The effects of anthropogenic forcing on the coupled climate system is a topic of great current interest. The first efforts along these lines have, naturally enough, been in terms of the widely-known observed record of global surface temperatures. As coupled models increase in sophistication it becomes useful to create more detailed and robust measures of the climate system's response. One such measure is the total global heat content of the world oceans, which (due to its integrative nature) is both sensitive and low-noise. State-of-the-art simulations using the parallel coupled model (PCM) show an increase in global ocean heat content very close to that observed over the past 50 years. Additional diagnostics of the surface heat budget show that this increase comes about as one would expect for increased atmospheric CO2, i.e., by an increased downward longwave heat flux from the atmosphere to the oceans that is fairly uniform globally, and drives a contemperaneous increase in ocean heat content. The relatively good observed record over North America also allows comparison of observed and model EOFs of variability in that region for surface temperature and precipitation. Additionally, we describe new model diagnostic techniques specifically targeted to the climate change problem, whereby the model directly calculates as it is running the probability density functions (PDFs) of variables, such as precipitation, whose PDFs must otherwise be obtained by saving at subdaily intervals and postprocessing. Having the model direclty accumulate the PDFs results in a substantial storage savings, making examination of PDF changes much easier and more practical.