Jim Stock: Global Temperature and Human Activity: "Two co-authors and I conducted an analysis published in 2006 that used cointegration methods to estimate the relationship between total energy impacting the earth (called radiative forcing) and global mean temperature. Radiative forcing is the sum of solar radiative forcing and radiative forcing from greenhouse gases. The figure plots temperature and its predicted value, using the relationship we estimated using data from 1860-1994 (all the data we had at the time, because of data availability lags). Because our data ended in 1994 (denoted by the vertical line in the figure), the past 20 years of data provide a true out-of-sample test of our regression relationship between radiative forcing and temperature. The red dashed line post-1994 is the predicted temperature, using our published regression estimate and post-1994 data on radiative forcing but not using temperature data for that period (the shaded region shows 67 percent confidence interval for the prediction). The figure illustrates that the published regression does a very good job predicting the path of temperature post 1994. For fifteen years starting in 1998, there was a warming 'hiatus', which has been used to argue against warming being linked to ever-increasing concentrations of carbon dioxide. In our model, the hiatus is due to increased emissions of sulfur oxides (which reflects sunlight back into space) from new coal plants in China, and a long lull in the solar cycle. Since then, temperatures have been rising, just as the regression predicts...

...The regression provides a simple way to estimate how much of the observed warming since 1870 is natural and how much is anthropogenic. To do so, consider the counterfactual in which anthropogenic emissions of greenhouse gases are held at their 1870-1890 average. The predicted value of temperature under this counterfactual is shown by the blue line in the figure; this is the temperature that that would have obtained, absent greenhouse gas emissions due to human activity. The human contribution is the difference between the total predicted value (red dashed line) and the green line that shows the actual temperature. Although natural sources are the major driver of predicted fluctuations in temperature before 1920, subsequently human emissions started to play a major role. Today, nearly all the increase in observed temperature is a result of anthropogenic emissions of greenhouse gases.

The results from our statistical analysis match results from very different modes of analysis based on climate models that include radiative forcing...

Global Temperature and Human Activity Econofact

R. Kaufmann, H. Kauppi, and J. Stock: The Relationship Between Radiative Forcing and Temperature: What Do Statistical Analyses of the Intrumental Temperature Record Measure?