Determining the umbrella cloud geometry of unwitnessed silicic explosive eruptions: A case study from Mount Mazama (Oregon, United States)

Abstract

Volcanic Ash Transport and Dispersal Models (VATDMs) make real-time forecasts of tephra fall resulting from explosive eruptions possible. However, these predictions still mainly rely on eruption source parameters, such as erupted mass, total grain-size distribution, and plume height, gathered via thorough studies of past eruptions similar in nature. This dependency of eruption source parameters to analogous eruptions becomes particularly challenging when there are limited instances of similar events. An example is rhyodacitic to rhyolitic eruptions. This type of volcanic eruption has only been witnessed twice, at Chaitén (2008–2009) and Cordón Caulle (2011−2012), both in Chile. Here, we examine the 7.7 ka Cleetwood eruption of Mount Mazama (Oregon, USA), as a case study. This rhyodacitic eruption started explosively with two initial VEI 4, subplinian phases, and ended effusively with the emplacement of a rhyodacitic flow. We use the results of a detailed study of the proximal and medial tephra deposits as input in a VATDM to investigate the geometry and dimensions of the main plume formed during the Cleetwood eruption. We 1) constrain the erupted mass and calculate a detailed total grain-size distribution, 2) explore the Reanalysis 2 wind database to determine the direction and velocity of the local wind at the time of the eruption, and 3) use the VATDM Tephra2 with a grid-search method to estimate plume height, mass distribution within the plume, and the characteristics of tephra diffusion. We find that a vertical release of the erupted mass along a single line above the vent adequately replicates the measured mass loads but fails to simultaneously fit measured grain-size distributions at the same locations. We thus devise a method that not only accounts for a customized total grain-size distribution, real 1D wind patterns, and variable mass distribution within the plume, but also allows for adjustments to the size and location of an elliptical umbrella cloud. Using this method, we successfully replicate both local mass loads and high-resolution grain-size distributions and show that particles ≥0.125 mm from the lower Cleetwood unit were likely deposited from a 5 × 45 km2 umbrella reaching 16 km a.s.l., elongated in the direction of main wind intensity. This research contributes to enhancing the accuracy of predicting tephra transport from silicic volcanic eruptions. Moreover, it underscores the importance of utilizing grain-size data in combination with mass loads at specific locations to gain insights into the characteristics of the eruption plume, especially for eruptions that have not been directly observed.