On the use of simulated photon paths to co‐register top‐of‐atmosphere radiances in EarthCARE radiative closure experiments

The Earth's Cloud, Aerosol and Radiation Explorer ( EarthCARE ) mission will retrieve vertical profiles of cloud and aerosol properties by combining data from active and passive instruments. The verisimilitude of retrievals will be assessed using data from its broad‐band radiometer (BBR), which measures top‐of‐atmosphere (TOA) short‐wave (SW) radiances at three along‐track viewing angles. BBR measurements will be compared with their modelled counterparts, simulated by a three‐dimensional (3D) Monte Carlo (MC) radiative transfer model acting on retrieved properties, thus defining a radiative closure experiment. Since cloud and aerosol microphysical and hence optical properties within each assessment domain vary horizontally and vertically, one challenge facing the closure process is selection of radiances that will foster the best assessments of retrievals. This study investigates whether co‐registration of radiances for closure assessment can be aided by information pertaining to photon paths from the MC model. Unlike methods that provide one effective reflecting layer (ERL), such as cloud‐top altitude, simulated photon paths can account for several reflecting layers. For this study, A‐Train satellite data provided cloud properties. The MC model was applied to this field to simulate BBR‐like measurements. Cloud properties were then perturbed randomly, to represent retrievals in approximate fashion, and the MC model reapplied to them. The resulting sets of radiances mimicked EarthCARE measured and modelled data, thus allowing a test of closure and co‐registration methodologies. Through the use of 3D photon path information, the rate of identification of inaccurate cloud retrievals improved over ERL approaches by ∼4% for cirrus clouds and ∼15% for broken clouds. For large‐scale deep convective clouds, however, inaccurate photon paths, ostensibly due to poor retrievals, reduced identification performance by 3%.