Use of cluster analysis for classification of Benchmark soil samples from India in different micronutrient availability groups

Publisher Website
Google Scholar
Abstract
Summary: In view of the increase in micronutrient deficiencies in crop plants, ans attempt is made to group 57 Benchmark Indian soils into different micronutrient availability classes (clusters) vis-à-vis some soil characteristics. Since most of the soil characters are expressed in proportions or percentages (a usual practice in soil studies), the data matrix was transformed into log10 values to bring it nearer to normality. The transformed data matrix was used for cluster analysis and subsequently also for discriminant analysis. By following the method of Euclidean cluster analysis, the 57 Benchmark soils could be subdivided into three clusters. The distinctness of clusters was proven by distance-matrix as well as discriminant analysis. All the soils of cluster I and II originated from arid and semi-arid climates and these were alkaline in reaction, low in organic carbon (OC) and high in total lime. On the contrary, cluster III comprised all but a few soils from humid-subhumid regions. These soils were acidic in reaction and relatively high in OC and total lime. The soils of cluster I and II were poorer in diethylene triamine penta-acetic acid (DTPA) extractable micronutrient cations than those of cluster III. This study thus reveals that it is possible to classify diverse soils by statistical means into distinct clusters based upon micronutrient availability and associated soil properties. Furthermore, it suggests that soils of arid and semi-arid climates as a group, because of their low micronutrient availability, are expected to be more prone to deficiencies than those of humid and subhumid zones.

This publication has 4 references indexed in Scilit: