Stochastic Meshless Uncertainty Gridding (SMUG) is a program to interpolate unevenly sampled data onto any grid, while also calculating the associated uncertianty in the interpolation. SMUG uses a Taylor series expansion to estimate the interpolated values, with the required derivatives estimates using a meshless method (see https://github.com/jlr581/finite_difference). Uncertainty in the input data is fully propogated through the interpolation using ball arithmatic. An ensemble of interpolations is used to capture the uncertinaty associated with neighbouring point selection and the Taylor series expansion.
If using SMUG for scientific studies, please cite
Jong et al (submitted), Antarctic bed topography estimation using a Stochastic Meshless1 Uncertainty Gridding (SMUG) method, Computers & Geosciences
The source code and Makefiles are included in the src directory. Both OpenMP and MPI versions of the code are included. To build the executable simply go to either the /src/OpenMP or src/MPI directory (depending on which version you wish to build). Then simply type "make" and the "smug" executable will be built. Example usage is provided in the example directory.
The example directory contains a text example based on the Ricker wavelet function. Details on running this example are included in the README.txt file in the example directory.