HR: 08:15h
AN: GP41A-02    [PDF]
TI: Toward Modelling Topsoil Magnetic Susceptibility for Demining Activities
AU: * Hannam, J A
EM: hannam@liverpool.ac.uk
AF: Environmental Magnetism Laboratory, Department of Geography Roxby Building University of Liverpool, Liverpool, L697ZT United Kingdom
AU: Dearing, J A
EM: j.dearing@liverpool.ac.uk
AF: Environmental Magnetism Laboratory, Department of Geography Roxby Building University of Liverpool, Liverpool, L697ZT United Kingdom
AB: The Landmine Monitor estimates that landmines cause up to 20,000 fatalities and casualties worldwide every year, in over 100 countries affected by landmine contamination. Although detection technologies have become more sophisticated, the metal detector still remains the most widely employed detection system in landmine affected regions. With increased use of minimum metal mines, the performance and sensitivity of metal detectors are increasingly challenged. In addition to mine constituents, depth of burial and orientation, soil properties significantly affect metal detection capabilities. Soils with high magnetic susceptibility, in particular those dominated by viscous components, interfere with the response signal in both frequency and time domain metal detection systems. Using Bosnia and Herzegovina (BiH) as a pilot region, we created an expert system to predict topsoil susceptibility from environmental information within a SOTER data base. Initially, the knowledge base is constructed from published relationships of environmental parameters and magnetic susceptibility and knowledge of experts in the field of soil magnetism. The knowledge base is underpinned by environmental conditions that are known to enhance or reduce magnetic susceptibility in topsoils. Where semi-quantitative data exists, transfer-functions are used to provide first approximations of susceptibility classes and offer a basis for a probability score for the susceptibility class. As a first approximation, susceptibility values are categorized into five continuous classes delimited by published magnetic susceptibility ranges in topsoils. The predicted susceptibility maps result in regional contrasts, delineated by the spatial scale of the environmental information. Further development of the model using a Baysean rule-based system with fuzzy boundaries is anticipated. Validation of the model is proposed using archived soil survey samples from BiH. In addition to providing essential data for effective planning of mine clearance in mine affected regions, prediction of topsoil magnetic susceptibility for demining missions provides a novel application of environmental magnetism.
DE: 1512 Environmental magnetism
SC: Geomagnetism and Paleomagnetism [GP]
MN: 2003 Fall Meeting