ALMA Observing Proposals
A General Overview
The majority of the following overview is from "Observing with ALMA - A Primer for Early Science", produced by the University of Calgary Department of Physics and Astronomy. It is recommended that potential observers read the ALMA Call for Proposals for the most up-to-date summary of ALMA capabilities. Additional documentation and tools for proposers are available on the Documents and Tools website.
Phase I consists of a detailed observational proposal with a scientific and technical justification, that is submitted to the Observatory through the Observing Tool (OT). The OT has a calculator for determining sensitivities and viewers for assisting with correlator setups. Additionally, an ALMA simulator (SIMDATA) is available to help predict what ALMA will see given various input models, observing conditions and observing strategies. Phase I submissions are peer-reviewed by committee and time is awarded appropriately.
Basic Set-up: Key Observational Parameters
While considering a possible ALMA project, it is important to understand that ALMA is a very flexible instrument. Data can be obtained over a wide range of observational parameters: angular resolution, field-of-view, spectral resolution, and sensitivity. These quantities must be specifically defined and justified for a given project during Phase I of the proposal process and proper choices are required to ensure the project's scientific goals can be met. These quantities are also used during Phase II, to guide in planning the execution of the project. Depending on the nature of a given project, these quantities may be inter-related. Below, we describe the basis for choosing these parameters.
is the minimum angular separation whereby adjacent and independent spatial features can be distinguished. Angular resolution fundamentally varies as the inverse of the product of observational frequency and distances between the antennas used to make the image; higher frequencies or longer inter-antenna distances result in data of higher angular resolution. An important concept to remember about interferometers is that they can only observe emission on a discrete set of spatial scales (i.e., frequencies), as measured by the antenna pairs making up an array. Since the number of spatial scales measured is finite, the resulting image is "filtered" spatially, and only reflects the emission on the observed spatial scales. The actual angular resolution achieved in a reconstructed image depends upon how the data are weighted during the imaging process. During imaging, interferometer data are commonly weighted "naturally," i.e., by the relative densities at which angular scales on the sky have been sampled in the data ensemble. Naturally weighted images have the highest sensitivity possible for a given band width and integration time but they may have lower angular resolution than would be expected simply from the separation of the furthest antennas. Other weighting schemes can yield images of higher resolution but at an expense to sensitivity.
Related to angular resolution is the largest observable scale that can be reconstructed during image processing. This quantity is related to the minimum distance between antennas. Interferometers are insensitive to emission on large angular scales because the telescopes cannot be moved arbitrarily close to one another. To recover emission that has been "resolved out," additional observations are needed, including observations with more compact arrays with smaller-sized antennas (such as the ACA) or large single-dish telescopes (ALMA uses its own antennas to sample large scale emission by scanning over fields in non-interferometric, "total power" mode).
is the area on the sky over which an interferometric image is obtained. The instantaneous FOV is formally the angular size of the half-power width of the Gaussian beam of the individual antennas, and is also called the width of the "primary beam". The size of the FOV depends on the inverse of the product of the frequency of the observation and the diameter of the individual antennas used; larger antennas or higher frequencies result in smaller FOVs. For a single pointing, the sensitivity of the observation is not uniform across the FOV; it declines with angular separation from the center position with the Gaussian responsivity of the main antenna beam. Larger FOVs and flatter map sensitivities across images can be attained by observing in series many adjacent locations on the sky (best separated by half the FOV to achieve Nyquist sampling), and using the resulting data to create a "mosaic" map. In order to have constant sensitivity across the mosaic, each pointing must be observed to the same relative sensitivity. Thus, mosaics can be quite costly in terms of observing time. Deciding on whether a mosaic or a single pointing should be observed requires an understanding of the expected source structure and size, i.e., whether or not the observed emission will be extended. Furthermore, if multi-band images over the same FOV are needed for a given project, mosaics may be required with higher frequency bands in order to match the aerial coverage of a single pointing with lower frequency bands. Note that mosaics can also aid in recovering some emission on scales larger than those that are sampled by single pointings.
is the minimum separation in frequency whereby adjacent independent features can be distinguished. The digitized data from ALMA allows for an incredible range in spectral resolution. Spectral resolution depends on how the correlator has been configured prior to observations. ALMA's correlator can be configured to provide data cubes with up to 8192 independent spectral channels. The width of these channels can be defined from 3.8 kHz to 25 MHz. For continuum observations, low spectral resolution (i.e., large band width) channels are averaged to achieve high sensitivity; the total band width of all correlator settings used cannot exceed 8 GHz. For line observations, high spectral resolution (i.e., small band width) is used to achieve high velocity resolution. For example, a 0.01 km/s velocity resolution of R=30,000,000 can be achieved for 13CO 1-0 at 110 GHz by utilizing the narrowest channels (3.8 kHz). There is, however, a cost to sensitivity in using small band width channels. Sensitivity can be improved after the observations by averaging channels together, i.e., by the inverse square root of the number of channels averaged, but at the expense of the spectral resolution. The ALMA correlator is highly complex and can be configured to observe simultaneously several spectral lines within the 8 GHz band at high spectral resolution while additional correlator channels can be simultaneously used to observe continuum emission at low spectral resolution. In addition, a combination of high and low resolution correlator windows can be chosen over the same band width to determine how emission from lines at these frequencies is contributing to the emission observed at low spectral resolution. Note that the frequencies of spectral lines accessible with ALMA can be obtained by making use of Splatalogue.
is the intensity level at which received emission is equal to the 1 sigma rms variations of noise in the data, and so serves as a threshold for the detection of emission. For ALMA, basic sensitivity depends strongly on receiver performance and atmospheric conditions (i.e., water vapour content, atmospheric turbulence, and target elevation). These effects are quantified by one parameter called "system temperature" (Tsys). High Tsys values (in K) indicate low sensitivity and vice versa. Note that the atmospheric conditions are very frequency dependent and thus the ability to observe with any particular receiver will usually depend strongly on the weather conditions. These conditions include both the water content of the atmosphere above the array that can attenuate astronomical emission but also the stability of the atmosphere above the array that can make it difficult to measure accurate phases. Both of these conditions affect higher frequencies more than lower frequencies.
Two other aspects of the observational set-up strongly affect sensitivity: spectral resolution and angular resolution. Continuum intensities are often given in units of Janskys per beam where 1 Jansky (Jy) = 10-26 W s-1 m-2 Hz-1 while line intensities are given in units of Kelvin (K). Converting from one unit to another requires knowledge about the angular resolution of the data, where the sensitivity in K is proportional to the sensitivity in Jy divided by the angular size of the beam. For a given ΔSν, the corresponding ΔK increases with decreases in beam size; it is harder to detect extended line emission at high angular resolution. The quantity ΔSν itself varies as the inverse-square root of the product of total integration time and the total band width of the observation. (How data are weighted during imaging also affects sensitivity; see above.) The total band width of the observation is determined by the correlator settings and how many spectral channels, i.e, resolution elements, are averaged together. For continuum data, up to 8 GHz band width can be used. Sensitivity also depends on the inverse square root of the number of observed polarizations; all ALMA bands have two polarization channels.