Ensemble modeling of CME propagation

Tuesday, 16 December 2014: 4:24 PM
Christina O Lee1,2, Charles Nickolos Arge1, Carl J Henney1, Dusan Odstrcil3, George H Millward4 and Victor J Pizzo5, (1)AFRL/RVBXS, Kirtland Afb, NM, United States, (2)Boston College, Institute for Scientific Research, Chestnut Hill, MA, United States, (3)George Mason University Fairfax, Fairfax, VA, United States, (4)Cooperative Institute for Research in Environmental Sciences, Boulder, CO, United States, (5)NOAA Boulder, Boulder, CO, United States
The Wang-Sheeley-Arge(WSA)-Enlil-cone modeling system is used for making routine arrival time forecasts of the Earth-directed “halo” coronal mass ejections (CMEs), since they typically produce the most geoeffective events. A major objective of this work is to better understand the sensitivity of the WSA-Enlil modeling results to input model parameters and how these parameters contribute to the overall model uncertainty and performance. We present ensemble modeling results for a simple halo CME event that occurred on 15 February 2011 and a succession of three halo CME events that occurred on 2-4 August 2011. During this period the Solar TErrestrial RElations Observatory (STEREO) A and B spacecraft viewed the CMEs over the solar limb, thereby providing more reliable constraints on the initial CME geometries during the manual cone fitting process. To investigate the sensitivity of the modeled CME arrival times to small variations in the input cone properties, for each CME event we create an ensemble of numerical simulations based on multiple sets of cone parameters. We find that the accuracy of the modeled arrival times not only depends on the initial input CME geometry, but also on the reliable specification of the background solar wind, which is driven by the input maps of the photospheric magnetic field. As part of the modeling ensemble, we simulate the CME events using the traditional daily updated maps as well as those that are produced by the Air Force data Assimilative Photospheric flux Transport (ADAPT) model, which provide a more instantaneous snapshot of the photospheric field distribution. For the August 2011 events, in particular, we find that the accuracy in the arrival time predictions also depends on whether the cone parameters for all three CMEs are specified in a single WSA-Enlil simulation. The inclusion/exclusion of one or two of the preceding CMEs affects the solar wind conditions through which the succeeding CME propagates.