The application of COSMIC data to global change research

The constellation observing system for meteorology, ionosphere, and climate (COSMIC) is well-suited to climate research, especially as it pertains to climate modeling. It presents a challenge to climate models, which are curr ently tuned to match climate mean states, by providing precisely calibrated data which can be analyzed according to two methods that are insensitive t o standard model tuning. Those two methods are climate signal detection and second-moment statistics, both of which consider the most useful climate m odel to be the one which provides the best predictions rather than the one which best recreates the current climate. In this paper we discuss these tw o new, alternative approaches to improving climate models and how COSMIC oc cultation data can be analyzed in this context. Climate signal detection is usually applied to determine what trends in a c limate data set can be described by external effects, such as increasing gr eenhouse gas concentrations, sulfur dioxide aerosols, etc. Here we show tha t it is actually a method to test climate models. By examining climate tren ds and anomalies as revealed by COSMIC data, we can test whether climate mo dels reproduce those trends and anomalies. We describe in detail how trends and anomalies can be extracted from COSMIC occultation data and the releva nce it should have to climate models. The fluctuation dissipation theorem, as applied to the climate, shows that a second-moment analysis of a climate model's output will reveal more about its physical soundness than does the mean states it produces. While this t heorem shows how a Green's function for climate change can be derived from the second-moments of the climate system, it is best applied by comparing l ike second-moments in data and in model output. This method of testing mode ls is most likely to reveal which parameterizations of convection and moist ure dispersion are most appropriate. The two methods of improving climate models are discussed in the context of COSMIC, showing how the occultation data can be processed to apply each of them.