U and Th concentrations and isotope ratios of samples from ODP Leg 166

Constraining the history of seawater (234U/238U) is important because this ratio is used to assess the validity of U/Th ages, and because it provides information about the past rate of physical weathering on the continents. This study makes use of U-rich slope sediments from the Bahamas in an attempt to reconstruct seawater (234U/238U) for the last 800 kyr. For the last 360 kyr, U/Th dating of these sediments provides ages and initial (234U/238U) values. Sixty-seven samples, largely from marine highstands, have initial (234U/238U) which scatter somewhat about the modern seawater value (~1.145) but neither this scatter nor the average value increases with age of sample. These data contrast with published coral data and suggest that seawater (234U/238U) has remained within 15‰ of the modern value for the last 360 kyr. This confirms the rejection of coral U/Th ages where the initial (234U/238U) is significantly different from modern seawater. Data from older highstands, dated with delta18O stratigraphy or by the presence of the Brunhes/Matuyama (B/M) reversal at 780 kyr, allow seawater (234U/238U) to be assessed prior to the range of the 230Th chronometer. Unfortunately, diagenetic scatter in the data between the B/M reversal and 360 kyr is rather large, probably relating to low U concentrations for these samples. But there is no indication of a trend in seawater (234U/238U) with age. High U samples from close to the B/M reversal show less diagenetic scatter and an initial (234U/238U) that averages 1.102. This lower value can be explained by lower seawater (234U/238U) at the time of the B/M reversal, or by progressive loss of 234U from the sediment by alpha-recoil. A simple box model is presented to illustrate the response of seawater (234U/238U) to variations in riverine input, such as might be caused by changes in continental weathering. Comparison of the Bahamas (234U/238U) data with model results indicates that riverine (234U/238U) has not varied by more than 65‰ for any 100 kyr period during the last 360 kyr. It also indicates that the ratio of physical to chemical weathering on the continents has not been higher than at present for any extended period during the last 800 kyr.

约束海水(234U/238U)比值的演化历史具有重要科学意义：该比值既可用于评估铀钍年代（U/Th ages）的有效性，也可为大陆尺度物理风化（physical weathering）的过往速率提供关键约束信息。本研究依托采自巴哈马（Bahamas）地区的富铀斜坡沉积物（slope sediments），尝试重建过去800 kyr间的海水(234U/238U)比值演化序列。
针对过去360 kyr的样品，通过对上述沉积物开展铀钍测年（U/Th dating），可获得其沉积年代与初始(234U/238U)比值。共计67个样品（绝大多数采自海洋高水位期（marine highstands））的初始(234U/238U)比值在现代海水比值（约1.145）附近呈现一定离散，但该离散程度与平均比值均未随样品年代的增加而升高。上述数据与已发表的珊瑚数据集存在显著差异，表明过去360 kyr间海水(234U/238U)比值始终与现代值的偏差不超过15‰。这一结论印证了对初始(234U/238U)比值与现代海水显著不符的珊瑚铀钍年代结果的否定。
通过δ18O地层学（delta18O stratigraphy）或780 kyr处的布容-松山（B/M）极性倒转（Brunhes/Matuyama (B/M) reversal）定年的更早海洋高水位期样品，则可用于评估230Th测年覆盖范围之外的海水(234U/238U)比值。遗憾的是，布容-松山极性倒转至360 kyr之间的数据集存在较为显著的成岩作用离散，该现象可能与这些样品的铀浓度偏低有关。但并未观测到海水(234U/238U)比值随年代演化的显著趋势。采自布容-松山极性倒转附近的高铀样品，其成岩离散程度更低，初始(234U/238U)比值的平均值为1.102。该偏低的比值可通过两种机制解释：一是布容-松山极性倒转时期的海水(234U/238U)比值更低，二是沉积物中的234U通过α反冲（alpha-recoil）作用发生持续性流失。
本文构建了一个简单箱式模型（box model），用以阐释海水(234U/238U)比值对河流输入（riverine input）变化的响应机制，而河流输入变化可由大陆风化（continental weathering）强度改变引发。将巴哈马地区的(234U/238U)比值数据与模型结果对比后发现，过去360 kyr间的任一10万年时段内，河流(234U/238U)比值的变化幅度均未超过65‰。同时该结果表明，过去800 kyr间的任何较长时段内，大陆物理风化与化学风化的强度比值均未高于当前水平。