Cointegration analysis of hemispheric temperature relations

Abstract

We use cointegration procedures that are designed to estimate and test relations among integrated time series to develop a model of the relation between surface temperature and the radiative forcing of solar irradiance, greenhouse gases, and tropospheric sulfates. We use this model to test some basic hypotheses regarding the relation between surface temperature and radiative forcing. We find that there is a statistically meaningful relation between surface temperature and changes in the radiative forcing associated with natural variability and human activity. We also find that hemispheric temperatures cannot be explained by hemispheric forcings alone: Hemispheric temperatures are linked. Differences in hemispheric temperatures are associated with differences in the hemispheric temperature effects of greenhouse gases, anthropogenic sulfur emissions, and solar irradiance. Estimates for the temperature sensitivity (ΔT 2 x ) are consistent with the middle and lower range of values estimated by physically based models.