Abstract: Measurements of ambient CO, NO_x and SO₂ made at Gucheng Station, Hebei Province, China from September 2006 to August 2007 are analyzed and applied to the validation of emission inventory data. It shows that the concentrations of these gases are significantly correlated among each other, suggesting that the gases have common sources. The concentration ratios of CO to NO_x, CO to SO₂ and SO₂ to NO_x for the entire day, daytime and nighttime in different seasons are obtained based on the slopes of the respective correlation lines. The results show that the seasonal variations of the ratios are not very high and the correlations of CO and NO_x, CO and SO₂, SO₂ and NO_x in winter are more significant than those in other seasons. There is large difference between the daytime and nighttime in the ratio of CO to NO_x, particularly in the warmer seasons, indicating strong photochemical impact on the ratio of CO to NO_x. There are certain dependences of the gas concentrations and the concentration ratios on wind directions, reflecting the different impacts of sources from different wind sectors. The prevailing surface wind directions at the Gucheng Station are northeast and southwest, facilitating the transport of plumes from Beijing and Baoding, respectively. When winds come from the north sector (Beijing), surface concentrations of CO and NO_x are significantly higher than those from other directions, and when the winds come from south sector (Baoding, Shijiazhuang), surface concentration of SO₂ is significantly higher than that from other directions. The characteristics of pollutants in different wind directions may reflect the characteristics of pollution sources in different directions. The concentration ratios from the observations are compared with emission ratios derived from the emission inventories (INTEX-B, TRACE-P, REAS, HB). To avoid the influences from inadequate vertical mixing and strong photochemistry, only afternoon data in seasons other than summer are considered. Under this condition, the observed concentration ratios are 43.7 and 31.6 for CO to SO₂ and for CO to NO_x, respectively, which are 2—12 times higher than the respective emission ratios derived from the emission inventories. This result suggests that CO emission may be underestimated by more than two times in the emission inventories. Further studies show that CO emission from biomass combustion, especially the large-scale straw combustion in harvest seasons would be the important but greatly underestimated source. The analysis of the observation data indicates that the average CO level in open straw burning periods is (90%±30%) higher than that in the other periods under similar meteorological conditions. So far, biofuels are still used for cooking and heating by rural families in many Chinese regions and open burning of agricultural residues is often not effectively controlled. In the future, the impacts of emissions from agricultural straw burning on the emission strengths should be given more attention in the process of making and using the emission inventories.