Extended Data Figure 7: Multivariable regression analysis to best fit the seismicity density, evaluating a subset of the models of Extended Data Fig. 4.
From: Western US intermountain seismicity caused by changes in upper mantle flow
The normalized contribution of each model vector, the importance wi , is indicated by filled and open symbols from a non-negative or regular least squares solution of equation (1), respectively. Results indicate that rate change of dynamic topography is the strongest driver of seismicity, with contributions from GPE, and modulated by Moho depth gradients and elastic thickness variations (compare with Extended Data Fig. 4).