| This dissertation addresses the problems of earthquake induced liquefaction. These problems of earthquake induced liquefaction can be broadly categorized as issues pertaining to pre- and post- liquefaction characterization. I believe that, for a comprehensive advancement of our ability to deal with the challenges of liquefaction, an improved understanding of both pre- and post-liquefaction characterization is critical. In specific, the improvements in pre-liquefaction characterization will better equip the geotechnical engineer to answer two critical questions: (1) given a likely seismic event, is the soil prone to liquefy; and (2) if liquefaction occurs, what consequences can be expected at the site in terms of ground movements or lateral displacements. Simultaneously, the advancements in post-liquefaction characterization will help in developing more complete datasets of occurrence/non-occurrence of liquefaction, which in turn will enable researchers to better validate and improve models developed for pre-liquefaction characterization.;The current models in pre-liquefaction characterization fall short to generate the necessary inputs to establish the interaction within and between the fields of pre- and post-liquefaction for its advancement. In addition, the current methods of field reconnaissance after an earthquake event do not provide a systematic approach to develop complete datasets of liquefaction effects. Therefore, this dissertation focuses on the application of pattern recognition/machine learning to improve and validate liquefaction models and to identify gaps in the data to strategize the scope for future data collection. Further, I also explore the potential to capitalize on the advancements in high spatial and temporal resolution satellite remote sensing in deriving surficial information for both pre- and post-liquefaction characterization. I hope that this work will provide the necessary first step to establish the interaction between pre- and post-liquefaction characterization. |
