Soil fertility spatial variability exists within and across smallholder farms, yet blanket recommendations are made across heterogeneous farms resulting in poor crop performance. A fertilizer management experiment was conducted on three smallholder farms within 1-km distance, positioned on upper, middle and lower catena, with soil organic carbon (SOC) of 3.9, 6.4 and 8.9 g kg-1, respectively; hereafter referred to as low, medium and high. The objective of the study was to investigate maize yield and weeds response to SOC and fertilizer management. A nutrient omission experiment with unfertilized control, NK, NP, PK, and NPK treatments were laid in a randomised complete block design and each treatment was replicated three times. Linear mixed-effects model (Restricted Maximum Likelihood (RELM)) in GEN STAT Discovery 14 was used to analyse data. Maize yield significantly increased by 2.53 and 2.62 times as SOC changed from low to medium and high, respectively. Maize yield in NP and NPK treatments did not significantly differ suggesting that granite derived soils have potassium levels adequate to support maize yield below 3 t ha-1, while N and P were the most limiting elements to maize yield. Multivariate, Principal Component Analysis (PCA), CANOCO 5 revealed that weeds species were strongly associated with SOC and weed density significantly (P< 0.05) increased from 317 to 332 weeds m-2 as SOC changed from low to high status. Compared to low SOC, weed biomass significantly increased by 4.5 times more in high SOC. Significantly higher weed density and biomass were recorded in treatments that contained P fertilizer (PK, NP, NPK) compared to NK treatment which was similar to unfertilized control, suggesting that P was limiting factor for weed density and biomass. In this study, we concluded that maize productivity, weed density, and biomass were influenced by SOC content, catena position, and fertilizer management. The strong association between weed species and SOC content, suggests that weed ecology could be used as a proxy for soil fertility status and fertilizer recommendations.

Keywords: Maize yields, macronutrient management, toposequence, weed density, and biomass.