A set of standardised aggregation methods to produce an ensemble of aerosol optical depth observations.

Over a rectilinear latitude-longitude grid (currently, taken from UKESM1), this calculates, for two wavelengths of aerosol optical depth the,

• Joint distribution over 44 bins roughly log-spaced between 0 and 5,
• Maximum observed aerosol optical depth,
• Mean;
• Mean of the base-10 logarithm. The precise wavelengths vary by instrument but aim to be 550 and 670 nm, with 440 nm substituting the later when unavailable.

The code currently supports the following datasets:

• Three datasets for AATSR and ATSR-2 developed by the European Space Agency's Aerosol CCI program:
  • ATSR Dual View (ADV), from the Finish Meteorological Institute;
  • Swansea University's (SU) retrieval, which is used operationally by the Copernicus Climate Change Service;
  • Optimal Retrieval of Aerosol and Cloud (ORAC), from the University of Oxford, RAL Space, and others.
• Three datasets for MODIS Terra and Aqua funded by NASA:
  • Dark Target;
  • Deep Blue;
  • MAIAC.

To Do:

• Remove the hard-coded paths.
• Complete code to determine the number of pixels observed by each sensor.
• Fit bi-log-normal distributions to the histograms.
• Convert current npz output files into CF-compliant netCDF files.
• Implement additional algorithms that are currently available on JASMIN:
  • Deep Blue for AVHRR and SeaWIFS;
  • The AVHRR product released by FIDUCEO;
  • MISR v23.