Sugar beet yield loss predicted by relative weed cover, weed biomass and weed density

Sugar beet yield loss was predicted from early observations of weed density, relative weed cover, and weed biomass using non-linear regression models. Six field experiments were conducted in Germany and in the Russian Federation in 2012, 2013 and 2014. Average weed densities varied from 20 to 131 with typical weed species compositions for sugar beet fields at both locations. Sugar beet yielded higher in Germany and relative yield losses were lower than in Russia. Data of weed density, relative weed cover, weed biomass and relative yield loss fitted well to the non-linear regression models. Competitive weed species such as Chenopodium album and Amaranthus retroflexus caused more than 80% yield loss. Relative weed cover regression models provided more accurate predictions of sugar beet yield losses than weed biomass and weed density.