Sub-Seasonal Prediction with CCSM4

There is a growing interest in a wide swath of user communities for forecast information on time scales beyond 10 days but less than a season. Moreover, there is compelling evidence that leveraging multiple modeling groups and centers to produce a multi-model ensemble is a practical approach to both increase the overall ensemble size and to represent at least some aspects of the forecast uncertainty due to model uncertainty. This proposal seeks to participate in a coordinated multi-model sub-seasonal effort. Specifically, the proposed work described here leverages the existing seasonal North American Multi-Model (NMME) effort at the University of Miami:

I. To produce retrospective CCSM4 sub-seasonal forecasts following an already agreed upon uniform protocol;
II. To participate in a 1-year experimental multi-model real-time CCSM4 prediction effort including the on-time delivery of real-time forecasts and the retrospective forecasts for calibration;
III. To participate in the coordination of research and rapid data sharing to support both research and real-time prediction needs.

In addition to the production of the retrospective and real-time forecasts, we will demonstrate the skill of CCSM4 in predicting sub-seasonal phenomena as sources of predictability, including MJO, blocking, NAO, sub-seasonal variability of ENSO and their impact on various hazards (e.g. precipitation, tropical cyclones, heat waves/cold spells). We will develop tools for performing quality control checks on CCSM4 forecasts based on the re-forecasts and apply them to the real-time predictions. We will also document the relative contributions of land surface vs. atmospheric initialization in forecast quality, diagnose how well the forecasts capture the interactions of the MJO and NAO, and how this impacts forecast quality of rainfall variability over the US. Moreover, we will examine how strong sub-seasonal drying events during spring along with land surface-cloud-precipitation feedback serve to initiate droughts in the US Great Plains. This later effort will leverage existing collaborations with Texas water managers to improve the use of sub-seasonal forecast information in decision support. The proposed research will be carried out as part of the CIMAS program, and addresses the CIMAS climate research and impacts theme in that the objective include improved sub-seasonal predictions using multi-model ensembles that directly serve NOAA’s goal of understanding climate variability and change.