cropping systems. Conservation tillage generally reduces runoff and erosion and may have lower rates than under perennial crops, especially during the establishment phase of the bioenergy crops. Furthermore, the high cost of perennials may lead to low establishment rates and more bare soil than with annual cropping systems.

Sustainability issues related to water will help determine the types of cellulosic feedstocks grown. For instance, due to higher biomass production, greater leaf area index, and longer vegetative growing season, M. x giganteus requires more water during the growing season than switchgrass or corn [40, 41]. Consequently, water availability strongly influences M. x giganteus yields, and the crop is thought to be best suited to locations that receive at least 30 inches of precipitation per year. In contrast to M. x giganteus, switchgrass yields are most strongly influenced by nitrogen availability [42]. This means that in areas like the Midwest where rainfall is generally adequate, but high nitrates in drain tile are an issue, M. x giganteus would be a better choice for producers than switchgrass. In drier areas of the country, where water is limiting, but groundwater quality is not an issue, adequately fertilized (i. e., 50-100 lbN/acre or 56-110 kgN/ha) switchgrass may produce higher amounts of biomass [42].