Search for: All records

For the past sixteen years under the support of NSF, NASA and most recently the US Geological Survey, we have been systematically measuring electric and magnetic field time series from moving arrays of magnetotelluric (MT) instrumentation spanning the conterminous US and the interior of Alaska. While originally motivated by questions of the structure and evolution of the North American continent, the resulting 3-D electrical conductivity structure of the Earth's crust and upper mantle and the electromagnetic impedance data derived from this work have in recent years proved of considerable importance to mitigating risk to critical infrastructure (most notably, the power grid) from geomagnetically induced currents caused by space weather and electromagnetic pulse events. Under current NSF support we are exploring how to combine real-time magnetic observatory data streams with this information and with power flow simulations of the power grid to provide real-time alerting information of GIC impacts on high-voltage transformers to electric utilities. In the present work we go beyond real-time and present preliminary results of our efforts to train neural networks to assimilate data from dense arrays of ground-based MT stations in Alaska to provide forecasts of ground electric and magnetic field time series that could in future, with installation of permanent MT arrays, provide actionable intelligence to utilities ahead of GICs impacting their networks.