Ana Barros : Down the Predictability Hole - Searching for Metrics to Understand and Replace Physical Parameterizations of Nonlinear Processes in Atmospheric Models

Short-term forecast skill in weather forecasting over the last 15 years has been achieved mostly through data assimilation. Predictive ability however has hit barriers that have not been overcome by increasing computer power and model resolution. Model tuning has come out from hiding, and it is arguably ``trending'' in peer-review over the last 2-3 years. The open question is what (and how) to do next. I will address this question relying on two-decades of research on the representation of moist processes in the atmosphere, specifically targeting the following issues:
1) Evaluating Models to Elucidate the Physics that Matter
2) Detecting and Isolating Sources, Sinks and Barriers of Predictability
3) Meeting the Utility Challenge - Projections vs Predictability

• Category: Nonlinear and Complex Systems
• Duration: 01:29:53
• Date:  November 29, 2016 at 2:55 PM
• Views: 274
• Tags:   seminar, CNCS Seminar