Principal component and cluster analyses based characterization of maize fields in southern central Rift Valley of Ethiopia

Abstract

Abstract Soil management decisions should consider physical potential of the environment, weather variability, and requirements of crops to maximize production to the potential limits. This calls for characterization of environments using selected input variables. Such studies are scanty in southern central Rift Valley of Ethiopia due to which the area is considered homogeneous and identical for agricultural planning, extension, and input delivery programs. Thus, to investigate the scenario, we employed principal component, clustering, and GIS analysis on geo-referenced physiographic and climatic attributes, and their statistical variables obtained from 43 stations with the objective of identifying homogeneous management units with similar physiography, weather pattern, and production scheduling. The analysis of principal components (PCs) indicated that three PCs explained 74.7% of variance in October, November, December, and January (ONDJ), four PCs explained 79.3% of variance in February, March, April, and May, and four PCs explained 80.5% of variance in June, July, August, and September (JJAS). Cluster-I was characterized by high altitude and low temperature in ONDJ season. Cluster-II was characterized by low altitude and high temperature across most seasons. Cluster-III was intermediate in altitude, temperature, and rainfall. Cluster-IV was characterized by high rainfall in JJAS. In all the clusters, PC1 was the mean rainfall component with strong association with altitude and longitude, while PC2 was the temperature component. PC3 is the statistical component with strong influence from mean rainfall. Thus the factors that determine the formation of clusters are reduced from 12 to 5 (T mean, latitude, longitude, altitude, and RFmean) and 43 stations are grouped into 4 clusters (Shamana, Bilate, Hawassa, and Dilla) which are geographically and ecologically distinct. These clusters require different sets of agro-meteorology advisory, maize management, and input delivery strategies.