Absolute photometric accuracy of the CFHTLS/T0003 release
The WIDE fields W1 and W3
We cross-identified the multi-band catalogs of the T0003 release with the SDSS data in order to check the absolute photometric calibration. Only stars outside of the masked regions (marked in the CFHTLS catalog with flag =1) werer used. Note that the star/galaxy separation is restricted to i* < 21. A search radius of 1 arcsec has been adapted between the two catalaogs to minimize possible fake associations. As the MEGAPRIME filter system is slightly different to the SDSS, we transformed the SDSS data to the MEGAPRIME system using following filter transformations resulting from the SNLS group (Regnault et al. )
|g_Megacam - g_SDSS = -0.156(0.002) * (g-r)_SDSS|
|r_Megacam - r_SDSS = -0.000(0.003) * (g-r)_SDSS|
|i_Megacam - i_SDSS = -0.094(0.005) * (r-i)_SDSS|
|z_Megacam - z_SDSS = +0.050(0.005) * (i-z)_SDSS|
For 22 WIDE fields of W3 and 2 fields of W1, an overlap with the SDSS data has been found, however no overlap with SDSS was found for those WIDE fields where u* and z measurments are available. Thus our comparison is restricted to the filters g', r' and i'. The differences between the SDSS and te CFHTLS photometry for each filter is shown in the histograms below for each of the 24 fields. We find an excellent agreement between the photometry of the CFHTLS WIDE fields and the SDSS photometry which is better than 0.03 mag in the filters g',r' and i'. No additional offsets has to be applied for those fields.