[Proposed by Gerard Heuvelink]
Measurement errors in soil observations can be large, but unfortunately they are often ignored. We need to work together with soil physicists, chemists and biologists to develop statistical methods that help characterise and quantify soil measurement error and make sure that measurement uncertainty is routinely stored in soil databases. We need to make sure that soil mapping algorithms take measurement uncertainty into account. All this is ever more important because we will get more of proximal soil sensing data and crowd-sourced and volunteered soil information, which all have substantial uncertainty. Measurement uncertainty also influences map validation strategies and sampling design optimisation.