[PhD’s Corner] Julien Sarron: Spatial estimation of a perennial crop yield in West Africa : mango case study in Senegal

Julien Sarron has successfully defended his PhD Thesis on 9 December 2019 at Montpellier SupAgro

Spatial estimation of a perennial crop yield in West Africa : mango case study in Senegal

  • Start Date: October 2017
  • Defence : 9 December 2019
  • University: MUSE Montpellier University of Excellence / Montpellier SupAgro
  • PhD SchoolGAIA , Montpellier
  • Field(s): Agronomy
  • Doctoral Thesis Advisor: :Eric Malézieux (Cirad HortSys)
  • Co-supervisors : Emile Faye (Cirad HortSys)
  • Funding: Cirad
  • #DigitAg: Labelled PhD- Axis 6 – Challenge 1 & 8

 

Keywords:  tree crop, agroforestry, cropping system, remote sensing, landscape, geostatistics, mango orchard, yield, digital agriculture, Sahel, West Africa, agroecology,

 

Abstract:

West Africa faced many challenges including demographic growth and climate change that put agriculture under pressure. Fruit trees, despite the numerous ecosystem services they provide (food, biodiversity, soil fertility, microclimate, etc.), suffer from low yields. In this context of food insecurity, it is essential to improve tools for agricultural yield monitoring to face development challenges. The objectives of this thesis were to develop tools for yield estimation of mango (Mangifera indica L.) at different scales and to identify drivers of yield variabilities in West Africa. These tools have been deployed to analyse yields of 30 orchards depicting three cropping systems (extensive, diversified, and intensive) in the Niayes region in Senegal. At the tree scale, an image analysis tool using deep learning algorithm combined with models allowed to accurately estimate the production of mango trees before harvest. The ‘yield gap’ concept was adapted to compute the tree ‘production gap’ as the difference between the attainable production of the tree (i.e., determined by the cultivar and the structure) and its actual production. This statistical method evidenced that the mango production was influenced by the cultivar and the cropping system. At the orchard scale, drone imagery was used to build land cover map and estimate the structure parameters (tree height, crown area, crown volume) of all trees in the orchard. These data allowed to model and spatialize the individual production of each mango tree. Yield estimation allowed the study of the variability of yields within and between orchards. Spatial distribution of tree structure and production appear more spread out in the extensive and diversified orchards. At the regional scale, regional agronomic diagnosis allowed to identify important drivers impacting orchard mango yield in the study area. While mango planting density is a key factor for yield in intensive orchards, the tree production is favoured in diversified orchards. The climate, the management practices, and the species diversity also induce yield variabilities. This thesis opens new methodological paths to compensate the lack of data for fruit tree yields analysis in West Africa. Finally, the study of factors impacting mango yields (climate, practices, and species diversity) will allow to sustainably improve practices and production of mango in West Africa.

 

Contact:  julien.sarron [AT] cirad.fr​ – Tél : +221 786338087

Networks: ResearchGateLinkedIn

 

Publications

Sarron J.; Sané C.A.B.; Borianne P.; Malézieux É.; Nordey T.; Normand F.; Diatta P.; Faye É. (2018) Is machine learning efficient for mango yield estimation when used under heterogeneous field conditions? XXX. International Horticultural Congress, 2018, Istanbul, Turkey [Acta Hortic., in press.].

Sarron, J.; Malézieux, É.; Sané, C.A.B.; Faye, É. (2018). Mango Yield Mapping at the Orchard Scale Based on Tree Structure and Land Cover Assessed by UAV. Remote Sensing, 10(2), 1900. https://doi.org/10.3390/rs10121900

Communications

Sarron, J., Sané, C. A. B.; Borianne, P., Malézieux, E., Nordey, T., Normand, F., Diatta, P. Is machine learning efficient for mango yield estimation when used under heterogeneous field conditions? Acta Horticulturae. (under review) – ISHS Young Minds Award : la communication présentée à HIC 2018, 30th International Horticultural Congress, 12-16 August 2018, Istanbul (Turkey) a été primée par l’International Society for Horticultural Science-https://agritrop.cirad.fr/592236/

Sarron J., Sané C.A.B., Diatta P., Malézieux É, Faye É, Plant diversity affects the productivity of Senegalese mango orchards: evidences from UAV photogrammetry, 4th World Congress of Agroforestry (2019) – https://agritrop.cirad.fr/592664/

P. Borianne, J. Sarron, F. Borne and E. Faye, Deep mangoes: from fruit detection to cultivar identification in color images of mango trees, DISP’19 – International Conference on Digital image and Signal Processing (2019) – https://hal.archives-ouvertes.fr/hal-02295256

PixYield – L’analyse d’images photographiques comme outil pour l’estimation de rendement : le cas du manguier au Sénégal (poster)