Hypothetical ensemble dispersion model runs with statistical verification

• Main Authors: Leadbetter, Susan, Jones, Andrew
• Format: Other/Unknown Material
• Language: English
• Published: Zenodo 2021
• Subjects: ensemble; dispersion; modelling; analysis; statistics; verification; volcanic dispersion; radiological dispersion; Hekla; Iceland
• Online Access: https://doi.org/10.5281/zenodo.4770066

This dataset contains output from the dispersion model NAME (Numerical Atmospheric-dispersion Modelling Environment)generated by modelling the hypothetical eruption of two volcanoes (Hekla and Oraefajokull both in Iceland) and a radiological release from 12 different locations across Europe. The hypothetical eruption of Hekla is a 12km eruption lasting 24 hours and the hypothetical eruption of Oraefajokull is a 25km eruption lasting 24 hours. Each of the radiological releases is a 1PBq Cs-137 release over 6 hours. The scenarios were modelling using three different sets of met data from the Met Office unified model; data from the global deterministic forecast, data from the global ensemble forecast and data constructed to form a global analysis. Output from the runs is stored in gzipped tar files labelled <name>_scenario_<year> .tar.gz where <name> is hekla, orae (short for oraefajokull) or radiological. Each of these tarballs contains the output from all the runs with a release start date in the year and month given in the file name. The data is stored in NetCDF format with each NetCDF file containing the run output from a single run with one type of met data. For example the NetCDF file:20181109T1200Z_engl_members.nc contains all the output from the run started at 12 UTC on 9 November 2018 using the ensemble global forecast met data. For the radiological scenario output is the total integrated air concentration and the total deposition after 48 hours. For the volcanic eruption scenarios output is the hourly air concentration of volcanic ash on 22 vertical levels, hourly ash column load and hourly ash deposits. Two additional files are included. These contain the Brier skill score computed by comparing the ensemble and deterministic output to the analysis output and the maximum distance at which threshold concentrations are exceeded. Full details of the computation can be found in a paper submitted to Atmospheric Chemistry and Physics. @Crown Copyright, Met Office