1Dipartimento di Fisica, Università di Bologna, Italy
*now at: University of Edinburgh, School of GeoSciences, Edinburgh, UK
Abstract. Measurements taken during the 2003 Pacific THORPEX Observing System Test (P-TOST) by the MODIS Airborne Simulator (MAS), the Scanning High-resolution Interferometer Sounder (S-HIS) and the Cloud Physics Lidar (CPL) are compared to simulations performed with a line-by-line and multiple scattering modeling methodology (LBLMS). Formerly used for infrared hyper-spectral data analysis, LBLMS has been extended to the visible and near infrared with the inclusion of surface bi-directional reflectance properties. A number of scenes are evaluated: two clear scenes, one with nadir geometry and one cross-track encompassing sun glint, and three cloudy scenes, all with nadir geometry.
CPL data is used to estimate the particulate optical depth at 532 nm for the clear and cloudy scenes and cloud upper and lower boundaries. Cloud optical depth is retrieved from S-HIS infrared window radiances, and it agrees with CPL values, to within natural variability. MAS data are simulated convolving high resolution radiances. The paper discusses the results of the comparisons for the clear and cloudy cases. LBLMS clear simulations agree with MAS data to within 20% in the shortwave (SW) and near infrared (NIR) spectrum and within 2 K in the infrared (IR) range. It is shown that cloudy sky simulations using cloud parameters retrieved from IR radiances systematically underestimate the measured radiance in the SW and NIR by nearly 50%, although the IR retrieved optical thickness agree with same measured by CPL.
MODIS radiances measured from Terra are also compared to LBLMS simulations in cloudy conditions, using retrieved cloud optical depth and effective radius from MODIS, to understand the origin for the observed discrepancies. It is shown that the simulations agree, to within natural variability, with measurements in selected MODIS SW bands.
The impact of the assumed particles size distribution and vertical profile of ice content on results is evaluated. Sensitivity is much smaller than differences between measured and simulated radiances in the SW and NIR.
The paper dwells on a possible explanation of these contradictory results, involving the phase function of ice particles in the shortwave.