Two-dimensional analytic weighting functions for limb scattering. (October 2017)

• Record Type: Journal Article
• Authors: Zawada, D.J.; Bourassa, A.E.; Degenstein, D.A.
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 200(2017)
• Page Start: 125
• Page End: 136
• Publication Date: 2017-10
• Subjects: Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2017.06.008
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: A method to analytically approximate weighting functions where changes in the atmosphere occur in both altitude and the along line of sight direction is presented. The method is implemented inside the radiative transfer model SASKTRAN-HR. The analytic method is shown to be orders of magnitude faster than traditional perturbation methods. The analytic method performs better than a single scatter approximation when the atmosphere is optically thick and in backscatter conditions. Abstract: Through the inversion of limb scatter measurements it is possible to obtain vertical profiles of trace species in the atmosphere. Many of these inversion methods require what is often referred to as weighting functions, or derivatives of the radiance with respect to concentrations of trace species in the atmosphere. Several radiative transfer models have implemented analytic methods to calculate weighting functions, alleviating the computational burden of traditional numerical perturbation methods. Here we describe the implementation of analytic two-dimensional weighting functions, where derivatives are calculated relative to atmospheric constituents in a two-dimensional grid of altitude and angle along the line of sight direction, in the SASKTRAN-HR radiative transfer model. Two-dimensional weighting functions are required for two-dimensional inversions of limb scatter measurements. Examples are presented where the analytic two-dimensional weighting functions are calculated with an underlying one-dimensional atmosphere. It is shown that the analytic weighting functions are more accurate than ones calculated with a single scatter approximation, and are orders of magnitude faster than a typical perturbation method. Evidence is presented that weighting functions for stratospheric aerosols calculated under a single scatter approximation may not be suitable for use in retrieval algorithms under solar backscatter conditions.