Toward a Climate OSSE Framework for Satellite Mission Design

Abstract The rich history of observing system simulation experiments (OSSEs) does not yet include a well-established framework for using climate models. The need for a climate OSSE is triggered by the need to quantify the value of a particular measurement for reducing the uncertainty in climate predictions, which differ from numerical weather predictions in that they depend on future atmospheric composition rather than the current state of the weather. However, both weather and climate modeling communities share a need for motivating major observing system investments. Here, we outline a new framework for climate OSSEs that leverages the use of machine learning to calibrate climate model physics against existing satellite data. We demonstrate its application using NASA’s GISS-E3 model to objectively quantify the value of potential future improvements in spaceborne measurements of Earth’s planetary boundary layer. A mature climate OSSE framework should be able to quantitatively compare the ability of proposed observing system architectures to answer a climate-related question, thus offering added value throughout the mission design process, which is subject to increasingly rapid advances in instrument and satellite technology. Technical considerations include selection of observational benchmarks and climate projection metrics, approaches to pinpoint the sources of model physics uncertainty that dominate uncertainty in projections, and the use of instrument simulators. Community and policy-making considerations include the potential to interface with an established culture of model intercomparison projects and a growing need to economically assess the value-driven efficiency of social spending on Earth observations. Significance Statement When planning a new satellite mission, it is important to first make sure that the new measurements will meet the science and end user goals of the broader community. While there are now well-established ways to quantify observation benefits for weather prediction, there is no such similar established framework for determining satellite measurement benefits for climate prediction. This article describes a new way to determine in advance whether new observations can reduce uncertainty in climate model projections.