Modeling volcanic ash and sulfur dioxide with the Weather Research Forecasting with Chemistry (WRF-Chem) model

• Main Author: Egan, Sean D.
• Other Authors: Cahill, Catherine, Stuefer, Martin, Webley, Peter, Lopez, Taryn, Simpson, William
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: 2019
• Subjects: volcanic gases; models; volcanic plumes; volcanic ash; volcanic eruptions; Kasatochi Volcano; Eyjafjallajökull Volcano; Iceland; Raikoke Volcano; Kuril Islands; Fairbanks; Alaska
• Online Access: http://hdl.handle.net/11122/10895

Dissertation (Ph.D.) University of Alaska Fairbanks, 2019 The Weather Research Forecasting with Chemistry (WRF-Chem) model is capable of modeling volcanic emissions of ash, sulfur dioxide and water vapor. Here, it is applied to eruptions from three volcanoes: the 2008 eruption of Kasatochi Volcano in Alaska, the 2010 eruption of Eyjafjallajökull in Iceland and the 2019 eruption of Raikoke in the Kurile Islands. WRF-Chem's ability to model volcanic emissions dispersion is validated through comparison of model output to remote sensing, in situ and field measurements. A sensitivity of the model to modeled plume height is discussed. This work also modifies the base WRF-Chem code in three ways and studies the effects of these modifications. First, volcanic ash aggregation parameterizations are added covering three modes of particle collisions through Brownian motion, differential settling and shear. Second, water vapor emissions from volcanic eruptions are added and coupled to the new aggregation scheme. The effects of these changes are assessed and found to produce volcanic ash concentrations in agreement with in situ measurements of plume concentrations and field measurements of tephra fallout. Third, the model is adapted to include multiple model initializations such that each is perturbed by selecting between two volcanic ash particle sizes and five initial plume heights. This modified WRF-Chem is nested in an application program interface that enables a new, automated, near real-time capability. This capability is assessed and the feasibility of its use as an augmenting tool to current operational VATD models is commented upon.