Prediction of water content and normalized evaporation from oil sands soft tailings surface using hyperspectral observations

Abstract

The paper addresses the challenge of measuring water content and evaporative fluxes from oil sands soft tailings surfaces using hyperspectral observations. Hyperspectral time-series laboratory observations were collected from four different mature fine tailings (MFT) samples displaying variations in swelling potential and bitumen concentration. The samples were allowed to evaporate from an initial state of water saturation to an air-dried state. From these data, several spectral features were evaluated to predict water content and normalized evaporation rate from the optically sensed portion of the tailings surface (less than a few hundred micrometres). For the samples tested, the best estimate of moisture content was achieved using the normalized soil moisture index (NSMI) index (coefficient of determination R2 = 0.97). The absolute reflectance at 1920 nm was found to be the best spectral estimator of normalized evaporation (R2 = 0.97), with the NSMI index also being valuable (R2 = 0.95). In both instances, the NSMI index may be of value for estimations attempted in the field. Remote estimation of moisture content and evaporation could help tailings managers assess the drying process to determine when the deposit has stopped drying at the surface and decide when the next lift should be deposited. In future efforts, the models obtained from this laboratory investigation will be assessed for their applicability in field settings and validated using concurrent sampling.