4 Further data processing

Comparison of the resulting map with maps at higher frequencies suggested that the brightness temperature calibration of the 22 MHz map had a zenith-angle-dependent error. We used the 408 MHz all-sky survey of Haslam et al. (1982) for comparisons because these single-paraboloid observations would likely be free of such defects, but comparisons with other similar data sets would probably have produced similar conclusions.

For comparison with the 22 MHz map we convolved the 408 MHz data from its original resolution of 51' to the variable beamwidth of our data. Calculation of a map of spectral index between the two frequencies indicated a systematic variation of spectral index with declination. Since such an effect is unlikely to be real, we suspected a calibration problem in the 22 MHz data. Although the cause is not fully understood, we believe that the response of the telescope to extended structure differed from its response to point sources in directions away from the zenith. This effect is discussed in Sect. 6.2. We believe, however, that the response of the telescope at the zenith (declination 48.8$^\circ$) is well understood. At this declination, we plotted brightness temperature at 22 MHz against brightness temperature at 408 MHz (a T-T plot). We used data at all right ascensions except (a) near the Galactic plane in the Cygnus region (at $\sim$21$^{\rm h}$, 48$^\circ$) where strong absorption features are evident in the 22 MHz map (see below) and (b) regions around a few strong small-diameter sources. Figure 1 shows this plot. The highest temperatures in Fig. 1 correspond to the Galactic plane in the anticentre (at 4$^{\rm h}$ 40$^{\rm m}$,48$^\circ$) where there may be a small amount of absorption, causing the T-T plot to deviate from a straight line. Using all the data shown in Fig. 1 we derived a differential spectral index of 2.52 $\pm$ .02. Restricting the fit to regions with T₂₂<50 kK (corresponding to right ascensions between 5$^{\rm h}$ 40$^{\rm m}$ and 18$^{\rm h}$10$^{\rm m}$) the correlation is very tight and the differential spectral index is 2.57 $\pm$ .02. These values of spectral index are close to the value expected in this frequency range from the work of Bridle (1967) and Sironi (1974) indicating that the temperature scale at 22 MHz is accurate. Furthermore, the extrapolation of the line fitted to the data points passes close to the origin of the T-T plot, giving us confidence that the zero level of the 22 MHz measurements is also well established at the zenith. T-T plot analyses at other declinations (again excluding areas of absorption) indicate that the zero level is acceptably correct throughout the declination range, but that the temperature scale varies (the accuracy of the zero level is discussed in Sect. 6.1).

  

Figure 1: A T-T plot of 22 MHz and 408 MHz data for declination 48.8$^\circ$and all right ascensions except a) for the range 20$^{\rm h}$ to 22$^{\rm h}$ and b) small regions around bright point sources

The temperature scale at declinations away from the zenith was adjusted using the following procedure. At each declination the average brightness temperature ratio between 22 MHz and 408 MHz was calculated (using T-T plots) over the range 8 to 16 hours in right ascension (to exclude extended absorption regions on the Galactic plane). The temperature scale at 22 MHz was then adjusted to make this ratio equal to that at the zenith (note that Fig. 1 shows data over a wider range of right ascension).

In order to cover the central regions of the Galaxy, we have included observations as far south as declination -28$^\circ$, where the telescope was operating at a zenith angle of 76.8$^\circ$. At these large zenith angles, there is some departure of the antenna gain function from its calculated value (see Costain et al. 1969). However, reliable flux densities of point sources have been obtained as far south as declination -17.4$^\circ$ (Roger et al. 1986) and we believe that our calibration procedure remains reasonably effective to the southern limit of our map.