Weather-Type Based Cross-Timescale Diagnostics of CMIP6-Era Models

The PIs proposed to implement the WT cluster analysis to develop a metric system for evaluating model bias or fidelity in the weather regimes and the related dry and wet extremes and identifying the source thereof. The proposed work is founded on the strong track record of the team, with a clear showing of evidence of the functionality of the analysis tool. The unique strength of the proposal is the combination of the CMIP data sets with perturbed-physics experiments with GFDL model, constituting an attribution chain from physics parameters to weather type regimes and to extremes. The deliverables consist of not only scientific publications, but also an open-source software package and online diagnostic atlas hosted at the IRI Data Library. The weather-oriented effort can be a merit to NOAA's mission. Although the weather type analysis is not entirely new, procrustes analysis (as illustrated in Figure 2) seems to be an informative way to characterize the shape and orientation of the weather types. More importantly, the likely outcome will also include actionable items on how to refine the physical parameters, especially those in convection and clouds, for better representation of the weather types and the associated regimes. The work leverages the existent physics-perturbation runs at GFDL, this makes the otherwise ambitious task achievable within the proposed budget. The panel finds the PIs are highly qualified for the proposed work and the requested amount are well justified.