Interface Tracking in Digitally Recorded Glaciological Data

As more data are recorded in digital form the importance of automatically extracting parameters of glaciological significance increases. This paper addresses the problem, with particular reference to tracking bedrock or internal reflectors in digitally recorded Radio Echo Sounding (RES) data. It has been found that the simplest solution to this problem is a “supervised” system, where operator decisions may be added interactively, either on operator command or upon loss of track. Increasing internal decision making within the program may reduce the number of operator interventions required, but is unlikely to eliminate them completely. Algorithms are presented and discussed for determining the position of the interface, for predicting the position of the interface in successive records, for determining the loss of track condition, and for re-acquiring track after loss of track.