243ronment. In West Asia and Northern Africa, Cooper et al. ( 1987 ) promoted manage-
ment practices favouring rapid canopy growth to reduce soil evaporation and
increase transpiration. Currently, the view of breeders in Australia has converged
with that of Cooper and his colleagues, to favour plant traits associated with rapid
water use thereby reducing soil evaporation in environments with dominant winter
rainfall (Richards et al. 1993 ; Rebetzke and Richards 1999 ; Richards and Lukacs
2002 ). In the prevalent climate of south-eastern Australia, saving water through
reduced water uptake and slow canopy growth produces little or no benefit, as most
of the water is lost through soil evaporation i.e. in environments with a high fre-
quency of small rainfall events (Sadras and Rodriguez 2007 ). This is not the case in
Australia’s northern region where winter rainfall is infrequent, and the fraction of
soil evaporation relative to total water use is small compared with southern loca-
tions (Mitchell et al. 2006 ; Sadras and Rodriguez 2007 ). For these reasons, traits
favouring slow canopy cover in northern locations and fast canopy cover in southern
locations could be adopted as a primary approach to overcome the constraints
imposed by rainfall patterns on water use and WUE (Sadras and Rodriguez 2007 ).
Similar examples from simple models which support the design of dryland agri-
cultural systems in Africa are in Dimes et al. ( 2015 ) (Fig. 2 ). Their work combined
Fig. 2 The relationship between simulated maize yield and seasonal rainfall at Mandela, Tanzania
for open pollination varieties and hybrid seed with and without weed competition and varying rates
of nitrogen fertiliser (0, 20, and 183 kg N/ha). Lines have slopes of 2.5 (solid), 4.1 (dashed) and 16
(dotted) kg/ha.mm (Source: Dimes et al. 2015 )
Modelling Dryland Agricultural Systems