5.3 Quantitative Assessment of Space Weather Forecasts Using Object Based Verification Techniques

Tuesday, 8 January 2013: 4:00 PM
Room 16B (Austin Convention Center)
Michael Wiltberger, NCAR, Boulder, CO; and W. Kleiber, V. Merkin, and B. Anderson

The quantitative assessment of spatial distributed data presents several unique challenges. In the terrestrial meteorology forecasts of precipitation include not only the region where the rainfall is likely to occur, but also its intensity. Object based verification techniques are used to evaluate the ability of numerical weather prediction models on both values using advance techniques that use computer algorithms to automatically identify similar regions in both the forecast and observations before proceeding to an evaluation stage that compares these regions. This comparison often includes metrics for size, shape, location, peak intensity, integrated precipitation and many others. Space weather forecasts contain similar challenges especially those related to auroral precipitation, polar cap potential patterns, and the distribution of field aligned currents. The recent deployment of Active Magnetosphere and Planetary Electrodynamics Response Experiment (AMPERE) measurements of the field-aligned currents provides an excellent opportunity for using object based verification methods to assess the performance of results of global scale numerical simulations of the geospace environment. In this work will compare the results from ultra high-resolution simulations of the magnetsophere-ionosphere system conducted with Lyon-Fedder-Mobarry global Magnetohydrodynamic simulation. This evaluation will be done using object based verification techniques and will include quantitative comparisons on the size, distribution, and intensity of the current systems present in the AMPERE observations and the simulation results. The presentation will conclude with comments on how this technique can be applied to other space weather parameters.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner