Thermochronological dating of detrital samples is an important tool for understanding the thermal history of basins and their source regions. Detrital thermochronology data are often complex with apparent ages ranging over hundreds of million years due to variations in source-rock cooling age and rate, and the influence of post-depositional burial. A common strategy to interpret such data is to split a dataset into several age components by finite-mixture modeling. Herein, we describe for the first time a remarkable pattern of age components in some thermochronological datasets: two or more components showing ages that are approximately multiples of each other. We apply finite-mixture modeling to a log-normal age distribution (a random effects model) and show that these “ghost age components” are artifacts of inappropriately fitting discrete components to continuous age ranges. The striking pattern of the age components is a direct consequence of the underlying age distribution. For a continuous log-normal age distribution, the ghost components have approximately similar age ratios that depend on the overdispersion and the number of components. Due to their artifactual nature, ghost components appear in a range of dating methods, including fission-track, UPb and (UTh)/He dating, and in equivalent doses of luminescence dating. As they are not geologically meaningful, their occurrence hampers the geological interpretation of thermochronological data. We thus propose to carefully decide for the number of fitted components based on grain properties and additional geological information and to screen modeling results for the constant age ratios that are characteristic of ghost components.

中文翻译：

---

碎屑热年代学中的幽灵年龄成分


碎屑样品的热年代学测定是了解盆地及其源区热历史的重要工具。由于烃源岩冷却年龄和速率的变化以及沉积后埋藏的影响，碎屑热年代学数据往往很复杂，表观年龄可达数亿年。解释此类数据的常见策略是通过有限混合建模将数据集分为多个年龄组成部分。在这里，我们首次描述了一些热年代学数据集中年龄成分的显着模式：两个或多个成分显示的年龄大约是彼此的倍数。我们将有限混合模型应用于对数正态年龄分布（随机效应模型），并表明这些“幽灵年龄成分”是离散成分不适当地拟合连续年龄范围的产物。年龄组成的显着模式是潜在年龄分布的直接结果。对于连续对数正态年龄分布，鬼成分具有近似相似的年龄比率，该比率取决于过度分散和成分数量。由于其人工性质，鬼成分出现在一系列测年方法中，包括裂变径迹、UPb 和 (UTh)/He 测年以及等效剂量的发光测年。由于它们没有地质意义，它们的出现妨碍了热年代学数据的地质解释。因此，我们建议根据颗粒特性和附加地质信息仔细决定拟合组件的数量，并筛选作为鬼组件特征的恒定年龄比的建模结果。