The Use of Monte-Carlo Techniques for the Estimation of Visibility

In archaeological research, visibility analysis is usually based on binary viewshed maps. However, Digital Elevation Models inherently contain errors due to inaccuracies in the original topographic data and the data structure used. Therefore the binary classification will be subject to an unknown amount of error. It would be more appropriate to use a method for visibility calculation which allows us to take the uncertainty on the elevation data into account. In this paper a further development of a method to calculate fuzzy viewsheds is proposed. The method is based on Monte-Carlo simulation of the DEM-error. From a large number of simulations, the visibility probability of a given cell can then be calculated. First experiments with this new methodology for our study area indicate that the hidden error in classical binary viewsheds is considerable. The deterministic use of such binary viewsheds for visibility analysis may therefore lead to erroneous conclusions. Furthermore, binary viewshed calculations appear to be rather sensitive to the algorithms used for DEM construction and visibility calculation. Fuzzy viewsheds appear to be much more robust. The application of fuzzy viewsheds requires information on the mean elevation error as well as an estimation of the number of Monte Carlo simulations needed. The first can be obtained by comparing interpolated elevation data with manually measured values. For the estimation of the required number of simulations two methodologies are proposed.