New atlas of open star clusters

Abstract Due to numerous new discoveries of open star clusters in the last two decades, astronomers need an easy-touse resource to get visual information on the relative position of clusters in the sky. Therefore we propose a new atlas of open star clusters. It is based on a table compiled from the largest modern cluster catalogues. The atlas shows the positions and sizes of 3291 clusters and associations, and consists of two parts. The first contains 108 maps of 12 by 12 degrees with an overlapping of 2 degrees in three strips along the Galactic equator. The second one is an online web application, which shows a square field of an arbitrary size, either in equatorial coordinates or in galactic coordinates by request. The atlas is proposed for the sampling of clusters and cluster stars for further investigation. Another use is the identification of clusters among overdensities in stellar density maps or among stellar groups in images of the sky.


Introduction
Open star clusters (OCls) are very important objects for astrophysics. It is well known that theories of stellar evolution are verified by star cluster colour-magnitude diagrams. OCLs show us the history of star formation in the Galactic disk, with young clusters being tracers of the Galactic spiral structure. Star clusters serve as laboratories for further stellar dynamics studies. The scientific interest in star clusters and for OCls in particular is growing.
A distinguishing feature of the last two decades is the rapid growth of the number of known open clusters, generally due to IR sky surveys, for example, Two Micron All Sky Survey (2MASS, Skrutskie et al. 2006), United Kingdom Infrared Digital Sky Survey (UKIDSS, Lucas et al. 2008), Visual and Infrared Survey Telescope for Astronomy Variables in the Via Lactea survey (VISTA-VVV, Minniti et al. 2010), and the Wide Field Infrared Survey Explorer (WISE, Wright et al. 2010). A large reference list with the discoveries of new clusters can be found in Carraro et al. (2016). As a result, the number of known OCls and candidates has in-creased significantly. The largest modern catalogue of star clusters (Kharchenko et al. 2016) contains 3208 objects, the catalogue of optically visible clusters and candidates (Dias et al. 2014) contains 2167 objects, and the catalogue of the Sternberg Astronomical Institute contains 168 new open clusters (Glushkova et al. 2012).
However these catalogues, accessible as tables, cannot provide all the practical needs for astronomers, because catalogues do not supply their users with a visual information about positions and sizes of star clusters. These catalogues are difficult to use without a visualisation, for example, when someone wants to compile a sample of clusters for investigation with clusters without close neighbour objects, or when someone selects stars from the cluster for investigation and tries to avoid contamination with stars from another cluster. In addition, one can face problems while attempting to identify overdensities on a map of stellar density or on an image of the sky.
Such a situation is illustrated in Figure 1. The upper panel of this figure shows the map of surface stellar density in a square field of about 2.7 degrees centred on the coordinates of star cluster NGC 7788. The map was plotted with the use of a kernel estimator (Seleznev 2016b) and the data from 2MASS (Skrutskie et al. 2006) for stars with J<16 magnitudes and with a kernel halfwidth of 5 arcminutes. It is clear that NGC 7788 is in the centre of this map, but what are other overdensities on this map? Of course, one can measure the coordinates of overdensities on the map and look for the known star clusters in the set of catalogues. However, it would be much more useful if we could get a graphical presentation of combined table with the cluster coordinates and sizes for the same field. The lower panel of Figure 1 shows the same map with overlapped positions and relative sizes of star clusters from the catalogue of Dias et al. (2014). Clusters marked by red have data on their sizes from the star counts of Danilov and Seleznev (1994). It is obvious that if the goal of someone is the study of the stellar content of NGC 7788, it is necessary to take into account the overlapping of projections of NGC 7788 and NGC 7790 on the celestial sphere.
Unfortunately, the existing sites and packages, which visualize the celestial sphere or its parts do not provide the solution to these problems. For example, the WEBDA online database¹ shows the image of the cluster field, but does not indicate the presence of close neighbour clusters and does not allow changes to the image size. A very useful Aladin package² has a greater functionality. But even if someone could open the catalogue of Kharchenko et al. (2016) in this package, they could not view the sizes of clusters in the sky image.
In addition, there is another problem. The sizes of open clusters are often underestimated in the existing literature, which was shown in Seleznev (2016a). Even if you could open some open cluster catalogues with the Aladin package with indication of cluster sizes, these sizes would be much smaller than the actual ones, as a rule.
The outlined problems can be solved by the new atlas of open star clusters and associations, similar to the old atlas of Alter and Ruprecht (1963), but with the data on newly discovered open clusters, their positions and actual sizes, and with modern technical realization. The atlas of Alter and Ruprecht (1963) contained only about 860 objects, now we have 4 times more clusters.
In this work we propose two realizations of the atlas of OCls. The first realization updates the atlas of Alter and Ruprecht (1963). It is a set of maps, with every map covering a field of 12 by 12 degrees with an overlapping of 2 degrees between adjacent maps. These maps are arranged in three strips along the Galactic equator from −16 to 16 degrees in Galactic latitude (the total number of maps is 108). The second realization is an online atlas. It shows a square field of an arbitrary size either in equatorial coordinates or in galactic coordinates.
The paper is organised as follows. In Sect. 2 we describe the procedure of combining data from the three catalogues mentioned above. Sect. 3 is devoted to the first realization of the atlas (the maps are attached to this paper 1 http://obswww.unige.ch/webda 2 http://aladin.u-strasbg.fr/ as a separate pdf file). Sect. 4 describes details of the online version of the atlas and contains brief instructions for the user. Sect. 5 summarises our results and provides some discussion and future prospects for development.

The combined table for the atlas
The atlas is based on a combined table, which contains cluster names, equatorial and galactic coordinates of cluster centres, and the cluster radii in arcminutes. To compile our atlas we used the following catalogues: All these catalogues were taken as tables and combined into one table. All coordinates were transformed into degrees and the decimal fraction of degrees. Then this combined table of 5587 rows was sorted in accordance to Right Ascension. The angular radii of clusters were then transformed into arcminutes. In catalogues 1-3 we used the r2 parameter as the angular radius.
At the next step, the rows with duplicated names were deleted. Catalogues 1-3 present the largest number of clusters with the largest number of uniform parameters. Then, we adhered to left rows from these catalogues in the combined table. After the removing of duplicated rows, the table had 3476 rows. After that, different cluster names were checked for cross matches with the help of the WEBDA database. This way we found 26 clusters, presented in the table with different cluster names. In the following, we compared objects in the table with a list of unconfirmed candidates by . This list was obtained through the difference between the complete list of MWSC objects (with 3784 rows) and the list of confirmed clusters (with 3006 rows) in . This way we removed 159 rows of the combined table, which corresponded to unconfirmed candidates. Finally, the combined table for the atlas consists of 3291 rows.
It is necessary to stress that cross-designation tables in the WEBDA database are very incomplete. There are many cases when one can see objects of different names in exactly the same position in our atlas. It would be RA (J2000), deg.  (Seleznev 2016b) and using the data from 2MASS (Skrutskie et al. 2006) for stars with J<16 magnitudes and with a kernel halfwidth of 5 arcminutes. (Lower panel) The same map with overlapped positions and relative sizes of star clusters from the catalogue of Dias et al. (2014). Clusters marked by red have data on their sizes from the star counts of Danilov and Seleznev (1994).
very important to have complete information about crossidentification of different cluster designations, because in some cases different catalogues list different coordinate values for the same cluster. For example, the cluster NGC 6664 has the declination coordinate of −8.21 degrees in the catalogue of  and −7.813 degrees in the catalogue of Dias et al. (2014). Then, it is quite possible that one cluster would be shown at different positions with different names in the atlas.
The combined table was supplemented by data on cluster radii from Danilov and Seleznev (1994) and from Seleznev (2016a). It has been done because cluster radii in the catalogues listed above are underestimated (Seleznev 2016a).

Collection of maps
galactic longitude and galactic latitude intervals for every page, which makes navigation easier.
This type of atlas is useful for fast surveying of clusters in relatively large fields. It can be efficient for sampling clusters without close neighbours for studying the cluster structure.

Web application
In order to identify the known clusters on the density maps or images of the sky, as in Figure 1, one needs to get a map just for the same field, which is covered by their density map or sky image. In order to make such an opportunity possible, we have designed an online application for the atlas. This application plots the map of the square field with an arbitrary size either in equatorial or in galactic coordinates by user request. The user can indicate the cluster name as the field centre, or an arbitrary point in some coordinate system (the equatorial or galactic one).
The combined table with the cluster data was transformed into a database with the use of MySQL, a free database service⁴. It makes data operation easier. Python language supports the use of this database. MySQL package is suitable for the creation of this online resource due to its good safety, the stable operation and its high operation speed. Transformation of coordinates is executed using the Astropy package within Python, and the plotting of the map is performed using Matplotlib package within Python.
The online resource is based on the Django frame-work⁵. The Django development environment has been chosen because it is a free framework, uses Python as a programming language, has detailed documentation with lots of examples, it can use MySQL as data storage, there are many ready-to-use templates, and it is fast and effective.
The order of the programme operation is the following.
1. The user indicates the centre of the field (by indicating the cluster name or the centre coordinates in the equatorial or galactic coordinate system). 2. The user indicates the size of the field in arcminutes, and the coordinate system they prefer. 3. If the user indicates the cluster name, then the cluster coordinates in the coordinate system indicated by the user are taken as the field centre coordinates. 4. If the user indicates the field centre coordinates in the galactic coordinate system, and wants to get a map in the equatorial coordinate system, the programme transforms coordinates from the galactic to the equatorial system. 5. If the user indicates the field centre coordinates in the equatorial coordinate system, and wants to get a map in the galactic coordinate system, the programme transforms coordinates from the equatorial to the galactic system.
6. Clusters, that fall into the field in accordance with the field size and coordinates of the field centre, are selected from the database and are plotted on the map. 7. The map in the coordinate system indicated by the user is plotted. Clusters are plotted by open circles, with radii taken from the database and plotted in accordance with the map scale. The short cluster name is displayed near the circle.
Cluster names in the combined table and in the database have been shortened in order to optimize the output on the map (to diminish the overlapping of names). Shortened names are listed in Table 2 online atlas should indicate cluster names in accordance with this table. Some shortenings are commonly used, some of them have been adopted by authors of this paper. Other cluster names are the same as in the catalogue of . The user of the online atlas should realize that every cluster name contains the underlined symbol between the proper name and the number (for example, NGC_7789; also see Table 2). This application will work at the following address http://astro.ins.urfu.ru/atlas.

Summary and discussion
This paper presents a new version of the atlas of open star clusters. The necessity for a modern atlas has risen due to a sufficient increase in the number of known clusters (by a factor of several) and a lack of possibility to get visual information on the actual positions and sizes of open star clusters with the existing tools used for visualizing the celestial sphere. The new atlas consists of two implementations. The first one is the collection of maps, the second Al_1 Ha_9 one is the online application. Each implementation serves different tasks. The collection of maps is useful for fast surveying of clusters in relatively large fields. It can be efficient, for example, for sampling clusters without close neighbours to study the cluster structure. The online application could be very useful for identification of the known clusters on the density maps or images of the sky. The atlas contains available information on cluster sizes, determined by detailed star counts Seleznev 1994, Seleznev 2016a). These cluster sizes are larger, as a rule, than cluster sizes determined with the automated reviews (Seleznev 2016a) and reflect the fact of the existence of vast cluster coronae (Danilov et al. 2014).
Further work needs to be done to continue the development of the atlas. Publications with newly found clusters are going on, for example, Loktin and Popova (2017) found 48 new possible candidates. Some objects are shown not to be real clusters after detailed investigations (for example, Pismis 14 in Carraro et al. (2017), and ESO131SC09, NGC 5284 and vdBergh-Hagen 164 in Carraro and Seleznev (2012)). It is necessary to find all double mentionings of the same cluster with different names. Maybe, it is worthy to indicate unconfirmed candidates with a different colour (it is possible, that some of them will restore their status in the future). The future Gaia catalogue will cause the number of known clusters to increase, and the revision of the status of many objects.
The authors realize that the terms 'cluster', 'cluster candidate' and 'unconfirmed candidate' are rather relative to some extent. The majority of objects in our atlas are from the catalogue of N.V. Kharchenko and her co-authors (Kharchenko et al. 2016). They refer to all these objects as 'clusters', and it is confirmed by their analysis, which in-cludes an analysis of proper motion data (see Kharchenko et al. 2016 and references therein). We are following their terminology. However, the data on many of these objects is very few, and the precision of proper motion data is often not high. Consequently, it is possible that the status of many objects in this catalogue could be revised in the future, especially taking into account future high-precision proper motions of Gaia and possible detailed investigations of open clusters by photometry and spectroscopy.