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

July 1, 2022
In Progress
Share this project


Fertilizer use in sub-Saharan Africa is currently low, which results in lower agricultural crop yields than in many developed nations. In the coming decades, fertilizer use is predicted to increase, in order to drive increasing agricultural productivity. If fertilizer application is carefully controlled, emissions of greenhouse gases from agricultural soils can be minimized while ensuring high productivity and food security: Climate-smart agriculture. The success of climate-smart agricultural practices depends on detailed knowledge of nutrient cycles in soils.

Nitrous oxide (N2O) is a potent greenhouse gas that makes a significant contribution to global warming. In addition, it is the most important substance currently released that contributes to the destruction of the ozone layer. N2O is produced by microbes in soils as they use nitrogen in two essential biochemical reactions: Nitrification and denitrification. Agricultural fertilization increases the amount of nitrogen available in soils, which leads to more significant emissions of N2O. N2O emissions have primarily been studied in temperate ecosystems, but the causes and magnitude of emissions in sub-Saharan Africa are not well known. Expected changes in N2O emissions due to climate change and increasing fertilizer use are very poorly constrained, making it particularly difficult to drive investment in climate-smart agriculture in this region.

In this project, we will use a novel measurement technique – laser spectroscopy – to study N2O emissions from agricultural soils in Kenya. Laser spectroscopy allows us to directly monitor the isotopic composition of N2O emitted from soils, and thus infer the microbial N2O production and consumption pathways occurring in the soils. We will monitor N2O emissions and isotopic composition through a dry-wet seasonal transition to understand how climate and soil moisture control the production of N2O. We will relate our results to observations of N2O emissions made at other sites in sub-Saharan Africa, and use machine learning techniques to develop a new model of N2O emissions. This will allow us to understand and predict emissions and thus contribute to the development and uptake of sustainable agricultural practices.



SDSC Team:
Eliza Harris
Fernando Perez-Cruz

PI | Partners:


  • Prof. Johan Six
  • Dr. Matti Barthel

More info

University of Eldoret, Kenya:

  • Dr. Abigail Otinga
  • Dr. Ruth Njoroge

More info

Funding Agency

  • Swiss National Science Foundation (Project Number 200021_207348)

More info



Drivers of N2O emissions in tropical soils are poorly known, making it difficult to accurately predict emissions in a changing climate.


We will advance our understanding of the major drivers of variability in tropical N2O fluxes and thus make improved predictions of N2O emission strength and variability in the coming decades.


In N2O-SSA, we will use a combined measurement-data science-modeling approach to understand the drivers of soil N2O emissions, in particular the influence of precipitation and soil moisture.

Specific project tasks include:

  • Develop TREX-QCLAS-chamber system for a high precision monitoring of N2O fluxes and emitted N2O isotopic composition from soils.
  • Conduct a 6-month measurement campaign in Eldoret, Kenya, to understand soil N2O sources and sinks across a dry-wet seasonal transition.
  • Compile N2O flux and isotope data from studies in sub-Saharan Africa and other tropical regions, and – in combination with campaign measurements – analyze data to understand the drivers of N2O production and consumption.
  • Use data science approaches for modeling and spatiotemporal upscaling, in order to predict future emissions and promote the development of targeted mitigation strategies.


Figure 1: Overview of the key ideas underlying N2O-SSA.

  1. Predicted growth in direct N2O emissions from increasing fertilizer application between 1995 and 2045 made using the IsoTONE model (Harris et al. 2022). The increased emissions are based only on increases in fertilizer use predicted in the Land Use Harmonization Database, therefore potential feedback due to climate change is not included in this estimate.
  2. Isotopic composition of N2O from different microbial sources. AOBs = Ammonia Oxidizing Bacteria. Isotopic signatures from Sutka et al. (2004, 2006), Toyoda et al. (2005, 2011), Wunderlin et al. (2013), Heil et al. (2014), Harris et al. (2015) and Denk et al. (2017).
  3. Simulated and measured spectra for dual-laser QCLAS monitoring of the 4 major isotopocules of N2O. Blue points indicate the laser measurements and the red line shows the fitted spectrum used to determine the mixing ratios of the isotopomers (see Harris et al. (2014, 2017) for further details).


Additionnal resources


  1. Harris et al. (2022) Warming and redistribution of nitrogen inputs drive an increase in terrestrial nitrous oxide emission factor, Nature Communications, doi: 10.1038/s41467-022-32001-z
  2. Barthel et al. (2022) Low N2O and variable CH4 fluxes from tropical forest soils of the Congo Basin, Nature Communications, doi: 10.1038/s41467-022-27978-6
  3. Harris et al. (2021) Denitrifying pathways dominate nitrous oxide emissions from managed grassland during drought and rewetting, Science Advances, doi: 10.1126/sciadv.abb7118


Related Pages

More projects


In Progress
Machine Learning for the Future Circular Collider Design
Big Science Data


In Progress
Real-time cleansing of snow and weather data for operational avalanche forecasting
Energy, Climate & Environment


AI-augmented architectural design
Energy, Climate & Environment


In Progress
Extracting activity from large 4D whole-brain image datasets
Biomedical Data Science


Latest news

PassGPT | Using language models to enhance password security
February 6, 2024

PassGPT | Using language models to enhance password security

PassGPT | Using language models to enhance password security

PassGPT is a Large Language Model for password generation trained on leaked passwords, which can outperform existing methods based on generative adversarial networks by guessing twice as many unseen passwords.
ADORE | A benchmark dataset in ecotoxicology to foster the adoption of machine learning
January 24, 2024

ADORE | A benchmark dataset in ecotoxicology to foster the adoption of machine learning

ADORE | A benchmark dataset in ecotoxicology to foster the adoption of machine learning

Applying machine learning to ecotoxicology could help reduce the number of animal tests, costs, and animals sacrificed while preserving the accuracy of the in vivo tests.
License Flowers | Art and AI at SDSC
February 21, 2024

License Flowers | Art and AI at SDSC

License Flowers | Art and AI at SDSC

An adventure to create art using AI to raise awareness on code licenses

Contact us

Let’s talk Data Science

Do you need our services or expertise?
Contact us for your next Data Science project!