Seminar

Abstract

The inability to make direct measurements and the complexity of the processes involved make it difficult to quantify all the components of the Earth's energy budget. Climate models have the potential to greatly inform our understanding of the relevant processes controlling the energy budget. There have been a series of climate model intercomparison projects (CMIP), the latest version of these projects, CMIP5, has contributions from dozens of global climate models. The spread in the output from different models running the same simulations highlights areas of consistency and uncertainty within the models. We use these models to understand how much energy has come into the Earth system since preindustrial times and the fate of that energy, whether it has warmed the oceans or been radiated back to space. Historical anthropogenic climate change is a relatively small signal on the global scale and robustly retrieving this can be difficult. Climate models have errors, biases and drifts that must be accurately accounted for to estimate anthropogenic climate change. Much of the drift can be accounted for by using anomalies from preindustrial control simulations. We have developed a new kernel method that avoids one source of bias in the calculation of radiative forcing. This technique more accurately estimates radiative forcing, especially at long timescales. Despite these adjustments, there is substantial variation between the various CMIP5 models of fundamental quantities such as ocean heat content and top of the atmosphere radiative imbalance. However, much of this variation is around a mean that is consistent with observations. One quantity of interest is the ratio of energy radiated back to space to that absorbed by the ocean. Since 1957, observations suggest that this ratio is 1.48 +/- 0.4. The CMIP5 model mean of this ratio is 1.8, within the uncertainty range of this observational estimate. Other quantities such as the magnitude of surface warming and cumulative forcing of green house gasses are estimated with similar success. With proper adjustments and corrections, the CMIP5 suite of climate models can robustly predict the historical trends of climate change and with less certainty the magnitude as well. The success of these models to predict historical changes lends credence to their predictions of an ever changing future Earth.