This performs the calculations in SDP memo 97.

• A phase screen at the height of the ionosphere (300km) moving at 150 km/s is simulated.
• A sky model is calculated from GLEAM.
• Nonisoplanatic visibilities from the screen are calculated taking into account the two pierce points for a given baseline.
• Isoplanatic visibilities are calculated using just a single pierce point for all stations.
• ICAL is run on both isoplanatic and non-isoplanatic data.
• MPCCAL is run on the non-isoplanatic data, starting from the model and gaintable from the ICAL.

• Install ARL from https://github.com/SKA-ScienceDataProcessor/algorithm-reference-library
• Install ARatmospy from https://github.com/shrieks/ARatmospy

The processing steps are in a python file and jupyter notebooks:

• To initialise the phase screen file: cd screens;python ArScreens-LOW.py
• To simulate the data: low-sims-mpc-10000-prep.ipynb
• To perform isoplanatic selfcal (ICAL) of isoplanatic data: low-sims-mpc-10000-iso.ipynb
• To perform isoplanatic selfcal (ICAL) of nonisoplanatic data: low-sims-mpc-10000-noniso.ipynb
• To create skymodel for mpc: low-sims-mpc-10000-create_skymodel.ipynb
• To perform non-isoplanatic selfcal (MPCCAL) of nonisoplanatic data: low-sims-mpc-10000-mpccal.ipynb

These steps can be run using make.

The output files are placed in a sub-directory results. The images are standard FITS files, and all the other files are HDF5. Figures are placed in the figures sub-directory.

• Change the number of Dask workers (n_workers) in each file to a value suitable for your machine. The default is 16 workers.