The development of a Decision Support Tool (DST) for fertilizer recommendations tailored to Ethiopia's farming system is crucial for boosting agricultural productivity and advancing the country toward precision agriculture. In pursuit of this objective, the Ethiopian Institute of Agricultural Research (EIAR), regional research institutes, and international collaborators have been engaged in ongoing research to create a machine learning-based DST model. This tool is designed to provide farmers with accurate, data-driven fertilizer recommendations, specifically tailored to individual farms or plots of arable land.

Dr. Birhanu Agumas, Senior Researcher at the Amhara Agricultural Research Institute (ARARI) and a member of the national research team’s technical committee working on the model-based DST, outlined the research process and its goals. He emphasized that, over the past three years, extensive research has been conducted to modernize fertilizer recommendations nationwide. The results are currently undergoing review at both national and regional levels.

Dr. Birhanu explained that the DST is based on machine learning algorithms, which utilize legacy data spanning several years, along with data from nutrient omission experiments and other fertilizer rate trials conducted over the past three or more years. The DST integrates crop response data from field experiments, as well as climate, soil, and topographic data, which serve as key variables. He further emphasized that the model was developed using comprehensive data representing Ethiopia’s major crop-growing areas, ensuring that it is relevant across diverse agricultural zones in the country.

Recalling the origins of the DST development process, Dr. Birhanu noted that international research institutions such as Alliance Bioversity International – CIAT, the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), and the International Maize and Wheat Improvement Center (CIMMYT) had previously developed their own fertilizer recommendation models. However, these models were not adopted by Ethiopia’s Ministry of Agriculture, EIAR, or regional research institutions. Dr. Birhanu explained that introducing multiple models would cause confusion among smallholder farmers, who need clear and unified guidance. As a result, the various institutions decided to collaborate on creating a single, harmonized DST that aligns with Ethiopia’s soil and climate conditions and meets the specific needs of local farmers.

In response, a collaborative team of researchers, data scientists, and modellers from national, regional, and international institutions came together to design the machine learning-driven DST. After extensive discussions, analysis, and data integration, the harmonized DST has been successfully developed and is now being tested on farmers' fields across the country.

ARARI, as a key regional research institute, has played a significant role in executing field experiments, conducting data analysis, and contributing to the development of the DST through its research centers in Adet, Sirinka, Gonder, Sekota, and Debirbirhan. These centers provided valuable data that fed into the model's development, strengthening its applicability and accuracy.

Dr. Birhanu concluded that the DST is currently undergoing rigorous validation on farmers’ fields throughout Ethiopia. After completing this validation and piloting phase, the tool will be officially rolled out. It will then serve as an essential resource for farmers, policymakers, agricultural experts, and researchers, providing them with reliable, data-driven recommendations on fertilizer use. This will ultimately enhance agricultural productivity and support sustainable farming practices across Ethiopia.