WVR Scaling Module
Software package to optimise the application of the WVR solutions for ALMA Observations from Maud et al. 2017 (A&A, 605, 121; arXiv). Note the astropy version 1.3.3 is required and CASA < 5.0 (CASA 4.7.2 is the latest working version): WVR_scaling_module_Sep2017.tar.gz
The Atacama Large millimetre/sub-millimetre Array (ALMA) makes use of Water Vapour Radiometers (WVR) which monitor the atmospheric water vapour line at 183 GHz along the line-of-sight above each antenna to correct for phase delays introduced by the wet component of the troposphere. Application of WVR derived phase corrections improve image quality and facilitate successful observations in weather conditions that were classically marginal or poor.
The phase fluctuations after WVR application are noticeably reduced, typically by a factor of 2 or more. In dryer conditions where the PWV <1mm, these corrections are become less significant (factors <2) as the water content causing phase fluctuation has decreased. At Allegro we have studied the effects of scaling the WVR solutions in an attempt to improve the phases of data taken in dryer, low PWV conditions. Maud et al. 2017 (A&A, 605, 121; arXiv) examined the Long Baseline Campaign Science Verification data and found that in dry conditions (PWV <1mm) a scaling factor (>1) applied to the WVR solutions in WVRGCAL can help to reduce the phase fluctuations, improve the coherence of the data and therefore improve the images of the science target. For the tested data the improvements range from 1-2% in most cases, but can be as high as >5-10%. One explanation for the WVR scaling factor is that the wet and dry air become mixed, such that the scaling accounts for extra delays induced by the dry air, that is not directly measurable.
A link is provided at the top of the page to a PYTHON module package that allows anyone with ALMA data to run the WVR scaling tests to see if any improvements can be made. At present and as detailed in the paper, we expect improvements for long baseline (>5km) and high frequency data (>230 GHz).
The figures below show an example of the plots the module produces to indicate what scaling value is needed and the expect coherence improvements. Left – tests made using baselines with the reference antenna only, Right – tests made using all baselines in the array. Tests using all baselines are more robust although are a factor of 5-10 times more time expensive. Please refer to the paper and the module README for more details.