Applying control theory to support the Geoengineering Model Intercomparison Project
Dave Leedal, Lancaster Environment Centre
Wednesday 11 January 2012, 1300-1345
The Geoengineering Model Intercomparison Project is a research group led by Ben Kravitz (Carnegie Institute for Science) with the aim of answering the question "What are the expected climate effects of geoengineering?". The approach the group is taking is to encourage the climate modelling community to take part in a model intercomparison exercise using a set of simple standardised experiments designed to replicate the broad characteristics of a geoengineering intervention in the climate system. The intercomparison should provide insight into where and how the various models agree or diverge on issues such as the predicted decrease in the strength of the African/Asian summer monsoon. Three of the GeoMIP experiment designs specify an adjustment in one model forcing term that must be countered by a geoengineering action in order to maintain a zero net radiative forcing. This is a classic feedback control problem although the GeoMIP organisers had not recognised it as such, instead an ad-hoc iterative approach had been suggested to find the required balancing geoengineering action. We have now designed a recursive control scheme that will identify the required geoengineering action during a single model integration thus reducing the overheads for research groups wishing to take part in GeoMIP. The approach has been adopted by GeoMIP. This provides a new application for control and a new approach for running GCMs. This talk gives a quick introduction to GeoMIP and our contribution to the project.