Skip to main content
University paper prices the next hole in a soil survey

University paper prices the next hole in a soil survey

A paper from the School of Underlying Conditions values each further auger hole in metres of error removed from a soil boundary, and finds a gradual boundary earns no more holes than a sharp one.

A stopping rule for the auger gives every boundary about one hole for each doubling of the ground it could be hiding in

Eighteen holes were bored inside six stretches of ground on one day of the University’s soil survey, and a paper published today has gone back to ask whether they were the right eighteen. One More Hole: A Stopping Rule for Locating a Soil Boundary by Auger Within a Fixed Field Day is by Dr Brynjar Obiora, Senior Research Fellow in the School of Underlying Conditions.

The paper aligns decision analysis, sequential search and field practice around a single unit of account, the metre of line. Each hole between two holes that disagree is valued by how far it moves the expected position of the boundary towards the true one, and a day’s holes are then allocated wherever the next is worth most. A boundary earns roughly one hole each time the ground it might lie in doubles, and a gradual boundary earns no more than a sharp one. Applied to the day in question, the rule moves four holes and lowers the summed expected error from 78.7 metres to 67.1. Three holes everywhere would have reached 71.2.

I have bored a good many holes in order to be sure. I expect to go on boring them, and I will now have a clearer idea of what each one was for.

— Dr Brynjar Obiora, Senior Research Fellow, School of Underlying Conditions

It is a source of some quiet satisfaction to the University that a judgement its crews have always made by feel now exists in a form the next crew can disagree with. Dr Obiora said the notebooks had been “more candid than I remembered being at the time”.

The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.ie9idi.