Estimation and Prediction for Remote Effects Spatial Process Models
Convenience function for stacking matrices into an array.
Reshape array of data matrices into long format
Make predictions using canonical correlation analysis (CCA)
Compute point estimates for parameters from posterior samples
Compute point estimates for parameters from posterior samples
Evaluate kron(A,B) * C without storing kron(A,B)
Performs an EOF decomposition of the data
Wrapper for a function to dump errors from C++
Extract region from a SpatialGridDataFrame
Basic extraction of SpatialGridDataFrame data for teleconnection analy...
Solves a triangular system with a Kronecker product structure
Compute Highest posterior density intervals from posterior samples
Samples an Inverse-Wishart matrix
Samples a multivariate normal with a Kronecker product covariance stru...
Formatting for longitude scales in ggplot spatial maps
Formatting for longitude scales in ggplot spatial maps
Matern covariance
Matern covariance
Compute effective range for Matern correlation to drop to a specified ...
Combine results from composition sampler
Combine sample covariance matrices from two samples
Combine sample means from two samples
Combine sample variances from two samples
Plot stData objects
Plot stFit objects
Plot stPredict objects
Plots teleconnection correlation maps
Simulate matrices from matrix normal distributions
Random wishart matrix
Basic evaluation of fit
Fit the remote effects spatial process (RESP) model
Compute log likelihood for model
Compute forecasts based on posterior samples
Simulate responses from the spatio-temporal teleconnection model
Computes variance inflation factors for fixed effects of the teleconne...
Summarize alphas
Summarize eof-mapped alphas
Plot stPredict objects
Fit a spatially varying coefficient model
Make predictions using a fitted varying coefficient model
Pointwise correlations for an exploratory teleconnection analysis
Tools for modeling teleconnections
Implementation of the remote effects spatial process (RESP) model for teleconnection. The RESP model is a geostatistical model that allows a spatially-referenced variable (like average precipitation) to be influenced by covariates defined on a remote domain (like sea surface temperatures). The RESP model is introduced in Hewitt et al. (2018) <doi:10.1002/env.2523>. Sample code for working with the RESP model is available at <https://jmhewitt.github.io/research/resp_example>. This material is based upon work supported by the National Science Foundation under grant number AGS 1419558. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.