In pursuit of a South African national soil database: potential and pitfalls of combining different soil data sets

Abstract

A national soil profile database will be a huge asset to South Africa, enabling improved soil information products to be created. However, numerous practical pitfalls exist in reliably creating such a database. This project aimed to identify such pitfalls, while showing the potential of a larger database. During a 2014 South African Soil Surveyors Organisation (SASSO) workshop in Cathedral Peak, 10 groups of soil surveyors were tasked to describe at least five profiles. The data from the groups were treated as different data sets, which needed to be merged into one SASSO database. Pitfalls to merging the data sets included misclassification of soils due to incomplete laboratory data, incorrect soil morphology description and incorrect soil description methodology. Furthermore, analysis by different laboratories and methods yielded dissimilar soil property values, while different data dictionaries collected diverse information. However, the enlarged database enabled the creation of a soil map that was more accurate than one created with a smaller, previously collected data set. A minimum set of criteria should accompany all data sets, including the soil surveyor’s name, the data collection date, the GPS position, the laboratory where the samples were analysed, and the analysis method with its standard error.