Package index • PESTO

Core ensemble solvers

C++ kernel update equations (Chen-Oliver IES, Evensen MDA) and weighted objective.

ensemble_solution(): Ensemble Solution Kernel (GLM form)
ensemble_solution_mda(): Ensemble Solution Kernel (MDA / Evensen form)
ensemble_solution_adaptive(): Ensemble Solution with Adaptive SVD Backend
compute_phi(): Compute Phi (Objective Function) for Ensemble

SVD backends

Automatic backend selection; randomised SVD; direct LAPACK.

adaptive_svd(): Adaptive SVD with Automatic Backend Selection
accelerate_svd(): Hardware-Accelerated SVD via LAPACK
rsvd(): Randomised SVD (Halko-Martinsson-Tropp Algorithm)

High-level PEST++ wrappers

Convenience wrappers driving PEST++ executables (pestpp-ies, pestpp-glm, …) from R.

pesto_ies(): Run PEST++ IES (Iterative Ensemble Smoother)
pesto_glm(): Run PEST++ GLM (Gauss-Levenberg-Marquardt)
pesto_sweep(): Run PEST++ SWP (Parametric Sweep)
pesto_sensitivity(): Run PEST++ SEN (Global Sensitivity Analysis)
pesto_surrogate_ies(): Surrogate-Accelerated IES Iteration
pesto_version(): Get PESTO package version information
pestpp_available(): Is a PEST++ Executable Available?

Forward-model contract and multi-fidelity

Typed forward-model object both adapter modes honour, and the multi-fidelity (cheap / expensive) bridge with its control-variate combiner.

pesto_forward_model(): Forward-Model Contract (S7 class)
as_forward_model(): Coerce an object into a pesto_forward_model
pesto_evaluate(): Evaluate a PESTO forward model
pesto_multifidelity_model(): Multi-Fidelity Forward Model (S7 class)
mf_control_variate(): Affine control-variate bias correction across fidelities

ODE / compartmental forward-model templates

Ready-to-use differential-equation forward models – a generic ODE builder plus crop-growth and SEIR specialisations – that plug into the IES driver as typed forward-model objects.

ode_forward_model(): Forward Model from a System of Ordinary Differential Equations
crop_growth_forward_model(): Crop-Growth Forward Model (Logistic / Expolinear Biomass)
seir_forward_model(): SEIR Compartmental Forward Model

In-process IES via R callback

Driver and APSIM adapter for forward-model callable IES, bypassing the .pst file cycle.

pesto_ies_callback(): Run IES with an In-Process R Callback Forward Model
pesto_ies_filter(): Run a Sequential (Filter-Mode) Iterative Ensemble Smoother
print(<pesto_ies_result>) plot(<pesto_ies_result>): Print and plot methods for PESTO ensemble-smoother results
apsim_callback(): apsimx Forward-Model Adapter for PESTO IES

Inflation and localisation

Finite-ensemble pathology countermeasures – covariance inflation against under-dispersion, localisation against spurious correlations, and the spread-ESS collapse diagnostic.

pesto_inflation(): Covariance Inflation Specification for IES
pesto_localisation(): Covariance Localisation Specification for IES
ensemble_spread_ess(): Spectral Spread Effective Sample Size of a Parameter Ensemble
correlation_localisation(): Correlation-Based Automatic Localisation Taper
gaspari_cohn(): Gaspari-Cohn Localisation Taper
ensemble_solution_localised(): Localised Ensemble Solution Kernel (explicit-gain GLM form)

Reference IES (pure R)

Textbook Chen & Oliver (2013) reference implementation used as the canonical comparison target by the comparison vignette.

pesto_reference_ies(): Pure-R Reference IES Update — Chen & Oliver (2013) eq. 12

Surrogates

Gaussian Process and Random Fourier Feature surrogates; convergence-aware ensemble sizing.

train_gp_surrogate(): Train a Gaussian Process Surrogate
predict_gp_surrogate(): Predict with GP Surrogate (with Uncertainty)
train_gp_surrogate_tuned(): Train a GP Surrogate with Maximum-Likelihood (Anisotropic) Length Scales
predict_gp_surrogate_tuned(): Predict from an MLE-Tuned GP Surrogate
train_rff_surrogate(): Train a Sparse GP Surrogate via Random Fourier Features
predict_rff_surrogate(): Predict with RFF Sparse GP Surrogate
surrogate_ensemble_update(): Surrogate-Accelerated Ensemble Update
adaptive_ensemble_size(): Adaptive Ensemble Sizing
check_surrogate_regime(): Check whether a surrogate-IES regime is favourable

PEST control file I/O

Read/write .pst files and ensemble snapshots; build scenarios programmatically.

read_pst(): Read a PEST Control File (.pst)
write_pst(): Write a PEST Control File (.pst)
create_pest_scenario(): Create a PEST Scenario Programmatically
read_ensemble(): Read an Ensemble File
write_ensemble(): Write an Ensemble File
print(<pesto_pst>): Print method for pesto_pst objects

Ensemble manifest (portable run record)

Versioned, hashed, self-describing ensemble-run record for reproducibility and downstream consumption by any tool.

pesto_ensemble_manifest: PESTO Ensemble Manifest (S7 class)
pesto_obs_schema(): Build a grounded obs_schema descriptor for a manifest
as_manifest(): Convert a PESTO ensemble result into a pesto_ensemble_manifest
write_manifest(): Write a manifest to YAML + sidecar data files
read_manifest(): Read a manifest from YAML + sidecar data files
verify_manifest(): Verify the integrity of a manifest

Visualisation

Publication-grade plots for convergence, ensembles, identifiability, surrogate diagnostics.

plot_phi(): Plot Objective Function (Phi) Convergence
plot_ensemble(): Plot Ensemble Parameter Distributions
plot_identifiability(): Plot Parameter Identifiability
plot_surrogate_diagnostics(): Plot Surrogate Diagnostics

Deprecated

Retained for backward compatibility; scheduled for removal in a future release.

ensemble_solution_gpu(): Deprecated: ensemble solution with adaptive SVD backend