AbstractsEarth & Environmental Science

Characterization and prediction of ephemeral gully erosion in Wallonia

by Alexandre Maugnard




Institution: Université Catholique de Louvain
Department: Environmental Sciences
Year: 2015
Keywords: Ephemeral gully erosion; Gully occurrence prediction; Aerial photographs; Topographic threshold; Gully initiation; Gully end; Uncertainty on gully identification
Record ID: 1075144
Full text PDF: http://hdl.handle.net/2078.1/158427


Abstract

Ephemeral gully erosion is responsible for large soil losses in agricultural settings. It contributes to a loss of agronomic value of the soil and causes severe damage to crops. It also induces an increased workload for farmers. By increasing the connectivity in the landscape, the development of ephemeral gullies also enhances sediment delivery to downstream sites. This aggravates off-site effects such as muddy floods and pollution of watercourses by sediment and associated contaminants. In Belgium most studies on ephemeral gully erosion have been performed in the loess belt of Flanders and little information are available for Wallonia. The aim of the thesis was to characterize and predict the occurrence of ephemeral gully erosion in Wallonia in order to improve our understanding of the phenomenon and to suggest relevant mitigation techniques. Gully identification based on interpretation of aerial photographs was shown to be operator-dependent and to affect gully characterization. The number, total length, dominant type as well as topographic conditions leading to gullying were found to vary greatly across Wallonia. A new, more robust method was proposed to characterize the topographic contexts prone to ephemeral gully initiation. The main factors controlling ephemeral gully ending and their spatial variability were also identified. This information may prove useful to manage gully erosion in watersheds sensitive to gullying. The occurrence of ephemeral gully erosion was predicted using a set of topographic or hydraulic indicators. However, their performances were moderate at best. This may be improved in the future by using high resolution DEMs and by ameliorating the hydrological production and transfer functions used. It may also be necessary to integrate information on linear landscape elements and tillage direction. (AGRO - Sciences agronomiques et ingénierie biologique)  – UCL, 2015