Can satellite data be used to constrain climate change feedbacks in GCMs?

Climate feedbacks associated with the radiative response to changes in water vapor, temperature, surface albedo and clouds are important to understanding the response of our climate to anthropogenic greenhouse gas increases. In spite of many improvements in recent years, climate models still exhibit a significant spread in feedbacks, contributing uncertainty to estimates of global warming. Comparison with observations might help narrow down this uncertainty. However, current observational data sets of suitable variables are not of sufficient length to allow for the computation of long-term climate change feedbacks. Shorter time scale proxies need to be explored. The seasonal cycle, as well as El Niño variability, have been used by different groups to compute individual feedbacks. How such short time scale feedbacks compare with long-term climate change feedbacks remains an important question. Here we use model simulations from the Coupled Model Intercomparison Project (CMIP) archive and satellite data including Atmospheric Infrared Sounder (AIRS) and Clouds and Earth's Radiant Energy System (CERES) products to investigate whether magnitudes of individual feedbacks are comparable across time scales and how short term model feedbacks compare with those obtained from observations. Using the radiative kernel technique, we compare feedbacks from CMIP Community Climate System Model (CCSM) simulations for CO₂ doubling and the seasonal cycle. We then compare the short term feedbacks to those computed from the available satellite data record. While we find good agreement between the seasonal cycle feedbacks in CCSM and satellite observations, the relationship between global warming and seasonal cycle feedbacks in CCSM requires further investigation. If robust time scale relationships can be established, they can be used to infer the length of the satellite data record that would be required for the estimate of climate change feedbacks.