【摘要】：CO_2 and temperature records at Mauna Loa, Hawaii, and other observation stations show that the correlation between CO_2 and temperature is not significant. These stations are located away from big cities, and in various latitudes and hemispheres. But the correlation is significant in global mean data. Over the last five decades, CO_2 has grown at an accelerating rate with no corresponding rise in temperature in the stations. This discrepancy indicates that CO_2 probably is not the driving force of temperature change globally but only locally(mainly in big cities). We suggest that the Earth's atmospheric concentration of CO_2 is too low to drive global temperature change. Our empirical perception of the global warming record is due to the urban heat island effect: temperature rises in areas with rising population density and rising industrial activity. This effect mainly occurs in the areas with high population and intense human activities, and is not representative of global warming. Regions far from cities, such as the Mauna Loa highland, show no evident warming trend. The global monthly mean temperature calculated by record data, widely used by academic researchers, shows R~2=0.765, a high degree of correlation with CO_2. However, the R~2 shows much less significance(mean R~2=0.024) if calculated by each record for 188 selected stations over the world. This test suggests that the inflated high correlation between CO_2 and temperature(mean R~2=0.765-0.024=0.741) used in reports from the Intergovernmental Panel on Climate Change(IPCC) was very likely produced during data correction and processing. This untrue global monthly mean temperature has created a picture: human emission drives global warming.