Lessons learned from the First Phase of the Continual Intercomparison of Radiation Codes (CIRC)

Results from Phase I of the Continual Intercomparison of Radiation Codes (CIRC) will be presented. CIRC distinguishes itself from previous intercomparisons such as the Intercomparison of Radiation Codes in Climate Models (ICRCCM) and the Radiative Transfer Model Intercomparison Project (RTMIP) by relying on an observationally validated catalog of cases. The CIRC Phase I cases are built around Atmospheric Radiation Measurement (ARM) climate research facility observations chosen on the basis of preset criteria fulfilling the goals of this initial phase of the project, namely to examine radiative transfer (RT) model performance under realistic, but not overly challenging atmospheric and surface states. CIRC Phase I consists of seven cases, five cloud-free, and two with overcast liquid clouds. All input typically needed by a GCM-type radiative transfer algorithm to calculate profiles of radiative fluxes and heating rates, as well as the reference output (“truth”) from line-by-line (LBL) models is openly available at the CIRC website, http://circ.gsfc.nasa.gov. Idealized subcases with spectrally constant surface albedo and without aerosol and/or cloud were also recently added in CIRC's Phase I collection in order to gain better understanding of RT model discrepancies from the LBL truth. In addition to providing an overview of model performance with respect to LBL and of the inter-model differences for the above cases, the presentation will also highlight many interesting aspects of model behaviour with respect to gaseous forcing (doubled CO2), treatment of the spectral surface albedo function, partitioning of total flux into direct and diffuse components, and inclusion of scattering in the thermal infrared. Since CIRC is intended as an evolving and regularly updated reference source for GCM-oriented RT code evaluation with the goal to contribute to further improvements of solar and thermal RT parameterizations, we will also discuss our plans for providing a broader set of observationally-based test conditions in future phases of the effort.