I'm curious about the final section of the article which refers to anomalous neutron readings. Two obvious possibilities are that the failure mode is producing bad data or that there is water somewhere unexpected. Does anyone know more?