There is a growing demand for reliable high-resolution, coupled-climate information in two communities: the predictions and projections communities. The climate prediction community conducts both basic and applied research on short-range climate predictability that directly benefits operational forecast capabilities, and the climate modeling and projection community focuses primarily on basic research concerning climate variability and long-term climate change. Despite the differences, there are several key parallels between these two research communities—a basic example being that both communities assimilate observational data into comprehensive physical climate or earth system models. The prediction community uses a variety of data assimilation techniques for initializing real-time forecasts and reforecasts, and for producing reanalysis, while the climate modeling and projections community started to adopt data assimilation techniques for basic research and for short-term reforecasts to diagnose model behavior. Both communities are also on the verge of increasing the resolution of the climate models while coupling with many more components of the climate and Earth system. While the predictions community explicitly aims to advance the development of operational products that are of the highest possible value to stakeholders and decision-makers at the weeks-toseasons timescale, the climate modeling community is implicitly involved in generating products that are used in assessments and to inform stakeholders and decision-makers about long-term climate change.
Recognizing the common challenges and capitalizing on the potential synergies, the U.S. Department of Energy (DOE) and National Oceanic and Atmospheric Administration (NOAA) jointly hosted the workshop High-Resolution Coupling and Initialization to Improve Predictability and Predictions in Climate Models. This workshop brought together two groups of scientific experts: one focused on sub-seasonal-to-seasonal (S2S) climate predictions and the other focused on using initialized simulations to identify biases in climate models, such as in the Cloud Associated Parameterization Testbed (CAPT).