In the course of catalytic combustion of hydrogen (1-5% H-2 in air) in monolith reactors, strongly localized stationary and traveling hot spots arise in response to a sudden and persistent rise of gas flow velocity. Such hot spots may occur, e.g. in a catalytic converter following the acceleration of a car or in a catalytic combustor as a result of a load increase. This phenomenon is illustrated by simulations using a two-phase reactor model. The temperature overshoot of the adiabatic limit is typically of the order of the adiabatic temperature rise itself. The following mechanism underlies this behavior. Light fuel is supplied to the catalytic wall by fast diffusion (in the direction perpendicular to flow), while the heat released by reaction is removed from the wall by the slower, mixture-averaged heat conduction. This leads to accumulation of heat at the catalytic surface that eventually saturates at high temperatures. The hot spots may exhibit intricate dynamics, propagating downstream or upstream, or they may remain stationary. The direction of propagation depends on the relative strength of convective downstream and conductive upstream contributions to the overall displacement of reaction fronts. Generally, the hot spot tends to drift downstream at low flow velocities, remain stationary at intermediate flow velocities, and drift upstream at high flow velocities.