Regional Time Trends in Climate Model Simulations

TIES 2007

Paul Switzer

Stanford University

The talk will consist of two parts, both related to climate model simulations. In the first part I examine regional climate time trends as seen in the model simulations. In the second part I examine methods for separating contributions of greenhouse gases from other determinants of climate variation.

I. General circulation models [GCM] are large and complex global geophysical models of the atmosphere and its ocean interface used to represent temporal dynamics of spatially distributed atmospheric variables that are important for climate. These GCMs are used to describe how past, present, and future climate has responded or will respond to increasing atmospheric concentrations of greenhouse gases. GCM outputs are concurrent time series of atmospheric indicators, such as surface temperature and precipitation, at each grid location on the earth’s surface. I will compare these climate simulations for selected geographic regions using dimensionality reduction methods to emphasize the dominant characteristics of modeled climate variation in space and time.

II. GCM climate simulations can be used to explore the separate contributions to regional climate of selected model forcings such as greenhouse gases and solar variability. The objective is to decompose the space-time climate simulations of surface temperature and rainfall in order to estimate component climate responses for individual forcings as well as their interactions. The potential number of statistical parameters is very large so parsimonious constrained representations are useful for characterizing the main features of space-time responses to model forcings.