|
Performance Evaluation of GARP Model in Forecasting the Probability of Gully Erosion Occurrence in Abgendi Basin, Kohgiluye & Boyerahmad Province
|
Iman Saleh *  , Omid Rahmati , Seyed Masoud Soleimanpour , Majid Khazaei , Samad Shadfar |
|
|
|
Abstract: (2674 Views) |
| The present study has evaluated the efficiency of GARP model in forecasting the probability of gully erosion occurrence in Abgendi basin, Kohgiluye & Boyerahmad province. After field visits, 81 gullies were recorded by GPS and surveying camera, then the map of gully erosion distribution was prepared in ArcGIS environment. Variables of elevation, slope, slope direction, distance to road, distance to surface streams, land-use, pedology, lithology, drainage density and Topographic Wetness Index (TWI) were selected as the factors affecting the occurrence of gully erosion and used as independent variables in the modelling process. The modelling of gully erosion occurrence was accomplished using Desktop Garp software. Also, Receiver Operating Characteristic (ROC) and True Skill Statistic (TSS) were used for validation of the prediction results. The results indicated that, the most likely probability of gully erosion occurrence is concentrated in the central part of the basin, and it is reduced by moving towards the border parts especially southeast of the basin. Accuracy evaluation criteria ROC=0.97 and TSS=0.84 also indicate excellent performance of the GARP model in forecasting the probability of gully erosion in Abgendi basin. |
Article number: 4 |
| Keywords: Land-use, Machine learning model, Soil conservation, Topographic Wetness Index |
|
|
Full-Text [PDF 1805 kb]
(58 Downloads)
|
Type of Study: Research |
Subject:
Special
Received: 2022/05/24 | Revised: 2023/06/19 | Accepted: 2022/09/10 | Published: 2023/04/20 | ePublished: 2023/04/20
|
|
|
|
|
|
