Determining minimum data input levels for reliable three-dimensional soil profiling

The ultimate goal of the project reported herein was to determine whether or not a readily quantifiable threshold data density could be established to accurately predict subsurface lithology. A study area in Dublin Ireland was selected for its high density of data (567 boreholes in less than 2km2) and for it being slated for the first phase of the upcoming Dublin metro. Analysis was conducted using the krieging function embedded in ArcGIS based on identifiable lithologies. Difficulties were encounter based on inconsistent labeling and the program's inability to cope with the presence of a single lithology appearing multiple times within a single soil profiles. In an area of 10,000m2, a minimumdata density of 8.0 boreholes predicted bedrock within 0.5m. Whereas in an area of 90,000m2, a sampling density of 8.3 was per 10,000m2 was needed to achieve the same accuracy, approximately double of British standards.