FIGURE 7.4 The extreme value distribution function for Dst as estimated from 20-day blocks. Each data point represents a minimum Dst measured during an independent, historical 20-day period. It predicts that the probability of Dst being less than −938 nT in ay future 20-day period is 10−4 SOURCE: T.P. O’Brien, Aerospace Corporation, “Extreme Events in Space Weather,” presentation to the space weather workshop, May 23, 2008.

FIGURE 7.4 The extreme value distribution function for Dst as estimated from 20-day blocks. Each data point represents a minimum Dst measured during an independent, historical 20-day period. It predicts that the probability of Dst being less than −938 nT in ay future 20-day period is 10−4 SOURCE: T.P. O’Brien, Aerospace Corporation, “Extreme Events in Space Weather,” presentation to the space weather workshop, May 23, 2008.

Perhaps more interesting is the application of extreme-value analysis to the Carrington event of 1859. In this case the parameter analyzed is 1-hour averages of Dst. Dst, the perturbation of the terrestrial magnetic field near the equator, is typically negative during magnetic storms. The extreme-value distribution function (H(x)) estimated from 1-hour averages of Dst organized into 20-day blocks is shown in Figure 7.4. This data set has k = −0.22. From this function the estimated lower limit to Dst is −938 n. That is, in any 20-day block of data the probability of Dst exceeding (being less than) −938 nT is 10−4. Now compare this estimate of the minimum possible value of hourly averaged Dst with that estimated from Colaba (Bombay) magnetometer data during the Carrington event: −883 nT (Tsurutani et al., 2003; X. Li, personal communication).13 Of course, there are multiple assumptions implicit in the conclusion that the Carrington event was nearly the extreme possible. These include the geophysics of the data set in Figure 7.4 being the same as that which produced the Colaba magnetometer extreme value during the Carrington event. Fennell remarked that “we many not be measuring what we would classically call Dst when you get down in this part of the probability distributions”. Additionally, the above conclusion assumes that the Sun’s variability is statistically unchanged over the time it has been observed and into the foreseeable future.

SUMMARY

As society becomes more interconnected, and as its systems become more efficient and connected, with risk transferred among them, as noted by James Caverly in an earlier session of the workshop, space weather impacts on electric power grids, satellites, and GPS are going to affect almost every area of our lives. The challenge for society is understanding the true nature of the vulnerability now and in the future.

A frequent theme throughout the workshop was the uncertainty in attempting to analyze future vulnerabilities. Uncertainty is introduced by the use of systems in ways not expected, or engineered for, by their designers. In some cases, the system providers may not even know who the users are. For example, the International GNSS Service



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement