The effect of managing improved fallows of Mucuna pruriens on maize production and soil carbon and nitrogen dynamics in sub-humid Zimbabwe

Abstract

Mucuna pruriens has emerged as a successful forage or green manure legume for use in the smallholder animal-livestock systems of Zimbabwe. The efficiency of N recovery from mucuna residues in subsequent maize crops can be low and the loss of nitrate nitrogen from the soil profile prior to maize N demand is proposed as a reason for this. An experiment was established in the 1999–2000 wet season at seven on-farm sites in a communal farming district of Zimbabwe (average rainfall 650–900 mm) on acidic (pH < 5), and inherently infertile soils with texture ranging from sandy/sandy loam (n = 5) to clay (n = 2). Improved fallows of mucuna grown for 19 weeks produced between 4.7 and 8.5 t/ha dry matter (DM) at the sandy/sandy loam sites and between 9.5 and 11.2 t/ha DM at the clay sites. This biomass was then either cut and removed as hay, or ploughed in as a green manure. Weedy fallow treatments, which represent typical farmer practice, produced 3.3–6.3 t/ha DM. A maize crop was then grown on these same sites in the following 2000–2001 wet season and the dynamics of soil N and C and maize production were investigated. Where mucuna was green manured, a positive linear response (r2 = 0.72) in maize yield to increasing mucuna biomass (containing 101–348 kg N/ha) was found. On the sandy sites, and where no P fertiliser was applied to the previous mucuna phase, a maize grain yield of 2.3 t/ha was achieved following the mucuna green-manure system; this was 64% higher than the maize yield following the weedy fallow and 100% higher than the maize yield following the mucuna ‘removed’ hay system. Apparent nitrogen recoveries in the range of 25 to 53% indicate that there are large quantities of nitrogen not utilised by the subsequent maize phase. The loss of 73 kg/ha of nitrate N from the soil profile (0–120 cm) early in the wet season and prior to maize N demand is proposed as a reason for low N recovery. No change in labile C (measured with 333 mM KMnO4) was detected through the soil profile at this time and it is suggested that labile C movement occurred between the sampling times.