Subseasonal Experiment (SubX): Do the Research-Center Models Improve the Skill of the Operational-Center Models for Multimodel-Mean Forecasting of Weeks 3 and 4?

The SubX project is a multimodel mean (MME) experiment designed around operational requirements with the goal of improving subseasonal forecasts, such as NOAA's Climate Prediction Center (CPC) Week-34 outlooks. The 7 participating global models are a combination of models from operational centers and models from research centers. An open question is whether or not a research-only model adds value, or skill, to a model from an operational center. In this work, a thorough examination of this question is addressed by verifying an operational-only MME and then comparing those skill scores to the same MME but with the addition of a research-only model. The SubX reforecast database will be used for the verification. Metrics for this comparison include the Heidke Skill Score and the Sign Test-Random Walk.