Inland notches are elongated C-shape indentations carved into hard, dense carbonate rocks of the Mediterranean climate zone. In recent years these unique features became of particular interest as geoindicators of denudation and slope erosion processes. The study goals are to learn the processes acting upon the notches cavity by (1) examining the micromorphology of the backwall, (2) identifying and classifying zones of erosion, (3) learning the surface flow pattern and, (4) proposing an erosion mechanism induced by exfoliation. To this end, a machine learning algorithm was designed, using the microstructure, texture, streamlines formation and potential focal flow. The workflow uses data from terrestrial laser scanning, applies high-resolution textural and curvature analyses and trains an artificial neural network to the different segments of the notch. The reconstructed surface model shows division into three distinct belts within the cavity surface: (1) The spalled belt [SB]; (2) The transition zone belt [TZ] and (3) The deposition belt [DB]. Each belt owes its distinct microtopography, implying on different geomorphic processes. The newly developed methodology improves the recognition of streamline networks along the cavity backwall, which mimic alluvial networks. The organization of the reconstructed micro channels is similar to drainage basin systems, where, first order channels developed upstream converge into larger ones and diverge as distributaries as the gradient decreases, depositing tufa along the notches base.
|State||Published - Apr 2020|
Bibliographical noteFunding Information:
This research was supported by the Israel Science Foundation (ISF-Grant No. 894/11 ). The authors are grateful to Atzmon Ben-Binyamin and Alik Chebotar for supporting the data collection in the field. Appendix A
© 2019 Elsevier B.V.
ASJC Scopus subject areas
- Earth-Surface Processes