The Method of Anchored Distributions (MAD) is a method for Bayesian inversion designed for inferring both local (e.g. point values) and global properties (e.g. mean and variogram parameters) of spatially heterogenous fields using multi-type and multi-scale data. Software implementations of MAD exist in C++ and C# to import data, execute an ensemble of forward model simulations, and perform basic post-processing of calculating likelihood and posterior distributions for a given application. This article describes the R package anchoredDistr that has been built to provide an R based environment for this method. In particular, anchoredDistr provides a range of post-processing capabilities for MAD software by taking advantage of the statistical capabilities and wide use of the R language. Two examples from stochastic hydrogeology are provided to highlight the features of the package for MAD applications in inferring anchored distributions of local parameters (e.g. point values of transmissivity) as well as global parameters (e.g. the mean of the spatial random function for hydraulic conductivity).
Supplementary materials are available in addition to this article. It can be downloaded at RJ-2017-034.zip
gstat, spBayes, spTimer, anchoredDistr, devtools, RSQLite, np, plyr, dplyr, ggplot2
Spatial, SpatioTemporal, Bayesian, Econometrics, Graphics, Phylogenetics, SocialSciences, TimeSeries
Text and figures are licensed under Creative Commons Attribution CC BY 4.0. The figures that have been reused from other sources don't fall under this license and can be recognized by a note in their caption: "Figure from ...".
For attribution, please cite this work as
Savoy, et al., "The R Journal: anchoredDistr: a Package for the Bayesian Inversion of Geostatistical Parameters with Multi-type and Multi-scale Data", The R Journal, 2017
BibTeX citation
@article{RJ-2017-034,
author = {Savoy, Heather and Heße, Falk and Rubin, Yoram},
title = {The R Journal: anchoredDistr: a Package for the Bayesian Inversion of Geostatistical Parameters with Multi-type and Multi-scale Data},
journal = {The R Journal},
year = {2017},
note = {https://doi.org/10.32614/RJ-2017-034},
doi = {10.32614/RJ-2017-034},
volume = {9},
issue = {2},
issn = {2073-4859},
pages = {6-17}
}