First Imaging of Coronal Mass Ejections in the Heliosphere Viewed from Outside the Sun – Earth Line

We show for the first time images of solar coronal mass ejections (CMEs) viewed using the Heliospheric Imager (HI) instrument aboard the NASA STEREO spacecraft. The HI instruments are wide-angle imaging systems designed to detect CMEs in the heliosphere, in particular, for the first time, observing the propagation of such events along the Sun – Earth line, that is, those directed towards Earth. At the time of writing the STEREO spacecraft are still close to the Earth and the full advantage of the HI dual-imaging has yet to be realised. However, even these early results show that despite severe technical challenges in their design and implementation, the HI instruments can successfully detect CMEs in the heliosphere, and this is an extremely important milestone for CME research. For the principal event being analysed here we demonstrate an ability to track a CME from the corona to over 40 degrees. The time – altitude history shows a constant speed of ascent over at least the first 50 solar radii and some evidence for deceleration at distances of over 20 degrees. Comparisons of associated coronagraph data and the HI images show that the basic structure of the CME remains clearly intact as it propagates from the corona into the heliosphere. Extracting the CME signal requires a consideration of the F-coronal intensity distribution, which can be identified from the HI data. Thus we present the preliminary results on this measured F-coronal intensity and compare these to the modelled F-corona of Koutchmy and Lamy (IAU Colloq. 85, 63, 1985). This analysis demonstrates that CME material some two orders of magnitude weaker than the F-corona can be detected; a specific example at 40 solar radii revealed CME intensities as low as 1.7×10⁻¹⁴ of the solar brightness. These observations herald a new era in CME research as we extend our capability for tracking, in particular, Earth-directed CMEs into the heliosphere.     

Numerical Magnetohydrodynamic Experiments for Testing the Physical Mechanisms of Coronal Mass Ejections Acceleration

Analysis of observations from both space-borne (LASCO/SOHO, Skylab and Solar Maximum Mission) and ground-based (Mauna Loa Observatory) instruments show that there are two types of coronal mass ejections (CMEs), fast CMEs and slow CMEs. Fast CMEs start with a high initial speed, which remains more or less constant, while slow CMEs start with a low initial speed, but show a gradual acceleration. To explain the difference between the two types of CMEs, Low and Zhang (2002) proposed that it resulted from a difference in the initial topology of the magnetic fields associated with the underlying quiescent prominences, i.e., a normal prominence configuration will lead to a fast CME, while an inverse quiescent prominence results in a slow CME. In this paper we explore a different scenario to explain the existence of fast and slow CMEs. Postulating only an inverse topology for the quiescent prominences, we show that fast and slow CMEs result from different physical processes responsible for the destabilization of the coronal magnetic field and for the initiation and launching of the CME. We use a 2.5-D, time-dependent streamer and flux-rope magnetohydrodynamic (MHD) model (Wu and Guo, 1997) and investigate three initiation processes, viz. (1) injecting of magnetic flux into the flux-rope, thereby causing an additional Lorentz force that will destabilize the streamer and launch a CME (Wu et al., 1997, 1999); (2) draining of plasma from the flux-rope and triggering a magnetic buoyancy force that causes the flux-rope to lift and launch a CME; and (3) introducing additional heating into the flux-rope, thereby simulating an active-region flux-rope accompanied by a flare to launch a CME. We present 12 numerical tests using these three driving mechanisms either alone or in various combinations. The results show that both fast and slow CMEs can be obtained from an inverse prominence configuration subjected to one or more of these three different initiation processes.     

An Analytical Model to Predict the Arrival Time of Interplanetary CMEs

Referring to the aerodynamic drag force, we present an analytical model to predict the arrival time of coronal mass ejections (CMEs). All related calculations are based on the expression for the deceleration of fast CMEs in the interplanetary medium (ICMEs), [(v)\dot]=-\frac115 700(v-V_SW)²\dot{v}=-\frac{1}{15\,700}(v-V_{\mathrm{SW}})^{2} , where V _SW is the solar wind speed. The results can reproduce well the observations of three typical parameters: the initial speed of the CME, the speed of the ICME at 1 AU and the transit time. Our simple model reveals that the drag acceleration should be really the essential feature of the interplanetary motion of CMEs, as suggested by Vršnak and Gopalswamy (J. Geophys. Res. 107, 1019, 2002).     

Empirical model of the transit time of interplanetary coronal mass ejections

We study the correlation between near-Earth observations of interplanetary coronal mass ejections (ICMEs) detected by the Wind and ACE spacecrafts and their counterparts of coronal mass ejections (CMEs) observed near the Sun by the SOHO/LASCO coronagraph during 1996–2002. The results have been compared with an empirical model given by Gopalswamy, et al. (2000; 2001) to predict the 1-AU arrival time of CMEs. In this paper, we use an expected data set with a wider range with initial velocities than that considered in previous models. To improve the accuracy of the predicted arrival time, we divided the CME events into two groups according to their effective acceleration and deceleration. The results show that our model works well for events with a negative acceleration in the initial velocity range between 500 and 2500 km/s, while the model described by Gopalswamy is better for events with initial velocities near the solar wind velocity. Published in Russian in Astronomicheskii Vestnik, 2009, Vol. 43, No. 2, pp. 137–144. The text was translated by the authors.     

Filament Eruption in NOAA 11093 Leading to a Two-Ribbon M1.0 Class Flare and CME

We present a multi-wavelength analysis of an eruption event that occurred in active region NOAA 11093 on 7 August 2010, using data obtained from SDO, STEREO, RHESSI, and the GONG Hα network telescope. From these observations, we inferred that an upward slow rising motion of an inverse S-shaped filament lying along the polarity inversion line resulted in a CME subsequent to a two-ribbon flare. Interaction of overlying field lines across the filament with the side-lobe field lines, associated EUV brightening, and flux emergence/cancelation around the filament were the observational signatures of the processes leading to its destabilization and the onset of eruption. Moreover, the time profile of the rising motion of the filament/flux rope corresponded well with flare characteristics, viz., the reconnection rate and hard X-ray emission profiles. The flux rope was accelerated to the maximum velocity as a CME at the peak phase of the flare, followed by deceleration to an average velocity of 590 km s⁻¹. We suggest that the observed emergence/cancelation of magnetic fluxes near the filament caused it to rise, resulting in the tethers to cut and reconnection to take place beneath the filament; in agreement with the tether-cutting model. The corresponding increase/decrease in positive/negative photospheric fluxes found in the post-peak phase of the eruption provides unambiguous evidence of reconnection as a consequence of tether cutting.     

Initial-Condition Influences on CME Expansion and Propagation

A general-purpose melon-seed-overpressure-expansion (MSOE) model is used to examine the dependence of the expansion and propagation of fast CMEs on average initial values of the magnetic field inside the CME, the magnetic field outside the CME, and the CME mass. The MSOE model extracts from the real situation only features that are essential in determining the expansion and propagation processes and idealizes them to obtain an adaptable “minimalist” set of equations that governs these processes yet retains enough relevant physics to give useful results. This minimalist set is used to carry out a systematic comparison of solutions against four sets of observed correlations between CME and ICME parameters in which the input parameters to the MSOE model are varied to achieve simultaneous fits to all four correlations. The fits impose relations between the three independent input variables: initial internal field strength, initial external field strength, and initial density. For example, light CMEs (e.g., those with no prominences) require the external field strength to nearly equal the internal field strength. However, heavy CMEs (e.g., those with prominences) require the external field strength to be much weaker than the internal field strength.     

Evolution of Coronal Mass Ejections in the Inner Heliosphere: A Study Using White-Light and Scintillation Images

Knowledge of the radial evolution of the coronal mass ejection (CME) is important for the understanding of its arrival at the near-Earth space and of its interaction with the disturbed/ambient solar wind in the course of its travel to 1 AU and further. In this paper, the radial evolution of 30 large CMEs (angular width > 150^∘, i.e., halo and partial halo CMEs) has been investigated between the Sun and the Earth using (i) the white-light images of the near-Sun region from the Large Angle Spectroscopic Coronagraph (LASCO) onboard SOHO mission and (ii) the interplanetary scintillation (IPS) images of the inner heliosphere obtained from the Ooty Radio Telescope (ORT). In the LASCO field of view at heliocentric distances R≤30 solar radii (R_⊙), these CMEs cover an order of magnitude range of initial speeds, V_CME≈260–2600 km s⁻¹. Following results have been obtained from the speed evolution of these CMEs in the Sun–Earth distance range: (1) the speed profile of the CME shows dependence on its initial speed; (2) the propagation of the CME goes through continuous changes, which depend on the interaction of the CME with the surrounding solar wind encountered on the way; (3) the radial-speed profiles obtained by combining the LASCO and IPS images yield the factual view of the propagation of CMEs in the inner heliosphere and transit times and speeds at 1 AU computed from these profiles are in good agreement with the actual measurements; (4) the mean travel time curve for different initial speeds and the shape of the radial-speed profiles suggest that up to a distance of ∼80 R_⊙, the internal energy of the CME (or the expansion of the CME) dominates and however, at larger distances, the CME's interaction with the solar wind controls the propagation; (5) most of the CMEs tend to attain the speed of the ambient flow at 1 AU or further out of the Earth's orbit. The results of this study are useful to quantify the drag force imposed on a CME by the interaction with the ambient solar wind and it is essential in modeling the CME propagation. This study also has a great importance in understanding the prediction of CME-associated space weather at the near-Earth environment.     

Kinematic Evolution of a Slow CME in Corona Viewed by STEREO-B on 8 October 2007

We studied the kinematic evolution of the 8 October 2007 CME in the corona based on observations from Sun – Earth Connection Coronal and Heliospheric Investigation (SECCHI) onboard satellite B of Solar TErrestrial RElations Observatory (STEREO). The observational results show that this CME obviously deflected to a lower latitude region of about 30° at the beginning. After this, the CME propagated radially. We also analyze the influence of the background magnetic field on the deflection of this CME. We find that the deflection of this CME at an early stage may be caused by a nonuniform distribution of the background magnetic-field energy density and that the CME tended to propagate to the region with lower magnetic-energy density. In addition, we found that the velocity profile of this gradual CME shows multiphased evolution during its propagation in the COR1-B FOV. The CME velocity first remained constant: 23.1 km s⁻¹. Then it accelerated continuously with a positive acceleration of ≈7.6 m s⁻².     

Bastille Day Event: A Radio Perspective

We describe the radio signatures that led up to and concluded the solar eruptive event of 14 July 2000 (Bastille Day Event). These radio signatures provide a means of remotely sensing the associated solar activity and transient phenomena. For many days prior to the Bastille Day Event kilometric Type III radio storm emissions were observed that were presumably associated with the active region NOAA 9077. These storm emissions continued until the X5.7 flare at ∼ 10 UT on 14 July 2000 that characterized the Bastille Day Event, then ceased abruptly. The Bastille Day Event itself produced very intense, complex, long-duration Type III-like radio emissions, which appear to have been associated with electrons generated (accelerated) deep in the solar corona. The coronal mass ejection (CME) associated with the Bastille Day Event generated decametric to kilometric Type II radio emissions as the CME propagated through the solar corona and interplanetary medium. The frequency drift of these Type II radio emissions are related to the dynamics of the propagating CME and indicate that the CME experienced significant deceleration as it propagated from the high corona into the interplanetary medium.     

Correlation Analyses Between the Characteristic Times of Gradual Solar Energetic Particle Events and the Properties of Associated Coronal Mass Ejections

It is generally believed that gradual solar energetic particles (SEPs) are accelerated by shocks associated with coronal mass ejections (CMEs). Using an ice-cream cone model, the radial speed and angular width of 95 CMEs associated with SEP events during 1998 – 2002 are calculated from SOHO/LASCO observations. Then, we investigate the relationships between the kinematic properties of these CMEs and the characteristic times of the intensity-time profile of their accompanied SEP events observed at 1 AU. These characteristic times of SEP are i) the onset time from the accompanying CME eruption at the Sun to the SEP arrival at 1 AU, ii) the rise time from the SEP onset to the time when the SEP intensity is one-half of peak intensity, and iii) the duration over which the SEP intensity is within a factor of two of the peak intensity. It is found that the onset time has neither significant correlation with the radial speed nor with the angular width of the accompanying CME. For events that are poorly connected to the Earth, the SEP rise time and duration have no significant correlation with the radial speed and angular width of the associated CMEs. However, for events that are magnetically well connected to the Earth, the SEP rise time and duration have significantly positive correlations with the radial speed and angular width of the associated CMEs. This indicates that a CME event with wider angular width and higher speed may more easily drive a strong and wide shock near to the Earth-connected interplanetary magnetic field lines, may trap and accelerate particles for a longer time, and may lead to longer rise time and duration of the ensuing SEP event.     

Sources of SEP Acceleration during a Flare – CME Event

A high-speed, halo-type coronal mass ejection (CME), associated with a GOES M4.6 soft X-ray flare in NOAA AR 0180 at S12W29 and an EIT wave and dimming, occurred on 9 November 2002. A complex radio event was observed during the same period. It included narrow-band fluctuations and frequency-drifting features in the metric wavelength range, type III burst groups at metric – hectometric wavelengths, and an interplanetary type II radio burst, which was visible in the dynamic radio spectrum below 14 MHz. To study the association of the recorded solar energetic particle (SEP) populations with the propagating CME and flaring, we perform a multi-wavelength analysis using radio spectral and imaging observations combined with white-light, EUV, hard X-ray, and magnetogram data. Velocity dispersion analysis of the particle distributions (SOHO and Wind in situ observations) provides estimates for the release times of electrons and protons. Our analysis indicates that proton acceleration was delayed compared to the electrons. The dynamics of the interplanetary type II burst identify the burst source as a bow shock created by the fast CME. The type III burst groups, with start times close to the estimated electron-release times, trace electron beams travelling along open field lines into the interplanetary space. The type III bursts seem to encounter a steep density gradient as they overtake the type II shock front, resulting in an abrupt change in the frequency drift rate of the type III burst emission. Our study presents evidence in support of a scenario in which electrons are accelerated low in the corona behind the CME shock front, while protons are accelerated later, possibly at the CME bow shock high in the corona.     

On the Coronal Mass Ejection onset and Coronal Dimming

A comprehensive case and statistical study of CME onsets has been conducted on the solar limb using the CDS, LASCO and EIT instruments aboard the SOHO spacecraft. This is the first dedicated campaign to establish firmly the EUV signatures of CME onsets and is based on a series of low-corona observing campaigns made in 2002. The event database consisted of 36 multiple emission line sequences observed with CDS and the study builds, in particular, on studies of EUV coronal dimming which have been associated with CME onsets. We witness a range of dimming events in EUV coronal emission line data. Shorter events, commonly of duration < 4 hours, we find are indirectly associated with CME onsets whereas longer-duration dimmings (> 4 hours) appear to be either due to coronal evolution or rotational effects. However, for some CME onsets, where the CDS pointing was appropriate, no dimming was observed. Dimming observed in EIT typically occurred immediately after the launch of a loop or prominence, and in 5 out of 9 events there is evidence of a matter buildup within the loop before launch. A total of 10 events occurred where CDS was used to directly observe the CME footprint, but no relationship between these events was found. The results suggest that the response of the corona to a CME launch differs between the low (1.0 R _⊙≤R≤1.2 R _⊙) and middle (1.2 R _⊙<R≤2.0 R _⊙) corona regions, hence implying a difference between dimming observations conducted with different instruments.     

Evidence for Tether-Cutting Reconnection in a Quadrupole Magnetic Configuration in the April 9, 2001, M7.9 Flare

We studied the M7.9 flare on April 9, 2001 that occurred within a δ-sunspot of active region NOAA 9415. We used a multi-wavelength data set, which includes Yohkoh, TRACE, SOHO, and ACE spacecraft observations, Potsdam and Ondřejov radio data and Big Bear Solar Observatory (BBSO) images in order to study the large-scale structure of this two-ribbon flare that was accompanied by a very fast coronal mass ejection (CME). We analyzed light curves of the flare emission as well as the structure of the radio emission and report the following: the timing of the event, i.e., the fact that the initial brightenings, associated with the core magnetic field, occurred earlier than the remote brightening (RB), argue against the break-out model in the early phase of this event. We thus conclude that the M7.9 flare and the CME were triggered by a tether-cutting reconnection deep in the core field connecting the δ-spot and this reconnection formed an unstable flux rope. Further evolution of the erupted flux rope could be described either by the “standard“ flare model or a break-out type of the reconnection. The complex structure of flare emission in visible, X-ray, and radio spectral ranges point toward a scenario which involves multiple reconnection processes between extended closed magnetic structures.     

Perfect Fluid Lrs Bianchi I With Time Varying Constants

It is investigated the behaviour of the “constants” G, c and Λ in the framework of a perfect fluid LRS Bianchi I cosmological model. It has been taken into account the effects of a c-variable into the curvature tensor. Two exact cosmological solutions are investigated, arriving t the conclusion that if q < 0 (deceleration parameter) then G, c are growing functions on time t while Λ is a negative decreasing function on time.     

Analysis of Solar Wind Events Using Interplanetary Scintillation Remote Sensing 3D Reconstructions and Their Comparison at Mars

Interplanetary Scintillation (IPS) allows observation of the inner heliospheric response to corotating solar structures and coronal mass ejections (CMEs) in scintillation level and velocity. With colleagues at STELab, Nagoya University, Japan, we have developed near-real-time access of STELab IPS data for use in space-weather forecasting. We use a 3D reconstruction technique that produces perspective views from solar corotating plasma and outward-flowing solar wind as observed from Earth by iteratively fitting a kinematic solar wind model to IPS observations. This 3D modeling technique permits reconstruction of the density and velocity structure of CMEs and other interplanetary transients at a relatively coarse resolution: a solar rotational cadence and 10° latitudinal and longitudinal resolution for the corotational model and a one-day cadence and 20° latitudinal and longitudinal heliographic resolution for the time-dependent model. This technique is used to determine solar-wind pressure (“ram” pressure) at Mars. Results are compared with ram-pressure observations derived from Mars Global Surveyor magnetometer data (Crider et al. 2003, J. Geophys. Res. 108(A12), 1461) for the years 1999 through 2004. We identified 47 independent in situ pressure-pulse events above 3.5 nPa in the Mars Global Surveyor data in this time period where sufficient IPS data were available. We detail the large pressure pulse observed at Mars in association with a CME that erupted from the Sun on 27 May 2003, which was a halo CME as viewed from Earth. We also detail the response of a series of West-limb CME events and compare their response observed at Mars about 160° west of the Sun – Earth line by the Mars Global Surveyor with the response derived from the IPS 3D reconstructions.     

Mass Loss from Rotating Hot-stars: Inhibition of Wind Compressed Disks by Nonradial Line-forces

We review the dynamics of radiatively driven mass loss from rapidly rotating hot-stars. We first summarize the angular momentum conservation process that leads to formation of a Wind Compressed Disk(WCD) when material from a rapidly rotating star is driven gradually outward in the radial direction. We next describe how stellar oblateness and asymmetries in the Sobolev line-resonance generally leads to nonradialcomponents of the driving force is a line-driven wind, including an azimuthal spin-down force acting against the sense of the wind rotation, and a latitudinal force away from the equator. We summarize results from radiation-hydrodynamical simulations showing that these nonradial forces can lead to an effective suppressionof the equatorward flow needed to form a WCD, as well as a modest (∼ 25%) spin-downof the wind rotation. Furthermore, contrary to previous expectations that the wind mass flux should be enhanced by the reduced effective gravity near the equator, we show here that gravity darkening effects can actually lead to a reducedmass loss, and thus lower density, in the wind from the equatorial region. Finally, we examine the equatorial bistability model, and show that a sufficiently strong jump in wind driving parameters can, in principle, overcome the effect of reduced radiative driving flux, thus still allowing moderate enhancements in density in an equatorial, bistability zone wind. This revised version was published online in July 2006 with corrections to the Cover Date.     

Coronal Dimmings and the Early Phase of a CME Observed with STEREO and <Emphasis Type="Italic">Hinode</Emphasis>/EIS

We investigate the early phase of the 13 February 2009 coronal mass ejection (CME). Observations with the twin STEREO spacecraft in quadrature allow us to compare for the first time in one and the same event the temporal evolution of coronal EUV dimmings, observed simultaneously on-disk and above-the-limb. We find that these dimmings are synchronized and appear during the impulsive acceleration phase of the CME, with the highest EUV intensity drop occurring a few minutes after the maximum CME acceleration. During the propagation phase two confined, bipolar dimming regions, appearing near the footpoints of a pre-flare sigmoid structure, show an apparent migration away from the site of the CME-associated flare. Additionally, they rotate around the ‘center’ of the flare site, i.e., the configuration of the dimmings exhibits the same ‘sheared-to-potential’ evolution as the postflare loops. We conclude that the motion pattern of the twin dimmings reflects not only the eruption of the flux rope, but also the ensuing stretching of the overlying arcade. Finally, we find that: i) the global-scale dimmings, expanding from the source region of the eruption, propagate with a speed similar to that of the leaving CME front; ii) the mass loss occurs mainly during the period of strongest CME acceleration. Two hours after the eruption Hinode/EIS observations show no substantial plasma outflow, originating from the ‘open’ field twin dimming regions.     

An Attempt to Detect Coronal Mass Ejections in Lyman-α Using SOHO Swan

In this study, the possibility that coronal mass ejections (CMEs) may be observed in neutral Lyman-α emission was investigated. An observing campaign was initiated for SWAN (Solar Wind ANisotropies), a Lyman-α scanning photometer on board the Solar and Heliospheric Observatory (SOHO) dedicated to monitoring the latitude distribution of the solar wind from its imprints on the interstellar sky background. This was part of SOHO Joint Observing Program (JOP) 159 and was an exploratory investigation as it was not known how, or even if, CMEs interact with the solar wind and interstellar neutral hydrogen at this distance (≈60 and 120 R _S). The study addresses the lack of methods for tracking CMEs beyond the field-of-view of current coronagraphs (30 R _S). In our first method we used LASCO, white-light coronagraphs on SOHO, and EIT, an extreme ultraviolet imaging telescope also on SOHO, to identify CME candidates which, subject to certain criteria, should have been observable in SWAN. The criteria included SWAN observation time and location, CME position angle, and extrapolated speed. None of the CME candidates that we discuss were identified in the SWAN data. For our second method we analyzed all of the SWAN data for 184 runs of the observing campaign, and this has yielded one candidate CME detection. The candidate CME appears as a dimming of the background Lyman-α intensity representing ≈10% of the original intensity, moving radially away from the Sun. Multiple candidate CMEs observed by LASCO and EIT were found which may have caused this dimming. Here we discuss the campaign, data analysis technique and statistics, and the results.     

Hα Dimming Associated With the Eruption of a Coronal Sigmoid in the Quiet Sun

We report on the occurrence of Hα dimming associated with a sigmoid eruption in a quiet-sun region on 14 August 2001. The coronal sigmoid in soft X-ray images from the Yohkoh Soft X-ray Telescope was located over an Hα filament channel. Its eruption was accompanied by a flare of GOES X-ray class C2.3 and possibly associated with a halo coronal mass ejection (CME) observed with the Large Angle and Spectroscopic Coronagraphs (LASCO) on board the Solar and Heliospheric Observatory (SOHO). During the eruption, coronal bipolar double dimming took place at the regions with opposite magnetic polarities around the two sigmoid ends, but the underlying chromospheric channel did not show observable changes corresponding to the coronal eruption. Different from the erupting coronal sigmoid itself, however, the coronal dimming had a detectable chromosphere counterpart, i.e., Hα dimming. By regarding the sigmoid as a coronal sign for a flux rope, these observations are explained in the framework of the flux rope model of CMEs. The flux rope is possibly deeply rooted in the chromosphere, and the coronal and Hα dimming regions mark its evacuated feet, through which the material is possibly fed to the halo CME.     

Propagation of the Bastille Day 2000 CME Shock in the Outer Heliosphere

The Bastille Day (14 July) 2000 CME is a fast, halo coronal mass ejection event headed earthward. The ejection reached Earth on 15 July 2000 and produced a very significant magnetic storm and widespread aurora. At 1 AU the Wind spacecraft recorded a strong forward shock with a speed jump from ∼ 600 to over 1000 km s⁻¹. About 6 months later, this CME-driven shock arrived at Voyager 2 (∼ 63 AU) on 12 January 2001 with a speed jump of ∼ 60 km s⁻¹. This provides a good opportunity to study the shock propagation in the outer heliosphere. In this study, we employ a 2.5-D MHD numerical model, which takes the interaction of solar wind protons and interstellar neutrals into account, to investigate the shock propagation in detail and compare the model predictions with the Voyager 2 observations. The Bastille Day CME shock undergoes a dramatic change in character from the inner to outer heliosphere. Its strength and propagation speed decay significantly with distance. The model results at the location of Voyager 2 are in good agreement with in-situ observations. Supplementary material to this paper is available in electronic form at http://dx.doi.org/10.1023/A:1014293527951