Reply to comment by Haaland et al. on “A new interpretation of Weimer et al.'s solar wind propagation delay technique”

Abstract

[1] We welcome the comments offered by Haaland et al. [2006] on Bargatze et al. [2005]. Our paper examines a technique developed by Weimer et al. [2003] that can be used to estimate solar wind propagation delay times by applying minimum variance analysis (MVA) to study interplanetary magnetic field (IMF) time series. Solar wind propagation delay is the term used to denote the time duration between the instants that related conditions are measured at a pair of monitors spatially separated in the interplanetary medium or equivalently the time duration between the moment that interplanetary conditions are measured at an upstream monitor and the time when the same conditions would putatively reach the nose of the magnetosphere. Correction for propagation delay is important because one must apply time shifting adjustments to observed solar wind time series to properly correlate changes in upstream interplanetary conditions with subsequent response variations of magnetospheric processes. [2] Tests show that use of the Weimer et al. [2003] technique in place of the convection or corotation delay correction techniques leads to a modest increase in the ability to correlate geomagnetic activity with measures of solar wind input to the magnetosphere [Bargatze and McPherron, 2004]. However, in the course of our effort to validate the Weimer et al. [2003] technique, an error was discovered in the computer algorithm they employ to perform MVA on the magnetic field variance matrix [Bargatze et al., 2005]. A correction paper has been published that addresses the error in the algorithm [Weimer, 2004]. Still, the error turned out to be serendipitous as use of the Weimer et al. [2003] algorithm yields propagation delay time estimates that agree well with delay times obtained using an independent technique based on multispacecraft observations [Weimer et al., 2002] when both techniques are applied to study the same data interval. [3] Now, in the work of Haaland et al. [2006] the authors state that their comment paper is motivated by a sentence that appears at the end of the abstract in the work of Bargatze et al. [2005]. The sentence in question states: ‘‘Given that the new technique appears to improve the accuracy of estimating solar wind propagation delays and that it requires IMF data from only one interplanetary monitor, testing it as a space weather forecasting tool is clearly motivated.’’ Unfortunately, the statement’s use of the pronouns is ambiguous and did not convey our recommendations clearly. Here we wish to clarify this statement. [4] It is clear that solar wind propagation delay techniques that are based on MVA provide the best means yet available to account for the propagation of solar wind phase fronts from an upstream monitor to the nose of the magnetosphere. Also, we would recommend using any MVA technique that produces results that mimic those produced by the Weimer et al. method that does not have the flaw present in their analysis algorithm. The MVAB-0 method suggested by Haaland et al. [2006] is one such option. MVAB-0 is the name that Sonnerup et al. [1998] assign to the form of MVA that one obtains under the condition that the average IMF unit vector, jBj, lies in the plane of the discontinuity. In fact, a technique that produces results that are identical to those found using the MVAB-0 method is presented by Bargatze et al. [2005]. It is based on applying MVA to analyze the magnetic field variance matrix calculated using B ! ?, the magnetic field fluctuations that are perpendicular to the average background IMF. The correspondence between MVAB-0 and our analysis technique can be confirmed by comparing the results presented in Table 4 of Bargatze et al. [2005] with the second set of results listed in Table 1 of the comment paper. The eigenvector values and the eigenvalues listed in the two tables agree to four significant figures. This agreement is expected given that both of these mathematical techniques are based on the same underlying assumption. That is, both assume that the phase plane normal points in a direction that is perpendicular to the average IMF direction. [5] Still though, there are many styles of MVA and it is still not clear whether other MVA variants could be used to obtain more accurate estimates of propagation delay times. It is also important to consider the fact that the solar wind contains many types of structures including rotational discontinuities, tangential discontinuities, and interplanetary shocks. Tangential discontinuities are those that have no normal magnetic field crossing the plane that defines the discontinuity surface. These discontinuities convect passively with the solar wind. The orientations of tangential JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 111, A06103, doi:10.1029/2005JA011557, 2006