Correlations between soil microbial and physicochemical variations in a rice paddy: implications for assessing soil health RYOICHI DOI* and SENARATNE LEELANANDA RANAMUKHAARACHCHI

Soil quality can be assessed by numerous soil variables (Mausbach and Seybold 1998). In addition to physicochemical variables, biotic variables also refl ect soil quality (Pankhurst et al. 1997). Soil variables often show different patterns of response to the same impact (Jha et al. 1992), refl ecting the multidimensional quality of soil health (van Straalen 2002). These soil variables are associated with different soil-related variables, including crop yields. Variations in soil-related variables can be empirically predicted from datasets consisting of multiple soil variables. Therefore, analysis of multiple soil biotic variables can provide comprehensive and multidimensional information (Sena et al. 2000). Spatial variations in soil-related variables have different patterns (Covelo et al. 2008). These differences are rooted in the complex relationships between each soil-related variable and the aspects of the soil that they describe (Mausbach and Seybold 1998). These different patterns suggest that there may be an ideal dataset to best describe each soil-related variable. Careful observation is necessary to determine which dataset is the most useful in describing each soilrelated variable. In addition to physicochemical profi ling of soils, various methods of multivariate profi ling of soils based on microbial observations have been proposed (Kirk et al. 2004). Changes in soil quality can be detected by observing community-level physiological profi les of soil bacterial communities (Garland and Mills 1991). Analysis of the physiological activity of soil bacteria may reveal important information about soil quality which may go undetected by physicochemical analysis, because soil bacterial activity responds differently to impacts than do physicochemical parameters (Doi and Puriyakorn 2007).