Fractures in arid regions, especially within low-permeability rocks, have a significant effect on the hydrological cycle. Recent studies have revealed large amounts of salt precipitates along fracture walls in the upper few meters of surface-exposed fractures (SEFs). However, measured salt-precipitation rates are far too high to be accounted for by diffusive venting of moist fracture air. An alternative mechanism that could explain enhanced evaporation from fractures is thermal convection, which also vents moist air from the fractures. This paper describes a series of experiments conducted in Hele-Shaw chambers aimed at quantifying the potential venting capacity of air convection through natural fractures as a function of thermal gradient and fracture aperture. Using smoke to visualize air flow, we quantified air velocities, mass flux, and convection-cell morphology as a function of Rayleigh number (Ra). Thermal differences of 0, 5, 10, and 13°C along the 50-cm-high Hele-Shaw chambers and apertures of 1 and 2 cm were explored. Velocity was found to increase linearly with Ra while mass transfer increased exponentially with Ra, likely due to the concomitant increase in the number of convection cells per chamber width. For example, a thermal difference of 13°C between ambient air and chamber air for a 2-cm aperture resulted in an increase of approximately two orders of magnitude in the mass-transfer rate relative to pure diffusive flux (i.e., when no thermal gradient is imposed). The air-mass-transfer rate in the Hele-Shaw chamber was found to be linearly proportional to the Sherwood number.