Climate-smart agriculture in sub-Saharan Africa: optimizing nitrogen fertilization with data science

Summary

Food insecurity in sub-Saharan Africa is widespread, with crop yields much lower than in many developed regions. Increased fertilizer application in the coming decades could improve yields, alleviating food stress; however, this can cause major increases in emissions of greenhouse gases such as nitrous oxide (N₂O). The N2O-SSA project aims to use laser spectroscopy to measure fluxes and isotopic composition of N₂O from maize and potato crops subjected to a range of fertilization levels. The data generated will allow us to develop a model to predict N₂O emissions under different climate and management scenarios, to guide future measurements, reduce uncertainty in these predictions, and identify potential avenues for sustainable agricultural development in sub-Saharan Africa.

The need for climate-smart agriculture

Food security is a major issue in sub-Saharan Africa, with 25-40% of households classified as food insecure, meaning they lack regular access to sufficient, safe, and nutritious food for growth, development, and an active, healthy life [1,2]. In recent decades, agricultural production in sub-Saharan Africa has rapidly increased; however, the majority of production increases come from the expansion of cultivated area, while yield per area (productivity) remains low. One of the major causes of low productivity is insufficient use of nitrogen fertilizer, with application rates often 10 times lower than recommended due to high costs and lack of access to fertilizer. In order to increase agricultural productivity, fertilizer use in sub-Saharan Africa is predicted to grow in the coming decades. Climate-smart agriculture requires careful fertilizer application control, minimizing greenhouse gas emissions from agricultural soils while ensuring high productivity and food security. However, the success of climate-smart agricultural practices depends on detailed knowledge of nutrient cycles in soils.

Nitrous oxide: A potent greenhouse gas and an indicator of agricultural nitrogen use efficiency

Nitrous oxide (N₂O) is a powerful greenhouse gas and the most potent stratospheric ozone-depleting substance released this century [3]. The largest anthropogenic source of N₂O is microbial production in agricultural soils following the addition of nitrogen (N) fertilizer. This source accounts for 65% of Switzerland’s total N₂O emissions [4], and while anthropogenic N₂O emissions are a direct environmental threat, N₂O is also a marker for N waste in agriculture. To reduce agricultural N₂O emissions, farmers must maximize the transfer of fertilizer N into crops, known as Nitrogen Use Efficiency. This can be achieved through climate-smart agricultural management practices, which aim to precisely match the needs of crops with management practices such as fertilizer application and timing.

The N2O-SSA project

We launched the N2O-SSA project with the full title “Combining measurements, modeling and machine learning to improve N₂O accounting for sustainable agricultural development in sub-Saharan Africa” in 2021. The project aims to measure N₂O emissions and isotopic composition from maize and potato crops over a range of fertilizer application levels. We will use data science techniques to understand drivers of N₂O emission peaks and construct modeling approaches to predict N₂O emissions in the coming decades. These models will provide a basis to develop sustainable agricultural practices that balance low greenhouse gas emissions with high yields while fostering soil and environmental health. 

N2O-SSA is funded by the Swiss National Science Foundation and led by the Swiss Data Science Center, in partnership with the Sustainable Agroecosystems research group at ETHZ, the Department of Soil Science at the University of Eldoret, and the International Livestock Research Institute.

Measuring nitrous oxide fluxes and isotopes using laser spectroscopy

Measuring the flux of a greenhouse gas such as N₂O tells us how much gas is emitted from or taken up by the soil; however, this does not tell us about the processes responsible for the gas production and consumption. In contrast, the isotopic composition of the gas (see Figure 1) can act like a fingerprint for different emission processes. In the case of N₂O, the isotopic site preference (SP) is particularly useful for distinguishing between the major microbial pathways of nitrification (high SP) and denitrification (low SP).

Figure 1: The four most common isotopocules of N₂O are ¹⁴N¹⁴N¹⁶O, ¹⁴N¹⁵N¹⁶O (𝛼), ¹⁵N¹⁴N¹⁶O (𝛽), and ¹⁴N¹⁴N¹⁸O. The isotopic site preference (SP) is the difference between the isotopic composition at the two sites and the bulk isotopic composition is the average of the two sites.

Fluxes of greenhouse gases like N₂O can be monitored in the field using chambers, which close over the soil to trap gases that are released from the soil (Figure 2). 

Figure 2: Chambers to monitor N₂O flux and isotopic composition are installed on soil columns at ETHZ’s Eschikon research site (© Turry Ouma).

We can measure the increase of N₂O in the closed chamber to estimate the flux (Figure 3a): When N₂O builds up quickly, this indicates a high flux. Using laser spectroscopy, we can also measure the isotopic composition of N₂O in the chamber headspace, because the different isotopocules absorb laser light at different wavelengths. As shown in Figure 3b, we can construct a time series of the flux and isotopic composition of emitted N₂O through repeated chamber measurements.

Figure 3: Measuring N₂O flux and isotopic composition. a) When the chambers close, N₂O builds up in the chamber headspace. When the chamber is closed, measurements of N₂O concentration (upper) and isotopic composition (lower) allow us to estimate the flux and the isotopic composition of emitted N₂O. b) Repeated measurements are used to construct a time series of flux and N₂O isotopic composition. This example shows a strong N₂O peak at around 17:00. The low site preference indicates the peak was due to microbial denitrification.

Although the example shown in Figure 3 looks simple, interpreting these isotopic datasets is challenging. The TimeFRAME tool, recently developed at SDSC, is an easy-to-use Bayesian hierarchical modeling approach to quantify the pathways producing and consuming trace gases based on isotopic measurements. You can find more information about TimeFRAME on the project repository and in an upcoming publication.

Pilot measurement campaign in Eschikon, Switzerland

In June 2023, we installed a laser spectrometer coupled to a preconcentration device interfaced with twelve automated chambers (Figure 4a,b) at the ETHZ Research Station for Plant Sciences in Eschikon. During this pilot measurement campaign, we aim to test the complete measurement setup, troubleshoot for potential problems, and develop automated data management and analysis capabilities. The pilot campaign has been a great learning and problem-solving opportunity for PhD student Turry Ouma and postdoctoral researcher Phillip Agredazywczuk, as field measurements with cutting-edge instrumentation rarely run smoothly. Having faced many problems with the system ranging from laser meltdowns to power issues, the system is now finally running (nearly) successfully four months after installation.

Outlook: Nitrous oxide measurements in Eldoret, Kenya

Following shipping to Kenya at the end of 2023, the instrumentation will perform measurements at an experimental farm site at the University of Eldoret (Figure 5). 

We aim to measure fluxes and isotopic composition from maize and potato plots with various fertilizer application levels over an entire growing season. This data will help us to understand the relationship between fertilizer application rate and N₂O emissions. Using data from other regional sites to extrapolate, we will develop a model to predict cropland emissions under different climate and management scenarios. A key focus will be understanding the sources of uncertainty in our predictions and guiding future measurements to reduce these uncertainties.

Feel free to reach out if you are interested in the project. 

Co-authors

• Turry Ouma, Swiss Data Science Center, ETH Zürich
• Philip Agredazywczuk, Swiss Data Science Center, ETH Zürich

References

1. S. Leitner et al. Closing maize yield gaps in sub-Saharan Africa will boost soil N2O emissions. Current Opinion in Environmental Sustainability, 47:95–105, 2020.
2. FAO, IFAD, UNICEF, WFP, WHO. The State of Food Security and Nutrition in the World, 2018: Building Climate Resilience for Food Security and Nutrition. Rome: Food & Agriculture Org., 2018.
3. H. Tian et al. A comprehensive quantification of global nitrous oxide sources and sinks. Nature, 586:248–256, 2020.
4. FOEN. Switzerland’s Greenhouse Gas Inventory, 1990–2021. Technical report, 2023.