Multivariate analysis of the physico-chemical properties of soils in selected locations of the floodplain of river Kaduna in Niger state, Nigeria

Abstract

Multivariate statistics was used to analyse the physico-chemical properties of soil in the selected locations of the floodplain of River Kaduna in Niger State, Nigeria. Samples were collected in March and September of the same year and the levels of some important soil quality parameters analysed. The data was subjected to basic statistics, cluster analysis (CA), Principal Component Analysis (PCA), Correlation Analysis, and Discriminant Analysis (DA). Descriptive statistics shows that, with the exception of silt and clay other parameters analysed decreased in value after flooding, while the predominant textual class of the soil is clay-loam. The PCA for March and September sampling periods extracted three components each which explained 84.10% and 78.90% respectively of the total variance. Cluster analysis yielded five distinct clusters for March: cluster 1 (K, Ca, EA, and pH); cluster 2 (OM, CEC, TN, Silt content, Clay content, and OC); cluster 3 (Na); cluster 4 (P), and cluster 5 (Mg and Sand content); and five separate clusters for September. From DA, seven variables (pH, OC, OM, TN, P, K and Mg) were the most significant parameters which accounted for the expected spatial/temporal variations in the soil of March, while CEC, Na, Ca, silt, sand and clay become an important discriminator in September. Flood has significant influence on the soil of the study area as level and distribution of the parameters changed from March to September.