Revealing Uncertainty in Maps of Glacial Lake Algonquin

Abstract

This paper examines the effects that uncertainty in digital elevation data has on results obtained from a spatial model. Ongoing analysis of glacial Lake Algonquin in northern Michigan has used a spatial model to interpolate unknown shoreline locations between known shoreline positions. Extant shoreline features were surveyed with global positioning technology and ancient shorelines were reconstructed using a DEM and a statistical model of isostatic rebound. However, shoreline data contain relatively small measurement errors and relatively large errors are associated with the DEM. These errors, when propagated through a series of GIS operations, may render uncertain results. This research recognises and attempts to assess these errors in order to produce a new map of Lake Algonquin shorelines — one that illustrates shorelines and areas of positional uncertainty. Results indicate that even small errors in input data can contribute uncertainty to model output. Understanding these uncertainties can be valuable to further research concerning these and other ancient shorelines.