In geology we define geological contacts between units and we place a line in a map when we have identified this in the field, in most places the contacts are inferred and are shown as dotted lines. The extrapolation to depth is even more uncertain and can be improved by geophysical measurements and validated by drilling.
In general that is how geology works, mapping of units from satellite images and aerial photos, validation in the field and presentation in a geological map; which is now a digital map and is increasingly a 3 dimensional digital geological map. Where resources need to be defined the mapping is followed up by geophysical images and ultimately by drilling to define the true geological contacts.
This is the bulk of the geologist's job and we know what the confidence is in our geological images and geophysical models. The challenge is communicating this to the public in a world where we will be increasingly using the subsurface space to build, to store things such as gas and heat, to extract resources and to secure nuclear waste and avoid emission of waste CO2 from power stations by underground storage. There is no doubt that public pressure on defining the environmental consequences of these activities is increasing and geologists will be required to provide robust models of the subsurface if we are to convince them about the science that we do.
BGS is working on creating a 3 dimensional geological map of the UK which will vary in resolution depending on where you are in the UK. It will be high resolution below the major cities where the upper 100's of meters are critical for construction, ultra-high resolution in the proposed UK nuclear waste repositories, potential CO2 storage reservoirs and in future areas of resource exploitation. In other areas BGS will focus on high level landscape models allowing for protection and understanding of the UK geological heritage, farming development and modelling of flood risk and other natural hazards.
Only where we have observed a geological contact will the locations be exact otherwise we will always need to infer contacts and build geological scenarios to the best of our abilities.