3D computational models for subsurface reservoir flow (simulation, data assimilation, and optimization)

The data set corresponds to the two examples of the paper "History matching production data and uncertainty assessment with an efficient TSVD parameterization algorithm".
Files in Example 1 contain 200 realizations of a subsurface reservoir of dimension 30x30x3 where each layer has a different covariance matrix. Each realizations include isotropic horizontal log-permeability field, vertical log-permeability field, and porosity field. The true permeability and porosity fields and simulation parameters (e.g., well locations and controls) are also included. The goal is to generate synthetic observed data from the true model, and then apply the algorithm to update/calibrate each of the realizations such that they match production data to within some tolerance.
Example 2 contains 1500 realizations of a 28x30x3 model. This example applies ensemble-based regularization to estimate the covariance matrix within the model calibration/estimation framework.