Experiments on gas-ash separation processes in volcanic umbrella plumes

We present a series of analogue laboratory experiments which simulate the separation of ash and gas and the formation of secondary intrusions from finite volcanic umbrella plumes. We examined the lateral spreading of mixtures of freshwater and particles released into a laboratory tank containing a uniformly stratified aqueous solution. For times smaller than the sedimentation time of particles through the intrusion, the current remains coherent and intrudes laterally. As some of the particles settle into the underlying ambient fluid, a layer of particle-depleted fluid develops below the upper surface of the current and the density of the residual fluid is reduced. Over longer times, the intrusion ceases to be coherent, with small fingers of relatively buoyant, particle-depleted fluid rising from the upper part of the intrusion into the overlying fluid. Meanwhile, the lateral motion of the injected solution induces a return flow in the ambient fluid which sweeps some of the particles sedimenting from the lower surface of the intrusion inwards. As a result, relatively dense particle-laden fluid collects below the intrusion and then sinks into the underlying fluid. Eventually this fluid reaches a new neutral buoyancy height, where it intrudes to form a second laterally spreading current below the original intrusion. The process then repeats to form further weaker intrusions below. results of the separation of the ash and volcanic gas in an umbrella plume are consistent with field observations at Sakurajima volcano where positively charged plumes, thought to consist of volcanic gas, have been observed above negatively charged plumes of ash. This work also suggests that volcanic aerosols may form up to a kilometer above the original injection height of the ash. In a strong wind shear, this could result in very different trajectories of the ash and gas and so be important for evaluating the impact of ash plumes on both aviation safety and volcanic aerosol formation. The separation of the ash from the umbrella plume, after a time of order 103−104 s, corroborates recent models of the May 18, 1980 eruption at Mount St. Helens that, after a few hours, the motion of the ash was simply controlled by wind dispersal and gravitational settling. The complex motion of the sedimenting ash may also explain some of the observations of multiple intrusions in short-lived umbrella plumes such as observed at Redoubt, Alaska in April 1990.