USING PHOTOGRAMMETRY TO QUANTIFY SURFACE ROUGHNESS OF PRE-HISTORIC BOULDERS: IMPLICATIONS FOR RELATIVE AGE DATING

This study tested whether boulder surface roughness could be used to determine the exposure age of prehistoric rockfall surfaces. Rockfall surface roughness was quantified through photogrammetry and ArcMap analysis of digital elevation models of prehistoric boulders at Mount Rapaki (TePoho o Tamatea) in Christchurch, New Zealand. In total nineteen (19) rockfall boulder (volcanic breccia) surfaces were analysed using two scale-independent metrics (area ratio, slope gradient) and four scale-dependent metrics (standard deviation of slope, standard deviation of elevation, standard deviation of curvature, and standard deviation of residual topography). Overall, the metrics confirmed a positive correlation between rockfall surface roughness and increasing exposure age, although the scatter in data is significant (correlation coefficients range from 0.31-0.87). Each metric shows an upward trend and in general, is best fit using a logarithmic or 2^nd order polynomial trendline. Standard deviation of curvature and standard deviation of residual topography performed best and several of the metrics indicate a slowing in roughness increase over time (after ~20-25 ka of exposure). Several challenges and limitations were encountered in quantifying boulder surface roughness, including: (1) determining the ideal ‘moving window’ size for analysis; (2) defining the optimal number of photos for surface resolution; and (3) the influence of textural heterogeneities in the volcanic rock. The use of photogrammetry and ArcMap analysis of digital elevation models representing surface roughness is an important first step in developing a time-efficient and cost-effective method for determining the exposure age of prehistoric rockfalls and, where appropriate, timing of causative prehistoric shaking events.