Postdoc in Data Science / Biogeochemistry in Zürich

The  (SDSC) is a joint venture between EPFL and ETH Zurich. Its mission is to accelerate the adoption of data science and machine learning techniques within the ETH Domain academic disciplines, the Swiss academic community at large, and the industrial sector. In particular, it addresses the gap between those who create data, those who develop data analytics and systems, and those who could potentially extract value from it. The centre comprises a multi-disciplinary team of data and computer scientists, and experts in select domains with offices in Zürich, Lausanne, and Villigen.

We are seeking a postdoc to work on the SNF-funded N2O-SSA project as well as on related SDSC research projects. The postdoc will be based at the Swiss Data Science Centre in Zürich and will work closely with the Department of Environmental Systems Science - in particular the Sustainable Agroecosystems Group (https://sae.ethz.ch/) - as well as the University of Eldoret, Kenya (). 

The position is available at 80-100% FTE for 3-4 years, ideally starting in early 2023. 

N2O-SSA: Combining measurements, modelling and machine learning to improve N2O accounting for sustainable agricultural development in sub-Saharan Africa 

N2O is a strong greenhouse gas emitted during microbial N cycling in soils, particularly N-fertilised agricultural soils. Fertiliser use in sub-Saharan Africa is currently relatively low, and increased fertilisation to drive higher agricultural productivity is predicted in the coming decades. There is great potential for climate-smart agricultural techniques in this region to ensure food security while minimising environmental impacts and greenhouse gas emissions. However, financial investment in sustainable agriculture in sub-Saharan Africa is very low, in part due to the scarcity of data, in particular regarding N2O fluxes and emission factors. This project aims to develop and implement a novel measurement technique to monitor N2O fluxes and isotopic composition, and thus infer N2O production and consumption pathways for the first time in sub-Saharan Africa. We will investigate various computational approaches to simulate N2O fluxes across sub-Saharan Africa and understand drivers and causes of uncertainty, with a particular focus on causal inference. We will develop an optimised modelling framework to allow robust predictions of future N2O emissions in a changing climate, to contribute to closing the data gap and fostering sustainable agricultural development in this region.

Further information about the project can be found on the website: 

This project will be divided into four parts: i) Development and testing of a coupled chamber-preconcentration- isotopic measurement system in Switzerland, including automated data analysis software to facilitate rapid quality checking and troubleshooting; ii) Deployment of the system at an experimental field site in Eldoret, Kenya, to monitor N2O flux and isotopic composition through a dry-wet seasonal transition for two crop types and two fertilisation levels, in close collaboration with project partners at the University of Eldoret. 

The first two parts of the project will be conducted mainly by the project PhD student, with guidance and collaboration with the postdoc. The advertised postdoctoral position focuses on the third and fourth parts of the project: iii) The application of machine learning to improve current N2O flux models, in particular IsoTONE (https://doi.org/10.1038/s41467-022-32001-z), to characterise uncertainty and improve computational efficiency, and thus allow robust predictions in a range of conditions; and iv) To investigate causal relationships between environmental parameters and N2O fluxes in tropical soils, and thus understand drivers of variability in N2O production pathways, using techniques such as causal invariance algorithms and a customized flux generative model for robust predictions of N2O. 

This interdisciplinary postdoc is a unique combination of biogeoscience and machine learning methodologies, thus offering a unique development platform for the postdoctoral researcher. 

The wider impact of this project will result from model development, which will provide new, robust data for N2O fluxes and emission factors in sub-Saharan Africa, and efficient modelling tools that can be used to investigate the impacts of climate change as well as management and mitigation policies on emissions. Furthermore, the project will strongly contribute to education and capacity building, strengthening ties between research communities in Europe and sub-Saharan Africa. 

SDSC projects: 

The SDSC works on research projects together with scientists from a diverse range of fields within the ETHZ and EPFL domain (). In addition to N2O-SSA, the postdoc will have the opportunity to work with SDSC data scientists on 1-2 ongoing SDSC projects, in particular: 

• MACH-Flow (external lead: Dr Lukas Gudmundsson, Prof. Dr Sonia Seneviratne, ETHZ): River flow is a key component of the terrestrial water cycle, and dependable information on river flow is essential. However, observation networks only monitor a small number of locations, leaving many gaps on maps of regional river flow. MACH-Flow aims to advance our capabilities for estimating daily river flow at ungauged locations in Switzerland using data science methods, with a special focus on making existing methods fit for application at the national scale at very high spatial resolution. 
• WATRES (external lead: Dr Paolo Benettin, EPFL): Rivers often react quickly to rainfall events, which can cause water quantity problems like floods; however, rivers also transport significant amounts of ‘old’ water, which is stored for years and strongly influences water quality. Quantifying the timing of watershed responses is complex because they can be nonlinear and time-variable, and may take unpredictable irregular shapes. The sensor revolution now provides both flow and tracer measurements at high resolution, and the goal of this project is to develop a new data-driven methodology to estimate the timing of watershed responses. 
• PHENO-MINE (external lead: Prof. Dr Achim Walter, ETHZ): Adapting field crops to a changing climate requires a detailed understanding of the relationship between growth dynamics and the environment. Analyzing phenotype data currently involves extracting low-level features from images, and relating dynamics in these features to target traits such as drought tolerance. Consequently, unforeseen associations between growth dynamics and target traits can be overseen. This project aims to use deep learning to better interpret phenotypic data, and thus gain a better understanding of growth dynamics and responses to environmental changes. 
• CLIMIS4AVAL (external lead: Prof. Dr Jürg Schweizer, SLF): Avalanche forecasting relies on snow, snow cover and weather data; measurement errors, anomalies and data gaps diminish forecast accuracy. Automated location-based avalanche forecasts, therefore, require real-time cleaning of online measurement data. This project focuses on real-time detection of anomalies in snow and weather time series data, as well as detection of outliers, and imputation of missing data using state-of-the-art machine learning approaches. 

• You will be responsible for advanced data analysis, data compilation, and model development and implementation within the N2O-SSA project 
• You will work together with N2O-SSA project partners in Switzerland and Kenya to achieve project goals, exchange knowledge and ideas, and synthesise and disseminate results 
• You will contribute to ongoing SDSC academic research projects together with other data scientists at SDSC, to use your knowledge in these projects, and build your skills in new data science approaches and methods
• You will participate in meetings, seminars and other activities within SDSC and USYS at ETHZ 
• You will have the opportunity to contribute to teaching activities at ETHZ if desired 

You must have a PhD in biogeochemistry, environmental sciences, machine learning, mathematics, or a related subject, and experience with advanced data analysis using python, R or similar. Ideally, you have conducted interdisciplinary work employing data science approaches in environmental sciences, or you have a strong interest in developing this profile. 

Regardless of your scientific background, you must be motivated to work in a challenging, cutting-edge research field, to work independently and develop your ideas, and collaborate with others from a range of backgrounds. Fluency in English (written and spoken) and strong communication skills, in particular experience in presentation and publication of research results at conferences and in peer-reviewed journals, is required. 

• A stimulating, collaborative, cross-disciplinary environment in a world-class research institution; 
• Flexible work arrangements, including remote working, flexible time, condensed week, and the opportunity to work part-time; 
• Exciting challenges, varied projects, and plenty of room to learn and grow; 
• An opportunity to follow your passion and use your skills to make an impact on research communities and society; 
• A possibility to spark your creativity by experimenting and learning new technologies;

We value diversity: In line with our values, ETH Zurich encourages an inclusive culture. We promote equality of opportunity, value diversity and nurture a working and learning environment in which the rights and dignity of all our staff and students are respected. Visit our Equal Opportunities and Diversity website to find out how we ensure a fair and open environment that allows everyone to grow and flourish.