A Hydraulic Tomography Experiment in Fractured Sedimentary Rocks, Newark Basin, New Jersey, USA

Abstract:

Hydraulic tomography was performed in July 2015 in contaminated fractured mudstone beds at the former Naval Air Warfare Center (NAWC) in the Newark Basin near Trenton, NJ using seven existing wells. The spatial arrangement of wells (in a circle of 9 m radius with one central well), the use of packers to divide the wells into multiple monitoring intervals, and the deployment of fiber optic pressure transducers enabled collection of a hydraulic tomography dataset comprising high-resolution drawdown observations at an unprecedented level of spatial detail for fractured rocks. The experiment involved 45-minute cross-hole aquifer tests, conducted by pumping from a given packer-isolated well interval and continuously monitoring drawdowns in all other well intervals. The collective set of drawdown data from all tests and intervals displays a wide range of behavior suggestive of highly heterogeneous hydraulic conductivity (K) within the tested volume, such as: drawdown curves for different well intervals crossing one another on drawdown-time plots; variable drawdown curve shapes, including linear segments on log-log plots; variable order and magnitude of time-lag and/or drawdown for intervals of a given well in response to pumping from similar fractures or stratigraphic units in different wells; and variable groupings of wells and intervals showing similar responses for different pumping tests. The observed behavior is consistent with previous testing at the NAWC indicating that K within and across individual mudstone beds can vary by orders of magnitude over scales of meters. Preliminary assessment of the drawdown data together with a rich set of geophysical logs suggests an initial conceptual model that includes densely distributed fractures of moderate K at the shallowest depths of the tested volume, connected high-K bedding-plane-parting fractures at intermediate depths, and sparse low-K fractures in the deeper rocks. Future work will involve tomographic inversion of the data to estimate the K distribution at a scale of ~1 m³ in the upper two-thirds of the investigated volume where observation density is greatest.