This RAPID (Rapid Response) project is a collaborative effort by scientists from University at Buffalo (UB), the Geophysical Institute at the University of Alaska - Fairbanks (UAF-GI) and University to make an attempt to model the ash dispersal due to the volcanic eruption at Eyjafjallajokull, Iceland, that has wreaked havoc on European aviation since ash emissions began on 14 April 2010. The advecting downwind plume is thought to be a serious hazard to aircraft, with potential for damage to jet engines and abrasion of airplane leading edges. The proposed work could have important ramifications for air traffic over the next months, as past Eyjafjallajokull eruptions have continued for over one year.
It is proposed to use ensemble forecasting to develop probabilistic ash cloud maps at different times. More specifically, the 'Puff' simulation code will be used to hindcast the motion of the Eyjafjallajokull ash cloud through time beginning on 14 April 2010. Variability in the height and loading of the eruption column will be introduced through the column model Bent. Windfield data from several sources and with differing resolutions will be used. Output variability due to uncertain input parameters and initial conditions, and random forcing, will be determined. The output will be validated through comparison with satellite data. The principal scientific goals of this project are: to understand how uncertain input variables and random forcing affect the output of Puff simulations of the Eyjafjallajokull cloud, to validate output against satellite data, and to explore the possibility of producing improved assessments of ash cloud hazards. It is expected that the work will have implications for ash cloud tracking as well as probabilistic hazards assessment and ensemble forecasting.
The eruption of Eyjafjallajokull, Iceland, caused havoc for European aviation, with ash emissions from 14 April 2010 into May, and peak emissions in the period 14--18 April. To make predictions of the likely position of the ash cloud and issue advisories to the airline industry, the LondonVAAC (Volcanic Ash Advisory Center) used mathematical models of advection and dispersion. The models require input data on source conditions with uncertainty, such as eruption plume height. The outputs of these models depend on input data that have variability, such as a varying windfield. Despite the potential risk to property and life from ash clouds, varying inputs to the models have never been combined in a rigorous fashion to produce ensemble forecasts. We introduce a novel ensemble method designed to minimize moment errors, to develop probabilistic ash cloud maps at different times. For validation, the "puff" trajectory model is used to hindcast the motion of the ash cloud through time, concentrating on the period 14--16 April 2010. Variability in the height and loading of the eruption is introduced through the volcanic plume model "bent". Output uncertainty due to uncertain input parameters is determined with a polynomial chaos quadrature-based sampling of the multidimensional "puff" input vector space. Theresults indicate that a 3D forecast envelope that includes the observed plume footprint can be generated.
