Comment on “Prediction of the 1‐AU arrival times of CME‐associated interplanetary shocks: Evaluation of an empirical interplanetary shock propagation model” by K.‐H. Kim et al.

[1] Recently, Kim et al. [2007] (hereinafter referred to as KMC) have evaluated the empirical shock arrival (ESA) model and found only about 60% of the observed shocks arrived within ±12 h of the model prediction. They also found the deviations of shock travel times from the ESA model strongly correlate with the CME initial speeds (VCME), suggesting that the constant interplanetary (IP) acceleration used in the ESA model may not be applicable to all CMEs. KMC further concluded that faster CMEs decelerate and slower CMEs accelerate more than that what is considered in the ESA model. In other words, the average speeds of slower CMEs must be higher than predicted, while those of the faster CMEs must be smaller than predicted. Even though they recognized that Kim et al. [2007, paragraph 22] ‘‘do not exclude the possibility that the projection effect would be the main cause of the deviations from the ESA model,’’ they did not include it in their comparison with the ESA model. We point out that such systematic deviations in arrival time arise owing to projection effects. [2] The key ingredient of the ESA model and the parent empirical CME arrival (ECA) model is the IP acceleration profile of the CMEs. The acceleration profile was first derived empirically by Gopalswamy et al. [2000] using CME observations from the Solar and Heliospheric Observatory (SOHO) and Wind observations of the corresponding IP CMEs (ICMEs). The ESA model is a simple extension of the ECA model, in that the arrival of shocks preceded the CME arrival by an interval given by the shock standoff distance [Gopalswamy et al., 2005a, 2005b]. Since SOHO provides CME information in the sky plane, the measured speeds are subject to projection effects, so Gopalswamy et al. [2001] obtained a new acceleration profile using data from Helios and P78–1 missions [Sheeley et al., 1985; Lindsay et al., 1999] when the spacecraft were in quadrature (so the projection effects were minimal). [3] Thus, the acceleration profile used by the ESA model requires that the CME initial speed be devoid of projection effects. The projection effects are severe for CMEs originating close to the disk center, many of which are expected to be halo CMEs. Xie et al. [2006] has already shown that when the deprojected CME speed obtained using a cone model is used as input to the ESA model, the prediction is substantially improved. For CMEs originating close to the disk center, the space speed is expected to be higher than the sky-plane speed, so the arrival times will be smaller and get closer the model curve. In fact, the ESA and ECA models require that CMEs originate close to the disk center of the Sun. This requirement has been stated by Gopalswamy et al. [2005a] as follows: ‘‘we have assumed that what we observe at 1 AU is the nose of the magnetic cloud and shock. This is mostly true for magnetic clouds, but appropriate modifications have to be made when the CMEs are ejected off of the Sun–Earth line.’’ The appropriate modification is to consider the Earth-directed speed of the CME, rather than the sky-plane speed of the CME. It is known that CMEs associated with magnetic clouds originate close to the central meridian, while those associated with noncloud ICMEs generally originate at large angles to the Sun-Earth line [Gopalswamy, 2006; Gopalswamy et al., 2008]. In fact, KMC used data from Manoharan et al. [2004], who considered only CMEs with solar sources located within ±30 from the Sun center (the central zone). Since the Earthward component is expected to be smaller than the sky-plane speed for CMEs at larger angles to the Sun-Earth line (i.e., outside the central zone), the Sun-Earth travel time will shift to larger values when the correction is applied. The combined effect of the two projection corrections (using space speed and Earthward speed) is that the lowspeed outliers move to the right and the high-speed ones move to the left producing better agreement with the ESA model. [4] In order to correct for the projection effects, we need to use a cone model. Several CME cone models exist in the literature (see Xie et al. [2006] for details), but here we consider a simple model to illustrate the importance of projection effects. The ultimate aim is to compute the Earthward speed (VE) from the sky-plane speed (VS). To do this, we need to input the heliographic coordinates of the JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, A10105, doi:10.1029/2008JA013030, 2008 Click Here for Full Article