SCIENTIFIC GOALS
What it's all about, in just one sentence...
The main goal of the VVV Templates Project is to derive well-defi ned light curve templates in the near-IR
for the automated classifi cation of VVV light curves.
for the automated classifi cation of VVV light curves.
VVV: the Vista Variables in the Vía Láctea (VVV) ESO Public Survey
[VVV official webpage]
The scientifi c goals of the VVV Survey are described in great detail in Minniti et al. (2010).
Here let us just say that...
The Vista Variables in the Va Lactea (VVV) ESO Public Survey is an infrared variability survey of the Milky Way bulge and an adjacent section of the mid-plane where star formation activity is high (Minniti et al. 2010; Saito et al. 2010). The survey is carried out on ESO's Visible and Infrared Survey Telescope for Astronomy (VISTA) 4m telescope, which is equipped with a wide- eld near-infrared (IR) camera (VIRCAM; Dalton et al. 2006) that is a mosaic of 16 2048x2048 detectors, each sensitive over the spectral region 0.8-2.5 micron, and with an average pixel scale of 0.34"/pixel. The total (effective) field of view (FoV) of the camera is 1.1x1.5 square degrees.
In addition to providing images covering a wide range in eff ective wavelength (ZY JHK lters), the VVV Survey is able to probe much deeper into high extinction regions than its predecessors. In particular, near-IR color-magnitude diagrams (CMDs) of some VVV fields extend about 4 magnitudes deeper than their 2MASS counterparts (Saito et al. 2010), and also about 1 magnitude deeper than the UKIDSS Galactic Plane Survey (GPS) in the regions of overlap (Minniti et al. 2010). Most importantly, the VVV Survey will also provide a window into time-variable phenomena, by repeatedly covering the Galactic bulge and regions of the inner plane over a timespan of about 5 years, with a total of 1929 hours of observing time. Of order 10^9 point sources within a total sky area of 520 deg^2 will thus be monitored. According to the December 2010 version of the Harris (1996) catalog, the surveyed region includes 36 globular clusters and about 314 known open clusters.
The final products will be a deep IR atlas in 5 passbands and a catalogue containing an estimated 10^6 variable stars.
Here let us just say that...
The Vista Variables in the Va Lactea (VVV) ESO Public Survey is an infrared variability survey of the Milky Way bulge and an adjacent section of the mid-plane where star formation activity is high (Minniti et al. 2010; Saito et al. 2010). The survey is carried out on ESO's Visible and Infrared Survey Telescope for Astronomy (VISTA) 4m telescope, which is equipped with a wide- eld near-infrared (IR) camera (VIRCAM; Dalton et al. 2006) that is a mosaic of 16 2048x2048 detectors, each sensitive over the spectral region 0.8-2.5 micron, and with an average pixel scale of 0.34"/pixel. The total (effective) field of view (FoV) of the camera is 1.1x1.5 square degrees.
In addition to providing images covering a wide range in eff ective wavelength (ZY JHK lters), the VVV Survey is able to probe much deeper into high extinction regions than its predecessors. In particular, near-IR color-magnitude diagrams (CMDs) of some VVV fields extend about 4 magnitudes deeper than their 2MASS counterparts (Saito et al. 2010), and also about 1 magnitude deeper than the UKIDSS Galactic Plane Survey (GPS) in the regions of overlap (Minniti et al. 2010). Most importantly, the VVV Survey will also provide a window into time-variable phenomena, by repeatedly covering the Galactic bulge and regions of the inner plane over a timespan of about 5 years, with a total of 1929 hours of observing time. Of order 10^9 point sources within a total sky area of 520 deg^2 will thus be monitored. According to the December 2010 version of the Harris (1996) catalog, the surveyed region includes 36 globular clusters and about 314 known open clusters.
The final products will be a deep IR atlas in 5 passbands and a catalogue containing an estimated 10^6 variable stars.
The VVV Survey area, with individual "tiles" (i.e., individual observation fields, obtained from a combination of "pawprints"; see, e.g., Minniti et al. 2010) numbered.
This is a plot of Galactic latitude b version Galactic longitude l, overplotted on a diff erential extinction contour
map.
Loading player...
The VVV Survey Templates Project
The Need for Near-IR Light Curve Templates
Unlike previous variability surveys, the VVV Survey is carried out in the near-IR. Despite several fundamental advantages, mostly due to the ability to probe deeper into the heavily reddened regions of the Galactic bulge and plane, the use of this spectral region also presents us with important challenges. In particular, the high-quality templates that are needed for training the automated variable star classi cation algorithms are not available. This is very diff erent from the situation of variability surveys carried out in the visible, for which vast quantities of high-quality template light curves are available (see, e.g. Debosscher et al. 2007; Blomme et al. 2011; Dubath et al. 2011).
To properly illustrate the importance of this point, one should consider that not only is the currently available number of near-IR templates inadequate for our purposes (since they do not properly sample the large variety of light curve shapes encountered for many of the most important variability classes). Also, many such classes have not yet been observed in a suciently extensive way in the near-IR, so that good light curves are entirely lacking for these classes. Therefore, in order to properly classify the 10^6 light curves produced by the VVV Survey, we need to build our own template light curves.
Unlike previous variability surveys, the VVV Survey is carried out in the near-IR. Despite several fundamental advantages, mostly due to the ability to probe deeper into the heavily reddened regions of the Galactic bulge and plane, the use of this spectral region also presents us with important challenges. In particular, the high-quality templates that are needed for training the automated variable star classi cation algorithms are not available. This is very diff erent from the situation of variability surveys carried out in the visible, for which vast quantities of high-quality template light curves are available (see, e.g. Debosscher et al. 2007; Blomme et al. 2011; Dubath et al. 2011).
To properly illustrate the importance of this point, one should consider that not only is the currently available number of near-IR templates inadequate for our purposes (since they do not properly sample the large variety of light curve shapes encountered for many of the most important variability classes). Also, many such classes have not yet been observed in a suciently extensive way in the near-IR, so that good light curves are entirely lacking for these classes. Therefore, in order to properly classify the 10^6 light curves produced by the VVV Survey, we need to build our own template light curves.
Current Status and Future Perspectives
In the more general framework of the VVV Survey, the VVV Templates Project has turned out to be a large observational eff ort in its own right, aimed at creating the fi rst database on stellar variability in the near-IR, i.e. producing a large database of well-defi ned, high-quality, near-IR light curves for variable stars belonging to di fferent variability classes. By monitoring hundreds of (optically well-studied) variable stars in the JHK bands, its primary goal is thus to provide a statistically signi fcant training set for the automated classi fication of VVV light curves.
The astronomical community as a whole has been very supportive, with the end result that we have secured time for this project using several IR facilities across the globe. The availability of telescope time is key to ensure that we produce a signi ficant number of template-grade light curves in time for classi fication of the VVV light curves, i.e. within the next 3 years or so, before the bulk of the VVV variability campaign is carried out. In order to ensure a homogeneous observational strategy and to optimize the use of the awarded time, each telescope/instrument combination is used to build template light curves for at most a very few specifi c variability classes (see the Observatories pages).
Importantly, the scientifi c return of the VVV template light curves will not be restricted to the automated classi fication of the 10^6 light curves produced by the VVV Survey. Rather, these light curves will provide us with a unique opportunity to expand our knowledge of the stellar variability phenomenon per se, by tackling the comparatively ill-explored near-IR regime. Consider, as an example, our upcoming monitoring, using the VISTA telescope, of the variable star content of Omega Centauri. Such a project will lead to high-quality light curves for more than 250 variables, greatly exceeding, both in quantity and in completeness of each individual light curve, what was achieved in previous near-IR studies of the cluster (Del Principe et al. 2006). In particular, since Omega Cen hosts the largest known population of SX Phoenicis stars, some 75 members, we will be in a position to directly derive, for the fi rst time, a period-luminosity relation for SX Phe stars in the near-IR (Freyhammer & Sterken 2003; Weldrake et al. 2007). In like vein, we will also improve the calibration of the RR Lyrae period-luminosity relation in the near-IR, given that we will obtain, for the fi rst time, complete light curves for most of the 170 RR Lyrae variables in the cluster. Similarly, we have recently started to monitor a series of open clusters known to host sizeable populations of delta Scuti stars. In addition to high-quality near-IR templates, this will also allow us to verify the period-luminosity relation in the near-IR that has been suggested for this important class of variables (King 1990).
In the more general framework of the VVV Survey, the VVV Templates Project has turned out to be a large observational eff ort in its own right, aimed at creating the fi rst database on stellar variability in the near-IR, i.e. producing a large database of well-defi ned, high-quality, near-IR light curves for variable stars belonging to di fferent variability classes. By monitoring hundreds of (optically well-studied) variable stars in the JHK bands, its primary goal is thus to provide a statistically signi fcant training set for the automated classi fication of VVV light curves.
The astronomical community as a whole has been very supportive, with the end result that we have secured time for this project using several IR facilities across the globe. The availability of telescope time is key to ensure that we produce a signi ficant number of template-grade light curves in time for classi fication of the VVV light curves, i.e. within the next 3 years or so, before the bulk of the VVV variability campaign is carried out. In order to ensure a homogeneous observational strategy and to optimize the use of the awarded time, each telescope/instrument combination is used to build template light curves for at most a very few specifi c variability classes (see the Observatories pages).
Importantly, the scientifi c return of the VVV template light curves will not be restricted to the automated classi fication of the 10^6 light curves produced by the VVV Survey. Rather, these light curves will provide us with a unique opportunity to expand our knowledge of the stellar variability phenomenon per se, by tackling the comparatively ill-explored near-IR regime. Consider, as an example, our upcoming monitoring, using the VISTA telescope, of the variable star content of Omega Centauri. Such a project will lead to high-quality light curves for more than 250 variables, greatly exceeding, both in quantity and in completeness of each individual light curve, what was achieved in previous near-IR studies of the cluster (Del Principe et al. 2006). In particular, since Omega Cen hosts the largest known population of SX Phoenicis stars, some 75 members, we will be in a position to directly derive, for the fi rst time, a period-luminosity relation for SX Phe stars in the near-IR (Freyhammer & Sterken 2003; Weldrake et al. 2007). In like vein, we will also improve the calibration of the RR Lyrae period-luminosity relation in the near-IR, given that we will obtain, for the fi rst time, complete light curves for most of the 170 RR Lyrae variables in the cluster. Similarly, we have recently started to monitor a series of open clusters known to host sizeable populations of delta Scuti stars. In addition to high-quality near-IR templates, this will also allow us to verify the period-luminosity relation in the near-IR that has been suggested for this important class of variables (King 1990).
Upcoming VISTA observations of the globular cluster Omega Cen. Left panel: an
example of variable star distribution across the cluster field: red circles mark the positions of known
RR Lyrae stars, while blue squares mark the positions of known eclipsing binaries. Right panel:
VIRCAM@VISTA's 16 detectors, with Omega Cen at the center of the focal plane.
Some of the first light curves obtained in the framework of the VVV Templates Project are shown in the figures below (K-band light curve for SX Phe, the prototype of its class, on the left; and for WY Sco, a classical Cepheid, on the right). The SX Phe light curve was obtained in the course of an observing campaign carried out with the SAAO 0.75m photometer in Swinburne, South Africa, whereas the WY Sco light curve has been obtained in the course of our ongoing observations with the REM 0.6m robotic telescope in La Silla, Chile. By the end of our monitoring campaigns for these and other stars, the number of data points per light curve will at least double, as required in order to properly defi ne light curve templates.
 K-band light curve of SX Phe, taken at the SAAO 0.75m telescope in Sept. 2010.
Note the overall amplitude of about 0.15 mag. |
 K-band light curve of the classical Cepheid WY Sco, currently being observed with
the REM 0.6m telescope. |
