TODO

Puzzles


Alternative formulations
  - Extra term to enable better fit to phase function in backscatter direction
  - Option to pass asymmetry factor too
  - Option in fitting to set narrow lobe to zero

Interface
  - C++ interface
  - Fortran-90 interface
  - Easier extraction of additional information

Features
  - Best diffuse flux profile modelling, e.g. cubic
  - Gas absortion for near IR
  - Add appropriate ice shape and test

Internals
  - More comments

Testing
  - Aerosol benchmark
  - Check automatic differentiation
  - Check "classic" tssf method
  - Check in CAPTIVATE

Optimization

Completed
  - Why for low Rayleigh optical depths, does diffuse contribution to
    flux (from integrating over radiance distribution) appear to be
    too low by a factor of two, when comparing to two-stream flux
    (which matches DISORT)? - DONE
  * Check Rayleigh scattering vs DISORT - DONE
  * Rayleigh scattering model - DONE
  - Check polar plot - DONE
  * Computing asymmetry factor properly from component shapes - DONE
  - Fitting of phase function components - DONE 
  - Removing of delta lobe and smoothing - DONE
  - Smoothing according to a specified lobe shape - DONE
  - AD on surface albedo - DONE
  - reflectance template instantiated for true & false arguments - DONE
  - Better surface albedo model - DONE
  - C++ object for ocean look-up table - DONE
  - C interface to ocean look-up table - DONE
  - Store exponentials on first use - DONE