Phosphorus runoff from a phosphorus deficient soil under common bean (Phaseolus vulgaris L.) and soybean (Glycine max L.) genotypes with contrasting root architecture

The selection and breeding of crop genotypes with root traits that improve soil resource extraction is a promising avenue to improved nutrient and water use efficiency in low-input farming systems. Such genotypes may accelerate nutrient extraction ("nutrient mining"), but may also reduce nutrient loss via soil erosion by producing greater shoot biomass and by direct effects of root traits on aggregate formation and water infiltration. Little is known about the effects of root architecture on phosphorus (P) runoff and soil erosion, and the relative importance of root and shoot traits on runoff P loss has not been determined. Four genotypes of common bean (Phaseolus vulgaris L.) and two genotypes of soybean (Glycine max) selected for contrasting root architecture were grown in a low P soil (Aquic Fragiudult, < 20 mg kg(-1) Mehlich-3 P, 3% slope) and subjected to rainfall-runoff experiments with and without shoot removal. Plots with intact shoots had significantly lower runoff volumes (1.3-7.6 mm) and total P loads in runoff (0.005-0.32 kg ha(-1)) than plots with shoots removed (7.0-16.8 mm; 0.025-1.95 kg ha(-1)). Dissolved reactive P leached from plant material did not contribute significantly to P loss in runoff. Total root length acquired from soil cores differed significantly among genotypes. Root length densities in the upper 15 cm of soil mid-way between rows were less than 4.0 cm cm(-3) and variation in root length density was not correlated with runoff or P loss. Root length density also did not affect rainfall infiltration or surface runoff volume. We conclude that for annual dicotyledonous crops such as bean and soybean with relatively low root length densities, root traits have little direct effect on soil erosion.