Predicting Off‐Fault Deformation From Experimental Strike‐Slip Fault Images Using Convolutional Neural Networks

Abstract Crustal deformation occurs both as localized slip along faults and distributed off of faults. While there are few robust estimates off‐fault in nature, scaled physical experiments simulating crustal strike‐slip faulting allow direct measurement the ratio fault to regional deformation, quantified kinematic efficiency (KE). We offer an approach predict KE using a 2D convolutional neural network (CNN) trained directly on maps produced by experiments. Experiments with different loading rates basal boundary conditions generate throughout evolution Strain us calculate its uncertainty, utilized loss function performance metric. The CNN achieves 91% custom accuracy prediction unseen data set. Although model is experiments, it can that matches available geologic estimates.