Spatial relation of weed competition and soil fertility in soybean farming

Abstract

Soybean is one of the main crops in Brazil, with a significant share of national agribusiness exports. Nonetheless, several factors such as weed competition and soil fertility directly affect soybean yield and productivity. This study aimed to analyse the spatial distribution of weeds as a function of soil fertility and soybean yield in farming fields. We carried out the experiment on a farm located in Brejo, Maranhão state, Brazil, through a geostatistical analysis of 60 sampling points on a regular grid of 10.0 m x 50.0 m. At these points, we collected phytosociological information on the weed community, soil fertility, and soybean yield. We performed principal component analysis (PCA) to determine the most responsive variables and to group them. We determined spatial dependence through geostatistical procedures, with the interpretation and adjustment of variogram components. We identified seven weed species, distributed across seven genera and six botanical families, of which 76.78% were eudicotyledons. In the cluster analysis, we grouped monocotyledonous species separately from eudicotyledons as explained by the morphophysiological contrasts between these botanical classes. Soybean yield did not correlate with soil fertility or weeds. These two factors can be considered only as a share of soybean productivity because their individual variations do not directly influence production factors. The efficient management of weeds and soil fertility should result in a more uniform and potencially more soybean yield when other conditioning factors are also effective