SciKit-GStat Uncertainty: A software extension to cope with uncertain geostatistical estimates

This study is focused on an extension of a well established geostatistical software to enable one to effectively and interactively cope with uncertainty in geostatistical applications. The extension includes a rich component library, pre-built interfaces and an online application. We discuss the concept of replacing the empirical variogram with its uncertainty bound. This enables one to acknowledge uncertainties characterizing the underlying geostatistical datasets and typical methodological approaches. This allows for a probabilistic description of the variogram and its parameters at the same time. Our approach enables (1) multiple interpretations of a sample and (2) a multi-model context for geostatistical applications. We focus the sample application on propagating observation uncertainties into manual variogram parametrization and analyze its effects. Using two different datasets, we show how insights on uncertainty can be used to reject variogram models, thus constraining the space of formally equally probable models to tackle the issue of parameter equifinality.