Translating amounts and rates of rock cooling derived from low-temperature thermochronometry into denudation requires assumptions about the local geothermal gradient. The temperature gradient in the crust depends on many factors, including basal heat flow, crustal heat production, and thermal conductivity. Consequently, geothermal gradients may be variable on time scales over which rock cooling is tracked by thermochronometry. Using one-dimensional numerical modeling of heat transfer in rocks of varying thermal characteristics, we show that the geothermal gradient of the eroded layer is the most important factor for accurate estimation of denudation amounts. Using a three-dimensional numerical model (Pecube), we demonstrate the impact of crustal heat production and thermal conductivity on estimates of total denudation derived from apatite fission track data from central west Britain. We show that the regional variation in cooling ages measured in Caledonian granites can be explained by geothermal gradient variation due to the presence of a heat-producing granite batholith and removal of insulating sedimentary rocks, and does not require variable denudation. Neglecting the blanketing effect leads to twofold overestimation of the amount of denudation. The occurrence of heat-producing basement that was once covered by a sedimentary blanket is common, in particular in the core of mountain belts. Accurate determination of the amount and rate of denudation from thermochronometric studies in these situations must take into account the composition of the eroded rocks.