A comparison of physical climate feedbacks between GCM and reanalysis datasets

In response to a radiative forcing, the Earth's climate system adjusts until reaching a new state of radiative equilibrium. For example, an increased amount of carbon dioxide in the atmosphere will absorb more terrestrial radiation thus decreasing the amount of outgoing longwave radiation (OLR) at the top of the atmosphere (TOA). Since more energy is kept in the system, the global temperature rises and the Earth emits more radiation until the OLR is in equilibrium with the incoming solar radiation. The magnitude of the climate response to an imposed forcing is dependent upon the strength of physical climate feedbacks within the system (i.e., water vapor, lapse rate, albedo, cloud) which act to amplify or dampen the response. Global climate model simulations of the climate response to an imposed forcing vary and that variance is due largely to differences in climate feedback strengths among individual models (IPCC AR4, 2007). One method to quantify individual climate feedbacks in models is the radiative kernel technique (Soden et al., 2008). This technique decomposes each feedback into two components: the TOA flux change due to a standard change in the feedback variable (radiative kernel), and the change in the feedback variable due to a particular climate forcing (climate response).

In this study, the radiative kernel technique is applied to ECMWF ERA-Interim reanalysis data spanning 1989-2008. Monthly departures from the 20-year reanalysis mean of the feedback variables (specific humidity, surface temperature, atmospheric temperature, albedo) are multiplied by the corresponding radiative kernel (Shell et al., 2008) to obtain the TOA radiative flux anomalies due to each variable. Preliminary results suggest an agreement in clear-sky globally averaged net TOA flux anomalies between the reanalysis data set and the kernel calculated flux anomalies of the feedback variables. This agreement suggests the radiative kernel technique can be used to decompose TOA flux anomalies into the contributions by each feedback variable. This is important in order to compare variability in climate feedbacks between reanalysis data and 20th century model runs. Results of this study will help identify areas in which some models perform better than others and directions for model improvement.