IMPROVING 3-D GEOLOGIC MAPPING: MOVING BEYOND THE SINGLE BEST PREDICTION

Geologic maps incorporate the geologists’ awareness of data accuracy, the consequences of the spatial distribution of the data, how the data represent likely geologic environments that created the observed system. They reflect the interpretation of features mapped at existing data locations and the prediction of feature distributions at unsampled locations. Despite the large amount of inference and conjecture that goes into a map, published maps typically only reflect the geologists’ single best prediction and they poorly reflect the geologists’ insights on predictive accuracy and plausible alternative geometries. This leaves map users with little guidance as to the varying reliability and uncertainty of the map predictions, and leaves decision makers with no guidance to limit their exposure to risk.

A fundamental change in mapping philosophy is needed to capture and communicate the insights already possessed by the mapper — the goal of mapping must be shifted from defining the single best prediction to that of predicting and characterizing multiple plausible scenarios. While we think the shift to a multi-scenario approach will be easy for most geologists to manage, there are still a few methodological problems to solve before this new philosophy can be effectively and efficiently implemented. Using a general workflow for geologic mapping as a guide, we clarify concepts of reliability and uncertainty in geologic modeling, and discuss the integration of multi-scenario thinking (e.g., probabilistic theory, fuzzy logic). We identify gaps in existing methods for characterizing uncertainty within a modeling project and we discuss ways in which a multi-scenario mapping philosophy can be applied by geologists with a range of quantitative skills.