Helioseismic constraints on the structure of the solar tachocline
Helioseismic constraints on the structure of the solar tachocline
复制标题
DOI：
10.1086/308050
复制
发表时间：
1999-12-10
影响因子：
4.9
通讯作者：
Tomczyk, S
中科院分区：
文献类型：
--
作者：
Charbonneau, P;Christensen-Dalsgaard, J;Tomczyk, S
本文介绍了一系列日震反演，目的是以尽可能高的置信度和精确度确定位于太阳对流包层底部下方的旋转切变层(塔可跃层)的结构。在不过度受分析方法选择的影响的意义上，我们特别感兴趣的是识别作为数据的稳健属性的倒置的特征。为此，我们实现了两种类型的二维线性反演，即正则化最小二乘(RLS)和减法最优局域平均(SOLA)，后者是根据旋转速度或其径向梯度来表示的。我们还使用基于遗传算法的正演模拟技术进行了非线性参数最小二乘拟合。每种方法的灵敏度都在合成数据上进行了彻底的测试。然后将这三种方法用于LOWL 2年频率分割数据集。该斜长岩的赤道厚度为W/R-。=0.039+/-0.013，赤道中心半径r(C)/R-。=0.693+/-0.002这三种技术也都表明塔可跃层是拉长的，但中心半径R(C)/R-不同。在纬度60度和赤道之间类似或等于0.024+/-0.004。假设误差不相关且呈正态分布，则可以99%的置信度拒绝严格的球面角跃层。塔可跃层厚度随纬度的变化没有统计学意义。讨论了这些结果对太阳长跃层的流体力学和磁流体动力学模型的影响。
This paper presents a series of helioseismic inversions aimed at determining with the highest possible confidence and accuracy the structure of the rotational shear layer (the tachocline) located beneath the base of the solar convective envelope. We are particularly interested in identifying features of the inversions that are robust properties of the data, in the sense of not being overly influenced by the choice of analysis methods. Toward this aim we carry out two types of two-dimensional Linear inversions, namely Regularized Least-Squares (RLS) and Subtractive Optimally Localized Averages (SOLA), the latter formulated in terms of either the rotation rate or its radial gradient. We also perform nonlinear parametric least-squares fits using a genetic algorithm-based forward modeling technique. The sensitivity of each method is thoroughly tested on synthetic data. The three methods are then used on the LOWL 2 yr frequency-splitting data set. The tachocline is found to have an equatorial thickness of w/R-. = 0.039 +/- 0.013 and equatorial central radius r(c)/R-. = 0.693 +/- 0.002 All three techniques also indicate that the tachocline is prolate, with a difference in central radius Delta r(c)/R-. similar or equal to 0.024 +/- 0.004 between latitude 60 degrees and the equator. Assuming uncorrelated and normally distributed errors, a strictly spherical tachocline can be rejected at the 99% confidence level. No statistically significant variation in tachocline thickness with latitude is found. Implications of these results for hydrodynamical and magnetohydrodynamical models of the solar tachocline are discussed.