Spatial Variation in CO2 Concentration Improves the Simulated Surface Air Temperature Increase in the Northern Hemisphere

The increasing concentration of atmospheric CO₂ since the Industrial Revolution has affected surface air temperature. However, the impact of the spatial distribution of atmospheric CO₂ concentration on surface air temperature biases remains highly unclear. By incorporating the spatial distribution of satellite-derived atmospheric CO₂ concentration in the Beijing Normal University Earth System Model, this study investigated the increase in surface air temperature since the Industrial Revolution in the Northern Hemisphere (NH) under historical conditions from 1976–2005. In comparison with the increase in surface temperature simulated using a uniform distribution of CO₂, simulation with a nonuniform distribution of CO₂ produced better agreement with the Climatic Research Unit (CRU) data in the NH under the historical condition relative to the baseline over the period 1901–30. Hemispheric June–July–August (JJA) surface air temperature increased by 1.28°C ± 0.29°C in simulations with a uniform distribution of CO₂, by 1.00°C ± 0.24°C in simulations with a non-uniform distribution of CO₂, and by 0.24°C in the CRU data. The decrease in downward shortwave radiation in the non-uniform CO₂ simulation was primarily attributable to reduced warming in Eurasia, combined with feedbacks resulting from increased leaf area index (LAI) and latent heat fluxes. These effects were more pronounced in the non-uniform CO₂ simulation compared to the uniform CO₂ simulation. Results indicate that consideration of the spatial distribution of CO₂ concentration can reduce the overestimated increase in surface air temperature simulated by Earth system models.