Abstract:

The thermochronology of detrital grains has quickly become a very popular technique for understanding some complicated geological processes, e.g. plate collision, orogeny and sedimentary provenance studies. Detrital single grains from well dated sedimentary formations contain different age components, serving as mineral tracers in sedimentary systems and record the sedimentforming processes. Therefore how to decompose and visualize the ages faithfully and extract possible geological information are keys to the understanding thermochronology result of detrital grains. Binned Frequency Estimation and Probability Density Estimation are two extensively using approaches. Unfortunately, both of them lack a firm theoretical basis and can produce inaccurate results when data quantity or quality is low. This paper introduces a statistical technique for analyzing isotope age, called Kernel Density Estimation (KDE), which involves the age data, but explicitly takes into account the analytical uncertainties by inherent adaptive bandwidth functions. Such adaptive functions vary the bandwidth according to the local density, avoiding the uncertainties from analytical errors and age abundance. Finally, we used the KDE to study two independent thermochronology date sets in northeast China: the zircon UPb ages of Phanerozoic granitoids in Great Xingan Range and the single grain fission track ages of detrital apatite in the Songliao Basin. The estimated age pattern coincides with the regional tectonics and tectonosedimentary records in the study area. It also gives more accurate chronological constraints to some major regional tectonic events. Thus, the KDE proves to be a valid statistical approach for geologic age decomposition.

Key words: detrital grains, thermochronology, Kernel Density Estimation, Songliao Basin, tectonic thermal evolution