Subsurface site characterization using multiple types of data
Investigators: Lance Besaw, Donna Rizzo

Numerous optimization technologies have been developed over the past two decades. These have been used to suggest ways to operate pump-and-treat remediation systems at effectively reduced costs. Simultaneously, characterization based on innovative methods, such as recurrent artificial neural networks and Extended Kalman Filtering has been developed and deployed. In addition to mapping property values, the characterization methods detect and describe (spatial and temporal) correlations among aquifer properties, and are used in uncertainty assessments. The figure on the right represents a block of Berea Sandstone that was acquired by New England Research (NER) from the Cleveland Quarries in Amherst Ohio. NER has supplied measurement data of permeability, resistivity and compressional and shear velocities using a proprietary software package that has been developed to link the underlying pore structure of a rock to diverse core scale properties.