Caractérisation agro-morphologique de la vesce (Vicia spp.), diagnostic de la situation fourragère et mise en évidence de l’importance des différents types de fourrages dans la région semi-aride de Sétif

Abstract

In agriculture, prevention performance of a parcel is paramount. In this study, we tried to predict the yield from species of varieties of leguminous species in a semi-arid region. Several variables are taken into consideration as the number of leaf, flower, and stem ...etc. Since these variables are characterized by uncertainty and imprecision, we proposed a tool for data analysis based on artificial intelligence techniques, including the principles of fuzzy logic. The result was very satisfactory. Program established for predicting the performance of a plot just from imprecise inputs variables of the system. Therefore in this context, and the main objective is to characterize morphologically some vetch species in the semi-arid region of Setif in Algeria, determine the most discriminate variables and find the varieties the most relevant and most efficient to satisfy the food needs of our livestock both quantitatively and qualitatively. As the data characterizing these parameters are distinguished by their vagueness and uncertainty, we found it useful to apply the principles of artificial intelligence including the principles of fuzzy logic in data analysis. The principles of fuzzy inference are perfectly suited to these data. The same variables were included for a second time for statistical analysis, the first descriptive and the second discriminant whose results corroborate. These analysis have detected significant differences of the input of the three species where the third one (V. sativa) which is obvious from all the variables by: Nr ,Lrp , Ngrg , Ngrp , Pgrp and yield and that at the two sites ( Northern and Centre of Setif ) . It is also clear from these analyzes that the variables that have the most weight or the most relevant in discriminating between the three species are Ngrp , Nfr , Lrp and Ngp. Variety evaluation allowed to see a lack of distinction between the varieties of the two introduced species 1 and 2 ( V.ervilia and V. narbonensis ) , a difference of 70 % was found between the varieties of local specie 3 V. sativa , so we can infer that these three species have different uses for their agronomic traits related to biomass production and grain yield , they have considerable potential fodder, including V.sativa which we can offer as complete feed for use on different forms. The second part of this work has been used to analyze the forage situation in the region of Setif and make a socio-geographical characterization of vetch in the same region. This was achieved through a series of surveys conducted over two seasons (2009/2010 and 2010/ 2011). The results of these investigations revealed that this region is characterized by the combination of cereals (especially durum wheat) for livestock whose diet suffers from a large deficit relying heavily on very modest resource production natural forage as the share of fodder crops is negligible compared to the rate of the needs of animals which are growing increasingly dependent on ever leaving the industry abroad. This situation prevailed for many years