Ash dispersal forecast and civil aviation workshop : Geneva, Switzerland, 18-20 October 2010 : consensual document

As a result of the serious consequences of the 2010 Eyjafjallajökull eruption on civil aviation, more than 50 volcanologists, meteorologists, atmospheric dispersion modellers, and space and ground-based monitoring specialists from 12 different countries (including representatives from 6 Volcanic Ash Advisory Centers and related institutions) gathered at the WMO headquarters in Geneva (for acronym definitions, see Appendix 1) to discuss the needs of the ash-dispersal modelling community, investigate new data-acquisition1 strategies and discuss how to improve communication between volcanological community and operational agencies. Based on a dedicated benchmark exercise and on three days of indepth discussion, recommendations have been made for future model improvements, new strategies of ash forecasting, multidisciplinary data acquisition, and more efficient communication amongst different communities. Issues addressed in the Workshop and key findings include: 1. Ash dispersal modelling. VATDM developers need to make a significant effort in collaboration with volcanologists and meteorologists to improve the definition of the source term (mainly mass eruption rate, grainsize distribution and mass distribution along the eruption column) and some critical aspects of particle sedimentation (i.e., particle aggregation and wet deposition), particularly if concentration has to be computed. 2. Uncertainty. VATDM developers need to make an effort to design models and forecasting strategies that can better characterize uncertainties. In fact, both the intrinsic behaviour of the natural system and input data (i.e., volcanological and meteorological data) are affected by various levels of uncertainties that need to be accounted for in order to compile comprehensive descriptions of particle transport and sedimentation. Stakeholders (e.g., aviation companies, decision makers) need to integrate probabilistic strategies into their processes of decision making. 3. Ensemble forecasting. Ash dispersal forecasting could be significantly improved through the implementation of ensemble forecasting strategies, namely: i) ensemble of input variables, ii) ensemble of VATDM (multi model), iii) ensemble on NWP and iv) ensemble on both VATDM and NWP. VATDM developers both from meteorology and volcanology fields need to explore and identify the best ensemble strategies that can be adapted to ash dispersal forecasting. 4. Combination of VATDM and observations. Real-time assimilation of observations into VATDM is crucial to model accuracy and hence to aviation safety. VATDM developers and monitoring specialists need to identify optimized strategies for the combination of models and observations.