Portal de Periódicos da CAPES

Comment on “Substorm growth and expansion onset as observed with ideal ground‐spacecraft THEMIS coverage” by V. Sergeev et al.

2012; American Geophysical Union; Volume: 117; Issue: A2 Linguagem: Inglês

10.1029/2011ja017254

2156-2202

Martin Connors,

Earthquake Detection and Analysis

[1] Sergeev et al. [2011a] give a comprehensive description of the initial phases of a substorm on 29 March 2009 (ca. 5 UT), with an uncommonly good spacecraft placement, both with respect to magnetospheric features and ground facilities. The emphasis is on the growth phase and initial onset of this “generic” event, which, however, had a “rather weak” ground signature. On the basis of THEMIS ground network data, Sergeev et al. [2011a, paragraph 37] found that “auroral zone bays…were below 50 nT during first 10 min of the expansion phase.” Other magnetic data from near the onset location, or the standard AL index, show perturbations, which significantly exceed this and well-known substorm signatures. These are detailed here to add to the description of this important event. This leads to a brief general discussion of best use of ground magnetic data to characterize substorms. [2] Use of the AE/AL indices has often been a starting point for substorm research [e.g., Weimer, 1994]. There are inherent problems with such indices [Gjerloev et al., 2004; Connors et al., 2011], especially in studies of individual events [Rostoker, 1972]. As an envelope, they discard what can be essential information about where activity was taking place. Further, a station among the limited number included in the index may not be near the electric current associated with the substorm, and may not indicate it well. This deficiency may occur in two forms: activity may not be indicated at all, or activity of a current that changes location may be indicated at a misleading time. Nevertheless, AE-type indices give a general indication of activity of auroral zone currents and are useful if used cautiously.Figure 1ashows the commonly used AL (lower), AU (upper) and AE (Electrojet: AU-AL) indices for a time period including that studied bySergeev et al. [2011a]. Also shown is the growth phase interval deduced by them, partly based on use of the AE index. The AU index turned upward at about 4:20 UT, shortly before the authors' start of growth phase. AU increase is often due to response of the evening sector eastward electrojet to solar wind driving [Baumjohann and Treumann, 1997], and marks the start of growth phase [Weimer, 1994]. Here, AE stations Barrow and College (Alaska), in the evening sector, showed very similar signatures (data not shown) and are sources of this AU change. The growth phase likely started slightly earlier than inferred. The growth phase ended near the time of rise of the AE index at 5:16 UT, marking the start of the expansion phase, as inferred by the authors from multiple sources. The origin of this rise to about 100 nT was westward auroral zone currents indicated by the AL component of AE. [3] In Figure 1b, X (northward) component signatures (1 s cadence) from AE stations Fort Churchill (FCC) and Sanikiluaq (SNKQ) are shown with the standard AL index near the time of onset. To facilitate comparison, the start time is that (5:13 UT) of Sergeev et al.'s Figure 12 [2011a], however, a slightly longer period (to 5:40 UT) is shown. Magnetic perturbations used here are referenced to the values at 4:10 UT, before the growth phase. Before onset, and again after 5:25 UT, the AL index (a one-minute average) followed very closely the SNKQ values. However, from 5:18 to 5:25 UT, the index value was clearly associated with FCC, the closest station to the onset location. InFigure 1c, the X component from Gillam (GILL) is shown with THEMIS AL. Perturbations at GILL clearly were much smaller than those at FCC. THEMIS AL appears to have mostly followed SNKQ, which is included in that index. The GILL perturbations, which Sergeev et al. [2011a] emphasized, at all times were weaker than those of nearby auroral zone stations. Figure 1d shows the Z components observed at the ground stations. Negative Z perturbations arise from a substorm electrojet north of a station. Only FCC, 265 km north of GILL, shows a positive Z perturbation at any time, immediately after onset. By 5:23 UT, this signal had gone through zero and become negative, indicating passage of the electrojet from the south, over the station, to north of it. The maximum perturbation at FCC, at 5:20 UT, took place when the electrojet was still south of the station, so that the 100 nT value underestimates auroral zone perturbations, but indicates they were larger than GILL alone would suggest. FCC data suggest a stronger substorm, and with standard features [see Rostoker et al., 1980] including a sharp onset and subsequent poleward expansion. One would expect the dense stations of the THEMIS ground array [Angelopoulos et al., 2009] to indicate onset better than standard AL. This counterexample cautions of the need to include as much ground data as possible. Quantitative evaluation of the strength of substorms appears lacking in the literature: AL values of −100 as occurred early in this event would not have even made the threshold of the statistical study of Weimer [1994]. The abundant evidence presented by Sergeev et al. [2011a] indicates that the physical processes associated with the substorm cycle took place. Despite improvements in the density of stations, substorm signals may not always be well indicated by a given index or even in data from a given multistation array. Proper characterization of even small substorms is important, as we attempt to understand the overall phenomenon. [4] Sergeev et al. [2011a]roughly estimated the current flowing in the SCW to be about 0.1 MA based on midlatitude perturbations. A determination of cross-meridian currents in the auroral zone based on the quantitative models ofKisabeth and Rostoker [1977] is useful for comparison. Perturbations at the stations Pinawa (PINA), GILL, FCC, and Rankin Inlet (RANK; 455 km N of FCC) allowed an optimization routine to find the north and south boundaries, and current, for an electrojet model crossing their 325° magnetic meridian. The routine gave indicative results for 5:20:30 UT (Figure 2a), near the time of largest magnitude of X perturbation at Fort Churchill, and 5:26:30 UT (Figure 2b), which appeared to mark the maximal expansion north during the initial phase of the event. The closeness of circles (data) and stars (value given by model) indicates that the simple electrojet model matched the data well. Small perturbations at PINA (not shown) were also well matched. The strengthened current soon after onset is found to be about 54 kA; at maximum poleward expansion of the electrojet, 0.106 MA. The center of the electrojet moved poleward about 2°. The current matches that found by Sergeev et al. [2011a], with electrojet motion comparable to that inferred from optical data. Since the auroral zone current was close to that deduced from midlatitude perturbations, which are sensitive to the field-aligned current of the SCW [Clauer and McPherron, 1974], that parameter of the SCW was independently and consistently derivable from two different aspects of ground data. It is a basic physical quantity, possibly usable in quantitative analysis [Sergeev et al., 2011b]. In discussing the weak detected current of this SCW, Sergeev et al. [2011a] discussed the possibility that R1/R2 currents could be intense but not observable from the ground. This cannot be easily be determined from ground magnetic data. [5] This brief comment shows that the ground signature of small substorms may be underestimated and not conform to recognized patterns, if data from near enough the onset is not available or not used. The best practice, suggested for studies of individual events, is to accumulate and interpret information from all stations reasonably expected to show signals. Improvements such as SuperMAG [Gjerloev, 2009; Newell and Gjerloev, 2011] assist in making this approach less onerous. Indices can be helpful but can be nonlocal and may omit information, and data inversion techniques could supplement them, allowing inclusion of more data in a physically meaningful way. [6] The THEMIS AL index was calculated by Xiangning Chu. NRCan FCC high-cadence data were supplied by D. Danskin. CARISMA and NRCan data supplied to THEMIS, and the AE indices from OMNI, calculated by J. H. King and N. Papatashvili, were obtained from CDAWeb. [7] Masaki Fujimoto thanks the reviewer for his or her assistance in evaluating this paper.