Prévision des potentialités de production de miel à l'échelle d'un rucher au moyen d'un système d'information géographique

Abstract

The objective of this work was to develop a method to predict the honey production of an apiary by means of landscape variables. A model was built using attractivity and productivity formulae based upon the flowering periods, the distances to the apiary, and the density and specific nectar productivity of the floral species surrounding the apiary. Honey harvests of 12 apiaries in Wallonie, Belgium (Fig. 1), were recorded from 1998 to 2000. The plant species, location and absolute percentage of cover of nectar-producing (melliferous) plants within a radius of two km of each apiary were surveyed during the summer 2000. These data were encoded and geo-referenced with a Geographical Information System (GIS). This software used tables containing data from field survey and the theoretical melliferous potentials of encountered plants (Tab. I). The model brought out the listing of foraged flowers and their potential honey productivity during flowering periods for each foraging area studied (e.g. Fig. 3). The predictions of honey production (kg/hive/year) were compared to the actual production and the model explained 60% of the variance of the observations (F1,10 = 37.76; P < 0.001; r2 = 0.60; Fig. 2). Some tests of the sensitivity of the results to changing parameters showed the importance of a minimal 2-km radius to model the foraging range (Tab. II). Taking into account landscape diversity, the abundance of areas containing dense melliferous flora was also essential. This work improves our understanding of the melliferous value of the landscape around an apiary. This model represents an interesting tool for beekeepers that would like to evaluate the potential honey production of an area and represents progress in the cartography of regional beekeeping potential. © INRA/DIB-AGIB/ EDP Sciences, 2006.