Effects of Periglacial Landforms on Soil Erosion Sensitivity Factors and Predicted by Artificial Intelligence Approach in Mount Cin, NE Turkey

Dede V., Turan I. D., Dengiz O., Serin S., Pacci S.

EURASIAN SOIL SCIENCE, vol.55, no.12, pp.1857-1870, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 55 Issue: 12
  • Publication Date: 2022
  • Doi Number: 10.1134/s106422932260138x
  • Journal Name: EURASIAN SOIL SCIENCE
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Agricultural & Environmental Science Database, Aquatic Science & Fisheries Abstracts (ASFA), BIOSIS, CAB Abstracts, Geobase, Pollution Abstracts, Veterinary Science Database
  • Page Numbers: pp.1857-1870
  • Keywords: periglacial processes, artificial neural network, soil erosion, Lesser Caucasus, AGGREGATE STABILITY, NEURAL-NETWORK, ROCK GLACIERS, MANAGEMENT
  • Ondokuz Mayıs University Affiliated: Yes

Abstract

In this study, the differences in soil properties formed on various periglacial landforms located on slope land and high elevation so, this case create main problem against to soil erosion. The main aims of the study are to determine the physico-chemical properties and some soil erosion sensitivity parameters of the soils formed on the different periglacial landforms of Mount Cin and to predict those soil erosion sensitivity factor using artificial neural network (ANN). It was detected three different periglacial landforms on the Mount Cin. Stony earth circles spread over Cin Hill which is on the summit plain of Mount Cin, while non-sorted steps are located on the northern slopes of Cin Hill and Topkaya Hill. In addition, mud circle landforms spread to the south of Karacakrak Hill. 25 soil samples were taken from the periglacial landforms in the study area. Afterwards, the physico-chemical properties of the samples were analysed in the laboratory. According to soil analysis from various periglacial landforms, the dominant soil texture is sandy loam: clay ranges from 5.61 to 16.79%, and sand from 48.61 to 76.72%. Also, the average soil erosion sensitivity factors, namely structure stability index (SSI), dispersion rate (DR), and crust formation (CF), were calculated at 29.65, 28.36, and 40.72%, respectively. Moreover, ANN is a model that can operate directly like the human brain. ANN uses the data of the current problem to make predictions. According to regression results of soil erosion sensitivity factors using ANN, the highest prediction rate was obtained for SSI (78%) and the lowest for DR (57%).