SEMI-AUTOMATICALLY IDENTIFYING UNCONFORMITIES FOR DIGITAL GEOLOGICAL MAPS

Unconformities are geological features which are defined by major changes in the age and/or the lithology within the rock record for a region. These features capture the temporal and the spatial location of past erosion or non-deposition episodes in the area. Unconformities provide another layer of information which is integrated with other geological information such as faults, ore deposits, geochronology, etc., to help one better understand and provide an accurate interpretation of the evolution of crustal domains. Although the spatial extension of unconformities is a valuable detail in any geological record, historically, geoscientists have found the manual process to identify unconformities on GIS maps challenging. First, the number of possible geological attributes and their relationships with each other created a complexity in the process of constructing queries. Then, from visualization, scientists manually identified appropriate cross-cutting relationships to constrain the location of unconformities on digital maps. We have designed a relatively simple algorithm and implemented a program to semi-automatically identify all non-intrusive polygon boundaries on GIS maps. These boundaries are defined by the difference in formation age of adjacent polygons greater than some user-defined threshold. The present method identifies both unconformities and thrusts/faults; however, one can eliminate the latter features by additional geometry computations with available GIS polyline data. As with any software development, we have tried to test the program with different inputs in various sizes and levels of detail to evaluate the robustness and efficiency of the program. Even though we only used a general personal laptop/desktop computer system to run the program, it was able to process even the most detailed and large regional maps. We are planning to investigate artificial intelligence techniques to provide additional constraints on the spatial extent of unconformities under younger sediments. Moreover, we aim to explore different machine learning algorithms for identifying shapes and for interpreting geological information so as to distinguish between unconformities and linear features such as thrusts and faults.