Farming system context drives the value of deep wheat roots in semi-arid environments.

The capture of subsoil water by wheat roots can make a valuable contribution to grain yield on deep soils. More extensive root systems can capture more water, but leave the soil in a drier state, potentially limiting water availability to subsequent crops. To evaluate the importance of these legacy effects, a long-term simulation analysis at eight sites in the semi-arid environment of Australia compared the yield of standard wheat cultivars with cultivars that were (i) modified to have root systems which extract more water at depth and/or (ii) sown earlier to increase the duration of the vegetative period and hence rooting depth. We compared simulations with and without annual resetting of soil water to investigate the legacy effects of drier subsoils related to modified root systems. Simulated mean yield benefits from modified root systems declined from 0.1-0.6 t ha(-1) when annually reset, to 0-0.2 t ha(-1) in the continuous simulation due to a legacy of drier soils (mean 0-32mm) at subsequent crop sowing. For continuous simulations, predicted yield benefits of >0.2 t ha(-1) from more extensive root systems were rare (3-10% of years) at sites with shallow soils (<1.0 m), but occurred in 14-44% of years at sites with deeper soils (1.6-2.5 m). Earlier sowing had a larger impact than modified root systems on water uptake (14-31 vs 2-17mm) and mean yield increase (up to 0.7 vs 0-0.2 t ha(-1)) and the benefits occurred on deep and shallow soils and in more years (9-79 vs 3-44%). Increasing the proportion of crops in the sequence which dry the subsoil extensively has implications for the farming system productivity, and the crop sequence must be managed tactically to optimize overall system benefits.