U-Th-Ra fractionation during crustal-level andesite formation at Ruapehu volcano, New Zealand

Quantifying the effects of crustal interaction on U-series disequilibria is critical if such data are to be used to obtain a correct interpretation of the time scales of magmatic processes. The 1945-1996 eruptions of Mount Ruapehu produced lavas that provide an exceptional, temporally controlled suite with which to examine U-series systematics in circumstances where crustal contamination is independently documented. Co-variations between SiO2 abundances, Th contents and Sr-87/(86) Sr ratios are consistent with crystal fractionation of a relatively primitive magma coupled with assimilation of a silicic crustal melt (AFC) and have been used to develop a quantitative model in which r (ratio of assimilant to melt fraction)=0.2. (Th-230/U-238) ratios increase from similar to 0.87 towards secular equilibrium with increasing Sr-87/ Sr-86, suggesting that the silicic contaminant was in U-Th secular equilibrium, as is observed, for example for Taupo Volcanic Zone (TVZ) rhyolites. The AFC model provides a good approximation of the (Th-230/U-238)-Sr-87/Sr-86 co-variation which suggests that AFC is likely to have taken place relatively rapidly, compared to the half-life of Th-230. (Ra-226/Th-230) varies from 1.19 to near secular equilibrium suggesting that the time required to produce the complete range of compositions erupted between 1945 and 1996 is of the order of a few thousand years rather than either very long (tens or hundreds of thousands of years) or very short (tens or hundreds of years) time intervals. However, using recent estimates of the rate of fractionation (4 x 10(-4) yr(-1)), the range in ((226) Ra/Th-230) cannot be accounted for by decay during fractionation even when the effects of assimilation of a secular equilibrium crustal melt are accounted for. Instead, AFC models with r= 0.2 which can replicate the data require that the rate of fractionation is an order of magnitude slower (1-5 x 10 (-5) yr(-1)). This could be explained by protracted cooling of magma in crust with an elevated geotherm consistent with the high surface heat flow in the TVZ to the north of Ruapehu. U-series disequilibria have become more variable with time suggesting either an increased role for open system magma replenishment in the last 50 years or that processes of crustal melting have become more variable or that the crustal section involved in assimilation is heterogeneous. The trend towards more Ra-Th variability is matched by decreasing Sr isotope ratios, implying progressive involvement of more mafic magma with elapsed time. (C) 2007 Elsevier B.V All rights reserved.