GRID enabled data mining on the Irish National Seabed survey

The Geological Survey of Ireland (G.S.I) at present has over 4 terabytes of multi-beam data gathered over the last five years from the Irish National Seabed and this data-set is expected to exceed 10 terabytes upon completion. The main problem that arises from having so much data is how to extract accurate information from the data-set. Geological Interpretation is carried out by visual inspection of bathymetric patterns. Due to the size of this data set and the emergence of similar data sets, the extraction of knowledge from such a data sets by human observers has become infeasible. The focus has turned to using artificial intelligence and computational methods for assistance. The data-set has a number of security concerns associated with it ruling out most distributed computing paradigms. This has lead to the creation of a Virtual Organisation MarineGrid[1] in the Grid Ireland project[2]. A new method has been developed for statistical analysis of bathymetric information specifically for automated geological interpretation of rock types on the sea floor and feature exteraction from the sea floor.In this paper we present a brief synopsis of both classification algorithm and feature extraction algorithm and the results obtained from running on two clusters connected via the Globus Toolkit. Globus is the de facto standard middle-ware upon which grids can be built.