Since the mid-19th century, the global atmospheric CO2 concentration (ACC) has increased dramatically due to the burning of fossil fuels. Because of unequal population growth and economic development among regions, the ACC increases possess strong spatial variability. Particularly, the increase in ACC has been larger in the mid-latitudes of the Northern Hemisphere (NH) than that at high- and low-latitudes. It is widely accepted that the ACC increase is the main reason for climate change, but the potential impacts of its spatial distribution on the climate system remain unclear. Therefore, we carried out two groups of 150-year experiments with the Community Earth System Model (CESM), using both spatially inhomogeneous (hereafter the SIC experiment) and homogenous (hereafter the SHC experiment) ACC increases in their settings. We found that the models' divergences occurred over the NH mid-latitudes, the Arctic and the western part of the tropical Pacific. SHC overestimated (underestimated) climate warming over the Artic (mid-latitudes), which may be induced by the intensified westerly and weakened meridional heat exchange between mid- and high latitudes in the NH. Over the tropical Pacific, the overestimation of climate warming may be induced by intensified Walker circulation coupled with the La Niña climate mode. For the entire NH, relative to SIC, SHC overestimated the climate warming from 1850 to 1999 by ~10%. Meanwhile, the SHC experiment also overestimated the interannual variabilities in temperature and precipitation, resulting in more serious extreme events. These findings suggest that human contributions to climate warming and increased extreme events since the industrial revolution may be overestimated when using a spatially homogenous ACC.