Prasicek Günther, Otto Jan-Christoph, Montgomery David R, Schrott Lothar
Department of Geoinformatics - Z_GIS, University of Salzburg, Hellbrunnerstr. 34, 5020 Salzburg, Austria ; Department of Geography and Geology, University of Salzburg, 5020 Salzburg, Austria.
Department of Geography and Geology, University of Salzburg, 5020 Salzburg, Austria.
Geomorphology (Amst). 2014 Mar 15;209(100):53-65. doi: 10.1016/j.geomorph.2013.11.026.
Erosion by glacial and fluvial processes shapes mountain landscapes in a long-recognized and characteristic way. Upland valleys incised by fluvial processes typically have a V-shaped cross-section with uniform and moderately steep slopes, whereas glacial valleys tend to have a U-shaped profile with a changing slope gradient. We present a novel regional approach to automatically differentiate between fluvial and glacial mountain landscapes based on the relation of multi-scale curvature and drainage area. Sample catchments are delineated and multiple moving window sizes are used to calculate per-cell curvature over a variety of scales ranging from the vicinity of the flow path at the valley bottom to catchment sections fully including valley sides. Single-scale curvature can take similar values for glaciated and non-glaciated catchments but a comparison of multi-scale curvature leads to different results according to the typical cross-sectional shapes. To adapt these differences for automated classification of mountain landscapes into areas with V- and U-shaped valleys, curvature values are correlated with drainage area and a new and simple morphometric parameter, the Difference of Minimum Curvature (), is developed. At three study sites in the western United States the thresholds determined from catchment analysis are used to automatically identify 5 × 5 km quadrats of glaciated and non-glaciated landscapes and the distinctions are validated by field-based geological and geomorphological maps. Our results demonstrate that is a good predictor of glacial imprint, allowing automated delineation of glacially and fluvially incised mountain landscapes.
冰川作用和河流作用造成的侵蚀以一种长期以来被认可的典型方式塑造着山地景观。由河流作用切割形成的高地山谷通常具有V形横截面,坡度均匀且适中,而冰川山谷往往具有U形剖面,坡度梯度不断变化。我们提出了一种新颖的区域方法,基于多尺度曲率与流域面积的关系自动区分河流和冰川山地景观。划定样本集水区,并使用多个移动窗口大小来计算从谷底流路附近到完全包括谷壁的集水区各部分等多种尺度上的像元曲率。单尺度曲率对于冰川覆盖和未冰川覆盖的集水区可能具有相似的值,但根据典型的横截面形状,多尺度曲率的比较会导致不同的结果。为了适应这些差异以将山地景观自动分类为具有V形和U形山谷的区域,将曲率值与流域面积相关联,并开发了一个新的简单形态测量参数——最小曲率差()。在美国西部的三个研究地点,根据集水区分析确定的阈值用于自动识别5×5千米的冰川和非冰川景观样方,并通过基于实地的地质和地貌图对这些区分进行验证。我们的结果表明,是冰川印记的良好预测指标,能够自动划定冰川和河流切割的山地景观。