Traditionally, tephra transport and dispersal models have evolved decoupled (offline) from numerical weather prediction models. There is a concern that inconsistencies and shortcomings associated with this coupling strategy might lead to errors in the ash cloud forecast. Despite this concern and the significant progress in improving the accuracy of tephra dispersal models in the aftermath of the 2010 Eyjafjallajökull and 2011 Cordón Caulle eruptions, to date, no operational online dispersal model is available to forecast volcanic ash. Here, we describe and evaluate NMMB-MONARCH-ASH, a new online multi-scale meteorological and transport model that attempts to pioneer the forecast of volcanic aerosols at operational level. The model forecasts volcanic ash cloud trajectories, concentration of ash at relevant flight levels, and the expected deposit thickness for both regional and global configurations. Its online coupling approach improves the current state-of-the-art tephra dispersal models, especially in situations where meteorological conditions are changing rapidly in time, two-way feedbacks are significant, or distal ash cloud dispersal simulations are required. This work presents the model application for the first phases of the 2011 Cordón Caulle and 2001 Mount Etna eruptions. The computational efficiency of NMMB-MONARCH-ASH and its application results compare favorably with other long-range tephra dispersal models, supporting its operational implementation.

The research leading to these results has received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007-2013) under the project NEMOH (REA grant agreement no. 289976). O. Jorba was partially funded by grant CGL2013- 46736 of the Ministry of Economy and Competitiveness of Spain. We are extremely grateful to the Argentinian National Meteorological Service for sharing data to validate this work; in particular we thank M.S. Osores for providing valuable insights into the eruption dynamics. Numerical simulations were performed at the Barcelona Supercomputing Center with the MareNostrum Supercomputer using 512 and 256 - 8x4 GB DDR3-1600 DIMMS (2GB/core) Intel Sandy Bridge processors, iDataPlex Compute Racks, a Linux Operating System, and an InfiniBand interconnection.

Peer Reviewed