Is there a global warming signal in hemispheric temperature series?

Global and hemispheric temperatures, greenhouse gas concentrations, solar irradiance, and anthropogenic sulfate aerosols all have increased during the last one hundred and fifty years. Classical linear regression techniques will indicate a positive relationship among such series whether or not such a relation exists. Such standard techniques cannot, therefore, show whether observed temperature increases are the result of anthropogenic climate change. However, recent developments in econometrics allow for the analysis of relationships between statistically non-stationary data. We apply some of these recently developed tests in order to uncover the presence of stochastic trends in global climate change variables. These tests indicate that the greenhouse gases are characterized by I(2) stochastic trends while they fail to find evidence of an I(2) stochastic trend in hemispheric temperature series. This would mean that there is no simple long-run equilibrium relationship between radiative forcing and temperature. We then use a multivariate structural time series model to decompose Northern and Southern Hemisphere temperatures into stochastic trends and autoregressive noise processes. This method does not suffer from some of the disadvantages of the standard tests. The results show that there are two independent stochastic trends. The first is I(2) and is shared by the Northern and Southern Hemisphere temperatures. It may be related to the to the radiative forcing variables and represent a global warming signal. The second trend is I(1) and is only present in Northern Hemisphere temperatures. This trend seems closely related to the radiative forcing due to tropospheric sulfates.