4 Construction of the reference catalogue

After reduction of the "long" images, as described in the previous section, we have separated and relabeled all the stars brighter than $m_{V\rm al} \, = \, 13$ that appeared in all the observational runs. They constituted the first set of candidates for a reference catalogue. A visual inspection was performed to eliminate those stars located near the edges of the images and those with very near apparent neighbours (which may have complicated the magnitude determination and introduced errors in position coordinates as well). The number of "long" observations was certainly enough to guarantee the astrometric accuracy of the candidates, but not their photometric stability. Thus, to detect possible variables among our reference candidates, we have employed 10 "short" exposure observations (in principle obtained for the monitoring project) in addition to the "long" ones to construct the appropriate light curves. For a few fields, the "short" observations did not contain the minimum of four Tycho stars, necessary for the reduction procedure. In these cases, the number of reference stars was increased with candidates stars from the "long" observations. On the other hand, this procedure allowed us to check the sensitivity of the results to a change in the input standard stars. Once we had constructed all the light curves for the remaining candidates we proceeded to eliminate those which had large magnitude and positions variations based on the derived parameters: position and magnitude average values, their corresponding standard deviations($\sigma_{\alpha}$, $\sigma_{\delta}$,$\sigma_{m_{V\rm al}}$), the average computed for the magnitude weighted with the error bars in each night and weighted error $\sigma_{\rm w}$. This selection eliminated stars having a magnitude standard deviation larger than 0.1; right ascention errors larger than $0.01^{\rm s}$ and declination errors larger than 0.1''. In addition, stars for which the mean value of the magnitude was significantly different from the average weighted with the nightly error bars were also discarded (see examples of the light curves in Appendix A, Figs. 4-6). In Table 2 we present the average values of the errors in position and magnitude and the number of stars in the final catalog.

  

Figure 4: Examples of light curves of the catalog stars in the windows BE, BG, BJ e LA

  

Figure 5: Examples of light curves of the catalog stars in the windows LB, LC, LD e LI

  

Figure 6: Examples of light curves of the catalog stars in the windows LR, LT, LU e LV

  

Table 2: Average position and magnitude precisions for reference stars in the low-extinction windows. The last column indicates the number of stars in the final catalogs