Conditions and timescales for welding block-and-ash flow deposits

Welding of pyroclastic deposits to reform a coherent rock mass is a common phenomenon, especially for pumiceous pyroclastic density current deposits (i.e., ignimbrites). However, and despite the pervasive abundance of block-and-ash flow (BAF) deposits in the geological and modern record, instances of strongly welded BAF deposits are few. Here, we present a series of high-temperature (800-900 °C) compaction experiments designed to map the conditions (deposit thickness/stress and temperature/viscosity) and timescales that permit or inhibit the welding of BAF deposits. Our experiments were performed on unconsolidated aggregates (containing an ash and lapilli component) derived from crushed and sieved lava blocks (containing 25% crystals) taken from the well-documented welded BAF deposit at Mount Meager volcano (British Columbia, Canada). The experiments demonstrate that welding efficiency increases with increasing time and temperature. Progressive welding is expressed by increasing axial strain, porosity loss, and bulk density. The rate of change of each of these physical properties reduces as welding progresses. Microstructural analysis of the experimental products shows that the loss of interclast porosity during welding results from the progressive sintering and amalgamation of vitric fragments, and that the pore shape changes from sub-equant pores to stretched lenses sandwiched between vitric and crystal fragments. The coincidence between the microstructure and rock physical properties of the natural and experimental samples highlight that we have successfully reproduced welded BAF in the laboratory. Furthermore, our permeability measurements highlight a hysteresis in the return journey of the “there-and-back-again” volcanic permeability cycle (expressed by an increase in permeability due to vesiculation and fragmentation followed by a decrease due to welding). This hysteresis cannot be described by a single porosity-permeability power law relationship and reflects the change in pore shape and connectivity during welding. Finally, we show that a simple model for welding can accurately forecast the welding timescales of the BAF deposit at Mount Meager (as reconstructed from the collapse of the Lillooet River valley dam) using our experimental data. We use this validation as a platform to provide a universal window for the welding of BAF deposits, also applicable for comparable welded deposits (e.g., welded autobreccias in block-lavas and lava domes), for a broad range of deposit thickness (or stress) and effective viscosity.