Opportunities and Challenges in the Analysis of Multi-model Ensemble Output

Claudia Tebaldi, National Center for Atmospheric Research

I will sketch the reasons for the challenging nature of multi-model ensembles and why analysis of their output provides both hurdles to overcome but also opportunities for innovative algorithms and data handling.

After reviewing the main sources of uncertainties and of the odd statistical characteristics of the data samples they generate, I will present a few distinct areas of research where I think the injection of fresh thinking and perspectives would have the chance of creating a big impact. Among those, model tuning; pattern scaling and emulators; statistical/dynamical downscaling method evaluation; and the more fundamental problem of model evaluation vis-a-vis model reliability for future projections (i.e., the search for emergent constraints). 

Link to Recording: http://video.ucar.edu/mms/image/CI2015_claudia_tebaldi.mp4