Astron. Astrophys. 321, 229-235 (1997)

4. A search for variability

We searched for a transient radio counterpart by looking for variability within the COMPTEL+IPN error region on three distinct timescales: long-term variations (about two years), variations within the period of the 92 cm observations (about three months) and short-term variations within a single observation (hours). Long-term variations were searched for by comparing the images of the 92 cm observations with those obtained for the Westerbork Northern Sky Survey (WENSS) ². The WENSS observations of the area around the position of GRB 940301 were taken during a period of 2 months more than 2 years before the GRB 940301 event.

We combined observations of GRB 940301 of April 2 and 11, and June 5, 11 an 25, 1994, and made a map of the 325 MHz data of these observations. These particular observations were chosen because of the quality of the data, the good coverage and because the WENSS survey was also done at 325 MHz. We obtained a difference map by subtracting the WENSS image (mapped onto an identical grid and corrected for the different primary beams) from the GRB 940301 image. In this way we found two objects that showed a very large increase in their flux density. Within an area of about diameter around the COMPTEL position these two sources were, in fact, the only ones showing a variation larger than 5% or a variation in excess of 5 times the noise. The brightest sources in the field typically agreed to within 1-2%.

We also searched for variability within the three month period of the 92 cm observations. Unfortunately, our observations are of various durations and were obtained with different configurations of the Westerbork array. Hence, the synthesized beams of the different images differ. The most sensitive way to look for variations in the data is via the construction of difference maps from residual maps, i.e. maps deconvolved with a model containing all real sources in the field. We proceeded as follows. We obtained equal synthesized beams in the residual maps of two observations by retaining only those points in common to both datasets. Subsequently, we subtracted the two maps from each other and obtained the difference map. We considered only pairs of observations that have at least two hours of overlap in hour angles. In these difference maps we did not detect any source variation above 10 mJy ), within the 92 cm period of observations (1 to 4 months after the GRB). The two sources found by comparison with the WENSS data, were not conspicuous in these difference maps.

Finally, we searched for variability on timescales of hours by looking for the characteristic artifacts that such rapid variations can cause in the case of a 1-dimensional earth-rotation synthesis instrument (see e.g. Van den Oord and De Bruyn 1994). In the June 26 1994 data we discovered a rapidly variable radio source which turned out to be associated with the nearby binary pulsar PSR B0655+64 which is located only from the pointing centre. These results are described in more detail in Galama et al. (1997).

© European Southern Observatory (ESO) 1997

Online publication: June 30, 1998
helpdesk.link@springer.de