Soil erosion modelling at European scale by using high resolution input layers

Soil erosion by water is one of the most widespread forms of soil degradation. Since soil erosion is difficult to measure at large scales, soil erosion models are a crucial estimation tool at regional, national and European levels. The high heterogeneity of soil erosion causal factors, combined with often poor data availability is an obstacle for the application of complex soil erosion models. Thus, the empirical Revised Universal Soil Loss Equation (RUSLE) (Renard et al., 1997), which predicts the average annual soil loss resulting from raindrop splash and runoff from field slopes, is still most frequently used at large spatial scales. The RUSLE is the simple multiplication of 5 soil erosion risk factors:
• Soil erodibility (K-factor)
• Rainfall erosivity (R-factor)
• Cover and management (C-factor)
• Support practices (P-factor)
• Slope length and Steepness (LS-factor)
The PhD study proposes a new soil erosion map of Europe (RUSLE2015) which has the following characteristics:
- It is based on peer review and high quality input factors
- The factors are composite layers: K-factor includes stoniness, C-factor includes Management Practices (tillage practices, cover crops, plant residues) and Vegetation fraction through remote sensing, R-factor includes high temporal resolution precipitation measurements of 1541 stations, P-factor includes support practices (contouring, stone walls, grass margins) and LS-factor is based on 25m Digital Elevation Model.
- has a very fine resolution of 100m
- allows land use and management scenarios and can be used by policy makers
- makes the data available in European Soil Data Centre (ESDAC)
- it is proposed in a transparent way and follows the literature principles
The first chapter makes a comparison of the pan-European soil erosion data collection (named EIONET-SOIL) with the modelled data from PESERA. This data collection concluded that almost all member states of the European Union are using (R)USLE models for the estimation of soil erosion. The paper identified the areas with high discrepancies between the two different soil erosion estimation approaches. By concluding this study, I have decided to use a RUSLE approach at European scale due to limitations of PESERA and data availability from Member states using RUSLE.
The second chapter includes the key parameter for modelling soil erosion which is the soil erodibility, expressed as the K-factor. The soil erodibility which expresses the susceptibility of a soil to erode, is related to soil properties such as organic matter content, soil texture, soil structure and permeability. With the Land Use/Cover Area frame Survey (LUCAS) soil survey in 2009 a pan-European soil dataset is available for the first time, consisting of around 20,000 points across 25 Member States of the European Union. The aim of this study is the generation of a harmonised high-resolution soil erodibility map (with a grid cell size of 500 m) for the 25 EU Member States.
The third chapter proposes a new soil erosion model named G2 which uses the empirical formulas of the Universal Soil Loss Equation (USLE). The difference is that G2 makes allows for the integrated spatio-temporal monitoring of soil erosion as the Rainfall erosivity (R-factor) and Vegetation retention (V-factor; known as C-factor in USLE) are proposed on a monthly temporal resolution. This study in Crete (Greece) allowed to deep the knowledge of each erosion factors which are to be modelled at European scale.
The fourth chapter assesses rainfall erosivity in Europe in the form of the RUSLE R-factor, based on the best available datasets. Data have been collected from 1,541 precipitation stations in all European Union (EU) Member States and Switzerland, with temporal resolutions of 5 to 60 minutes. The R-factor values calculated from precipitation data of different temporal resolutions were normalised to R-factor values with temporal resolutions of 30 minutes using linear regression functions.
The fifth chapter assesses rainfall erosivity in Greece on a monthly basis in the form of the RUSLE R-factor, based on 30-minutes data from 80 precipitation stations covering an average period of almost 30 years. The proposed R-factor spatio-temporal analysis shows a high intra-annual variability of rainfall erosivity which should also be investigated in whole Europe.
The sixth chapter presents the cover-management factor (C-factor) which is considered to be the most important because policy makers and farmers can intervene and, as a consequence, may reduce soil erosion rates. In arable lands, the C-factor was estimated using crop statistics (% of land per crop) and management practices data such as conservation tillage, plant residues and winter crops. The C-factor in non-arable lands was estimated by weighting the range of literature values found by fractional vegetation cover, which was estimated based on the remote sensing datasets.
The seventh chapter assesses support practice factor (P-factor) which is rarely taken into account in soil erosion risk modelling. The P-factor model considers the latest policy developments in the Common Agricultural Policy (contour farming) and the impact of stone walls and grass margins in reducing soil loss. The P-factor modelling tool can potentially be used by policy makers to run soil-erosion risk scenarios.
The eight chapter proposes an overview of the RUSLE2015 model and presents the final soil erosion map of Europe based on the input layers discussed in previous chapters. This concluding chapter makes an assessment of the soil erosion map per land use, region and per class of soil erosion. The verification of the map with other data sources has been satisfactory. Finally, this chapter proposes the use of soil erosion map for policy making in European Union and predicts the soil erosion trends based on land management and land use changes.
The first 4 chapters are published in peer review journals. The 5th chapter is under revision after initial acceptance, the 6th and 7th chapters have initially been accepted (under second revision) and editors have requested some changes. The concluding chapter (9th) has been submitted in February in a high impact factor peer review journal.