WFI Photometry Release (#26)

13 December 2004

Technical Summary


CURRENT RELEASE
Telescope .................. ESO/MPG 2.2m
Instrument ................. WFI
Origin ..................... ESO/EIS
Passbands .................. U (2); B (2); V; R; I (2)
Number of Filters .......... 8
Version .................... 0.9
Release Date ............... December 2004
Release prepared by ........ EIS team, A. Mignano

PRODUCTS
Product Type ............... Photometric Solutions
Number of solutions ........ 414

PREVIOUS RELEASE
None

Abstract

In this survey release, a first version of photometric solutions computed using the EIS Data Reduction (Survey) System for the wide-field imager (WFI), mounted on the ESO/MPG 2.2m telescope at La Silla, are given. These solutions were obtained as a by-product of the several surveys (Pilot, Pre-FLAMES, DPS, XMM and GALEX) conducted by the EIS team since the commissioning of WFI. In this release, solutions covering the period May 12, 1999 to October 17, 2003 are presented, comprising 148 nights. This provides a 4-year baseline and an excellent benchmark for the EIS Data Reduction System. In principle, it can be extended to January 1999 and to November 17, 2004, the latter date corresponding to the most recent EIS observations using WFI, available at the time of writing. Solutions are presented for the 8 filters used at different times by the EIS surveys. Even though spanning a large time interval the present release is neither complete nor definitive, but is intended to serve as a reference for WFI users, as it covers the lifetime of the instrument, and an illustration of some of the capabilities of the EIS software system. Note, however, that as emphasized in previous releases, the calibration of a wide-field imager is complex and requires much more investigation than the time available to the team for the present release.

The present release of photometric solutions extends the suite of EIS deliverables which have so far included: 1) software; 2) reduced images on a nightly basis; 3) stacks 4) single-passband catalogs; and soon 5) color catalogs, derived from the association of single-passband catalogs.

For more information about the terminology and conventions used in this document refer to the WEB README pages.

Contents of this Release

This release provides a list of the photometric solutions for 8 WFI filters, obtained from 1- to 3-parameter fits to measurements of Landolt stars. Only solutions for which the rms of the fit is less than 0.1 mag have been included. These solutions can either be read from the table accessible from the WEB pages, via the Survey Release pages, or from the associated XML which can be requested from the release page (see Photometry Release Information section). The original intention was to provide for each instrument, filter and survey (EIS) night an XML (and ASCII) file comprising the best solution, as determined by the EIS system, together with all the measurements available for standard stars in the night and filter considered. Unfortunately, due to time constraints it was not possible to complete the distribution infrastructure as desired. So instead, only the solutions are being provided.

Retrieving EIS Products

Description

The section "contents" of the present release lists:

  1. Table entry number;

  2. Passband (Band);

  3. ESO's filter number;

  4. MJD of the noon preceding the night of observation, which identifies a night for the system;

  5. Civil Date (YYYY-MM-DD) at the start of the night;

  6. Number of independent parameters in the solution (# par);

  7. Type of solution which can be one of the following: 1) a 3-parameter fit represented by Z-K-C; 2) 2-parameter fit solutions represented by Z-K or Z-K-CM, where CM is a median color term computed from previously available solutions chosen by the user; and 3) the one-parameter fits represented by Z-KT, Z-KM and Z-KM-CM, where KM and CM are again median values computed from previously available solutions and KT is a theoretical value computed from a mean extinction curve for a given site (CTIO is currently used).

  8. Zeropoint (Zp);

  9. Zeropoint error;

  10. Extinction (K);

  11. Extinction error;

  12. Color term (CT);

  13. Color term error;

  14. Color index used;

  15. Value of the scatter (rms) of the final fit;

  16. Number of versions, namely the number of different solutions obtained for the same night and passband. The different versions may represent solutions obtained using different configuration parameters or different input set of measurements. A new version is created whenever a new solution is found for the same night, instrument and filter. Whether a solution is new is determined by comparing the checksum computed for the ``object'' describing the photometric solution just created with those already stored in a table in the database; the large number of versions for some nights (up to 13) is due to several tests that were carried out and do not reflect any particular problem with that night. Note that the EIS Data Reduction System also provides the infrastructure to examine the HISTORY of a photometric solution and compare its different versions.

  17. Hyperlink to the product log (see below)

  18. Interactive check-box which enables the individual product to be selected/de-selected before finally submitting the selection to the ESO Science Archive Facility using "Request Marked Products" at the foot of the page. As explained above the product was intended to be an XML containing not only the solution for that night but also all the measurements of standard stars available for that night and filter. Even though most of the infrastructure is in place the implementation had to be halted. Therefore, in the present release the boxes are not activated and the XML containing all solutions should be requested from the link available in the Photometry Release Information section. The current XML file is a rudimentary implementation and should not be considered final.

Note: All data products are subject to revision and update once released.

Product Logs

While the infrastructure is in place for a more extensive log, which will include access to the measurements of the standards, currently it only shows the residual of the fits as a function of UT, airmass (X), color and magnitude as tabulated in the original reference (Landolt 1992). It is also envisioned to show the residual as a function of position of the standard star in the camera either as a function of right ascension and declination as well as a function of the chip in which the standard is located. This essential in order to investigate in more detail the chip-to-chip variations.

Plots

In the upper right corner of the data release WEB page for catalogs, one finds a button (properties) which link to plots showing the time dependence of the zeropoints for each filter. This is shown as a function of the time interval relative to the indicated value of MJD (left panel) and as a function of the night index (right panel) . Note that in the second plot all nights can be seen but the real time information provided in the first plot is lost. The different symbols represent: 3-parameter fits (solid circles); 2-parameter fits without (solid square) and with a color term (open square), with the latter being a median value of previous values available in the database; and a 1-parameter fit with the extinction coefficient and with (skeleton triangle) and without (solid triangles) the color term, both assumed to be a median of a trend analysis, or computed from a mean extinction curve for the site (open triangles).

Note that while the V and R filter have always been the same, the filters in all other bands have changed in time with #877 replacing #841 in the U-band, #878 replacing #842 in B-band, and #879 replacing #845 in the I-band. These changes further complicate the administration of the surveys and in the future should be avoided.

For each filter the distribution of zero-points, extinction coefficients and color term for the available solutions are also shown. The vertical lines indicate 3-sigma intervals relative to a median value of all solutions obtained with at least one free parameter. Note that in the histogram of the color term zero values are also shown but do not affect the calculation of the median.

Comments Specific to this Release

Photometric solutions are available for the following WFI filters:


Passband ESO Filter Name ESO Identity Number
U U350/60 #877
U U/38 #841
B BB#B/123_ESO878 #878
B B/99 #842
V V/89 #843
R Rc/162 #844
I BB#I_EIS #879
I Ic/Iwp #845

Comparison to Previous Release(s)

This is the first release of photometric solutions.

Data Reduction

The photometric pipeline being used to automatically extract catalogs, measure fluxes at different apertures, identify objects corresponding to known standard stars, make linear fits to these measurements and derive zeropoints, extinction and color terms (if applicable) was originally developed to deal with single chip imagers - only recently systematic tests have been conducted for multi-chip cameras. A number of improvements to the code still remain and more extensive tests with the code and with the many configurable parameters involved are required. Therefore, the solutions presented in this release should not be considered final.

In accordance with the established EIS standard procedures, photometric calibration of the science fields is performed on the Vega magnitude system. Zeropoints are obtained from linear fits to the measured instrumental magnitudes of standard stars in Landolt (1992) fields taken at the same night as the science frames. These fits were obtained using from one (taking some specified values for the extinction and color term) to three free parameters depending on the available airmass and color coverage. Most observations were carried out in visitor mode, by a large number of people and over a long stretch of time, making it difficult to ensure a homogeneous calibration plan.

In practice, the photometric pipeline computes photometric parameters for all possible types of fits (one to three-parameters) and assigns the solution with the smallest scatter, to be the ``best solution'' for the night. A night is considered photometric if this scatter is less than a pre-defined value, which at the present time is taken to be below 0.1 mag. If none of the solutions satisfy this criteria and/or the solution found yields unrealistic results (e.g. negative extinction) then the night is considered non-photometric and a default value for the zeropoint is adopted and its error set to -1. These solutions are not shown here.

For homogeneity, the default value normally adopted is the median of the zeropoints reported in the trend analysis kept by either the telescope team (depending on the instrument) or the internal EIS database. In the case of WFI, only one solution is currently reported in the WFI WEB pages.

For the nights considered, one finds a total of 414 solutions, of which 124 are 3-parameter fits, 37 use 2-parameters and the remaining 253 are fits only for the zeropoint. Even though most of the WFI observations were for deep observations, involving stacks of several images, and therefore not requiring the calibration of all nights, the large number of 1-parameter fits indicates a less than ideal calibration plan. In particular, inspection of the nights indicates that the major problem has been poor airmass coverage. The breakdown of solutions by filter and solution type is as follows:


Passband 1-par 2-par 3-par total
U#877 19 7 13 39
U#841 20 0 1 21
B#878 25 0 11 36
B#842 37 0 22 59
V#843 70 8 31 109
R#844 47 14 13 74
I#879 0 4 10 14
I#845 35 4 23 62

Using only 3-paramter fit solutions, one obtains the following median values:


Passband Zp k color term
U#877 21.91 0.46 0.05
U#841 21.78 0.48 0.01
B#878 24.71 0.18 0.19
B#842 24.65 0.23 0.24
V#843 24.19 0.15 -0.12
R#844 24.47 0.12 0.01
I#879 23.34 0.05 0.03
I#845 23.12 0.08 0.18

General Comments

As mentioned in earlier releases, as well as by other authors, the proper calibration of a multi-chip camera is complex and requires a thorough investigation. In this context the following comments should be made:

  1. From the trend analysis, shown under properties, occasionally one finds solutions that deviate significantly from the average behavior. During image calibration, these outlying solutions are identified, according to user-defined criteria, discarded and replaced by a default solution, with the night being then considered non-photometric.

  2. The scatter of the zeropoint of the U-filters, especially #877, is large. Inspecting of the distribution of zeropoints for this filter one finds that it is bi-modal (see plots under properties). Note that this feature is also visible when only 2- and 3-parameters fit solutions are considered, discarding the possibility that this is an artifact of inadequate choices of default parameters. Examining the time variation of the zeropoint suggests sudden rises followed by a steep decline. The variation is large, being of at least 0.6 mag peak-to-peak. The same behavior seems to be present in other filters, but the amplitude of the effect is considerably smaller. One possible interpretation is that it reflects an efficiency loss in the reflectivity of the system. These discontinuities would then be associated with the re-aluminization of the mirror. Unfortunately, the sampling of survey nights is not sufficiently dense to fully characterize the time dependence.

  3. Preliminary attempts to use the EIS DPS catalogs to compute photometric redshifts have indicated some unexpected behavior as well as departures from the expected loci of stars in color-color diagrams predicted by the Girardi et al. (2004) model. Both of these effects are associated to the U-band and requires further investigation. Possible alternatives are to use different color indices (e.g. U-V) or the addition of a quadratic color term. Since the matched objects are stored in the database such investigation can be carried out, requiring only the retrieval of the instumental magnitudes of objects with matching standards.

  4. A more detailed investigation should be undertaken to determine the impact on the solutions due to: 1) residual chip-to-chip gain differences even though these have been corrected for by the pipeline; 2) the illumination correction.

  5. More extensive tests of the automatic extract/match/measure and fit procedures adopted and the associated configuration parameters (including SExtractor) should be carried out .

  6. A more thorough investigation of the pros (denser sampling, larger number of standards per chip) and cons (different system, possible variability) of using secondary standards, already available in the internal database but not yet used.

Data Quality Assessment

The quality of the photometric calibration of the reduced and stacked images has been assessed in the EIS release numbers 19, 20, 22and 23 and will not be repeated here.

Comparison with other authors

A crude evaluation of the absolute calibration of the EIS products has been made in previous releases by comparing the number counts of galaxies. In addition, the photometric parameters obtained by the system have been compared to:

Next Release(s)

Originally, the following releases were scheduled to take place:

Conclusions

A first version of the photometric solutions covering nearly 150 nights over a 4 year period for 8 WFI filters were presented. These results will probably evolve in time since the proper photometric calibration of a multi-chip camera is perhaps one of the most challenging and time-consuming tasks for the preparation of science grade products from wide-field cameras. Having in mind the new generation of survey cameras/telescopes (WFCAM, VST and VISTA) and its large data flow, a considerable effort has been made to provide the required tools to allow not only for the un-supervised photometric calibration but also mechanisms to monitor and iteratively explore the nature of the derived solutions, as well as ways of improving the procedure used to automatically derive them. The results presented here provide a quick glimpse of the infrastructure being built for that purpose. It is important to emphasize that different parts of the code have been developed separately and by different people. It is during a dry-run of a release that a number of issues are resolved and incorporated into the system. Unfortunately, in this release these iterations were not possible due to time constraints beyond the control of the EIS team, and parts of the envisioned infrastructure were left incomplete.

References

Girardi, L., et al. , 2004, A&A, submitted

Landolt, A. U., 1992 Astronomical Journal 104, 340

Radovich et. al., 2004, A&A, 417, 51


Remco Slijkhuis 2004-12-13