Melissa Clutter, Ty Ferré, and Fred Zhang1
Surface barriers are commonly installed to reduce downward water movement into contaminated zones. Specifically, evapotranspiration (ET) barriers are used to store and release water (via ET) prior to percolation to an underlying waste zone. To measure the effectiveness of the Prototype Hanford Barrier (PHB), several neutron probes access tubes were installed throughout the cover to monitor store-and-release mechanisms. Our aim was to use existing PHB data, simulated model data, and a dimensionality reduction approach called universal multiple linear regression (uMLR) to optimize the number of sensors in the PHB field site. We compare the network designs based on the downsampling of existing data to the recommended sensor design based on model simulated data, to understand the usefulness of implementing uMLR prior to sensor installation. We found that uMLR, combined with robust decision making (RDM), provided a simple, robust network design for monitoring total water stored at the PHB site.
1Pacific Northwest National Laboratory, Richland, WA