in the context t of the Soil Thematic Strategy t Andreas Günther 1, Miet Van Den Eeckhaut 2, Paola Reichenbach h 3, Javier Hervás 2 & Jean-Philippe Malet 4 1 Federal Institute for Geosciences and Natural Resources (BGR) Hannover, Germany 2 Joint Research Centre (JRC), European Commission Ispra, Italy 3 National Research Council, Research Institute for Hydrogeological Protection (CNR-IRPI) Perugia, Italy 4 School and Observatory of Earth Sciences (CNRS) & University of Strasbourg (EOST) Strasbourg, France 1/ 20
Content Background: landslides and the EU Soil Thematic Strategy Attempt for a climato-physiographic landslide terrain delineation over the EU EU-level landslide location information EU-level landslide conditioning parameter data Parameter class analysis and example for a model calibration Work in progress and conclusions 2/ 20
EU Soil Thematic Strategy Objectives Soil protection against degradation Preservation of soil capacity to develop its functions: - Ecological l - Economic - Social Soil Threats Erosion Organic matter decline Contamination Salinisation Compaction Soil biodiversity loss Landslides - Cultural Flooding COMMUNICATION COM(2006) 231 on the Thematic Strategy for Soil Protection DIRECTIVE COM(2006) 232 establishing a framework for the protection of soil and amending Directive 2004/35/EC IMPACT ASSESSMENT SEC(2006) 620 of the Thematic Strategy for Soil Protection http://ec.europe.eu/environment/soil/index.htm 3/ 20
Schematic approach to address soil threats Applicable to erosion, SOM decline, salinisation, compaction, and landslides COMMON CRITERIA set in Directive Adoption of a coherent and comparable approach to delineate priority it areas using a RISK common methodology Model or empirical MONITORING Member States establish ACCEPTABILITY Consideration of known factors determining loss of soil quality For many soil threats, EU-level assessments do exist COM(2006)231 Risk Area IDENTIFICATION Member States establish a TARGET for Risk Area Member States adopt MEASURES to achieve target REPORT to Commission 4/ 20
Tiers in risk/priority area identification Adoption of nested ( tiered ) geographical approaches for spatial evaluation of several soil threats, including landslides (Soil Information Working Group of the European Soil Bureau Network) Tier Description Characteristics Tier 1 Risk area identification -Data must be available -Small scale (1:11 Mil.) -Qualitative approach, or Model approach combined with thresholds Tier 2 Detailed analysis within the risk zones -Larger scale -Any approach (or combinations) -Enhanced data need to allow model application Eckelmann et al., 2006 5/ 20
Tier 1 approach for landslides General specifications as formulated by the JRC landslide group Tier 1 to be carried out for the whole of EU Small scale ( ~ 1:Million), mapping unit: 1 km grid cell Use of a reduced, common set of data as conditioning factors already available for the whole of Europe: lithology, slope, and land cover Application of a heuristic indexing approach for landslide susceptibility mapping Calibration and validation using representative inventory data for the main geological, geomorphological and environmental scenarios in Europe Hervás et al., 2007 6/ 20
Climato-physiographic landslide terrains At the continental level, a terrain differentiation is needed Global susceptibility models do not reflect changes in landslide conditions determined by climato-physiographic setting throughout the continent Specific climato-physiographic i hi conditions result in different importance of the conditioning parameter classes to landslide susceptibility Different climato-physiographic conditions reveal different landslide controlling characteristics Even when considering generic landslides, coastal landslides must be treated separately 7/ 20
Climato-physiographic landslide terrains Climato-physiographic subdivision based on aggregation of climatic zones of the Soil Regions Map 1:5 M of BGR (http://www.geoshop-hannover.de/) Zone 0 Coastal areas (1km buffer) Zone 1 Subpolar and boreal oceanic climate Zone 2 Boreal continental climate Zone 4 Temperate oceanic climate Zone 5 Temperate continental climate Zone 7 Temperate and steppic arid climate Zone 8 Temperate and Mediterranean mountain climate Zone 9 Mediterranean climate 8 / 20
Climato-physiographic landslide terrains Elevation [m] Some basic zonal statistics 1800 1200 600 0 0 1 2 4 5 7 8 9 Z Zone 1 PGA [g] Zone 2 2 1 0 0 1 2 4 5 7 8 9 Mean precipitation [mm/month] 125 Zone 5 Zone 0 (coast) 100 75 50 Zone 4 Zone 8 25 0 1 2 4 5 7 8 9 Zone 7 Mean temperature [ C] 15 Zone 9 5-5 0 1 2 4 5 7 8 9 9 / 20
Landslide location information >100,000 000 landslide locations collected by 6/11 from various sources with different levels of detail Data sources National level inventories, detailed Regional level inventories, detailed National level compilations, sparse Map picked locations Googled locations 10 / 20
Landslide conditioning parameter data: slope Classified integer slope map derived from BGR EU27 DEM (resolution: 100 m, produced from SRTM and Russian topography) p Landslide % Slope classes 0 7 1-3 6 4-6 5 7-10 11-15 4 16-20 3 21-30 31-90 2 1 0 Total landslide frequency over slope integer 11 / 20
Landslide conditioning parameter data: lithology Dominant soil parental material classification based on aggregation of JRC SGDBE 1:1M STU s dominant parental material level 2 alluvium/colluvium calcearous rocks clayey materials crystalline rocks detrital formations glaciofluvial materials loamy/silty marl other/organic sandstone/flysch/molasse sandy materials schists volcanic rocks 12 / 20
Landslide conditioning parameter data: land cover Land cover map based on aggregation of ESA GlobCover global land cover dataset (resolution: 300 m) (http://ionia1.esrin.esa.int/) Artificial Bare Closed Forest Cropland Open Forest Pasture/Meadow Shrub 13 / 20
Heuristic model calibration Zone-specific allocation of the standardized parameter class weights and spatial combination for each calibration-zone using S n 3 j 11 w j x ij S = Grid cell (1km) susceptibility index w j = Global weight of parameter (0.58 for slope, 0.28 for lithology, 0.14 for land cover; Malet et al., 2010) x ij = weight of class i in parameter j Evaluation and reclassification of the zone-specific index grids into four susceptibility levels using ROC curves with TPR 0.5, 0.75, 0.9 Spatial combination of the reclassified susceptibility maps of the eight zones 14 / 20
Conclusions At the continental scale, terrain-differentiated susceptibility assessments for generic landslides are mandatory EU-wide datasets for terrain differentiation and the characterization ation of major landslide conditioning parameters are available A wide range of different landslide inventories do exist throughout Europe, where location information can be used to obtain a signal for parameter classification and weighting More control on heuristic parameter class weighting is needed to obtain more reliable estimates In a future stage, the inventory data should be exploited in such that individual susceptibility maps for different landslide types can be generated (e.g., for falls/topples and flows/slides) 15 / 20
Work in progress Alternative (and possibly more appropriate) climato-physiographic subdivisions are currently elaborated and tested Collection of landslide locations is not finished yet, need for enhanced data standardization Alternative lithology coverages will soon become available and will be tested SMCE will be used to improve heuristic modelling In representative areas with detailed locational information, multivariate statistical assessments (LR) will be employed for crossvalidation of the heuristic model Special attention will be given to the combination of the zone-specific susceptibility maps into one coherent pan-eu map 16 / 20