Realistic propagation of uncertainties in geologic thermobarometry.
Realistic propagation of uncertainties in geologic thermobarometry.
复制标题
地质温压测量中不确定性的现实传播。
作者：
K. Hodges;L. Mckenna
地质温压测量中两个最重要的不确定性来源是分析的不精确性和与实验校准技术相关的系统误差。分析不确定性是样品特异性的，并决定了P-T估计的精度。校准不确定性是反应特异性的，有效地限制了估计的准确性。我们描述了一个系统的方法，传播这两种类型的不确定性，通过温压计算，以便把现实的置信限P-Z估计。作为一个例子，我们评估的准确性和精度的石榴子石黑云母，石榴子石斜长石蓝晶石石英，和石榴子石金红石蓝晶石钛铁矿石英热压法的泥质样品从东南部加州的葬礼山。校准和分析不确定性共同传播到几百兆帕斯卡和超过100 K的绝对压力和温度不确定性（950/o置信水平）。分析不精确度仅占压力不确定度的10 - 20度，占温度不确定度的300度以下。我们自信地计算地质样本的平衡压力和温度的能力似乎相当有限，但通过额外仔细的实验工作可以显着提高。比较温压法是将一套温压计应用于不同的样品，以计算P-T条件下的差压，它消除了与实验校准有关的系统误差。通过仔细的分析工作，有可能自信地解决P-7的差异，小到几十度和几十兆帕斯卡。
Two of the most significant sources of uncertainty in geologic thermobarometry are analytical imprecision and the systematic error associated with experimental calibration techniques. Analytical uncertainties are sample-specific and dictate the precision of a P-T estimate. Calibration uncertainties are reaction-specific and effectively limit the accuracy of an estimate. We describe a systematic method of propagating both types of uncertainty through thermobarometric calculations in order to place realistic confidence limits on P-Z estimates. As an example, we evaluate the accuracy and precision of garnet-biotite, garnet-plagioclase-kyanite-quartz, and garnet-rutile-kyanite-ilmenite-quartz thermobarometry for a pelitic sample from the Funeral Mountains of southeastern California. Calibration and analytical uncertainties together propagate into absolute pressure and temperature un@r-tainties (950/o confidence level) of several hundred megapascals and more than 100 K. Analytical imprecision accounts for only l0-20o/o of the pressure uncertainty and less than 300/o of the temperature uncertainty. Our capacity to confidently calculate equilibration pressures and temperatures for geologic samples seems rather limited, but it can be improved significantly through additional careful experimental work. Comparative thermo-barometry, which involves applying a single set of thermobarometers to different samples in order to calculate dffirences in P-T conditions, eliminates the systematic error associated with experimental calibrations. Through careful analytical work, it is possible to confidently resolve P-7 differences ofas little as a few tens ofdegrees and a few tens of megapascals.