3 Data analysis

The analysis process primarily fell in three parts: firstly, we performed basic photometric calibration of the data, secondly we made high signal-to-noise images suitable to locate any interesting patches at the site of (or close to) the supernova, thirdly we performed photometric measurements of features visible in the SN environments.

3.1 Calibration

First, following standard procedures, the images were debiassed and flat-fielded (using IRAF arithmetic routines). Then spatial registration of the images was performed. For each galaxy, one image was taken as reference. Then the IRAF task IMEXAM was used to measure the pixel coordinates of the peak intensity of an object (or objects) visible in all images (and for all bands), using a centering algorithm. If the nucleus of the galaxy was not suitable, i.e. it was saturated, then the positions of typically three stars were located and taken as reference, otherwise the position of the galaxy nucleus itself was used. All images of each galaxy were then registered with linear shifts and linear interpolation of fluxes (IMSHIFT or IMLINTRAN in IRAF).

After the registration, the sum of the images in the same band was derived to maximize the signal to noise ratio of each color. For each galaxy one B sum, one V sum and one R sum were produced. In addition, we also simply summed all available data in all bands to get a maximum $\rm S/N$ intensity image.

The sky levels were estimated with the use of the "median'' option of IMSTAT (in IRAF) in an empty window of these images selected by eye. The photometric measurements were made by using PHOT (in noao.digiphot.apphot). PHOT was used for two different operations: to use the CCD image to simulate aperture photometry of the entire galaxy and hence, by comparison to published photometry, derive an absolute calibration of the data; and secondly to obtain aperture photometry of all patches visible around the location of the supernova to investigate the SN environment and locate candidate light echoes. The photometry used for absolute calibration was from the Longo & de Vaucouleurs (1983) and the de Vaucouleurs & Longo (1988) aperture photometry catalogs (respectively B and V, and R), not corrected for galactic extinction. The photometric calibration in each band for each individual galaxy was derived by taking the mean value of all apertures in that band. Magnitudes indicated as "uncertain'' in the catalogs were not used.

Five galaxies (NGC 3115, 3627, 4303, 4382 and 6384) were saturated in their nuclear regions in the long exposures. Thus the calibration for these was bootstrapped from a single unsaturated short exposure, which was calibrated as above.

No aperture photometry was found in the literature for UGC 2069, UGC 2259, IC 4237, IC 4798 and NGC 4674. The calibration was realized by adopting the total-extinction corrected blue magnitude provided by NED (the NASA Electronic Database) and by adopting standard B-V and V-R colors appropriate to the galaxy morphological types. Aperture magnitudes were derived for aperture values appropriate to each run pixel scale.

The calibration for NGC 1058 and NGC 1073 was done with respect to two calibration stars (one observed in each night of observations). In the case of NGC 1058 this was necessary because literature photometric apertures fell partially outside the region imaged.

The calibration for field A of NGC 6946 was obtained by scaling it to the calibration of field B, because the galaxy nucleus does not appear in field A, although the two fields imaged do overlap.

Finally for a number of galaxies, indicated with an asterisk in Tables 1a and 1b, R magnitudes were not available. In these cases we assumed a V-R appropriate to the morphological type of the galaxy, using the relations given by Buta & Williams (1995; their Table 6 and Fig. 3).

3.2 Optical structure

To locate patches of faint light in order to investigate the environments of the supernovae and seek candidate echoes, we firstly simply summed all images in all bands of a given field to make an image with essentially maximum signal to noise (S/N): the "total sum'' image. We filtered the image with a 3 $\times$ 3 median to eliminate cosmic rays. To isolate compact, discrete patches of emission from the underlying larger scale complex emission of the host galaxy, we used a digital "unsharp masking'' method. The high S/N galaxy image was differenced with a 17 $\times$ 17 median filtered version of the same image. The final output is an image well suited to identify patches of faint light against the background.

In addition to unsharp mask images of the high S/N total sum, we also derived in the same way, unsharp mask images of each individual bandpass (in B, V and R). Figures 1 through 36 show the environments of the individual supernovae as derived from these unsharp masked images: Figs. 1 through 13 are for the JKT observations, the remaining for the ESO. The figures are in order of the galaxy Right Ascension. Each supernova environment is shown in 4 different images: the total sum, followed by the three bands, B, V, and R. A 30 arcsec side box is shown centered at the SN position and displayed between ${-2}{\sigma}$ and ${+3}{\sigma}$where ${\sigma}$ is a robust estimate of the dispersion of intensity levels in the box (specifically half the percentile width corresponding to $\pm {1 {\sigma}}$ for a Gaussian distribution).

3.3 The supernova environments

Each of the unsharp masked total sum images was visually inspected and the position of all distinct objects or patches within a 10 arcsec diameter circle centered at the supernova were tabulated. In a few instances the position of patches located outside of the ring was tabulated when these patches appear blue in the images. Keeping these tabulated positions fixed, we then measured their brightness in the other three B, V and R images using a fixed aperture of 3 pixel radius for the JKT run and a fixed aperture of 2 pixels for the ESO run (corresponding to 0.99 and 1.22 arcsec respectively).

The limiting magnitude for each box is different due to different photometric conditions, different telescopes, exposure times, and different levels of contamination from the underlying host galaxy. Approximate B, V and R limiting magnitudes are respectively given in the last three columns of Tables 2a and 2b. They correspond to a ${\rm 5{\sigma}}$ detection, unless otherwise indicated. Each limiting magnitude is calculated over an aperture of 3 and 2 pixels in radius respectively for the JKT and the ESO runs using the ${\sigma}$ which is the robust estimate of the dispersion of intensity levels in the box (as discussed above). A complete discussion of Tables 2a and 2b is given in Sect. 4.

Local Galactic extinction corrections were also applied to the resulting photometry, using values taken from the A_B of the Third Reference Catalog (RC3; de Vaucouleurs et al. 1986). The corresponding V and R corrections were derived by taking the following relationships from Cardelli et al. (1989): ${{A_B} = {1.324}\times{A_V}}$ and ${ {A_R} = {0.84}\times{A_V}}$.