Biotechnologie, Agronomie, Société et Environnement (Jan 2018)
L'expertise pour prédire la production cotonnière en Afrique de l'Ouest : est-elle une solution face aux aléas climatiques émergents ?
Abstract
Expertise for forecasting cotton production in Western Africa: a solution to emerging climate uncertainty?. Description of the subject. Climate uncertainty is expected to have a severe impact on the economy of West African countries. The use of forecasting in agricultural production has become a priority for those working in the field of development. This article focuses on the early prediction of cotton production levels before harvesting in Senegal, using expertise from the cotton development company, SODEFITEX. Objectives. This paper aims to analyze the knowledge employed to predict yield, given the uncertainty about future climate conditions, and to evaluate the quality of such predictions, depending on the territorial scale of the expertise considered. Method. The methodology used in this study was a statistical analysis of cotton production in Senegal between 2004 and 2015, combined with a survey of 32 SODEFITEX experts. Results. Analysis of the survey of experts revealed that their evaluation of the cotton production was based mainly on the state of the crops and on crop management, with the planting date being the priority. For some experts, the history of production and its organizational and social environments were also important factors. A comparison of forecasts with actual production showed that the experts had systematically overestimated the levels of production when actual yields were below 1,000 kg·ha-1, under adverse climatic conditions. Conclusions. Within a context of poor scientific documentation of the early estimation of production by experts, the results of the study shed light on the knowledge drawn upon and the performance of the forecasts delivered, thus paving the way for a rational criticism of the expertise employed. The diversity of points of view highlighted by the multi-criteria analysis, between hierarchical levels and between agents at the same level, underlines the need for a closer look at the knowledge being utilized, in order to bring out "what matters" in these forecasts and to move towards greater objectivity of expertise in the face of recurrent climate hazards.
