A transient climate change simulation with greenhouse gas and aerosol forcing: experimental design and comparison with the instrumental record for the twentieth century

The Canadian Centre for Climate Modelling and Analysis (CCCma) global coupl ed model is used to investigate the potential climate effects of increasing greenhouse gas (GHG)1 concentrations and changes in sulfate aerosol loadin gs. The forcing scenario adopted closely resembles that of Mitchell et al. for both the greenhouse gas and aerosol components. Its implementation in t he model and the resulting changes in forcing are described. Five simulatio ns of 200 years in length, nominally for the years 1900 to 2100, are availa ble for analysis. They consist of a control simulation without change in fo rcing, three independent simulations with the same greenhouse gas and aeros ol changes, and a single simulation with greenhouse gas only forcing. Simul ations of the evolution of temperature and precipitation from 1900 to the p resent are compared with available observations. Temperature and precipitat ion are primary climate variables with reasonable temporal and spatial cove rage in the observational record for the period. The simulation of potentia l climate change from the present to the end of the twenty-first century, b ased on projected GHG and aerosol forcing changes, is discussed in a compan ion paper. For the historical period dealt with here, the GHG and aerosol f orcing has changed relatively little compared to the forcing changes projec ted to the end of the twenty-first century. Nevertheless, the forced climat e signal for temperature in the model is reasonably consistent with the obs erved global mean temperature from the instrumental record. This is true al so for the trend in zonally averaged temperature as a function of latitude and for some aspects of the geographical and regional distributions of temp erature. Despite the modest change in overall forcing, the difference betwe en GHG + aerosol and GHG-only forcing is discernible in the temperature res ponse for this period. Changes in precipitation, on the other hand, are muc h less evident in both the instrumental and simulated record. There is an a pparent increasing trend in average precipitation in both the observations and the model results over that part of the land for which observations are available. Regional and geographical changes and trends (which are less af fected by sampling considerations), if they exist, are masked by the large natural variability of precipitation in both model and observations.