| Autor: |
Drid, Ilhem, Achour, Yacine, Zighmi, Karim, Nouioua, Thamer, Khababa, Oussama, Rahmani, Samra, Ouadfel, Oumelkheir, Atoui, Bachir |
| Zdroj: |
Arabian Journal of Geosciences; 7/15/2022, Vol. 15 Issue 14, p1-18, 18p |
| Abstrakt: |
Gully erosion is a pervasive problem which frequently develops in semi-arid environments. The objective of this study is to test the efficiency of three individual models for designing gully erosion susceptibility maps (GESM), to assess the performance of the generated models, and to explore the utility of these maps to help decision makers and local land users in the evaluation of land degradation. The K’sob watershed (1570 sq km) is an agricultural land with the most affected areas by soil erosion in the Algerian Eastern Tell, NE Algeria. The prevalence of alluvial terraces, clays, and marls in the area led to the development of very large gullies. The most significant consequences of this phenomenon are the silting up of the K’sob dam, the loss of agricultural lands, and the nonprotection of the watershed. To evaluate soil erosion by gullying in this watershed, random forest, extreme gradient boosting, and logistic regression models were applied. In the entire area, 565 (1475 gridded cells) gully erosion were identified by a hard field investigation. Eleven gully erosion-conditioning factors were selected and assessed to find the most susceptible areas to this hazard. Then, the performance of the applied models was assessed. Results show that the XGBoost model has the best performance (AUC = 0.942, Std. error = 0.009, and CI “0.893–0.971”). Requiring land users in the evaluation of land degradation can supply awareness into local degradation processes and could improve the local pertinence of indicators for investigating progress towards land degradation. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
Full text from SpringerLink