Stats4Grads: Emulating an Ocean Ecosystem Model
9 December 2009 14:15 in CM221
Bayesian emulation provides a tool for analysing complex simulators.
When there are many input parameters over a large space, and model runs
are costly, emulation enables us to approximate the simulator across the
space, and gives a measure of our uncertainty at each point.
This talk introduces emulation and then investigates how it can be
applied to HadOCC, the Hadley Centre Ocean Carbon Cycle model. The
eventual goal of the project is to be able to jointly emulate two
simulators of the same system, and this idea will be introduced in the
See the Stats4Grads page for more details about this series.