Recent advances in our understanding of sub-seasonal phenomena and improvements in our ability to simulate sub-seasonal variability along with the demand for sub-seasonal forecasts for decision support suggest that now is the time to develop an experimental multi-model sub-seasonal predictive capability. Indeed, preliminary multi-model sub-seasonal hindcasts conducted by the North American Multi-Model Ensemble (NMME) team indicate that a multi-model prediction system has more overall skill than any single system alone.
The core team proposed here includes the leadership from the seasonal NMME team, and will organize and coordinate the sub-seasonal multi-model experimental reforecasts and real-time predictions that will be tested under separate awards as part of MAPP CTB announcement of opportunity. The coordination efforts will use the lessons learned from the NMME seasonal prediction project to ensure that sub-seasonal hindcasts and real-time forecasts are available to the entire research community in a timely fashion. We will establish robust links with the operational forecasters at the outset of the project so that forecast products developed meet operational needs. We propose to develop rigorous techniques to combine the multi-model predictions into reliable probabilistic forecasts, and to comprehensively evaluate the skill of the forecasts. Particular emphasis will be place on forecast products most relevant for drought prediction. This proposed team is ideal to lead this effort given our leadership the NMME seasonal prediction project and the associated sub-seasonal experiment, and that we led the development of the sub-seasonal prediction protocol that is to be followed in this experiment.
The coordination activities proposed here have four basic elements: (i) Collecting and serving data both internally at CPC for use by operational forecasters and for the external community via the IRI data library largely serving research needs; (ii) providing a baseline verification particularly for the weeks 3-4 temperature and precipitation probability forecasts; (iii) multi-model evaluations and combinations including selecting suitable models, optimizing the design of the system, and evaluation of the prediction products; and (iv) enhancing communications especially between operational forecasts and the forecast producers.