An approach to estimating the uncertainty in model descriptions based on a landscape space to model space mapping concept is described. The approach is illustrated by an application making use of plot scale geophysical estimates of changes in water content profiles to condition a model of recharge to the Sherwood Sandstone Aquifer in the United Kingdom. It is demonstrated that the mapping is highly uncertain and that many different parameter sets give acceptable simulations of the observations. Multiple profile measurements over time offer only limited additional constraints on the mapping. The resulting mapping weights may be used to evaluate uncertainty in the predictions of vadose zone flow dynamics for the site.