Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Belgique → Estimating and understanding smallholder crop production variability from pixel to village levels thanks to satellite observation in Mali

Université catholique de Louvain (2018)

Estimating and understanding smallholder crop production variability from pixel to village levels thanks to satellite observation in Mali

Lambert, Marie-Julie

Titre : Estimating and understanding smallholder crop production variability from pixel to village levels thanks to satellite observation in Mali

Auteur : Lambert, Marie-Julie

Université de soutenance : Université catholique de Louvain

Grade : Docteur en sciences agronomiques et ingénierie biologique 2018

Résumé
In Sudano-Sahelian Africa, crop production is very variable across fields and years due to spatial rainfall heterogeneity, soil fertility constraints and agricultural practices. Crop production estimates are often available too late and at large administrative unit levels, preventing their use by the international community to support food security. This thesis aims to monitor local crop growing conditions and estimate crop production at village level. Unlike current agriculture monitoring systems, local crop growth variability was well depicted by moderate resolution imagery. This required an up-to-date cropland map successfully derived for the entire region thanks to a knowledge-based automated classification. Combining results from crop classification and yield estimation, production for four crops was accurately estimated at village level. This thesis also aims to better understand the smallholder crop production variability. We estimated the locally attainable yield gap at field, farm and village levels for four crops. Hierarchical analysis of farmer survey data highlighted the driving forces behind this yield gap, pinpointing inputs such as fertilization, farmer socio-economic conditions and farmer strategies. While most generic yield gap approaches assume that all farmers maximize their productivity, we showed that low income farmers put the emphasis on managing food shortage risks, while wealthier farmers allocate resources to manage the long-term fertility of their fields. These methodological developments are very promising to design, target and evaluate agriculture interventions according to the actual farmer’s needs

Présentation

Version intégrale (37,3 Mb)

Page publiée le 16 janvier 2020