Paving the way for more accurate earth system modelling in Malaysia

Abstract

Variations in soil structural and physiochemical properties can impact earth system models related to climate change. However, in most of the countries including Malaysia, soil data are available as low resolution and semi-detailed soil series maps, which hinder their applicability. Therefore, the objective of this study was to develop a framework to compare the conventional soil mapping with digital soil databases and to determine the accuracy of digital soil mapping. Kedah province located in the northern region of Peninsular Malaysia was selected as the study area. Observed soil (Cation exchange capacity – CEC, pH, organic matter as chemical and clay, sand and silt content as physical properties) data published by the Department of Agriculture, Malaysia was compared with SoilGrid (soillgrid.org) dataset. The depths of observed data were harmonised to international standard soil depths as; 0-5, 5-15, 15-30, s30-60, 60-100, and 100-200 cm using the equal quadratic spline approach. Global data were compared with harmonised 16 selected soil series data in Malaysia. Observed soil chemical properties (CEC, pH, and organic carbon) were within the range of SoilGrids data which allows the selected chemical properties to be used where observed data are limited or not available. In contrast, most of the soil physical properties from SoilGrids were not accurate when compared with the observed data. The developed framework in this study can be used to strengthen the accuracy of the available soil data in Earth system models, particularly those that are related to organic carbon and soil structural properties.