Impact of an improved longwave radiation model, RRTM, on the energy budget and thermodynamic properties of the NCAR community climate model, CCM3

The effect of introducing a new longwave radiation parameterization, RRTM, on the energy budget and thermodynamic properties of the National Center for Atmospheric Research (NCAR) community climate model (CCM3) is described. RRTM is a rapid and accurate, correlated k , radiative transfer model that has been developed for the Atmospheric Radiation Measurement (ARM) program to address the ARM objective of improving radiation models in GCMs. Among the important features of RRTM are its connection to radiation measurements through comparison to the extensively validated line‐by‐line radiative transfer model (LBLRTM) and its use of an improved and validated water vapor continuum model. Comparisons between RRTM and the CCM3 longwave (LW) parameterization have been performed for single atmospheric profiles using the CCM3 column radiation model and for two 5‐year simulations using the full CCM3 climate model. RRTM produces a significant enhancement of LW absorption largely due to its more physical and spectrally extensive water vapor continuum model relative to the current CCM3 water continuum treatment. This reduces the clear sky, outgoing longwave radiation over the tropics by 6–9 W m −2 . Downward LW surface fluxes are increased by 8–15 W m −2 at high latitudes and other dry regions. These changes considerably improve known flux biases in CCM3 and other GCMs. At low and midlatitudes, RRTM enhances LW radiative cooling in the upper troposphere by 0.2–0.4 K d −1 and reduces cooling in the lower troposphere by 0.2–0.5 K d −1 . The enhancement of downward surface flux contributes to increasing lower tropospheric and surface temperatures by 1–4 K, especially at high latitudes, which partly compensates documented, CCM3 cold temperature biases in these regions. Experiments were performed with the weather prediction model of the European Center for Medium Range Weather Forecasts (ECMWF), which show that RRTM also impacts temperature on timescales relevant to forecasting applications. RRTM is competitive with the CCM3 LW model in computational expense at 30 layers and with the ECMWF LW model at 60 layers, and it would be relatively faster at higher vertical resolution.