Abstract

The genus Eragrostis tef Zuccagni is commonly known as “Teff”, is an indigenous cereal crop and is the major staple food crop in Ethiopia. It is mostly used to prepare a spongy flatbread called “Injera” and is consumed by more than 70% of the Ethiopian people. This study is conducted at nine Teff-dominated zones of the country to examine whether geospatial technology can serve to estimate the productivity of crop yield. For this, ground truth sample plots were used for nine zones, and geospatial technology and machine learning were applied for upscaling to the whole study area’s scale. Very good correlation results were obtained from spatial predictions of Teff yield for 2015 and 2020 with ROC-AUC of 89 and 91% and R2 of 0.67 and 0.73, respectively. The average predicted yields of Teff were about 1.37 t/ha and 1.99 t/ha for 2015 and 2020, respectively, indicating that such technology can offer a very good result to estimate yields for unreachable areas in the case of either during unfavorable political or other natural conditions. By doing so, we can plan to apply such technologies that can serve to save time, effort, and resources.

This work is licensed under a Creative Commons Attribution 4.0 License.