1 Introduction

The interaction between the Galactic magnetic field and interstellar clouds leads to a variety of radio emitting structures, shells, filaments, and loops. Many of these features are linearly polarized. They extend up to high Galactic latitudes but they are very faint, an order of magnitude fainter than the large-scale diffuse Galactic emission. However, the study of these structures may provide important clues for the understanding of the "disk-halo connection''. The interpretation favours chimneys, fountains, Parker loops or superbubbles created by OB associations to transfer material from the disk into the halo. They will deform and compress the Galactic magnetic field. If these structures account for the rather smooth nonthermal background, they have to be sufficiently numerous and large.

The Galactic region above $b = +4\hbox{$^\circ$}$ and below $b = -4\hbox{$^\circ$}$, where the just mentioned structures play an important role, has never been studied in a systematic way. Large-scale surveys covering this area either suffer from low angular resolution or they miss the diffuse emission as it is the case for the 1.4 GHz VLA-survey (Condon et al. 1996). A combination of 1.4 GHz data from the Effelsberg telescope with high angular resolution data from the VLA is of particular importance to separate compact background sources from faint extended Galactic structures. An example has already been shown and discussed by Fürst et al. (1998).

One of the milestones in the research of Galactic magnetic fields is due to Brouw & Spoelstra (1976) and Spoelstra (1984). They have surveyed the northern sky in linear polarization with the Dwingeloo 25-m telescope at 1.411 GHz at an angular resolution of about half a degree. They also compiled maps of the linear polarization at frequencies between 408 MHz and 1.411 GHz. At 2.7 GHz Junkes et al. (1987) published a survey of the linear polarization of the Galactic plane ($\vert b \vert \leq 1\hbox{$.\!\!^\circ$}5$ and $4\hbox{$.\!\!^\circ$}9 \leq \ell \leq 76\hbox{$^\circ$}$). Recently, the southern Galactic plane ($\vert b \vert \leq 5\hbox{$^\circ$}$) was surveyed at 2.4 GHz by Duncan et al. (1995). Aside from the Dwingeloo data (see the discussion in Sect. 5), all these polarization surveys are constrained to the Galactic plane. On the low frequency side, at 327 MHz, Wieringa et al. (1993) observed small-scale variations in polarization at high latitudes with the Westerbork synthesis telescope.

The knowledge of the magnetic field in the Galactic halo even in the immediate vicinity of the Galactic plane is limited, because of the lack of observations of the linear polarization at the proper angular resolution and sensitivity. To fill this gap new observations at 1.4 GHz are being carried out to cover the entire Galactic plane at medium Galactic latitudes ($\vert b \vert \leq 20\hbox{$^\circ$}$) visible at Effelsberg. This survey will reach a sensitivity at total intensity close to the confusion limit ($\sim 15$ mK $T_\mathrm{B}$)and an even higher sensitivity ($\sim 8$ mK $T_\mathrm{B}$)at linear polarization.

We discuss the general survey parameters and observing method in Sect. 2. An absolute calibration method for the total intensity data is described in Sect. 3. The method developed to overcome the usual problem of polarimetric observations, the instrumental polarization, is demonstrated in Sect. 4. In Sect. 5 we propose a procedure to adjust the polarization data to an absolute temperature scale.