Abstract
Multivariate profiling of soil bacterial communities provides information that will possibly contribute to the conservation and rehabilitation of the environment. In the Sakaerat Environmental Research Station, Thailand, the land degradation gradient, represented by a dry evergreen forest (DEF) and bare ground (BG) as a result of deforestation and subsequent land degradation was determined. Profiling was done with the antibiotic resistance most probable number (MPN) method (method 1), the antibiotic disc diffusion method (method 2) and the Biolog method. Compread to the DEF soil, the BG soil had significantly lower moisture content and pH values (at p=0.05), and total C and bacterial MPN count (at p=0.10). Profiles of the soils obtained by the antibiotic and the Biolog methods were analyzed. The profiling methods scored Wilk’s lambda values of 0.001 (method 1, p=0.086), 0.004 (method 2, p=0.157) and 0.000 to 0.110 (Biolog, p=0.040 to 0.708 at 0.2 to 1.4 average well color development values, AWCDs), then gave principal component score plots showing that all the methods successfully determined the degradation gradient with comparable efficiency. Redundancy analysis ordination diagrams for the above data sets indicated that the soil environmental factors that significantly correlated to the bacterial community profiles were: moisture content (p<0.05, method 1); moisture and total nitrogen (TN) contents (p<0.05, method 2); moisture content and C/N ratio (p<0.05, Biolog 0.2 AWCD); moisture and TN contents (p<0.05) and bulk density (BD, p<0.1, 0.6 AWCD); pH (p<0.05, 1.0 AWCD); and moisture and TN contents (p<0.05, 1.4 AWCD). The soil moisture gradient was shown to be the most decisive determinant of the changes in soil bacterial community profile associated with deforestation. The changes in TN, pH and BD were thought to be possible causes and/or effects of the variations of multivariate soil bacterial profiles in association with the degradation.
