Élaboration du rendement du blé dur en conditions semi-arides en Tunisie : relations entre composantes du rendement sous différents régimes de nutrition azotée et hydrique

Abstract

Summary — Analysis of yield components to diagnose limiting factors of durum wheat yield in semi-arid conditions: relationships under various water and nitrogen supplies. We studied the yield components of durum wheat (Triticum durum Desf) under semi-arid conditions in Tunisia (fig 1). The relationships between several yield components were established in conditions where nitrogen and water were unlikely to be limiting. The level of each component was fitted to that of components determined during previous periods of the crop cycle. Resulting relationships were considered as describing the maximum values of each component. They were plotted together with coupled values of components observed in conditions where water or nitrogen could be limiting. This allowed diagnosis of the effects of varying levels of nitrogen and water supplies on each yield component. The experiments were carried out in two soils (Tell and Hamri, see table II) and several levels of nitrogen fertilization (fig 2), for two years with contrasting rainfall. The relationships between yield components, observed in the wet year and for high levels of nitrogen supply had similar patterns compared with those established for wheat grown in temperate regions (table I). However, the values of fitted parameters (table III) were appreciably different. In Tell soil, the maximum levels of all yield components were reached during the wet year for nitrogen fertilization as low as 60 kg haapplied at tillering and beginning of the stem elongation; in Hamri soil, the maximum levels could not be reached, even for a total nitrogen amount of 100 kg hayield components had very low levels during the dry year in the two soils, with no appreciable effect of fertilization (table IV). In both dry and wet years, yields correlated better with the number of kernels per m (fig 3) than with the kernel weight. This suggests that pre-anthesis environmental conditions had a overriding effect on yields (figs 4 and 5). The observed range of kernel weight was not large enough to induce high variations in yield, in spite of the wide range of water and nitrogen supply. We used these results to propose a way of deciding nitrogen fertilization in relationship with yield objective in wet conditions (table V).