Fits a Bayesian hierarchical AR(1)-driven state-space model
relating annual global mean surface temperature (GMST) to a reduced
paleoclimate proxy (computed from cached or user-specified paleoclimate
multi-proxy network data), instrumental temperature observations, and
radiative forcings (greenhouse gases, volcanic aerosols, and solar
irradiance). Unobserved pre-instrumental temperatures are estimated
jointly with the model parameters using 'Stan' via the 'instantiate' and
'cmdstanr' packages. Includes tools to reduce a multi-proxy network to a
single representative proxy series (via principal component regression,
partial least squares, LASSO, or supervised PCR), estimate the parameters
in the reconstruction model, run k-fold cross-validation, provide
uncertainty quantification on the reconstructed temperatures, and
visualize reconstructions and posterior draws. The model and methods
follow Barboza et al. (2014)
R package implementing the BayGMST model of Bagwell et al. (in prep.): a Bayesian hierarchical AR(1) state-space model reconstructing global mean surface temperature (GMST) from a reduced paleoclimate proxy network, instrumental observations, and radiative forcings.
Status: restructured (branch CRAN) from a collection of top-level
analysis scripts into an installable R package aimed at eventual CRAN
submission. See NEWS.md for exactly what changed in that restructuring, and
CRAN-READINESS.md for the remaining steps before an actual submission
(this restructuring was done in an environment with no R installed, so
several files -- NAMESPACE, man/*.Rd -- are hand-authored placeholders
that must be regenerated and verified with real R tooling; see that file).
Fitting models requires cmdstanr, which is not on CRAN, and a CmdStan installation. Install both before BayGMST: BayGMST compiles its Stan models when it is installed, and skips that step if they are missing (proxy reduction and forcing transforms still work). If you add them later, reinstall BayGMST.
install.packages(
"cmdstanr",
repos = c("https://stan-dev.r-universe.dev", getOption("repos"))
)
cmdstanr::install_cmdstan()
Then install BayGMST itself:
install.packages("BayGMST") # once on CRAN
# remotes::install_github("paleopresto/BayGMST_R") # development version
library(BayGMST)
rp <- reduce_proxies(proxy_matrix, years, temp_calib, calib_years, method = "PCR")
forcing <- transform_forcings(G = co2, V = volcanic, S = solar)
fit <- fit_baygmst(
proxy = rp,
instrumental_T = instrumental_T,
forcing_G = forcing$G, forcing_V = forcing$V, forcing_S = forcing$S,
years = years
)
reconstruct(fit)
plot(fit) # reconstruction + posterior densities, as in the PReSto manuscript
summary(fit)
The proxy input does not have to come from reduce_proxies():
as_baygmst_proxy() accepts bare vectors, data.frames, time-by-ensemble
matrices, and composite objects from other packages (such as
compositeR, the intended primary
pathway once it is published), converting ages in yr BP and collapsing
ensembles as needed. See the "Using composites from other packages" section
of the vignette.
See vignette("baygmst-intro", package = "BayGMST") for a full worked
example using small bundled example datasets.
Note: inst/legacy-scripts/config.yml is the configuration file of the
original pre-package scripts, kept only for provenance; no package function
reads it.
The underlying model builds on the reduced-proxy Bayesian framework of
Barboza et al. (2014, 2019) and Wang (2020); see
vignette("baygmst-model", package = "BayGMST") for the full statistical
description (equations, priors, forcing data provenance, and the mapping
to the Stan code). For per-function documentation, see the package help
(?BayGMST); inst/legacy-scripts/ holds the original, pre-packaging
scripts this was built from.