Graduate Student / Postdoc Seminar

Ensemble Estimation Algorithms for Weather and Climate Models

Speaker: Lauren Padilla, Princeton University

Location: Warren Weaver Hall 1302

Date: Friday, November 5, 2010, 1:30 p.m.


Observational data may be used to improve estimates of the parameters and state of weather and climate models of varying complexity. High dimensional, chaotic, multiscale models parameterize small-scale processes for computational feasibility. Simple, globally-averaged models capture climatic trends but lack detailed physics and require coarse parameterization. Model errors due to parameterization lead to large uncertainties in model forecasts. Our work uses a class of ensemble filters, known as sigma-point Kalman filters, to correct or constrain model parameters and output with observations. I will present an application of the filter to estimation of parameters in a simple model of the global energy balance and introduce a modification to the filter algorithm for application to large, nonlinear systems.