3 Measurements and raw results

In this paper we investigate 18 coronal type II bursts recorded in the decimetric-metric (dm-m) wavelength range. The basic rule in selecting the events was that they clearly show a simple band-split pattern obeying the criteria defined in Paper I. Out of the 18 chosen events, 12 events were recorded by the radio spectrograph of the Astrophysikalisches Institut Potsdam covering the frequency range 40-800 MHz (Mann et al. 1992) and 6 were observed by Culgoora Solar Observatory radio spectrograph sweeping over the frequency range 18-1800 MHz (Prestage et al. 1994; see http://www.ips.oz.au/culgoora/spectro/index.html). The measurements were performed at the harmonic emission band since in dm-m type II bursts it is usually stronger and better defined then the fundamental band (see, e.g., Fig. 1 where a part of the fundamental band can be seen in the 50-40 MHz range around 10:00 UT). In some cases measurements of equal accuracy were possible at both emission bands, showing no significant difference in the outcome (see also Smerd et al. 1974, 1975; Mann et al. 1996).

In Fig. 1 the procedure of measurements is illustrated by showing one of the analysed type II bursts (see also Fig. 2 in Paper I). The LFB and UFB of the band-split emission are marked by the lines that follow the two emission ridges. The period of type II burst emission is divided into a number of roughly equidistant subintervals, defined by the times t₁ to t_N. At each of these N moments the frequencies of the emission maximum at LFB and UFB ($f_{\rm L}$ and $f_{\rm U}$, respectively) were measured, providing the relative band-splits BDW(t_i) defined by Eq. (1).

The number of $[f_{\rm U}(t_i)$, $f_{\rm L}(t_i)]$ data pairs varied from 3 to 22 per event, depending on the duration and frequency range covered by a particular type II burst. From the obtained data the frequency drifts:

$$% D_f\left(\overline t_i\right) \equiv \left(-\frac{\partial ... ... _i = \frac{f_{\rm L}(t_{i})-f_{\rm L}(t_{i+1})}{t_{i+1}-t_i},$$ (4)

and the relative band-splits:

$$% BDW_i\equiv BDW\left(\overline t_i\right)=\frac{BDW(t_{i+1})+BDW(t_{i})}{2}$$ (5)

are determined for the depicted time intervals, where $\overline t_i=$ (t_i+1+t_i)/2. These measurements are presented in Fig. 2, where the frequency drifts and band-splits are shown as a function of the LFB frequency $f_{\rm L}$.

Figure 2a shows that the frequency drift is a distinct power-law function of the frequency. In Paper I it was shown that the power-law dependence D_f(f) extends to the kilometric wavelength range. The slope of D_f(f) found therein for the frequency range from 10 MHz to 30 kHz is almost the same, amounting to $\approx$1.9.

Figure 2b shows no correlation between the band-split of metric type II bursts and the frequency. In the inset of Fig. 2b the distribution of values of BDW is presented showing that the majority of the analysed type II bursts has the relative band-split between 0.15 and 0.35.

Figure 2: Measured values of a) frequency drift and b) band-split, presented as a function of the (fundamental) frequency at LFB, $f_{\rm L}$. The power-law least squares fits and the corresponding correlation coefficients C are shown. The histogram in the inset in the lower panel shows the distribution of relative band-splits. The presented data are based on the measurements of 18 type II bursts.