An analytical expression, derived in this paper, for the steady state thermal structure of thrust belts and subduction systems shows that surface erosion rate, basal accretion rate, and upper plate heat production are extremely important in controlling the temperature and metamorphic history of orogenic belts and accretionary prisms, especially where upper plate thickness (h) is large. For example, even at rapid erosion and accretion rates and with large upper plate heat productions, temperatures near the toe of a thrust wedge (e.g., at h=5 km) are nearly the same as those that would be present in the absence of erosion, accretion and upper plate heat production; in thicker parts of the same wedge (e.g. at h=30 km) the combined effects of erosion, accretion upper plate heat production can be dramatic. For sufficiently large magnitudes of upper plate heat production or sufficiently rapid erosion rate inverted (negative) geothermal gradients are produced in the lower part of the upper plate and the upper part of the lower plate, although it is difficult to produce inverted geotherms where h≤20 km. The potential importance of erosion and accretion can be illustrated through two examples: (1) beneath the Himalayas, increasing the upper plate radiogenic heat production from 0 to 2.5 mW/m3 and increasing the rate of surface erosion from 0 to 1 mm/yr raises the computed steady state temperatures on the Main Central Thrust, at h=30 km, from 265¿C to 600¿C, and produces a local temperature maximum of 635¿C located 6 km above the fault; and (2) within the upper plate of the Peru trench, increasing the rate of surface erosion and basal accretion from 0 to 2.4 mm/yr (or increasing the upper plate radiogenic heat production from 0 to 1.3 mW/m3) increases the computed steady state surface heat flow at h=30 km from 10 to 33 mW/m2, but only increases temperatures at the fault surface from 125¿C to 170¿C.

Shear heating along thrust faults or subduction boundaries is not necessary to produce the high metamorphic temperatures or surface heat flows observed in these two examples, and the generation of large amounts of heat by shear stresses on the faults (more than a few tens of bars) appears to be inconsistent with some of the observations made in these areas. This study indicates that many of the major discrepancies between existing thermal models of overthust belts and observed temperatures, as inferred from field data, can be attributed to neglect of horizontal and vertical advection of heat caused by surface erosion and by accretion of material from lower plate to upper plate across the fault surface. ¿American Geophysical Union 1993