We present an experimental study of atmospheric gravity currents advancing over a heating ground. The study aims at identifying the processes that characterize this flow, to gain better insight into atmospheric phenomena like the inland penetration of sea-breeze fronts or the evolution of thunderstorm outflows. Particular attention is given to the advancing velocity of the current front, as a function of the nondimensional parameters that characterize the problem. The experimental setup consists of a 4-meter-long channel in which the bottom wall can be heated by resistive fabrics generating a homogeneous heat flux. A sustained dense gravity current, created by a mixture of air and carbon dioxide, is injected at the channel inlet and visualized through laser tomography. The different heating intensities are characterized by a non-dimensional parameter, referred to as B, and defined as the ratio between the vertical buoyancy flux, induced by the imposed heat flux per unit area φ, and the horizontal buoyancy flux that generates the current. We find that, regardless of the Richardson number at the inlet, the advancement velocity is reduced in the heated-ground case compared to the adiabatic experiment (see Figure). The slowdown of the front in the heated case is due to i) the interaction between the current front and the vertical hot plumes created by natural convection and ii) the heating of the current flow, causing a reduction of the density difference between the current and the ambient. Intense heating can cancel out the density differences between the current head and the ambient, which can eventually stop the current and induce the liftoff of the current head. For different initial conditions and heating intensities, the stopping position xstop follows the relation xstop /hs =5/B, where hs is the height of the channel inlet.