I've compared the 816 processed images from the July 2009 data release to SDSS to get zero points for each image. For y band I fit y_{PS} - z_{SDSS} versus i_{SDSS}-z_{SDSS} and got a slope of -0.247. Using this I made a simulated y-band magnitude for each SDSS star as y_{SDSS-sim} = z_{SDSS}-.247*(i_{SDSS}-z_{SDSS}), assuming that stars of zero i-z color should have zero z-y color. I used the resulting simulated y-band magnitudes to compute the PS zero points.

See PS1_depth for magnitude histograms and zpt_variation_july09 for visual representations of the chip-to-chip zero point variations.

The summary table is:

filter	zero point	FWHM	DRM zero point
g	24.58	~.06	24.90
r	24.80	~.06	25.15
i	24.74	~.05	25.00
z	24.26	~.05	24.63
y	23.41	~.05	23.03

I'm just pulling these zero points and FWHM by eye off the following plots, looking for the good photometric clusters of points.

The zero points are well-behaved enough that we can see the effects of atmosphere. We get airmass extinction terms:

filter	k_{PS}	k_{SDSS}
g	.136	.17
r	.182	.10
i	.122	.06
z	.079	.06
y	.085	---

Obviously there is a lot of uncertainty in the values I've given, and I haven't been very careful. However, it is nice to see that they agree vaguely.

The zero points have negligible temperature dependance:

The straight lines through the data are arbitrary lines I've added to make looking for deviations from flatness easier; temperature and zero point look awfully uncorrelated.