TY - CPAPER AB - Field observations from thirteen recent case histories from the UK and France are described which are used to assess an approach for predicting ground movements and soil-structure interaction. High quality site investigations, including very careful sampling and advanced testing are seen to be crucial to success. Providing there are adequate geotechnical data, the non-linear FE predictive approach advocated by the Authors is shown to be sufficient for most engineering purposes, considering typical project time scales. Factors that merit particular attention include: the geological model; sampling and testing quality; anisotropy and pressure dependency; a fully non-linear modelling approach; addressing behaviour from very small to large strains; the 'structure' of natural and artificial geomaterials; the possibility of brittle failure; three dimensional geometry and construction details; permeability and groundwater variations. It is shown, however, that the present approach does not capture some important long term effects of time, creep and ageing. Comments are also made on the use of such advanced techniques in combination with an 'observational approach', and on pos-sible avenues for further research and development. © 2005 Taylor & Francis Group. AU - Jardine,RJ AU - Standing,JR AU - Kovacevic,N EP - 245 PY - 2005/// SP - 201 TI - Lessons learned from full scale observations and the practical application of advanced testing and modelling ER -