100-year GIC event scenarios. Antti Pulkkinen and Chigomezyo Ngwira The Catholic University of America & NASA Goddard Space Flight Center

Transcription

1 100-year GIC event scenarios Antti Pulkkinen and Chigomezyo Ngwira The Catholic University of America & NASA Goddard Space Flight Center 1

2 Contents Objectives. Approach. Identification of four key factors to be addressed. Results addressing the four identified key factors. Future work. Summary 2

3 Objectives Extreme event scenarios of significant current science and engineering interest. Use state-of-the-art understanding of the physics of the problem and extensive data sets to address the issue. Initial work Work carried in support of the NERC GMD TF activity. 3

4 Objectives Objective to rigorously characterize the properties of key physical parameters associated with rare/extreme GIC events. Generate regional extreme geoelectric field and GIC scenarios that can be used in further engineering analyses. Our definition for the rare/extreme event: 100-year maximum amplitude of the 10-s resolution geoelectric field. (Note: this does not imply 10-s pulse lengths) It is straightforward to map the geoelectric field into 4 GIC.

5 Four key geophysical factors that need to be addressed The effect of the ground conductivity structure on the extreme geoelectric field amplitudes. The effect of the geomagnetic latitude on the extreme geoelectric field amplitudes. Temporal scales (e.g. pulse lengths) of the extreme events. Spatial scales of the extreme events. 5

6 Selected approach 1. Assume regionally ( km) uniform geoelectric field. Assumption may not always hold especially at high-latitude locations. 2. Select a storm event and super- and sub-threshold geomagnetic observatories to obtain representative temporal storm profiles. Compute the geoelectric field and normalize the amplitudes. 3. Scale the normalized geoelectric field amplitudes obtained via 2) to obtain 100-year maximum amplitude event. Extrapolate previously computed statistics to determine the maximum amplitudes (originally Pulkkinen et al., 2008 now updated). 4. Compute GIC. Direct linear mapping or quasi-dc power grid 6 model can be utilized.

7 Extrapolation of statistics 101 If the stochastic process generating the studied signal does not change, knowledge about the shape of the probability distribution allows us to carry out extrapolation. In another words, if we have data for X years, we can infer information about statistical occurrence of amplitudes occurring less frequently than (probabilistically speaking) every X years. Rigorous study of tails of the probability distribution is carried out in extreme value theory. Less rigorous extrapolations reasonable if we don t go too far out of the range of available data. 7

8 Extrapolation of statistics 101 For demonstrational purposes we model the geoelectric as a lognormal stochastic process. We first generate 10 s synthetic data for 18 years. Example of synthetic 10 s geoelectric field data for 24 hours E [V/km] UT [hours] 8

9 Extrapolation of statistics 101 Data for 18 years # of 10 s values per 100 years Visual extrapolation to 100-year amplitudes E [V/km] 9

10 Extrapolation of statistics 101 We then generate 10 s data for 500 years and generate the statistics again. We now have observed 100-year events. 10

11 Extrapolation of statistics 101 Data for 500 years # of 10 s values per 100 years E [V/km] 11

12 Statistics and the effect of the ground conductivity Statistical occurrence of modeled geoelectric field in Quebec and British Columbia (Pulkkinen et al., 2008). 10-s IMAGE magnetometer chain data for years used for computing the geoelectric field. This was a lot of data. We have extended the statistics with data for year

13 Statistics and the effect of the ground conductivity (original) British Columbia Quebec Different curves represent different IMAGE stations. Visual extrapolation to 100-year amplitudes (well-justified if physics of the process remains the same) Data for

14 Statistics and the effect of the ground conductivity (update) a) British Columbia b) Quebec Different curves represent different IMAGE stations. Visual extrapolation to 100-year amplitudes (well-justified if physics of the process remains the same) Data for

15 The effect of the geomag. latitude Global 60-s magnetometer data from of the order of 100 stations processed for March 1989 and October 2003 storms. 15

16 The effect of the geomag. latitude 50 deg. of geomag. latitude Approx. order of magnitude drop in max. amplitudes Max. values over March 1989 storm 16 Max. values over October 2003 storm

17 The effect of the geomag. latitude 17

18 The effect of the geomag. latitude OK OK, this is only two storms. However, we do not have too many extreme storms with comparable data coverage. Consequently, we carried out the analysis of March 1989 auroral boundaries and review of literature regarding the boundaries during some of the historical events (e.g. Carrington event). 18

19 The effect of the geomag. latitude Approx. 40 deg. of geomagnetic latitude DE-I UV imaging data for March 14, :51 UT 19

20 The effect of the geomag. latitude Approx. 40 deg. of geomagnetic latitude Auroral sightings during the September 1859 storm (Kimball, 1960). Closed circles overhead 20

21 The effect of the geomag. latitude Pulkkinen et al. (2011)

22 The effect of the geomag. latitude So, although much more analysis is needed, there is some indication about the universality of the geomagnetic latitude threshold for at least some extreme storms. Further analysis by means of (even more) extended geomagnetic database and large-scale simulations of the magnetosphere-ionosphere system underway. 22

23 Temporal scales Need to capture great variety of geospace processes and temporal scales associated with extreme storms. Chose to use observational data for a major event to capture the variability. We use two representative stations. One from sub-threshold latitudes (Memanbetsu, Japan, 37 deg. of geomagnetic latitude) and another one from super-threshold latitudes (Nurmijärvi, Finland, 57 deg. of geomagnetic latitude). 10-s geomagnetic field observations from the two stations for 29-31, October 2003 provide representation of the temporal profiles. Geomagnetic field mapped into geoelectric field using the plane wave method and the Quebec ground model. Amplitudes normalized for scaling. 23

24 24

25 Spatial scales 25

26 Spatial scales Two-fold nature of major and extreme events: a) large geoelectric field magnitudes can be experienced across the globe in the region covered by the auroral current system, b) spatial correlation lengths associated with the field fluctuations can be short. Spatial GIC and geoelectric field correlations on global scale not well-known. We will assume spatially uniform geoelectric field on regional ( km ) scales. 26

27 Summary of the scenarios 27

28 This data is publicly available for further engineering analyses. 28

29 Mapping to GIC Apply extreme geoelectric field scenarios to quasi-dc models of the regional grid obtain GIC through each node of the system. Or apply simply (shown to hold to a good approximation): GIC(t) = ae x (t) + be y (t) System parameters Geoelectric field scenario 29

30 Mapping to GIC: application to Dominion s Virginia Power grid 30

31 Mapping to GIC: application to UK grid 31

32 Ongoing work We will use a large number ( 10) of major or extreme storm events to better quantify the location of the identified geomagnetic latitude threshold. We will use more extensive 10-s magnetometer data to establish more robust base statistics used in deriving the 100-year storm amplitudes. The new statistics will be used along with state-of-the-art large-scale magnetospheric simulations at NASA GSFC to explore the upper physical/theoretical limit for the geoelectric field magnitudes. An attempt will be made to reconstruct the geoelectric field and GIC waveforms of the Carrington event using available historical records. We will generate a collection of waveforms that represent the most significant drivers of large GICs. These will include waveforms for substorms, geomagnetic pulsations and sudden impulses. 32

33 100-year GIC scenarios Extreme regional geoelectric field scenarios generated as a function of ground conductivity structure and geomagnetic latitude. Numerical data for scenarios publicly available for mapping to GIC and further engineering analyses. Further work on extreme GIC ongoing. All the details can be found from: Pulkkinen, A., E. Bernabeu, J. Eichner, C. Beggan and A. Thomson, Generation of 100-year geomagnetically induced current scenarios, Space Weather, in press,