A segmentation approach for the reproducible extraction and quantification of knickpoints from river long profiles
A segmentation approach for the reproducible extraction and quantification of knickpoints from river long profiles
复制标题
DOI：
10.5194/esurf-7-211-2019
复制
发表时间：
2018-09
影响因子：
3.4
通讯作者：
B. Gailleton;S. Mudd;F. Clubb;D. Peifer;M. Hurst
中科院分区：
文献类型：
--
作者：
B. Gailleton;S. Mudd;F. Clubb;D. Peifer;M. Hurst
抽象的。河流剖面陡度的变化或陡峭的垂直台阶（即瀑布）被认为表明侵蚀率、岩性或影响景观演化的其他因素的变化。这些变化被称为临界点或临界区，在基岩河流系统中普遍存在。这些特征被认为揭示了有关景观演化和侵蚀模式的信息，因此它们的位置经常在地貌文献中报道。报告断点和断点带的研究必须使用可重复的方法来量化其位置，因为它们的数量和空间分布在解释构造活动景观中发挥着重要作用。在这篇文章中，我们介绍了一种可重复的拐点和拐点区域提取算法，该算法使用通过沿河道长度整合流域面积来转换的河流剖面（所谓的积分或 χ 方法）。然后对剖面进行统计分段，并使用这些分段的不同坡度和高程变化来识别拐点、拐点区域及其相对大小。所识别的 knickpoints 和 knickzone 的输出位置与人类测绘相比毫不逊色：我们使用之前报告的 knickzones 在加利福尼亚州圣克鲁斯岛测试了该方法，并针对巴西 Quadrilátero Ferrífero 的新数据集测试了该方法。该算法允许提取不同的拐点形态，包括阶梯式、正坡折（向上凹）和负坡折拐点。我们确定了对所产生的 knickpoint 和 knickzone 位置影响最大的参数，并为该方法的使用和输出提供了指导，以生成可重现的 knickpoint 数据集。
Abstract. Changes in the steepness of river profiles or abrupt vertical steps (i.e. waterfalls) are thought to be indicative of changes in erosion rates, lithology or other factors that affect landscape evolution. These changes are referred to as knickpoints or knickzones and are pervasive in bedrock river systems. Such features are thought to reveal information about landscape evolution and patterns of erosion, and therefore their locations are often reported in the geomorphic literature. It is imperative that studies reporting knickpoints and knickzones use a reproducible method of quantifying their locations, as their number and spatial distribution play an important role in interpreting tectonically active landscapes. In this contribution we introduce a reproducible knickpoint and knickzone extraction algorithm that uses river profiles transformed by integrating drainage area along channel length (the so-called integral or χ method). The profile is then statistically segmented and the differing slopes and step changes in the elevations of these segments are used to identify knickpoints, knickzones and their relative magnitudes. The output locations of identified knickpoints and knickzones compare favourably with human mapping: we test the method on Santa Cruz Island, CA, using previously reported knickzones and also test the method against a new dataset from the Quadrilátero Ferrífero in Brazil. The algorithm allows for the extraction of varying knickpoint morphologies, including stepped, positive slope-break (concave upward) and negative slope-break knickpoints. We identify parameters that most affect the resulting knickpoint and knickzone locations and provide guidance for both usage and outputs of the method to produce reproducible knickpoint datasets.