Pore pressure response to barometric pressure change in Champlain clay: Prediction of the clay elastic properties

• The loading efficiency, LE, is about 0.8–0.95 for the Champlain clay.
• Visual inspection, linear, and multiple regression provided roughly similar values for LE.
• The clay compressibility obtained with LE is smaller than that found with pulse and oedometer tests.
• The elastic parameters obtained with LE are for in situ very small strain.
• The hydraulic gradient in the clay layer fluctuated markedly during the 2-year monitoring period.

Champlain Sea clay deposits play a crucial role in limiting contaminant migration to underlying aquifers in the Saint Lawrence River Valley. Champlain Sea clays are well known as being prone to landslide and sample disturbance. Estimation of their elastic properties using their in situ loading efficiency would provide mechanical properties for the soil dynamic analyses and immediate settlement calculations encountered in geotechnical projects. Determination of loading efficiency can be envisioned as a very small strain test. In this paper we describe a detailed monitoring program for 10 vibrating wire piezometers installed within sealed boreholes on 5 study sites near Montreal in Quebec, Canada. The pore pressure response to barometric pressure variation was analyzed with three methods: linear regression, visual inspection, and multiple regression, which gave loading efficiencies between 0.7 and 0.95. These values were then used to estimate the elastic properties of the clay formation. Vertical compressibility and specific storage on the order of 1 × 10− 6 kPa− 1 and 1 × 10− 5 m− 1 were respectively obtained for most of the monitoring sites. A comparison of compressibility values derived from the in situ loading efficiency method, with those obtained with pressuremeter, oedometer, and pulse tests confirmed that clay compressibility is strain dependent. The pore pressures measured on the study sites have been corrected by applying the multiple regression and visual inspection methods. For most piezometers, both methods provided nearly identical results. However, for deep piezometers, the multiple regression method provided smoother pore pressure time series than the visual inspection method. The vertical hydraulic gradient obtained with corrected pressure data was shown to change significantly on some test sites during the monitoring period.