Reflectance simulations using the Optimal Columns, Stochastic Columns and Streams schemes for treating subgrid cloud structure in solar radiance models

This dataset contains the reflectance fields used in the paper "Generating optimal columns to represent cloud structure in solar radiance models" by Robin J. Hogan and Leonhard Scheck. The paper demonstrates that when assimilating solar radiances into weather forecast models, the effect of complex subgrid cloud structure can be well approximated by the use of around four subcoluns with optimally chosen optical depths and area fractions. It compares simulated reflectances using this approach with reference calculations using the "Stochastic Columns" approach with thousands of subcolumns, and an older "Streams" scheme that is found to be less accurate and less efficient than the new Optimal Columns scheme. The reflectances are provided on the unstructured global deterministic 13km ICON model grid for 12 atmospheric snapshots covering the full annual cycle, and were computed using the RTTOV-MFASIS model for those grid cells where the solar and the satellite zenith angles are less than 79 degrees. Each file contains reflectances for one visible channel and one infrared channel, for one particular subgrid scheme. Multiple "Optimal Columns" files are provided corresponding to different numbers of subcolumns and different assumptions on the shape of the column optical depth distribution, along with "Stochastic Columns" files using between 10 and 10000 subcolumns, and two "Streams" files. These data were used to select the Optimal Columns configuration that was recommended in the paper.