Conclusion

Here, we have tested a discrete cosine-tranform flattening method. It results in computational cost savings of approximately a factor of two to four in 2D and three to five in 3D as compared to the FFT flattening method. The memory savings are about a factor of two in 2D and four in 3D.

We intend to leverage the computational advantages of this DCT method by incorporating it either as an initial solution or a preconditioner into other flattening methods, possibly including flattening with geological constraints Lomask and Guitton (2006).