Finding Nothing: Algorithms to search for circular features across geospatially integrated datasets

Counting craters is the primary technique for age-dating planetary surfaces; thus identifying craters can be an important, if tedious, task. While fresh craters are easy to identify, erosion, burial, and tectonic activity can obscure topographic and other signatures of older craters. One of these signatures is a crater's circularity and so a suite of algorithms are developed to search for circularity using multiple types of data. In this study, these algorithms are applied to the Crooked Creek, Missouri and Avon, Missouri areas. These areas are part of a line of features extending from southern Illinois through central Missouri into eastern Kansas, several of which have been proposed as impact sites. Two (Crooked Creek and Decaturville, MO) are generally agreed upon as impact craters, while debate still surrounds the others. Crooked Creek, as a known impact structure, will be used to test algorithms before searching for potential impact sites at Avon. In this case it would be of interest to show for the Avon area that a circular landform or feature does or does not exist within the given area. For the latter it is particularly difficult to make a convincing case based on inductive reasoning. If, for example, a particular data set shows many areas where no circular feature exists, can it be said that no circular features exist in the overall region? It is true that the more cases one can present, the more confidence one has in the conclusion, but still a single counter example is sufficient to negate the premise. While "proof" can never result from induction, the more negative results from diverse data sets, the greater the confidence in the inductive premise. Towards this end, seven data sets have been geospatially integrated for the test areas. Four of these sets (elevation, orthoimagery, geographic names and hydrography) come directly from The National Map of the U. S. Geological Survey, the other three include surface geology (integrated from historical geologic maps ranging in age from 1922 to 1981), structural elements (from the same maps) and gravity (from integrating a local gravity survey with stations from the Gravity and Magnetic Data Repository). For each set, an algorithm is proposed to seek circular features. Successful solutions from each set are compared and those meeting an integration criterion are flagged for visual inspection. If no viable candidates are located with the search region, then confidence is high that the feature does not exist at the resolution of the search.