Adapting food systems to future climates needs quick crop varietal replacement informed by climatic information. Large-N on-farm trials in which triplets of new crop varieties are being tested under different environmental conditions have proven to be a valuable information source. However, owing to its methodological novelty, the experimental data has not been used to its full potential. The objectives of this thesis topic is to enhance the analysis of data from the on-farm trials in combination with improved environmental data and/or to facilitate interpretation and decision making using suitable reporting and presentation of the analysis results.
So-called large-N on-farm trials follow a citizen science philosophy and involve large groups of farmers who volunteer to test on their farms. The approach takes advantage of variation between heterogeneous farms under diverse environmental conditions to determine varietal performance under varying environmental conditions. The methods extract much more information about environmental adaptation than conventional participatory methods and allow more farmers to engage with the technology testing process. Bioversity has developed innovative methods to make this possible through “triadic comparisons of technologies” (tricot). Accordingly, farmers test three crop varieties on their farms and assign ranks to these. The full experiment involves distributing several varieties in triplets over many farms covering a wide range of environmental conditions. Environment-specific scores for the full set of tested varieties are obtained by rank comparisons methods, such as (extensions of) the Plackett-Luce model while the environmental interactions are currently accounted for by model-based recursive partitioning using environmental covariates. Earlier MGI thesis work has analysed data from participatory common bean variety trials in Nicaragua, Central America. The current thesis will extend this work to different, even larger data sets and/or develop and try new methods to apply to the data, for example to recommend portfolios of complementary varieties to address climate risk.
Objectives (choose from):
- Select spatially explicit co-variates for use in the variety ranking analysis
- Model-based recursive partitioning of a large set of partial comparisons of wheat varieties under different environmental conditions
- Validation of analysis results Presentation of analysis results, for example using variety portfolio recommendations
- To be discussed ...
- Coe, 2002. Analyzing ranking and rating data from participatory on-farm trials. In: Bellon, M.R. and Reeves, J. (eds). Quantitative analysis of data from participatory methods in plant breeding. CIMMYT, Mexico, pp. 44-65.
- Turner, H., van Etten, J., Firth, D., Kosmidis, I., 2019. Introduction to PlackettLuce. https://cran.rstudio.com/web/packages/PlackettLuce/vignettes/Overview.html
- Van Etten, J., Beza, E., Calderer, L., Van Duijvendijk, K., Fadda, C., Fantahun, B., . . . Zimmerer, K. (2016). First experiences with a novel farmer citizen science approach: crowdsourcing participatory variety selection through on-farm triadic comparisons of technologies (tricot). Experimental Agriculture, 1-22. doi:10.1017/S0014479716000739
- van Etten, J., de Sousa, K., et al. 2019. Crop variety management for climate adaptation supported by citizen science. Proceedings of the National Academy of Sciences, 116(10), pp.4194-4199.
- MGI theses of Elise van Tilborg and Meng Zhang
- Strong analytical skills (including statistical analysis)
- Scripting skills
Theme(s): Modelling & visualisation, Empowering & engaging communities