Development and Evaluation of a Volcanic Ash Ensemble Forecasting System Using the NOAA HYSPLIT Model

Abstract:

Volcanic ash poses severe risks to aviation safety and human health. It is critical to predict how fine ash is transported once erupted from volcanoes. Most of the published studies only focus on deterministic models, and therefore, fail to address the nature uncertainty in atmospheric transport. This work presents the development and preliminary results of a new ensemble forecasting system for volcanic ash dispersion using NOAA’s Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model. The dispersion ensembles are constructed using the current HYSPLIT model configuration implemented in operations, driven by multiple meteorological predictions: the Global Ensemble Forecasts (GEFS) v11 and v12. Each of them has one control member, and 20 and 31 perturbation members for GEFS v11 and v12, respectively. Since the ensemble forecasting system models uncertainties, the variability of weather fields – a major source of uncertainty – can be used to help address the uncertainty related to transport. Satellite retrievals of ash from ABI-GOES-16, MODIS and HIMAWARI for three historical volcano eruptions, in the NOAA Volcanic Ash Advisory Centers’ area of responsibility, were used for evaluating HYSPLIT ensemble output. The Model Evaluation Tools (MET) verification package developed by the Developmental Testbed Center (DTC) version 9 was applied for ensemble statistical analysis. With satellite observed ash data, we compared the performance of the volcanic ash mass loading predicted by the HYSPLIT ensemble to several deterministic HYSPLIT runs driven by the Global Data Assimilation System (GDAS) and Global Forecast System (GFS), which were used operationally by NOAA. The preliminary results from this analysis show that using the ensembles can reflect some of the meteorological forecast uncertainty in the ash plumes, and hence may improve the transport and dispersion forecasts.