VISTA Near Infrared Transit Survey

PI: David Pinfield

List of interested people

Survey structure developed

After a variety of useful input and discussion with many of you, I have designed the transit survey as follows:

In the end I opted for the poor weather option (using 1.2-2 arcsecond seeing and/or thin cloud). Simon and I are satisfied (from analysis of WFCAM data) that the accuracy of the photometry will not be compromised. And the whole survey could be done in 2 years using only these type of conditions (this approach would also make for efficient Vista queue observing).

Single filter J-band observations (except for 2 regions in the bulge; see below) will allow for sufficient cadence to maximise transit detection. This was the concensus of opinion, and my simulations agree.

20 separate survey regions, each of 4.8 sq degs (16 separate Vista paw-prints, untiled to maximise sky coverage), giving a total coverage of ~100 sq degs.

Field coordinates are in the dec=-30>dec>-45 range, to deal with alt-az and wind issues. Galactic latitudes all |b|=10-20 (except the 2 bulge fields). This will give us a high density of comparison stars, to minimize systematics (e.g flat field variations).

The regions will be monitored with 50 separate 1 hr observing blocks spread across 2 yrs. Each block will do 5-point jitters on each of the 16 paw-prints in turn, and then cycle this process 7 times in the hour, giving 35 observations per star per block. 50 1 hrs blocks for each of the 20 regions requires 100 nights over 2 years.

2 of the fields have been located in the bulge. This will allow a microlensing search for planets (Eamonn Kerins is coordinating the details for this, and is likely to use K instead of J, and do 20 minute blocks to increase triggering frequency) as well as transits (albeit brighter star transits - M dwarf monitoring will be rather compromised towards these 2 fields).

6 of the fields are contained in the scattered Kuiper belt region, and will allow solar system objects to be discovered as well (e.g. KBOs and Plutinos). Stephen Lowry is providing input here.

Simulation

The simulated planet catch from the survey is incouraging, and will be included in the draft proposal that I will circulate soon.

Team contributions

However, it would be useful for people to put themselves forward at this stage, to be formally associated with certain aspects of the proposed project. Below is a list of areas that I think should be covered in the survey plan. I have penciled in some names already based in part on people's previous comments... please say if you agree. But I would like others to let me know where they feel they would like to contribute most. Some additional information could also be useful. e.g. if you have guaranteed time, paly a leading role, or have lots of experience with particularly useful followup facilities, this would be beneficial to the program and could strengthen the proposal.

General areas

• Creation of archive tools to make light curves for instance. To some extent handled by pipeline/archive people (Simon Hodgkin, Mike Irwin?). But others may wish to do something (for instance, setting up an "orphan catalogue" for transient work -- Pete Wheatley?).

• Writing codes to search for transits within light curve data in an optimal way.

• Simulation codes to interpret transit survey results. I have a code that simulates the survey. However others have done similar simulations which which would be useful (Michael Gillon and Cristina Afonso).

• Spectral typing followup of candidate transit sources. Easy to do, but may need a good chunk of time on 2-4m optical or NIR scopes.

• Multi-band photometry to confirm transits and remove contamination. Observations would obviously need to be well timed (Robotic facilities?).

• Radial velocity followup. For instance, Hugh Jones is leading a Gemini bid for the PRVS near infrared echelle spectrograph.

• There will also be some high proper motion astronomy that would come out of the survey (reaching a depth of J=22.2). I would be interested in making a search for nearby brown dwarfs, and Ralf Napiwotzski would be interested a search for helium rich white dwarfs.

• Transient phenomina and supernovae will also be found (~14 Sn for e.g.). Although the turn around time for data processing (~10 days or so) will preclude very rapid followup, other studies could be made.

...and anything else you think I have missed.

(please email me with comments)

Initial discussion points:

The main science driver

By carrying out a transit survey (ie. repeat observations of some regions of sky) in the near infrared, one is sensitive to companion transits of M dwarfs. Because M dwarfs have radii about 3-10 times smaller than solar type stars, detectable transits can have radii as small as 10% that of the hot Jupiter transits found previously in the optical. Further more, the habitable zones around M dwarfs are much closer (0.02-0.4AU) than around solar type stars, and despite the likelihood that such close systems will be tidally locked, stable atmospheres are possible (Joshi Haberle & Reynolds 1997), and could be studied (Segura et al. 2005). M dwarf transits could thus be habitable rocky planets.

Other areas of astronomy that could benefit from a transit survey

Time variability astronomy

(Linking in with Robonet)

• Super novae
• Novae
• Variable stars
• Micro-lensing

Other types of transit

• Brown dwarf/white dwarf - late star transits
• Double star transits

Faint moving objects

• Very faint high proper motion objects
  • Nearby brown dwarfs and free-floating planets
  • Cool white dwarfs
• Solar system objects

Deep image stacking

• Extra-galactic high z surveys
• Very faint companions to stars at wide separation

Are there any others ?

An initial working model and optimisation issues for debate

While the precise form of the proposed Vista Transit Survey is very much open to discussion, I would start by suggesting that we follow, to some degree, the method currently being used by Hodgkin & Pinfield with WFCAM on UKIRT (see below, and our WFCAM proposal). Early analysis of our WFCAM data will provide some further information on optimised observing strategy for transits. However, some flexibility will remain in;

• Coordinates of the fields (around the ideal locations; see below)
• Individual Field size
• Cadence
• Length of observation block
• Filters used
• Total number of observing blocks
• Weather conditions

I am in the process of re-running my WFCAM simulation for a VISTA survey, which will provide one "take" on how the survey could go.

But please have a read below, and provide input that you feel will allow the survey to be optimised for both transit detection as well as other important areas of astronomy. I give some specific questions in italics.

The targeted transit sources

It is vital to be able to follow up transit sources with radial velocity measurements. A Vista Transit survey would produce accurate light curves for a magnitude limited sample (to J~16) that could be followed up and confirmed with existing or planned instrumentation on 4-8m class telescopes (e.g. PRVS+Gemini, CRIRES+VLT, NAUHAUL+GranTeCan, NIRSPEC+Keck, HARPS).

Other telescopes/instruments that could be used to confirm transits ?

Sky coverage

The survey would consists of a number of fields (each of some chosen size) with galactic latitude b<20, and declination around -25. The lowish value of b is to insure that we have the maximum number of M dwarfs contained in the field, and also that we have a sufficiently large number of comparison stars to facilitate very accurate light curves. There will be more details on this soon, as we are planning to analyse our early WFCAM data. The declination should be approximately matched to the latitude of the telescope (i.e. -25 for VISTA) for scheduling reasons - so that the fields are available for maximum time in a night. There is some flexibility here. For example, with WFCAM, we moved a field slightly closer to the Pleiades to provide deep NIR coverage and potentially pick up transits in the cluster. One would also move fields to avoid certain regions. For example, it would be undesirable to point towards star forming regions due to the presence of e.g. disk induced variations (e.g. Caballero et al. 2004, 2006).

Any suggestions of coordinates for fields ?

Observing blocks

Repeat observations of any one of the survey fields are made in a series of observing blocks (ie MSBs), designed to be completed in some specified time. This allows for flexibility in scheduling - blocks may be observed back to back, or could be done more randomly in time, but either way one should be able to measure a whole transit (or at least a good fraction of one) if it occurs during the block. With WFCAM we chose 1hr blocks. This length of time was the maximum that allowed blocks to be easily attempted as queued UKIRT observations. It is also fairly well matched to the expected transit durations - between 30 minutes and an hour for habitable M dwarf planets.

One must adjust the size of survey fields (to be covered in a block) so as to give the best balance between cadence and coverage. With WFCAM we went for 4 tile areas (ie. 16 paw-prints). With Vista, we could obviously cover alot more sky per field.

The total number of observing blocks determines the overall time allocation required for the survey. This number is essentially what dictates the "planetary mass - orbital separation" parameter space that the survey will explore. More on this to come via simulation.

The length of observing blocks ?

The area covered by each observing block ?

How many observing blocks in total ?

Observing conditions

On UKIRT, our program (at least partly) takes advantage of "bad weather" time. Bad weather means thin cloud (~20% variations), and/or poor seeing (>0.8"). Relative photometry should not be a big problem through thin cirrus, although poor seeing and thin cloud will limit photometric depth somewhat. But there is, of course, a distinct advantage in being able to do good science with poor weather conditions, and one can expect a certain fraction of the nights in any year to fall into the "bad weather" category. Weather stats for the VISTA site suggest seeing worse than 0.8 arcseconds 25% of the time and cloudy conditions 23% of the time (not clear what fraction would be "thin" cloud).

One can always build in an option to force observing blocks to the top of the queue (in good or bad conditions) to improve phase coverage if poor.

As an alternative, or an addition, one could schedule longer survey periods (e.g a week at a time, during which time blocks are done back to back). Such a tactic would pick up short period transits, but could reduce some systematic uncertainties.

Bad weather program ? Good weather program ? or a combination ?

If some bad weather time used, how should we define "bad" ?

Filters used

The WFCAM program uses only the J-band filter. M dwarfs are detectable out to the greatest distance in this band. Using only one band minimises filter change overheads during an observing block, and provides rapid cadence, more potential "in transit" measurements, and higher signal-to-noise for any detected transit. A one filter survey will produce a larger primary data-set of transit candidates, and leave any multi-band analysis to followup observations. Note that all potential transit sources have 2MASS JHK, and M dwarf targets can thus be identified from their photometry.

However, one might consider doing multiple band observations. This could allow for the early identification of potential sources of contamination amongst low-level transit candidates (e.g. photometric variability, pulsation). However, filter changes within each block would limit cadence and thus limit sky coverage for a block. This would reduce the number of transits found.

One filter or more ?

If more, which ones ?

VISTA consortium web pages

Click here to return to main VISTA Twiki Page