Accelerating improvements in process-representation in weather and climate models through active community engagement

Abstract:

Weather and climate model development focused along the dimensions of both complexity and increased resolution seeks to improve the representation of key processes and address long-standing biases in comparison with observations. It is important to assess the performance of processes that are introduced into models while ensuring that improvements in models are not simply a cancellation of underlying process-level biases. Now in its second generation, NOAA’s Model Development Task Force (MDTF) Diagnostics Package has sought to engage the academic community in developing process-oriented diagnostics to better understand the simulations and performance of Earth system models developed by NOAA-GFDL and NCAR. Existing diagnostics focus on atmospheric processes and results using NOAA-GFDL’s CMIP6-generation simulations will be presented. Diagnostics currently under development further expand the range of weather to climate timescales, include more impact-relevant diagnostics, and expand beyond the atmosphere to include coupled ocean, land, and biogeochemical processes. Recent changes were made to the software framework that drives the analyses and are centered around an open-source development model with significant investments in improving portability, data handling, and documentation. While the field of model diagnostic packages continues to grow throughout the community, the MDTF Diagnostics Package is actively seeking coordination with other modeling centers and agencies to eliminate redundancy while ensuring the software meets its mission-critical objectives for both NOAA and NCAR.