Principal component and cluster analysis of coffee production environments in the Gedeo Zone of southern Ethiopia

Abstract

Despite the importance of coffee, there is a lack of comprehensive analysis to understand the key factors influencing coffee production across the Gedeo Zone of southern Ethiopia. We used principal component and cluster analysis approaches to minimize the dimensionality of the data gathered from the study area's topography, climate, soil analysis, farmers' interviews, and geographic information systems. Data from coffee farm records spanning ten years (2013−2022) were collected from 18 rural kebeles in the area. The first 12 principal components explained 95.4% of the total variation. The cation exchange capacity of soils contributed 22.5% of the variation in the first principal component, while evapotranspiration and shade trees explained 16.1% and 11.9%, respectively of variability in coffee productivity. Total nitrogen, altitude, and ash explained 10.5%, 8.1%, and 5.8% of variability whereas organic carbon, iron, soil water conservation, Normalized Difference Vegetation Index, variety, and clay explained 5.0%, 4.0%, 3.3%, 3.2%, 2.9%, and 2.1%, respectively. The 18 kebeles were reduced to five clusters (Tumata chiracha, Hama, Dumerso, Konga, and Wotiko), which differed for all characteristics (p < 0.001), and Konga and Wotiko clusters concentrated 39.0% (2.184) of the information with higher productivity of coffee. Similarly, the 58 variables were reduced to cation exchange capacity, clay content of soils, total nitrogen, evapotranspiration, shade trees, altitude, ash, organic carbon, iron, soil water conservation, NDVI, and variety, which were considered the most representative traits explaining the variability of the data set. Thus, farmers, agricultural planners, and policymakers shall embark on a different set of planning schemes, varietal choices, and management practices that maximize the quality and yield of coffee for each cluster.