Reply to "The new assessment of soil loss by water erosion in Europe. Panagos P. et al., 2015 Environ. Sci. Policy 54, 438-447-A response" by Evans and Boardman [Environ. Sci. Policy 58, 11-15]

The new assessment of soil loss by water erosion in Europe (Panagos et al., 2015a) was commented by Evans and Boardman (2016), who raised not only concerns related to the spatial differences outlined by our work compared to their visual semi-qualitative assessment conducted in Britain during the late eighties, but also generally to the suitability, validity and scientific robustness of the applied modelling approach. The objective of the pan-European assessment using the Revised Universal Soil Loss Equation (RUSLE) was not to outcompete any regional- or national-scale modelling, but to harmonize and improve our knowledge and our understanding of current soil erosion rates by water across the European Union. The focus of such a modelling project is on the differences and similarities between regions and countries beyond national borders and nationally adapted models. In order to do so, a state-of-the-art large-scale spatially distributed modelling exercise using harmonized datasets and a unified methodology to suit the pan-European scale was carried out. We reply that the semi-qualitative approach proposed by Evans and Boardman (2016) is not suitable for application at the European scale because of work force and time requirements, input data accessibility issues, accuracy of field-based estimates, subjectivity of soil loss estimates during the aerial and terrestrial photo interpretation, impossibility of upscaling or downscaling, inadequate representation of sheet erosion processes, lack of spatial and temporal representativeness, and lack of detailed description expressing the risk level. As such, their methodology has limited applicability, with today's financial resources it is not feasible at European or at national scale and, most important, cannot respond to policy requests regarding scenarios of climate and land cover/use change. In contrast to Evans and Boardman (2016), we do know that RUSLE, like probably any other approach, is not able to reproduce "reality". The latter is actually a misjudgment which has been extensively discussed 20 years ago. Modelling in general and large-scale modelling specifically can per se not aim at an accurate prediction of point measurements, but tests our hypothesis on process understanding, relative spatial and temporal variations, scenario development and controlling factors (Oreskes et al., 1994). As such, our approach can be offered as a helpful tool to policy makers at pan-European scale. We are confident that the simple transparent structure of RUSLE as well as the discussion of the uncertainties of each modelling factor will help to supply objective guidance to policy makers