Cumulate Thermobarometry

小安的列斯火山弧为活岛弧俯冲带提供了一个不同寻常的完整和可接近的研究案例。活岛弧俯冲带是由美洲板块至少自始新世以来在加勒比海板块东部边缘下的俯冲形成的。这条弧为在相对简单的海内环境中探索俯冲带过程提供了一个世界级的机会。弧岩浆作用过程发生在深部，必须通过地表的信息推断出来。这类信息的一个主要来源是岩浆岩的性质和组成，这些岩浆岩要么作为熔岩从火山喷发出来，要么在火山管道系统中结晶，被上升的岩浆带到地表，并在其中作为捕虏体喷发出来。熔岩和捕虏体都反映了发生在俯冲板块上方地幔楔中富含挥发物的岩浆熔融和地表熔岩喷发之间的过程。捕虏体的一个重要类别，被称为堆积体，是由在冷却的岩浆中生长的晶体形成的，这些晶体下沉并积聚在岩浆库的底部，从而表明岩浆储存在火山系统的离散位置。这些捕虏体的母体堆积序列位于地壳中一个未知的深度，但它们的晶体组成和捕获的熔体气泡中一定包含了大量编码的信息。这些信息包括关于岩浆来源的线索，这些岩浆是由这些岩浆结晶而成的，以及它们在储存和运输过程中与其他岩浆的混合。目前，研究人员不得不根据累积捕虏体的数量进行推断，而没有弧下深处的亲本累积序列的连贯图像，也没有各种累积样本之间的关系。在这项研究中，我们将尝试获得埋藏堆积层序分布的连贯理解，包括沿弧的横向和深度。PI及其同事先前的工作详细描述了沿弧喷发的堆积捕虏体的矿物学、结构外观和组成的变化；这些解释表明，在弧的北部，为单个火山提供岩浆的岩浆房往往位于较浅的水平，岩浆中含有较高比例的水。在沿750公里弧的累积捕虏体取样中，仍然存在小而重要的差距，我们建议纠正这些差距。需要一种新的方法来从捕虏体中的矿物组成中精确地得出累积结晶的深度和温度。我们将利用一个大型的实验数据库，其产品近似模拟观察到的累积捕虏体；我们将补充额外的实验，以确保在控制压力、温度、挥发性含量和氧逸度的条件下获得累积成分和矿物学的完整观察范围。然后，我们将根据数据库校准热力学模型，代表每一个观察到的矿物相。这些模型将描述平衡矿相在受控条件下组成的变化。它们将被校准，以期能够通过严格的统计分析恢复实验条件。然后，完成的模型将应用于自然样本的数据库，目的是基于到达地表的样本，建立一个地下岩浆储存系统的近似三维图像。原则上，这些信息可以与地震成像相结合，并与地震成像相验证，地震成像利用地震活动性来探测地壳结构，也可以与辐射测年相结合，为分析提供时间维度。从这个项目中得到的模型和解释将与世界上许多其他俯冲带的设置有关。

The Lesser Antilles volcanic arc provides an unusually complete and accessible case study of an active island arc subduction zone, formed by the subduction of the American plate under the eastern edge of the Caribbean plate since at least the Eocene. The arc offers a worldclass opportunity to explore subduction zone processes in a relatively simple intraoceanic setting. Arc-magmatism processes take place at depth and must be inferred through information available at the surface. A major source of such information is the nature and composition of magmatic rocks, which may either be erupted as lava from the volcanoes, or crystallized within the volcanic plumbing system, carried to the surface by ascending magma, and erupted as xenoliths within it. Both lavas and xenoliths reflect aspects of the processes that take place between the melting of volatile-rich magma in the mantle wedge above the subducting plate, and the eruption of lavas on the surface. An important category of xenoliths, known as cumulates, are formed from crystals that grew in a cooling magma, sank and accumulated at the bottom of the magma chamber, thereby indicating magma storage at discrete locations in the volcanic system. The cumulate sequences that are parental to these xenoliths lie at an unknown depth in the Earth's crust, but must contain extensive information encoded in the compositions of their crystals and captured bubbles of melt. This information includes hints about the source of the magma from which the cumulates crystallized, and its mixing with other magmas during storage and transport. Investigators are currently obliged to make inferences based on the population of cumulate xenoliths, without having a coherent picture of the parental cumulate sequences at depth under the arc, or the relationships of the various cumulate samples to each other. In this study we will attempt to gain a coherent understanding of distribution of buried cumulate sequences, both laterally along the arc and at depth. Previous work by the PI and co-workers has detailed the variation in the mineralogy, textural appearance and composition of cumulate xenoliths erupted along the arc; these are interpreted as showing that, towards the north of the arc, the magma chambers feeding the individual volcanoes tend to lie at shallower levels, and the magmas contain a higher proportion of water. Small but important gaps remain in sampling of cumulate xenoliths along the 750km length of the arc, and we propose to rectify these. A new method is required to derive precise depths and temperatures of cumulate crystallization from compositions of minerals in the xenoliths. We will make use of a large database of experiments with products that approximately mimic the observed cumulate xenoliths; we will supplement it with additional experiments, to ensure that the full observed range of cumulate compositions and mineralogies has been obtained under controlled conditions of pressure, temperature, volatile content and oxygen fugacity. We will then calibrate thermodynamic models against the database, representing each of the observed mineral phases. The models will characterize the variation of composition with the controlled conditions among mineral phases in equilibrium. They will be calibrated with a view to being able to recover the experimental conditions via a rigorous statistical analysis. The finished models will then be applied to the database of natural samples, with the aim of building up an approximate three-dimensional picture of the subsurface magma storage system based on the samples that reach the surface. In principle this information may be combined with and verified against seismic imaging, which uses seismicity to probe crustal structure, and also with radiometric dating to provide a time-dimension to the analysis. The models and interpretations derived from this project will be relevant to a number of other subduction zone settings worldwide.