Improved targeting of maize varieties to different on-farm production conditions in East and Southern Africa based on crowdsourced varietal performance data

Project: PhD

Project Details

Description

Slow varietal turnover is partly the cause of low crop yields in in Sub-Saharan Africa (SSA). Inability to identify suitable and superior technologies is partly attributed to poor definition of recommendation domains (also referred to as TPEs –Target Population of Environments), poor performance of new varieties relative to existing ones, and farmers’ perception of new varieties and technologies as riskier. TPE definition and evaluation of crop designs (i.e., combinations of Genotype and Management, G×M) in on-farm conditions are critical in improving varietal targeting and turnover. This project will combine data analysis, participatory on-farm trials, and modelling, to redefine TPEs for maize in East and Southern Africa and use them to assess the performance of crop designs, socioeconomic drivers behind farmer choices of crop designs and risks associated. First, systematic review and unsupervised statistical modelling using biophysical and socioeconomic variables will be used to define TPEs and their potential use in maize breeding (Chapter 1). Then, empirical tricot experiments will be conducted in redefined TPEs, to validate the TPEs, and estimate performance, i.e., worth (probability of winning) and yield of genotypes and G×M combinations and assess how covariables (weather and socioeconomic) influence varietal and G×M choices (Chapter 2). Using this field performance data, Chapter 3 will use crop simulation to scale and evaluate both biophysical (yield variability) and economic (return) risks of G×M combinations. Finally, Chapter 4 will combine all previous results to assess how farmers’ risk attitudes and perceptions and their socioeconomic context explain farmers’ varietal and G×M choices.
StatusActive
Effective start/end date1/01/22 → …

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.