The dynamics and thermodynamics of volcanic clouds: Theory and observations from the april 15 and april 21, 1990 eruptions of redoubt volcano, Alaska

Abstract

On April 15, 1990 and April 21, 1990, two relatively small explosive eruptions occurred at Redoubt Volcano, Alaska, lasting about 4 and 8 minutes respectively. On both occasions the erupted material travelled as a pyroclastic flow down an ice canyon on the north flank of the volcano. Using slow-scan television recordings of the eruption and the seismic record in both the near- and far-field, we deduce that after a few minutes, the upper parts of these pyroclastic flows became buoyant and a large and hot ash cloud ascended off the flow. These thermals rose to a height of about 12 km, at which point they began to spread laterally, as umbrella clouds. Using a simple thermodynamic model, we estimate that the clouds had a temperature in the approximate range 600–700 K as they rose buoyantly from the flow after entraining and heating air, and melting and vaporizing ice. We also estimate that in each eruption approximately 10 9 kg of fine ash was injected into the atmosphere. During the April 15 eruption, a slow-scan television camera recorded the ascent of the cloud. These observations are described, analysed and compared with the predictions of a new model for the dynamics governing the ascent of such clouds. In accord with the observations, our model predicts that the cloud initially ascended rather sluggishly, since it is only just buoyant on rising from the pyroclastic flow. As it ascends, it entrains and heats up more air, and hence generates more buoyancy. Therefore, it accelerates upwards; only much higher in the cloud does the velocity decrease again, as the thermal energy of the cloud becomes exhausted. The model also predicts that the height of rise of such coignimbrite thermals is a function of the initial mass and temperature of the cloud, but is almost independent of the initial velocity. During the April 21 eruption, a sequence of photographs recorded the lateral spreading of the umbrella cloud during an interval of about 10 minutes after the eruption. These photographs are analysed and successfully compared with a simple model for the spreading of the umbrella cloud as a gravity current in a stratified environment. Using an AVHRR satellite image of the air-fall deposit from the April 15 eruption combined with a simple model, we suggest that the primary mechanism of ash dispersal was transport by the ambient wind.