Using geographic information systems to map the prevalent weeds at an early stage of the cotton crop in relation to abiotic factors

Abstract

Cotton is still one of the most important crops in Greece despite the changes in the country’s socioeconomic status which have reduced the total cultivated area. In order to minimize yield losses, weed control is essential during the cultivation period. The aim of this study was to determine the distribution of the prevalent weeds that escape the usual herbicide application in the main cotton zone, located in the Karditsa prefecture. The weed densities and the irrigation methods used were recorded in 101 sampling sites of 25 m2; the cotton crop had been grown for the last 5 years using similar weed control techniques. Existing soil maps of the area were also used, through which soil data (texture and carbonates content) were accessed. Among the 14 weed species that have been recorded, four were perennial (Cyperus rotundus, Convolvulus arvensis, Cynodon dactylon, Sorhum halepense) and were ranked as first, second, fourth and fifth, respectively, according to the mean density, indicating the inefficient herbicidal control. In the fields irrigated by sprinklers, the weeds occurred in greater populations than those that occurred in fields irrigated by drippers, at values of 4.64 and 3 weeds m-2, respectively. In terms of the studied soil properties, the distribution of C. arvensis was significantly correlated with carbonate content and soil texture in the surface soil layer. The autocorrelation analysis showed that only perennial weeds are spatially correlated whereas the interpolated maps showed this spatial trend of weed appearance.