Multivariate Approaches for Soil Fertility Characterization of Lower Brahmaputra Valley, Assam, India Reza S.K.*, Baruah Utpal1, Singh S.K.2 ICAR-National Bureau of Soil Survey and Land Use Planning, Sector-II, DK-Block, Salt Lake, Kolkata, West Bengal 1ICAR-National Bureau of Soil Survey and Land Use Planning, Jamuguri Road, Jorhat, Assam 2ICAR-National Bureau of Soil Survey and Land Use Planning, Nagpur, Maharashtra *Corresponding author Email: reza_ssac@yahoo.co.in
Online published on 14 July, 2016. Abstract The aim of this study was to characterize the soil fertiltiy like pH, organic carbon (OC), available nitrogen (N), available potassium (P) and available phosphorus (K), and DTPA extractable iron (Fe), manganese (Mn), zinc (Zn) and copper (Cu) of lower Brahmaputra valley of Assam using multivariate statistics (principal component, correlation matrix and cluster analysis). A total of 2753 soil samples from a depth of 0–25 cm at an approximate interval of 1 km were collected from Barpeta, Bongaigaon and Nalbari districts of Assam. Soil properties showed large variability with greatest variation was observed in DTPA-Zn (120%), whereas the smallest variation was in pH (17.5%). The principal component analysis (PCA) applied on the investigated soil properties identified three components with eigen values greater than 1, which explained 65% variability and same grouping was also obtained from cluster analysis. Cluster 1 includes Zn, Cu, pH, Fe, P and Mn, which has highest loading in PC1 and also showed strong significant relationship. Cluster 2, which contains OC and N, and had highest loading in PC2 and also showed the significant positive relationship with each other. Cluster 3 contain only K, which is equally distributed both in PC1 and PC2 and also significantly and positively correlated with pH, Fe, Zn and Cu of cluster 1 and OC and N of cluster 2. Top Keywords Lower Brahmaputra valley, soil properties, principal component analysis, correlation matrix, cluster analysis. Top |