[论文解读] R scripting libraries for comparative analysis of the correlation methods to iden- tify factors affecting Mariana Trench formation

Polina Lemenkova|arXiv (Cornell University)|Jan 1, 2018

Geological Studies and Exploration参考文献 10被引用 21

一句话总结

本研究采用R编程库比较皮尔逊、斯皮尔曼与肯德尔相关性方法，以识别影响马里亚纳海沟形成的环境因素。基于海深、构造与地质数据，研究识别出沉积物厚度、坡度角与构造板块动力学为最具影响力的三个因素，统计可视化方法（如相关图与主成分分析）证实其在塑造海沟地貌中的主导作用。

ABSTRACT

Mariana trench is the deepest place on the Earth. It crosses four tectonic plates of the Pacific Ocean: Mariana, Caroline, Pacific and Philippine. The formation of the trench is caused by the complex interconnection of various environmental factors. The aim of this study was to describe and characterize various impact factors affecting formation of the Mariana trench geomorphology and continental margin environments using R programming language and mathematical algorithms of correlation methods written on R code. To record the system of geological, tectonic, geographic, oceanological and bathymetric features affecting Mariana trench, a combination of statistical methods, GIS and R programming codes were applied. The questions answered are as follows: which factors are the most influencing for the Mariana trench morphology, and to what extend do they affect its development? Is sedimental thickness of the ocean trench basement more important factors for the trench formation comparing to the steepness slope angle and aspect degree? Three methods of computing were tested: Pearson correlation, Spearman correlation, Kendall correlation, numerical correlogram, correlation matrix and cross-correlatios to analyze environmental impact factors. The correlogram matrices are computed and visualized by R scripting libraries. Complex usage of programming tools, mathematical statistics and geospatial analysis enabled to get a differentiated understandings of the hadal environments of the Mariana trench. The results revealed following three types of factors having the highest score: geometric (tg° slope angle), geologic (sedimental thickness) and tectonic structure. The results furthermore indicated that tectonic plates, sedimental thickness of the trench basement and igneous volcanic areas causing earthquakes play the most essential role in the geomorphology of the trench.

研究动机与目标

• 识别并排序影响马里亚纳海沟地貌的最关键环境因素。
• 比较皮尔逊、斯皮尔曼与肯德尔相关性方法在分析地理空间影响因素时的有效性。
• 开发并应用R编程库，对相关性矩阵与交叉相关性进行统计可视化。
• 评估沉积物厚度、坡度陡峭度、坡向角度与构造结构在海沟形成中的相对重要性。

提出的方法

• 使用R编程结合统计库计算皮尔逊、斯皮尔曼与肯德尔相关系数。
• 利用数值相关图与相关性矩阵可视化环境因素之间的关系。
• 对海深剖面进行交叉相关性分析，以检测时间与空间上的依赖性。
• 通过主成分分析（PCA）识别25条剖面中海深数据的主导模式。
• 将GIS数据与R脚本整合，处理并可视化构造板块边界与火山活动距离等地理空间变量。
• 使用ggcorr与自定义R函数生成影响因素的散点图矩阵与层次树状图。

实验结果

研究问题

• RQ1 哪些环境因素对马里亚纳海沟的地貌影响最为显著？
• RQ2 沉积物厚度、坡度角与坡向角度在海沟形成中的影响如何比较？
• RQ3 构造板块动力学与火山活动在多大程度上与海沟发育相关？
• RQ4 不同相关性方法（皮尔逊、斯皮尔曼、肯德尔）在识别地理空间变量间显著关系时表现如何？
• RQ5 构造、海深与地质因素对海沟整体形态结构的相对贡献是什么？

主要发现

• 海沟基底的沉积物厚度被确定为三大最具影响力因素之一，与海沟地貌具有强统计相关性。
• 坡度角（tg°）成为关键几何因素，平均坡度约为4–5度，且在海底峡谷区域更陡峭。
• 构造板块边界与俯冲动力学，特别是马里亚纳、太平洋与菲律宾板块的相互作用，与海沟形成的相关性最高。
• 岩浆火山活动及其伴随的地震活动被发现是重要贡献因素，尤其在引发构造不稳定性方面。
• 肯德尔tau-b相关性方法有效处理了结 tied 数据，并证实了构造结构与地貌演化之间存在强序数关联。
• 主成分分析与散点图矩阵显示，25条分析剖面的海深剖面高度相关，表明存在一致的地貌模式。