Climate Observing System Simulation Experiments: Reducing the Uncertainty in Climate Uncertainty

Climate Observing System Simulation Experiments (COSSEs)The climate observing system of the future will need to support three types of goals: monitoring the Earth, advancing models of climate processes and improving climate projections. Each of these areas is expected to drive quantified science hypotheses and goals, and would in most cases be relevant for COSSEs to better understand requirements. Tools to evaluate proposed climate observing systems are under development and cover a wide range of approaches; these tools can be used to address the three major climate observing system elements: climate change monitoring, climate process understanding, and climate prediction uncertainty. Efforts have already begun to develop COSSEs for aspects of climate observations including ocean heat content (Argo), carbon cycle sources and sinks (OCO2 and Methane CarbonTracker), temperature trends using radio occultation (COSMIC) and cloud feedbacks (CLARREO/CERES). COSSEs can be used to evaluate many aspects of climate observations, ranging from instrument accuracy requirements (CLARREO) to estimations of retrieval uncertainty (OCO2) and/or sampling uncertainty (CALIPSO, OCO2, COSMIC, CLARREO). Development of COSSE capabilities is not trivial.Some aspects of COSSEs are highly model dependent, particularly when COSSEs are used in a method of reanalysis or when global climate models are used to identify sensitivities to specific parameters. For these reasons, model independent COSSE efforts will be used as often as possible to isolate the observational capabilities as appropriate