首页 | 本学科首页   官方微博 | 高级检索  
     检索      


Assessment of sidewall erosion in large gullies using multi-temporal DEMs and logistic regression analysis
Authors:J A Martínez-Casasnovas  M C Ramos  J Poesen
Institution:a Department of Environment and Soil Science, University of Lleida, Rovira Roure 191, E-25198, Lleida, Spain;b Laboratory for Experimental Geomorphology, K.U., Leuven, Belgium
Abstract:Although in the last decades gully erosion has been a thriving research field, few studies have specifically addressed the contribution and location of sidewall erosion processes in gullies. In this paper, sidewall erosion in some large gullies in a Mediterranean area (Anoia-Penedès, NE Spain) is mapped and assessed for two time intervals (1975–1995 and 1995–2002), using detailed digital elevation models derived from aerial photographs at a scale of 1:5000 to 1:7000. Logistic regression analysis is applied to compute the probability of occurrence of gully sidewall erosion from terrain variables. The results confirm the complex nature of sidewall processes, whose intensity is most probably related to rainfall characteristics. Prolonged wet soil conditions in the period 1995–2002, together with the large and high-intensity rainfall of an extreme event occurred on 10th June 2000, help to explain the different sediment production rates: 16±0.4 Mg ha−1 year−1 in 1975–1995 and 83±6.3 Mg ha−1 year−1 in 1995–2002. The logistic regression analysis revealed that gully-wall slope angle was the main factor controlling gully sidewall failure. In gully walls with high slope angles, tension crack development is the main process promoting wall collapse. The application of the logistic regression model showed a high overall accuracy (87%) but over 50% of commission and omission errors for the class of interest (sidewall erosion), in agreement with the variance explained by the model.
Keywords:Author Keywords: Gully erosion  Sidewall processes  Digital elevation models  Logistic regression analysis
本文献已被 ScienceDirect 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号