Command line interface to libAIR and water vapour column retrieval

Since version 0.12, libAIR comes with a simple command line interface to the routines which retrieve the water vapour and the phase correction coefficient. Currently only the very simplest model is supported, in which the all of pressure, temperature and water vapour column are free parameters, and they are inferred from the four observed water vapour radiometer readings.

The programs are located in the cmdline directory of the library package. See Building libAIR for instructions for building of the library. You need to have at least libAIR V0.12 and BNMin1 V1.8 to use the command-line programs.

Usage & Example

To get help on usage of the program use the –help option:

./wvrretrieve --help
Make a retrieval of water vapour and correction coefficients from WVR data

This program is part of LibAIR version 0.19
GPL license -- you have the right to the source code. See COPYING

You should supply the observed WVR temperatures as four positional arguments:
e.g.:

./wvrretrieve <Ch1Temp> <Ch2Temp> <Ch3Temp> <Ch4Temp>
where channels are numbered from line centre out

Allowed options:
  --version             Print the name and version of the program
  --help                Print (this) help message
  --wvrtemp arg         Temperature recorded by WVR, from innermost channel
                        outward
  --elevation arg (=90) Elevation of observation in degrees
  --cont                Try to constrain the continuum
  --osf                 Apply priors suitable for the OSF (if this option is
                        not use then priors are for AOS)

An example invocation of the program is:

./wvrretrieve 280.457 228.251 150.674 111.533
Evidence: 2.07525e-14
 -- Parameters --
    c: 5.31315 +/- 0.118196
 -- Phase Correction Coefficients --
dTdL0: 0.279412 +/- 0.000844754
dTdL1: 2.34219 +/- 0.00567481
dTdL2: 2.95852 +/- 0.00458467
dTdL3: 2.46791 +/- 0.00316159

Here 280.457 228.251 150.674 111.533 are the four brightness temperatures recorded by the WVR (this was relatively wet weather at the OSF). The remainder of the text is the output from the program:

  • Evidence is the Bayesian evidence for the model & data: p(D|H). This is a very useful quantity in diagnosing problems or strange results
    1. Evidence much smaller than usual probably indicates a problem with the WVR reading
    2. Evidence value somewhat smaller than usual can indicate effects that are not modelled such as continuum due to cloud or several different water vapour layers and significantly different pressures/temperatures
  • The line-of-sight water vapour column in millimetre c with first-order estimate of error. Note that future versions will support an elevation parameter, and when that is supplied c will refer to the zenith column.
  • The four phase correction coefficients, expressed as ratio of change in temperature in K for a given change in non-dispersive path in millimetres

Table Of Contents

Previous topic

ALMA WVR phase correction algorithms and software

Next topic

Atmospheric Modelling for ALMA