Overview

The goal of the ali-retrieval package is to implement a full end-to-end simulator that test the performance of the proposed instrument at all levels, from raw level 0 measurements to level 2 retrieval products. The end-to-end simulator is composed of four primary components:

  1. Atmosphere The “true” atmospheric state, used as input to the simulation. The quality of the end-to-end simulation relies on an accurate and fully determined atmosphere.

  2. Radiative Transfer Model The RTM solves the radiative transfer equation given the input atmosphere to produce radiance at the front aperture of the instrument.

  3. Instrument Model Simulate the optics, electronics and processing of an instrument to convert radiance into the level 1 measurement.

  4. Retrieval Inverse solver to retrieve the atmospheric state from the level 1 measurements. Comparison of the retrieved state with the ``true’’ atmosphere shows error in the retrieval due to measurement noise, retrieval assumptions, and simulation errors.

Atmosphere

To fascilate accurate simulations, various atmospheric scenes are implemented including ice clouds, water vapour and aerosol:

  1. CALIPSO-OMPS

  2. EarthCARE

  3. GEM-MACH

Radiative Transfer Model

The ali-retrieval package relies on SASKTRAN for the radiative transfer modelling. It is a line-by-line model excellent for modelling of limb measurements with both 1, 2 and 3D atmospheres.

Instrument

Builds on the skcomponents package to compose optical components, sensors and electronics. The instrument model takes the input radiance at the front aperture and using the instrument location and orientation (geometry) produces a level 1 product. This is done by first transforming the radiance through the instrument optics to account for transmission and polarization responses. The radiance then impinges on a sensor which integrates the radiance over solid angle, wavelength, and time. The signal from the sensor then is passed through the electronics to account for noise and conversion to digital units to produce the level 0 product. The level 0 data is then calibrated to produce the level 1 measurement.

Retrieval

Builds on the skretrieval package to implement particle size and extinction retrievals.