Layertran

Layertran is a plane-parallel and spherical-shell Monte-Carlo model intended as a reference code for radiative transfer models, or for high-accuracy modeling and optical closure work.
Layertran is currently in beta-test for commericial release, please request a quote for details.
Layertran is an optimised Monte-Carlo code for scalar and vectorial radiative transfer in layered media. It can be set up as a coupled ocean-atmosphere system, as it allows for refractive and reflective non-flat interfaces between layers (e.g. the ocean surface). Layertran can be configured to run in plane-parallel or spherical-shell geometries.
Layertran has been in use for over 10 years and was originally designed as a reference code for comparing other radiative transfer models to, but it is efficient enough for general use and building look-up tables for atmospheric correction and other purposes. Outputs automatically include both the mean and the variance due to Monte-Carlo noise. The philosophy behind Layertran is simplicity and transparency of design, implemented as a straightforward forward-radiative transfer model without any approximate variance reduction techniques, or complicated algorithms that obscure the behavior. Instead, perfomance relies on efficient software engineering. The design focus is that correct functioning should be easily verified, and this is supported by software design best practices such as an exhaustive suite of unit tests.
A useful aspect of Layertran is that the outputs can be compartmentalised according to the number and nature of the scattering events that have occurred. For example, it is possible to quantify the radiance paths that include scattering from aerosols versus those that only include Rayleigh scattering, or the independent contribution of reflections from the water surface can be isolated. In a marine remote sensing context this makes the code particularly useful for developing atmospheric correction schemes. Layertran is coded in c++ and is modular in design, so additional components can be added if the available functions are not sufficient. Commercial licences include access to the full source code.
Layertran has been rigorously compared to other Monte-Carlo models and invariant-imbedded solutions for both scalar and vectorial radiative transfer. Layertran has also been compared to empirical data from polarised radiometry images of sky radiance and sub-surface oceanic light fields. See the paper cited below for details.
Layertran is currently in beta-test as a commercial product. Interested parties are recommended to get in touch to find out the current status.