1 Introduction

The term interplanetary scintillation (IPS) phenomenon refers to the random fluctuations in the intensity flux of radio waves propagating from a distant compact radio source. This phenomenon of radio wave scattering is caused by electron-density fluctuations in the solar wind [ΔN e ]. In brief, electron-density irregularities modify the plane wave fronts from a distant radio source, causing a diffraction pattern that moves across the antenna system at a speed close to that of the solar wind (Rickett and Coles 1991).

IPS is a remote-sensing technique that allows us to monitor regions in the inner heliosphere that are inaccessible to spacecraft. Figure 1 shows the basic geometry of IPS observations. Since the turbulence density of interplanetary plasma rapidly decreases with heliocentric distance [≈ R −4], there is a limited effective scattering region for every elongation angle [ϵ]. This region is located around the minimum distance between the line of sight (LOS) and the Sun (P-point in Figure 1). The strongest scintillation is associated with P and, by geometry, the heliocentric distance to P [AU] is: p=sinϵ (Manoharan and Ananthakrishnan 1990).

Figure 1
figure 1

Configuration of IPS observations in the plane that includes the Sun, Earth, and the line of sight (LOS) to the radio source. In this scenario, there is a plasma cloud (solar wind or ICME) crossing the LOS. The angle between Sun, Earth, and LOS is the elongation angle [ϵ]. The dotted line is the perpendicular distance [p] that corresponds to the minimum distance between the Sun and LOS, where at around P-point the contribution to the diffraction pattern is strongest.

Full size image

The scintillation index is related to electron-density fluctuations [ΔN e ] along the LOS of the radio source. This index varies with heliocentric distance, reaching the highest value at a certain distance from the Sun. This distance depends on the observational frequency and indicates the transition between the weak- and the strong-scattering region nearest to the Sun (Manoharan 1993).

The IPS technique assumes that ΔN e is proportional to density variations in the solar wind. Strong increments in solar-wind density may be associated with two general types of large-scale solar-wind perturbations: i) the region around a stream interface between a fast solar-wind stream overtaking a slow ambient wind (stream interaction region), and ii) density enhancements associated with solar transient events such as interplanetary shocks and plasma sheaths driven by fast interplanetary coronal mass ejections (ICMEs). We can use IPS observations to identify large-scale solar-wind compressive perturbations propagating from the Sun to 1 AU (Hewish and Bravo 1986; Manoharan et al. 1995; Janardhan et al. 1996; Tokumaru et al. 2003b; Manoharan 2006).

This article has two aims: to report i) the first measurements around the solar maximum at 140 MHz of the variations of the scintillation index with respect to the heliocentric distance for two strong IPS radio sources, and ii) the first identifications of ICMEs detected by MEXART during the solar maximum of Cycle 24. The outline of the article is as follows: Section 2 describes the MEXART characteristics and the observations for this study, Section 3 presents the variations of the scintillation index with respect to the heliocentric distance for 3C48 and 3C147, Section 4 reports the ICMEs detected by MEXART, Section 5 presents the discussion, and Section 6 the conclusions.

2 MEXART Observations

MEXART is a single-station observatory dedicated to performing solar-wind studies employing the IPS technique. It is located in the state of Michoacan, Mexico (19∘48′ N, 101∘41′ W), at an altitude of about 1960 meters. This radio telescope has an operational frequency of 139.65 MHz with a bandwidth of 1.5 MHz, the sampling rate is 50 measurements per second, and the receiver time constant is 47 ms. The time constant is the interval required to reconstruct the signal (integration time). The antenna is a plane array of 64 east–west lines of 64 dipoles each (64×64=4096 dipoles in total), covering a physical area of about 9600 m2. MEXART is a transit radio telescope deploying 16 fixed latitudinal beams produced by a Buttler matrix. The instrument uses the Earth’s rotation to scan the sky daily, detecting the transit of several cosmic radio sources that are well known in the IPS catalogs. A detailed description of the array is provided elsewhere (Gonzalez-Esparza et al. 2004; Mejia-Ambriz et al. 2010). The data employed for this study were obtained during the 2013 and 2014 observational campaigns. These observations were made employing half of the total array (2048 dipoles≈4800 m2). Table 1 shows the radio source list followed by MEXART and the observational intervals.

Table 1 IPS radio sources used for this study as followed by MEXART during 2013 and 2014. The columns correspond to the source coordinates (right ascension and declination), type (QSO=quasar) and electromagnetic flux at 159 MHz (3C catalog). The final column corresponds to the period when the radio source has elongation angles smaller than 90 degrees.
Full size table

3 Variation of Scintillation Index with Heliocentric Distance; Observations at 140 MHz

We employed a new methodology based on the wavelet transform (WT) function to calculate the scintillation index (Aguilar-Rodriguez et al. 2014). This technique allows us to obtain, in a straightforward way, a normalized scintillation index and the frequency power spectrum (Fourier knee). The first step in our methodology is to remove low-frequency signals in the source flux record (time series). We subtracted a running average of 10 seconds (500 data points for 0.02 seconds of temporal resolution). This procedure acts as a high-pass filter that cleans the data from possible ionospheric scintillation. After that, we calculate the WT of the data series and plot it to generate the WT images. These images are similar to a dynamic spectrum where the x-axis is the time, y-axis is the period or frequency, and the power associated with every frequency is indicated in a color code. In this way, we can identify and select the interval where the IPS fluctuations dominate and remove intervals with interference from the data series. In general, observing at 140 MHz, the IPS frequency domain covers a region within 0.3 – 1 Hz. From the WT image we selected and calculated the average power within this range of IPS frequencies. On the other hand, we selected, in the same time interval, another high-frequency range (2<f<10 Hz), which corresponds to the instrumental noise, and also calculated its average power. Finally, the normalized scintillation index of a given source observation is the average power of the IPS range minus the average power of the instrumental noise range (see Figure 2 of Aguilar-Rodriguez et al. 2014).

Employing the methodology described above, we calculated the scintillation indices for two strong radio sources (3C48 and 3C147) as detected at different elongation angles (see Table 1). We performed the observations in 2013 and 2014. Figure 2 shows the variation of the scintillation normalized index with the heliocentric distance [p] for the two IPS sources. The log–log plots show a distribution of points with an expected tendency and dispersion. A simple power law gives a good fit in the weak-scattering regime, and the radial dependence of the scintillation index can be expressed as m=m 0 p −b, where p is in AU, and m 0 is the scintillation index at 1 AU (Rao, Bhandari, and Ananthakrishnan 1974; Manoharan 1993). We applied a least-squares fit to each log–log plot to obtain the power index [b]. Figure 2a shows normalized scintillation indices of 3C48 obtained from the 2013 measurements at different heliocentric distances, where we obtained b=−1.26. Figure 2b shows the scintillation indices of the same source from 2014 measurements, and in this case, we obtained b=−2.02. The fitting in the two log–log plots gave different results in the two years. When we joined the two 3C48 data sets, 2013 and 2014 (Figure 2d), we obtained b=−1.69. Figure 2c presents the corresponding log–log plot for 3C147 based on MEXART observations in 2014, where we obtained b=−1.88. In general, the indices of the power-law functions had values similar to those reported before at other observational frequencies (Manoharan 1993; Fallows, Williams, and Breen 2002; Manoharan 2006; 2010).

Figure 2
figure 2

Variation curves of normalized scintillation index as a function of heliocentric distance [p=sinϵ] (log–log plots). The panels refer to the normalized scintillation index corresponding to 2013 and 2014 observations, for the astronomical radio sources 3C48 and 3C147. We applied a least-squares fit to each plot to obtain a power-law function and its index [m=m 0 p −b]. The boxes in the panels indicate the strongest scintillations in our data.

Full size image

These power-law function indices commented on here represent the average density fluctuations in the inner heliosphere (i.e. turbulence level for the “quiet” (undisturbed) solar-wind (Tokumaru et al. 2003b)). These results can be compared with previous studies at other observational frequencies and studying other sources, where other authors reported power-law function indices between −1.3 and −2.2 (Manoharan 1993; Fallows, Williams, and Breen 2002; Manoharan 2006; 2010). Based on this yearly fitting, we can estimate the g-scintillation index, which is the scintillation index normalized with respect to the average scintillation value at the given elongation angle (Gapper et al. 1982; Manoharan et al. 1995). This index allows us to distinguish regions (in the inner heliosphere) where the density fluctuations are stronger (g>1, compression regions) or weaker (g<1, rarefaction regions) than the expected density fluctuations of the ambient solar-wind (g≈1).

4 First ICMEs Detected in MEXART IPS Data

We assumed that there are frequent fast ICMEs around solar maximum, driving interplanetary shocks that cause large-scale regions of significant density enhancements in the interplanetary medium. Figure 2 shows some scintillation data points that appreciably exceed the least-squares fitting (i.e. the expected scintillation average value), indicating that they were associated with strong compressive events in the solar wind. These events by definition have g-indices greater than one. We identified 17 of these IPS events [g>1.5] and investigated whether they could be associated with preceding CME counterparts.

To associate observations of IPS with CME counterparts, we should keep in mind that IPS solar-wind observations are significantly affected by two effects: i) the LOS integration, and ii) the plane-of-sky projection (Tokumaru et al. 2003a). To estimate the solar-wind velocity using IPS results, we implicitly assumed that the velocity is radially directed and is constant along the LOS. Since the scattering power (of the electron-density irregularities in the solar wind) falls off with heliocentric distance approximately as R −4, observations of IPS are heavily weighted around the P-point with an asymmetric Gaussian weighting, and most of contribution comes from near to the point of closest approach (P-Point) (Manoharan 1997; Fallows, Williams, and Breen 2002). On the other hand, with respect to the projection effect, the plane-of-the-sky and the plane of the IPS geometry (i.e. the plane perpendicular to the discontinuous p-line in Figure 1) are close to one another only when the elongation angle [ϵ] is small. However, when ϵ>30∘ (R>0.5 AU), the plane of the IPS geometry moves closer to the Earth–Sun line.

The first step to associate observations of IPS and CME counterparts was to estimate the solar-wind speed from our IPS data by applying the single-station technique [V IPS] (Manoharan and Ananthakrishnan 1990). We used the power spectra (Fourier knee) obtained from the WT and the velocity model fitting from the single-station technique (see Aguilar-Rodriguez et al. 2015 for further details). For the velocity model fitting, we used a radio source angular width = 0.25 arcseconds for 3C48 and = 0.35 arcseconds for 3C147 (Janardhan and Alukar 1993). We computed the solar-wind speed for each of our observations of IPS, with an associated occurrence time and heliocentric distance (elongation angle) (Table 2). Based on this initial information, we established an onset temporal window for a possible preceding CME counterpart taking in account four possible scenarios: i) CMEs that according to their initial speed might be accelerated or decelerated due to its interaction with the background solar-wind, ii) CMEs propagating close to their initial speed (which was closer to the background solar-wind speed), iii) fast CMEs that might be decelerated by catching up to a slower preceding CME (marked as a complex event in Table 2), and iv) slow CMEs that might be accelerated after being overtaken by a subsequent faster CME (also marked as complex events in Table 2). We selected as candidates for CME counterparts those for which we found the best travel-time agreement with the IPS observation after we considered the different cases described above.

Table 2 List of CME–IPS events. The columns from left to right correspond to association confidence [AC] parameter, CME date, CME angular width [AW], CME polar angle [PA], CME type [H: halo CME, PH: partial halo CME, L: limb CME, m: minor CME], CME initial velocity [V 0], IPS radio source observed, IPS observation date, IPS source heliocentric distance [R], IPS source polar [PAIPS], g, IPS velocity [V IPS], and transit velocity [〈V t 〉] between the CME onset time and IPS event time.
Full size table

The heliographic location of the P-point of the radio source (all-sky, Figures 3b and 4b) indicates the location of the source with respect to eastern or western sides of the Sun. We searched the SOHO/LASCO data for CMEs with a reasonable agreement given by this rough propagation direction (eastern or western side) and the temporal window given by the occurrence and velocity of the IPS event. These CME counterpart candidates also required an initial velocity [V 0, from the coronograph field of view] consistent with the IPS solar-wind speed at the P-point considering the four possible scenarios described above. Finding these feasible CME counterparts, we additionally checked that the transit velocity [〈V t 〉] (given by the CME onset time seen by LASCO, the IPS event time at MEXART, and the heliocentric distance of the P-point [p=sinϵ]) was physically consistent with the other CME and IPS velocities.

Figure 3
figure 3

Association of MEXART IPS observation on 23 May 2013 with a preceding CME counterpart detected on 22 May 2013 by SOHO/LASCO. Panel a shows the white-light image from SOHO/LASCO, (b) all-sky map of the heliographic location of 3C48 (blue diamond) at the moment of the MEXART observation, (c) wavelet analysis of the IPS record image, and (d) power spectrum and the model fitting to calculate the solar-wind speed, where we indicate the parameters employed to derive the fitting.

Full size image
Figure 4
figure 4

Association of MEXART IPS observation on 25 July 2014 with a preceding CME counterpart detected on 22 July 2014 by SOHO/LASCO. Panel a shows the white-light image from SOHO/LASCO, (b) all-sky map of the heliographic location of 3C147 (blue diamond) at the moment of the MEXART observation, (c) wavelet image of the IPS record image, and (d) power spectrum and the model fitting to calculate the solar-wind speed, where we indicate the parameters employed to derive the fitting.

Full size image

Table 2 shows the IPS events detected by MEXART and their probable preceding CME counterparts associated with SOHO/LASCO observations. In all cases, by design, the CME initiation at the solar surface and its propagation direction are consistent with the heliographic latitude of the IPS source. On the other hand, the initial CME velocity given by SOHO/LASCO observations [V 0] agrees reasonably well with the IPS speed at the P-point calculated from MEXART observations [V IPS]. The CME data of the events in 2013 were obtained from the SOHO/LASCO catalog ( cdaw.gsfc.nasa.gov/CME_list/ ); whereas the CME data in 2014 were taken from the CACTUS LASCO catalog ( sidc.oma.be/cactus/ ). The first column of Table 2 refers to the association confidence (AC) parameter, where we estimated how confident we are in the CME–IPS event association. Columns 2 – 6 present the CME characteristics (date, hour, angular width, position angle, type, and onset velocity). Subsequently, columns 7 – 13 show the IPS event characteristics (source, date, hour, heliocentric distance, position angle, g-index, and solar-wind velocity). The last column presents the calculated transit speed [〈V t 〉] from this CME–IPS association. For the events with higher association confidence (AC=1) we found a good agreement between all the parameters of a unique CME and the IPS event. For the events with a medium AC (= 2) we did not find such a good agreement in all the parameters between the CME and IPS characteristics (although it is reasonable). For events with a low AC (= 3), we were unable to associate a unique CME because there were various CMEs propagating in similar directions around the expected CME onset time (complex events). For these cases, we report the most likely CME involved; however, these IPS events were most likely associated with a combination of ICMEs interacting in the interplanetary medium (complex CME events marked with an asterisk in Table 2).

To first approximation, the CME events shown in Table 2 correspond to the most likely CME counterparts. Nevertheless, it would be necessary to use a 3D CME model to corroborate these associations. We present two case events to illustrate our methodology.

4.1 Case event 1. CME on 22 May 2013

Figure 3 shows the analysis of the CME limb event on 22 May 2013. Figure 3a presents the LASCO-C2 image. The SOHO/LASCO Catalog reports the CME onset time at 08:48 UT with an initial speed of 687±196 km s−1 (the uncertainty in the velocity was obtained from sharpp.nrl.navy.mil/cgi-bin/swdbi/lasco/cmes/form , where we had used the initial and final speed estimated as a proxy for the measurement uncertainty). Figure 3b indicates the heliographic location of the IPS source (3C48) on the all-sky map (blue diamond), were the P-point was at a radial distance of 0.59 AU with a heliographic latitude of + 19 degrees on the western hemisphere. The LOS of the IPS source intersects the CME propagation direction; however, we should keep in mind that in this case ϵ=35∘ and the plane of the IPS geometry is oriented closer to the Earth–Sun line. The CME angular width was about 210∘. The MEXART IPS event occurred on 23 May 2013 at 16:18 UT. From our analysis we obtained a IPS solar-wind speed of about 575±45 km s−1. The transit speed [〈V t 〉] from this CME–IPS association was 778 km s−1 (suggesting a clear deceleration in the interplanetary medium). Figure 3c shows the WT analysis of the IPS record, indicating the location of the IPS frequency box and the instrumental noise frequency box that we used to calculate the scintillation index and the frequency power spectra (see methodology above). Figure 3d shows the model fitting to the observed power spectra (Fourier knee) and the resulting solar-wind speed and its associated uncertainty. We characterized this event with a low association confidence (AC=3). This is because although the velocities [V 0, 〈V t 〉, V IPS] agreed well, there was a subsequent fast halo-CME (V=1466 km s−1) with an onset time of about five hours later. This following, faster halo CME might have interacted partially with our slower CME counterpart along the LOS of the radio source. In this case, we cannot associate a unique CME event with the IPS observation.

4.2 Case event 2. CME on 22 July 2014

Figure 4 shows the analysis of the CME limb event on 22 July 2014. Figure 4a shows the LASCO C2 image. The CACTUS Catalog reports the CME start time at 18:00 UT with an initial speed of 437±92 km s−1. Figure 4b shows the heliographic location of the IPS source (3C147), where the P-point was at a heliocentric distance of 0.68 AU (ϵ=42∘) with a heliographic latitude of 47 degrees on the western hemisphere. As in the previous event, the LOS of the IPS source intersected the CME propagation direction. The MEXART IPS event occurred on 25 July 2014 at 16:16 UT. From our analysis we obtained an IPS solar-wind speed of about 420±70 km s−1. The transit speed [〈V t 〉] from the CME initiation to the MEXART IPS detection was 402 km s−1. Figure 4c shows the WT analysis of the IPS record, and Figure 4d shows the fit of the velocity model to observed power spectra. We characterized this event with a high association confidence (AC=1) because the three velocities [V 0, 〈V t 〉, V IPS] agree well and we did not find other CMEs that could have interacted with our CME counterpart along the LOS of the source. This was a medium-speed limb CME propagating near the expected ambient solar-wind velocity, which did not suffer a significant deceleration in the interplanetary medium.

5 Discussion

MEXART has a unique operation frequency (140 MHz) and geographic longitude to perform observations of IPS, and MEXART observations of IPS sources can be compared with previous studies using different frequencies. In Section 3, we presented the analysis of the scintillation index where we found an expected tendency and variation in the log–log plots (Figure 2). The yearly tendency indicates that, in general, the IPS sources scintillate more strongly closer to the Sun (all of the observations were made within the weak-scattering region). The observations were taken during the solar maximum of Cycle 24. The two IPS sources covered a wide range of heliolatitudes (0∘−65∘) during the interval of observations. In the two cases, large elongation angles (ϵ≈80∘) correspond to positions close to the Ecliptic (≈5∘), whereas small elongations angles (ϵ≈23∘) correspond to positions out of the Ecliptic (≈ 66∘). The dispersion of the scintillation indices with respect to the least-squares fit reveals fluctuations from the ambient density variations at that particular heliocentric distance (i.e. variations from the expected scintillation of the quiet solar wind). From the least-squares fitting in the variation plots we obtained a power-law function and its index, which characterize the yearly scintillation behavior of that particular IPS source. We used this power-law function to calculate the g-index of our observations of IPS. We found a significant variation of the power-law function index of 3C48 comparing 2013 (b=−1.26) and 2014 (b=−2.02) observations. On the other hand, when we compare the results from 2014 observations, we found similar power-law function indices for 3C48 (b=−2.02) and 3C147 (b=−1.88).

In Section 4 we used the g-index to identify 17 IPS events with the strongest scintillation in our data, and we investigated whether they could be associated with preceding CME counterparts (Table 2). For 3 out of the 17 IPS events we were unable to find a possible CME counterpart. All of the preceding CMEs in Table 2 show a very well defined and massive front in SOHO/LASCO observations and, in all of the events, the ICME front crossed the LOS of the IPS source. For the 14 IPS events associated with CME counterparts, five had the highest association confidence parameter (AC=1), two of which were associated with halo CMEs, one with a partial halo CME, and two with CME limb-type events. On the other hand, most of the IPS events associated with CME counterparts (8 out of 14) were related with complex CME events. Although we were expecting that most of our strong IPS events would be related to fast CMEs driving interplanetary shocks (solar maximum of Cycle 24), all of the events in Table 2 had CME onset velocities lower than 1000 km s−1 and IPS velocities lower than 700 km s−1. In general, we do not expect strong IPS scintillation from slow- or medium-speed CMEs because these events do not cause significant compression in the solar wind. However, since most of our IPS events were associated with complex CMEs, this suggests that two or more ICMEs intersecting the LOS of the IPS source interacted, and this interaction caused the compression region around the P-point that produced the strong scintillation.

Figure 5 shows 14 velocity histograms of the IPS events associated with CME counterparts in Table 2. Each histogram has three bars corresponding to the three velocities associated with each event: CME initial velocity [V 0], transit velocity [〈V t 〉], and IPs velocity [V IPS]. The plots also indicate the error in V 0 and V IPS. In general, for a fast CME decelerating in the interplanetary medium, we expect that V 0>〈V t 〉>V IPS. Each histogram shows the association confidence parameter [AC] and the heliocentric distance [p] of the P-point of each event. For the five IPS events with AC=1 (IPS event associated with a unique CME), we obtained the expected decreasing velocity tendency commented on before. We found this trend even in those cases where the CME initial velocity was close to the expected ambient solar-wind speed (V 0<450 km s−1). This tendency is related to the deceleration process of fast ICMEs in the interplanetary medium, as has been pointed out by several studies (Gopalswamy et al. 2000; Manoharan 2006; Corona-Romero and Gonzalez-Esparza 2011). We recall, however, that in this analysis we did not consider the projection effect and the LOS integration effect that produce uncertainties in the estimations of the CME and IPS velocities.

Figure 5
figure 5

Velocity histograms of the 14 IPS events associated with CME counterparts. The three bars in each histogram correspond to CME initial velocity [V 0], transit velocity [〈V t 〉], and IPs velocity [V IPS] (Table 2). The bars show the error in V 0 and V IPS. Each histogram indicates the association confidence parameter (AC) and the heliocentric distance [p] of the P-point for each event.

Full size image

6 Conclusions

We reported MEXART observations of IPS of two strong quasars, 3C48 and 3C147, during 2013 and 2014. We applied a methodology based on the wavelet transform to calculate g-scintillation indices and solar-wind velocities. We found, based on 3C48 observations, different average conditions of the ambient solar-wind density in 2013 and 2014. We reported the first ICMEs detected by MEXART, associating the IPS observations with preceding CME counterparts detected by SOHO/LASCO. For the IPS events associated with single CME counterparts, we found a deceleration tendency of the CMEs as they propagate in the interplanetary medium. In a future work, we will explore the applicability of a 3D CME model to validate the IPS-CME counterpart association. This study shows that this instrument, fully dedicated to IPS observations, is performing continuous and reliable solar-wind measurements. The MEXART will have new capabilities in 2015, upgrading the amplification from the antenna signal, doubling the antenna collection area, and having a new receiver with wider bandwidth. All of these improvements will allow the daily tracking of a few tens of IPS sources and the production of daily maps of g-values and solar-wind velocities.