...
So this seems to suggest that the error in color is not a function of stellar color? I will try this again with unstacked ipp data soon.
...
Aug 22
I've tried looking for the effects of water vapor (greater deviation from mean i-z color for redder stars) in the unstacked data, but found nothing. Here's the graph (only stars which have no close neighbors and accurately calibrated colors from stacked data were plotted):
To ensure that this is no fluke, I looked at the changes in stellar color of a particular star (chosen because we have many observations of it). I found that on a single night, the g-r color (which is not supposed to be affected by water vapor) varied as much as (if not more) than the i-z color (which water vapor is supposed to affect):
I also checked for the possible effects of airmass. That does seem to affect anything either:
So unfortunately, I haven't found any evidence that water vapor affects stellar color (or I guess I should say, I have not found any evidence that water vapor is the limiting factor on the precision of PS1 photometry) so far 
I made a presentation of my results for PRISE, and the powerpoint is available here: https://www.dropbox.com/s/7larxgd1jiun4k5/gutter.pptx
...
Data access on Odyssey:
- Run JAuth.jar to get login key
- ssh -Y into to odyssey.fas.harvard.edu, or herophysics.fas.harvard.edu, using the electronic key.
- run tcsh
- source .myrcstubbs
- data are at /n/panlfs/data/MIRROR/ps1-md/gpc1/
- nightly science uses individually warped images, nightly stacks run on stacked frames
- image types: wrp is warped.
- see available modules with "module avail"
- load a module with "module load hpc/ds9-5.6"
- photometry is in .cmf files, as FITS tables.
- in python:
- import pyfits as p
- p.open('filename')
- print a[0].header
- or, imhead on command line
- a[1].data.AP_MAG for aperture magnitudes
- PSF_RA and PSF_DEC are in the skycell files.
- make a scratch directory for data in /n/panlfs
...