| | 7 | |
| | 8 | The past 2 weeks have had minimal development work from the IPP team, with both the PS1SC meeting in Boston and the AAS meeting the following week, as well as illness among the IPP team members. |
| | 9 | |
| | 10 | Here are the few highlights: |
| | 11 | |
| | 12 | * We had a crisis with the RAID on ipp005 for most of this period. We earlier had suspected the raid card was flaky, and intended to replace it. The machine forced our hand: the raid card started to report errors on one of the drives, and would fail to rebuild the drive. After several iteration of attempting to finish the rebuild, we decided to bite the bullet and replace the raid card first. Fortunately, after Cindy installed the new raid card, it recognized all of the disks, the raid finished the rebuild, and the system came back to us intact. |
| | 13 | |
| | 14 | * I looked into the causes for the photometry calibration to fail in some portions of the sky. The causes are fairly clear: these are regions where there are chunks that are not well-connected to other parts of the sky, and in these regions the weather was mostly non-photometric. It seems pretty clear that, without additional overlap data, we will have a hard time tying down the zero points of those fields. what is important then in the meanwhile is to have the ability to determine that certain regions are not well measured. D.Finkbeiner suggests that the eigenvectors of the chi-square matrix can be used to identify isolated regions, so I am exploring this. |
| | 15 | |
| | 16 | * We ran staticsky single-image photometry on the LAP stacks which overlapped the SDSS Stripe 82 region, and have pointed DRAVG folks at those results. |
| | 17 | |
| | 18 | Feedback from the PS1SC meeting gives us guidance on work in the upcoming weeks. For the near-term, we have the following algorithmic issues to address: |
| | 19 | |
| | 20 | * zero point offsets relative to SDSS : is there an offset in reported photometry now compared to earlier in the year? |
| | 21 | * kron magnitudes of bright galaxies are too faint : this is probably caused by a limitation on the size of the kron aperture |
| | 22 | * astrometry variations in the stack : still unclear if this caused by the stability of the calibration or the analysis |
| | 23 | * mis-clipping in stacking : this may be addressable by a modification of the initial rejection pass |
| | 24 | * sources in the wings of stars and bright galaxies : flag these by adding a bit noting that an object is in the footprint of a brighter object |
| | 25 | * background error when re-processing the log-flux images |
| | 26 | * change log-flux to inverse sinh |
