For a sustainable management of the soil resource, we need to quantify the relevant soil properties. In many cases, this involves interpolation and extrapolation of soil-related information to unvisited locations. Since we can only sample a limited number of locations in space and/or time, we need to combine well-designed soil sampling schemes with adequate (statistical) modelling techniques to derive parameter estimates and to compute spatial or temporal predictions. Moreover, it is often useful to incorporate ancillary information, for instance from proximate or remote sensing, into the modelling framework. The latter is particularly true for digital soil mapping, where the challenge is to develop and apply tools, for instance based on nonlinear regression, that make best use of the available data. Additionally, it is crucial how remotely sensed soil-related information can be best combined with proximally sensed and traditionally measured soil properties. In other cases, we are rather interested in global quantities such as the spatial mean or its temporal trend, which we want to estimate with efficient sampling or monitoring schemes. Sampling, sensing and statistical modelling all induce uncertainty, which needs to be quantified in order to permit reliable assessments and to improve future investigations. In this session, we are interested in all aspects related to sampling for soil survey and monitoring as well as in the statistical approaches to incorporate the sample data and relevant ancillary soil sensing information into models and maps of soil status and change.