Large-scale yield gap estimation and characterization with multisource remote sensing data – Case study of rainfed wheat in Ethiopia

My name is Habtamu Sewnet Gelagay, I am attending my PhD in EFSA-Agronomic Sciences at GAIA doctoral school, SupAgro institute, university of Montpellier. I have been hosted at CIRAD, UR AIDA, Montpellier, France for the last six months, and now I am seconded to IITA ,Nairobi, Kenya. I’m from Ethiopia and my background is around the use of geospatial technologies for agricultural and environmental monitoring. Before starting this PhD I have been working for several years in the private and public sector in Ethiopia (Ethiopian Space Agency, GIZ, World Bank, CIAT etc). With my ever-growing research interest, I am strongly convinced to start a PhD to advance my research work, build enough confidence to solve scientific problems, and pursue an academic career.
I’m working on large-scale yield gap estimation and characterization with multi-source remote sensing data – Case study of rainfed wheat in Ethiopia. In Ethiopia, wheat yield has improved in recent years, but farmers harvested only 20% of the potential yield. To address food security and environmental pressures, we need spatially explicit information on the yield gap and its determinants. However, conducting large-scale spatial analysis of the yield gap has been challenging due to the lack of robust frameworks. This Ph.D. aims to enhance existing methodologies by leveraging Earth observation, geospatial technologies, and field data to quantify and characterize the yield gap of rainfed wheat in Ethiopia. The objectives include developing a methodology for agroecological spatial unit delineation, improving large-scale yield estimation, and conducting a comprehensive analysis of the yield gap. This research will utilize multi-source remote sensing, geospatial technologies, and the analytical capabilities of the Google Earth Engine platform.

The worthwhileness and relevance of the PhD research project for data poor region and countries like Ethiopia interests me to undertake the thesis. To this end pursuing this PhD program will have a wonderful societal influence and decision supporting role by precisely figuring the gap between actual and potential yield along with proper characterization of the causes of the gaps. Additionally, the study could be an exemplary in the area by filling the gaps between the havoc of plenty of earth observation data and decision-making through an intelligent and spatio-temporally deepened analysis of rainfed wheat yield gap. Overall, the topic presents an opportunity to make a valuable contribution to both the academic community and society as a whole. This Ph.D. will also offer multiple benefits for my professional and personal growth, including expertise and knowledge, intellectual growth, career advancement, and networking opportunities.

Currently, my research focuses on mapping rain fed wheat crop types and delineating Agro-Ecological Spatial Units to scale up the analysis of wheat yield gaps. I have obtained the following preliminary results which will be presented at the IALE 2023 conference :1). Developed a national scale map of rain fed wheat crop land in Ethiopia, which includes information on crop rotation conditions 2). Established a methodological framework for the delineation of Agro-Ecological Spatial Units.

To obtain the above results we incorporate  time series multisource remote sensing data, along with other biophysical and environmental variables through leveraging advancements in data science and machine learning within the Google Earth Engine cloud computing platform.
Regarding the prospects for my thesis, these initial results are promising and lay the foundation for further analysis. As I continue my research, I expect to gain more comprehensive insights into the wheat yield gap and contribute to addressing food security challenges in Ethiopia.

• Starting date: 10 October
• Research unit: Aïda
• University: Institut Agro
• PhD school: GAIA
• Scientific field:  Agrnomic sciences and digital agriculture
• Thesis supervisors:  Marc Corbeels and Louise Leroux, UMR Aïda, Cirad
• Funding : Cirad - OneCGIAR EiA Initiative
• #DigitAg : Axe 6 : Modélisation et simulation (systèmes de production agricole), Axe 5 : Fouille de données, analyse de données, extraction de connaissances, Challenge 1 : Le challenge agroécologique, Challenge 8 : Développement agricole au Sud

Keywords: Yield gap, multi-source remote sensing, Google Earth Engine, Smallholder agriculture, Ethiopia

Abstract: In Ethiopia; despite wheat yields improvement observed in recent years, the current wheat yield is only 20% of its potential. To be able to develop sustainable production systems that allow to address food security issue while decreasing environmental pressures, spatially explicit information on yield gap and its determinants are needed. However, spatially explicit yield gap (YG) analysis at large scale has long been a challenging exercise due to the lack of reproducible and robust spatial frameworks. This PhD will address issues related to the improvement of existing methodologies to quantify and characterize the yield gap of rain fed wheat in Ethiopia. We hypothesize that this challenge can be met with the new Earth observation and geospatial technologies combined with field data. The goal of this Ph.D. research study is to assess the yield gap of rain-fed wheat in Ethiopia on a large scale, using a data-driven approach that integrates advanced data science, machine learning, and up-to-date remote sensing from multiple sources, with analysis conducted on the Google Earth Engine cloud computing platform. Specifically, this PhD aims to achieve the following objectives: 1) to develop a methodology for the delineation of agroecological spatial unit (ASU) for scaling up rain-fed wheat crop YG analysis. Under this objective, accurate and reliable rain-fed wheat crop land will be produced, and will serve as a stepping point for ASU delineation and YG analysis. 2) to improve large scale actual rain-fed wheat yield estimation and 3) to spatially deepen the YG analysis through (1) a temporal perspective by disentangling the persistent YG from the transient YG and (2) a structural perspective by decomposing the YG in to technical efficiency, resource and technological YG.

Contact: habtamu_sewnet.gelagay [AT] cirad.fr
Social networks: Google Scholar - ORCID