Abstract

Soil spatial patterns in tropical forest may impact, or to some extent control, important characteristics of tropical forests such as forest biodiversity and carbon emissions. However, tropical forests are highly heterogeneous and therefore determining soil spatial patterns using traditional methods is highly labor intensive and destructive. As a result, rapid soil pattern-property assessment methods for use under tropical conditions are of great interest for evaluating soil patterns and properties at the plot scale. We evaluated the potential of electromagnetic-induction imaging (EMI) for determining spatial variation of soil properties in native tropical forest and teak plots during the dry and wet season. Geostatistical analysis of apparent electrical conductivity (ECa) measured by the EMI sensor revealed consistent semi-variogram ranges of ?30 m over both time (dry season, wet season) and sites (native forest, teak). Analyses of measured soil properties with principal component analyses were similar for both plots: clay-silt content, sand content and soil hygroscopic water coefficient (SHC) were the leading components of the main axes. However, multiple linear regression models revealed a more complex picture. In the teak monoculture, ECa was correlated with soil texture (i.e., clay-silt content) and electrical conductivity of the soil saturation extract (ECe) (R2 = 0.93), while in the native forest ECa was correlated with texture and residual soil water repellency (SWR) (R2 = 0.85). This finding indicates that SWR may play an important role in determining wetting patterns in tropical forests, with implications for resource distribution and gas emissions and may provide a basis for future investigations of soil-tree relationships.

You do not currently have access to this article.