DETAILED CATALOG OF LANDSLIDES TRIGGERED BY THE 2008 WENCHUAN EARTHQUAKE AND STATISTICAL ANALYSES OF THEIR SPATIAL DISTRIBUTION

The latest research results indicate that the 2008 Wenchuan Earthquake(MS 8.0)triggered more than 197000 landslides. Base on Geographic Information Systems(GIS) and remote sensing technologies, three(nearly) complete landslide catalogs are constructed firstly. The three catalogs are corresponds to respective landslide polygons, landside centroid points, and landslide top points respectively. These catalogs indicate that there were 197481 landslides triggered by the Wenchuan earthquake throughout an area of about 110000km2.These landslides are delineated as respective individual solid polygons, landslide central dots, and landslide top dots using visual interpretation of high resolution aerial photographs and satellite images taken pre-and post-earthquake, verified by selected field checking. These landslides cover a total area of about 1160km2.An area covers about 44031km2 is selected as study area. There are 196007 landslides. They have a total area of 1150.6km2 occurred in the study area. This landslide catalog is the most detailed and complete landslide inventory related with the 2008 Wenchuan earthquake. It is also the largest landslide inventory record related to a single earthquake event. Subsequently, spatial distribution statistical analysis of landslides triggered by the 2008 Wenchuan earthquake are carried out in the study area. Base on the inventories of landslide polygons and landslide centroid points, two density maps, named landslide area density map and landslide point density map were constructed, respectively. Both the two landslide density maps show that landslide activities primarily concentrate along the Yingxiu-Beichuan co-seismic surface fault-rupture, and most of the landslides occurred on the hanging wall. The highest landslide density areas appear in the northwest direction areas, on the hanging wall, of a section of the Yingxiu-Beichuan fault(from the Yingxiu Town to the Beichuan County).Those high intensive landslide areas correspond to high deformation areas during the earthquake. So, it can be concluded that the earthquake triggered landslides is very strongly affected by co-seismic deformation. Furthermore, correlations of landslide occurrence with seismic parameters, geologic and topographic conditions are analyzed using the three landslide catalogs. Statistical analysis of these landslides spatial distribution using (a) landslide area percentage(LAP),defined as the percentage of the area affected by the landslides,(b) landslide centroid number density(LCND),defined as the number of landslide centroid per square kilometer, and (c) landslide top number density(LTND),defined as the number of landslide top point per square kilometer, is carried out to determine how the occurrence of landslides correlates with elevation, slope angle, slope aspect, slope position, slope curvature, lithology, distance from epicenter, seismic intensity, distance from the Yingxiu-Beichuan surface fault-rupture, peak ground acceleration(PGA).It is observed that the LAP,LCND, and LTND values have positive correlations with slope angle, PGA, and seismic intensity, and have negative correlations with distance from epicenter, and distance from the Yingxiu-Beichuan surface fault-rupture. More slope curvature close to zero, more difficult to landslide occurrence. The highest values of one LAP and two LND appear at range from 1200m to 3000m in elevation. The landslides have preferred orientations, dominated by the eastern, southeastern, and southern directions. The sandstone, siltstone(Z) and granitic rocks have more concentrated landslides. No apparent correlations were found between LAP,two LND values and slope position. Statistical results also show that differences between LCND and LTND is very small for earthquake and geologic parameters, rather more difference relate with topographic parameters. Some qualitative comparisons between the results in this paper and others published papers are carried out finally. There would be obvious differences of statistical results related to correlations of landsides and elevation, lithology, distance from the epicenter, distance from the Yingxiu-Beichuan fault from the present results and in other publications. In addition to those parameters, the statistical result from other factors including slope angle and slope aspect show similar tendency. So, it can be concluded that a complete, detailed and accurate landslide catalog is an essential part for seismic landslide hazard analyses and spatial distribution statistical analyses. Otherwise, parts of results obtained from analyzing uncompleted landslide distribution data would be controversial. It is rather necessary and important to construct complete landslide catalogs for seismic landslide spatial distribution statistical quantitative analyses, hazard assessment, and landscape evolution controlled by earthquake triggered landslides in earthquake struck area, et al.