Soil Erosion map of Europe based on high resolution input datasets

Transcription

1 Soil Erosion map of Europe based on high resolution input datasets Panagos Panos 1, Pasqualle Borrelli 1, Katrin Meusburger 2, Cristiano Ballabio 1, Christine Alewell 2 1 European Commision Joint Research Centre 2 University of Basel, Environmental Geosciences

2 Policy: Soil Thematic Strategy Sealing Soil Biodiversity loss Erosion Decline of Soil Organic Matter Soil Threats Salinization Compaction Landslides Contamination 2/22

3 Why we do soil erosion modelling? Develop Indicators for DG ESTAT: Aggregated data at different regional levels (per country, Region, province - NUTS0, NUTS2, NUTS3) Policy support to DG AGRI: Common Agricultural policy impact. Provide aggregated data for agricultural areas with soil loss rates > 10 t ha -1 yr -1 Policy Support to DG ENV: Monitor the state of soils (Aggregated data & maps of soil erosion) European Environmental Agency (EEA): State of the Environment report EUROPE 2020: Develop soil erosion indicator for Resource efficiency scoreboard Development of Green Growth Indicators for the Organisation for Economic Cooperation and Development (OECD) Open Data Access through the European Soil Data Centre (ESDAC): Distribute of Soil erosion data plus the independent layers (soil erodibility, rainfall erosivity, support practices, Topography, etc.) to the scientific community and policy makers in EU-countries >200 datasets of PESERA estimates ( ), >280 datasets of soil erodibility(april 2014 Mar 2015); 100 datasets of Rainfall erosivity(jan - Mar 2015)

4 Why to model the factors of soil erosion? Which are the gaps? Developments of indicators for Policy makers: Request for maps and aggregated data on soil erosion for application of policies in European Union: Agriculture, Environment, Climate Change, etc. Previous attempts of European Soil Erosion maps were not convincing neither the scientific community nor policy makers (not even representatives from the countries) Gaps in the modelling of erosion factors: Soil Erodibility (K-factor) did not consider stoniness; K-factor was estimated using as input European Soil Database (low resolution; outdated data 1960 s) Rainfall erosivity (R-factor) was estimated based only on low resolution precipitation data; R-factor was estimated arbitrary based on local functions. Cover-Management (C-factor) was estimated based only on CORINE Land cover Classes ignoring vegetation density; C-factor did not consider the management (tillage, crop residues). Support practices (P-factor) was never estimated at European scale How to model erosion in Europe in different temporal scales?

5 Soil Erosion EIONET Map ( ) Panagos et al. (2014), Soil Science & Plant Nutrition

6 EIONET data collection and PESERA EIONET: The first attempt to collect soil data from Member States No legal obligation Limited participation (8 countries) The Metadata (model used, land uses, period, input factors) accompanied the data proved of extreme importance for the correct interpretation of the data Build a network of contacts in Europe for soil erosion Certain under-estimation of soil loss rates on behalf of PESERA The Sediment module in PESERA results in higher values in plains (Po in Italy and plain in Slovakia) and lower values in mountain areas. Different input datasets and Scale: Finer resolution in EIONET, old input layers in PESERA (CORINE CLC 1990, GTOPO30) Use EIONET data for future validation Which is the most used Model at National level? (R)USLE Panagos et al. (2014), Soil Science & Plant Nutrition

7 RUSLE2015: New soil erosion model A = K * R * C * LS * P Panagos et al (2015) In Review

8 Soil Erodiblity (K-factor) Soil Erodibility is an integrated annual value of the soil profile reaction to the process of soil detachment and transport by raindrops & surface flow Combines the influence of Texture, Organic carbon, soil structure, Permeability, coarse fragments and Stone cover 20,000 Land use/cover survey (LUCAS) samples with measured data Regression interpolation using Terrain features, Lat/Long, vegetation covariates Spatial Resolution: 500m Verified against 21 local, regional and national datasets from 13 countries Panagos et al (2014), Science of Tot. Env.

9 Soil Erodibility (K-Factor) incorporating Stone cover Stone cover effect: 15%; Important effect in Mediterrean 9/22

10 Rainfall erosivity and data collection Panagos, P., Ballabio, C., Borrelli, P.,...et 14 others Rainfall erosivity is the kinetic energy of rainfall (MJ mm ha -1 h -1 y -1 ) Combines the influence of precipitation duration, magnitude and intensity Participatory approach: Environmental & Meteorological Services from all Member States (Mar 2013 Jun 2014). 1,541 Precipitation stations with detailed rainfall intensity (all countries) Calibration requested: 5 min, 10-min, 15 min, 60 min. Temporal Resolution: 30-Minutes Time series: 7 56 Years (Mean: 17.1yr; 75% of time series in ) Data: 26,394 years of High Temporal resolution precipitation records REDES: Rainfall Erosivity Database on the European Scale from MS (2015). Science of Total Env.

11 Rainfall Erosivity (R-factor) Resolution: 1Km; Robust Geostatistical model Highest R-factor in Mediterranean & Alpine regions and lowest in Scandinavia R-factor not dependent only from precipitation Panagos et al.(2015)

12 Cover Management (C-factor) Differentiate between Arable lands & Non-Arable lands Non arable: Forest Shrub sparse vegetation Heterogeneous Permanent crops - pastures/grasslands Use of CORINE Land Cover classes Calibrate the C-factor from literature: 20 major published studies with Remote Sensing(RS) images from Copernicus Programme: Vegetation Density layer: RS every 10 days Example: Pastures C-factor Range from literature: Each pixel gets a value in this range depending on its Vegetation Density (0-100%) Pastures (mean) C-factor in Ireland: Pastures (mean) C-factor in Cyprus: Panagos et al.(2015). In Revision.

13 Topography (LS-factor) 25m DEM resolution 25m LSfactor (capture geomorphological features compared to 100m DEM) Desmet & Govers algorithm (1996) Fast process with SAGA software 50GB of dataset available in European Soil Data Centre (ESDAC) No arbitrary limitations in slope length Slope cutoff: 50% (after literature review & experimental results in Switzerland) Panagos et al.(2015), Geosciences, MDPI

14 Support Practices (P-factor) Data input from: - Good Agricultural Environmental Conditions (GAEC) plus - LUCAS 270,000 earth observations Support practices Impact: Contour farming (5%) Stone Walls (38%) Grass Buffers (57%) P-factor in EU-28: 0.97 P-factor in arable: 0.95 Panagos et al.(2015). Environmental Science and Policy.

15 EU Soil loss Map with RUSLE2015 Average EU-28: 2.46 t ha -1 yr -1 (in the erosive prone areas: 91% of EU) Total Soil loss: 970 Mt annually Spatial resolution: 100m Reference year: % of EU lands have rates >2 t/ha 5.2% of EU lands suffers from severe erosion (>10 t/ha c.a 1mm of soil) Panagos et al (2015) In Review

16 EU Soil loss Map with RUSLE2015 Panagos et al (2015) In Review

17 High resolution input layers: Research improvements (1) Before K-factor improvements: 5 textural classes vs. Soil properties measurements Large polygons (1960 s) vs. Soil samples (2009) Stoniness effect Verification with local/regional/national studies Now Before R-factor improvements: 50kmx50km vs 1kmx1km (2500 times finer Scale) Based on high temporal resolution measured data R-factor is not dependent only on precipitation Erosivity density trends in Europe Seasonal (and monthly) maps of R-factor Calibration rules in different temporal resolutions Now

18 High resolution input layers: Research improvements (2) Before C-factor improvements: 13 C-factor values (classes) vs. vegetation density Forests have double protection against erosion in Finland than Cyprus C-factor:0.20 for arable lands vs. crop composition Incorporation of management: Tillage practices, cover crop, plant residues Now Before? P-factor =1 P-factor improvements: First time estimated at European scale Incorporation of support practices: Contour farming, maintenance of stone walls, grass margins Scenarios for increasing soil protection Now

19 RUSLE2015 & Soil Loss Map: Concluding remarks Trend: Decrease of 7% (20% in arable lands) due to impact of Common Agriculture Policy (CAP) and soil protection measures: reduced tillage, plant residues, cover crop, contour farming, maintenance of stone walls, increase of Buffer strips. Very good correspondence with EIONET (7 out of 9 Member States): The European model is as robust as national ones. High resolution & best available input data in EU Transparent way & easily parameterization Peer-reviewed following literature Replicable & comparable with national estimates Participatory: involvement of countries (R-factor, K-factor, Statistics - EUROSTAT) Scenario analysis: Land cover change (Land Use Modelling Platform - LUMP) Management changes (Policies: CAP , Biofuels directive) Climate Change (precipitation & intensity trends in 2050) Panagos et al (2015) In Review

20 Soil Loss Map: Dynamic tool for policy makers and further research RUSLE2015 delineates hot spots that requests special protection measures Land use scenarios: LUMP predicts increase in forest 2.2% in the expense of seminatural areas 5.8% reduction in soil loss (contradictory trends of arable lands?) Climate change scenarios: IPCC HadGEM2 predicts 8.2% decrease of precipitation in Europe. Erosivity density & future precipitations R-factor decrease (unknown the trends in precipitation intensity?) Policy developments: Biofuels directive pushes for replacing cereals with energy crops: sugar beet, sunflowers (more erosive). i.e. 10% transformation increase 3.8% of soil loss in arable lands Common Agricultural Policy : Duplicate the grass margins and apply contour farming in arable lands > 5% Reduce soil erosion by 5%. Soil Organic Carbon detachment due to soil loss in agricultural lands: 14.8Mt per year (1% of topsoil SOC in 12 years) Outlook & future plans: phosphorus budget in agricultural lands sediment yields predictions from catchment area seasonal variability of the R- and C-factors at European scale. Panagos et al (2015) In Review

21 Information in peer review publications: 5 published 2 accepted Revisions done 1 in Review European Soil Data Centre: Panos.panagos@jrc.ec.europa.eu