Modélisation de la variabilité spatiale de quelques propriétés du sol dans la région de Ghardaïa (cas de Hassi F'Hel)

Abstract

The productivity of agrosystème or natural ecosystem depends essentially on variability of soil properties on field scale. This work is for objective the modelling of soil reaction spatial variability and electric conductivity by geostatistical approach. The work was realized on Hassi Fehal region (Ghardaïa), samples were taken from depth from 0 to 25 cm and analyzed for the salinity and pH of soil. The analysis combined by variance and annual variation rate of studied parameters showed a very highly significant decrease of pH and salinity. The geostatistical analysis revealed diverse models and levels for spatial distribution of studied properties. The results showed a spatial dependence moderated in strong for pH and salinity. The variographic analysis showed that the effect of nugget is very low for the ground reaction and nothing for salinity. The reach varies of 30,25m for soil reaction to 48.13m for salinity. So confirming validity of adopted and allowing future plans optimization of sampling. The results of crossed validation ended on variograms models selection of most reliable to be known, circular and exponential, which were used to realize of spatial variability cards by ordinary krigeage. Synthetically, the results indicate that the agricultural intensification by cereals cultivation under pivot on Hassi Fehal region influenced quantitatively and qualitatively the content and spatial variability of soil properties. These results must be considered for an integrated and long-lasting lands management on these similar regions.