Geologic and geomorphic controls on rockfall hazard: how well do past rockfalls predict future distributions?

To evaluate the geospatial hazard relationships between recent (contemporary) rockfalls and their prehistoric predecessors, we compare the locations, physical characteristics, and lithologies of rockfall boulders deposited during the 2010-2011 Canterbury earthquake sequence (CES) (n = 185) with those deposited prior to the CES (n = 1093). Population ratios of pre-CES to CES boulders at two study sites vary spatially from similar to 5 : 1 to 8:5 : 1. This is interpreted to reflect (i) variations in CES rockfall flux due to intra- and inter-event spatial differences in ground motions (e.g., directionality) and associated variations in source cliff responses; (ii) possible variations in the triggering mechanism(s), frequency, flux, record duration, boulder size distributions, and post-depositional mobilization of pre-CES rockfalls relative to CES rockfalls; and (iii) geological variations in the source cliffs of CES and pre-CES rockfalls. On interfluves, CES boulders traveled approximately 100 to 250 m further downslope than prehistoric (pre-CES) boulders. This is interpreted to reflect reduced resistance to CES rockfall transport due to preceding anthropogenic hillslope de-vegetation. Volcanic breccia boulders are more dimensionally equant and rounded, are larger, and traveled further downslope than coherent lava boulders, illustrating clear geological control on rockfall hazard. In valley bottoms, the furthest-traveled pre-CES boulders are situated further downslope than CES boulders due to (i) remobilization of pre-CES boulders by post-depositional processes such as debris flows and (ii) reduction of CES boulder velocities and travel distances by collisional impacts with pre-CES boulders. A considered earth-systems approach is required when using preserved distributions of rockfall deposits to predict the severity and extents of future rockfall events.