Rates and spatial variations of soil erosion in Europe: A study based on erosion plot data

Abstract

An extensive database of short to medium-term erosion rates as measured on erosion plots in Europe under natural rainfall was compiled from the literature. Statistical analysis confirmed the dominant influence of land use and cover on soil erosion rates. Sheet and rill erosion rates are highest on bare soil; vineyards show the second highest soil losses, followed by other arable lands (spring crops, orchards and winter crops). A land with a permanent vegetation cover (shrubs, grassland and forest) is characterised by soil losses which are generally more than an order of magnitude lower than those on arable land. Disturbance of permanent vegetation by fire leads to momentarily higher erosion rates but rates are still lower than those measured on arable land. We also noticed important regional differences in erosion rates. Erosion rates are generally much lower in the Mediterranean as compared to other areas in Europe; this is mainly attributed to the high soil stoniness in the Mediterranean. Measured erosion rates on arable and bare land were related to topography (slope steepness and length) and soil texture, while this was not the case for plots with a permanent land cover. We attribute this to a fundamental difference in runoff generation and sediment transfer according to land cover types.

On the basis of these results we calculated mean sheet and rill erosion rates for the European area covered by the CORINE database: estimated rill and interrill erosion rates are ca. 1.2 t ha^− 1 year^− 1 for the whole CORINE area and ca. 3.6 t ha^− 1 year^− 1 for arable land. These estimates are much lower than some earlier estimates which were based on the erroneous extrapolation of small datasets. High erosion rates occur in areas dominated by vineyards, the hilly loess areas in West and Central Europe and the agricultural areas located in the piedmont areas of the major European mountain ranges.

Introduction

Landscapes where human activities are expanding commonly witness a shift from natural to accelerated erosion (Wilkinson, 2005) which threatens the soil resources and the sustainability of natural ecosystems (CEC, 2006). The negative effects of soil erosion include water pollution and siltation, crop yield depression, organic matter loss and reduction in water storage capacity (e.g. Pimentel et al., 1995, Bakker et al., 2004, Boardman and Poesen, 2006), which may lead to fundamental social challenges such as land abandonment and the decline of rural communities (Bakker et al., 2005). The protection of soil resources has therefore been recognised as an important objective of environmental policy (CEC, 2006): this requires a correct assessment of erosion rates and their geographical distribution. Such an assessment is also important from a purely scientific point of view: soil erosion affects global biogeochemical cycles (Van Oost et al., 2007, Quinton et al., 2010) and understanding the human impact on river sediment fluxes requires that sediment mobilisation by soil erosion is correctly quantified (Hooke, 2000, Wilkinson, 2005).

The United States is probably the only country where soil erosion data have been collected over a long period of time using standardized methodology in relation to the development of the (Revised) Universal Soil Loss Equation (Wischmeier and Smith, 1978, Nearing et al., 2000). For other areas including European, national or continental erosion estimates have mostly been based on sporadic evidence. For example, Pimentel et al. (1995) extrapolated erosion rates measured on a few plots on a single location in Belgium to estimate average erosion rates for the whole continent, a procedure which was later criticized by Boardman (1998).

Various models of erosion assessment for large areas in Europe have been developed (e.g. Gobin et al., 2004, Kirkby et al) and such models provide valuable information on the spatial distribution of soil erosion and how it may be affected by land use and/or climate change. However, model evaluation has hitherto been limited (e.g. Tsara et al., 2005, Licciardello et al., 2009). As demonstrated by Favis-Mortlock (1998), results of erosion models are often poor when they are applied to other areas than where they were tested. Such models may also require input data that are not always available for large areas with the required accuracy (Jones et al., 2003).

In Europe as a whole, existing field data have hitherto been used to a limited extent only to assess national or continental soil erosion rates because a transnational programme of soil erosion measurement using a standardized procedure has never been conducted. Nevertheless, erosion data have been collected in Europe through a number of research projects based on various technologies including measurements of changes in surface height, rainfall simulation experiments and tracer analyses. The majority of data have been collected on bounded field plots under natural rainfall, more or less comparable to those used by the USDA for the calibration and validation of the (R)USLE. Although plot size, methodology, and the length of the measurement period vary considerably, these erosion plot data present a wealth of information about actual erosion rates in Europe.

We compiled a large database of erosion rates measured on medium-sized plots (> 3 m and < 200 m in length) under natural rainfall conditions on 81 sites in Europe. We used this database to assess the overall intensity and spatial distribution of sheet and rill erosion under various land use types. Specific objectives of this paper are: 1) to develop a documented database on sheet and rill erosion rates under various land use types; 2) to identify and, if possible, quantify the effect of factors controlling erosion rates; 3) to estimate average sheet and rill erosion rates under various land use types, and 4) to map the spatial distribution of erosion based on the measured erosion rates, topography, land use and soils.