Simulation of 3D block populations to charaterize outcrop sampling bias in bimrocks

Geotechnical and hydrogeological characterization of block-in-matrix rocks (bimrocks) such as melange, fault gouge, till, and landslide debris can be difficult, but accurate and reliable characterization is important because block or block size distributions are known to influence factors such as permeability, shear strength, and the choice of construction methods. Geotechnical and hydrogeological exploration methods such as drilling and outcrop mapping, however, produce biased results because they yield 1D or 2D samples of 3D populations. Monte Carlo computer simulations can be used to explore the amount of bias introduced when 3D block distribution information is inferred from 2D projections such as outcrop maps or photographs. Simulations of the 2D outcrop projections of 3D blocks show that outcrop mapping has the potential to overestimate or underestimate mean block sizes and total block volumes by tens of percent, although the tendency will be towards underestimation for blocks that are not highly elongated. The magnitudes of errors introduced by 2D outcrop sampling can be on the order of ± 50 % for mean block sizes and ± 80 % for total block volumes. Carefully designed statistical sampling of block sizes and orientations combined with numerical simulations, however, has the potential to yield valuable information about the statistics of block distributions that may have significant effects on the design and construction of engineered works.