Long-term low latitude cosmogenic 3He production rate determined from a 126 ka basalt from Fogo, Cape Verdes

An ankaramite lava from Fogo, Cape Verdes, has been used to determine the long-term cosmogenic 3He (3Hecos) production rate at low latitude (14.9°N). 40Ar/39Ar incremental heating yield an eruption age of 125.8±7 ka. Cosmogenic 3He abundances yield local production rates of 75.6±7.0 atoms/g/year in olivine (1σ, n=10) and 70.1±2.9 atoms/g/yr in pyroxene (1σ, n=11). The lower 3Hecos production rate in pyroxene is consistent with compositional control on production. When scaled to sea level high latitude using available algorithms, the olivine 3Hecos production rate ranges from 86–109 atoms/g/yr. These values overlap the long-term production rate determined at the Canary Islands (99 atoms/g/yr). They are, however, lower than previously reported long-term production rates from Hawaii and southern South America (129–152 atoms/g/yr), and the range of <20 ka calibration sites (110–130 atoms/g/yr). Erosion-burial and high atmospheric pressure appear to be of negligible importance. The likeliest explanation is that the geomagnetic field strength was higher in the past. The agreement between the Fogo geomagnetic field strength-corrected production rate scaled using various published scaling algorithms suggests that differences in incorporation of the non-dipole component of the Earth magnetic field in the scaling algorithms are negligible for long-term production rate determinations at low elevation and low latitude. Longitudinal variation in long-term cosmogenic nuclide production rates between the central Atlantic Ocean and the Americas (Hawaii and Patagonia) may exist. This has implications for exposure dating of old surfaces. It will require that locally determined production rates should be used where available, and that scaling schemes incorporating longitudinal variations should be used if nearby calibration sites are not available. Furthermore, if long-term longitudinal production rate variations exist, it implies that variation in the geomagnetic field may not average out after around 20 ka, but may require significantly longer periods (up to 100 ka).