Cycle 2 Capabilities Copy
In the Observing Tool (OT) an observing proposal is specified in terms of Science Goals. A single Science Goal (SG) is constrained to include one set of observational parameters that apply to all sources included in that goal. This includes a single angular resolution, sensitivity, Largest Angular Structure (LAS), and receiver band. For Cycle 2, there is no restriction on the number of Science Goals per proposal.
1. Antennas
All proposers should assume that observations in Cycle 2 will have available thirty-four 12-m antennas in the 12-m main array (hereafter 12-m Array), and that the Atacama Compact Array (ACA, also called the Morita Array) will have available nine 7-m antennas (for short baselines; hereafter 7-m Array) and two 12-m antennas (for single-dish observations; hereafter the Total Power or TP Array). The ACA is used for short baseline interferometry and single-dish observations, and will only be offered to complement observations with the 12-m Array, and not as a stand-alone capability. The use of the TP Array is limited to spectral line observations in Bands 3–8, to the exclusion of continuum observations and no spectral line observations using Band 9.
It may be that, due to problems with the equipment or other reasons, the number of antennas available will sometimes be less than these numbers. In that case the ALMA support staff will endeavor to schedule observations that they believe will not be seriously affected by having a slightly smaller number of antennas. The integration times or uv coverage might be increased to compensate where that is practical.
2. Receivers
Bands 3, 4, 6, 7, 8 and 9 will be available on all antennas. However, observations with Bands 8 and 9 will only be offered for configurations up to ~1 km (see Section A.3). For all bands both linear parallel-hand polarizations of the astronomical signals (XX, YY) are received and processed separately (dual polarization).
The heterodyne receivers multiply the incoming sky signals with a signal from a Local Oscillator (LO) in order to down-convert the signal to an Intermediate Frequency (IF) range, which is more practical to digitize. This results in two 4 GHz-wide Sidebands (per parallel polarization) which are equidistant in frequency from the LO signal. There are two types of receivers: dual-sideband (2SB), where the upper and lower sidebands are separated in the receiver and then processed separately, and double-sideband (DSB), where the sidebands are super-imposed coming out of the receiver but may be separated in later processing.
Table 1. Properties of ALMA Cycle 2 Receiver Bands
| Band |
Frequency range1 |
Wavelength range | IF range | Type |
|---|---|---|---|---|
| (GHz) | (mm) | |||
| 3 | 84 – 116 | 3.6 – 2.6 | 4 – 8 | 2SB |
| 4 | 125 – 163 | 2.4 – 1.8 | 4 – 8 | 2SB |
| 6 | 211 – 275 | 1.4 – 1.1 | 5 – 10 | 2SB |
| 7 | 275 – 373 | 1.1 – 0.8 | 4 – 8 | 2SB |
| 8 | 385 – 500 | 0.78 – 0.60 | 4 – 8 | 2SB |
| 9 | 602 – 720 | 0.50 – 0.42 | 4 – 12 | DSB |
Notes for Table 1:
- These are the nominal frequency ranges for continuum observations. Observations of spectral lines that are within about 0.2 GHz of a band edge are not possible at present in Frequency Division Mode (FDM, see section A.6.1), because of the responses of the spectral edge filters implemented in the correlators.
Although up to three receiver bands will be available at any time, the capability to rapidly switch between them within the same Science Goal (except for the purposes of data calibration) is not offered in Cycle 2.
Water Vapor Radiometer (WVR) measurements to correct for errors due to fluctuations in atmospheric water vapor will be available for all 12-m antennas. These corrections will be applied when it improves the phase coherence. No WVRs will be installed on the ACA 7-m antennas and users should assume that in Cycle 2 no WVR corrections will be applied to 7-m Array observations.
2.1. Band 9 considerations
For Band 9 observations, additional uncertainties will affect the data. Since the sidebands can be only separated reliably in interferometric observations, single-dish Band 9 spectral line observations with the TP Array will not be offered in Cycle 2. Also, owing to the complexity of the atmospheric absorption in Band 9, calibration will be compromised (this also applies to Band 8 and the high frequency end of Band 7). Band 9 ACA 7-m Array observations are more compromised than Band 9 12-m Array observations, since the lack of WVR means that rapid atmospheric phase correction will not be available, and the smaller collecting area will limit the network of available calibrators; in particular bright calibrators will be sparse at these high frequencies. All of these factors will affect imaging at Band 9 during Cycle 2 and will in particular limit the achievable dynamic range with the ACA 7-m Array. Spectral dynamic ranges up to 500 are offered for this Band in Cycle 2 (see Section 8.1 for details).
3. 12-m Array configurations
The antennas of the Cycle 2 12-m Array can be staged into distinct configurations intended to smoothly transition from the most compact (with maximum baselines of ~160 m) up to the most extended (maximum baselines of ~1.5 km, offered for Bands 3, 4, 6 and 7; and ~1 km for Bands 8 and 9). Seven configurations have been defined to represent the possible distribution of 34 antennas over this range of maximum baselines. The detailed properties of these configurations are given in Chapter 7 of the Cycle 2 Technical Handbook.
For all observations, the relevant parameters used by the OT in deciding the required array components for the representative frequencies of a given project are: (1) the Maximum Recoverable Scale (MRS) that can be imaged without the need for the ACA (defined by the shortest baseline of the most compact 12-m Array configuration); (2) the coarsest angular resolution obtainable with the 12-m Array (defined by twice the resolution of the most compact 12-m Array configuration to avoid significant loss of sensitivity); and (3) the finest angular resolution obtainable (defined by the longest baseline of the most extended 12-m Array configuration). These quantities are given in Table 2. Sources with a user-specified LAS larger than the Maximum Recoverable Scale listed in this table will require the addition of ACA observations. Observations with a requested angular resolution either coarser or finer than the values listed in Table 2 (scaled to the appropriate frequency) are not allowed. Values that are inconsistent with any Cycle 2 limitations for the above parameters will result in a warning or a validation error in the OT.
Table 2. Maximum Recoverable Scale1 and Coarsest and Finest Angular Resolutions1 for the Cycle 2 12-m Array configurations
| Frequency |
Maximum Recoverable Scale without ACA2,3,4 |
Coarsest allowed angular resolution2,3,5 |
Finest achievable angular resolution2,3,6 |
|---|---|---|---|
| (GHz) | (arcsec) | (arcsec) | (arcsec) |
| 100 | 25 | 7.5 | 0.41 |
| 150 | 17 | 5.0 | 0.27 |
| 230 | 11 | 3.3 | 0.18 |
| 345 | 7.2 | 2.2 | 0.12 |
| 460 | 5.4 | 1.6 | 0.12 |
| 650 | 3.8 | 1.2 | 0.09 |
Notes for Table 2:
- See Chapter 7 of the Technical Handbook for relevant equations and detailed considerations.
- Computation for source at zenith. For sources transiting at lower elevations, the North-South angular measures will increase proportional to 1/sin(ELEVATION).
- All angular measures scale inversely with observed sky frequency.
- “Maximum Recoverable Scale” is the largest angular structure that can be observed effectively. It is defined by the shortest baseline of the most compact 12-m Array configuration (14 meters).
- Coarsest allowed angular resolution is twice the resolution of the most compact 12-m Array configuration (maximum baseline of 166 meters).
- Finest achievable angular resolution is defined by the resolution of the most extended 12-m Array configuration (~1.5 km for Bands 3-7, and ~1 km for Bands 8 & 9), assuming uniform weighting.
4. ACA
The ACA in Cycle 2 is composed of nine 7-m antennas for the 7-m Array and two 12-m antennas for the TP Array. Two 7-m Array configurations will be offered in Cycle 2, one compact for regular observations and another with significantly less antennas packed close together for observations of objects that transit at low elevations. Both configurations render similar performances, and the decision on which to use for a given project is at the discretion of the Observatory. For more on the ACA see Chapter 7 of the Cycle 2 Technical Handbook.
The TP Array is used to recover, at low angular resolution, all the angular scale information up to the size of the mapped areas. For Cycle 2, the TP Array can only be used for spectral line observations (not continuum) in Bands 3–8. No TP Array Band 9 observations are offered for this cycle. This means that angular scales greater than those listed in Table A 3 cannot be recovered for any observations in Band 9, or for continuum observations in any band. For projects that require both 7-m Array and TP Array, observations will be carried out in parallel as much as possible to optimize the use of the ACA.
Table 3. Maximum Recoverable Scales for ACA 7-m observations
|
Frequency (GHz) |
Maximum Recoverable Scale1,2,3 (arcsec) |
|---|---|
| 100 | 42 |
| 150 | 28 |
| 230 | 18 |
| 345 | 12 |
| 460 | 9.1 |
| 650 | 6.4 |
Notes for Table 3:
- Computation for source at zenith. For sources transiting at lower elevations, the North-Source angular measures will increase proportional to cosec(ELEVATION).
- All angular measures scale inversely with observed sky frequency.
- “Maximum Recoverable Scale” is the largest angular structure that can be observed effectively. It is defined by the shortest baseline of the 7-m Array (8.9 meters). See Chapter 7 of the Technical Handbook for details.
Observations with the 12-m Array and the ACA will be conducted independently, and the data from the different arrays will be calibrated separately and combined during data reduction.
5. Time estimates for multi-configuration observations
Images that obtain a high fidelity over a broad range of angular scales require observations taken with a continuous range of antenna baseline separations. The user-requested angular resolution (θ) determines the most extended 12-m configuration that is needed (up to the “finest allowed angular resolution” listed in Table 2), and the user-requested sensitivity plus calibration requirements determine the amount of observing time needed in this configuration (Δtextended). For point sources, only a single configuration is sufficient to reach the user-requested resolution. For non-point sources, the OT uses the user-provided LAS and angular resolution to determine all the array components needed. Interested users should refer to Chapter 7 of the Cycle 2 Technical Handbook for a table of the array combinations needed to recover various angular scale ranges.
For the purposes of proposal preparation, the time needed for the different array components (including calibrations), referenced to the time needed in the most extended 12-m configuration, has been defined as 4Δtextended for the TP Array, 2Δtextended for the 7-m Array and 0.5Δtextended for a more compact 12-m Array configuration (if needed).
The total time required by a proposal is estimated in the OT by adding the expected observing times for both the 12-m Array and the ACA. For Cycle2, this total time must be less than 100 hours. Table 4 lists the total observing time estimates for the different array combination possibilities offered in Cycle 2. For this computation the fact that 7-m Array and TP Array observations will be done, as much as possible, in parallel has been taken into account, i.e. the ACA time is the TP Array time if this array is used or otherwise the 7-m Array time. Please note that there are two project execution queues, one for the 12-m Array and one for the ACA. Therefore, the time available for Cycle 2 observations is about 2000 hours for the 12-m Array and the same for the ACA. The time accrued by a proposal using the 12-m Array and the ACA will therefore be charged to the two queues separately, as per the time requirements estimated by the OT for each array.
Table 4. Total Time multiplication factors for multi-array observations
| Array Components needed (based on θ and LAS) | Total Time estimate |
|---|---|
| Single 12-m Array configuration | 1.0 Δtextended |
| Two 12-m Array configurations | 1.5 Δtextended |
| Single 12-m Array configuration and 7-m Array | 3.0 Δtextended |
| Two 12-m Array configurations and 7-m Array | 3.5 Δtextended |
| One 12-m Array configuration and 7-m Array and TP Array (spectral line, Bands<9) | 5.0 Δtextended |
| Two 12-m Array configurations and 7-m Array and TP Array (spectral line, Bands<9) | 5.5 Δtextended |
The fixed time ratios adopted in Table 4 were selected to get good imaging performance between the different array components (see Chapter 7 of the Cycle 2 Technical Handbook for more details). However, given the many fewer baselines in the 7-m Array compared to the 12-m Array, good imaging may not be obtainable with short observations (estimated ACA times of less than ~1 hour). If, because of this issue, users want to request longer 7-m Array observations, a detailed request must be added to the Technical Justification.
Based on the LAS, the OT will advise whether the ACA is needed for a given project. If the user chooses not to follow this recommendation, it must be explained in the Technical Justification.
6. Spectral capabilities
6.1. Spectral windows, bandwidths and resolutions
The ALMA IF system provides up to four basebands (per parallel polarization) that can be independently placed within the two receiver sidebands. For 2SB receivers (Bands 3–8 – see Table 1), the number of basebands that can be placed within a sideband is 0, 1, 2, or 4. For DSB receivers (Band 9), any number of basebands (up to 4) is acceptable.
The 12-m Array uses the 64-input Correlator, while the 7-m and TP Arrays use the 16-input ACA Correlator. Both correlators offer the same spectral set-ups. They operate in two main modes: Time Division Mode (TDM) and Frequency Division Mode (FDM). TDM provides modest spectral resolution and produces a relatively compact data set. It is used for continuum observations or for spectral line observations that do not require high spectral resolution. FDM gives high spectral resolution and produces much larger data sets. It is used for observations of spectral lines in all sources except when coarse spectral resolution is sufficient. Six FDM set-ups with different bandwidths and spectral resolutions are available (see Table 5).
For each baseband, the correlator resources can be divided across a set of spectral “windows” (spw) that can be used simultaneously and positioned independently. For Cycle 2 up to four spectral windows per baseband are allowed. A fraction of the correlator resources are assigned to each spectral window, which sets the number of channels and bandwidth of the spectral window. The sum of the correlator resources spread across all spectral windows must be less than or equal to one.
The correlator can be set to provide between 128 and 3840 channels within each spw, and the fraction of correlator resources that are assigned to each spw sets the number of channels and the bandwidth available within it.
Cycle 2 allows the data to be pre-smoothed in the correlator by averaging (or binning) spectral channels in powers of 2. This allows one to reduce the data rate without increasing the sampling integration time (see Chapter 4 of the Cycle 2 Technical Handbook for more information). In Cycle 2, the maximum data rate is 60 MB/s, with the expected average of 6 MB/s. If the spectral setup ends up with a data rate that is more than twice the expected average of 6 MB/s, the user will need to technically justify this.
Different correlator modes can be specified for each baseband, but all spws within a given baseband must use the same correlator mode. For example, a high-resolution FDM mode can be used for spectral line observations in one baseband (with up to 4 differently placed FDM spectral windows), while the other three basebands can be used for continuum observations using the low-resolution TDM mode. And while each spw within a baseband must use the same correlator mode, they can each be assigned a different fraction of the correlator resources and each use a different spectral averaging factor, providing a broad range of simultaneously observed spectral resolutions and bandwidths. Spectral windows can overlap in frequency, although the total continuum bandwidth for calculating the sensitivity is set by the total independent continuum bandwidth.
Table 5. Properties of ALMA Cycle 2 Correlator Modes, dual-polarization operation1,2
| Bandwidth(3) | Channel spacing(4) | Spectral resolution | Number of channels | Correlator mode |
|---|---|---|---|---|
| (MHz) | (MHz) | (MHz) | ||
| 20003 | 15.6 | 31.2 | 1283 | TDM |
| 1875 | 0.488 | 0.976 | 3840 | FDM |
| 938 | 0.244 | 0.488 | 3840 | FDM |
| 469 | 0.122 | 0.244 | 3840 | FDM |
| 234 | 0.061 | 0.122 | 3840 | FDM |
| 117 | 0.0305 | 0.061 | 3840 | FDM |
| 58.6 | 0.0153 | 0.0305 | 3840 | FDM |
Notes for Table 5:
1. These are the figures for each spectral window and for each polarization, using the full correlator resources and no on-line spectral binning.
2. Single-polarization modes are also available, which gives twice the number of channels per spw, and half the channel spacing of the above table.
3. The "Bandwidth" given here is the width of the spectrum processed by the digital correlator. The usable bandwidth in all modes is limited to a maximum of about 1875 MHz by the anti-aliasing filter, which is ahead of the digitizer in the signal path. For TDM modes, the anti-aliasing filter also limits the total bandwidth to about 1875 MHz and the number of channels to about 120.
4. The “Channel Spacing” is the separation between data points in the output spectrum. The spectral resolution – i.e., the FWHM of the spectral response function – is larger than this by a factor that depends on the “window function” that is applied to the data in order to control the ringing in the spectrum. For the default function – the “Hanning” window – this factor is 2. See the Technical Handbook for full details.
6.2. Polarization
For Cycle 2, on top of the dual polarizations (XX, YY) and single polarization modes (XX), observations to measure the full intrinsic polarization (XY, YX) of sources will also be offered for TDM observations in Bands 3, 6 and 7, with some restrictions. Please note that only linear polarization has been commissioned. While PIs will receive data which will allow them to generate circular polarization data, the quality and/or accuracy of that data at this time is not assured, and such data should not be used for scientific purposes.
When a Dual Polarization setup is used, separate spectra are obtained for each linear parallel-hand polarization of the input signal. This will give two largely independent estimates of the source spectrum that can be combined to improve sensitivity.
In Single Polarization mode, only a single input polarization (XX) is analyzed. For a given resolution, this provide √2 worse sensitivity than the Dual Polarization case, but one can use either a factor two more bandwidth for the same spectral resolution or a factor of two better spectral resolution for the same bandwidth. Single polarization should therefore be used in cases where having a large number of spectral channels is more important than having the best sensitivity.
Full Polarization continuum measurements will be offered in Cycle 2 only for 12-m Array observations in Bands 3, 6 and 7. Sources must be centered and have a user-specified largest angular structure that is less than one-third of the 12-m Array primary beam at the frequency of the planned observations. Observations shall be single-field, but measurements of individual sources within a 10-degree area on the sky are possible (one field per source; see below). The continuum polarization measurements are offered only for specific frequency settings, as detailed in Table 6.
Table 6. Fixed frequencies for Polarization Observations1
| Band | SPW1 | SPW2 | LO1 | SPW3 | SPW4 |
|---|---|---|---|---|---|
| (GHz) | (GHz) | (GHz) | (GHz) | (GHz) | |
| 3 | 90.5 | 92.5 | 97.5 | 102.5 | 104.5 |
| 6 | 224.0 | 226.0 | 233.0 | 240.0 | 242.0 |
| 7 | 336.5 | 338.5 | 343.5 | 348.5 | 350.5 |
Notes for Table 6:
- Fixed central frequencies for four TDM spectral windows, each of width 1.875 GHz, and the corresponding LO1 setting. Frequencies were chosen to optimize spectral performance, and they are centered in known low noise and low instrumental polarization tunings of the receivers.
It should be noted that full polarization observations require sufficient parallactic angle coverage for calibration (about 3 hours). Science Goals with properties that lead to a total observing time estimate that is less than 3 hours will have the time estimate set to 3 hours to ensure sufficient parallactic angle coverage is obtained.
7. Source restrictions
Sources can be designated by a fixed RA and Dec, or can include moving targets (including the planets, their moons, asteroids and comets). Observations of the Sun, however, are not supported in Cycle 2.
Sources are selected in one of two ways: by specifying a single rectangular field, or by specifying one or more source positions, with or without offsets. Each involves some restrictions. The total number of positions in a Science Goal (SG) must be less than or equal to 150 and all must lie within 10 degrees of each other. Pointings with the ACA, if used, do not count against the 150 pointing Science Goal limit.
7.1. Single rectangular field
A rectangular field (also referred to as a mosaic) is specified by a field center, the length, width and orientation of the field, and a single spacing between the pointing centers. Observations are conducted using the “mosaic” observing mode. This repeatedly cycles through all the pointings in the mosaic so that the imaging characteristics across the map are similar.
One rectangular field is allowed per Science Goal. A single mosaic can have up to 150 pointings and must be done in a single frequency set-up. If ACA observations are requested as part of a mosaic, then a corresponding 7-m Array mosaic will also be observed. If these are spectral line observations, the full mosaic area can also be covered by the TP Array using On-The-Fly mapping.
The OT will set up a uniform mosaic pattern based on a user-specified pointing separation, and will calculate the time to reach the required sensitivity considering any overlap. Non-Nyquist spatial samplings are allowed (up to HPBW/√2), where HPBW is the angular size of the half-power width of the Gaussian primary beam of a single 12-m antenna, at the frequency of the observations. Sparser samplings must be justified in the proposal. If the position separations are not too large, then the interferometric data are combined in post-processing to produce a single image. Mosaics in different Science Goals will not be combined during post-processing.
7.2. Individual pointings
If a user does not wish to specify a rectangular field, they may include in a single Science Goal a mixture of sources and offsets, provided that:
- They are not separated by more than 10 degrees on the sky;
- They can be observed with one spectral setup (relative placement and properties of spectral windows);
- They can be observed with no more than five separate frequency settings that all fall within the same receiver band;
- The sum over all sources, offsets, and frequency settings is less than or equal to 150.
For more widely spaced targets, such as wide-area surveys, additional SGs may be used for sources separated by more than 10 degrees.
Offsets can be specified for all sources within a Science Goal, but the 150 pointing limit applies. If ACA observations are requested for the Science Goal, then the corresponding ACA observations will be obtained for each source. Sets of offsets are designated either as a “Custom Mosaic” or a “Pointing Pattern”. The former are observed using the “mosaic” observing mode and can be separated by no more than HPBW/√2). The interferometric data will be combined in post-processing to produce a single image. The latter are not observed as mosaics, do not have a separation constraint (apart from the 10-degree separation limit of a Science Goal), and will not be combined to produce a single image.
For offsets, the OT does not consider the effect of overlapping pointings; users must take this into account when specifying the required sensitivity.
7.3. Spectral scan mode
Proposers who wish to carry out spectral surveys or redshift searches can do so using the “Spectral Scan” option in the OT to automatically set up a set of contiguous spectral windows to cover a specified frequency range, provided that:
- All targets are separated by less than 10 degrees on the sky;
- Angular resolution and LAS are computed for the Representative Frequency of each SG;
- No more than 5 frequency tunings are used, all in the same band;
- Only one pointing per target (no mosaics or offsets allowed);
- The sum, for all targets, of the number of separate tunings required per target does not exceed 150 (i.e., the maximum number of targets, for 5 tuning for all targets in a SG, is 30);
- Only 12-m Array observations are required (the ACA is not offered for this mode).
8. Calibration
Absolute amplitude calibration will be based on observations of objects of known flux, principally solar system objects. It is expected that the accuracy of the absolute amplitude calibration relative to these objects will be better than 5% for Bands 3 and 4. Calibration in the higher frequency bands is likely to be less accurate. The goal is for it to be better than 10% in Bands 6 and 7. Calibration at Bands 8 and 9 will be challenging even at the 20% level owing to the high atmospheric opacity, particularly so for Band 9 Total Power observations due to the Double-Sideband nature of the receiver (see Section 2).
The ALMA Observatory has adopted a set of strategies to achieve good calibration of the data (see Chapter 10 of the Cycle 2 Technical Handbook). Requests for changes in these strategies will only be granted in exceptional circumstances and must be fully justified by the requester. Some flexibility exists in choosing the actual calibrator sources. The default option is automatic calibrator selection by the system at observing time. If users opt for providing their own calibrators, justification will be needed. This may result in decreased observing efficiency and/or calibration accuracy.
8.1. Bandpass accuracy
The detailed shape of the spectral response of the arrays during observations depends on many factors. This shape particularly affects projects that intend to observe spectral features that cover a significant fraction of a spw, and/or study spectral features with small contrast with respect to a strong continuum. It has been determined that, for Cycle 2, projects that require spectral response accuracies (i.e., the desired signal-to-noise ratio per spectral resolution element), per observation execution, of up to 1000 for ALMA Bands 3,4,6 and 500 for Bands 7, 8 and 9 are feasible. Requests for higher accuracies may be the grounds for rejection of the proposal.
9. ToO and time-constrained observations
Observations of ToO, monitoring and time-constrained projects will be offered in Cycle 2 with a few restrictions:
- Observations must be done in only one 12-m Array configuration; time constraints cannot be imposed with the ACA;
- Monitoring projects that require observation windows spaced by less than two weeks are unlikely to be fully observed;
- ToO projects that require observations within two weeks of contact with the ALMA observatory cannot be guaranteed to be executed.