A new Poissonian algorithm for the determination of fission-track ages

Abstract

Fission-track (FT) thermochronology has been widely used since the 1970s to constrain the low-temperature thermal history of rocks in numerous geological settings. Statistical problems associated with this method have been addressed since the 1980s. Today, with the improvements brought to the technique, new questions arise which require the adaptation of previously established statistical methods. Samples used for FT thermochronology are typically composed of minerals (e.g. apatite, zircon, titanite, etc.) with a wide variability in uranium (U) concentration. In this work we propose a new statistical methodology, called p-partition, that even assuming grains of a same sample could have different uranium concentrations. We apply our methodology to fission-track ages standards, in-situ and detrital samples. For standards we found that the new estimated age is closer to the independently determined K–Ar age, for detrital samples we applied our methodology to existent and published data for the case of the Venezuelan Andes. Our results were compared with other decomposition methods for the fission-track ages distributions, they show robustness, and in some particular case our method has more resolution in comparison with conventional methods. In this work we present a methodology and algorithms that attempt to solve mathematically the problem of decomposition of data with Poisson mixtures.