On systematic errors in parametrized calculations of longwave radiation transfer

• Published in: Quarterly Journal of the Royal Meteorological Society
• Main Authors: Morcrette, Jean‐Jacques, Fouquart, Yves
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 1985
• Subjects: Subarctic
• Online Access: http://dx.doi.org/10.1002/qj.49711146903; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fqj.49711146903; https://rmets.onlinelibrary.wiley.com/doi/pdf/10.1002/qj.49711146903

Abstract A detailed narrow band model of longwave radiation transfer in a clear‐sky atmosphere is used to study the impact on cooling rate profiles and radiative fluxes of some approximations commonly used in GCM‐type radiation codes. Systematic errors in the results arise when economies are made in the vertical and spectral resolution. A coarse discretization of the vertical profiles of temperature and humidity overestimates atmospheric absorption (by as much as +5 W m −2 ), as does (by as much as +5 W m −2 ) the use of mean values of the temperature and humidity in each model layer. Overestimation of the absorption (up to +7 W m −2 ) also results from the use of the strong line approximation for H 2 O and CO 2 , or of the weak line approximation for O 3 . Decreasing the order of the vertical quadrature used for the vertical integration tends to decrease the absorption. Effects of the water vapour continuum absorption outside the 800‐1250 cm −1 window region cannot be neglected, as this absorption accounts respectively for 21 and 8 W m −2 of the downward radiation at the surface for a tropical and a subarctic winter atmosphere, and decreases the corresponding outgoing radiation at the top of the atmosphere by 5 and 0.3 W m −2 respectively. Neglecting N 2 O and CH 3 decreases absorption by about 5 W m −2 . Restricting CO 2 absorption to the 500‐800 cm −1 interval, and O 3 absorption to the 970‐1110 cm −1 interval further decreases by 5 W m −2 the absorption in the tropical atmosphere. For transmission functions described by statistical models, a widening of the spectral intervals reinforces the influence of the strong lines and leads to an overestimation of the absorption, which is partly compensated by a weaker temperature dependence.