All Climate Events

General Bayesian Learning in Dynamic Stochastic Models: Estimating the Value of Science Policy

Download as iCal file

Thursday, 12 April 2018, 2:00

Thursday, April 12, 2018. 2:00PM. General Bayesian Learning in Dynamic Stochastic Models: Estimating the Value of Science Policy. Derek Lemoine, University of Arizona. Sponsored by Geophysical Fluid Dynamics Laboratory. More information here.


[They] integrate climate scientists into an economic model of climate change by extending recursive dynamic programming methods to allow for nonconjugate learning about an uncertain parameter. This nonconjugate learning enables us to calibrate a statistical model for updating beliefs about the climate's sensitivity to greenhouse gas emissions to the actual history of scientific progress. Improvements in scientific monitoring and modeling of the climate system are valuable because the policymaker can better adapt emission policy to the true state of the climate system. We find that such improvements can be worth up to $600 billion per year and can double the welfare gain from adopting optimal emission policy.

Location  NOAA GFDL, Smagorinsky Seminar Room, Princeton, NJ.