A Reexamination of the Radiative Balance of the Stratosphere

Abstract

Previous diagnostic calculations of the stratospheric radiation budget using observed temperature and absorber distributions produce net heating rates that, although qualitatively similar in their overall patterns, differ quantitatively from each other. Furthermore, when horizontally averaged over the globe, most heating rates reveal significant departures from radiative equilibrium. It is shown that globally averaged infrared cooling and solar heating should theoretically be in balance to within 0.03 K day?1 throughout the stratosphere over monthly means and to within smaller ranges over longer time periods. Such accuracies cannot be attained with current methods and available data. Since it is shown here that distributions of important chemical tracers are sensitive to diabatic transport differences larger than 0.1 K day?1 in the lower stratosphere, global radiative imbalances should at least be kept to within 0.1 K day?1. This last, less ambitious goal appears to be almost achievable using current technology.A comprehensive radiative transfer algorithm has been constructed based on accurate and efficient methods for use in coupled stratospheric models of chemistry, dynamics, and radiative transfer. The individual components of our code are validated here against available line-by-line calculations. Compared to line-by-line calculations, the total IR heating has an accuracy of 10% in the stratosphere and the errors are less than 0.07 K day?1 in the lower stratosphere. Our result happens to have radiative heating and cooling rates that are globally balanced to about 0.1 K day?1 in the lower stratosphere. Furthermore, the net heating rate in the tropical lower stratosphere, which controls the strength of the important Brewer?Dobson circulation, is found to be weaker than previously thought, with important implications for the global distribution of chemical species. In an attempt to explain the differences among existing models, a set of nine case runs is performed with different input datasets of temperature and ozone, with various degrees of degradation of accurate methods and physical parameterizations, and with different numerical implementations. Although it is difficult to perform a true intercomparison of existing models based solely on published material, one finds, based on experiments conducted using our model, that some commonly adopted approximations in IR schemes and physical parameterizations tend generally to increase the tropical net heating in the lower stratosphere by over a factor of 2 and also to significantly increase the global radiative imbalance at other heights.