The sources of energy for crustal melting and the geochemistry of heat-producing elements

Abstract

Crustal melting to produce granite magmas requires a tremendous amount of energy. In principle, there are three main mechanisms of heating that can extensively melt a fertile crust: radiogenic heat caused by the decay of 40K, 230Th, 235U and 238U; increased subcrustal heat flux caused by the upwelling of deeper, therefore hotter, mantle materials; advection of heat caused by the emplacement and crystallization of hot mantle magmas. Two-dimensional finite elements modeling reveals that a fertile crust thickened to 65–70km would produce copious granite magmatism after 30–40M.y. if its average heat-production is A>1.2μWm−3, but it would scarcely melt if A<0.65μWm−3. Increasing the subcrustal heat flow from the normal value of QM≈0.025Wm−2 to QM≥0.04Wm−2 may also lead to extensive crustal melting, especially if the crust does not thin to less than 30–35km. Very high QM (≥0.06Wm−2) affecting the continental crust is unlikely, but the combination of moderately high QM (≈0.04Wm−2) and a thick fertile crust with A<1.2μWm−3, such as often happens in the volcanic and back-arc areas of subduction zones, is ideal to produce copious granite magmatism. Lastly, the emplacement of hot mantle magmas in a fertile crust can produce crustal melts in just a few thousand years, but the volume of these is equal to or less than the volume of the intruding magma. A clue for understanding the relative importance of each of these three mechanisms comes from the radiogenic heat production of granite rocks calculated from the concentration of 40K, 230Th, 235U and 238U at the time of their formation. This parameter estimated on more than 3400 granites samples of different ages and provenance reveals a strongly asymmetric distribution peaking around 2.4μWm−3, a value much higher than the average continental crust (about 1–1.2μWm−3) and certainly much higher than the average lower continental crust (about 0.4–0.8μWm−3). Only those granite rock types that are clearly connected with mantle heat sources such as the Archean TTG, post-Archean subduction-related trondhjemites, and recent adakites have a heat production equal to or smaller than the lower continental crust. Since the bulk melt/solid partition coefficient of the heat-producing elements (HPE: K, Th and U) is k≤1, the elevated HPE contents of granites indicates that most of them have been derived from HPE-rich sources. We conclude that radiogenic heating is often essential, and always advantageous, for generating large volumes of granite magmas, and that granite magmatism is the main cause of the accumulation of HPE in the upper crust.