Two field data sets are used to model near-surface soil temperature profiles in a bare soil. It is shown that the commonly used solutions to the heat flow equations by Van Wijk perform well when applied at deeper soil layers, but result in large errors when applied to near surface layers, where more extreme variations in temperature occur. The reason for this is that these approaches do not consider heat sources or sinks below the surface. This paper proposes a new approach for modeling the surface soil temperature profiles from a single observation depth. This approach consists of two parts: 1) modeling an instantaneous ground flux profile based on net radiation and the ground heat flux at 5 cm depth; and 2) use of this ground heat flux profile to extrapolate a single temperature observation to a complete surface temperature profile. The new model is validated under different field and weather conditions showing low RMS errors of 1-3 K for wet to dry conditions. Finally, the proposed model is tested under limitations in input data that are associated with remote sensing applications. It is shown that these limitations result in only small increases in the overall error. This approach may be useful for satellite-based global energy balance applications. Copyright 2008 by the American Geophysical Union.