BayGMST: Bayesian Reconstruction of Global Mean Surface Temperature
Source:R/baygmst-package.R
BayGMST-package.RdFits a Bayesian hierarchical AR(1) state-space model relating global mean
surface temperature (GMST) to a reduced paleoclimate proxy network,
instrumental temperature observations, and radiative forcings
(greenhouse gases, volcanic aerosols, and solar irradiance). See
vignette("baygmst-intro", package = "BayGMST") for a worked
end-to-end example, and
vignette("baygmst-model", package = "BayGMST") for the full
statistical model description, which builds on the reduced-proxy
Bayesian framework of Barboza et al. (2014, 2019) and Wang (2020).
Model fitting requires CmdStan
fit_baygmst() and cv_baygmst() compile and sample a 'Stan' model via
the instantiate and cmdstanr packages, which in turn require a
working CmdStan installation on the user's machine. cmdstanr is not
on CRAN; install it from https://stan-dev.r-universe.dev, then run
cmdstanr::install_cmdstan(). BayGMST compiles its Stan models when
it is installed, and only if cmdstanr and CmdStan are available at
that moment, so install them first, or reinstall BayGMST after adding
them. Use instantiate::stan_cmdstan_exists() to check whether
CmdStan is available. reduce_proxies() and transform_forcings() do
not require CmdStan.
Provenance
This package was restructured from a collection of top-level analysis
scripts; see NEWS.md for exactly what changed (fixed bugs, dropped
dead code, and modeling decisions reviewed by the authors) and
system.file("legacy-scripts", package = "BayGMST") for the
original, unmodified scripts.
Author
Maintainer: Nick McKay nick@nau.edu
Authors:
Tyler Bagwell teb6@rice.edu
Julien Emile-Geay julieneg@usc.edu
Frederi Viens