The development and use of a soil database with quantified uncertainties

Most soil information systems store soil observations without storing the associated uncertainties. Some have an ‘accuracy’ field, but this is usually unused because accuracy information about soil observations that are entered into the databases is rarely available. It is possible that the metadata that go with the soil data give some kind of summary measure of the accuracy, for instance the root mean squared error of a dataset, but this does not fully capture the uncertainty and is too little information to incorporate uncertainty information in digital soil mapping. The aims of this research project are to (1) design and implement a database for uncertain soil data, (2) analyse how quantitative uncertainty information can be included in digital soil mapping, and (3) analyse how the digital soil mapping results differ from a case in which uncertainty in soil observations is ignored.