Effect of soil nutrient management on soil weed seed bank dynamics across a soil fertility gradient in smallholder farms, Eastern Zimbabwe

Abstract

Weed seed banks are latent biotic stress contributors to low crop production among smallholder farms. The objective of this study was to investigate effect of repeated soil nutrient amendments on weed seed bank dynamics in eastern Zimbabwe. Soil samples were taken from three farms with low, medium and high soil organic carbon (SOC) along the catena. Sampled treatments included varying combinations of compound D (7%N, 14%P, 7% K) fertiliser, cattle manure and lime. Weeds that emerged per plot were counted by species and data was tested for normality before running a linear mixed-effects model and Restricted Maximum Likelihood. Principal Component Analysis (PCA) was performed to establish the relationship between seasons, SOC, soil nutrient amendment, and emerged weeds. Significant (p < .05) species richness (Margalef index), and Shannon Weiner index were recorded in the medium (6.4g kg-1soil), high (8 g kg-1 soil) and low (< 3.9 g kg-1soil) SOC. Weed species emergence was significantly (p < .05) influenced by the level of SOC along the catena positions, representing results from the source. Emergency and abundance of weeds such as Richardia scabra, was associated with low SOC acidic sandy soils from the upper catena. However, application of nutrients (NPK+lime treatment) reduced weed species counts from 1.96±0.12 to 1.4±0.12. Cattle manure and NPK+CM treatments recorded significantly higher weed emergence, and weed biomass; compared to the control. Clearly, cattle manure treatments significantly increase the soil weed seed bank; thus, weeding intensity is likely to increase in cattle manure treated fields.