I. Advocate Operational R&D Forecast Needs
Promote science-related forecast challenges identified by our operational forecasters and end users, and invite our partners in the research community to work together on improvement of national climate prediction services.
II. Carry Forward Scientific Integrity in Service Development for Decision Support
In climate decision support, mitigation is about climate trends, and adaptation (to the impacts of future climate conditions) is about climate variability, which is place-based. The variability is changing when globe is warming. The regional adaptation requires prediction information about the variability of modes and the stochastic component of climate. Large uncertainties and model problems make adaptation more difficult. With decreasing spatial scale, the year-to-year variability becomes large compared to the long-term trends. It has been a challenge to attribute anthropogenic causes to the local climate variation. Besides, inadequate climate observing system and no reliable downscaling would add additional service limitations.
NOAA Scientific Integrity Policy - NAO 202-73D.2
III. Advocate Promising S&T Development in Research That Has Great Potential to Improve Operation
Neural Net (NN) stemming from the character of human brain activities is designed to make nonlinear mapping without any presuppositions. It attracts more interests accompanied by data explosion and rapid advancement in parallel computing capabilities. Recent in-depth development of NN applications to assist conventional physics-based model outcomes for operational S2S climate prediction shows promising improvement in products and services.
2. Integration of Systems Engineering into Weather-Climate Model Optimization
The Orthogonal Array Test (OAT), a systems engineering approach of fractional factorial design, is widely used in industrial and agricultural production and proven to be effective to deal with multiple factors, levels and interactions with reliability and sensitivity analysis. It has been very successful in system configuration, parameter level selection and tolerance design etc.
