Recent observations have shown that low-level clouds have significant impact on snow surface temperature in the Arctic by enhancing downwelling longwave radiative flux at the surface. This study focuses on the detailed interactions among the longwave radiation, clouds, turbulent structure, and snow surface temperature. The approach is to perform sensitivity simulations with a coupled one-dimensional planetary boundary layer (PBL) turbulence closure-snow model. The numerical experiments show that the responses of the snow surface temperature to the cloud longwave radiation may depend on how the clouds are related to the boundary layer turbulent structure. For a cloud-topped boundary layer without upper level clouds, longwave radiative cooling at the cloud top is a main source for the turbulence. In this situation the energetic turbulent eddies effectively transport the radiatively cooled air near the cloud top down to the ground, resulting in a strong sensible heat flux at the surface. When multilayered clouds exist, the radiative cooling at the lower-level cloud top is significantly reduced due to the enhanced downward longwave radiative flux above the cloud, leading to a higher temperature of surface air and weaker positive sensible heat flux at the surface. Consequently, the snow surface temperature is higher in the presence of the multilayered clouds than in cases of only boundary layer clouds. Therefore the boundary layer clouds, dependent on their vertical distribution, may not only increase the snow surface temperature by increasing the downwelling longwave radiative flux but also create negative feedback mechanisms to reduce this temperature increase. A key process in this feedback mechanism is turbulent transport as it directly links the longwave radiative cooling at the cloud top to the surface air temperature. ¿ 2001 American Geophysical Union