Résumé:
This work, which falls within the framework of a research program on the issue of the introduction of simplified cultivation techniques in Algeria focuses on an analysis of the effect of three cultivation techniques for the establishment of a cereal, on soil physical properties and the effects on root development and on the durum wheat yield. And the impact of these techniques on weed infestation. Tests carried out at the experimental station of the Agricultural National School (EL HARRACH) and the Experimental Station of ITGC shows clearly that the technique used has an effect on the evolution of moisture, on the porosity and on the cone index of the soil. The water content in the soil is preserved better with the direct drilling, its value reaches from 12,97% before sowing to 16,51% at the ear development stage on the plowed parcel (TC); it passes from 13,83% to 16,73% on the minimum tilling (TM) and from 14,28% to 18,27 on the direct drilling (SD). Porosity is slightly more important with the conventional technique. As regards penetrometric resistance, the soil is more resistant on the parcel in direct drilling where it reaches 15,84 daN/cm ². Rp is also strongly influenced by the water content. The mean values show that the penetrometer resistance is inversely proportional to the water content. These changes are illustrated by the relation: Rp (TCNT) = 11,95 + 0,006*n - 0,31*H. The results obtained also show that the roots system is developed more on parcel worked with the conventional method than on those worked in direct drilling and minimum work. The estimated yield are respectively 36,19 q/ha for the direct drilling and 50,03 q/ha for conventional work and it is only 35,7 q/ha on the parcel worked with the chisel. Statistical analysis showed that there was a highly significant effect of treatment with the herbicide on weed density in direct seeding P-value <0.0001. From the analysis of the correlation is observed that there is a strong correlation between:
H% TC = 14,18 - 0,027D.m.h (TC), Hauteur tige (SD) = 33,43 + 5,75 H% SD, RDT (SDT) = 32,27 - 0,16*D.m.h (SDT), Biomasse (SDT) = 287,83 - 1,50 D.m.h (SDT). The analysis of multiple regression between the dependent variable and the independent variables RTD Dmh, H% and the stem diameter gave the following equation of the model:
RDT (SD) = 14,96 + 0,10*D.m.h(SD) - 1,26*H%(SD) - 11,70*Diamétre (SD). The coefficient of determination R² = 0.99, the variable density of weeds, humidity and stem diameter can explain 99% of the variability in yield. These three variables have some effect on the yield in plots conducted till, for plots worked we could not obtain an explanatory model like direct seeding. This leads us to say that the mastery of weeds certainly leads to obtaining a good yield.
