Rill erosion on cultivated hillslopes during two extreme rainfall events in Normandy, France

Abstract

In the Normandy region of France, two extreme runoff events took place during winter of 1999 and spring of 2000 that caused flooding and considerable on-site as well as off-site damages. After each event, erosion damage was mapped on an experimental cultivated catchment (94 ha). The location and extent of rill, ephemeral gullies and deposits were measured. For each field, information on land use and soil surface characteristics were also collected. Since 1991, when experimental work and survey campaigns were initiated on this catchment, interrill erosion dominated over rill or ephemeral gully erosion. Our objective was to link information on topography, soil surface characteristics, and land use with intensity and type of erosion that developed. Erosion features that were most related to topographic attributes and hence less affected by seasonal variations were ephemeral gullies and some predefined deposit types. Topographic attributes alone were not sufficient to determine development of rill erosion, due to seasonal differences in vegetative cover. At the catchment scale, total erosion varied from 10 t/ha in December with 93% of the catchment area with vegetation cover ≤20%; to 1.5 t/ha in May, with 73% of the catchment area with vegetation cover >60%. The relative importance of ephemeral gully erosion out of total linear erosion varied from ca. 24% for the rainfall events of December to more than ca. 83% for the rainfall events of May. These results also highlight the fact that average annual sediment delivery as well as the relative importance of different erosion forms at the catchment scale cannot be generalised. Erosion prediction and erosion assessment risks are strongly dependent on catchment land use, morphology and storm characteristics.

Introduction

Particle removal from the soil mass is controlled by the balance between the erosive forces of overland flow and rainfall and the topsoil erosion resistance. Interrill erosion is driven by rainfall impact leading to shallow runoff. Rill or ephemeral gully erosion is localised in areas where the shear stress caused by concentrated flow is the main source of energy. Detachment capacity of the flow is related to its hydraulic characteristics. Many proposed hydraulic indices of rill initiation in the literature are related directly to critical tractive force. These include shear velocity (Govers, 1985, Rauws and Govers, 1988), bed shear stress (Chisci et al., 1985, Torri et al., 1987), stream power (Rose, 1985), unit stream power (Govers and Rauws, 1986, Moore and Burch, 1986), and either unit or total discharge (Meyer et al., 1975).

The validity and effectiveness of these hydraulic indicators in predicting rill erosion are dependent upon methodology, which includes field plots and laboratory flumes of diverse design and dimension, slope intensity and shape, pedological material type, water condition and rainfall characteristics. In addition, even for what is accepted as “satisfying” results there is not only a scattering in the data which reflects some system variability, but also the difficulty of obtaining accurate hydraulic measurements in very shallow flows and identifying the initial stage of rill initiation (Bryan, 2000). Moreover, as stated by Römkens et al. (1997) these concepts are more readily applied to non-cohesive soils where interparticle or interaggregate bonding is of no or minimum concern. When a cohesive soil condition prevails the validity of these relationships for a broad application need to be examined. One such situation occurs when soils are susceptible to sealing, like on the agricultural plateau of the European Lœss Belt. On cultivated lands, rill and ephemeral gully erosion involve flow concentration along preferential pathways, generally controlled by topography (slope aspect, length and intensity), vegetation patterns, natural or agricultural-induced microtopography (furrows, deadfurrow, backfurrow), tracks or ditches. However, flow concentration alone does not necessarily cause rill incisions (Emmett, 1970, Dunne and Dietrich, 1980; cited by Bryan, 2000). Characterisation based on hydraulic parameters alone may not always be sufficient to fully account for rills and ephemeral gullies at the catchment scale. Both flow characteristics, soil surface state and land use have to be considered. In order to study the magnitude of infiltration and soil loss and their controlling factors in a catchment area, Bohm and Gerold (1995) related several soil physical and chemical properties to various soil loss and infiltration variables by regression analysis. The regression equation revealed that neither texture nor shear strength, but plant cover, was the most important factor. In the same study, an erosion damage survey of a 10 ha area attributed 93% of the losses from seedbed plots (mean soil loss of 3.4 t/ha), with the remaining losses from stubble and fallow (mean soil loss of 0.1 and 0.6 t/ha, respectively).

In the Normandy region (France), two extreme runoff events took place over a 6-month period that caused flooding and considerable on-site as well as off-site damages. After each event, erosion damage was mapped from two experimental cultivated catchments. On the smallest catchment (94 ha), location and extent of rill, ephemeral gullies and deposits were mapped precisely with a tacheometer. For each field, information on land use and soil surface characteristics was also collected. Since experimental work and surveys were initiated on this catchment in 1991, water erosion had been reported to be mainly due to interrill processes, whereas rills or ephemeral gullies had been observed only occasionally.

The aim of this study was to describe and understand the controlling factors that triggered the appearance and development of ephemeral gullies and incisions during these extreme events by comparing information on soil surface characteristics and land use with the intensity and type of erosion features.