FRACTURING OF PUMICE FLOW DEPOSITS: LASCAR VOLCANO, CHILE

Unconsolidated pyroclastic flow deposits of the 1993 eruption of Lascar volcano, Chile, are, with time, becoming increasingly dissected by a network of deeply penetrating fractures. Now visible on the deposit surface, the fracture network on the northern deposit comprises orthogonal sets of decimeter wide sub-parallel linear voids that form a pseudo-polygonal grid. Using a combination of field and remote-sensing techniques, the fracture pattern has been related to surface slope and deposit type. Ground penetrating radar has been used to image fractures, which have propagated up to a few meters into the deposit. In addition, orbiting radar interferogrametry shows deposit subsidence of up to 1 cm/yr occurred 3 years after emplacement. Recent in-situ measurements found that the deposits retain moderately elevated temperatures, a few tens of centimeters beneath the surface. The fractures represent 7% (along flow direction) and 4% (perpendicular to flow) linear extension (a measure of strain). However, the low slope of the substrate (3 to 4°) inferred to be beneath the deposit, and pseudo-polygonal nature of the fracture network suggests that the fractures are not dominantly the result of extension. Thermal contraction can be inferred to have played a role in deposit shrinking but cannot be solely responsible because the fractures are an order of magnitude too wide given the coefficient of thermal contraction expected for an andesitic ignimbrite. The process of clast settling and compaction is considered to be active over this time scale as the deposit degasses (seconds to minutes) and settles (minutes to years), and this is considered to have caused the majority of the contraction observed. Uniquely at Lascar, the arid environment has allowed these slow processes to occur for more than 15 years without being overprinted by fluvial erosion. The observed fracture network is therefore interpreted to be unique to deposits from hot, fluidized flows in very arid environments where sufficient thermal contraction, deposit expansion, and clast settling can occur while subsequent fluvial erosion is slow. Understanding the generation of these features in recent deposits where surface expression is well-displayed, might be key to identification and understanding the origin of similar features in ignimbrites elsewhere.