The New Milky Way

a wide-field survey of optical transients
near the Galactic plane

Project summary

Currently, it may take days for a bright nova outburst to be detected. With the few exceptions, little is known about novae behavior prior to maximum light. A theoretically-predicted population of ultra-fast novae with t2 less than one day is evading observational discovery because it is not possible to routinely organize fast follow-up observations of nova candidates. With the aim of brining the detection time of novae and other bright (V<13.5) optical transients from days down to hours or less, we develop an automated wide-field (8x6 degrees) system capable of surveying the whole Milky Way area visible from the observing site in one night. The system is built using low-cost mass-produced components and the transient detection pipeline is based on the open source VaST software. We describe the instrument design and report results of the first observations conducted in October-November 2011 and January-April 2012. The results include the discovery of Nova Sagittarii 2012 No. 1 as well as two X-ray emitting cataclysmic variables 1RXS J063214.8+25362 and XMMSL1 J014956.7+533504. The rapid detection of Nova Sagittarii 2012 No. 1 enabled us to conduct its X-ray and UV observations with Swift 22 hours after discovery (about 31 hour after the outburst onset).

A detailed description of the survey is presented in this arXiv e-print.
Please cite it if you use NMW data in a publication. A short presentation describing the project may be found here.

Online image access

All images obtained during the transient search survey are available online through this webform:

Please click only once and be prepared that image search and resampling may take up to a few minutes!

You may get direct access to the image files here.


The transient detection system consists of a wide-field survey camera: an F=135 mm f/2.0 telephoto lens attached to an unfiltered ST8300M CCD installed on a HEQ-5 Pro robotic mount, a Windows-based computer controlling the mount and the camera through a custom-made software developed by Vasily Vershinin and a Linux-based data-reduction computer. The system is operated from an altitude of 2000 m by Ka-Dar Observatory at Karachay-Cherkessia (MPC code C32) at the Russia's North Caucasus mountains. The system parameters are summarized in the Table.

The New Milky Way (NMW) survey parameters:

Optics: Canon 135 mm f/2.0 Accuracy at V=11: 0.1m absolute
CCD camera: SBIG ST-8300M 0.05m internal
(blue-sensitive chip) Images per field: 2 or 3 per night
Image size: 3352 x 2532 pix (dithering)
Optical filter: none Stars per frame: 20000
Equatorial mount: HEQ-5 Pro Images per night: 200
Field of view: 8 x 6 degrees Milky Way imaging time: 5 hr (January)
Pixel scale: 8.4"/pix 10 hr (April)
Exposure time: 20-40 sec Processing time: up to 7 hr/night
Limiting magnitude: V<14.5 Results inspection time: up to 4 hr/night
Transient detection limit: V<13.5

For data reduction we choose object detection rather than image subtraction, because the former approach is computationally more efficient and provides better photometric accuracy despite being less sensitive in crowded fields. The transient detection pipeline is implemented as a BASH script controlling the VaST software which performs the following steps. (i) Star detection and circular aperture photometry using SExtractor. (ii) Cross-matching the lists of stars detected on reference and second-epoch images. Transient candidates are identified as objects detected on all second-epoch images that were either not detected on the reference image or were at least 1m fainter. The second criterion is needed to identify a low-amplitude flare or a high-amplitude flare of a faint object that is blended with a brighter one visible on the reference image. Small movement of the telescope between images (dithering) is used to to reject CCD artifacts. (iii) Images are plate-solved using the software and celestial positions of the detected stars obtained. (iv) The instrumental magnitude scale is calibrated using V magnitudes of unsaturated Tycho-2 stars in the field of view.

The main problem of our transient detection system is that it currently lacks an automated enclosure. While being able to perform observations automatically, it still has to be set up, started and stopped manually. Thus, observing time is limited by the availability of a qualified observer on site.

The images obtained within the NMW survey are available at this webpage. The image archive server is written in Haskell functional programming language with Yesod web framework and is using SWarp for fast image resampling. The archive of photometric measurements is currently under development.


The "New Milky Way" camera.

Example image of the Lagoon (M8) and Trifid (M20) nebulae region in Sagittarius.

The observing site.

The observing site location.

Questions or comments? Please write to
kirx [at]