One of the major ways to increase the water use of the crop itself is by increasing the depth of rooting. In many dryland environments, crops do not use all the water available in the soil profile because of restrictions to root growth. These restrictions may be physical, chemical, or biological. Agronomic practices that reduce the physical impedance to root growth can benefit yields of dryland crops in water-limited environments. Deep ripping to about 30 cm has been shown to increase yields and hence rainfall-use efficiency on deep sandy soils (Jarvis, 1982; Delroy and Bowden, 1986; Asseng et al., 2002; Asseng and Turner, 2003). Other physical soil constraints such as compacted subsoils can be alleviated by the application of gypsum to flocculate the soil particles, and to increase water penetration and root growth (Hamza and Anderson, 2002, 2003). Chemical constraints are not as easily remedied, but soil acidity at depth that constrains root growth can be ameliorated by liming, particularly with deep placement of lime. However, soil alkalinity that restricts the growth of lupin roots (Atwell, 1991; Tang et al., 1992), soil sodicity, and boron toxicity are more difficult to ameliorate agronomically and may need to rely on the use of different species or tolerant genotypes (Tang et al., 1993). Finally, root diseases and nematodes can constrain root growth and are most easily controlled by rotations to reduce the disease- and nematode-incidence and by cultivation techniques that minimize fungal activity (Roget et al., 1996).
It should be noted that deeper roots are not always beneficial. In environments in which the seasonal rainfall and soil characteristics are such that the depth of soil wetting is restricted, deeper rooting will be of no benefit. A simulation analysis by Asseng et al. (2002) showed that deeper roots gave the greatest benefit on sandy soils, particularly in the high-rainfall zones where nitrogen can leach below the root zone, and had smaller or even negative effects on yields for wheat growing on clay soils with limited wetting to depth (Smith and Harris, 1981). The analysis also demonstrated the role of nitrogen application in overcoming restrictions to rooting depth, particularly in sandy soils (Asseng et al., 2001b).
Rotations also provide an opportunity to increase water use by a crop. Roots of some species have the potential to penetrate deeper into the soil than others (Hamblin and Hamblin, 1985), and this may provide ‘biopores’ for a subsequent crop. It has been suggested that both narrow-leafed lupin (Lupinus angustifolius) and canola/oilseed rape (Brassica napus) develop ‘biopores’ in the soil that allow easier root penetration by the water and roots of a subsequent crop (Angus et al., 1991; Cresswell and Kirkegaard, 1995). However, results have been equivocal. Nevertheless, there is considerable evidence that lucerne (Medicago sativa) has roots that penetrate deep into the soil over 2–3 years and allow deeper water penetration and deeper root penetration by a subsequent crop (Ward et al., 2002).
However, the major impact of agronomic management on rainfall-use efficiency has not arisen from increasing total water use by the crop in evapotranspiration, but from increasing water use by the crop itself in transpiration at the expense of water loss by weeds or from the soil by soil evaporation, deep drainage, surface runoff, or lateral throughflow. This increase in water use by the crop at the expense of other losses generally results in significantly increased yields, with only a 5–10% increase in total evapotranspiration (Asseng et al., 2001c).