fig1_3.jpg
fig4_7.jpg
fig2_5.jpg
fig4_6.jpg
fig4_8.jpg
fig4_9.jpg
R. Ungruhe, W.C. Seitter and H.W. Duerbeck
We present a large-scale galaxy catalogue for the red spectral region which covers an area of 5\,000 square degrees. It contains positions, red magnitudes, radii, ellipticities and position angles of about 5.5 million galaxies. Together with the APM catalogue (4,300 square degrees) in the blue spectral region, this catalogue forms at present the largest coherent data base for cosmological investigations in the southern hemisphere.
217 ESO Southern Sky Atlas R Schmidt plates with galactic latitudes -45 degrees were digitized with the two PDS microdensitometers of the Astronomisches Institut Münster, with a step width of 15 microns, corresponding to 1.01 arcseconds per pixel. All data were stored on different storage media and are available for further investigations.
Suitable search parameters must be chosen in such a way that all objects are found on the plates, and that the percentage of artificial objects remains as low as possible. Based on two reference areas on different plates, a search threshold of 140 PDS density units and a minimum number of four pixels per object were chosen. The detected objects were stored, according to size, in frames of different size length. Each object was investigated in its frame, and 18 object parameters were determined.
The classification of objects into stars, galaxies and perturbed objects was done with an automatic procedure which makes use of combinations of computed object parameters. In the first step, the perturbed objects are removed from the catalogue. Double objects and noise objects can be excluded on the basis of symmetry properties, while for satellite trails, a new classification criterium based on apparent magnitude, effective radius and apparent ellipticity, was developed. For the remaining objects, a star/galaxy separation was carried out. For bright objects, the relation between apparent magnitude and effective radius serves as the discriminating property, for fainter objects, the relation between effective radius and central intensity was used. In addition, a few regions of enhanced object density like dwarf galaxies and star clusters were removed from the catalogue.
Because error estimates of the automatic classificationprocedure are very uncertain, an extensive visual check of the automatic classification was carried out. A large number of objects previously classified automatically - 1.3 million galaxies, 815,000 stars and 647,000 perturbed objects - was re-classified by eye. We found that galaxies suffer most from misclassification. Down to magnitude 13, the error is, independent of galactic latitude, at least 60%. Between13th and 17th magnitude, the percentage of misclassified galaxies for b < -45 degrees drops continuously to between 15% and 30%, and is clearly dependent on galactic latitude. The classification of galaxies at low galactic latitudes is most strongly affected; in these regions only half of the galaxies are correctly classified. Errors found in this work thus lie by a factor 2-3 above values quoted in the literature.Stars show classification errors of at most 10%, whose level increases towards fainter magnitudes. The classification accuracy is less dependent on galactic latitude than in the case of galaxies. As concerns artifacts, noticeable classification errors occur only for objects brighter than magnitude 15, which is mainly caused by saturation effects of the photographic emulsion. At magnitudes fainter than 15th,the error is below 5%. No dependence from galactic latitude is seen.
These investigations show that the automatic classification yields satisfactory results only in certain magnitude intervals, which depend on galactic latitude.
The object magnitudes are influenced by the desensitization of the emulsion during exposure and by the vignetting of the telescope. Objects at the plate margins appear systematically too faint. The magnitudes were corrected by means of measured number densities of galaxies and stars, which were determined in 63 fields around the galactic southpole. The difference of the magnitude zero-point between the center and the margin of a plate amounts to approximately 0.05 mag after correction of the margin desensitization.
Because of their high central intensity, stars reach the saturation limit of the emulsion already at magnitude 17. Thus bright stars appear systematically too faint. The saturation effect can be corrected by means of a point-spread function, which is calculated from the unsaturated parts of the stellar intensity profiles. The magnitude corrections for the saturation are carried out for each plate separately.
In order to establish a unique magnitude zero-point for the 217 single plates, a mutual adjustment of neighbouring fields by means of their overlap regions was done. The procedure was carried out separately for stars and galaxies. In total, 1,005 overlap regions for galaxies and 1,103 regions for stars were available. The zero-point error after adjustment amounts to 0.06 mag for galaxies and 0.07 mag for stars. The external calibration of the photographic rF magnitudes was carried out by means of CCD sequences obtained with three telescopes in Chile and South Africa. In total, photometric V, R data for 1,037 galaxies and 1,058 stars in 92 fields are available.
The transformation between photographic and CCD magnitudes requires a relation between F and VR.It was carried out separately for stars and galaxies, because they show different colour transformations. After the transformation of the photographic rF magnitudes to the standard Johnson R system, the errors of the local magnitude zero-point amount to 0.11 mag. for galaxies and 0.15 mag for stars. Because of the large areal extent of the catalogue, the galaxy magnitudes must be corrected for interstellar extinction. Magnitude corrections are based on hydrogen column densitiesof interstellar dust. Extinction corrections amount to up to0.1 mag for 55% of galaxies, and 0.2 mag for another 35%. For the remaining 10%, the corrections are above 0.2 mag.The iteration procedure for the indirect adjustment of single platesmay cause a magnitude gradient in north-south or east-west direction. Investigations of the magnitude differences between photographic and CCD magnitudes versus right ascension and declination show no significant gradients.In order to generate a complete catalogue of galaxies and stars, all double or multiple objects that occur in overlap regions have to be excluded. After the merging of all single plates (including half of the overlap regions), both catalogues contain 5.5 million galaxies and 20.2 million stars.
The completeness of the catalogues was investigated from the comparison of counts of stars and galaxies with simulations. The limit of completenessis at magnitude18.3 for galaxies, and at 18.8 for stars. In the case of galaxies, a clear deficit is seen for galaxies down to magitude 16 in comparison with the simulation. Neither by taking into account galaxy evolution, nor by changes in the cosmological parameters, an adjustment of the simulation to the catalogue counts was possible. These results and those of others support the assumption that we are dealing with a real galaxy deficit. The determined slope of 0.66 of the galaxy counts is, within the limits of accuracy, in agreement with the measured values of other authors. No comparable star counts are available.
The N-point angular correlation function were determined from various sub-catalogues consisting of 9, 25, 63, 121 and 152 fields as well asfor limiting magnitudes from magnitudes 15.0 to 19.0. The computation of chance distributions was carried out for galaxy counts in cells with side borders from 25''to 28.4''. Averaged correlation functions and their coefficients were determined by means of factorial moments. The delete-d jackknife procedure was applied for the error estimate, with 200 replications per subcatalogue.
The 2-point angular correlation function shows a linear trend in logarithmic plots for all sub-catalogues on scales from 0.02 to 2degrees.Within this range, it can be parametrized by a power law omega2 = A theta exp(1-gamma). Depending on sub-catalogue, the gamma values scatter between 1.63 and 1.73. They show a good agreement with the EDSGC, APM and MRSP catalogues. The parametrization of the amplitudeof the 2-point angular correlation versus apparent magnitude yields beta values between 0.267 and 0.322, which are in accordance with beta values from model calculations. The curve form of the 2-point angularcorrelation function shows a significantly flatter decline on scales exceeding 2 degrees which cannot be reproduced by the standard CDM-model.
The correlation functions of higher order intersect at a point theta_S, whose position depends on the limiting magnitude. For scales theta <theta_S , the high orders dominate, while for theta > theta_S,they decay very quickly. Since the higher orders only occur in galaxyclusters, the intersection can be taken as a measure for the cluster size.Correlations for the third to fifth order that still exist on large scalesindicate that, compared to the size of galaxy clusters, larger structuresexist, the superclusters.
The correlation function coefficients with N ge 3 show characteristicplateaus for all limiting magnitudes, whose length depends on the order considered. Simulations have shown that the plateaus ofsmall scales point towards a strong non-linear cluster formation. The meaning of plateaus on large scales is not yet known. Comparisons with the APM- and MRSP-catalogues show a good agreement, both in the shape as well asin the amplitudes of the coefficients. Only the EDSGC shows significantly higher amplitudes on small scales, which are likely caused by wronglyclassified and/or doubly counted galaxies.
This paper is an edited and translated version of a Ph.D. thesis, submitted by R. Ungruhe to Münster University in 1998. It is released to make the results from this work available to a larger scientific community. Since the data of the galaxy catalogue will also be made available through the NED/IPAC database, its users can make themselves familiar with the methods of analysis and of the construction of the catalogue.
Another paper, Angular and three-dimensional correlation functions determined from the Muenster Red Sky Survey} by P. Boschán (the second referee of the Ph.D. thesis), published in Monthly Notices of the Royal Astronomical Society, Vol. 334, pp. 297-309, is to a very large part based on the contents of the thesis. It should be noted that it was written without the knowledge and without the permission of the author of the Ph.D. thesis.
Manuscript in gzipped postscript: jad9_1.ps
Principal Figures:
| fig1_3.eps fig1_3.jpg |
|
fig4_7.eps fig4_7.jpg |
|
fig2_5.eps fig2_5.jpg |
|
| fig4_6.eps fig4_6.jpg |
|
fig4_8.eps fig4_8.jpg |
|
fig4_9.eps fig4_9.jpg |
|
ASCII catalogue is available in files (gzipped)
Abell cluster photometry (ASCII files) in directories