Boosting iterative stochastic ensemble method for nonlinear calibration of subsurface flow models

► A practical nonlinear parameter estimation algorithm is proposed. ► Gradients are approximated using an ensemble of directional derivatives within a Gauss-Newton iteration. ► The ensemble output covariance matrix is regularized using the iterative ℓ2 Boosting method. ► The number of regularization iterations is selected based on Akaike information criterion (AIC). ► The proposed algorithm is successfully applied to the parameter estimation of several subsurface flow models.