abstract
Geohazards in the offshore environmental are generally associated with large-scale potential failures such as large submarine landslides. In order to properly characterize the offshore deposits associated with such potential failures, techniques to integrate diverse data (as acquired in Research Themes 1 to 4) are needed. By applying our combined expertise in geomorphic interpretation, geotechnical characterization,geostatistics, and Geographic Information System (GIS), we will develop an integrated method for regionally characterizing offshore sediments. This integrated method will take into account the spatial variability of sediment properties in order to determine required/recommended sampling density and provide interpolated regional models of geotechnical properties. The regional models of geotechnical properties can then be displayed in conjunction with multibeam bathymetry and other geophysical data in a three-dimensional GIS. Because lithologic and engineering properties of sediments vary considerably laterally and with depth, extrapolation of properties from sparse datasets to undersampled locations is necessary for the evaluation of large scale geohazards. The geologic/geomorphic evaluation of seafloor morphology provides a basemap for subsequent extrapolations of sediment properties. Geostatistics provides a methodology to take advantage of the inherent spatial correlations of geologic data to estimate properties in undersampled areas and to quantify the uncertainty (Chiles and Delfiner 1999; Goovaerts 1997). Geostatistical techniques are derived from the assumption that geomaterials exhibit spatial patterns and that these spatial patterns, if known, can be used to improve predictions of geo-properties at unsampled locations. For the problem at hand, the subsurface structure and soil properties need to be estimated over a broad area in order to evaluate susceptibility to submarine slope failures. Geostatistical techniques will be used to quantify spatial correlation and to estimate geotechnical properties and strata interfaces. Luna and Frost (1998), Lee et al. (1999), Parsons and Frost (2000), and Dawson and Baise (2005) have developed similar integrated approaches using GIS, geostatistics, visualization software, and engineering computations to evaluate the liquefaction potential of a site, slope stability, and site investigation quality. The Dawson and Baise (2005) approach uses 3D GIS to define volumes of liquefiable soil. Using a 3D GIS framework linked with a relational database, we plan to assemble multiple data types into a single visual interface. The 3D GIS will therefore allow visualization of multiple datasets simultaneously. We will develop a set of GIS tools to visualize subsurface (geotechnical) and GIS layers (topography, multibeam bathymetry, etc.) in 3D. We are currently developing with the Computational Geometry Group at Tufts a customized 3D GIS that displays geotechnical point data with layer data (topography, street maps, etc). For the proposed project, we will further customize the 3D GIS to incorporate any additional datasets such as scanning and profile data generated in earlier tasks. The customized 3D GIS will have geostatistical analysis, statistical analysis, and spatial clustering capabilities. We will evaluate the spatial variability of offshore sediments. Geostatistics will be used to assess spatial correlation of offshore sediments. Clustering techniques will also be investigated to determine zones of similar materials as an alternative to geostatistical Kriging, which can break down when the spatial heterogeniety is high (Baise et al. 2005). Alternatively, clustering allows for short scale variability within regionally correlated points. As an outcome of the spatial variability evaluation, sampling scheme recommendations will be developed to insure that offshore investigations optimally sample the sediment properties. The geotechnical sampling scheme will be based on morphology characterization of the region (McAdoo et al. 2000) as well as geostatistical characterization of the geotechnical properties (relevant strength properties). This theme will also build on the extensive project data collection at ICG/NGI.
