Variation in shortwave water vapour continuum and impact on clear-sky shortwave radiative feedback

This work assesses the impact of the current differences in the strength of the shortwave water vapour continuum on clear-sky calculations of shortwave radiative feedback. Three continuum models were used: the MT_CKD (Mlawer–Tobin–Clough–Kneizys–Davies; versions 2.5 and 4.1.1) and CAVIAR_updated (Continuum Absorption at Visible and Infrared Wavelengths and its Atmospheric Relevance) models. Radiative transfer calculations were performed with the ECMWF radiation scheme (“ecRad”). The correlated k-distribution gas-optics tables required for ecRad computations were trained with each of these continuum models using the ECMWF correlated k-distribution software tool. The gas-optics tables trained with the different continuum models were tested individually in the shortwave. The atmosphere configuration was as follows: fixed surface temperatures (TS) between 270 and 330 K, fixed relative humidity at 80 %, a moist adiabatic lapse rate for the tropospheric temperature, and an isothermal stratosphere with the tropopause temperature fixed at 175 K. At TS = 288 K, it was found that the revisions of the MT_CKD model in the shortwave over the last decade and a half have a modest effect (∼ 0.3 %) on the estimated shortwave feedback. Compared to MT_CKD_4.1.1, the stronger CAVIAR_updated model has a relatively greater impact; the shortwave feedback is ∼ 0.006 W m−2 K−1 (∼ 1.6 %) more positive. Thus, for the clear-sky situation investigated, uncertainties due to the shortwave continuum are insignificant for the present-day climate. The uncertainty in the shortwave feedback increases up to 0.008 W m−2 K−1 (∼ 2.0 %) between the MT_CKD models and 0.018 W m−2 K−1 (∼ 4.6 %) between the CAVIAR_updated and MT_CKD_4.1.1 models at TS ≈ 300 K. At TS ≈ 300 K, a large portion of shortwave feedback originates at windows where CAVIAR_updated and MT_CKD_4.1.1 differ substantially, leading to the larger feedback uncertainty.