Preliminary comparisons of the typical polarized radiative transfer models: precision and efficiency
Gao Yang, Minzheng Duan, Xingyou Huang
Abstract
Gao Yang, Minzheng Duan, Xingyou Huang
Abstract
Radiative transfer models are key tools in the remote sensing and parameterization of climate radiative forcing, while polarized radiative transfer models can provide more accurate insights into the radiation processes in the earth-atmosphere system. PolRadtran/RT3 (Polarized Radiative Transfer, based on the adding-doubling method), SOSVRT (Vector Radiative Transfer, based on successive order of scattering), and VDISORT (Vector DIScrete Ordinate Radiative Transfer, a polarized version of DISORT based on the inverse of matrix method), are three of the most common radiative transfer models, each with polarization based on different physical principles. A comparison of their accuracy and efficiency reveals that SOSVRT is the most efficient, with the time consumed remaining almost invariable with the increase of stream numbers, but increasing with the optical depth of the layered atmosphere. For example, the time consumed for an optical depth of 1.0 was found to be two times that for an optical depth of 0.5 for the Mie scattering atmosphere. The efficiencies of RT3 and VDISORT in modeling polariza- tion with a large stream number were found to be low. For example, under the Rayleigh scattering atmosphere at 400nm and a stream number of 40, the time consumed was 23 times and 7 times as much as that of SOSVRT, respectively. The computation time for the two models was found not to be sensitive to the optical depth, but increased greatly with the increase in stream number. All three models were found to be of the same order of accuracy, but VDISORT showed a fluctuating result for simula- tions with large streams.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Radiative transfer models are key tools in the remote sensing and parameterization of climate radiative forcing, while polarized radiative transfer models can provide more accurate insights into the radiation processes in the earth-atmosphere system. PolRadtran/RT3 (Polarized Radiative Transfer, based on the adding-doubling method), SOSVRT (Vector Radiative Transfer, based on successive order of scattering), and VDISORT (Vector DIScrete Ordinate Radiative Transfer, a polarized version of DISORT based on the inverse of matrix method), are three of the most common radiative transfer models, each with polarization based on different physical principles. A comparison of their accuracy and efficiency reveals that SOSVRT is the most efficient, with the time consumed remaining almost invariable with the increase of stream numbers, but increasing with the optical depth of the layered atmosphere. For example, the time consumed for an optical depth of 1.0 was found to be two times that for an optical depth of 0.5 for the Mie scattering atmosphere. The efficiencies of RT3 and VDISORT in modeling polariza- tion with a large stream number were found to be low. For example, under the Rayleigh scattering atmosphere at 400nm and a stream number of 40, the time consumed was 23 times and 7 times as much as that of SOSVRT, respectively. The computation time for the two models was found not to be sensitive to the optical depth, but increased greatly with the increase in stream number. All three models were found to be of the same order of accuracy, but VDISORT showed a fluctuating result for simula- tions with large streams.
Key concepts: Radiative transfer, Atmospheric radiative transfer codes, Scattering, Atmosphere (unit), Polarization (electrochemistry), Rayleigh scattering, Radiative forcing, Computational physics