Based on the heat engine framework, a simple scaling theory for dust d
evils is proposed and compared to observations. This theory provides a
simple physical interpretation for many of the observed characteristi
cs of dust devils. In particular, it predicts the potential intensity
and the diurnal variation of dust devil occurrence. It also predicts t
hat the intensity of dust devils depends on the product of two thermod
ynamic efficiencies, corresponding respectively to vertical and horizo
ntal temperature gradients.