Abstract
Digital soil nutrients (DSN) information plays a critical role in enhancing management of different ecological and agricultural systems. Mapping soil nutrient parameters at regional scale countries like South Africa that rely hugely on agricultural production but lack DSN datasets products that can enhance usage and accessibility of such information products to a wider range of communities. This study aims to contribute towards the GlobalSoilMap initiative by utilizing the South African Soil Database and MODIS datasets, to analyze and map selected soil nutrients using the Quantile Regression Forest model across South Africa. The soil organic carbon, pH-soil, cation exchange capacity (CEC), clay, sand, and silt parameters are mapped at 0–30 cm depth, including their uncertainties. The local DSM were comparable to the iSDA products with correlations above 80 % for selected soil parameters which proves that the country level DSM product can provide locally relevant information than global products. Variable importance showed that vegetation indices, land surface temperature and precipitation have a strong relationship with soil properties. Our results will empower policymakers to incorporate soil nutrient distribution information in their decision-making processes and farmers can use the same for improved soil management practices.