资源环境科技发展态势分析平台

The accuracy of radiation calculations matters for climate simulations. Radiation code assessments typically validate such quantities as radiative fluxes and heating rates. However, it is not clear how these quantities or their uncertainties affect climate sensitivity-the extent of warming driven by the CO2 radiative forcing. Here, we assess the temperature response uncertainty by comparing simulations based on parameterized radiation codes to that based on a benchmark line-by-line model, in an idealized global warming (quadrupling CO2) experiment. We present an idea to quantitatively relate the radiative quantities to equilibrium temperature response. We find that when the temperature change is solely driven by the radiative process (a pure radiative adjustment), the temperature change and its uncertainty can be well diagnosed by the proposed method. Under this situation, the temperature response uncertainty is found to result from the uncertainties in both CO2 forcing and the radiative Jacobians. This calls into attention the importance of Jacobians in the radiation code intercomparison and validation. When the temperature change is driven by both radiation and convection (a radiative-convective adjustment), the temperature change can no longer be predicted by a simple diagnostic equation. Nevertheless, the validation against the benchmark simulation provides an estimate of the temperature response errors that may be attributed to radiation code inaccuracy. We find that such errors may reach several 10th degrees Kelvin for surface temperature and more than 1 degrees Kelvin for atmospheric temperatures.