[Statlist] Postdoc position: spatial statistics and machine learning for choosing crop varieties

[David Ginsbourger]

*Postdoc position: spatial statistics and machine learning for choosing 
crop varieties*

In the framework of the �Wheat Advisor� project coordinated by 
Swissgranum and involving various parties including researchers from the 
Swiss centre for agricultural research (Agroscope) and the University of 
Bern (UniBE), we are calling for applications for a postdoctoral 
position (80%-100%) to be funded subject to successful completion of the 
collaboration contract.

The main focus of this collaboration is to leverage recent progresses in 
statistical modelling and in machine learning to help more efficiently 
recommending which crop variety to choose in farms depending on measured 
and indirectly inferred co-variables.

Population growth and climate change increasing pressures on our global 
food systems call for a sustainable intensification of food production 
while increasing the systems� resilience to climatic risks. Recommending 
crop varieties with optimum yield potential given a particular 
environmental setting and management is key to achieving these goals. 
However, evidence-based decision-support tools that could help farmers 
choose the most suitable crop varieties for their fields are lacking so 
far. This postdoc position addresses this gap via the investigation of 
different prediction approaches to optimize variety-specific wheat 
yields given information on local climate, nitrogen supply, soil and 
topography.

This endeavor is quite challenging as available data presents 
variability due not only to the latter co-variables but also due to 
climatic fluctuations, unobserved properties of individual crops, and more.

The aim of this position is to evaluate and develop novel approaches 
borrowing the best from both distance/kernel-based prediction and mixed 
effects statistical modelling for improving decision-making regarding 
which wheat variety to grow in specific environments and designing more 
efficient experimental networks. In particular, the recruited 
postdoctoral researcher will be involved in designing a campaign of 
novel crop experiments, hence going all the way from statistical 
modelling to experimental design, data collection, and ideally 
prototyping a recommendation tool for wheat producers. The outputs will 
hence provide valuable insights to increase both food security and the 
ecological sustainability of wheat production in Switzerland.

This work will be developed mainly through a collaboration between the 
Institute of Mathematical Statistics and Actuarial Science of UniBE and 
Agroscope at Changins. The ideal candidate is a statistician with a 
taste for large-scale agronomical applications or an agronomist with 
outstanding statistics skills. The position is for 1 year, with 
possibility of an extension on the project towards further practical 
implementations of the investigated and developed approaches in 
agronomical contexts. The starting date is as soon as possible in 2020. 
Applications will be reviewed swiftly and selected candidates will be 
contacted for interviews.

*CV, publication list and motivation letter (with contact information of 
up to 3 persons accepting to be asked for recommendation letters) to be 
sent jointly to:*

Prof. Dr. David Ginsbourger: david.ginsbourger using stat.unibe.ch 
<http://mailto:david.ginsbourger@stat.unibe.ch/>

and Dr. Juan Herrera: juan.herrera using agroscope.admin.ch 
<http://mailto:juan.herrera@agroscope.admin.ch/>


	[[alternative HTML version deleted]]